Tissue Engineering

Preparation Copolymers of Collager&lycosaminoglycan for Tissue Regeneration Lila J. Chamberlain and loannis V.Yannas ...

Author: Jeffrey R. Morgan | Martin L. Yarmush

45 downloads 3312 Views 41MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Preparation Copolymers

of Collager&lycosaminoglycan for Tissue Regeneration

Lila J. Chamberlain

and loannis V.Yannas

1. Introduction Certain analogs of the extracellular matrix (ECM) have been shown to possess surprising morphogenetlc activity during healing of lessons m various anatomlcal sites This chapter describes methods for synthesis of the two ECM analogs that have been studied most extensively. The reader 1s referred to descriptions of these methods m the original literature (Z-3) The blologlcal actlvlty of ECM analogs has been revlewed elsewhere (4). One of these analogs, referred to as the skm regeneration template (SRT), has Induced regeneration of dermis in full-thickness skin wounds m the guinea pig model (2,.5-7), the porcme model (8), and m humans (S,!Wl). Smce it IS well known that the dermis of the adult mammal does not regenerate spontaneously (12,13), the SRT 1s required for dermal regeneration in all commonly encountered skm wounds that are sufficiently deep to have compromised the dermis. The SRT 1s currently used as a dermal regeneration treatment for patients who have sustained deep burns or deep mechanical trauma, mcludmg trauma from elective surgery, and who would otherwise have been treated with autografts (10) In the clinical settmg or m animal models, the SRT 1s apphed on wounds as a bllayer graft; the proximal layer 1s the highly porous ECM analog and the distal layer 1s a slhcone film (Fig. 1). The latter has no blologlcal actlvlty, but serves as a temporary dressing that protects the proximal layer from dehydration and bacterial invasion, and also converts the bdayer mto a mechamcally competent sheet, capable of being handled conveniently and sutured on the patlent’s tissues Another ECM analog, referred to as the nerve regeneration template (NRT), has induced regeneration of a functional peripheral nerve across a 15-mm gap From

Methods m Molecular E&ted by J R Morgan

Me&me, Vol 18 Tmsue Engmeenng Methods and and M L Yarmush 0 Humana Press Inc , Totowa,

3

Protocols NJ

Chamberlain and Yannas

4

Silicf

18 Tissue

Matrix

Fig. 1. SRT as it would be implanted in the skin wound model

Nerve

Stump

Collagen-GAG

Matrix

Fig. 2. NRT, ensheathed by an implantation tube, as it would be implanted in the peripheral nerve wound model. in the rat sciatic nerve model (14). In this model, the highly porous ECM analog is used to till the lumen of a tube, made either of silicone (nondegradable) or collagen (biodegradable); the nerve stumps are inserted into the tube and are prevented from being displaced further by two sutures at each stump (Fig. 2). If the silicone tube is used without an ECM analog, the maximum gap distance that can be bridged by a functional peripheral nerve in this animal model is only 10 mm (15). The structure of the NRT has been determined by selection of the network structure that resulted in maximum regenerative activity in the sciatic nerve model, using a gap of 10 mm (3,16). The selection of the collagen tube instead of the silicone tube, based on superior regenerative activity, has also been made in the same animal model (17). Another ECM analog, similar in structure to the SRT and the NRT, has been shown capable of regenerating the canine knee meniscus (18). The structure of these biologically active ECM analogs has been characterized on the scale of the nanometer, as well as on the scale of the micrometer. In the former scale, both ECM analogs referred to above are graft copolymers of type I collagen and chondroitin 6-sulfate, which are crosslinked covalently, and can therefore be described as insoluble macromolecular networks. On the larger scale, the analogs are highly porous matrices that are characterized in terms of the pore volume fraction, the average pore diameter and the average orientation of pore channel axes. Being insoluble, the ECM analogs cannot be isolated and characterized structurally using common biochemical techniques

CoNagen-Glycosaminoglycan Table I Structural

Properties

Copolymers

of Two Regeneration

5 Templates

Design parameter of ECM analog

Skm regeneration template (SRT)

Nerve regeneration template (NRT)

Type I collagen/chondromn 6-sulfate, (w/w) Degradation half-life, wk Average pore diameter, pm Pore channel orientation

9812 1.5 2(1-l 25 Random

98/2 6-8 5-l 0 Axral

for the structural analysts of protems. However, structural methodology that has been used to characterize synthetic polymeric networks, including infrared and Raman spectroscopy, rubber elasticity analysis of network structure and

various forms of microscopy, have been employed m the charactertzatton of the ECM analogs . The structures of the SRT (2) and the NRT (3,16) (Table 1) have been identified by selecting the analogs of maximum activity from a large number of ECM analogs with related structure Inspection of Table 1 shows that the NRT is significantly different from the SRT regarding network cross lmk density, average pore diameter, and average ortentation of pore channel axes (Fig. 3). 2. Materials 2.1. Preparation

of Collagen-GAG

Suspension

Type I collagen, from bovine tendon, (Integra LifeSciences, Plamsboro, NJ), m the form of hydrated fibrillar granules, is divided mto 14-g aliquots and stored at 0°C Freeze-thaw cycles during storage should be avoided. If dry collagen IS used, it should be kept refrigerated at 4°C Cooled overhead blender (Granco overhead blender, Granco, Kansas City, MO), mcludmg a coolmg system (Brmkman cooler model RC-2T, Brmkman, Westbury, NY) The blender is used to mix the collagen-glycosammoglycan (GAG) suspension, which must be kept at 4°C during the entire preparation. 0 05 Macetic acid solution Add 8 7 mL glacial acetic acid (Mallmckrodt Chemtcal, Paris, KY) to 3 L dH,O This solution has a shelf life of approx 1 wk Peristaltic pump (Manostat Cassette Pump, cat no 75-500-0.00, Manostat, New York) 0 11% w/v chondroitm 6-sulfate solution Dissolve 275 mg chondromn 6-s& fate (from shark cartilage, cat no C-4384, Sigma, St Louis, MO) m 250 mL 0.05 M acetic acid solution. The chondroitm 6-sulfate solution is stored at 4°C and has a shelf life of 1 d The chondroitm 6-sulfate powder is stored m a desiccator at 4°C


Fig. 3. Environmental scanning electron micrographs of the (A) skin regeneration template (scale bar = 100 pm) and the (B) nerve regeneration template (scale bar = 20 pm).

2.2. Formation of Matrix Pore Structure 2.2.1. Skin Regeneration Template 1. Freeze dryer (VirTis Genesis, VirTis, Gardiner, NY). Required to freeze the suspension and to sublimate the ice crystals, leaving behind a highly porous matrix structure. The freeze dryer is equipped with trays that are pressed against the chamber shelves when placed in the freeze dryer. These trays ensure good contact between the cooled shelf and the product, and are important for proper pore formation in the skin regeneration template.

2.2.2. Nerve Regeneration

Template

1. Polyvinylchloride (PVC) tubing (0.125 in. id, 0.25 in. od), cut into 12-cm lengths. 2. Silicone processing tubes (model 602-235 medical grade Silastic, 0.058 in. id, 0.077 in. od, Dow-Corning, Midland, MI) cut into 15-cm lengths. 3. Silicone adhesive (Medical GradeSilastic, Dow-Corning, MI). 4. Liquid nitrogen: 160-L canister. 5. Axial freezing bath: This custom-made device (Fig. 4) is required to freeze the suspension for a nerve regeneration template (19). To achieve the appropriate

Collagen-

Glycosaminoglycan

Copolymers

Fig. 4. Schematic of the custom-made axial freezing bath (19). Not to scale. (A) Electric timing motor. (B) Gear drive chain. (C) Drive gear. (D) Idler pulley. (E) String. (F) Guide rod. (G) PNS Graft tube carrier. (H) PNS Graft tubes. (I) Counterweight. (J) Thermometer. (K) Thermocouple. (L) Temperature controller. (M) Insulate glass vessel. (N) Silicone heat transfer fluid (0) LN2 Solenoid valve. (P) N2 Throttle valve. (Q) One-quarter inch copper cooling (R) Circulator.

pore structure, the suspension is injected into tubes and lowered into a freezing bath. The freezing apparatus consists of a liquid nitrogen-controlled cooling system and a gear train arrangement, which allows for variable lowering velocities. The cooling system uses liquid nitrogen, traveling through coiled copper tubing, to cool the heat transfer fluid inside the bath (Silicone Oil, Syltherm XLT Heat Transfer Liquid, Dow Corning, MI). A simple temperature controller is used to regulate the flow of liquid nitrogen. The freezing bath is insulated with hard Styrofoam and capped with an acrylic disk. 6. Freeze dryer (VirTis Genesis). Required to sublimate the ice crystals, leaving behind a highly porous matrix structure.

2.3. Cross/inking, 2.3.1.

Sterilization,

Skin Regeneration

and Hydration

Template

1. Vacuum oven (Fisher Isotemp Vacuum Oven, Fisher Scientific, Boston, MA; VacTorr 150 Vacuum Pump, GCAlPrecision Scientific, Chicago, IL).


8 2 3

4 5

6

7 8

9 10

Slhcone adhesive (SIlastic, Dow-Cornmg, MI), sterilize by autoclavmg Stenle implements 5-L plastic tub (approx W 11 x L14 x D4 m ) with Teflon cover (does not need to seal, only cover the tub), gauze, Teflon working surface, forceps, metal spatulas, rulers, scalpel blade holder, and scalpel blades Sterilize by autoclaving Lammar flow bench (Rehalab, Tenney Engineering, Umon, NJ) All sterile procedures are performed m the lammar flow bench 0 05 M acetlc acid solution Add 2 9 mL glacial acetic acid (Mallmckrodt) to 1000 mL dH,O Sterilize by filtration using a 0 2-p filter (cat. no 83 10, Costar Scientific, Cambridge, MA) This solution has a shelf life of approx 1 wk 0 25% glutaraldehyde m 0 05 M acetic acid* Combme 10 mL of 25% glutaraldehyde and 3 mL glacial acetic acid Add distilled water to 100 mL Add an addltlonal900 mL of dH,O This solution has a shelf hfe of about 1 wk, and 1s stored m a dark contamer at room temperature Sterilize by filtration using a 0 2-pm filter 4000 mL dH,O: Sterlhze by filtration usmg a 0 2-pm filter Teflon cutting template. Make a matrix-cutting template by cuttmg a piece of Teflon the size and shape of the desired matrix sheet. Using thrs type of template the matrix 1s cut without tearing, and is ensured the proper size matrix sheet Sterilize template by autoclavmg Phosphate buffered salme (PBS) (cat no. P-3813, Sigma), 1000 mL Sterlhze by filtration using a 0.2~pm filter 70% lsopropanol m dH,O, 1000 mL Sterdize by filtration using a 0 2-w filter

2.3.2. Nerve Regeneration

Template

1 Implantation tubes (see Note 1). For implantation, the nerve regeneration template 1s ensheathed by an lmplantatlon tube (Fig. 2) Tubes that can be used include porous collagen tubes (1.5 mm Id, 3 0 mm od, Integra), nonporous collagen tubes (1.5 mm id, 1 8 mm od, Integra), and silicone tubes (model 602-235 medical grade SIlastic, 0 058 m id, 0 077 m od, Dow-Corning) 2 Vacuum oven (Fisher Isotemp Vacuum Oven, Fisher Sclentlfic, VacTorr 150 Vacuum Pump, GCA/Preclslon Scientific) 3 Sterile implements several pair of forceps, scalpel blade holder, scalpel blades, ruler, specimen Jars, and a Teflon working surface Sterilize by autoclavmg 4 PBS (cat no P-38 13, Sigma), 1000 mL Sterilize by filtration using a 0 2-pm filter 5 70% lsopropanol m dHzO, 1000 mL Sterdize by filtration using a 0 2-p filter

3. Methods 3.1. Preparation of Collagen4A G Suspension The techmque for preparing the collagen-GAG suspension 1sidentical for the SRT and NRT It 1simportant that the collagen and GAG components remam refrigerated; therefore, the entire suspension preparation must take place at 4OC 1 Defrost a 14-g ahquot of frozen hydrated tendon collagen for 3&60 mm at room

temperature

Collagen-Glycosam~noglycan

Copolymers

9

2 Turn on coolmg system for blender and cool to 4°C (takes about 30 mm) 3 Add 13 69 g of defrosted hydrated tendon collagen (or 3 6 g of dry collagen), all at once, to 600 mL of 0.05 Macettc acid in one blender, and blend at high speed setting (approx 20,000 RPM) for 90 mm (see Note 2) 4 Calrbrate the peristaltic pump to 40 mL/5 mm 5 Add 120 mL of 0 11% w/v chondrottin 6-sulfate solution dropwtse to the blendmg collagen dtsperston over 15 mm, using the pertstaltrc pump (mamtam blender at 4°C and htgh-speed setting) 6 Blend the mixture for an addmonal90 mm on high-speed settmg (approx 20,000 RPM) 7 Pour out the collagenGAG suspenston and store m a capped bottle at 4°C The suspension has a shelf hfe of about 4 mo (see Note 3) If stored more than 4 wk, reblend for 15 mm at low speed (approx 10,000 RPM), in cooled blender (4”C), before usmg

3.2. Formation of the Matrix Pore Structure 3 2 1 Skrn Regeneration Template 1 Remove the air from the collagen-GAG suspenston by placmg it mto a 1500-mL Erlenmeyer flask under vacuum for 10 mm wtth agttatton, or unttl bubbles are no longer vistble 2 Set the shelf temperature of the freeze-dryer to -45°C 3 Turn on the condenser of the freeze-dryer. 4 Allow at least 1 h for the shelf temperature to reach -45°C 5. Pour the collagen-GAG suspenston mto an alummum VirTts freeze-dryer tray The depth of the suspension can be varied to change the thickness of the resultmg dry matrix 6 Place the suspenston-filled tray on the freeze-dryer shelf, and close the chamber door Be sure that the tray and the shelf are m good contact 7 Wait for approx 1 h (or longer as necessary) until the collagenGAG suspenston 1s frozen (see Note 4) 8 Check the condenser temperature It must be at -50°C or below before proceedmg to the next step 9 After the suspension 1sfrozen, turn on the freeze-dryer vacuum pump Make sure the chamber door makes a good seal 10 Once the vacuum IS below 200 mtorr, increase the shelf temperature to 0°C 11 Leave overmght (at least 15 h). 12 Increase the shelf temperature to 20°C 13. When the chamber reaches 20°C turn off the vacuum pump and condenser Release the vacuum m the chamber and remove the dry collagen-GAG mati~x in the form of a white, highly porous sheet

3.2 2. Nerve Regenerabon

Template

1 Prepare vented PVC Jackets by heating 12-cm sections of flexible PVC tubing at 105°C for 2 h, to straighten Puncture each tube with a 25 gage needle at 90-degree Intervals around the tube, spaced 1 cm apart for the length of the tube


10

...... ....... .......... ......... 9

Collagen-GAG Suspens10n

\

Conical

Plug

Pressurized, Processing

Conical

Silicone Tube

Plug

/

Fig. 5. Preparation of the PVC jacket assemblies for freezing the NRT. (A) Diagram of PVC assembly during injection of the collagenGAG suspension. (B) Final PVC jacket assembly, ready for freezing. Not to scale.

2. Flush silicone processing tubes (15 cm in length) with dHzO, and let dry. 3. Seal one end of each silicone processing tube with silicone adhesive. Inject a cylindrical plug of adhesive, approx 5 mm in length, into the end of silicone tube and allow the excess to stay on the outside of the tube. The excess is important for adhesion and can be cut off later. Let cure for 24 h at room temperature to a tack-free, elastomeric state. 4. Prepare for use a 160-L liquid nitrogen tank for the bath cooling system. 5. Remove the air from the collagenGAG suspension by placing into a 1500-mL Erlenmeyer flask under vacuum for 10 min with agitation, or until bubbles are no longer visible. 6. Turn on the cooling system of the axial freezing bath and set the bath temperature to -80°C (see Note 5). It will take approx 45 min of liquid nitrogen cooling for the bath to reach this temperature. 7. Insert each plugged silicone processing tube into a prepared PVC jacket. 8. Draw collagen-GAG suspension into a IO-cc syringe (Becton Dickinson model 5604, Becton Dickinson, Rutherford, NJ) and expel all the air bubbles. Attach a 25-gage needle (Becton Dickinson model 25G518, Becton Dickinson) to the syringe and insert the needle carefully into the plugged end of the silicone tube. The needle should be inserted far enough so that a needle length of about 3-5 mm extends beyond the Silastic plug into the tube (Fig. 5A). 9. Inject collagen-GAG suspension until the tube is full and no air remains in the tube. Pinch the free end of the silicone processing tube against the wall of the PVC jacket using a conical, plastic plug (the end of a pipet tip works well). Insert the plug far enough so that the silicone processing tube is sealed, and no suspen-

Collagen-Glycosaminoglycan

Copolymers

11

Conical Plug $ ?

i%

PVC Jacket

i-

Liquid CollagenGAG Suspension

ilicone Processing Tube

iilicone ‘IProceswe Freezing Front Frozen Collage;v GAG Suspension

\

Tray-Tube Interface

Frozen CollagenGAG Suspension

Freezing Bath Heat Transfer Fhud

Fig. 6. Manufacturing assembly diagram for the peripheral nerve regeneration template. (A) Freezing orientation. (B) Freeze-drying (sublimation) orientation. Not to scale.

sion can leak out. Insert another conical, plastic plug into the needle end of the tube. The plug at the needle end should not block the flow of the suspension into the silicone tube via the needle (Fig. 5A). 10. Inject additional suspension until the silicone processing tube becomes pressurized and expands to fill the entire PVC jacket (Fig. 5B). The silicone tube will inflate because of pressure from the injection of additional suspension. The end of the needle should be inside the PVC jacket to help prevent pressure build up at the needle tip. When the silicone tube has completely filled the PVC jacket, carefully remove the needle; simultaneously, press the conical plug into the end of the tube until the silicone processing tube is pinched against the PVC jacket and sealed. Pressure should be kept on the syringe plunger until the needle is completely out of the tube. Check to make sure the silicone processing tube is still filling the entire PVC jacket (Fig. 5B). 11. Attach the drive gear to the electric timing motor on the axial freezing apparatus (Fig. 4). Place prepared PVC jackets, up to four at a time, on the tube carrier (Fig. 4). Place the tube carrier on the gear train and manually lower until the bottom of the PVC jacket assembly is just touching the freezing bath. Start the motor and let the tubes lower into the bath at a velocity of lo4 m/s (see Note 6; Fig. 6A).

12

12 13 14

15

16 17 18

Chamber/a/n and Yannas Monitor the process of lowermg to ensure that the tubes do not stick to the copper tubmg m the freezing bath Turn on the freeze dryer and set the shelf temperature to -20°C Turn on the condenser of the freeze dryer. When the PVCJackets are fully immersed m the freezmg bath, turn off the ttmmg motor and remove the tubes from the bath. Qmckly separate the tubes and remove the comcal plugs Cut off the plugged end of the srhcone tube and cut each PVC Jacket assembly approxtmately m half with a sharp razor blade Thts process provides more exposed surface for subhmatton of the me crystals Lay the PVC Jacket assemblies on a freeze dryer tray and place the tray m the -20°C freeze dryer (Fig. 6B) This step must be done as quickly as possible (wtthm a mmute) to ensure that the tubes stay completely frozen Seal the chambers on the freeze dryer and close the vacuum outlet tube Check to be sure the condenser temperature IS below -45°C (if not, watt for the condenser temperature to reach 45°C before proceeding to the next step) Turn on the vacuum pump and wart for the vacuum to reach 200 mtorr Make sure that the chamber door IS sealed Once the vacuum reaches 200 mtorr, increase the shelf temperature to 0°C Leave the PVC Jacket assemblies m the freeze dryer for 17 h at this temperature and pressure Increase the temperature to 25°C then turn off the vacuum pump and the condenser Release the vacuum and remove the PVC Jacket assembhes, whtch contam the dry, whrte, hrghly porous matrix msrde the &cone processmg tubes

3.3. Cross/inking, Sterilization, and Hydration 3 3 1 Sk/n Regeneration Template 1 After removmg the dry collagenGAG matrix from the freeze-dryer, inspect the matrix for any uregularmes, using a scalpel blade, remove any regtons that appear to be dtstmctly different m appearance from that expected of a very htghly porous solid of uniform thtckness Usually these regions ~111 be located near the pan edges Take note of the dtfference between the pan side (the side that was m contact with the horrzontal pan surface) and the an stde (the stde that was m contact wtth the environment) of the dry matrix The pan stde has a much smoother surface Future steps ~111 requrre dtstmguishmg between the pan and an stdes of the dry matrtx. 2 Make an alummum foil pouch large enough to tit the sheet of dry matrix Take a large piece of foil and fold it m half The folded edge 1s now the bottom of the pouch. Take the left edges and fold, at least twice, to form a sealed side. Repeat on the right side of the pouch Insert the dry matrix mto the pouch (one sheet of matrix per pouch) and leave the top open 3. Place the matrix-filled pouch (top open) m the vacuum oven for dehydrothermal (DHT) treatment (see Note 7). The condtttons of treatment m the vacuum oven are’ 30 mtorr, lOS’C, 24 h 4 After 24 h, remove the pouch and tmmedtately seal the top by folding the top edges of the for1 pouch at least twice If the matrtx IS not bemg prepared for

Collagen-Glycosam~noglycan

5

6

7

8

9

10

11

12

Copolymers

13

immediate use, it can be stored m the foil pouch m a desiccator (up to 1 yr) until needed The matrix is now sterile and must be handled using sterile procedure from this pomt After the DHT process, or after storage, sterile silicone adhesive is placed on the dry matrix Prepare a sterile field m the lammar flow hood and place m It all sterile implements, mcludmg the sterile silicone adhesive and the dry matrix (dropped m the sterile area out of the foil pouch, the pouch can be discarded) Prepare operator for sterile work m gown, cap, coat, and sterile gloves Pour 1000 mL of 0.05 M sterile acetic acid solutton mto the sterile 5-L plastic tub Cover tub loosely with the stertle Teflon cover Place the matrix on the sterile Teflon working surface with the air side up (the silicone adhesive is placed on the an- side of the dry matrix) Squeeze a long bead of the vtscous s&cone fluid along one edge of the dry matrix sheet Holding the matrix with one finger on the edge, spread the silicone In a thick layer, usmg straight sweeps of the spatula toward the opposite edge of the sheet Use only one sweep of the spatula for each portion of the dry matrix Wipe the spatula clean with sterile gauze Another bead of silicone adhesive may be necessary m the center of the dry matrix to reach the opposite edge, if the matrix sheet is large Spread along the pore channels, if possible, so the matrix does not tear Next, use the spatula at a 90-degree angle to remove the excess silicone and create a thm layer of approximately constant thickness (approx 1 mm) Use single strokes to cover each area Exercise care, smce the dry matrix tears easily If tearing occurs, it will be necessary to discard the torn portion of the coated sheet Wipe the spatula often with the gauze to remove excess silicone adhesive Immerse the matrix, s&one side up, m the acetic acid solution for 15 s Flip the matrix over, silicone side down, and leave It for 20 h m the covered tub filled with acetic acid After 20 h in the acetic actd solution, the air bubbles must be removed from the hydrated matrix Prepare operator for sterile procedure (as m step 5) Remove the an m the hydrated matrix by holding one edge of the matrix (wearing sterile gloves) and gently rubbing the hydrated matrix with one finger until air bubbles come out. Be careful not to tear the hydrated matrix Turn the matrix over so that the silicone side faces up; allow it to rehydrate m the acetic acid for 4 h Remove any trapped an bubbles After the 4-h rehydration, remove the acetic acid from the tub wtth suction If the matrix IS allowed to dehydrate, its activity will be lost Withm 1 mm or less after removing the acetic acid, begin pouring 1000 mL of sterile 0 25% glutaraldehyde m 0 05 Macettc acid into the tub. Pour carefully, to avoid air bubbles Remove any air bubbles by gently rubbmg the foam (see step 9) Soak the matrix m the glutaraldehyde solution for 24 h m the covered tub at ambient temperature (20-22°C) (see Note 8) Remove the excess glutaraldehyde solution with suction Rinse the matrix 3 x by adding three lOOO-mL sterile water rmses to the tub, allowmg the matrix to soak for 10 mm m each rinse, then remove the rinse water with suction. Add a fourth

14


rinse of sterile water (1000 mL) to the tub, cover, and soak the matrix for 24 h Remove any an bubbles (see step 9) 13 After the 24-h period, the matrix must be cut and stored Remove the hydrated, porous matrix from the water and place It, sthcone side down, on the Teflon sheet Be careful not to let the matrix fold Cut away any Jagged edges with a scalpel Place the Teflon cutting template on the matrix and cut around tt gently with a scalpel Do not press down on the template while cutting Hold only the edges of the matrix This step must be completed wtthm less than 1 mm to prevent dehydration (which leads to deacttvatton of the matrix) 14 For mnnedtate use, place the cut matnx m a storage container tilled with stenle PBS The matrix can be stored m this medmm for 1 d For longer term storage, place the cut matrix mto a storage container with 70% sterile tsopropanol Store at 4°C m this medium up to 30 d Place the matrix m PBS for 12-24 h prior to use as an Implant

3 3 2 Nerve Regeneration

Template

Make an alummum foil pouch for each PVC Jacket assembly (containing the silicone processmg tube and the matrix) Take a large piece of foil and fold rt m half The folded edge 1s now the bottom of the pouch Take the left edges and fold, at least twice, to form a sealed side Repeat on the right side of the pouch Insert the PVC Jacket assembly mto the pouch (one tube per pouch) and leave the top open The PVC Jacket ts left in place to protect the matrix from damage durmg handling and storage Prepare fotl pouches (see step 1) for the implantation tubes Place each tmplantanon tube m a pouch, leaving the top open Place the matrix-filled pouches (top open) and the rmplantatton tube pouches (top open) m the vacuum oven for DHT treatment (seeNote 7) The condmons of treatment m the vacuum oven are 30 mtorr, 105°C 24 h After 24 h, remove each pouch and munedrately seal the top by folding the top edges of the for1 pouch, at least twice If the matrix 1s not being prepared for immediate use, tt can be stored m the foil pouch m a destccator (up to 1 y) until needed The matrix IS now sterile and must be handled using sterile procedure from this point After the DHT process 1s complete, the matrix can be cut and hydrated for use Prepare a sterile field and place m rt all sterile Implements, including the PVC Jacket assembhes and rmplantatron tubes Prepare the operator for sterile work m gown, cap, coat, and sterile gloves Under sterile condtttons, trim each tmplantatron tube to a length of 20 mm, using a scalpel Remove the matrix from the sthcone processmg tube by making a careful slit with the scalpel down the length of the srlrcone tube and gently pulling out the matrix with forceps Discard the sthcone processmg tube Trim off any crushed or otherwtse damaged pieces of the dry matrrx Cut the remaining portron of the dry matrix into IO-mm segments The exact length depends on the experimental design for use of the implant

Collageffilycosam~noglycan

Copolymers

15

9 Insert each IO-mm segment of matrix mto the center of a trimmed implantation tube (Fig. 2) 10 Place each Implant mto a specimen Jar filled with sterile PBS for hydration and short term (less than 2 d) storage If implants ~111not be used immediately, store at 4°C m 70% lsopropanol for up to 30 d Transfer the Implants to sterile PBS solution 1 d prior to implantation

4. Notes Using a degradable, collagen lmplantatlon tube yields a superior regenerated nerve than when usmg a nondegradable, slllcone tube for regeneration across a lo-mm gap m the rat sciatic nerve This finding 1s based on hlstologlcal data at 30 wk (17) The blending time IS critical because this 1s the step that induces swellmg of the collagen flbrlls and conversion of about 90% of banded collagen fibers to unbanded structures (20) The blending procedure does not, however, remove the triple helical structure of collagen (2) Banded collagen induces platelet agglegatlon, by removing the bandmg, the collagen fibrlls do not aggregate platelets (22) Blending for shorter times, or at slower speeds, does not eliminate the bandmg to the desired extent CollagenXiAG suspension that IS stored for over 4 mo may be contaminated with fungus It IS recommended to prepare the suspension shortly before matrices are to be manufactured to avoid possible contammatlon Freezing the collagen-GAG suspension on a cool, flat surface creates randomly orlented pore channels with approximately circular cross sections The average pore diameter can be mampulated by varying the shelf temperature For optlmum dermal regeneration the pores must be between 20 and 125 pm (2) Freezmg at -45°C typically results m pores with an average channel diameter of 70 + 30 pm (Fig. 3A) The temperature of the axial freezmg bath controls the average pore diameter In the case of NRTs, poles with an average dlametel of 5-10 pm were determined to be optimal for the regeneration of axons (3,16,22) Freezing at -80°C results m pore channels that average 5-10 pm m diameter (Fig. 3B) Higher freezing temperatures result m larger pore diameters (19) By slowly lowering the suspension-filled tubes mto the freezing bath, the pores form as axially aligned channels The degree of axial alignment can be modified by changing the lowering velocity For example, quenching the tubes m the freezing bath (V = 1 m/s) creates radial pores, while lowering the device into the bath much more slowly, say at lOA m/s, results m highly aligned axial pores (19,23), which were optimal for regeneration (3,26,22) DHT treatment in a vacuum oven at 105°C for 24 h serves as a method of crosslmkmg (24-25) and sterlltzmg (I) the prepared matrices Treatment at 105°C does not affect the triple-hehcal structure of the collagen, provided the moisture content at the begmnmg of DHT treatment IS below 10 wt%, which IS achieved by freeze-drying the matrlces (I)

16


8 Glutaraldehyde is used to covalently crosslmk the matrix m addition to crosslmkmg introduced by DHT treatment and may be omrtted rf a lower level of crosslmk density (leading to a more rapid degradation rate in VIVO) is desired (26) (I) For the SRT, a 24-h treatment m glutaraldehyde resulted m a degradation rate favorable for dermal regeneration The maximum allowable degradation rate is 140 enzyme units, as determmed by using an m vitro collagenase digestion assay (2) Glutaraldehyde treatment is not used for the NRT, since a rapidly degrading collagenGAG matrix was found to be optimal for pertphera1 nerve regeneration (3,16,22) Although the SRT hasa higher crosslmk density becauseof glutaraldehyde treatment, the degradation half-life of the SRT is much shorter than the NRT probably owmg to a stgmticantly lower collagenalytic activity m a nerve lesion compared to a skm lesion (Table 1)

References 1 Yannas, I V , Burke, J F , Gordon, P L , Huang, C , andRubenstem,R H (1980) Design of an artifictal skm II Control of chemical composmon J Bzomed Muter Res 14, 107-131 2 Yannas, I V , Lee, E , Orgtll, D P , Skrabut, E M , and Murphy, G F (1989) Synthesis and characterization of a model extracellular matrix that induces partial regeneration of adult mammalian skm Proc Nat1 Acad Scl USA 86, 933-937 3 Chang, A S , Yannas, I V , Perutz, S , Loree, H , Sethi, R , Krarup, C , et al (1990) Electrophysiologtcal study of recovery of peripheral nerves regenerated by a collagen-glycosammoglycan copolymer matrix, m Progress znBlomedlcal Polymers (Gebelem,C G. and Dunn, R. L , eds), Plenum,New York, pp 107-120 4 Yannas, I V (1995) Regeneration templates, m The Bzomedzcal Enguzeerzng Handbook (Bronzmo, J D , ed ), CRC, Boca Raton, FL, pp 1619-1635 5 Yannas, I V , Burke, J F , Warpehoski, M , Stasikelts, P , Skrabut, E M , Orgill, D , and Guard, D J (198 1) Prompt, long-term functional replacement of skin Trans Am Sot Artlf Intern Organs 27, 19-22. 6 Yannas, I V , Burke, J F , Orgill, D P , and Skrabut, E M (1982) Wound tissue can utilize a polymeric template to synthesize a functional extension of skm Sczence215, 174-l 76 7 Murphy, G F , Orgill, D P , and Yannas, I V (1990) Partial dermal regeneration is induced by biodegradable collagen-glycosammoglycan grafts Lab Invest 63, 305-3 13 8 Orgill, D P , Butler, C E , and Regan, J F (1996) Behavior of collagen-GAG matrices asdermal replacementm rodent and porcine models Wounds8, 15l-l 57 9 Burke, J F , Yannas, I V , Qumby, W C , Jr, Bondoc, C C , and Jung, W K (1981) Successfuluse of a physiologically acceptable artifictal skm m the treatment of extensive skin mJury Ann Surg 194,413428 10 Helmbach, D , Luterman, A , Burke, J , Cram, A, Herndon, D , Hunt, J , et al (1988) A multi-center randomized clnucal trial artificial dermis for maJorbums Ann Surg 208,3 13-320

Cotlagen-Glycosaminoglycan

Copolymers

17

11 Stern, R., McPherson, M , and Longaker, M T (1990) Htstologtc study of arttficoal skm used m the treatment of full-thtckness thermal mJury J Burn Cure Rehabzl 11,7-l 3 12 Brllmgham, R E and Medawar, P B (195 1) The technique of free skm grafting m mammals J Exp Blol 28,385-394. 13 Btllmgham, R E and Medawar, P B (1955) Contracture and mtussusceptive growth m the healing of extensive wounds in mammahan skin. J Anat 89, 114-123 14 Yannas, I V , Orgill, D P., Silver, J., Norregaard, T V , Zervas, N T , and Schoene, W. C (1987) Regeneration of sciatic nerve across 15mm gap by use of a polymeric template, in Advances zn Blomedzcal Polymers (Gebelem, C G , ed ), Plenum, New York, pp l-9 15 Lundborg, G , Dahlm, L B , Danielsen, N , Gelberman, R H , Longo, F M , Powell, H C , and Varon, S (1982) Nerve regeneration m silicone chambers Influence of gap length and distal stump components Exp Neural 76, 361-375 16 Chang, A S and Yannas, I V (1992) Pertpheral nerve regeneration, m Neuroscience Year (Supplement 2 to the Encyclopedia ofNeurowence), (Smtth, B. and Adelman, G., eds ), Bit-khauser, Boston, pp 125-l 26 17 Chamberlam, L J ( 1996) Long term functional and morphological evaluation of peripheral nerves regenerated through degradable collagen implants S M Thesis, Massachusetts Institute of Technology 18 Stone, K R , Rodkey, W G , Webber, R J , McKmney, L , and Steadman, J R (1990) Future directions, collagen-based prostheses for memscal regeneratton Clin Orthop 252, 129-l 35 19 Loree, H (1988) A freeze-drymg process for fabrlcatlon of polymeric bridges for peripheral nerve regeneration S. M Thesis, Massachusetts Institute of Technology. 20 Forbes, M J (1980) Cross-flow filtration, transmission electron mtcrographtc analysts and blood compattblhty testing of collagen compostte materials for use as vascular prosthesis S. M Thesis, Massachusetts Institute of Technology 21 Sylvester, M , Yannas, I V , and Salzman, E W (1989) Collagen banded fibrtl structure and the collagen platelet reaction Thromb Res 55, 135 22 Chang, A S -P (1988) Electrophystologtcal recovery of peripheral nerves regenerated by biodegradable polymer matrix S M Thesis, Massachusetts Instttute of Technology. 23 Loree, H. M , Yannas, I V , Mrkic, B , Chang, A S , Perutz, S. M , Norregaard, T V , and Krarup, C (1989) A freeze drying process for fabrication of polymeric bridges for pertpheral nerve regeneration Proc NE Bloeng ConJ 53-54 24 Yannas, I. V and Tobolsky, A V. (1967) Cross-lmkmg of gelatme by dehydration Nature 215,509-5 10 25 Yannas, I. V (1972) Collagen and gelatm m the solid state, J Mucromol Scl Rev Macromol Chem C7,49-104 26 Yannas, I V , Burke, J F , Huang, C , and Gordon, P L 91975) Correiatton of m vlvo collagen degradation rate wtth m vitro measurements J Blamed Mater Res 8,623%628

Preparation and Use of Tethered Ligands as Biomaterials and Tools for Cell Biology Susan J. Sofia, Philip R. Kuhl, and Linda G. Griffith 1. introduction The munoblllzatlon of blomolecules to various supports has been an lmportant research area for many years. Molecules such as heparm (1,2), as well as various enzymes (3,4), antibodies (5,6), and adhesion hgands (7), have been bound to such supports as slllcon or glass, agarose gels, polyethylene oxide (PEO) gels, poly(vmy1 alcohol) (PVA) gels, and polymer surfaces. These studies have had important appllcatlons m the creation of antlthrombogemc surfaces for blood contact, affinity chromatography, and tissue growth and regeneration. Most apphcatlons of surface-bound blomolecules m biotechnology and medicine have involved soluble llgands or substrate mteractlons with the bound molecule (e.g., affinity chromatography, blosensors, lmmoblllzed enzymes), and have thus employed lmmoblllzatlon techniques m which these molecules are indeed essentially lmmobllized (covalently bound directly to the substrate or linked via short spacers) A major concern in such applications 1sorientation of the molecule when lmmobillzed (4), particularly with protems, smce multiple reactive sites (typically amme or carboxyl side chains) are usually present. Apphcatlons of surface-bound bromolecules m blomaterlals and tissue engmeermg, m contrast to those In biotechnology, usually involve mteractlons of the bound molecules with receptors m the cell membrane. In biotechnology appllcatlons, the density of bound blomolecule may affect the rate of reaction or the efficiency of a packed-column process, but varlatlons in density of surface-bound hgands for cell receptors can have profound effects on cellular responses, e.g , engendermg cell adhesion and spreading as threshold phenomena (S), or enabling formation of specialized cell adhesion structures (9). From

Methods in Molecular Edlied by J R Morgan

Medune, Vol 18 Twue Engmeenng Methods and and M L Yarmush 0 Humana Press Inc Totowa.

19

Protocols NJ

20

Sofia, Kuhl, and Griffith

Increasing the density of substrate-bound hgand presumably allows receptors, which bind the ligands, to aggregate or cluster together because of the decreased distance between receptor-hgand complexes. In adhesion, clustering of ligand-bound mtegrms is associated with acttvation of signaling pathways (IO,II), and controllmg the spatial organization of receptors IS thus of Interest. It seems likely that, for some aspects of cell behavior mediated by adhesion, the effects of increased hgand concentratton on cell processes arise mainly from altering the ability of receptor-ligand complexes m the cell membrane, and not because of greater number of receptor-hgand bonds bmdmg For example, by increasing the length of the spacer linking the ligand to the substrate, cell behavior, such as spreading, can be observed at a lower hgand density (12) The ability to control ligand spatial organization independent of hgand density-and thus, presumably, the spatial organization of receptor-hgand complexes-could be a useful tool m studying receptor-mediated processes, mcludmg interactions with cytokmes and growth factors. Receptor dimerizanon has been suggested as a necessary event m signal transduction by several growth factors, but remains controverstal for some growth factors, mcludmg eptdermal growth factor (EGF) (13-15). We thus focus on procedures to tether, rather than immobilize, biomolecules on substrates suitable for cell culture, and present procedures for creating clusters of tethered hgands, using EGF as an example In the followmg procedures, we use PEO star molecules as linkers to tether the btomolecule to the surface. PEO star molecules have a poly(dtvmy1 benzene) core, with PEO arms extending radially outward from the core (16-18). The length (M,,,) and number (functionality) of arms can vary. The multiplearm nature of stars allows a single molecule to have several ligands bound (bmdmg occurs at the free ends of the PEO arms), and to retam sites for lmkmg the molecule to the substrate. PEO 1sperhaps ideally suited as a tether since it is extremely flexible and mobile when m aqueous solutton (19). It therefore gives a very mobile presentation of the ligand to the surrounding environment PEO 1s also well known to be protein- and cell-adhesive (20). This results m the hgand and cell mteractmg m a specific manner, as desired, without nonspecific mteractions with the PEO tether The advantages of usmg PEO star molecules over linear PEO to tether the hgand are twofold. First, because of sterlc hindrance m the core region, the PEO arms of a star molecule are m an extended conformation, thus making the arm chain ends more accessible to the outer regions of the star molecule and surroundmg environment (21,22). This is unlike the chain ends of a lmear molecule, which can exist anywhere wtthin the volume of the molecule Second, because of the larger number of chain ends per star molecule than per linear (depending on the star functionality), star

Tethered Ligands

21

molecules have a higher probability of binding to a surface via these arm chain ends, as well as leaving a large number of free chain ends, once the star is bound, for the attachment of a hgand. We present here methods for bmding EGF to glass surfaces, using PEO star molecules as tethers With this method, the ligand 1s bound first to the stars, and the hgand-star coqugate is then bound to the surface. The stars are also reacted with ethylene diamme, which functtonalizes some of the arm ends with an amme, which are then used to bind the star to the surface. This approach mmimizes nonspecific adsorption of protein on the surface and provides a facile means for independently varying the ligand density and the llgand spatial organization. Two similar procedures are presented. Procedure I describes a tethering scheme more specttic to EGF; m this scheme, each star molecule bound to the surface has only one EGF molecule bound to it. Procedure II is more general, we use EGF as an example, but the procedure can be applied generically to biomolecules that can be linked via a free amine. Also, the number of molecules bound to a star molecule 1snot limited to one. We use murme EGF for tethering, because rt possessesonly a single amme, the amme terminus, to which the PEO tether can bind. The amine terminus 1s not involved m receptor bmdmg, and thus EGF has the potential to be fully active when tethered, as confirmed by its ability to stimulate DNA synthesis m cultured hepatocytes (23) 2. Materials 2.1. Tresylation

of PEO Star Molecules

PEO star molecules (Shearwater Polymers, Huntsvtlle, AL). Oven-dried, glass, round-bottom (rb) flask (100 mL) and cap with stir bar Triethyl amme (TEA) (Aldrich, Milwaukee, WI) Tresyl chloride (TrCl, trifluoroethanesulfonylchloride) (Aldrich). Caution: Use extreme care when using TrCl, as it is very toxic and corrosive TrCl is also moisture sensitive Unused material can be stored m an oven-drred glass vial with plastic cap m a desiccated Jar at 4°C. Dichloromethane (CH,Cl,), methanol (MeOH); concentrated hydrochloric acid (HCl, -37%), hquid nitrogen Molecular sieves, maintained m convection oven at 120°C cooled to room temperature (RT) before use.

2.2. Preparation

of Glass Surface

1 Glass slides, 1 x 3 m , with Teflon-lined wells (Cel-Line, Newfield, NJ) 2 Concentrated sulfuric acid, concentrated hydrochloric acid, glacial acetic acid, methanol 3. 18 MOhm purified water

Sofia, Kuhl, and Gnfflth

22 4 5 6 7 8 9

Trtmethoxystlylpropyldtethylene triamme, stored desiccated away from light (Huls Sthcon Compoundscatalog, United Chemical Technologtes, Bristol, PA) Phosphatebuffer 0 1 MNaH,PO,, pH 7 (Procedure I only) Glutaraldehyde 70% aqueoussolutton (Sigma, St LOUIS,MO), dilute to 2.5% with phosphatebuffer (Procedure I only) Reductant solutton (Caution: Prepare m fume hood). Dissolve 12 8 mg sodium cyanoborohydride (Aldrich) per 1 mL 10 mMNaOH (Procedure I only) Succmtc anhydride, stored under nitrogen (Aldrich) (Procedure II only). ECD (I-ethyl-3-[3-dimethylammopropyl] carbodnmide hydrochloride), stored desiccated at -20°C NHS (N-hydroxysuccmtmtde), stored desiccated at -20°C (Pierce, Rockford, IL) to be mixed m dioxane m chemical Fume hood (Procedure II only)

2.3. Preparation

of lz5/-EGF (lodination)

1 Eptdermal growth factor (mouse, Collaborattve Btomedtcal, Bedford, MA) Resuspendlyophthzed material m 100 mM NaH,PO,, pH 7 0, to concentratton of 200 pg/mL Ahquot m 0 25 mL portions, store at -20°C 1.5 mL screw-cap mtcrocentrtfuge tubes Iodo-beads (Pierce), stored desiccatedat 4°C Iodmation reaction buffer 100 mMNaH,PO,, pH 7.0. Iodmatton column buffer 1 mg/mL poly(ethylene oxide), mol wt 1000 g/mol, dtmethyl ether (DMPEO) (Polysciences, Warrmgton, PA) in iodmation reaction buffer 6 Chase solution. one crystal KI (potassium iodide, Aldrich) m 20 mL column buffer 7 Carrier protein solutton Dissolve 100 mg bovine serumalbumin (BSA) and one crystal KI m 20 mL Dulbecco’s phosphatebuffered salme(1X, without Ca*+and Mg*’ tons, Sigma) 8 12 mg/mL sodium metabtsulfite m todmatton reaction buffer, made up on the day of use, 1 mL 9 KwtkSep Excellulose disposablegel filtration columns (5 mL) (Bto-Rad, Hercules, CA) Equthbrate with 2 column volumes of iodmation column buffer 10 1% phosphotungstic acid m 1 N HCl 11 Na1251,2-8 mC1, depending on the specific activity desired (NEN Life Science, Boston, MA) 2 mCi reacted with 10 ug EGF results m specific activity 3&40 pCi/pgEGF

2.4. Preparation and Purification of Amine-Star-EGF 2.4 1 Preparation Usmg 1251-EGF 2 4 1 1 PROCEDURE I 1 ‘251-EGF 2 Tresylated star PEO 3. Star PEO drssolutton buffer, 10 rnA4sodium acetate, pH 4

Tethered Ligands

23

4 Ethylene dramme (EDA) (Sigma) 5. Econo-column, 2.5-cm diameter, lo-cm bed hetght (Bto-Rad) 6 Gel filtratron column packmg* Bra-gel P-60, fine (mol wt cutoff 60 k, Bto-Rad) (see Note 2) 7 GuHCl gel tiltratton buffer. 10 mMNaH,PO,, 0 5 Mguamdme HCl (purest grade, Sigma), with 1 mg/mL dtmethoxy PEO (mol wt 1000 g/mol, Shearwater Polymers), pH 6.2. 8 Amon exchange column Hi-Trap Q, 1 mL (Pterce) 9 Amon exchange start buffer 10 n&I piperidine, pH 10 8 (Sigma) 10 Amon exchange regeneration buffer 10 r&4 bu-Trts propane, 0 5M NaCl, pH 7 11 Grafting buffer: Dissolve 1 mg/mL dtmethoxy PEO m 0 1 M NaH2P04, adjust pHto70

2.4.2. Preparation using 1251-EGFand Nonrodlnated EGF 2 4 2 1 PROCEDURE II ‘251-EGFand nomodmated EGF Tresylated PEO stars Ethylated dtamme(EDA), ethylene dtammedthydrochlorlde (EDA-HCl), sodium aztde (NaNs) (Aldrich) Dulbecco’s phosphate buffered salme (DPBS), 102 (0 1 M phosphate, without Ca2+andMg2+ tons), titrated to pH 7 4 (Sigma). (When diluted to 1X, no need for tltratron-already at pH 7 4) Econo-column 2 5-cm diameter, IO-cm bed height (Bto-Rad) Bra-gel P-30, fine, gel-filtratron packmg (mol wt cutoff 45k, Bto-Rad), prepared with 1X DPBS with 0 04% (w/v) NaNs) (seeNote 2)

3. Methods 3.1. Tresylation of Star PEO 3 1 1 Preparation 1 Dry lOO-mL rb flask, wtth stir bar, m convectton oven, at 120°C for at least 24 h 2 Dissolve 1 g star PEO m 12 mL CH2C12(Caution: m fume hood) Once fully dtssolved, add 1 g molecular steves,wart - 30 mm for bubblmg to cease,sealthe vessel, and store at 4°C for 24 h 3 Add lg molecular stevesto 10 mL of CHzCl, and to 10 mL TEA Watt -30 mm for bubbling to cease,tightly sealthe vial, and store at 4°C for 24 h

3.1.2. React/on (Caution:

in fume hood)

1 Cool the rb flask and stir bar, while capped, to room temperature 2 Decant the PEO star solutton mto the rb flask Decant the excess 10 mL CH,Cl, mto the star/sieves vesseland swirl to recover most of the restdual star solution left m the container

Decant this Into the rb flask

Only uncap the rb flask for as

long as necessary to add the required soluttons, otherwtse, keep It capped (see

24


Note 1) Set the stn bar rotating at a slow to intermediate speed, to create a small vortex 3 Ptpet dropwrse mto the rb flask the required volume of TEA, to be 1 5x the number of mol of -OH groups m the 1g star sample, I e , there should be a 50% excess number of moles of TEA as -OH groups (The groups of moles of -OH groups IS determined by dtvtdmg the mass of star to be tresylated (1 g) by the mol wt of an arm, Ma,,.) 4 Ptpet dropwlse mto the rb flask the required volume of TrCI, so that it IS also at 50% excess over the number of moles of OH groups. 5 Cap tightly and let react at room temperature for 90 mm

3 1.3. Purification of Tresyl-PEO Star Caution:

Except when centrifugmg,

work m a chemical

fume hood

Evaporate off the CH$l, using a liquid nitrogen trap Redissolve the drred residue m 50 mL acidified MeOH (250 pL cone HCl m 50 mL MeOH) Place the solutton mto centrifuge tubes and maintain at -20°C for at least 3 h, to allow the star PEO to prectpttate Centrrfuge at 7000 t-pm, at -20°C, for 25 mm Immediately decant off the top solvent Redrssolve the precipitated residue wrth moderately acidified MeOH (50 $ cone HCI m 50 mL MeOH) Reprectprtate at -20°C for at least 3 h, centrifuge, and tmmedtately decant off the top solvent Repeat steps 5 and 6 3x for a total of five prectpttattons With final precrpttate, evaporate off the solvent using a lrqutd nttrogen trap Etther use the tresyl-PEO tmmedtately, or store under nitrogen, destccated, at -70°C

3.2. lodination

of EGF

Several methods for todmatlon of proteins are avatlable. The use of Iodobeads from Pierce IS a mild method applicable for proteins contammg tyrosme residues, like EGF. The only departure from the standard technique IS that no protein stabilizer (usually BSA) 1s added to the protein mixture m order to avoid interference with the later couplmg and grafting reacttons. This procedure should, of course, be performed only m authorized areas by researchers tramed m the use of radioactive materials. Care IS taken to maxlmlze recovery of EGF, m order to increase the accuracy of the SA measurement The followmg procedure IS gtven for todmatmg 10 pg of protem. More or less protein can be todmated as needed, with the specific actrvity mcreased or decreased by varying the amount of Na’251 used m the todmation reaction.

25

Tethered Llgands 3.2. I. lodinatlon Reaction

1 Eqmhbrate KwikSep column with 2 column volumes (10 mL) of column buffer 2 Place an Iodo-bead m a mrcrocentrlfuge reactton tube and rinse wtth 1 mL reac-

tton buffer Pour off the buffer 3 Drlute the Na’251 wtth 20 pL of reactton buffer, then add the solution to the reac-

tton tube 4. Let stt for 5 mm 5 Add 100 pL EGF solution (0 1 mg/mL) to reaction tube Save ptpet ttp and EGF

tube (if now empty) 6 Watt 15 mm (durmg thus time, prepare the two precolumn tubes, see step 14). 7 Using the saved tip, transfer the contents of the reaction vial (-120 pL) to the

8 9

10 11 12 13 14 15 16 17

18

19

origmal EGF tube (or fresh tube, tf EGF tube not empty) Save tip and reaction tube Add 80 pL sodium metabrsulfite solution to the reaction tube, and swirl gently. With saved tip, transfer the 80 l,tL to the EGF tube Save tip and reaction tube again Add 60 & chase solution to the reaction tube, swirl gently With saved tip, transfer the 60 pL to the EGF tube Save tip and reaction tube Repeat steps 10 and 11 once Mix the contents of the EGF tube Take two precolumn samples 3 pL added to 997 pL carrier protein solution Load the labeled protein mixture onto the Kwtk Sep column, but do not open outlet yet Rinse the EGF tube twice with 150 pL column buffer each rinse, and add the rinses to the sample on the column before opening the outlet Open the outlet and collect fractions m screw cap mtcrocentrrfuge tubes Collect 5 mm m the first tube, then take 2-mm fractions untrl 25 mm, adding column buffer to the top of the column once the sample fully enters the column gel Count the fractions (a pancake-type, hand-held probe works well for this purpose, or an extremely low-efficiency NaI detector) and combme the hottest 3-5 fractions m one tube MIX well. Take two postcolumn samples Add 2 pL combined fracttons solution to 998 I.LL carrier protein solutron Mix well

3.2.2. Determination

of Specific Activity

Specific activity of the todinated protein is determined the protean wtth phosphotungsttc actd.

by precipitation

of

1 Vortex the two precolumn and two postcolumn samples 2 Nonprectprtated samples a Put 995 pL reaction buffer mto each of eight gamma tubes b Transfer 5 pL from each pre- and postcolumn sample mto each of the gamma tubes, do m duplicate

26

Sofia,

Kuhl, and Griffith

3

Precipitated samples a Put 20 ccl, carrier protein solution mto each of eight screw-cap tubes. b Add 5 p.L from each of the pre- and postcolumn samples to the carrier protem tubes, do m duplicate c Add 1 pL phosphotungstlc acid to each tube, vortex d Centrifuge 2-3 mm at high speed on desktop centrifuge e Discard supernatant and add 1 mL 1 N NaOH f Vortex to dissolve pellet g Transfer to labeled gamma tubes 4 Count all nonpreclpltated and precipitated tubes m gamma counter 5 To calculate specific activity m counts/mm/rig (CPM/rig), multiply the CPM average value obtained for the precipitated precolumn samples by (200/3) x (sample volume loaded on column)/(ng EGF lodmated) To convert CPM to $1, divide CPM by the efficiency of the gamma counter, then divide by 2 22 x lo6 The concentration of protein m the postcolumn combined fractions In CPM/mL can be determined by multlplylng the average nonprecipitated postcolumn CPM by lo5 The specific activity can then be used to convert from CPM/mL to ng/mL

3.3. Activation

of Glass Slide Surfaces

In these first two steps, the glass slides are cleaned and reacted with an ammosllane to obtain reactive primary ammes, -NH*, on the surface. These ammes are used m subsequent chemistry to bmd the hgand-star conjugate to the surface The S1-0 bond that binds the ammosllane to the glass surface IS sensitive to degradation by base. Therefore, whenever usmg the ammosllanetreated samples, the pH should be mamtamed below pH 8.0. 3.3.1.

Cleaning

of Glass Slides

1 Immerse the slides m a 40% cone H2S04 m water solution 30-60 mm.

Let soak for

2 Rinse slides well with purified water, then lmerse m a 1 1 cone HCI methanol solution. Let soak for 3&60 mm 3 Rinse slides well with purified water, then rinse once m methanol 4 React slides with ammosllane lmmedlately after cleanmg

3.3.2

Reaction

with Amlnosiiane

1 Prepare a 1 mM acidified methanol solution, using glacial acetic acid 2 Prepare a 95% (v/v) acidified methanol, 1% (v/v) ammosllane, 4% (v/v) water solution, adding first the ammosllane, mixmg for several seconds, then lmmedlately adding the water and mlxmg

3 Immediately add the cleaned glass shdes and let react for 15 mm at room temperature 4 Rinse the slides well with 34 washes of methanol. Cure m a convection oven, 12O”C, for 15 min, or let sit covered, m air, overnight to cure

Tethered Ligands

27

3.3 3. Conversion of Surface -NH, to -COH 3.3.3.1.

PROCEDURE I

1, Rinse slides with grafting buffer 2 Apply glutaraldehyde solution to each Teflon-lined well and incubate at RT m a humid box overmght 3 Remove glutaraldehyde from wells and rinse thoroughly with grafting buffer. 4 React immediately with amine-star-EGF conjugate (see Subheading 3.4.3.)

3.3.4. Conversion of Surface -NH, to -COOH 3 3 4.1 PROCEDURE II 1 Dissolve 0 3 g succmic anhydride m 100 mL 5X DPBS (0 05 A4 phosphate) titrated to pH 6 0 It takes several minutes of stnrmg for the succnnc anhydride to fully dtssolve 2 Place cured slides m the solutton and let react, at room temperature, for 15 h 3 Rinse well with water, then rinse once with methanol

3 3 5. Conversion of -COOH to Activated Ester Surfaces activated with EDC/NHS are moisture- and UV-sensttive They should be stored desiccated and sealed from light and used no more than 1 wk from when first made They can also be stored m dried dioxane for l-2 mo. Use care when runmng the activation step, as dioxane is extremely toxic. Caution: Work m a chemical fume hood 1 Place slides m the appropriate volume of dioxane (e g , 100 mL), and set stir bar spmnmg at intermediate speed 2 Slowly add m NHS and EDC, each to a concentration of 0.1 A4 (e.g., m 100 mL dioxane, add 1 2 g of NHS and 1 9 g of EDC) A white slurry results 3 Let react at room temperature for 90 mm By the end of the reaction, some residue may adhere to the slides 4. Rinse the slides once m dioxane, then 34x m methanol. Place immediately m desiccated container until ready for reaction with amme-star-hgand conjugate

(see Subheading

3.4.7.)

3.4. Formation and Grafting of Amine-Star-EGF Conjugate Of key importance m the formation of amme-star-ltgand coqugates is the number of ligand molecules that covalently bind to a single star molecule The number of hgands/star is mainly governed by the ratio of moles of tresylated star to moles of hgand when binding the hgand to the star molecules (seeNote 3). Procedure I provides a method of bmdmg only one EGF molecule to a star molecule, Procedure II gives a method of bmdmg one or more EGF (or other ligand) molecules to a star molecule. Procedure II also gives a means for estimatmg the number of ligand molecules bound to each star It is written


28

wtth a variable mole ratio of hgand to starsnot fixed, so that the experimenter has the freedom to choose this parameter. The resultmg number of hgand molecules per star molecule for Procedure II is a calculated average, m actuality, there 1sa distribution in the number of protein molecules bound to the star molecules 3.4 1 Synthesis of Amine-Star-EGF The amount of EGF to be reacted should be such that there 1s 1000x molar excess of star -OH groups over moles of EGF 1 Add 104 x molar excess EDA (excess over moles EGF) to reactton veal contammg todmated EGF and vortex (e.g , add 42 ~01 EDA to 4 2 nmol 125I-EGF) 2 Dissolve tresyl-acttvated star PEO m small volume (300 pL for 4 2 pm01 star OH groups for the example given m step 1) 10 mM acetate (pH 4) Take care that the star PEO completely dissolves (see Note 4) 3. Add completely dissolved tresyl-star solutton to the reaction vial and vortex again 4 Incubate at 4°C overnight

3.4.2. Purification of Amine-Star-EGF 1 Eqmltbrate gel filtratron (GF) column with GuHCl gel filtration buffer 2 Dilute reaction mixture 1.2 with GuHCl gel filtration buffer 3 Load dtluted reaction mtxture on GF column Watt for sample to completely enter the gel, then load GF column buffer to the top of the column. 4 Collect 3-mm fractions m gamma tubes for 7&90 mm 5 Count the fractions on gamma counter and pool the fractions from the first peak These are the PEO star fractions 6 Dialyze pooled PEO fractions against anion exchange start buffer at 4OC, usmg, for example, small volume dialysis cassettes (Pierce) Load pooled fractions mto cassettes using syringe and needle, and dialyze against I-2 L amon exchange buffer Change buffer twice 7 Amon exchange chromatography a Equilibrate anion exchange column with 5 mL start buffer, followed by 5 mL regeneration buffer, and then 5 mL start buffer again b Load dialyzed star fractrons onto column and wash through with 5 mL start buffer c Elute the amine-star-EGF conjugate with 5 mL regeneration buffer d Dtalyze the amme-star-EGF against grafting buffer at 4°C (l-2 L), replace buffer three ttmes

3.4 3 Grafting of Amine-Star-EGF

to Glass Surface

1 Add amme-star-EGF conjugate m grafting buffer to -COH activated wells and incubate at 4°C for 2 h 2 Remove conlugate solutton (It may be reused) 3. Rmse wells thoroughly with grafting buffer to remove nonspectfically adsorbed conjugate

Tethered Ligands

29

4 To quantify star-EGF bmdmg to solids, place them face up m thm-walled, flatbottom plastic containers (to mamtam sterihty) 5 Place the samples, along with standards of known 1251activity, on a phosphorimager plate (Phosphorimager with reusable phosphor screens, Molecular Dynamics, Sunnyvale, CA) The isotope emits radiation of sufficient energy to penetrate the glass and plastic and expose the plate on the order of hours to days Star-EGF slides should be stored at 4°C.

3 4.4. Synthesis of Amine-Star-EGF 3.4 4.1 PROCEDURE II 1 Dissolve the appropriate amount of EGF m 1 mL purified water Use 500 pL to first dissolve the EGF, then rinse the EGF product vial(s) with the remaining 500 l.tL to recover asmuch of the EGF as possible 2 Add 500 uL of the ‘251-EGFsolutron to the unlabeled EGF solution Mix well 3. Dissolve the appropriate amount of tresyl-star m 1 mL of 10X DPBS, titrated to pH 7 4 The solution ~111bubble asthe PEO dissolves 4. Immediately upon dtssolvmg all the tresyl-PEO, add it to the EGF solution. Mix well and let react for 15 mm 5 Add 100 times molar excessof EDA over the total number moles of-OH groups, with 75% of the added moles of EDA from EDA-HCl, and the remammg 25% from EDA, with the EDA-HCl being addedfirst (to prevent the pH from becommg too basic). Mix well, and let react overnight at 4°C

3.4.5. Punficatlon of Amine-Star-EGF 1 Prepare gel filtratron column with Bio-Gel P-30 packing, bed volume 50-70 mL, with 1X DPBS buffer with 0.04% NaNs 2 Run the samplethrough the column, collecting fractions every 2 mm m gamma tubes, for 70-90 mm If possible, wrap the column m lead while the sample 1s runnmg through, as well as using lead to shreld the collected samples,for the safety of those workmg m the area 3 Countthe fractionson agammacounterandpoolthe highestfractionsfrom the fast peak

3.4.6. Calculation of EGF Functionality on Stars 1 Weigh an empty glassvial and cap Put the pooled fractions m the veal and weigh the vial again to obtain the approximate volume. Lyophihze to recover the dried amme-star-EGF conmgate and buffer salts, 2. Reweigh the vial and cap with solids. Subtracting the weights of the vial, cap, and buffer salts gives the massof conlugate recovered 3 Knowmg the total radtoactivlty from the counts of the pooled fractions allows a calculation of the total moles of EGF present. A calculated estimate of the number of EGF moleculesper star molecule can then be obtained No mol EGF per star = (moles EGF ion pooled fractions)(M, star)/ (massconmgate recovered).

Sofia,

30 3.4 7. Grafting

of Amine-Star-EGF

Kuhl, and Griffith

to Glass Surfaces

Dissolve the lyophihzed amme-star-EGF product with purified water to the desired concentration, generally, between 1 and 10 mg/mL should be sufficient (see Note 6) Titrate the solution of pH 7 4, if needed, using 0 1 MNaOH or HCI Pipet the solution onto the activated slides and let react for 4 h at room temperature (For a 4-cm2 activated surface area, 25&300 pL of solution are needed to fully cover the surface for reaction.) Rmse the slides thoroughly with 1X DPBS, using either a gentle stream pipeted onto the surface (e g , a dozen pipet rinses with l-l 5 mL buffer per rinse), or swirled gently m 3-4 washes of 1X DPBS If pipetmg, the rmses can be counted m a gamma counter to check when radioactivity commg off the surfaces reaches a minimum To determine the amount of EGF present on the surface, radioactivity can be measured using phosphorimagmg, as described m Subheading 3.4.3. step 5, or sample slides not to be further used can be carefully cut mto pieces and the counts on each mdivtdual piece measured m a gamma counter Knowing the surface area treated, a surface density of EGF can then be calculated

4. Notes 1 When tresylating PEO, it is important that all glassware and reagents be dried, free of moisture, and that the reaction solution be exposed to an as little as possible, because the extent of tresylation is very dependent on the absence of water Of course, more extensive measures could be taken, such as runmng the reaction under mert gas, but with the method outlined here, yields of 75-80% can be obtained when starting with eqmmolar quantities of -OH groups, TrCl, and TEA 2 Using different gels to separate stars and free hgand depends on the relative sizes of two species 3 In bmdmg bgands to star molecules, the relative amounts of both star and hgand to be used need to be taken mto consideration In the case of EGF, because EGF IS a rather expensive protein, smaller quantities of materials were generally used As an example, for a star molecule with a functionality of 70 and arm mol wt of 5200 gimol (therefore, M, = 350,000 g/mol), if 20 mg of tresylated star material is used to react with EGF, then an eqmmolar amount of EGF would be -350 pg Considermg that the yield of this particular reaction (I e , the fraction of starting EGF that binds to the stars and recovered from the GF column) IS about 30%, then roughly three times this amount of EGF, -1 mg, would be needed to ensure that all star molecules, on average, have one EGF molecular bound to them If 10 mg tresyl-star were reacted with 1 mg EGF, then on average -2 EGF molecules would bind to each star. For other hgand/star molecule systems, these amounts would probably change 4 If tresylated PEO has been stored for a period of time, one way to determine if the stored material has become unusable (because of crosslmkmg) is by watching for the presence of tmy microgels when dissolved m aqueous solution In dissolvmg

31

0 ! 0

I 1

I

I

2

3

4

Star-amlne Concentration (wt %) Fig 1 Dependence of PEO dry layer thickness on star-amine

concentratton

the material, once the bubbles rise to the top, check the solutton to see that tt 1s clear, free of floating microgel parttcles 5 Another issue to consider IS the particular star type to use, 1 e , star functionality and arm mol wt As to the former, with functionalmes greater than -18-20, star molecules m good solvent behave as hard spheres (18) They have very strong excluded volume effects, so that they are very difficult to overlap, both m solution and on a surface. Larger functtonalmes provide for more sues per star to whtch EGF and EDA can bmd As stated m the mtroductton, the star arms are m an extended conformatton as a result of sterlc hmdrance in the core region This arm extension increases with functionality The accessibility of the arm ends, however, decreases with mcreasmg arm mol wt, because of a return to more randomly cotlmg characterlsttcs We therefore recommend using shorter arm mol wt (ca 500~10,000 g/mol) and larger functionalmes (20 or more) for use m tethering hgands to surfaces 6 When working with the protein, tt 1s of course important to maintain the pH of solutions around neutral pH, to prevent the protem from denaturing Generally, a range of pH that 1s usually safe IS pH 6-8, but the stabtllty of the specific hgand to be used should be investigated before use 7 To gave an idea of the dependence of star-amine coupling to an EDUNHS acttvated surface on couplmg concentratton (1 e , concentration of star-amine m solution), Fig. 1 shows a plot of the dependence of PEO dty layer thickness, as measured by XPS, as a function of coupling concentration for two specttic star molecules Both star types have functlonalmes m the hard-sphere regime, with arm mol wt m the range of 5000-10,000 gimol The data was obtained for star molecules reacted with EDA only, no hgand

32


References 1. Tay, S. W., Merrill, E. W., Salzman, E. W., and Lindon, J. (1989) Activity towards thrombin-antithrombin of heparin immobilized on two hydrogels. Biomaterials 10, 1 l-15. 2. Lindberg, B., Maripuu, R., Siegbahn, K., Larsson, R., Golander, C.-G., and Eriksson, J. C. (1983) ESCA studies of heparinized and related surfaces. J. Coil. Int. Sci. 95,308-32 1. 3. Trevan, M., ed. (1980) Immobilized Enzymes: An Introduction and Applications in Biotechnology, Wiley: New York. 4. Leckband, D. and Langer, R. (1991) An approach for the stable immobilization of proteins. Biotechnol. Bioeng. 37, 227-231. 5. Huang, S.-C., Caldwell, K. D., Lin, J.-N., Wang, H.-K., and Herron, J. N. (1996) Site-specific immobilization of monoclonal antibodies using spacer-mediated antibody attachment. Langmuir 12,4292-4298. 6. von Sommern, A. P. G., Machielsen, P. A. G. M., and Gribnau, T. C. J. (1993) Comparison of three activated agaroses for use in affinity chromatography. J. Chromatog. 639,23-3 1. 7. Massia, S. P. and Hubbell, J. A. (1990) Covalent surface immobilization of ArgGly-Asp and Tyr-Ile-Gly-Ser-Arg-containing peptides to obtrain well-defined cell-adhesive substrates. Analyt. Biochem. 187,292-30 1. 8. Weigel, P. H., Schnaar, R. L., Kuhelnschmidt, M. S., Schmell, E., Lee, R. T., Lee, Y. C., and Roseman, S. (1979) Adhesion of hepatocytes to immobilized sugars: a threshold phenomenon. J. Biol. Chem. 254, 10,83&10,838. 9. Massia, S. P. and Hubbell, J. A. (1991) An RGD spacing of 440 nm is sufficient for integrin alpha V beta 3-mediated fibroblast spreading and 140 nm for focal contact and stress fiber formation. J. Cell Biol. 114, 1089-l 100. 10. Kornberg, L. J., Earp, H. S., Turner, C. E., Prockop, C., and Juliano, R. L. (199 1) Signal transduction by integrins: Increased protein tyrosine phosphorylation caused by clustering of beta-l integrins. Proc. Natl. Acad. Sci. USA 88, 8392-8396. 11. Plopper, G. E., McNamee, H. P., Dike, L. E., Bojanowski, K., and Ingber, D. E. (1995) Convergence of integrin and growth factor receptor signaling pathways within the focal adhesion complex. Mol. Biol. Cell 6, 1349-1365. 12. Lopina, S. T., Wu, G., Merrill, E. W., and Griffith-Cima, L. G. (1996) Hepatocyte culture on carbohydrate-modified star polyethylene oxide hydrogels. Biomaterials 17,559-569. 13. Sorkin, A. and Carpenter, G. (1991) Dimerization of internalized epidermal growth factor receptors. J. Biol. Chem. 266,23453-23460. 14. Mohammadi, M., Honegger, A., Sorokin, A., Ullrich, A., Schlessinger, J., and Hurwitz, D. R. (1993) Aggregation-induced activation of the epidermal growth factor receptor protein tyrosine kinase. Biochemistry 32, 8742-8748. 15. Carraway, K. L. and Cerione, R. A. (1993) Inhibition of growth factor receptor aggregation by an antibody directed against the epidermal growth factor receptor extracellular domain. J. Biol. Chem. 268,23,860-23,867.

Tethered Ligands

33

16. Merrill, E. W. (1993) Poly(ethylene oxide) star molecules: synthesis characterization, and applications in medicine and biology. J. Biomater. Sci., Polymer Ed., in press. 17. Gnanou, Y., Lutz, P., and Rempp, P. (1988) Synthesis of star-shaped poly(ethylene oxide). Makromol. Chemie 189,2885-2892. 18. Lutz, P. and Rempp, P. (1988) New developments in star polymer synthesis: star-shaped polystyrenes and star-block copolymers. Makromol. Chemei 189, 1051-1060. 19. Wong, J. Y., Kuhl, T. L., Israelachvili, J. N., Mullah, N., and Zalipsky, S. (1997) Direct measure of a tethered ligand-receptor interaction potential. Science 275, 82@-823. 20. Harris, J. M., ed. (1992) Poly(ethylene glycol) Chemistry. Topics in Applied Chemistry (Katritzky, A. R. and Sabongi, G. J., eds.), Plenum, New York, p. 385. 2 1. Forni, A., Ganazzoli, F., and Vacatello, M. (1996) Local conformation of regular star polymers in a good solvent: a Monte Carlo study. Macromo/ccu/es 29, 2994-2999. 22. Grest, G. S., Kremer, K., and Witten, T. A. (1987) Structure of many-arm star polymers: a molecular dynamics simulation. Macromol. 20, 13761383. 23. Kuhl, P. R. and Griffith-Cima, L. G. (1996) Tethered epidermal growth factor as a paradigm for growth factor-induced stimulation from the solid phase. Nature Med. 2, 1022-1027.

3 Nondestructive Evaluation of Biodegradable Porous Matrices for Tissue Engineering Leoncio Garrido 1. Introduction Slgmfkant advances m transplantatton therapy for end stage organ failure or ttssue repair have led to a shortage of donor organs and reconstructtve tlssue Alternatives to remedy this deficrency are being mvesttgated In parttcular, the engineering of ttssue using natural or synthetic polymers and Isolated cells from the organ or ttssue of interest has received increasing attention m the past few years (I-9). The idea behind this approach 1s that the extracellular matrices made of biodegradable and btocompattble polymers can provide mltral structural support for cell adhesron and growth, and gradually reabsorb during organ/ttssue generation without any detrimental effect to formed tissue. In this chapter, we describe the preparation of btoabsorbable extracellular matrices of poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBHV) copolymers by solvent casting (10,11). This family of polymers degrades slowly, and, m the process, produces a less acidic environment than other polymers currently used for tissue scaffolding (12). Because of these properties, PHBHV copolymers may be alternatives to polylactrdes and polyglycohdes for those applications m which prolonged cell support to maintain cell vtabtlrty and dtfferenttated functron IS required. The porous matrix for tissue engineering should have hrgh porosity to be able to accommodate a large number of cells, as well as large interconnected pores to facthtate a umform dtstrrbutlon of cells and the diffusion of oxygen and nutrients to them (1.?,14) In this context, tt would be advantageous for destgnmg and opttmrzmg extracellular matrices to be able to study the proliferation, vtabihty, and function of cells m well-characterized matrtces. To From

Methods m Molecular Edlted by J R Morgan

Medmne, Vol 18 Tfssue Engmeenng Methods and and M L Yarmush 0 Humana Press Inc , Totowa,

35

Protocols NJ

36

Gamdo

Implement this approach, the use of nondestructive methods of analysis is required. The ability of nuclear magnetic resonance (NMR) to nondestructively determme the compositton of a given compound is well known. Moreover, the NMR signal can be spatially encoded using magnetic field gradients (15). NMR imagmg (MRI) techmques can produce pictures m which the visual contrast, i.e., difference m brightness between separate regions, is related to the NMR properties of the sample (i.e., spin-lattice, T,, and spm-spin, T,, relaxation times) and the concentratton of the nuclei bemg observed (NMR is chemically specific), and it depends on the parameters of the pulse sequence used for the imaging experiment (e.g., flip angle, echo time, repetttton time, and so on) The nondestructtve and nomnvasive nature of the MRI experiment (it does not require sectionmg or ultrahigh vacuum, and does not cause radiation damage, as does, for example, scanning electron microscopy) offers a unique opportunity to characterize samples prior to being used, and to study zn situ timedependent chemical and morphological processes. In this chapter, we describe the use of proton MRI to determine the porosity of btoabsorbable extracellular matrices of PHBHV copolymers prepared by solvent casting (10,11). The proton MRI method works by tmagmg the content of a fluid, m this case water, introduced into the pore space of the matrix (16,Z 7). For the apphcatton discussed here, this does not represent a problem, since the extracellular matrices are immersed m aqueous media at several stages during the preparation process, i.e., to remove the salt particles, and, often, prior to cell seeding or implantanon. The size of the features, e.g., pores, agglomerates, on so on, detected by the MRI method depends on the spatial resolution achieved m the Image. The acquisition of images with high spatial resolution, or small volume elements (voxels), requires long imaging times. This is because the mass of material giving NMR signal m a voxel decreases as the cube of the increase in linear resolution (mcreasmg resolution means voxels with smaller drmenstons) Since the signal-to-noise ratio obtainable from a fixed voxel is proportional to the square root of the time spent signal averaging, there 1s a sixth-power dependence between the linear resolution and the time required to obtain a fixed signal-to-noise ratio per image voxel Here, we present a two-steps MRI approach for efficiently screening extracellular matrices to assessthe umformtty m porosity and the pore structure First, a fast MRI protocol (total tmagmg time less than 30 min), for mapping gradual varrattons m porosity and macroscopic (> 100 pm) heterogeneities or defects m the matrix caused by salt agglomerates and nondissolved polymer particles, is used. Second, highresolution MR images, 200 x 40 x 40 pm, are acquired (total tmagmg ttme of 4.5 h) to illustrate the pore size m those matrices free of defects.

hodegradable

Porous Matrices

37

The correlation of the MRI findings with the set of condltlons used to prepare the matrices allows us to determine, m a reproducible manner, the values of experimental parameters that yield matrices with the desired mlcromorphology Although the specific details are not included in this chapter, the proton MRI methods described here can be used to characterize morphological changes m the bloabsorbable matrices zn situ during aging. In addition, m vitro cellular prohferatlon and m vivo tissue growth m the matrices can be momtored with MRI. 2. Materials 2.1. Po/y(3-Hydroxybufyrate-co-3-Hydroxyvalerate) (PHBHV) Porous Matrices 1 Poly(3-hydroxybutyrate-co-3-hydroxyvalerate) with 24% hydroxyvalerate, mol wt average, M; 297,000 g/mol, mol wt number average, M; 104,000 g/moI, and a meltmg temperature, T,, of 112’C (Aldrich, Milwaukee, WI) 2 Chloroform Optima (Frsher Sclentlfic, Pittsburgh, PA) 3. Whatman No 5 filter paper (Fisher Scientific). 4 Methanol, ACS reagent (Fisher Scientific). 5 Sodmm acetate anhydrous, 99%, cell culture reagent (Sigma, St. Louis, MO) 6 Carbon tetrachlonde, 99.9%, ACS reagent (Aldrtch) 7 USA standard testmg sieves, ASTM-l 1 specification (Fisher Sclentlfic) with openings 180 pm (No 80) and 2 12 pm (No 70) 8 Phosphate-buffered salme (PBS) 138 mMNaC1,2 7 mMKC1, 10 mMphosphate buffer (Sigma)

2.2. Nuclear Magnetic Resonance:

lnstrumen ta tion

1 2 0 T superconducting magnet (Nalorac, Martmez, CA), horizontal bore of 18 3 cm m diameter, operatmg at a proton frequency of 84.742 MHz, and Interfaced with a SIS spectrometer/lmager (Vanan, Palo Alto, CA, formerly SISCO, Fremont, CA). The system 1s equipped with Oxford Instruments (Osney Mead, UK) gradlent coils with id 11 cm and a maximum gradient amplitude of 12 g/cm m each of the three orthogonal axis 2 Home-built NMR probe conslstmg of a four-turns solenoIda ~011, 14 5 mm m diameter, 23 mm m length, tuned and matched to the proton frequency at 2 0 T, mounted on a Lucite@ tube (id 10 mm, od 14 mm), with Its axis perpendicular to the mam static magnetic field (E I du Pont de Nemours, Wllmmgton, DE) 3. The sample holder consists of a 10 mm od NMR tube, 6 5 cm long, with a fourlegged plastic Insert to hold the sampleIn the desiredposItIon (see Fig. 1)

2.3. Nuclear Magnetic Resonance: Pulse Sequence The pulse sequence used m the present experiments consists of a conventional two-dimensional multlsllce spin-echo acquisition, with short

38

Gamdo

A ---

CD

Fig 1 Sketch illustratmg the sample holder used to mamtam the matrix m the desired position during imaging (B) shows a cross-section of the holder, as mdtcated by dashed lme m (A)

echo time to mmlmtze attenuatton (16, I 7)

the effects of magnetic

susceptlbtllty-induced

signal

1 For low resolution images, the pulse sequence consists of two smc-shaped, 2 ms long, 90- and 180-degree radiofrequency (rf) pulses The echo time is 11 ms and the repetition time is 3 s The slice thickness is 1 mm, and gap between slices is 0 1 mm Total number of slices varies between 13 and 17 (sufficient to cover the entire sample) The digital resolution m the images is 128 X by 128 Y pixels, correspondmg to 86 X by 86 Y pm The total imaging time is 26 5 mm 2 For high-resolutton images, the same pulse sequence is used, with the followmg changes the length of the rf pulses is 5 ms, the echo time is 35 ms, the repetition time is 2 s, and the slice thickness is 200 pm, with no gap The digital resolution m the images is 256 X by 256 Y pixels, correspondmg to 39 X by 39 Y pm The total imaging time is 4 5 h mm

2.4. Image Processing

and Display

1 Spectrometer/tmager software, VNMR 2. NIH image, version 1.55, National Institutes of Health, Bethesda, MD

Biodegradable 3. Methods 3.1. Purification

39

Porous Matrices of PHBHV

Prior to being used for the preparation 24% HV, 1s purified (see Note 1)

of matrices, PHBHV

copolymer,

with

1 Dissolve 4 g of the copolymer m 200 mL of hot chloroform, m a Soxhlet apparatus, for 4 h under nitrogen atmosphere 2 Filter the solution to remove impurities 3 Precipitate the copolymer m the filtered solution with 300 mL of methanol 4. Filter and wash the preclpltate with cold (4°C) dHzO at least 2x 5 Dry the precipitate under vacuum to constant wetght for approx 48 h Store the copolymer m a desiccator under vacuum at room temperature (22°C) until use

3.2. Preparation

of Salt Particles

1 Sieve sodium acetate particles with sieve no 70 (nominal opemng of 2 12 pm) 2 Sieve the fraction havmg a particle size ~212 pm with sieve no. 80 (nominal opening of 180 pm), and collect the salt fraction retained

3.3. Preparation

of PHBHV Porous Matrices

PHBHV porous matrices are prepared by a solvent casting techmque described elsewhere (10,11) To obtain highly porous matrices (>90% porostty) with large interconnected pores, high concentration of salt particles (size -200 pm) m the final mixture 1s required, which leads to several problems The method was optimized to improve the mechanical stab&y of the porous matrix and to reduce the agglomeration and sedimentation of salt particles during casting (see Note 2) 1 To prepare the PHBHV porous matrices, dissolve 2 g of purified copolymer In 20 mL of a mixture of carbon tetrachlorlde (42% v/v) and chloroform (58% v/v) (see Note 2), and mix with 8 g of salt (sodmm acetate anhydrous, with particle sizeof196+ 16~) 2. Stir and cast the mixture on a polypropylene mold with dlmenslons of 5 2 x 4 2 x 1 cm

3 Cover with tissue paper to reduce the evaporation rate of the solvent Let the mixture dry for 72 h at room temperature. After drying, the thickness of the resultmg cake 1s -3 5 mm 4 Cut the cake m pieces with dimensions of approx 15 x 5 x 3.5 mm3, and Immerse the samples m dH,O (change every 12 h) at 37°C m a shaking water bath to remove the salt, until constant weight IS reached (3-5 d) 5 Store dry matrices m a desiccator under vacuum at room temperature until use

3.4. MR imaging of Water Content To map the internal empty volume of a matrix, using proton MRI, the sample must be filled with water. Thus, the slgnal intensity in the MR images repre-

Fig. 2. Multislice proton MR images corresponding to a PHBHV (24% HV) matrix acquired with the pulse sequence parameters described in Subheading 2.3.1. Regions of low NMR signal attributed to nondissolved polymer particles are observed in images 4, 5, 6, 9, 11, and 12. Also, large voids are observed in images 2,4, 11, and 12. The three bright spots surrounding each of the polymer particles is an artifact caused by the magnetic susceptibility difference in the water concentration. sents the concentration of water at each location in the matrix. The porosity at a specific location in the sample is given by the fractional image intensity at that location normalized to the intensity of pure water, which represents 100% porosity.

3.4.1. Detection of Defects 1. Place the sample with the plastic sample holder in the NMR tube and add PBS until the sample is fully immersed. 2. Fill the open space in the sample by using vacuum to remove the air and facilitate the penetration of water in the porous structure (see Note 3). 3. Place the NMR probe with the sample in the imager and proceed to acquire a lowresolution MR image, using the pulse sequence with the set of imaging parameters described in Subheading 2.3.1. 4. After the acquisition is completed, Fourier-transform the data using the imager software (see Note 4). Figure 2 shows an example of a multislice data set corre-

Biodegradable

Porous Matrices

41

Fig. 3. Detail of image #12 shown in Fig. 2, illustrating two classes of defects observed in PHBHV matrices prepared by the solvent casting method described in Subheading 3.3.

sponding to a PHBHV porous matrix. The presentation of the data in two dimensions, using a gray scale, allows visual discrimination of signal voids caused by nondissolved polymer particles and very large pores formed by trapped air or solvent vapor during casting (see Fig. 3). 5. At this stage, it can be decided if samples are accepted or discarded, depending on the size and distribution of the defects observed.

3.4.2. Measurement of Porosity After screening the samples for macroscopic defects, the quality of a sample is further assessedby measuring the total porosity and its variation across the sample. 1. Read image data with NIH Image software and smooth sharp variations in signal intensity (see Note 5). 2. To determine the open porosity of the sample, normalize the signal intensity in the image by the mean signal intensity measured in a region with only water (100% porosity). The information obtained is presented in two ways to facilitate the evaluation of this parameter. a. Histograms (see Note 6). After selecting a region of interest, defined by the sample contour, a histogram of this region is used to construct the distribution function for porosity (see Fig. 4). b. Surface plots (see Note 6). A surface plot allows easy visualization of gradual variations of porosity across the sample (see Fig. 5).

3.4.3. In-Plane Pore Size and Connectivity 1. High-resolution MR images are acquired using the pulse sequence with the set of imaging parameters described in Subheading 2.3.2.

42

Garrido 250

0

20

40

60

80

100

Open Porosity (%) Fig. 4. Histogram representation of the signal intensity data corresponding to a region of interest containing 95% of the total cross-section of the sample shown in image 10 (see Fig. 2). The horizontal axis is the signal intensity normalized to the mean intensity of the water distribution and, therefore, representsthe NMR-derived porosity distribution function.

10 (Fig. 2). The plot illustrates the gradual variation in porosity acrossthe sample.

Biodegradable

43

Porous Matrices

Fig. 6. High-resolution MR image (200 x 39 x 39 pm3 per voxel) of water in a PHBHV (24% HV) porous matrix acquired with the pulse sequence imaging parameters described in Subheading 2.3.2.

ediutn

Fig. 7. Surface plot representationof cross-sectionof the sampleshown in Fig. 6. The plot illustrates the distribution and connectivity of empty volume in an image plane with a thickness of 200 pm. 2. After Fourier transforms the raw data using the imager software (seeFig. 6), the imagesare processedand displayed using NIH Image software, as described in Subheading 3.4.2. (seeFig. 7).

Gamdo

4. Notes 1 Poly(hydroxybutyrate-hydroxyvalerate) copolymers are of bacterial origin. Generally, polymer samples as received from supplier contam cellular debris from synthesis, which can be removed by the treatment described 2 The agglomeration of salt particles was reduced using sodium acetate, instead of sodium chloride. To mmimtze the sedimentation of salt particles durmg casting, the den&y of the orgamc phase was adJusted to match that of the sodium acetate anhydrous (1 528 g/cm3) by mixing carbon tetrachloride and chloroform with densities of 1 594 and 1 472 g/cm3, respectively, m a ratio of 42.58 In addition, we found that the amount of polymer should be at least 20% of the total solids present m the mmal mixture, m order to obtam matrices dimensionally stable after removmg the salt 3 Instead of vacuum, the impregnation of the PHBHV matrices with water is faciletated by filling the samples first with a fluid having low mterfacial tension with the microscopic surfaces of the matrix, such as methanol or ethanol Then, the alcohol m the matrix is replaced by immersmg the sample m water, allowmg sufficient time for the exchange between the two fluids to occur The immersion of samples m alcohol does not add unnecessary steps to the preparation of porous matrices for cell implantation, since this IS a usual procedure for sterihzation of the matrices 4 The data-processmg software varies between NMR spectrometers, depending on the manufacturer, but the outcome should be independent of the mstrument used, assuming that the experimental conditions and processmg parameters used are the same Similarly, instead of the NIH Image software, other image processing software commercially available may be used to process and display the MRI data with comparable results, provided that the same algorithms and procedures are used 5 The smoothmg of the image data will reduce the artifacts caused by regional differences m magnetic susceptibtltty, and will allow a better visualization of the variations of signal mtensity attributed to the changes m the sample geometry 6 The width and shape of the porosity distribution function, as well as the percentage of regional variation m porosity with respect to the mean value of porosity may be used to set a criteria for quality control of the extracellular matrices

References 1. Langer, R and Vacanti, J (1993) Tissue engmeermg. Sczence 260,92&926 2 Vacanti, J P , Morse, M , Saltzman, M A , Domb, W M , Perez-Atayde, A , Langer, R (1988) Selective cell transplantation using bioabsorbable artificial polymers as matrices J Pedtatr Surg 23, 3-9 3 Mooney, D J , Coma, L , Langer, R , Johnson, L , Hansen, L K , Ingber, D E , Vacanti, J P (1992) Prmciples of tissue engineering and reconstruction using polymer-cell constructs Mater Res Sot Symp Proc 252, 345-352 4 Coma, L G , Ingber, D E , Vacanti, J P , and Langer, R (1991) Hepatocyte culture on biodegradable polymeric substrates Blotechnol Bzoeng 38, 145-l 58

Biodegradable

Porous Matrices

45

5 James, K and Kohn, J (1996) New biomaterials for tissue engmeermg MRS Bull 21,22-26 6 Lu, L and Mikes, A G (1996) The importance of new processing techniques m tissue engmeermg. MRS Bulletzn 21, 28-32. 7 Ripamonti, U. and Duneas, N (1996) Tissue engineering of bone by osteomductlve biomaterials MRS Bull 21, 36-39 8 Dunn, M G (1996) Tissue-engineering strategies for ligament reconstruction MRS Bull 21,4346 9 Borkenhagen, M and Aebischer, P (1996) Tissue-engineering approaches for central and peripheral nervous-system regeneration MRS Bull 21, 594 1 10 Mikes, A G , Wald, H , Sarakmos, G , Leite, S , and Langer, R (1992) Biodegradable cell transplantation devices for tissue regeneration Mat Res Sot Symp

Proc 252,353-358. 11 Mikes, A G , Thorsen, A J , Czerwonka, L A, Bao, Y., Langer, R , Wmslow, D N., Vacanti, J P. (1994) Preparation and characterization of poly(L-lactic acid) foams Polymer 35, 1068-1077 12 Holland, S. J , Jolly, A M., and Tlghe, B J (1987) Polymers for biodegradable medical devices II Hydroxybutyrate-hydroxyvalerate copolymers* hydrolytic degradation studies Bromatertals 8, 289-295 13 Mikes, A. G , Sarakmos, G , Lyman, M D , Ingber, D E , Vacant], J P , and Langer, R (1993) Prevasculanzation of porous btodegradable polymers Bzotechnol Btoeng 42, 716-723 14. Rotem, A, Toner, M., Sangeeta, B , Foy, B. D., Tompkms, R. G , and Yarmush, M. L. (1994) Oxygen 1s a factor determining m vitro tissue assembly. effects on attachment and spreading of hepatocytes Btotechnol Btoeng 43, 654-660 15 Mansfield, P and Moms, P G (1982) NMR imaging m biomedicme, m Advances tn Magnetic Resonance (Waugh, J S , ed ) Suppl2. Academic, New York 16. Ackerman, J L , Gamdo, L , Ellmgson, W A , and Weyand, J D (1987) The use of nmr imagmg to measure porosity and binder distribution m green-state and partially smtered ceramics, in American Ceramtc Soctety Proceedtngs Jotnt Conference on Nondestructtve Testing of High-Performance Ceramics (Vary, A. and Snyder, J , eds.), American Ceramic Society, Boston, pp. 88-l 13 17 Ackerman, J. L., Garrido, L., Moore, J R , Pfleiderer, B., and Wu, Y (1992) Fluid and sohd state mri of biological and nonbiological ceramics, m Magnetic

Resonance Mtcroscopy Methods and Appltcattons tn Materials Sctence, Agrzculture and Bzomedzctne (Blumtch, B. and Kuhn, W , eds ), VCH, Wemheim, Germany, pp 237-260.

4 Fabrication of Biodegradable Polymer Foams for Cell Transplantation and Tissue Engineering Peter X. Ma and Robert Langer 1. Introduction Organ transplantation has been successful since the early 1960s as a result of the success m lmmunologlc suppression m the clmlcal setting (I), and has saved, and 1s contmumg to save, countless lives, but IS far from a perfect solution to tissue losses or organ failures By far the most serious problem facing transplantation IS donor scarcity. Approximately 30,000 Americans need liver transplantation each year, but only about 10% of the patients have the chance to receive a donated liver transplant (2). There is a total of approx 100,000 people m the Umted States with transplants, but there are more than 1 mllhon with blomedlcal implants (3). Tissue engmeermg and cell transplantation are fields emerging to resolve the mlssmg tissue and organ problems. There are three strategies m cell transplantation and tissue engineering* 1 The use of isolated cells or cell substltutes to replace those cells that supply the needed function, mcludmg various mampulatlons before the cell mfuslon (#), 2 The dellvery of tissue-Inducing substances, such as growth factors, to targeted locatlons (S), 3 Growmg cells on three-dImensIona matrlces or devices (2) The advantage of the use of isolated cells IS to avoid surgical compllcatlons The disadvantages include cell death and loss of function because many cells are substrate-dependent, 1 e , they can grow and function only when they are attached on an appropriate substrate. The uses of tissue-mducmg substances are dependent on the development of delivery systems that can dehver bloactlve molecules and the economical large-scale production of the bloactlve agents Therefore, there has been very active research aimed at the regeneration of new tissues From

Methods

In Molecular

Edited by J R Morgan

Medune,

Vol

18

and M L Yarmush

47

Tissue

Engmeenng

0 Humana

Methods

and

Press Inc , Totowa,

Protocols

NJ

Ma and Langer and organs by culturing mammalian cells on three dimensional matrices or devices (2). Some of these devices work as an immunoprotective barrier, which allows nutrients and metabolic wastes to diffuse through, but not antibodies or immune cells. The core of this technology is a semipermeable membrane with a well-defined mol wt cutoff or size cutoff (6,7), These technologies are leading to the development of immobilized cell systems, either as implants or extracorporeal devices (6,840). Some other three-dimensional matrices work as templates to guide cells to grow, to synthesize biologically functional molecules and extracellular matrices in three dimensions (II, 12). The unique feature in this approach is the control over cell adhesion and function by choosing or developing the right substrate materials, and the control over new tissue or organ structure by scaffold design and processing. Matrix materials can be further classified as natural materials or synthetic materials. They can also be classified as degradable or nondegradable. The most frequently used matrix materials are either natural macromolecules, such as extracellular matrix components and polysaccharides (13,14); engineered macromolecules from natural origins, such as collagen-glycosaminoglycan copolymers (15); and synthetic polymers, either degradable (11,12) or nondegradable (16). However, nondegradable materials carry a risk of infection and permanent connective-tissue reaction (17). Natural extracellular matrix materials provide the advantage of cellular recognition. However, there is less control over their mechanical properties, biodegradability, and batch-to-batch consistency. They might exhibit immunogenicity (IS). Many of them are also limited in supply and can therefore be costly. Synthetic biodegradable polymers offer control over structure and properties such as chemical and physical structure, crystallinity, hydrophobicity, degradation rate, and mechanical properties. They can be processed into various shapes and microstructures, such as desired surface area, porosity, pore size, and pore size distribution. Their surface properties can be altered to adapt to the biologic requirements for cell adhesion, growth, and function. Therefore, synthetic biodegradable polymers have been widely used as vehicles for cell transplantation and templates for tissue engineering (2). There are several basic requirements for biodegradable polymer scaffolds. They need to be highly porous, with the proper pore size distribution to give cells room to grow, and paths for nutrients and metabolites to permeate. They should have a high surface area (surface:volume ratio) to give cells sufficient surface to adhere. They should have the right degradation rate to match the tissue regeneration process. They should have the structural integrity and mechanical strength to maintain the desired shape before the new tissue takes over. They should be biocompatable and their degradation products should not be cytotoxic. Their surface properties should be adaptable to the requirements for cell attachment, growth, and function.

Biodegradabie

Polymer

49

Foams

Various biodegradable polymers have been processed into scaffolds to engineer new tissues and organs (11,12,29-23). Biodegradable foam processing via salt-leaching technology has been reported earlier (II). The foams have also been used in bone- and liver-tissue engineering (2425). The fabrication processes have been further modified, improved, and standardized to better satisfy scaffolding requirements, multi-specimen preparation, and reproducibility. The poly(L-lactic acid) (PLLA) foams fabricated using the improved procedure are currently being used for liver tissue engineering (26-28). In this chapter, the improved biodegradable polymer foam processing will be described in detail, using PLLA as an example. However, the method should be of general use for other biodegradable polymer or nondegradable polymers.

2. Materials 2.1. Salt Crystals 1. NaCl crystals: Mallinckrodt, Paris, KY (see Note 3). 2. Grinder: Scienceware Micro-Mill Grinder, Fisher Scientific,

Pittsburgh, PA.

2.2. Polymer Solution 1. Polymer: poly(L-lactic acid) with an inherent viscosity of approx 1.6, Boehringer Ingelheim, Ingelheim, Germany (see Notes 1 and 5). 2. Solvent: chloroform, EM Science, Gibbstown, NJ. 3. 20-mL glass scintillation vials: Cole-Parmer Instrument, Niles, IL.

2.3. Surface-Coating

Solution

1. Poly(viny1 alcohol) (PVA) partially hydrolyzed from poly(viny1 acetate): 75% hydrolyzed, average mol wt 3,000, Aldrich, Milwaukee, WI (see Note 6). 2. Membrane filter: Schleicher and Schuell (Keene, NH) nylon membrane filters, pore size 0.45 pm, diameter 47 mm, VWR Scientific, Boston, MA.

2.4. Fabrication 1. Clean room ionizing air gun: model M-l 205wc, Terra Universal, Anaheim, CA. 2. Teflon PFA vials: 18 mm id, 5 mL, Cole-Parmer. 3. Falcon polystyrene tissue culture dishes: 35 x 10 and 60 x 15 mm, Becton Dickinson, Lincoln Park, NJ. 4. Blotting paper: Gel Blot Paper, size 20 x 20 cm, Schleicher and Schuell.

2.5. Sterilization

and Packing

1. Ethylene oxide sterilization agent: Anprolene gaseous sterilant, H. W. Andersen, Haw River, NC. 2.. Sterilization pouches: Chex-all II instant sealing pouch, Propper Manufacturing, Long Island, NY. 3. Aluminum sealing pouch: Kapak Pouches, 6.5 x 8 in., Kapak, Minneapolis, MN.

50

Ma and Langer

3. Methods 3.1. Salf Particles NaCl crystalsare ground into smallerparticlesusing amtcro-ml11gnnder (seeNote 3). The mixture ofthe groundparticlesISseparatedwith sievestntodifferent sizeranges The saltparticlesof the destredstzerangeare collectedandare sealedwith parafilm tn a glassbottle to prevent moisture adsorptionandparticle aggregation. 3.2. Polymer Solution The biodegradable polymers, contamed m glass bottles and sealed with parafilm, are stored m freezers (C-20°C) to prevent hydrolysis. They are taken out and put on the lab bench for at least I h to equilibrate with ambient temperature, before opening the bottles to prevent water from condensing on the polymers, which could lead to weight change (water adsorption) and molecular weight decrease (hydrolysis). The PLLA is weighed accurately mto a scmtillatton vial. In a fume hood, the calculated amount of chloroform IS added mto the glass vial to make a 5% solution (w/v) (see Note 6) The vials are kept on an orbital shaker (approx 100 rpm) at 25°C and are vortexed from time to time to facilrtate polymer drssolution. It takes about 2 d to dissolve the polymer well 3.3. Preparation of Surface Coating Solution PVA is weighed mto a 500-mL bottle with deionized water, and stirred with a magnetic stirrer on a stnrmg hot plate, to make a 1% (w/v) solution The heater is periodically switched between off and low to speed up the dissolution. However, the temperature is kept lower than 60°C. High temperature can cause decomposttton and color change (yellow or brown) It takes about 2 d to dissolve the PVA well The PVA solution is then filtered through a nylon membrane filter (pore size 0.45 pm), using a regular lab vacuum filtration setup 3.4. Fabrication of Foam The whole process to fabricate one batch of samples takes approx 2 wk after the polymer solutions and salt particles have been prepared. A schematic illustration of the fabrication procedure IS shown m Fig. 1. The detailed fabrication processes are described below. 3.4.1 Preparatron Teflon vials are washed with solvent to remove residual polymer and stuck salt particles. Dry vials are blown with high-pressure an to clean away loose dust. An ionizing an gun is connected with a nitrogen gas tank. The gas pressure is set at 10 psi with a gas regulator. The vials are ionized thoroughly by aiming and circling the nozzle at the entire surface area of both the inside and outside of each vial, to eliminate static.

Biodegradable

Polymer Foams

c

Solutlon

51

-

c Drying

/

Casting

Evaporation

Leaching

n;;+;q;;;; Salt Particles

Fig 1 SchematIc lllustratlon (not proportlonal) of basic fabrxatlon processesof biodegradablepolymer foam with salt-leaching techmque

3 4 2. Castmg 0.400 f 0.005 g of salt of the desired size are added mto each Teflon vial (see Note 3). The salt surface is leveled by first tapping the vial on edge to roughly spread out salt, then pressing it down onto a thm plate adapted on top of a vortex shaker. The vibration setting of the vortex shaker should be adjusted to shake mdlvldual grams of salt without disturbing already leveled salt. A flat metal plate with three adjustable legs (homemade) IS set m a fume hood, and 1s leveled with a lab level The vials are transferred carefully onto the leveled metal plate and are placed as far apart from each other as possible. 0.24 mL prepared PLLA solution IS plpeted onto the salt In order to evenly dlstrlbute the polymer solution m the salt, it IS better to drip a few drops equally apart on the salt instead of one large drop (five drops usually work fine, one m the center and four at corners of an imaginary square); or allow the polymer solution to drip slowly onto salt, using the free hand to rotate the vial, so that the solution covers nearly all surface area. The polymer solution should not drip when transferring from the stock solution to the vials, nor should the solutton drip down sides of vials

3.4.3. Evaporation The vials are covered with their caps to slowly evaporate the solvent for 24 h. The covers are removed carefully, without disturbing the salt and polymer dlstnbutlon The solvent continues to evaporate from the mixture of polymer solution and salt m the fume food for another 24 h. The metal plate should be covered to prevent foreign matter from settling on the polymer salt mixture (Fig. 2)

3.4.4 Leaching The polymer-salt composite disks are removed from the vials onto a clean sheet of paper. Handling carefully with forceps, the disks are placed

52

Ma and Langer Cover

Leveled lMeta1 Plate Frg 2 Second-stage evaporation setup mto 35 x 10 mm culture dishes (one piece per dish), and five dishes are banded together with elastic bands. An 1-L beaker is filled with drstrlled water. Holding a banded group of dishes under the water surface with gloves on, each dish IS opened slightly to allow air bubbles to escape. The foams at the edges should not be squeezed by the covers. Then the banded dishes are allowed to smk to the bottom. When the beaker 1s full, a clean glass plate is put on top of the dishes to prevent floating and all the dishes are submerged m the water The beaker IS placed on an orbital shaker and the shaker is turned on (setting roughly at 75 rpm) The foams are leached for 2 d, and the water IS changed 3x/d. The foams are taken out one by one (handling lightly and carefully) and placed on blottmg paper, smooth side down. The paper absorbs water, then the foams are placed m 60 x 15-mm culture dashes, rough side down (five pieces per dash) The covers are placed on, and au drying under a hood occurs for 24 h (see Note 4).

3 4.5. Surface

Coatmg

The dashes contammg the half-dry foams (rough srde up) are placed on a flat surface. 5 mL 1% PVA solutton IS slowly added to each dash by ptpetmg over the foams (see Note 6). No piece of foam should be allowed to float over others or to overlap. The dish IS covered and coatmg occurs for 24 h without drsturbmg. 3.4.6.

Drying

The PVA-coated foams are removed and placed on blotting paper, smooth side down. The paper IS allowed to absorb the excess PVA solutton. The foams are placed m clean 60 x 15-mm culture dishes, rough stde down (five pteces a dish) The dishes are covered and an-drying occurs for 24 h. The dishes are banded with elastic bands and transferred m freeze-drying flasks. The foams are lyophrlrzed for 72 h. Porous PVA-filtrated PLLA disks, l-mm thrck and 18 mm m drameter, with approx 95% porosity, are obtained (see Notes 2-4).

3.5. Sterilization

and Packaging

Defective foams are discarded. The remaining foams are rearranged m dishes, based on convenience for future use (such as five m a dash), and are

Biodegradable

Polymer Foams

covered. The foams (m their dishes) are sealed in gas-permeable sterlhzatlon pouches (Chex-all II instant-sealing pouches). The pouches are placed m a plastic liner bag. A plastic gas-release bag containing one ethylene oxide ampule 1s unrolled and placed m the liner bag alongside the seam. Excess air IS pressed out of the liner bag and the bag 1s tied shut with a twist-tie. The ampule is broken inside the bag and the bag 1splaced in a sterlhzation metal box under a fume hood. The lock hd to the box is closed; ethylene oxide sterilizes the foams m the bag for 24 h. The bags are untied and the contents are shaken into the sterilization box; ethylene oxide is allowed to diffuse away for 24 h. All materials are left m the hood until the end of the sterlllzatlon period. Caution: The operator should have protective gloves and a face mask on while handling the sterlhzatlon agent. The sterilized foams (m the gas-permeable pouches) are placed on a sheet of dry filter paper mto a Kapak alummum sealing pouch, and the pouches are thermally sealed with an Iron or a pouch sealer to prevent moisture adsorption and degradation of the fabricated biodegradable polymer foams

4. Notes 1 The methods described are of general use for fabricating porous structures, either from degradable polymers or from nondegradable polymers, as long as the polymers can be dissolved m a good solvent, but not m water Poly(lactlc acid), poly(glycohc acid), and their copolymers are a series of polymers that are approved by FDA for human use as suture materials and some Implantable devices. They are considered biocompatable, with mmimal tissue reaction (29,30) This chapter IS almed at the need for cell transplantation devices and porous scaffolds for tissue engineering, therefore, a blocompatable and bladegradable polymer PLLA is used as an example, and the sterlhzation and packagmg procedures are included With these techniques, highly porous foams can be reproducibly fabricated, and multiple samples can be fabricated m one batch These techniques do not require comphcated equipment or large lab space, and therefore are useful for research sample preparations These methods can be eas11y extended to large scale productions The methods can also be extended for melt processmg, as well 2. The geometry and size of porous devices can also be altered to satisfy specific needs In this chapter, a cylmdrlcal vial IS used as an example to make round porous disks In order to fabricate a device of a particular size and shape, a mold of that particular size and shape 1s required. Lamination techmques can be used to fabricate very thick devices by use of many layers of thin-contour, shaped porous sheets (31) The lammatlon techmque also permits the fabrication of more complex scaffolds for complex tissue and organ engmeermg 3. The porosity, pore shape, pore size, and pore-size dlstrlbutlon can be controlled with the salt*polymer ratio, particle shape, particle size, and size dlstrlbutlon For example, mcreasmg the initial salVpolymer ratio can increase the porosity of the

54

Ma and Langer foams to be fabrtcated Increasing the salt parttcle size can mcrease the pore stze of the foam (12) NaCl 1s easily available, btocompatable and economtcal, and therefore 1s used m this example However, other salts and substances can also be used. The porosity, surface volume ratio, pore size, and pore-stze dtstrtbutton can be measured wtth mercury-mtruston porosimetry The mtcrostructure of the devices can also be manipulated by various treatments used in polymer science and engmeermg For example, heat treatment can be used to control the crystallmtty of the porous polymer foams (II), and thereby then mechanical properties and even degradation rate It 1sknown that water has easier access to the amorphous regions Therefore, polymer chains m the amorphous region are believed to degrade first (22) By mantpulatmg the crystallmtty wtth thermal treatment, degradation rate and mechamcal properties can be mampulated to a certain extent However, m order to drasttcally change the properttes, such as degradatton rate, tt would be wtse to modify the chemical structure or to choose a different polymer For example, the degradation rate can be changed an order of magnitude by Just changing from poly(glycolic acid) to poly(L-lactic acid) A whole range of degradation rates can be achieved by choosmg poly(glycohc acid-co-lactic acid) with varying monomer ratios There are also other synthetic biodegradable polymers, such as polyanhydrtdes (32,33) and poly(ortho ethers) (34)), to choose from There are endless posstbtlmes m modtfymg extstmg polymers and synthestzmg new polymers There are countless chotces for surface coatmg materials PVA IS used here as a hydrophthc coatmg for easier filtration of cell suspensions, for cell loading m cell transplantatton, and cell seeding m tissue engineering There are many other hydrophtltc materials that can be used for the same purpose. There are also biologtcally active components, such as peptides, proteins, and growth factors, that can be potentially attached on the surfaces to control cell attachment and dtfferenttation.

References Couch, N , Wtlson, R , Hager, E., and Murray, J (1966) Transplantatton of cadaver kidneys experience with 2 1 cases Surgery 59, 183-l 88 Langer, R. and Vacantt, J (1993) Tissue engineering Sczence 260, 92&926 Nerem, R M and Sambarns, A (1995) Ttssue engineering from biology to btologtcal substitutes &sue Eng 1, 3-13 Brtttberg, M , Lmdahl, A , Ntlsson, A , Ohlsson, C , Isaksson, O., and Peterson, L. (1994) Treatment of deep carttlage defects m the knee wtth autologous chondrocyte transplantation N Engl J Med 331,889-895 Aebtscher, P , Salesstotts, A , and Wmn, S (1989) Basic fibroblast growth factor released from synthetic guidance channels facilitates peripheral nerve regeneration across long nerve gaps. J Neurom Res 23,282-289 Colton, C (1995) Implantable btohybtld arttficial organs Cell Transplant 4,4 l-36

Biodegradable

Polymer

Foams

55

7 Darquy, S and Reach, G (1985) Immunotsolatlon of pancreatic B cells by mtcroencapsulation An m vitro study Dlabetologla 28, 776780 8 Nyberg, S , Shn-abe, K , Peshwa, M , Stelaff, T , Crotty, P , Mann, H , et al (1993) Extracorporeal apphcatton of a gel-entrapment, broartttictal hver demonstratron of drug metabohsm and other blochemtcal functions Cell Transplant 2,441452 9 Reach, G (1993) Btoarttfictal pancreas Dzabet Med 10, 105-109 10 Colton, C and Avgoustmlatos, E (1991) Bloengmeermg m development of the hybrtd arttfictal pancreas J Bzomech Eng 113, 152-170 11 Mtkos, A G , Thorsen, A J , Czerwonka, L A , Bao, Y , Langer, R , Winslow, D N , and Vacantt, J P (1994) Preparation and charactertzatlon ofpoly(l-lacttc acid) foams Polymer 35, 1068-1077 12 Ma, P X and Langer, R (1995) Degradation, structure and properttes of fibrous nonwoven poly(glycohc acid) scaffolds for tissue engmeermg, m Polymers zn Medune and Pharmacy (Mtkos, A G , Leong, K W , Radomsky, M L , Tamada, J A , and Yaszemskt, M J , eds ), MRS, Pittsburgh, pp 99-104. 13 Bell, E , Rosenberg, M , Kemp, P , Gay, R , Green, G , Muthukumaran, N , and Nolte, C ( 199 1) Rectpes for reconstttutmg skin J Bzomech Eng 113, 113-l 19 14 Krewson (nee Beaty), C E , Chung, S W , Dai, W , and Saltzman, W M (1994) Cell aggregation and neurlte growth m gels of extracellular matrtx molecules Blotechnol Bloeng 43,555-562 15 Yannas, I V (1994) Apphcattons of ECM analogs m surgery J Cell Bzochem 56,188-191 16 Pongor, P , Betts, J , Muckle, D , and Bentley, G (1992) Woven carbon surface replacement m the knee Independent clinical review Bzomaterzals 13, 1070-l 076 17 Grtstma, A (1987) Btomatertal-centered mfectlon nncrobtal adheston versus ttssue mtegratton Science 237, 158881595 18 Coma, L , Vacantl, J , Vacanti, C , Ingber, D., Mooney, D , and Langer, R (199 1) Tissue engmeermg by cell transplantatron using degradable polymer substrates. J Bzomech Eng 113, 143-151 19 Ma, P X , Schloo, B , Mooney, D and Langer, R (1995) Development of btomechamcal properttes and morphogenesls of m vttro tissue engmeered cartilage. J Blamed Mater Res 29, 1587-1595 20 Freed, L. E , Marques, C J , Nohrta, A., Emmanual, J., Mlkos, A G , and Langer, R (1993) Neocarttlage formatron m vztro and m vlvo using cells cultured on synthetic biodegradable polymers .I Blamed Mater Res 27, 1 l-23 21 Vacant], C., Kim, W , Upton, J , Vacant], M , Mooney, D , Schloo, B , and Vacantt, J (1993) Tissue-engineered growth of bone and carttlage Transplantatzon Proc 25, 1019-1021 22 Organ, G , Mooney, D , Hansen, L , Schloo, B , and Vacant], J (1993) Enterocyte transplantanon usmg cell-polymer devices to create Intestinal eptthehal-lmed tubes Transplantation Proc 25,998-1001 23 Shmoka, T , Ma, P X , Shum-Tim, D., Breuer, C K , Cusrck, R A, Zund, et al ( 1996) Ttssue-engmeermg heart valves* autologous valve leaflet replacement study m a lamb model Czrculatzon 94 (Suppl.), 11-164-11-168

56

Ma and Langer

24 Thomson, R , Yaszemskl, M., Powers, J , and Mlkos, A (1995) Fabrication of biodegradable polymer scaffolds to engineer trabecular bone J Blomater Scz Polym Ed 7,23-38 25. Mooney, D., Park, S., Kaufmann, P., Sano, K., McNamara, K , Vacantl, J , and Langer, R (1995) Biodegradable sponges for hepatocyte transplantation J Blamed Mater Res 29,959-965 26 Cuslck, R A , Lee, H , Sano, K , Pollok, J M , Utsunomiya, H , Ma, P X , Langer, R , and Vacantl, J P (1997) The effect of donor and reclplent age on engraftment of tissue engineered liver J Pedlat Surg 32(2), 357-360 21 Lee, H , Cuslck, R A, Browne, F , Kim, T H , Ma, P X , Utsunomlya, H , Langer, R , and Vacanb, J P Increased anglogenesIs by local dellvery of bFGF increases survival of transplanted hepatocytes, to be published 28 Lee, H , Cuslck, R A , Utsunomlya, H , Ma, P X , Langer, R , and Vacantl, J P Effect of lmplantatlon site on hepatocytes heterotoplcally transplanted on blodegradable polymer scaffolds, to be published 29 Matlaga, B and Salthouse, T (1983) Ultrastructural observations of cells at the Interface of a biodegradable polymer Polyglactm 910 J Bzomed Mater Res 17, 185-197 30 Craig, P , Wllhams, J , Davis, K , Magoun, A , Levy, A, Bogdansky, S , and Jones, J J (1975) A blologlc comparison of polyglactm 910 and polyglycollc acid synthetic absorbable sutures Surg Gynecol Obstet 141, l-10 31 Mlkos, A , Sarakmos, G , Lelte, S , VacantI, J , and Langer, R (1993) Laminated three-dlmenslonal biodegradable foams for use m tissue engmeermg Blomaterlals 14,323-330 32 Leong, K , Brott, B , and Langer, R (1985) Bloerodlble polyanhydrldes as drugcarrier matrices I Charactenzatlon, degradation, and release characterlstlcs J Blamed Mater Res 19,941-955 33 Domb, A and Langer, R. (1987) Polyanhydndes. I Preparation of high molecular weight polyanhydrldes J Polymer hence 25, 3373-3386 34 Chow, N S and Heller, J (1978) US Patent 4,093,709

5 Formation of Highly Porous Polymeric Foams with Controlled Release Capability A Phase-Separation

Technique

S. Kadiyala, H. Lo, and K. W. Leong 1. Introduction The successof many tissue engmeermg apphcations depends on a scaffold with the suitable physical properties, one of which might be a macroporous structure that allows cellular ingrowth. Such a porous implant further raises the posstbthty of delivering chemotactic or growth factors to Influence the course of cell proliferation and differentiation in sztu. The scaffolds can also be preseeded with cells to accelerate tissue growth or repair. Even in the absence of these payloads, they still provide the benefit of introducing the mmlmal amount of foreign material into the tissue. Furthermore, by making the porous scaffold, or foam, from biodegradable polymers, the regenerated tissue would be rid of any synthetic component, leading to a more functtonal biological equivalent, and elimmatmg concerns of long-term tissue compatibibty. Because of the inherent hmitations of polymer strength and the requirement for high porosity, tt IS not feasible to use the foams m conditions in which they are subjected to large mechanical loads. However, it is important that the foams can maintain then- form under physiologic conditions and do not collapse prior to any tissue mgrowth. For cellular infiltration to occur, the pore sizes would have to be larger than the cells-actually, substantially larger, considering the effect of tortuostty on cell migratton. It is unlikely that pore sizesbelow 40 mm can support tissue mgrowth, though such foams may still be used to deliver growth factors or to act as tissue barriers. The optimal pore size range of the foam is probably tissue-dependent. For instance, tt has been hypothesized that, in orthotoptc sites, pore sizes below 400 mm lead to bone formatton and pore From

Methods Edlted

by

in Molecular Me&me, Vol 18 Tmue Engmeermg Methods and Protocols J R Morgan and M L Yarmush 0 Humana Press Inc , Totowa, NJ

57

58

Kadiyala, Lo, and Leong

sizes above 400 mm lead to fibrous tissue mgrowth (Z-4). In addition to pore size, porosity, which more reflects the mterconnectivity of the foam, IS also important. High porosity maxtmtzes the volume of tissue mgrowth and minimizes the amount of polymer used It also facilitates transport of nutrients and cellular waste products Another parameter 1s the pore morphology, whtch may be meamngful m favoring the mgrowth of certain cell types. There exist a large number of methods for producing a macroporous structure from polymers (5), many of whtch are developed for non-btological apphcations We have studied the technique of foam fabrication by phase separation from a homogenous polymer solution (6,7), a versatile method that appears capable of addressing the important issues of a desirable porous scaffold

1.1. Theory of Foam Fabrication

by Phase Separation

The foam fabrication technique is based on the phase separation of the polymer from a homogenous solutton m the spmodal region of the phase dtagram. As the polymer solution is cooled down to the soltdtfication temperature of the solvent, the homogenous polymer solution goes through a phase-separation region, called the spinodal region where two distinct hquid phases emerge a polymer phase and a solvent phase The stze of the phase domains depend on the temperature and the duration at which the solution IS maintained at that temperature (8-10). The closer the temperature to the homogenous region and the shorter the duration, the finer 1s the phase structure obtained. The morphology at any stage can be preserved by rapid quenching of the solution to prevent any coarsenmg of the microstructure. When the solvent is subsequently removed without disturbmg this morphology, a highly porous structure will remam. For our purpose, the key to success is that the solvent can be removed without aqueous washing, because that would leach out any embedded btoactive agents This can be accomplished by using a solvent that has a high vapor pressure at room temperature, and hence can be removed by sublimation. In addition to the factors of temperature and time mentioned, there are other factors that can be vaned to change the pore structure of the foam, mcludmg the solution concentration of the polymer, the mol wt of the polymer, the chemical cornpositron of the polymer, and the choice of the solvent. To predtct the mtcrostructure of the foam made from a particular polymer-solvent, detailed phase dtagrams for each system need to be constructed by studymg the phase behavror under different temperature, pressure, and anneahng penods In the absence of these phase diagrams, the foam microstructure can be opttmtzed empmcally, and a variety of foams can be fabricated by changmg some of the factors descrtbed above (6,7, II). The choice of solvents for this techmque of foam fabrication 1sbased on the followmg considerations. The polymer of interest should be soluble m the solvent, the melting point of the solvent should be low enough to mmrmize the

Highly Porous Polymeric Foams

59

thermal degradation of the polymer and the incorporated bioacttve agent; the solvent should have a low energy of sublimation, and its freezmg pomt should be high enough to allow ease of subhmatton; the solvent should not pose a toxictty rusk at the residual amounts left at the end of the fabrication process After an extensive search of most of the organic solvents, we have tested naphthalene and phenol, two solvents that satisfy most of these crtteria. The melting points of naphthalene and phenol are 80°C and 42°C respectively. Although the temperatures are higher than those that have been shown to denature many bioactive proteins m aqueous conditions, if the fabrication can be performed under anhydrous condttions, much btoacttvity, in prmciple, can still be retained (12,13). Thus, any protein added to the molten polymer needs to be m an anhydrous form, easily achieved, for instance, by lyophtlizatton. 2. Materials 1. 2 3 4

Naphthalene and phenol (99% pure) Aldrich, Milwaukee, (PLA) Zrmmer, Warsaw, IN Atomizer. General Glassblowmg, Richmond, CA. Mercury porostmetry Porous Materials, Ithaca, NY

WI

3. Methods The foam fabrication methodology will be described here with specific examples (6,7, II) to illustrate the points: 3.7. Method A: Foam Fabrication by Atomization (Naphthalene System) Chromatography sprayers are used to provide a fine, uniform spray The sprayer consists of an atomizer top attached with a screw top and 0 rmg to glass reservoir flasks. The sprayer is connected to a mtrogen gas tank with a pressure regulator. The spray pattern, which m turn influences the morphology of the foam, can be adjusted by the spraymg pressure and the coverage of the vent hole. In the followmg procedure, the vent hole is fully covered. Into a 125mL reservon flask contammg a magnetic stirrer, 57 g of solid anhydrous naphthalene was added, followed by addition of another 3 g of anhydrous (PLLA) (mol wt 500,000) over the naphthalene solid Note that the order of addttton will reduce the chance that the polymer adheres to the surface of the flask, thereby facilitating drssolutton of the polymer 2 Stopper the flask loosely Heat the flask containmg the sohd naphthalene and polymer to 90°C by a stir plate Stir the contents at 200 rpm for 20 mm, or until a homogeneous melt IS obtamed Wrapping the flask with a heating tape preheated at 90°C would ensure an uniform temperature of the molten mixture and facrhtate polymer drssolutron. If

Kadlyala, Lo, and Leong

60

necessary, add fresh naphthalene to compensate for any loss caused by sublimation and evaporation of naphthalene If a drug or any excipient IS to be mcorporated mto the foam, the agent should be added into the melt at this time Most hydrophilic drugs will not be soluble m the naphthalene melt Stir another 5-l 0 mm to achieve uniform suspension of drugs for each addition If decomposltion of material is noted (e g , yellowmg), the stock should be remade 3 Wrap the atomtzer top with a heating tape at a temperature of 90°C This is a crmcal step Allow plenty of time to heat the atomizer tip InsufIicient heatmg will cause deposition of naphthalene on the sprayer nozzle and clog the spray Should that happen, a heat gun can be used to dissolve the naphthalene 4 Cut a Teflon sheet to the desired shape for the deposition of naphthalene mixture The Teflon surface allows easy removal of samples followmg the spray Connect the heated atonuzer top to the reservoir flask. Connect warm mtrogen (4&5O”C) to the sprayer at a pressure 5 psig, and begin to spray by closmg the vent hole with thumb A disk of 3 cm m dtameter and 3 mm in thtckness can be obtained m less than a minute. 5 Place samples mto a vacuum oven heated to 50°C Apply vacuum overmght at less than 1 torr to remove the naphthalene

3.2. Method B: Foam Fabrication Method by Casting (Naphthalene

System)

Sample cups made of tmplate with slip-on covers (VWR Sctentrfic, Boston, MA) were used for casting naphthalene mixture 1, A stock solution of 5 wt% PLLA m naphthalene is prepared accordmg to step 1 of Method A 2 Heat a 1-oz tm cup to 90°C on a heat plate 3 Charge 10 g of stock solution mto the heated tm cup from step 2 Close the cover immediately 4 Immediately transfer the tm cup to liquid mtrogen It will only take less than 10 s to solidify the naphthalene melt 5 Place samples mto a vacuum oven heated to 50°C Apply vacuum overnight at less than 1 torr to remove the naphthalene

3.3. Method C: Foam Fabrication Method by Casting (Phenol System) The procedures are srmtlar to Method B, with the followmg excepttons The temperature of the stock solutton of the mold should be heated to 50°C instead of to 90°C. The subhmatton of phenol should be done at room temperature, instead of 50°C m the oven.

3.4. Other Considerations The prmcrple of foam fabrtcatton has been demonstratedby the work described m this review, but the relattonship between the thermodynamtcs of phase

Highly Porous Polymeric Foams

61

separation and the final porous architecture of the foam remains to be established. A detailed study of the phase diagram of a given polymer-solvent system will be required to allow prediction of the microstructure of the foam, based on the chosen parameters. Also, additional work IS required to evaluate the biocompatlblhty of the foams. One might argue that if a controlled-release function IS desired, one can inject drug-loaded mlcrospheres mto the foam. However, it might be difficult

to uniformly distribute the mlcrospheres m the foam and might compromise the goal of leaving the most volume for tissue mgrowth. Even not considering the merit of drug mcorporatlon, the phase-separation technique IS still attractive, compared to some of the other methods, m terms of slmphclty and versatility.

4. Notes Characterlzatlon

of Foams Two methods that can be utlhzed to characterize the

foams are mercury poroslmetry

and scanning electron microscopy

(SEM)

Mercury porosimetry can yield data on the porosity and pore-size dlstrtbution In tissue-engmeermg apphcatlons, only pore sizes above a certain threshold (10 pm) should be used to calculate the porosity, since pores below that threshold may not contribute to the healing response. SEM IS an effective tool to determine the pore morphology, onentatlon, and, to a limited extent, pore sizes

Modlficatlons

to the system. The effect of changes of mol wt of the polymer,

concentration of the polymer m the fabrication mix, the solvent used, and the polymer used are shown m Fig. lA-H. As can be seen in these SEM mlcrographs, the foam microstructure can be readily changed by changing any of the aforementioned parameters Change of foam morphology Fig. 1E illustrates the result of directional coolmg on the foam morphology To obtain this kmd of morphology, the polymer solution 1s poured mto a mold preheated to the temperature of the polymer solution, and coolmg 1s applied from one surface of the mold, while the other surfaces of the mold are kept insulated. Such an architecture would be useful m cases like tendon regeneration, m which the fibers are aligned m a longltudmal direction Inclusion of bioactlve moieties The bloactive agent to be included m the foam IS added to the polymer mix in the molten state As mentioned previously, when operating at high temperatures, it 1s essential that the system be maintained m an anhydrous state to retain the bloactlvlty of the protein. Sustained release of small molecules, such as bromothymol blue and sulphorhodamme (61, or large proteins, such as alkaline phosphatase (7) or bone morphogenetlc protein (BMP) (II), have been achieved The release kinetics ofthe embedded agent 1sgoverned by many parameters The hydrophoblclty/hydrophlllclty of the agent relative to the solvent and the polymer determines its partition or dissolution m the polymer

phase Hydrophilic

compounds are likely to just adsorb on the polymer phase of

the foam and will not be released m a sustained manner The loading level (wt% of agent m the polymer phase) determines the fraction of the agent that IS dls-


62

1OOpm

lmm

PLLA

5oO,M?O/ naphthalene

PLU

500,000/

naphthaiene

PLLA

500,000

/ naphthalene

PLLA

130,CNWpheno~

PUA

500,0001 naphthalens

5%

5%

PLLA

2,OOff nsphthalene

5%

1%

BPA-PP

10%

5%

1OOpm

lO$m -

50,000/

lO%HA~~l~er/

naphthalene

5%

phenol

5%

Fig. 1. Scanning electron micrographs of assorted biodegradable foams made by phase separation. Foams are identified by polymer type and mol wt, solvent used, and percentage of polymer in solvent. Foams A, B, C ,and G were fabricated by atomization; D, E, F, and H were fabricated by casting. Reproduced with permission from ref. 7.) solved in the polymer phase and the fraction that is adsorbed on the surface. A biphasic release profile can be expected in this situation, the first phase being release of the surface exposed drug, and the second phase controlled by diffusion and matrix degradation. Morphology of the foam-not so much the pore size and

High/y Porous Polymeric Foams

0 1'1'~/~~,',,,~~,'~'~,~'~',,~'~,~~~,l 50 loo 0

63

150 Time

200

250

300

350

(hours)

Ftg 2 (A) In vitro release curve of FITC-labeled alkaline phosphatase (FITC-AP) from 10 wt% FITC-AP-loaded PLLA foams (a-, % AP released) m 0 1 M pH 7 4 phosphate buffer at 37°C (upper curve) (B) Changes m btoactivtty of FITC-AP after bemg released from PLLA foams (-0-) with pure FITC-AP placed under the same condtttons as a control (lower curve) --@-, AP control Reproduced with permisston from Lo ref. 7)

porosity, but the thickness of the Individual cells, affects the diffusional distance for release Finally, degradation of the polymer may overshadow many of these factors tf the degradation rate 1s high The release profile of alkalme phosphatase from a PLA foam fabricated with a PLA-naphthalene system, along with assessment of the btoactivtty of the released alkaline phosphatase 1sshown m Fig. 2 The large burst and btphasic release seen m this case is probably caused by the 10% loading level, as explained above A sustained and steady release, however, can be obtained with a lower drug loadmg, as illustrated m the delivery of BMP (II).


64

2

18

35 220 Time (hours)

Ether

Wash

Fig. 3. Levels of residual naphthalene in the foams after sublimation and ether wash. 5. Biocompatibility issue. A major concern for this fabrication technique, as is the case with all other techniques that might require contact with organic solvents, is the residual levels of solvents that might influence the biocompatibility of the implant. The residual amounts of naphthalene in the foam as a function of the sublimation time are shown in Fig. 3, and translate to ug quantities in a typical implant. Moreover, the hydrophobicity of naphthalene would mean that it would be released only when the polymer degrades, over a long period of time. In vitro culturing of cells in the foam (6) and in vivo implantation (22) have revealed no acute problems of biocompatibility. However, long-term studies will need to be conducted. One way to reduce the amount of residual naphthalene might be to include a secondary vacuum exposure at a higher temperature and higher vacuum. Naturally, this may not be advisable if delicate bioactive agents are embedded in the foam.

References 1. Ducheyne, P. (1985) Success of prosthetic devices fixed by ingrowth or surface interaction. Acta Orthop. Belg. 51, 144-161. 2. Schliephake, H., Neukam, F. W., and Klosa, D. (1991) Influence of pore dimensions on bone ingrowth into porous hydroxyapatite blocks used as bone graft substitutes. A histometric study. Int. J. Oral. Maxillofac. Surg. 20, 53-58. 3. Eggli, P. S., Muller, W., and Schenk, R. K. (1988) Porous hydroxyapatite and tricalcium phosphate cylinders with two different pore size ranges implanted in the cancellous bone of rabbits. A comparative histomorphometric and histologic study of bony ingrowth and implant substitution. Clin. Orthop. Rel. Res. 232, 127-138. 4. Collier, J. P., Mayor, M. B., Chae, J. C., Surprenant, V. A., Surprenant, H. P., and Dauphinais, L. A. (1988) Macroscopic and microscopic evidence of prosthetic fixation with porous-coated materials. Clin. Orthop. 235, 173-180.

High/y Porous Polymeric Foams

65

5. Kadiyala, S., Lo, H., and Leong K. W. (1994) Biodegradable polymers as synthetic bone grafts, in Bone Formation and Repair (Brighton, C. T., Friedlander, G., Lane, J. M., eds.), AAOS, Boca Raton, FL. 6. Lo, H., Kadiyala, S., Guggio, S. E., and Leong, K. W. (1996) Poly(L-lactic acid) foams with cell seeding and controlled-release capacity. J. Biomed. Mater. Res. 30,475-484. 7. Lo, H., Ponticiello, M. S., and Leong, K. W. (1996) Fabrication of controlled release biodegradable foams by phase separation. Tissue Eng. 1, 15-28. 8. Tanaka, H. and Nishi, T. (1987) Direct determination of the probability determination of concentration in polymer mixtures undergoing phase separation. P&s. Rev. Lett. 59, 692495. 9. Siggia, E. D. (1979) Late stages of spinodal decomposition in binary mixtures. Phys. Rev. A20,595-605. 10. Aubert, J. H. (1988) Interfacial tension of demixed polymer solutions. Polymer. 29, 118-122 1 1. Kadiyala, S., Lo, H., Ponticiello, M. S., Reddi, A. H., and Leong, K. W. (1996) Bone induction achieved by controlled release of BMP from PLA/Hydroxyapatite foams, in Transactions of the Fifth World Biomaterials Congress, Toronto, p. 289. 12. Zaks, A. and Klibanov, A. M. ( 1984) Enzymatic catalysis in organic media at 100 degrees C. Science 1249-25 1. 13. Zaks, A. and Klibanov, A. M. (1988) Enzymatic catalysis in nonaqueous solvents. J. Biol. Chem. 263.3 1943201.

Magnetic-induced Alignment of Collagen Fibrils in Tissue Equivalents Timothy

S. Girton, Narendra

Dubey, and Robert T. Tranquillo

1. Introduction The use of reconstttuted type I collagen gel as a scaffold for engmeered soft ttssues IS a htghly attractrve prospect, given that collagen IS the prmctpal component of the extracellular matrix (ECM) m VIVO,providing a mechamcally suttable and mformatton-rtch scaffold for cell-ECM mteractrons. It has the advantage that cells can be directly entrapped wtthm the comprlsmg collagen fibrrls as they grow mto an entangled network from a cell contammg solution of monomertc type I collagen. These tissue equivalents have the further advantage that the collagen fibrtls can be aligned by applying a magnetic field during fibrillogenesrs Then, through a process termed “contact guidance,” the cells align with the fibrtls by drrectmg then- motthty. Such alignment is charactertstic of many trssues,and may provide microstructural and mechamcal cues for regulation of cell phenotype and functton, as well as influence the gross mechanical properties of the ttssues.Recent research m our laboratory has used magnetic-induced alignment m the fabrtcatton of tissue-equivalents, notably cncumferenttal alignment m tubes, and longttudmal alignment m rods (patent pending). The former 1s armed at reproducmg the architecture of the arterial media; the latter IS aimed at providing a bridge that promotes directed axonal growth between severed nerve ends These tissue engineering apphcatrons exploit the finding of Torbet and Ronztere (I) m then cell-free studies that forming fibrrls tend to align in the plane normal to the direction of the field (because of the negative diamagnetic amsotropy of collagen) and parallel to the mold surfaces (because of an uncharacterized mterfactal effect). The eventual entanglement with other growing frbrrls mechanically stabtlrzes the From

Methods III Molecular Me&one, Vol 18 Tissue Engrneermg Methods and Edlted by J R Morgan and M L Yarmush 0 Humana Press Inc , Totowa,

67

Protocols NJ

Girton, Dubey, and Tranquillo magnetic-induced altgnment, m that the alignment is sustained after the formed network is removed from the magnetic field. There is srgnificant potential for using magnetic processmg of type I collagen for engmeermg the ahgnment m tissue equivalents. It is simple: The mold is merely placed with appropriate orientation m the magnetic field during fibrillogenesrs; it is reproducible, as documented m our previous studies of axially aligned tissue-equivalent slabs (2), and it 1scontrollable, via the rate of fibrillogenests (e g., via temperature) and the magnetrc field strength (2) The method is suttable for mass production, since it is essentially scale-mdependent, the alignmg effect of the magnetic field being mdependent of the mold dimensrons, except for the gap size, which must be small, so that the mterfacial effect, which selects the alignment direction from the plane, IS stgmticant. In practice, “small” 1s on the order of a few mtlhmeters, which, fortunately, 1s suitable for many tissue-engmeermg apphcations. The only practical method reported outside of our group for rmposmg ahgnment m tissue equivalents is stretch-induced alignment (3). Other methods reported for (acellular) collagen gel that might be adapted include ordered convection, driven by a collagen concentration gradient (4), or temperature gradtent (5) and weak electric currents (6), however, the fibrtl alignment m these methods was very nonuniform Although unnecessary for explortmg magnetic-field-induced collagen fibrll alignment, knowledge of the mechanism is of fundamental interest. A collagen molecule exhibits negative diamagnetic amsotropy primarily because of its peptide bonds In a magnetic field, a molecule possessmg a negative diamagnetic amsotropy experiences a torque that tends to orient the molecule normal to the field direction. However, the ratio of ahgnrng energy to randomrzmg thermal energy (Browman motion) is only 5 x 10m3for a single collagen molecule at room temperature m the highest magnetic field available (about 20 T). It has been proposed that this ratio should exceed 6, in order to have more than 75% of the maximum magnetic ahgnment (1). It was therefore concluded that the high degree of alignment observed is a result of the summation of diamagnetic amsotropies of the collagen molecules (monomers) that comprise the fibril (polymer). For a rotationally symmetric, rigid polymer comprised of monomers whose symmetry axes are parallel, the diamagnetic amsotropy of the polymer can be considered to a good approxrmation as the sum of the diamagnetic amsotropies of the monomers. Thus, for such monomers that selfassemble in a linear fashion (fibrm, actm, and tubulm being other blopolymer examples), the additivrty of the diamagnetic amsotropres causesthe polymer to experience an aligning torque in the presenceof a strong magnettc field However, once extensive fibrillogenesis has occurred, the fibrils are too constrained within the entangled network to be further aligned. Alignment, therefore, is

Ahgnment of Collagen Fibrils

69

restricted to a window of time during fibrillogenesis m which the growing tibrils are sufficiently large to experience a torque able to overcome Browman motion, yet small enough that they are not too constrained by neighboring Iibrtls. This time-window can thus be extended by conditions that slow the rate of fibrillogenests, such as decreased temperature. In this chapter, we detail our methods for generating magnetic-induced alignment m tissue equivalents, in particular, circumferentially aligned tubes and longttudmally aligned rods. We dtscussthe preparatton of the collagen gel, generation of the required geometry, and opttmization and control of the desired ahgnment. 2. Materials 2.1 Monomeric

Collagen Solution

1 2. 3 4 5 6

Vitrogen 100 monomeric type I collagen (Collagen, Palo Alto, CA) (see Note 1) 1 M HEPES buffer solution (Gibco-BRL, Grand Island, N Y ) 0 1 Nendotoxm free NaOH solution (Sigma, St LOUIS, MO.) 1OX mmimum essential medium (MEM) (Gibco-BRL) Fetal bovme serum (FBS) (HyClone, Logan, UT) Pemcillm-streptomycm solution (5000 U of pemcillm and 5000 ~18streptomycm/mL m 0 85% saline solution) (Gibco-BRL) 7 L-Glutamine solution 29 2 mg/mL of L-glutamine m 0.85% saline solution (Gibco-BRL).

2.2. Hardware 1 Molds (comprised entirely of autoclavable parts) a. Tubes an 8-mm diameter Teflon rod, two 12-mm diameter 4-mm-tall rubber cylmders, and a 12-mm id clear polypropylene tube (Fig. 1A) b Rods. a 3-mm id, 2-cm-long Teflon tube, and two 3-mm id, j-mm-long Teflon caps (Fig. 1B) 2 ~-CC syringe with 2 1-gage needle 3 2 1-gage needle 4 A horizontal bore (preferably 6-cm diameter or larger), high-strength (preferably 4 T or higher ) magnet (see Note 2). 5 A watertight acryhc tube, whose dimensions allow it to be placed m the bore of the magnet and hold the desired number of molds (see Note 3) 6 A large-capacity hot-water heater and recirculator (Fisher, Sprmgfield, NJ) 7 A tilled icebox large enough to hold all of the molds

3. Methods 3.1. Collagen Solution Preparation The preparation of the collagen solution is rather straightforward and simple. However, the temperature of the reagents and the order of mixmg, as well as

Girton, Qubey, and Tranquil/o

70

Teflon rod

Field

Polypropylene

tube

Teflon tube

Teflon endcaps Fig. 1. (A) The tubular mold comprised of two rubber cylinders, a Teflon rod, and a polypropylene tube. Shown are the two needles used to fill the mold, as well as the necessary orientation of the mold relative to the magnetic field for circumferential alignment. (B) The rod mold comprised of two Teflon cylindrical endcaps and a Teflon tube. Shown is the needle and syringe used to fill the mold, as well as the necessary orientation of the mold relative to the magnetic field for longitudinal alignment.

the source of the collagen, all prove to be important. (Vitrogen 100 is preferred, because its fibrillogenesis rate is relatively slow (7), thereby enhancing the degree of magnetic-induced alignment, as previously explained.) If the temperature of the reagents and the resulting collagen solution are kept close to 5°C turbidity experiments (a good indication of both the rate and extent of fibrillogenesis) indicate that no significant fibrillogenesis occurs. This allows time for the injection of the solution into the molds and their transfer to the magnetic field before significant fibrillogenesis. The temperature maintained during fibrillogenesis is discussed subsequently. To make 1 mL of 2 mg/mL collagen solution, thoroughly mix the following cold (5°C) ingredients, in the order listed below, in a sterile container that is kept on ice until tibrillogenesis is desired:

Alignment of Collagen Fibrils

71

1 20 $ HEPES-buffer solution (helps mamtain the correct pH). 2 132 & 0.1 N NaOH (raises the pH of the solution from 2 to about 7.6) Vltrogen 100 collagen fibrlllogenesls begins once both the pH and the temperature are raised The ideal pH for fibrlllogenesls 1s about 7 6, increasing temperature increases the rate of fibrlllogenesls Note that Vltrogen 100 is 3 mg collagen/mL and shipped at 5°C and pH 2 3. 50 & 10X MEM (used prlmarlly as a source of nutrients for cells entrapped or postseeded (see Note 4), but also provides a phenol red pH Indicator) 4 60 & FBS (used primarily for cell survival, however, FBS may alter the dlameter and structure of the collagen fibrils formed) 5 1 & Pemcllln-streptomycm solutlon (used to ensure sterlhty of the gel) 6 10 $ L-glutamme solution (plovldes cells with needed nutrients, so can be omitted if there are no cells present) 7 677 pL Vltrogen 100 (see Note 5).

3.2. Preparation

of Molds

After the preparation of the collagen are assembled and the mold filled.

solution,

the autoclaved

mold pieces

1 Tube mold Fit a sterile rubber cylinder onto each end of the Teflon rod and slide the resulting piece mto the clear polypropylene tube (Fig. 1A) Pierce one of the rubber cylinders with a 21-gage needle, so that the needle tip 1s exposed m the tubular cavity between the Teflon and the polypropylene outer cylinder Fill the ~-CC syringe with the collagen solutlon, and pierce the other rubber cylinder with the syringe needle as before. Fill the tubular cavity with the collagen solution m the syringe (the other needle provides an evacuation hole for the air) 2 Rod mold Insert a Teflon endcap into one end of the Teflon tube (Fig. 1B) Fill the tube with the collagen solution and cap the end with the remaining Teflon endcap After

fillmg,

wrap the molds with Parafilm

and place them on ice.

3.3. Magnetic-hcfuced Alignment Once the molds are full, they are placed m the temperature control chamber, and with the use of a hot-water recirculator, the temperature of the samples IS mamtamed durmg fibnllogenesls. As noted previously, the degree of altgnment obtained Increases with decreasing temperature, however, for Vltrogen 100 fibrlllogenesls to occur within 2 h when prepared as described, the temperature must typlcally be 33-37°C . Placement of the molds m the proper orientation in the magnetic field IS essential for achieving the desired fib4 alignment (cf. Figs. lA,B). Collagen will align m the direction parallel to the surfaces that define the gap that IS m the plane normal to the magnetic field direction; thus, for circumferential ahgnment of the collagen m the tubular mold, the axis of the mold must be placed

72

Glrton,

Dubey,

and Tranqutllo

parallel to the direction of the magnetic field; for longltudmal alignment m the collagen rod, the axis of the mold must be placed normal to the dn-ectlon of the magnetic field. 1 Connect the temperature control Tygon tubing 2 Place the molds m the appropriate and start the hot-water recirculator the magnet. 3 Leave the chamber m the magnet 4 Remove the chamber from the (see Note 6)

chamber to the hot water recu-culator

with

ollentatlon m the chamber Seal the chamber, Place the chamber unmedlately m the bore of until fibrlllogenesls IS complete bore, and then the molds from the chamber

After removal of the tissue equivalents from the molds, it IS often desirable to verify that the collagen fibrlls are aligned (see Note 7). Viewing the samples between crossed polarizmg elements IS a simple and nomnvaslve method of checking for alignment The tissue equivalents are too delicate right after fab-

rication to be removed from the medium without disruption; therefore, crosslmkmg

of the collagen

by the addition

of 3.7% formalm

1s first recom-

mended. Rotate the slide holding the sample, and verify that the intensity of transmitted

light varies with the angle of rotation

(there should be alternating

maxima and minima every 90 degrees). The associated blrefrmgence can be measured to quantify the degree of alignment (2), but this IS substantially more complicated

and unnecessary for the simple purpose of verlficatlon

of alignment

4. Notes 1 Vltrogen 100 IS a pepsin-solublhzed collagen from bovine dermis Depending on the source of the collagen and the manner of extraction , the rate of gel formation (collagen fibnllogenesls) may vary drastically Acid-extracted type I collagen from rat tall tendon can form a gel m a few minutes, m comparison to 3&60 mm typically required for Vltrogen 100 2 Safety should be a primary concern when using such high-strength magnets Be

sure that no metallic parts are used in the molds or temperature-control

chamber

3 A temperature-control chamber should be used to keep the samples at the optimal fibnllogenesls temperature m the magnetic field Our chamber IS a watertight acrylic tube with a removable endcap possessmg an entrance and exit port for the warm water from the recirculating water bath There are few restrictions on the type of device, as long as it maintains the temperature and can fit m the bore of the magnet

4 Cells can be entrapped wlthin the collagen gel by the addition of a small amount of a concentrated suspension of cells m culture medium to the collagen solution Cells are entrapped within the fibnls of the gel during fibnllogenesls, and subsequently spread and migrate during mcubatlon of the tissue equivalent, with an

associated cell-Induced compaction of the network of collagen fibrlls

Alignment of Collagen Fib&

73

5 The collagen concentratton may be altered, although this may also affect the rate of tibrtllogenesls. 6 Immediately after fibrtllogenests (I e , before cell-induced compactton), the ttssue equrvalents are eastly dtsrupted, and must therefore be removed carefully from the molds For the tubular molds, stmply insert a plunger mto one end of the polypropylene tube, making contact with the rubber cylinder, and force the rubber cylmders, the Teflon rod, and the tissue eqmvalent out mto warm cell-culture medmm For the rod molds, stmply remove both Teflon caps and allow the tissue eqmvalent to shde out of the Teflon tube mto the medmm All of these mampulattons must be done m a sterile environment 7 Collagen alignment can be Inferred from cell alignment, since contact guidance appears to be a property of all mottle blood and tissue cells, but a period of mcubatton IS first required for cell spreadmg to occur, so that cell ortentatton can be measured

References Torbet, J ( 1984) Magnettc ahgnment of collagen durmg self-assembly Bzochem J 219, 1057-1059 Gmdo, S and Tranqmllo, R T (1993) A methodology for the systematrc and quantltatrve study of cell contact gmdance m ortented collagen gels Correlatton of fibroblast orrentatton and gel btrefrmgence J Cell Scz 105, 3 17-33 1 Kanda, K and Matsuda, T (1993) Behavior of artertal wall cells cultured on pertodxally stretched substrates Cell Transplantation 2,475484 Ghosh, S and Comper, W D (1988) Oriented fibrtllogenests of collagen m vitro by ordered convectton Conn Tissue Res 17, 33-41 Hughes, K E , Fmk, D J , Hutson, T B , and Vets, A (1988) Hugh-strength collagen bromatertals JALCA 83, 372-378 Becker, R 0 , Bassett, C A , and Bachmann, C H (1964) Btoelectrlcal factors controllmg bone structure, m BoneB~odywnm, (Frost, H. M , ed ), Little, Brown, Boston, pp 209-230 Bell, E , Ivarson, B , and Merrtll, C (1979) Productton of a tissue-like structure by contractton of collagen latttces by human fibroblasts of drfferent prolrferattve potential m vttro Proc Nat1 Acad Scl USA 76, 1274-1278

Preparation

and Use of Thermosensitive

Yi-Lin Cao, Clemente

Polymers

Ibarra, and Charles Vacanti

1. Introduction Restoration of organ structure and functron, utrhzmg tissue engmeermg technologies, often requires the use of a temporary porous scaffold. The function of the scaffold 1sto direct the growth of cells mtgratmg from the surroundmg tissue (trssue conductron), or of cells seeded withm the porous structure of the scaffold. The scaffold must therefore provtde a suitable substrate for cell attachment, drfferentrated functron, and, tn certain cases, cell prohferatton (I-3). These crmcal requirements may be met by the choice of an appropriate material from which to construct the scaffold, although the suttabthty of the scaffold may also be affected by the processmg technique. There are many brocompattble materials that could potentially be used to construct scaffolds, however, a biodegradable material is normally desirable, since the role of the scaffold 1susually only a temporary one. Many natural and synthetic biodegradable polymers, such as collagen, poly(ahydroxyesters), and poly(anhydrrdes), have been widely and successfully used as scaffolding materials because of then versatrhty and ease of processmg. Many researchers have used poly(a-hydroxyesters) as starting materials from which to fabricate scaffolds, using a wide variety of processing techniques. These polymers have proven successful as temporary substrates for a number of cell types, allowing cell attachment, proliferation, and mamtenance of differentiated function (IO) Poly(a-hydroxyesters) such as the polyoxamers are a family of more than 30 different nontoxrc, nomomc surface active agents. These compounds are made at elevated temperature and pressure by the sequential addrtton of propylene oxrde and then ethylene oxrde to neutralize the salt that is generally retained m the final product. The polyoxamer serves cover a wide range of hqutds, pastes, and solids, with mol wt varying from 1100 to about 14,000 The ethylene oxrdepropylene oxide weight ratios range from about 1:9 to about 8:2 (4) From

Methods Edited

m Molecular

by J R Morgan

Medune,

Vol 18 Tissue Engmeenng

and M L Yarmush

75

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

Cao, Ibarra, and Vacanti

76

Concentrated aqueous soluttons of the polyoxamers form gels. A unique phystcal property of these gels ISthen reversible nature Upon lowermg the temperature, the gels revert to a liquid state.Upon rewarming, the gels reform. This change in physical state can occur mnumerable times and IS temperature-dependent Polyoxamer 407 was found to form a gel at only 20% concentration by weight at 25°C which IS less than that of the other members of the polyoxamer series. Because of this, polyoxamer 407 has been selected for further study as an mnocuous carrier for medrcattons to treat burn wound mJurres, as well as other dermatologrc or surgical disorders (4-9) A desirable feature of a polyoxamer gel for the clnncal treatment of burns, IS its ability to adhere to the surface to which it 1sapplied. This 1sa physical mteractton between the aqueous polyoxamer gel and the surface to which tt adheres. A novel approach to the formation of a layer of tissue engineered cartilage on host bone using synthetic polymers, is described The ability of thermosenstttve, btocompattble, and btodegradable liqurd polymer, a synthetic copolymer of ethylene and propylene oxide, whrch transforms to gels at phystologrc temperatures to support carttlagmous tissue formatton when mixed with rsolated chondrocytes on a viable osseous surface, was mvestrgated. 2. Materials 1 Chondrocyte culture medium (complete medium) The medium used to culture chondrocytes is composed of Ham’s F- 12 nutrient mixture (Gibco-BRL, Grand Island, NY, # 21250-089), supplemented with 10% fetal bovine serum (GibcoBRL, # #OOOO), pemcillm (100 U/mL, [Sigma, St Louis, MO], #A-9909), gentamycm (100 ug/mL, [BtoWhittaker, MD], # 17-5 1SZ), streptomycin (100 pg/mL [Sigma, Ann Arbor, MI], #A9909), amphotericin B (0 25 )rg/mL [JRH Biosciences, Lenexa, KS], #59-604-076), L-glutamme (292 pg/mL [Gtbco], #OOOO), ascorbtc acid (5 Clg/mL [Sigma], #A-0278) After sterthzatlon through a 0 2 pm filter (Gelman Sciences, Ann Arbor, MI, #66234), the culture media is stored at 4°C m 500-mL ahquots 2 Dulbecco’s phosphate buffered salme (PBS) (Gibco, # 21600-010) Ca*+- and Mg2+-free Chondrocyte wash solution 1X solution is made by mixmg 9 6 g of powder with 1 L deionized ddH,O pH was adJusted to 7 0 with NaHCO, if necessary, and filter-sterilized through a 0.2-l-m filter (Gelman Sciences) pH will rise 0 l&O 3 units after filtering Store at room temperature 3 PBS/antibiotic solution is made by mixing 1X PBS (described above) with 1 g/L cefazobn sodium (Apothecon, Princeton, NJ, #NDC 0015-7339-12) Filter through a 0 2 pm filter (Gelman Sciences), and store at 4°C

4 Collagenasesolution is madewith 300mg collagenaseclass2 powder (Worthmgton Btochemtcals, Freehold, NJ, #LSO4 177), mixed m 100 mL plam Ham’s F- 12 medium (Gibco-BRL), prefiltered through Fisher P4 paper disks (Fisher Sclentlfic, Pittsburgh, PA, #09-8036H), and finally filtered through a 0 2+m filter (Gelman Sciences), and stored frozen at -5 to -8°C

Thermosensitive

Polymers

77

5 Pluromc gels Pluromc F 127 NF (Polyoxamer407NF) surfactant (BASF, Mount Oltve, NJ, #549926) will form a gel at concentrations greater than 20% w/v (g/mL) m plain Ham’s F-12 medium at temperatures above 15°C Gels ~111 reltquefy when cooled below 15°C. Concentratrons of 3040% (w/v) are Ideal for tissue engmeermg 6 Trypsm 0 25%/EDTA solutron Gtbco-BRL, #27250-042/l 1267-O 10 7 Trypan blue vttal dye (Sigma, # T-8 154) 8 Tissue-culture dashes (100 x 20 mm). Falcon, Becton Drckmson, Lmcoln Park, NJ, #3003. 9 50-cc Conical (centnmge) tubes. VWR Sctentrfic, Westchester, PA, # 2 1008- 178. 10 0.2~pm Filter disks (47~mm diameter) (Gelman Sciences, Ann Arbor, MI) 11 250 pm Polypropylene spectra mesh (autoclavable). Ftsher Screntrfic#08-670-25Oc, Pittsburgh, PA

3. Methods 3.1. Pluronic

(Polyoxamer)

Gels

The techmque used for makmg aqueous polyoxamer gels ts very simple A weighed amount of the polyoxamer 407 IS slowly added to a known weight of cold water (less than 10°C) Sttrrmg rate should be controlled to mamtam a slight vortex m the hqutd, Too rapid a stnrmg rate wtll cause aeration and the

formation of foam When all the polymer has been added, sttrrmg can be contmued (whtle keeping the solutron cool), untrl a clear solution IS formed, or the container can be placed n-ra refrigerator and left undisturbed for several hours, at which time solutron IS complete. Known burn medtcattons, such as silver

salts, antibtottcs, and so on, can be added to the cold solution and dissolved or suspended therem, while warming to induce gel formation. The temperature at whrch the hquid gels will vary, depending on the poloxmar concentratron and on the type and amount of additives. Thus, the addition of a humectant, such as glycerm, will increase the gel strength more than the addition of the same

amount of propylene glycol. The polyoxamer gels can be autoclaved and are stable on storage (Fig. 1) Protocol to prepare Pluromc (polyoxamer) gels. 1 On a magnetic stirrer, place an autoclavable bottle with a stir bar m a beaker wrth Ice and water to keep the mrxture cold Alternatrvely, this procedure can be performed m a cold room at 5°C 2 Add cold Ham’s F-12 medium to the bottle Use a bottle that 1s large enough to allow 30% headspace above the mrxture 3 Add Pluronic F 127 powder slowly through a dry funnel, while stirring, gradually addmg l-2 g at a time every l&15 mm Avoid allowing large clumps to form 4 Keep the mixture stlrrmg until clumps of powder are smaller than 2-3 mm This

may take up to 16 h The mrxture must be kept cold throughout thts process


78

Fig. 1. Photographic close-up view of a drop of Pluronic gel in liquid state at the distal end of an 1S-gage needle.

When the Pluronic powder has completely dissolved, remove the stir bar and autoclave the mixture in a pan of water for 20 min. Some researchers report successful results sterilizing these gels with UV light and filtering through a 0.25~p filter in the liquid (cooled) stage. Store at 443°C.

3.2. Isolation

of Chondrocytes

Newborn calf forelimbs were obtained from a local abattoir within 6 h of animal sacrifice. Cartilage was harvested by sharp dissection under sterile conditions from the articular surfaces of the humerus and scapula. Tissue fragments were washed twice with 1X PBS solution with antibiotics in 50 cc conical tubes. PBS was then replaced with a 3% solution of class 2 collagenase to digest the tissue at 37°C for 12 h in a shaker. The solution and tissue debris was filtered through a 250~w-pore nylon mesh, and was centrifuged to obtain a cell pellet, which was then washed twice with PBS with antibiotics solution. The cells were finally suspended in 10 cc of complete culture media and were counted in a hemocytometer. Trypan blue vital dye was used for counting to assess cell viability.

3.3. Cell-Polymer

Suspension

After counting the chondrocytes, the cell-polymer suspension is centrifuged again and the supernatant is removed. Chilled 30% (w/v) copolymer solution is added to obtain a cellular concentration of 5 x lo7 cells/ml. The conical tube

Thermosenslbve

Polymers

79

contammg the cell pellet must be maintained m Ice at the moment of adding the chilled polymer solutton, to prevent tt from setting before the cells can be evenly suspended m tt by vortexmg. The cell-polymer suspension is kept at 4°C to mamtam the hqutdtty of the copolymer.

3.4. Tissue-Engineering Applications 3.4.1. Cartilage Growth on Bone by Painting 3 4 1 1 METHODS

Under general anesthesia with inhaled methoxyflurane, the calvartae of 12 male athymtc mice were exposed through a longttudinal mctston on the head, and the surface of the bone was abraded using a drill bit. With a sterile fine brush (0 5 x 1 cm), each calvarmm was pamted with several overlying appltcattons of the chondrocyte-copolymer admixture described above. After each apphcatton, the hqurd was allowed to polymerize over the calvartal surface as rt was exposed to body temperature. A total of five layers were applied to each animal. Subsequently, the mctsron was closed with 5-O nonabsorbable sutures Four animals were sacrtticed at 4,6, and 8 wk Twelve addmonal animals were used as controls, applymg the ltqutd copolymer without chondrocytes or chondrocytes suspended m tsotomc salme, m similar fashion. The animals were euthamzed by anesthetic overdose at the different time mtervals and the calvartae were harvested for gross and mtcroscoptc exammatron 3 4 1 2 RESULTS

The pamted chondrocyte-copolymer solutton gave rise to a semtsohd sticky gel a few minutes after apphcatton, and showed evidence of cartilage formation m all the animals m the experimental groups at all time-points. The new carnlage had been mcorporated to the surface of the underlying bone, as a 3-5 mm thick layer of a white, opaque coating that was difficult to dislodge mechamtally. Htstologtc sections stained with hematoxylm and eosm (H&E) demonstrated the presence of hyalme-like cartilage Remarkably, the pamted cartilage mtiltrates the underlying osseous substrate, formmg bone-carttlage interface None of the control animals demonstrated any cartilage formatton either grossly or on hlstologrc exammatton. The calvarrae of animals pamted with either chondrocytes, suspended m isotomc salme or llqutd copolymer alone, showed no changes m theu bony contour. 3 4 1 3 CONCLUSIONS

This study demonstrated the formatron of new cartilage using isolated chondrocytes suspended in a thermopolymerrzable gel, applied with a brush on a bony surface. A relatively thick surface of arttcular-like cartilage with a near-


80

normal, strong, and stable bonecartrlage interface was formed This technique shows potential usefulness m Jomt resurfacmg, with the unique charactertsttc of allowing the formation of new cartilage in the desired shape and locatron, with potential mmtmal mvastveness.

3.4 2. Inlectable Cart//age 3 4.2 1 METHODS Twelve athymtc mice were anesthetrzed with Inhaled methoxyflurane. Five 100~p.L ahquots of a chondrocyte-gel suspension of 30% (w/v) polyoxamer, with a cell concentratron of 5 x lo7 cells/ml were injected subcutaneously, through a 22-gage needle, mto five dtfferent locations on the dorsum of each mouse. Twelve addmonal mice were injected with the same amount of cells suspended m tsotomc saline, and 12 mice were injected with gel wtthout cells, m stmtlar locattons, as controls. Three animals from each group were euthamzed by anesthetic overdose after 2, 4, 6, and 8 wk respectively The presence of areas of mduratron under the skm was assessed prevtous to euthanasia m all animals. Specimens harvested at thts time were evaluated grossly and htstologlcally m search of evidence of new cartilage formatton. 3422

RESULTS

All the animals m the experimental groups presented wrth five subcutaneous nodules at the moment of euthanasta Upon palpation through the skm of the mice, the nodules had firm rubbery consistency. There were no apparent nodules or sites of mduratton under the skm of any of the control animals After surgical exposure, the nodules on the animals m the experimental groups were white, opalescent, amorphous masses of varying sizes (7-10 x 57 x 2-5 mm) (Fig. 2). The nodules had mamtamed the size and shape that the gel adopted rtght after mJectton, as tt dissected its way through the subcutaneous space. They were all enclosed wtthm a well-vasculartzed capsule that could be eastly dissected from the underlymg fascia and overlying skm. After removal of the vascular capsule, the texture and consistency of the nodules resembled that of cartilage. Specimens harvested after 2 wk were relatively softer at palpation, compared to specimens harvested at later ttme-points. Specimens harvested after 8 wk presented the greatest apparent resistance to manual compresston Htstologrc exammatton of all specimens, stained with H&E, demonstrated the presence of new cartilage formatton (Fig. 3). Specimens at the earher trme-points showed the presence of very cellular munature-appearing cartilage After 8 wk, the htstologtc appearance of the nodules was that of more mature-appearmg, hyalme-like cartilage, with chondrocytes m lacunae, surrounded by extracellular matrix with a higher matrrx:cell ratio. New cartt-

Thermosensitive

Polymers

81

Fig. 2. Gross appearance of tissue-engineered cartilage specimens, harvested after 4 wk of subcutaneous injection of porcine chondrocytes suspended in Pluronic gel in nude mice.

Fig. 3. Microphotograph (x10) of histologic section stained with H&E of tissueengineered cartilage, after 8 wk of subcutaneous injection of porcine chondrocytes suspended in Pluronic gel in nude mice.

82

Cao, Ibarra, and Vacant/

lage was not found m the subcutaneous space after dissection of animals m which plain polymer gel was inJected Small remnants of polymer scattered m some areas could only be found m animals m this group at the earlier timepoints (24 wk). Very small nodules of cartilage-appearmg tissue were found durmg dissection of animals m which chondrocytes suspended m salme were inJected Such small nodules could not be appreciated through the skm of these mice Histologic appearance of this nodules resembled that of the new tissue formed m the ammals m the experimental groups

3 4 2 3 CONCLUSIONS New cartilage can be obtained by mJectmg isolated chondrocytes suspended m a thermopolymerizable gel as a carrier m the subcutaneous space of athymic mice The volume of new tissue formed was significantly greater than that obtained by the simple mJectron of cells suspended m salme. The new tissue matured over time, but by 2 wk, the presence of new cartilage formation was evident both grossly and histologrcally with this technique. The copolymer gel utthzed is degraded and disappears almost completely after 2 wk m the subcutaneous space, if it is not combined with cells that can replace its presence with newly synthesized extracellular matrix, and thus, new tissue. In summary, this work demonstrates the abilny of these relatively new thermosensmve polymers to act as a cell-dehvery device that can be utilized m a number of different ways to form new cartilage. This system shows great promise m the fields of orthopedics and reconstructive surgery as a means to form new cartilage, perhaps utrhzmg autologous cells, for the repair or reconstruction of cartilage surfaces and structures employing mimmally mvastve techniques

References 1. Langer, R. and Vacanti, J. P (1993) Tissue engineering Science 260, 920-926 2 Vacantr, J P and Vacanti, C A. (1996) The challenge of tissue engineering, in Text Book of Tzssue Engrneerzng (Lanza, R , Langer, R , and Chick, W , eds ), R G Landes, Austin, TX, pp l-4 3 Vacanti, C A , Langer, R , Schloo, B , and Vacanti, J P (1991) Synthetic polymers seeded with chondrocytes provide a template for new cartilage formation Plastic Reconstruct

Surg 88, 753-759.

4 Wong, W H and Mooney, D J (1997) Synthesis and properties of biodegradable polymers used as synthetic matrices for tissue engineering, in Synthetic Bzodegradable Polymer Scaffolds (Atala, A , Mooney, D J , Vacanti, J P , Langer, R , eds ), Bnkhauser, Boston, pp 5 l-82 5 Schmolka, I. (1972) Arnficial Shn I Preparatton and propernes of Pluronic F-127 gels for treatment of burns J Bzomed Mater Res 6, 57 l-582

Thermosensitive

Polymers

83

6 Faulkner, D M , Sutton, S. T , Hesford, J D., Faulkner B C , MaJor, D A , Hellewell, T B , Laughon, M M , Rodeheaver, G T., and Edhch, R F (1997) A new stable Pluromc F68 gel carrier for anttbtottcs m contammated wound treatment Am J Emerg A4ed 15,2&24 7 Mtyazakt, S , Tobtyama, T , Takada, M., and Attwood, D. (1995) Percutaneous absorption of Indomethacm from Pluromc F 127 gels m rats. J Pharm Pharmacol

47,455-457 8 Pausttan, P W , McPherson, J C , III, Haase, R R , Runner, R R., Plowman, K. M , Ward, D F , Nguyen, T H , and McPherson, J C (1993) Intravenous Pluromc F- 127 m early burn wound treatment in rats Burns 19, 187-l 9 1 9 Rice, V M , Shanti, A, Moghtsst, K , and Leach, R E (1993) A comparative evaluation of Polyoxamer 407 and oxidrzed regenerated cellulose (intercede TC7) to reduce postoperative adhesion formation m the rat uterine horn model Fed Sterd 59, 901-906 10 Thomson, R C , Yaszemskt, M J , and Mtkos, A G (1996) Polymer scaffold processing In Text Book ofPruwples of Tusue Engmeermg (Lanza, R , Langer, R , and Chick, W , eds ), R. G Landes, Austin, TX, pp 263-272

ln Situ Photopolymerized Hydrogels for Vascular and Peritoneal Wound Healing Amarpreet

S. Sawhney and Jeffrey A. Hubbell

1, Introduction 1.1. Wound Healing in the Peritoneal Cavity The process of wound healmg in internal organs lmed wrth mesothehal cells is different in some respects from healing in cutaneous mjurres Hertzler (I) noted that cutaneous wound re-epitheliahzatron takes place from the wound borders, but perttoneal defects become re-mesothelialrzed simultaneously. Most organs m the pelvic cavrty possessa mesothehal lmmg that can be injured as a result of trauma from surgical intervention or handling. It 1s generally agreed that the remesothehahzatron of the peritoneum can take place over 5-8 d (2) During this period, events of inflammation and mesothebal repair occur that can set the course for postsurgrcal adhesion formation. Postsurgical adhesion formation, or the formation of scar tissue bridges between organs in the pelvic cavity, 1san undesirable result of a natural healing process Ryan et al. (3) showed experimentally that the surface of bowel 1ssubject to demesothehahzatron and loss of native fibrmolytrc activity by surgical mampulatron, whether rt 1sallowed to dry or 1skept moist with salme. This denuded surface could allow blood clots to form and act as precursors to adhesron formation before remesothehahzatron can occur The formation of adhesions can distort the normal anatomy and natural posmomng of organs, and has often been rmphcated m postoperative pelvic pam and bowel obstructions, m addition to other complicatrons Ryan recommended that to minimize adhesron formatron, tissue exposure and manipulation should be mimmrzed (3) However, this 1soften not possible, and so the use of barriers to Isolate injured organs from their surroundmgs has been proposed as a way to mmtmize adhesion formatron. From

Methods m Molecular EdIted by J R Morgan

Medicine, Vol 18 Tissue Engmeenng Methods and and M L Yarmush 0 Humana Press Inc , Totowa,

85

Protocols NJ

86

Sawhney and Hubbell

Barrier materials, such as Gore-TexTM surgical membrane, have been shown to reduce the instance of adhesion formation (4) However, this material requires retrieval, because it IS not absorbable. The barrier material should also be extremely nomnflammatory, so as not to further enhance the already active mflammatory processes. Therem lies a challenge for absorbable btomaterials to degrade within the time frame of a few weeks, and yet be inert and nomnflammatory during this degradation process. Fabrics such as those made from oxidized regenerated cellulose absorb within the required time frame but were seen by Haney et al. (5) to cause peritoneal m3ut-y 1.2. Vascular Wound Healing The process of balloon angioplasty of artertes that have been narrowed because of artheroscelerotic plaque is used to restore patency of the vessels The balloon dilation process is traumatic and causes denudation of the endothelium, as well as mlury of the mttmal and medial tissues as a result of dtlatton The clmical significance of the ensuing events of wound heahng that occur are the renarrowing (or restenosis) of approx 3040% of coronary arteries dilated by balloon angioplasty within 6 mo of treatment (6) Wound healing m this case is a complex series of events mvolvmg both lummal and medical events It has been observed that smooth-muscle cell migration, and prohferatton and elaboration of matrix proteins, contribute to mtimal thickening and consequent lummal narrowing. Several growth factors, cytokmes, and chemoattractants play a role m smooth muscle-cell sttmulatton Some of these are lummal m origm, such as platelet-derived growth factor and thrombm, and may be deposited locally as a result of the thrombosis at the angioplasty site (7,s). This mural thrombus, which forms because of the denudation of the artery and the exposure of underlymg thrombogemc medial tissues, may also serve as a matrix for smooth muscle cell prohferatton. Other factors, such as basic fibroblast growth factor, are mtrmsic to the vessel wall and are also mitogens for smooth-muscle cell proliferation (9,ZO). 1.3. Photopolymerized Bioresorbable Hydrogels Hydrogels are materials that have water as a significant component of their structure and have the capacity to swell in an aqueous surrounding. Because of the high water content of these materials, and then low mterfacial tension with water, they are often extremely blocompatrble, and can serve as barriers and scaffolds over injured tissue to help direct the wound healing process. Noniomc hydrogels, such as the poly(ethylene glycol) (PEG)-based gels discussed herem, contam large amounts of water, and have hydrophtlic, nomomc surfaces. This lowers the driving force for protein adsorption onto these surfaces from the physiological surroundmgs (11). Cell adhesion is mediated by protein

In Situ Photopolymerized

87

Hydrogels

CH3 Lactlde Stannous 2-Ethylhexanoate I

PEG - Lactate Diol

Acryloyl Chloride

m=2-6

Trlethyl Amine 1

Macromer

Frg 1 Schematrc showtng the synthesis of the photosensmve macromers

adsorptton and/or proteoglycan mteractton with charged surfaces (12,13). The use of PEG to increase biocompatrbtlity is well documented in literature. PEG has been used m numerous studies to increase btocompattbrltty of bloodcontactmg materials (14) and soft-tissue contactmg materials (15). The precursor molecules (called “macromers” here), that can be used to form the crosslmked hydrogels by polymerization, are novel molecules that have three domains along the polymer chain. The synthesis and structure of these macromers 1s illustrated n-t Fig. 1. The water-soluble central domain enables the entire polymer to be water soluble. Short a-hydroxyester oltgomertc domains that flank the central water-soluble domain are terminated with acrylates as polymertzable end groups. The presence of hydrophobic end regions

88

Sawhney and Hubbell

on the central PEG cham has been seen to result m the formatton of mtcellar structures m aqueous solutions (26). The formation of such preorgamzed configurattons m aqueous soluttons of these precursors results m the rapid polymertzatton and gelation of these multtfunctional precursors. The btoresorbability of the hydrogels described above is afforded by hydrolysis of the well-known poly(a-hydroxy acids), such as poly(lacttc acid) and poly(glycohc acid) The use of a light-sensitive mmatton mechanism to provide the rapid and spatially controllable gelatton of a hqutd precursor, m direct contact with the treated ttssues, provides a conformal lubrtctous and adherent hydrogel film on the tissues. This is presumably caused by the flow of the liquid precursor mto crevices on the surface of the tissue, with subsequent mechanical mterlockmg upon converston mto a gel In thts chapter, we describe the synthesis and use of such hydrogels, and then use m guiding wound healing m the peritoneal cavity and m vascular tissues Photopolymertzed bioresorbable hydrogels have been shown to be effective m gutdmg wound healing and m prevention of postsurgtcal adhesion formation m several animal models (17-19). These gel barriers are resorbed over a time frame of less than 2 wk m VIVO. Studies conducted m rats, using only the ungelled viscous precursor solutton, have failed to prevent the formatton of adhesions This IS quite expected, because the use of vtscous macromolecular soluttons, despite the coatmg and lubrtcatmg effects, has been documented to be msuffictent to prevent the formatton of adhesions (20) The gel degrades mto water-soluble substances, primarily PEG, lactic acid and its ohgomers, and ohgomers of acrylic actd. These components are all water-soluble, and are either naturally existent m the body or are rapidly cleared by excretion through the kidneys Photommatton 1s used to induce gelation This reaction takes place under very mild condmons, with no excessive hberatton of heat or toxic factors to cause tissue necrosis This is because of the relatively low acrylic group concentratton, compared to the larger nonreactive central PEG segment, which forms the bulk of the material. The photoimttator used m this work presents a very low toxtcological burden The oral LD,O m rats is m excess of 6 g/kg (21), and, typtcally, e.g., m a rabbit model, the total dose 1s about 0.6 mg/kg. The photomttiator used m this study is sensitive to long-wavelength ultraviolet (LWUV) light, which 1s essentially lacking n-r cytotoxic or mutagenic potency. The total dose of light used is about four orders of magmtude less than that at which an effect would be expected to be observable (22) The approach of applying conformal barrier layers m the liquid state, and then gelling the liquid by application of light, provides several conceptual advantages over the applrcation of preformed barriers, such as fabrics First, the liquid may be readily deployed through a laparoscope, whereas a fabric must be rolled, pushed through the laparoscope, unrolled, and placed on the


Hydrogels

89

tissue to be treated. Moreover, the ltqutd is inherently conformal, i.e., it adapts its shape precisely to that of the ttssues; by contrast, fabric barriers are not conformal to the macrotopology of the ttssue surface The methods descrtbed here can be used to study the formation and prevention of the adhesions between the injured tissue and neighboring tissues. Isolatton of the injured tissues after angioplasty, to prevent local thrombosis, may be a way to change the course of events leading to restenosts. In this chapter, a nonpharmacological approach to prevent mtimal thtckenmg, by inhibitmg contact between blood and subendothelial tissues with a resorbable polymeric barrier, 1s demonstrated. A method 1s described to synthesize a thm nonthrombogemc, conformal hydrogel barrier on the arterial wall by mterfactal photopolymerization of a macromolecular precursor 112situ. This novel barrier has been previously shown to virtually eliminate thrombosis in injured arteries The transient presence of the barrier has also been shown to preserve patency and reduced intimal thickening long after the dtsappearante of the barrier (23). It is critical to achieve this coating process with careful control of the thickness and growth of the coatings, e.g., with mmima1 loss m lummal diameter of a treated artery. Polymertzatton m contact with ltvmg tissue in a phystologtcal environment necessitates that the chemtcal entities used and generated during the polymertzation process be nontoxic,

and that the polymertzatton

conditions

be mild.

In this chapter,

we

describe techniques to reproducibly and accurately deposit hydrogel films on the surface of vascular tissue

2. Materials All chemical, reagents, and solvents can be purchased from Aldrich (Mtlwaukee, WI), unless otherwise mentioned. 2. I. Peritoneal

Wound Healing

2.1.1. Synthesrs of PEG-co-Poly(a-Hydroxy Acid) Copolymers 1 Reagents ~1,w-dihydroxy PEG with mol wt 10,000 (PEG lOkDa), oL-lactide, stannous 2-ethylhexanoate, trrethyl amme, acryloyl chloride 2 Solvents+ benzene, ethyl acetate, dtchloromethane, anhydrous dtethyl ether. 3 Glassware round bottomed flasks, tubing, Y-connectors, gas mamfold, beakers, crystallizatron dashes, 1-cc syrmge with needle, fretted glass filter 4. Other supplies heating mantle, magnetic stirrer and sttrbars, vacuum pump.

2.1.2. Synthesis and Characterization

of PEG Macromers

1 Polyethylene films for infrared spectra, NMR tubes 2 Tetramethyl sllane, CDCI,

90

Sawhney and Hubbell

2.1.3. L WUV-induced Photopolymerization 1 Phosphate-buffered salme (PBS) 0.2 g/L KCl, 0 2 g/L KH,PO,, 8 g/L NaCl, 1 15 g/L Na2HP04, pH 7.4 2 Inmator solutton 300 mg 2,2 drmethoxy 2-phenyl acetophenone drssolved m 1 mL of N-vinyl pyrroltdmone. 3 LWUV-emrttmg lamp (Blak-Ray, model B-100A with Flood, VWR Screntific, Boston, MA) 4 Sterrlrzed alummum for1 (autoclaved) 5 Polymerrzable solution A 10% w/v solution of macromer m PBS 1s used This solution can be filtered through a 0 45-pm filter (Mrlhpore, Bedford, MA) to sterrltze It. To 1 mL of thus solution, 3 ltL of mmator solutton IS added, and the solutron 1s vortexed brrefly. This solutron is then drawn up n-r a 10 cc syringe that 1s shielded from light using the alummum for1

2.7.4. Rabbit Uterine Horn Model 1 New Zealand whrte rabbits between 2.5 and 3 5 kg m weight (HRP, Denver, PA) 2 Anesthesia A cocktail contammg equal parts of ketamme (“Ketaset” 100 mg/mL m lo-mL vials Fort Dodge Labs, Fort Dodge, IA, dose at 100 mg/kg), Xylazme (100 mg/mL 50-mL vials, Fermenta Ammal Health Care, Kansas City, MO, dose at 5 mg/kg) and Acepromazme (10 mg/mL m 50-mL vials, Fermenta, Kansas City, MO, dose at 0 5 mg/kg) 3 21-gage mtravenous catheter (Baxter Healthcare, MC Gaw Park, IL), l-, 5-, and 1O-cc syrmges, 2 1-gage hypodermtc needles 4 Betadme solutrons (Perdue Fredericks, Norwalk, CT) for scrubbing and preparmg, and hair clipper 5 Standard abdominal surgery pack containing abdominal scissors, self-retamlng retractors, surgical blade (#lo) and handle, suture scrssors, needle holder, towel clamps, tissue forceps. (Johnson & Johnson Professtonal, Raynham, MA), 6 Sutures 3-O Vrcryl and 2-O Vrcryl (Ethicon, Sommerville, NJ) 7 Monopolar electrocautery (Hyfrecator Plus, Butcher Medical Systems, Irvine, CA)

2.2. In terfacially Formed Thin Hydrogels for Vascular Wound Healing 2.2.7. lnterfaclal Photopolymerization of Thin Hydrogel Films 1 Reagents. eosm-Y, trrethanolamme, N-vmylpyrrohdmone 2 Argon ton laser (American Laser, Salt Lake City, Utah) and 200~)*m core optical fiber (Spectra-Physics, Cranbury, NJ) 3 Solution 1. eosm-Y is dissolved m PBS at a concentration of 0 05% w/w Solution 2: to 50 mL of 0.67 g vacuum-dtstrlled trtethanolamme 1s added, mixed well, and pH 1s adJusted to 7 4 using 0 6 N hydrochlorrc acid The photopolymertzable macromonomer IS drssolved m the PBS-trtethanolamme solutron at a concentratron of 10% w/v N-vinyl pyrrohdone is dtsttlled over argon, and then added at a concentratton of 0.5 PL/l mL, to form the macromonomer solutron


Hydrogels

91

4 Inverted mtcroscope (Ntkon TMS, Avon, MA), mechamcal sizer (Boeckeler Instruments, Tucson, AZ), and microscope slides.

2.2.2. Rabbit Balloon Angioplasty injury Model 1 2. 3 4 5 6 7 8

9 10

Rabbits, anesthesia, betadine scrub solution, and hair clippers, as m Subheading 2.1.4. Polyethylene tubing (PE-50) 4F arterial embolectomy catheter (Baxter Healthcare) Silk suture ttes (4-O silk, Ethtcon) Vascular Clamps (Johnson & Johnson Professional) HEPES buffered salme solution (HBSS): 10 mn/p HEPES and 0 9% NaCl (w/w) m deionized water, pH adJusted to 7 4 using HCl and NaOH Neutral buffered formalm (10% w/w) Laser, soluttons, and mtcroscopy setup are as described m Subheading 2.2.1. All solutions are filter-stertlized m a lammar flow hood, and drawn up m a sterile fashion m syrmges Scanning electron microscope, critical point drier, and graded series of ethanol Heparm solutton 1000 UimL m 1 mL vials (Elkms-Smn, Cherry Hill, NJ)

3. Methods 3.7. Peritoneal Wound Healing 3.1.1. Synthesis of Photosensrtive Macromers 3.1 1.l . SYNTHESIS OF PEG-CO-P• LY(IX-HYDROXY ACID) COPOLYMERS (24) 1 The synthesis of the PEG-co-poly(a-hydroxy actd) copolymer is illustrated below, and the synthesis scheme is outlined m Fig. 1 PEG is dried by azeotroptc distillation with benzene as a 10% solution oL-Lacttde IS recrystallized from ethyl acetate prtor to use 2 50 g of dry PEG 10 kDa, 3 60 g of m.-lactide (5 mol oL-lactide per mol of PEG), and 15 mg of stannous 2-ethylhexanoate are charged mto a 100-mL round-bottom flask under a nitrogen atmosphere The reaction mixture is stirred under vacuum at 2OO’C for 4 h, and at 160°C for 2 h, and is subsequently cooled to room temperature 3 The resulting copolymer is dissolved m dichloromethane, precipitated m anhydrous ether, filtered, and dried (yield typically 95%) The c1- and o-hydroxyl end groups of PEGS with various mol wt can be used as rmg opening reagents to initiate the polymerization of either DL-lactide or glycohde, to similarly form several other copolymers (see Note 1)

3 1 .I .2. SYNTHESIS AND CHARACTERIZATION OF PEG MACROMERS 1 The above-described PEG-co-poly(a-hydroxy acid) copolymers, which are themselves a- and o-terminated by hydroxyl groups, are end-capped with acrylate groups to form a polymertzable macromer, the synthesis of the macromer with the PEG 1OK is illustrated 50 g of the aforementioned mtermedtate copolymer is dissolved m 500 mL of dtchloromethane m a 500-mL round-bottom flask, and IS cooled to O’C m an ice bath.

92

Sawhney and Hubbell

2 1 3 1 mL of trtethylamme and 1 58 mL of acryloyl chlortde are added to the flask, and the reaction mixture 1s stn-red for 12 h at 0°C and for 12 h at room temperature The reaction mixture IS filtered to remove trtethanolamme hydrochlortde, and the macromer IS obtained by pourmg the filtrate m a large excess of dry dtethyl ether It 1sfurther purified by dtssolutton and reprecrpttatton once, using dlchloromethane and hexane, respectively Finally, it 1s dried at 70°C under vacuum for 1 d This PEG macromer 1s called lOKL5 This indicates that the acrylate-terminated macromer IS synthesized from PEG segments of mol wt 10,000 (UL5), usmg nL-lactrde as the extension of the PEG a- and o-hydroxy end groups (lOKL_5), and the degree of polymertzatron of the lactoyl repeats (not lactidyl repeats, the cyclic dtmer of lactic acid, which 1sactually polymerrzed) is, on average, 5 per hydroxy end group (IOKLS) Several other macromers can be similarly synthesized by termmatmg PEG-co-poly(a-hydroxy acid) copolymers with acrylate or methacrylate groups (see Note 2). 3 The structure of the synthesized macromers can be verified using IR and NMR spectroscopy. NMR spectra can be recorded m CDCI,, and tetramethyl stlane 1s used as internal standard IR spectra can be run on a polyethylene film on which a drop of chloroform solutton of the macromer 1s placed and dried The dtsappearance of the -OH absorbance at 35 IO/cm confirms the complete acrylatton of the macromer. The presence of lactate ester 1s confirmed by an absorbance at about 1750/cm The degree of mcorporatton of lactate morettes can be calculated using the integral NMR mtensmes of the -CHs (1 3 ppm) and the -CH,O (3 63 ppm) of PEG

3.1.2 L WUV Induced Photopolymerizatlon The macromer, synthesized as described above, 1s used to prepare a photopolymerrzable macromer solution, as described m Subheading 2.1.3. 0.1 mL of this photopolymertzable macromer solutton 1s poured onto an 18 x 18-mm glass cover slip, and n-radiated using a low-intensity, portable, LWUV lamp, until gelation occurs. The time required to induce nontacky gelation 1s noted. This can be accomplished by continuously scratching a film of the macromer solution with a sharp mstrument, until the scratched film retains the scratch mark. The chemical reactions occurring during this photopolymerizatton process are outlmed m a schematic m Fig. 2 Vrstble ltght can be used for the photopolymertzatton process, also. (see Note 3, Fig. 3).

3 1 3. Rabbit Uterine Horn Model The animals are placed in two groups: 7 for control (injured, but not treated), and treatment (injured and treated) for the evaluation of the wound healmg tn the absence and presence of the photopolymerized hydrogels 1 Anesthesia* The rabbtts are anesthetized with an mtramuscular mlectton of the anesthetic cocktall. An intravenous catheter IS placed m the ear vem and the am-

In Situ Pho topolymerized Hydrogels

2,2-Dlmethoxy,

93

2-phenylacetophenone

3

+

l CH3

Methyl Benzoate

Photopolymerizatlon and gelatron

Fig 2. Schematic showing the mltlatlon and photopolymerlzatlon sitive macromers using LWUV light

of the photosen-

mals are maintained under anesthesia on periodic infusions of ketamme and Xylazme, as necessary 2 Surgery* The abdomen IS shaved and prepared with Betadme. A midline mclslon 1smade in the lower abdominal region, using the scalpel Any bleeding encountered during muscle mCislon IS controlled usmg electrocautery The abdominal retractor 1splaced to allow good vlsuallzation of the visceral organs. The uterine horns are located and the vasculature to both horns 1s systematlcally cauterized using monopolar eiectrocautery (setting at 30), to induce an lschemlc qury to the horns. After cautenzatlon, 1 0 mL of the polymerlzable solution IS applied along the obverse side of both horns, and also along the cauterized injury area, by dripping the solution through the syringe and gently rubbing the solution mto the tissue, usmg a gloved finger After uniform application of the solution, the horns are exposed to the LWUV lamp, at an lrradlance of 8-10 mW/cm*, for 1 mm, to

Sawhney and Hubbell

94

COOH Light

-*

Br

[3*1

+

Br

:N(CH,CH,OH),

514 nm

Excited

triplet

f3*1 COOH

M -

H+ -CHCH20H COOH Macromer

Polymerlzatlon

NaO Br

Fig 3 SchematIc showing the vlslble-light-initiated

photopolymerlzatlon

Br Almost colorless

process

induce gelatlon The procedure IS repeated on the reverse side of both horns and the cauterized area The mclslons are then closed using a contmuous 2-O Vlcryl suture for the musculoperltoneal layer, and a 3-O Vlcryl suture for the cutaneous layer No prophylactic antlblotlcs are administered For the control group, the lschemlc injury 1s made as described above, and the mclslon IS closed without the apphcatlon of the polymerlzable solution or illummatlon with the LWUV light All techniques should be identical between the treatment group and the control group, and all animals should be operated on by the same researcher 3 Evaluation The animals should be reoperated at the end of 2 wk under ketamme anesthesia (35 mg/kg, administered mtramuscularly) to evaluate adhesion formation, and are sacrificed by IV rnJectlon of euthanasia solution Gross observations are recorded as to sites of adhesion formation and organs involved m adhesion formation Presence or absence of hydrogel material should be noted, as well as any gross evidence of inflammation, necrosis, hematoma or seroma formation, and presence of cysts or infection Mlcroscoplc confirmation of any gross changes should be carried out by exclslon of site involved and hlstologlcal processing Quantitative assessment of adhesion formation 1s evaluated using scoring sys-

In Situ Photopolymenzed Table 1 Grading of Postoperative Grade of Adhesion Grade 0 Grade 1 Grade 2 Grade 3 Grade 4

Hydrogels

95

Adhesions

Type of Adheston Formed No adhesions Tentattve transparent adhesions that frequently separated on therr own Adhestons that gave some resistance, but can be separated by hand Adhesions that require blunt mstrument dissection to separate Dense, thick adhesions that required sharp instrument dtssectron m the plane of the adhesion to separate

terns for extent and tenacity Extent of adhesion IS evaluated by measuring the length of the uterine horn that has formed adhesions, and/or measuring the total length of the uterine horn after lysts of adhesions Adhesions are described as a fraction of the length of the horn that IS mvolved m adhesion formation Tenacity of adheston 1s classified according to the numerical scale depicted m Table 1 At least two researchers should be present for independent scormg of all ammals. Results are analyzed using ANOVA, with the Fisher PLSD post hoc test used to assess slgmficance, comparing the control group (InJured, but not treated) and treatment group (inmred and treated) The sues of adhesion formation can be excised for htstology and fixed m 10% neutral buffered formalm Gomort’s or Masson’s trtchrome stain are used for the vtsuahzatton of collagen m the adhesrons The sites of dense adhesion will stain green and show well-organized, collagen-rich tissue

3.2. In terfacially Formed Thin Hydrogels for Vascular Wound Healing The mterfaclal photopolymerlzatlon process takes advantage of the fact that the eosm-Y 1s antomc m nature, and thus can adsorb romcally to the surface of

tissue through mteracttons with postttvely charged functional groups present The polymertzable soluttons are prepared so as to Isolate the photomltiator and adsorb tt on to the vessel wall. Thereafter, the polymerrzable solution contammg the photosensmve macromer, but lacking the photomttlator, IS infused, and the tissue IS rllummated. All components needed for the photopolymertzatton process, i.e , light, photomtttator, and the polymerrzable solution, are present only at the vessel wall. This causesthe mterfactal growth of a thm hydrogel at the vessel wall. As the photomrtrator 1sdepleted (see Fig. 3 for reaction schematic), the reaction becomes self-hmmng and ceases after the hydrogel has grown to a certain thickness.

96

Sawhney and Hubbell

3 2 1 Interfacial Polymerization of Thin Hydrogel Films on Aork Tissue The lummal tissue surface of a 1-cm2 ptece of porcme aortx tissue 1spatted dry with a paper towel. Solutton 1 1sapplied to the entire lummal surface, and then coprously rinsed with PBS. The eosm-Y-stained tissue IS submerged m 0 5 mL of solutton 2 and tllummated with an argon ion laser operated at multiline, multimode through an optical fiber at an lllummation mtensity of 70-100 mW/cm2. The mterfacrally polymerrzed piece of tissue 1s placed m PBS for at least 1 h to allow the hydrogel to completely hydrate. A vartety of forrnulatton and photopolymertzatton condmons can be changed to mampulate the thtckness of the resulting hydrogel. A summary of the photopolymerrzatron condmons that can be employed to produce films of various thicknesses appears m Table 2 (see Note 4) Sections,lessthan 0.5~mmthick, of fully hydrated, mterfacially formed hydrogel-porcine aortrc tissue laminate, prepared as described above, are cut with a razor blade applied roughly to the center of the sample and movmg from the hydrogel side downward to the tissue side. The cross-section IS placed on a glass mrcroscope slide, flooded with PBS, and observed under an mvertedstage mtcroscope equipped with a mechanical sizer wrth a resolutron of 1 pm Five thickness measurements are taken along the length of each sectton, using the mechanical sizer 3.2.2. Rabbit Balloon-Angioplasty

Iqury Model

A balloon denudatron mJury m the carotid artery of the rabbit IS employed as a model more closely related to balloon angroplasty Male New Zealand white rabbits are used, and no anticoagulant 1sadmmrstered prtor to surgery 1 Anesthesia. Induction and maintenance of anesthesia is identical to that m Subheading 3.1.3. 2 Surgery and mp.uy process The neck region of the rabbits 1s shaved and prepared using a Betadme scrub solution The left carotid artery is surgically exposed through a midline mcision. Any bleeding from the skm mcision or the muscle layers IS controlled using electrocautery and a zone is isolated by placing atraumatic vascular clamps on the internal, external, and common carotid artertes The PE-50 tubing 1s inserted and the isolated zone is flushed with HBSS several times, to remove blood from the isolated zone The tubmg is removed, and a 4F arterial embolectomy catheter is Inserted through the internal carotid artery The balloon on the catheter IS inflated and dragged through the common carotid artery This process IS performed a total of three times by a surgeon who is blinded to treatment grouping 3 Hydrogel deposition* The balloon catheter is removed, the tubing is replaced, and the isolated zone is flushed with solution 1 (containing the eosm-Y) This solution 1s allowed to contact the vessel for about 30 s, and is then flushed away

Table 2 Thickness

Values of Interfacially

Formulation

Fi

Photopolymerized

Hydrogels

for a Variety of Conditions Illumination intensity

Component concentrationa

Plateau thickness

Eosm Y (mg/mL)

TEOA (mM)

(J.lLEL)

Macromonomer (w/v%)

(mW/cm2)

1

0 02

90

2

10

100

81 + 13

2

0 05

90

2

10

100

107 * 12

3

0 10

90

2

10

100

191 f 24

4

0 02

90

2

23

100

140f27

5

0 05

90

2

23

100

181 I!I 12

6

0 10

90

2

23

100

257 + 24

“TEOA,

tnethanolamme,

VP, N-vmylpyrrohdone

(w)

98

Sawhney

and Hubbell

using the HBSS The zone IS then filled with solution 2 Excess solutton IS allowed to flow out of the vessel and the external surface of the vessel IS lightly massaged with a gloved finger to contact the solutton well wtth the vessel wall Illummatton IS then performed externally, usmg the argon ton laser, as described above, for about 30 s, at an Intensity of 7&100 mW/cm2 The tubmg IS removed, the Internal carottd artery IS ligated, and the clamps are removed to allow blood flow The unpolymerized precursor IS rmsed mto the vasculature The mctsrons are then closed using a contmuous 2-O Vtcryl suture for the musculopernoneal layer, and a 3-O V~cryl suture for the cutaneous layer No prophylacttc antibtottcs are administered For the control group, the balloon mJury is made as described above, and the mcislon IS closed wtthout the applrcatton of the polymertzable solutton or rllummatton All techmques should be tdenttcal between the treatment group and the control group, and all ammals should be operated on by the same researcher 4 Re-exploration The snes of the injured segment of the common carottd artery are re-explored etther 2 h or 14 d after mmry, under ketamme anesthesta (35 mg/kg) An mtravenous dose of heparm (100 U/kg) IS grven to the ammal about 10 mm before re-exploration On vtsuahzatton of the m~ury sites, the patency of the arteries IS establtshed by clamping them distal to the sate of InJury and deposmon, gently mtlkmg them to remove blood present, and then releasmg the clamps to vtsuahze refill of the vessels wtth blood 5 Htstology The arteries are then clamped proxtmally and drstally, and flushed clean with salme and fixed using neutral-buffered formalm (10% w/w) under 1 5 m of aqueous solutton column pressure The tissues are allowed to fix m this fashion for 2-3 h and are then harvested and allowed to addmonally fix for 3-5 d prior to processmg for htstology Sbdes are stamed with etther Verhoeff s elastm stam or Masson’s trlchrome stam Mtcroscopy using the mechamcal sizer IS used to measure the cross-secttonal areas of the thrombus and lumen (2 h), and of the mtrma and medta (14 d) Half of each of three vessel segments from each group, randomly selected, are reserved for scannmg electron mtcroscopy to determme the presence of endotheha1 cells at the 14-d ttme-point The tissue IS cut open longltudmally, dehydrated m a graded ethanol serves, and drted by crmcal-pomt drying Samples can be coated with gold and Imaged by scannmg electron microscopy A region m each vessel segment, correspondmg to the approxtmate center of the Injured zone m the ammal, IS exammed

4. Notes 1 In addmon to using DL-lactlde, several other a-hydroxy ester linkages can be used to control the rate of absorptton of such hydrogels For example, glycoltde can be used to gave faster absorbmg hydrogels, and a-caprolactone can be used to form slower absorbmg hydrogels 2 In addmon to using acrylate motetles as the polymertzable end groups, other end groups, such as methacrylate or cmnamate, can also be used

99 3 Visible light from a xenon arc lamp or an argon ton laser can be used to nutlate the polymertzatron of these solutions. For example, a 10% w/v solution of macromer m PBS with eosm Y (0 5 mM) and triethanolamme (0.5 M> containing 0 5 pL/mL N-vinyl pyrrolidmone can be irradiated with an argon ion laser Light output should be m the range of 400-600 mW. 0 I mL of this macromer solutton is poured onto an 18 x 1s-mm glass cover slip and tllummated. The fiber 1sheld an appropriate distance from the sample, to get a 2-3-cm spot size Gelatton is noted as described. 4 The thickness of the resulting mterfactally polymerized thm hydrogels can be controlled by varying several components m the formulatton A samplmg of some of these formulatton condmons 1sgiven m Table 2

References 1 Hertzler, A E (1919) The Perztoneum, Mosby, St Louts, MO 2. DiZerega, G S and Rogers, K. E (1992) The Peritoneum, Springer-Verlag, New York 3 Ryan, G B , Grobety,-J , and MaJno, G (1973) MesothelialmJury and recovery, Am J Pathol, 71,93-l 12 4 Surgical Membrane Study Group (1992) Prophylaxis of pelvic sidewall adhesions with Gore-Tex surgical membrane. a multtcenter clinical mvestigation, Fertzl Sterll 57, 92 1-923

5 Haney, A F and Doty, E (1992) Murme peritoneal mJury and de novo adhesion formation caused by oxtdtzed regenerated cellulose (INTERCEED TC7) but not expanded polytetrafluoroethylene (Gore-Tex surgical membrane), Fertzl Sterzl 57,202-208 6 Schwartz, R S , Holmes, D R , and Topol, E J (1992) The restenosts paradigm revlstted. an alternative proposal for cellular mechanisms J Am Co11 Carhol 20(S), 1284-1293 7 Fuster, V , Badlmon, L , Badtmon, J J , and Chesebro, J H (1992) The pathogenests of coronary artery disease and the acute coronary syndromes N Eng J Med 32(S), 242-250 8 Ross, R , Rames, E W , and Bowen-Pope, D F (1986) The biology of plateletderived growth factor Cell 46, 155-169. 9 Edelman, E. R., Nugent, M A., Smith, L. T., and Karnovsky, M J (1992) Basic fibroblast growth factor enhances the coupling of mtimal hyperplasta and proltferation of vasa vasorum m qured rat arteries. J Clan Invest 89,465473 10 Linder, V , Lappr, D A , Band, A , Majack, R A , and Reidy, M A (1991) Role of basic fibroblast growth factor m vascular lesion formation. Czrc Res 68, 106113. 11 Lee, S. H and Ruckenstem, E ( 1988) Adsorptton of proteins onto polymeric surfaces of different hydrophthclties a case study with bovine serum albumm, J Collold Interface Scz 125, 365-379. 12. Buck, C. A and Horwttz A F (1987) Cell surface receptors for extracellular matrix molecules, Ann Rev Cell Bzol. 3, 179-205 13. Andrade, J D and Hlady,V (1986) Protein adsorption and materials btocompatrbihty a tutorral review and suggested hypotheses, Adv Polymer Scl 79, 163

100

Sawhney and Hubbell

14 Desai, N P and Hubbell, J A (199 1) Solution techmque to mcorporate polyethylene oxide and other water-soluble polymers into surfaces of polymeric biomatenals, Bzomaterzals 12, 144-153 15 Desat, N P. and Hubbell, J. A. (1992) Ttssue response to mtraperttoneal tmplants of polyethylene oxide-modified polyethylene terepthalate, Bzomaterlals 13,505-5 10 16 Ito, K , Tanaka, K , Tanaka, H , Imat, G , Kawaguchi, S , and Itsuno, S (1991) Poly(ethyleneoxide) macromonomers 7 Mlcellar polymerization m water, Macromolecules 24,2348-2354 17 Sawhney, A S , Pathak, C. P , van Rensburg, J J , Dunn, R C , and Hubbell, J A (1994) Optimization of photopolymerized broerodible hydrogel properties for adhesion prevention, J Blamed Mater Res 28, 83 l-838. 18 Hill-West, J. L , Chowdhury, S. M , Sawhney, A S., Pathak, C P , Dunn, R C , and Hubbell, J A. (1994) Prevention of postoperattve adhesions m the rat by m situ photopolymenzation of btoresorbable hydrogel barriers, Obstet Gynecol 83,5!&64 19 Hill-West, J L , Chowdhury, S. M., Sawhney, A. S., Pathak, C P., Dunn, R C , and Hubbell, J A (1996) Efficacy of a novel hydrogel and conventional hydrogel barriers, J Reprod Med 41, 149-154 20 Verreet, P. R , Faku Ohmann, C , and Roher, H D. (1989) Preventing recurrent postoperative adhesions an experimental study in rats, Eur Surg Res 21,267-273 2 1 Cuba-Geigy Corporation, Material Safety Data Sheet for Irgacure 651, Hawthorne, NY 22 Coohill, T P , Peak, M J , and Peak, J G (1987) The effects of the ultraviolet wavelengths present m sunlight on human cells zn wtro, Photochem Photobzol 46,1043-1050 23. Hill-West, J L , Chowdhury, S M , Slepran, M J , and Hubbell, J A (1994) Inhrbmon of thrombosis and mtimal thickenmg by m situ photopolymerization of thm hydrogel barriers, Proc Nat1 Acad Scz USA 91,5967-5971 24 Sawhney, A S , Pathak C P , and Hubbell, J A (1993) Bioerodible hydrogels based on photopolymerized poly(ethylene glycol)-co-poly(a-hydroxy acid) dtacrylate macromers, Macromolecules 26, 58 l-587

9 Fabrication and Implantation of Gel-Filled Nerve Guidance Channels Ravi V. Bellamkonda

and Robert F. Valentini

1. Introduction In human adults, the peripheral nervous system (PNS) IS capable of healing and regeneration. In order to reestablish function, nerve tissue must heal by true regeneration of a functtonal structure, since healing by simple scar will not reestablish electrical connectivity. Nerve guidance systems have been used experimentally to enhance regeneration through the use of functtonallzed gels, the delivery of growth-promotmg molecules, and the use of neuronal support cells or genetically engineered cells. The objectives of this chapter are to overview the methods used to construct gels for nerve sttmulatmg regeneration and to outline the surgical techniques to implant nerve guidance systems. The behavior of cells is influenced by both their intrinsic genetic programs and their extracellular environment. Humoral factors, cells, and the extracellular matrix (ECM) are the mam components of a cell’s external environment. Extracellular matrices modulate the organization of mtracellular cytoskeleton, cell differentiatton, and the spatial architecture of cells and tissues. The protein and proteoglycan macromolecules of the extracellular matrix are highly regulated and are thought to be organized to provide permissive and nonpermissive three-dimensional (3-D) pathways for neural cell migration and axon guidance during development (1,2). 1.1. The Problem of Regeneration It has been estimated that more than 200,000 nerve repair procedures are performed annually m the United States alone (3). However, complete recovery after damage to the nervous system is rare. Various treatments to prevent nerve degeneration have been attempted (45). These include coolmg of inlured From

Methods in Molecular Medmne, Vol 18 Tissue Engmeenng Methods and Edited by J R Morgan and M L Yarmush 0 Humana Press Inc , Totowa,

101

Protocols NJ

702

Bellamkonda

and Valentmr

nerves m an attempt to avoid ischemia (6), and the application of agents like naloxone (7) to avoid the consequencesof trauma. In most nerve mjuries, however, degeneration of the distal neural elements cannot be prevented. Neurons m adults are not capable of prohferatmg. Therefore, recovery of the injured neuron depends on its abihty to survive and regenerate Nerve regeneration thus may be defined as the regrowth of mjured axons, followed by restoration of their origmal synaptic connections and the recovery of then physiological functions. Neurons of differing species, and different neurons within the mammahan species, have differing abilities to regenerate. Older animals regenerate less well than younger ones. In mammals, the neurons of the PNS have a relatively greater ability to regenerate, compared to neurons of the central nervous system (CNS), which seldom regenerate (8,9) 1.2. Lack of Appropriate Environment Contributes to Regenera five Failure One of the mam factors that contributes to the differential regenerative abillties of peripheral and central neurons 1s the environment m the adult PNS, compared to the environment m the adult CNS The environment, m particular, the extracellular matrix (ECM), plays a critical role m both development and regeneration of the PNS and CNS (1,10). One of the major differences m the PNS and CNS environments is m the cells that myelinate neurons m the two systems. The PNS IS myelmated by Schwann cells, which, m response to Injury, secrete growth factors and neurite promotmg basement membrane proteins, such as lammm (LN) (1,11-13). Studies show that the use of PNS nerve grafts (which carry Schwann cells) may facilitate limited regeneration of lesioned nerves m the CNS (14,15). In contrast to the PNS, CNS neurons are myelmated by ohgodendrocytes, which express molecules that inhibit neurite growth and are thought to contribute to the regenerative failure m the CNS (16) Even within the PNS, the spinal dorsal roots, which lie close to the junction of PNS and CNS, do not regenerate (17), and present a challenge to regeneration efforts These studies mdicate that the environment of regeneration is a determmmg factor influencing the regenerative fate of the particular neural system It thus has been hypothesized that when polymer scaffolds contammg ECM protems, such as lammm and collagen, are presented to a severed nerve m VIVO,the regeneration of injured or transected nerves would be significantly enhanced. Hydrogels are popular materrals for 3-D polymer scaffolds for soft tissues such as nerves, because then- physical properties match well with soft tissues m viva. Permissive, hydrogel-based 3-D scaffolds are often functionalized with neurite-promoting proteins/agents, such as Lammm (18) or Schwann cells (9) to enhance the scaffold’s ability to support and promote

103

Gel-Filled Nerve Gudance Channels

Polymer Scaffolds for Nerve Regeneration

:1:

Synthetic Scaffolds

funct10n enhanced by

?

I

Collagen

Fig 1 Llstmg of 3-D polymer scaffolds covered m this chapter

axonal regrowth. Small, synthetic peptldes mimicking the function of larger ECM protems are also used. The important 3-D polymeric scaffolds used to support neurlte extension and growth m vitro and m VIVO are shown m Fig. 1 The followmg sectlons ~111 detail the synthesis and functlonallzation methods. 2. Materials

2.1. Polyhydroxye

thy/me thacrylate

(pHEMA)

1 Tns-(hydroxymethyl) ammomethane,sodium chloride, ammomum persulphate, sodmmblsulphlte, and sodium bicarbonate (Sigma, St LOUIS,MO) 2. Pemcllln-streptomycm (Glbco-BRL, Grand Island, NY) 3 2-Hydroxyethylmethacrylate (Polysclences,Warrmgton, PA) 4 Ethylene glycol, sodium hydroxide, and hydrochloric acid (Fisher Sclentlfic, Fair Lawn, NJ) 5 Untreated 24-well tissue-culture plastic trays (Corning, Cornmg, NY)

2.2. Po/yglycery/methacry/ate

(pGMA)

1 Glycldyl methacrylate (GdMA, 2,3-epoxypropyl methacrylate, Aldrich, Mllwaukee, WI) 2 (Dlmethylammo) ethyl methacrylate (DMAEMA, Aldrich)

704

Bellamkonda and Valen tlnl

3 Methacryhc acid (MAA) (Aldrich) 4 Ethylene glycol dlmethacrylate (EGDMA, Aldrich Chemical Co , Milwaukee, WI) 5 Ammonium persulphate and sodium metablsulphlte (Aldrich Chemical Co , MIXwaukee, WI)

2.3. MA TRIGEL@ 1 Matrlgel (Collaborative Research, Bedford, MA) Store at -20°C, thaw overnight at 4°C. Gels rapidly and irreversibly at room temperature of 22°C 2 Dulbecco’s modified Eagle’s medium (DMEM, Glbco-BRL) Store at 4°C Must be kept cold (on ice) before using to dilute Matrlgel 3 Calf Serum (FCS, Glbco-BRL). 4 Penclllm-streptomycin (Glbco-BRL)

2.3 1 /so/at/on of Schwann Cells (9). 1 Adult male inbred Fisher rats (250-300 g, Tacomc, Germantown, NY) 2 Vltrogen 100 (Collagen, Palo Alto, CA) Must be stored at 4°C Gels irreversibly at room temperature of 22°C 3 Ca*+/Mg*+-free Hanks’ balanced salt solution (HBSS, Sigma) 4 Trypsm (Sigma). 5 Collagenase 6 Hyaluromdase 7 Forskolm (2 M, Sigma) 8 Pltultary extract (10 g/mL, Sigma) 9 EDTA (Sigma)

2.4. Collagen

Type I (Vitrogen

100)

1 Sodium chloride, sodium hydroxide, and monophosphate (Sigma) 2 Type I collagen (Vltrogen 100) IS purchased from Collagen Store at 4°C

2.5. Agarose

Gels

1 SeaPrep@ agarose powder (FMC Bloproducts, Rockland, ME) 2 Dulbecco’s phosphate buffered salme (Glbco-BRL)

2.5.1. Peptlde and Laminln lmmobilizatlon to Agarose Gels Usmg Imidazole-Based Solution Chemistry 1. Agarose (SeaPrep, FMC) 2 Dulbecco’s PBS. 3 Lammm (mouse, Glbco-BRL) or lammm peptldes CDPGYIGSR, SIKVAV, or GRGDSP (Anawa, Wagen, Switzerland) Store at -20°C 4 1 carbonyldilmldazole (CDI), albumin (bovme), acetone, and sodium blcarbonate (Sigma). Avoid exposure to moisture during storage. Moisture-sensitive, 5 Molecular sieves (4 A\, Fisher Scientific) Dry sieves by heating at 350°C for 2 h

Gel- Filled Nerve Guidance Channels

105

2.5.2. Peptlde and Laminin Immobilization to Agarose Gels Using Benzophenone-Based Photo Chemistry 1 2 3 4 5 6.

Agarose (SeaPrep, FMC) Dulbecco’s PBS Lammm (mouse, Glbco-BRL) Albumin (bovine), acetone, and sodium bicarbonate (Sigma) Molecular sieves and 1,4 - dloxane (Fisher Scientific) 4-benzoylbenzolc acid, succmlmldyl ester (Molecular Probes, Eugene, OR) Light-sensitive Store m dark 7 Gate MR-4 UV lamp (Glo-Mark Systems, Saddle River, NJ)

2.6. In Vitro Analysis of Polymer Scaffolds: Visualization of 3-D Neurite Extension It is important to quantify neurlte extension m 3-D polymer scaffolds, to optimize their design and functlonalizatlon with neurlte promotmg agents, such as lammin or Schwann cells. This 1s achieved by capturing images of neurlte growing m 3-D polymer scaffolds, using the system described below

2.6 1. DigItal Image Analysis System 1 Nlkon inverted light microscope model TMS-F (Nlkon, Tokyo, Japan) 2 Javelm Electronics JE7862 high-resolution monochrome CCD camera (Javelm, Japan) 3 Scion LG-3 frame grabber (Scion, Frederick, MD) 4 Power Macintosh 7200 computer (Apple Computer, Cupertino, CA) 5 NIH Image Software Package (developed at the U S National Institutes of Health, and available on the Internet at http //rsb mfo.mh.govlmh-image/)

2.6.2. Other Materials 1 Hemacytometer (Hausser Scientific, Horsham, PA) 2 Nerve guidance channels can be fabricated from a range of commercially able, tubular materials, as listed m Table 1

3. Methods 3.7. Hycfroxyethy/methacry/ate

Gel Synthesis

avad-

(see Note 1) (19,20)

1 2-hydroxyethylmethacrylate, ethylene glycol, and an aqueous buffer, usually 0 05 A4 Tris-HCl, 0 15 M NaCl, are mixed at 1 1 1 (by vol, so that the final HEMA concentration IS 33% w/v) 2 After degassmg under low pressure, and while on ice, 0 1 mL of 12% sodium blsulphlte and 0 1 mL of 6% ammomum persulphate solution are added to the mixture 3 Gels m the shape of flat sheets are prepared by using a form that consists of a sandwich of two precleaned glass microscope shdes spaced by two no 1 covershps

Bellamkonda and Valen tini

106 Table 1 Materials

Used for Nerve Guidance

Channels

Nonporous Polytetrafluoroethylene (PTFE) Polyethylene (PE) Sillcone elastomers (SE) Polyvmyl chloride (PVC) MIcroporous “Gortex,” expanded polytetrafluoroethylene (ePTFE) “Mllhpore” (cellulose filter) Semipermeable Polyacrylomtrlle (PAN) Polyacrylomtr~le/polyvmyl chloride (PAN/PVC) Polysulfone (PS)

4 This sandwich IS held m posltlon with an alligator clip, and the separation of the slides IS approx 0 5 mm 5 The polymer mixture is inJected between the glass shdes with a Pasteur plpet, and IS drawn mto the space by capillary action 6 Followmg InJection of the shdes with polymer, the shdes are polymerized at 37°C for 2 h 7 When polymerlzatlon IS complete, the sandwich form 1s separated and the sheet of polymer IS peeled from the form 8 These sheets are then exhaustively dialyzed agamst 0 05 M Tris-HCl and 0 15 A4 NaCl buffer solution, pH 7.4 9 When all of the ethylene glycol had been removed from the polymerized gel, the gels became crystal clear and are then ready to be prepared for cell-culture experiment 10 Disks or buttons of polymer are cut from the polymerized sheets with a no 10 cork borer, the radms of the resultant buttons IS 140 mm. 11 The buttons are then placed, l/well, into untreated Commg 24-well tissue culture trays 12 To sterlllze the gels, 1 mL of a 10% pemclllm-streptomycm antlblotlc solution, made m Puck’s Ca- and Mg-free salme, 1s added, and the trays are exposed for 2 h to UV germicidal lrradlatlon 13 The antlblotlc solutions are removed and the gels are incubated for 30 mm with media, prior to cell seeding

3.2. Polyglyceryl

Mefhacrylate

Gels (21,22)

1 GMA is obtamed by hydrolysis of glycldyl methacrylate m aqueous sulphurlc acid, according to the method described by RefoJo (1965) 2 The GMA IS further purified by dlstlllatlon under dynamic vacuum (1 33 x 10m3pa) to remove epoxy-type residues and traces of polymer 3 Hydrogels based on pGMA are prepared by radical polymerlzatlon with the crosslmkmg agent EGDMA m dH,O.

107

Gel-Filled Nerve Guidance Channels

4 Polymerlzatton 1s mtttated m a small glass tube, with the redox mtttator ammomum persulphate (6% v/v) and sodmm metablsulphtte (12% v/v) added m a 0 37 wt% ratto to GMA, at 60°C for 12 h 5 The hydrogels are washed m dH,O for a mmtmum of 4 wk and kept m water solution 6. The hydrogels are sterthzed by botlmg m dH,O

3.27. Copolymeric Poly(g/ycery/ Hydrogels (see Note 2) (27-23)

5

6.

8 9 10 11

mefhacrylafe)-Collagen

GMA ts obtained by hydrolyses of glyctdyl methacrylate m aqueous sulphurtc acid, according to the method described by RefoJo (1965) The GMA 1sfurther put tfied by dtstillatton under dynamic vacuum (1 33 x 1Om3pa) to remove epoxy-type restdues and traces of polymer Hydrogels based on pGMA are prepared by radical polymertzatton with the crosslmkmg agent EGDMA m dH,O. Polymertzatton IS mtttated m a small glass tube wtth the redox mtttator ammomum persulphate (6% v/v) and sodium metabisulphtte (12% v/v) added in a 0 37 wt% ratio to GMA, at 60°C for 12 h The hydrogels are washed m dH,O for a mmtmum of 4 wk, and are kept m water solutton The hydrogels are sterilized by botlmg m dHzO Collagen is mixed with a concentrated salt solution of 1 3 MNaCl m 0 2 Mmonophosphate, and with 0 1 MNaOH, m the proportton of 8 1 1 (v/v), and the final concentration of collagen is 2 4 mg/mL. The pH of the collagen IS adjusted to 7 2 The solutton 1s kept on ice to prevent spontaneous gelatton Dehydrated pGMA are re-equtltbrated by swelling m the collagen solutton for 2 dat4”C The gels are then mamtamed at 37°C m a humidified incubator, to allow polymertzation of collagen

3.3.1 Loading of pHEMA and pGMA Gels Into PA N-PVC Guidance Channels 1 Because pHEMA and pGMA both involve m situ free-radical polymertzatton, the authors think tt best to polymerize the gels after loading the PAN-PVC gutdante channels with the monomertc reagents (step 4 of Subheading 3.2.) 2 For pGMA-collagen gels, guidance channels loaded wtth pGMA may be lyophtltzed and reequtltbrated with collagen solutton, as described above

3.4. Matrigep Loading (see Note 3) (9)

into Nerve Guidance

Channels

Matrigel IS derived from an Engelbreth-Holm-Swarm (EHS) mouse sarcoma. All plpets and tissue-culture materials that contact Matrlgel must be cold, to prevent premature gelatlon

Bellamkonda and Valentmi

108 3.4 1. Schwann Cell /so/at/on

1 Matrlgel IS diluted with cold DMEM, so that the ratio of DMEM Matrlgel IS 7 3 (v/v). 2 The 30% Matrtgel solutton IS then loaded mto IO-cm long-nerve guide tubes by attaching a syringe with a 22-gage needle to one end and drawmg the Matrtgel mto the tube Care should be taken to avoid an bubbles 3 The tubes are then cut mto lo-mm-long tubes, then ends sealed with polymer glue, and stored for 24 h at 37°C m an mcubator to allow for Matrtgel gelatton 4 Sclattc nerves are collected mto DMEM, stripped of their epmeurmm, and chopped mto 1-mm2 pteces 5 The nerve pteces are placed on Vttrogen loo-coated Petrt dishes m DMEM supplemented with 10% fetal calf serum and penicrllm streptomycm (1000 U/mL) 6 Every 5 d, the nerve pieces are transferred to new Petri dishes, leaving behind the fibroblasts that migrate out onto the Vttrogen 100 After four such transfers, almost all of the fibroblasts leave the nerve explant and the explants are then dlssoctated by incubating the nerve chunks at 37°C for 2 h m Cat2- and Mg+*-free HBSS containing 0.3% trypsm, 0 1% collagenase, and 0 1% hyaluromdase The cells are then tnturated, washed, and cultured m DMEM with 10% FCS for 1 d The followmg day, the culture medium IS replaced with motogemc medium contaming DMEM, FCS (lo%), forskolm (2 n&I), and pttuttaty extract (10 mg/mL) For the next 4-6 d, Schwann cells are cultured m mltogemc medium at 37°C m a humidified atmosphere with 5% C02.

3.5. Schwann Ceil Seeding onto Matrigei Scaffolds: Sch wann Ceil Functionaiized Ma trigei: Loading into Nerve Guidance Channels (see Note 4) 1 Schwann cells are used to functlonallze and enhance the neurtte-promotmg properties of Matrtgel Here, Matrlgel IS used as a scaffoldmg material that presents Schwann cells to the regeneration environment 2 Schwann-cell-seeded Matrtgel scaffolds are introduced mto nerve guidance channels using a method similar to the one described above Except, instead of using a mixture of DMEM*Matrtgel at 70 30 (v/v), 80 1O6to 120 x 106/mL of Schwann cells are suspended m DMEM, and this cell-DMEM suspenston IS used to dilute Matrtgel m a ratio of 70.30 (v/v) 3 The cell-Matrtgel suspenston IS then loaded mto 1O-cm tubes, as described above

3.6. Type I Collagen

Gels (see Note 5) (24)

1 Collagen solutton (2 0 mg/mL) IS mixed with a concentrated salt solutton of 1.3 M NaCl m 0 2 M monophosphate, and with 0 1 M NaOH, m the proportion of 8 1 1 (v/v), and the final concentratton of collagen 1s 2 4 mg/mL 2 The pH of the collagen IS adjusted to 7 2 3 The solutton IS kept on Ice to prevent gelatton unttl cells are entrapped

Gel-Filled Nerve Guidance Channels 3.7. Loading

Collagen

109

Gels into PAN-PVC Tubes

1 Collagen gels may be loaded mto nerve guidance channels m a procedure simtlar to the loading of Matrigel mto tubes, using a 22 gage needle and syrmge to draw ltqutd collagen mto the tube (at 4°C) 2 The guidance channels’ ends are sealed with polymer glue and the channels placed at 37°C for 24 h 3 The channels are then cut to desired length to bridge a nerve gap

3.8. Functionalking Agarose Gels with Laminin or LamininPeptides, Using lmidazole Chemistry (see Note 6) (25,26) Four mL gels of 1% agarose in PBS are dehydrated by repeated washes m acetone, followed by dry acetone (dried under 4 A molecular sieves) A CD1 solution prepared m dry acetone (5 mg/mL) IS added to the acetone-washed agarose gels (5 mL/4 g gel) The activation reaction is allowed to proceed for 5 mm with gentle agitation Gels are then washed 5x with dry acetone for 5 mm per wash to remove unbound CD1 CDI-activated gels are made into 2% solution m 2 mL, 100 mA4 sodium bicarbonate buffer (pH = 8 5) by heatmg and agitation Lammm (0 5 mg) or peptide (2 0 mg) is dissolved mto 2 mL of the sodmm btcarbonate buffer, and the solutton mixed with the CDI-activated gel solutton. The couplmg reaction is allowed to proceed for 36 h under gentle agitation at room temperature The gels are further quenched m sodmm bicarbonate for 6 h at room temperature, and then washed thoroughly with PBS for another 48 h, lyophihzed, and redissolved to the desired gel concentration, typically 1 0% m PBS

3.9. Functionalking Agarose Gels with Laminin or LamininPeptides, Using Benzophenone-Based photochemistry (see Note 7) (27,28) 1 Four mL gels of 1% agarose m PBS are dehydrated by repeated washes m acetone, followed by dry acetone (dried under 4 A molecular sieves) 2 Benzophenene (BP) solution prepared m dioxane (1 mg/mL) is added to the acetone-washed agarose gels (2 mL/4 g gel), and the mixture is exposed under UV light (wave length 250 nm) for 1 h 3 Gels are then washed 5x with dtoxane for 5 mm per wash to remove unbound BP 4. Lammm (0 5 mg) or peptide (2.0 mg) is dissolved mto 3 mL of 100 mA4 sodium bicarbonate buffer (pH = 8 5), and this solution is added to the BP-activated gels (4 g gel) 5 The coupling reaction IS allowed to proceed for 36 h under gentle agitation at room temperature 6 The gels are further quenched m sodium bicarbonate for 6 h at room temperature, and then washed thoroughly with PBS for another 48 h, lyophilized, and redissolved to the desired gel concentration of 1.O% m PBS

170

Bellamkonda and Valentml

proteitt Ftg. 2 Schemattc deptctton of neural cell entrapment m peptide/protem-modified 3-D hydrogels

3.10. Preparation of Agarose Gels and Loading of Neural Cells into Agarose Gels for In Vitro Analysis of Hydrogel Scaffold (see Fig. 2 for schematic; see Note 8) 3 IO 1. Preparation of Agarose (for 10 mL of 1 0% Agarose Solutron) 1 Add 0 10 g of SeaPrep agarose powder to 10 mL PBS 2 Star the mixture over medium heat until the agarose is completely dissolved (usually 1&15 mm) 3 Sterilize the liquid agarose through a 0 2 pm syrmge filter mto a sterile container

3.10.2. Cell Entrapment In Agarose Gels 1 Coat the bottom of each well of a sterile 24 well cell culture dish with 200 pL of the sterilized liqutd agarose 2 Swirl the dish, evenly distributmg the liquid over the bottom of the well 3 Chill the dish at 4°C until the liquid agarose gels (usually about 15 mm, dependmg on the concentration). 4 Draw 200 r.lr, of the sterilized liquid agarose mto a l-cc syringe, set aside 5 Take up the desired volume of the cell and medium mixture mto a micropipet (typically 4.5 dorsal root ganglia pieces or 5O,OOC-100,000 embryonic cortical cells from rodent brains m 10 p,L or less of sterilized cell culture medmm) 6 Insert the tip of the micropipet mstde the bevel of the syrmge, creating a seal

Gei- Filled Nerve Gudance Channels

111

7 Quickly pull back on the syrmge plunger until the cell and medium mixture IS enttrely transferred mto the syringe 8 Carefully pump the plunger m and out to thoroughly mix the cells and the llquld agarose. 9 Add the contents of the syringe to the previously gelled agarose in the 24-well cell-culture dish 10 As described above, swirl the dish and chill until the entire well has completely gelled 11 Add 1 mL of cell-culture medium to each well 12 Add 80 & NGF to each well (when required) 13 Every 48 h, the cell-culture medium and the NGF should be replaced and the cells should be vlsuahzediquantlfied

3.11. In Vitro Analysis and Quantification of Neurite Extension in Hydrogel Scaffolds (29-31) 3. I 1 7. Cahbratron of the Digrtai image Analysis System 1 Place the hematocytometer under the hght microscope 2 Focus on the smallest area of the grid under the desired magmficatlon (one side of the smallest squares IS 50 pm long) 3 Switch the optlcal path of the mlcroscope to the CCD camera 4 Start the NIH Image software package. 5 Select Start Capturing from the Special menu (an image of the hematocytometer ~111 appear on the monitor) 6 AdJust the light and focus the microscope 7 Trace along a known length of a hematocytometer grid lme with the straight lme tool (typlcally, a 600 pm length) 8 Select Set Scale from the Analyze menu 9 In the set scale dialog box, change the units to micrometers and the known length to 600 0 10. Click OK and the caltbratlon IS complete

3.11.2. Measuring the DRG New-de Length Calibrate the system Place the gel to be imaged under the microscope. Using the binocular eyepieces, focus on a DRG. Switch the optlcal path of the mlcroscope to the CCD camera Start the NIH Image software package. Select Start Capturing from the Special menu. Adjust the light power and the focus of the microscope Trace along the length of one of the longest neurites with the freehand line tool Select Measure from the Analyze menu. 10 Select Show Results form the Analyze menu (the length 1s given m the units specified durmg the cahbratlon).

2. 3 4. 5 6 7 8 9

Bellamkonda and Valentini

112

11 Repeat this procedure for seven more of the longest neurttes (the average of these eight lengths 1s the average neurtte length for that parttcular DRG) 12 Repeat the entire process for 2-3 more DRGs

3.11.3. Measuring the DRG Spread

8 9 10

11 12 13 14

Calibrate the system. Place the gel to be imaged under the microscope Using the binocular eyepieces, focus on a DRG Switch the optical path of the microscope to the CCD camera Start the NIH Image software package Select Start Capturing from the Special menu Adjust the light power and the focus of the microscope (at thts point, image enhancement procedures can be performed) Trace along the outer edge of the cell body with the freehand selectron tool Select Measure from the Analyze menu. Trace along the outer edge of the gangltomc area with the freehand selectton tool (the ganghomc area IS defined as the area covered by the cell body and all of the neurites) Select Measure from the Analyze menu. Select Show Results form the Analyze menu (the area IS gtven m the square of the unit specified during the cahbratton). The spread IS defined as the difference of the total gangltomc area and the cellbody area divided by the cell body area Repeat this process for 2-3 more DRGs

3 11.4. Computing the Percent of Cells Extending Neurrtes (NOTE* The dtgnal image analysis system ts not required for this procedure) 1 Place the gel to be imaged under the microscope 2 Focus on a typical area of cells 3 With two counters, simultaneously count the total number of cells present m the gel and the number of cells that are extending neurltes 4 Repeat this procedure on several different areas of the gel, keeping a runnmg total (typically, at least 900 total cells are counted) 5 The percent of cells extending neurttes 1s calculated by dtvtdmg the number of cells extendmg neurttes by the total number of cells

3.12. implantation of Nerve Guides (see Note 9) In repair procedures using nerve guidance channels, the free or mobibzed ends of a severed nerve are postttoned withm the lumen of a tube and anchored m place wrth sutures (Fig. 3). Tubulatton repan- provides prevention of scar trssue mvasron mto the regenerating envtronment, provlsron of dn-ecttonal guidance for elongatmg axons, and maintenance of endogenous trophtc or growth factors (32,33)

Gel- Filled Nerve Gudance Channels 1. CREATION

2. PLACEMENT

3. NERVE-CHANNEL

OF 3-4

mm

NERVE

OF 6 mm GUIDANCE

RELATIONSHIP

113

DEFICIT

CHANNEL

AND

GAP

v/

1 O-O

SUTURES

DISTANCE

Ftg 3. Schemattcshowmg surgical placementof nerve gutde. Guidance channels are useful m experimental procedures evaluating the process of nerve regeneratton, because the gap distance between the nerve stumps can be precisely controlled, the fluid and tissue entering the channel can be tested and evaluated, the properties of the channel can be varied, and the channel can be filled with various gels, and so on as described ear-her Nerves of dtfferent sizes and with vartous dimenstons can be repatred wtth appropriately sized guides In the early 1980s the rat sciatic nerve model emerged as a useful system for studying nerve repair and for evaluating guidance channels. The spatialtemporal progress of nerve regeneration across a IO-mm rat sciatic nerve gap repaired with a silicone elastomer tube has been described m detail and provides a point of reference for evaluating other systems. During the first hours following repair, the tube fills with a clear, protein-containmg fluid exuded by the cut blood vessels m the nerve ends. The fluid contains the clot-forming protein, Iibrm, as well as neurotrophrc factors known to support nerve survival and outgrowth. By the end of the first week, the lumen is filled with a longitudinally oriented fibrm matrix, which coalescesand undergoes syneresis to form a continuous bridge between the nerve ends. The fibrin matrix is then Invaded by cells mtgratmg from the proximal and drstal nerve stumps. Thts includes Iibroblasts (which first organize along the periphery of the fibrm matrix), Schwann cells, macrophages, and endothelial cells (which form capillaries and larger vessels) At 2 wk, axons advancmg from the proximal stump are engulfed

114

Bellamkonda and Valentlnl

m the cytoplasm of Schwann cells. After 4 wk, some axons have reached the distal nerve stump and many have become myelmated The number of axons reaching the distal stump is related to the distance the regenerating nerve has to traverse and the length of original nerve resected. Guidance channels without filling gels do not support regeneration if the nerve gap 1sgreater than 10 mm, and if the distal nerve stump 1sleft out of the guidance channel. In these experimental models, the morphology and structure of the regenerated nerve IS far from normal. 3.12 1. Animals 1 Rats are the most commonly used ammal model for experimental nerve repair, and will be the focus of this methodology It should be noted that numerous small and large animal models exist 2 Male or female adult (25&350 g) animals,either outbred (e.g., SpragueDawley, CD) or inbred (e.g , Fisher) from a certified animal supplier, may be used

3 12.2 Anesthesia 1 Rats can be anesthetized using mtraperitoneally qected Nembutal (pentobarbitol sodium) at a dose of 50-60 mg/kg (Abbott Labs, Chicago, IL) 2 Supplemental anesthesia, if needed, can be achieved with Metofane (methoxyflurane) applied carefully vta a nose cone

3.12.3 Animal Prep The hair is clipped around the leg (hip, thigh, lower leg) to be operated on The ammal is positioned on an operating platform and tied or taped in place with the operated leg left free, to avoid tension. The shaved area IS bathed with an iodophore (e g., Betadme solution, Purdue Frederick, Norwalk, CT) and an dried Instruments and drapes should be sterilized using an autoclave 3.12 4. instruments Instruments for microsurgical procedures should be used, and include, at a mmimum: two pairs of #5 forceps, one or two rat tooth forceps, one blade handle, one pair of dissectmg scissors, one pair of fine scissors for cutting the nerve, sterile 5-O and 10-O monofilament nylon suture, and an operating or dissectmg microscope with magnification to x20 or x40. One or two pan of angled forceps is also useful. All instruments should be stainless steel and sterihzable, and are available from a variety of manufacturers. 3.12 5. Surgical Exposure An mcisron is made along the lateral thigh, from the proximal femur to the knee, usmg a number 10 blade. The femur can be used as a guide to approachmg the nerve and can easily be palpated through the muscle/soft tissue of the

Gel- F//led Nerve Gudance Channels

115

thigh. The underlying fascia IS incised along the whitened border indicating the Juncture of the vastus lateralis and biceps femorls muscles. The femur lies deep to this avascular plane and can be exposed without cutting mto the body of any muscle group, thus mmlmizmg blood m the field. The sciatic nerve can be exposed by pushing the hamstring muscle group posteriorly to expose the nerve. A retractor or suture can be used to retract the muscle while surgery IS underway. The nerve is encased in a mass of loose connective tissue, fat, and small blood vessels. 3 12.6 Nerve Guide Placement 1 At this point, an operating or dlssectmg microscope should be used. At least 1 cm of nerve is isolated by carefully dissecting away everythmg except the outer connective tissue sheath (e g , epmeurmm), which 1sghstening white m appearance Care should be exercised so as not to cut any blood vessels or to handle the nerve except by the epmeurmm usmg 5-O forceps The nerve should be handled gently, because axonal damage can occur 2 The exposed nerve IS divided with mlcrosclssors at the level of the mid-thigh and the nerve guide 1s positioned. Sutures (10-O) are placed as m Fig. 3. The first suture IS placed 1 mm m from the proximal end of the nerve guide, pulled through the lumen, placed through the epmeurmm 1 mm m from the end of the proximal nerve stump, placed back mto the lumen, and carried through to the outer wall, close to the original entry point. Pull the suture up fully after each placement The nerve IS pulled mto the tube by gently pulling on the suture/nerve until it moves into the tube The knot is tied using three ties. It may be easier to place the sutures loosely m the channel and nerve at one end before cutting the nerve The tube IS flushed with sterile salme to remove any trapped air bubbles A segment of distal nerve IS resected, to correct for the length of nerve guide used, and the same procedure IS used to posltlon and anchor the distal stump through outer tube to lumen to epmeurmm to lumen through tube

3 12.7 Wound closure The field IS inspected for bleeding, disrupted sutures, air bubbles m lumen, foreign material, and so on. The fascia 1sclosed using 3-4 interrupted sutures (5-O). The skin 1sclosed using interrupted 5-O sutures, or running, subcutlcular 5-O bloresorbable sutures. Wound chps may also be used. 4. Notes 1 HEMA Gels.When the HEMA gel concentrationwas droppedbelow 33%, In our experience, HEMA gels lost their ability to become optically clear after the ethylene glycol was washed away This of course poses problems for visuahzatlon of cells m vitro via optical mlcroscopy. On the other hand, 33% HEMA gels may be too stiff (or their porosity too low) for these gels to allow neurlte extension when neural cells are trapped within the gel matrix. This may be a problem with HEMA gel synthesis by this method

116


2 pGMA gels. pGMA gels, along with other acrylic gels, such as HEMA, are polymerized at the site of final use (or mstde nerve guidance channels), because they are trreversible gels and cannot be reformed mto another shape once polymerized This may pose practtcal difficulties for their use and application Also, the polymerized gels should be washed thoroughly to remove any remammg unpolymerized monomeric compounds to ensure biocompatibility, because acrylic monomers are often toxic 3 Matrigel Although Matrigel has been shown to be a potent promoter of neurtte extension from neural cells m vitro, its application m cluncal medicine IS m doubt, because it is derived from a mouse sarcoma, rendering difficult its FDA approval Also, m some instances it has been shown to mhibit nerve regeneration In addition, Matrigel, because of its source, may have great variattons m batch to batch lots at the manufacturer, and this factor should be considered m the experimental destgn usmg Matrigel Care also should be taken whtle handlmg Matrigel, because it is an n-reversible gel that gels rapidly at room temperature, and hence it should always be handled on ice and thawed at 4°C overmght, not at room temperature 4. Schwann cell loading onto Matrtgel scaffolds Care should be taken that tibroblast contammation is avoided m the Schwann cell culture Also, sufficient numbers of Schwann cells are critical to observe enhanced regeneration. A munmum of 80 x lo6 cells/ml may be required to observe an effect in viva when Schwann cells are seeded onto 3-D Matrigel scaffolds. 5 Collagen gels Collagen gels at 2 4 mg/mL sometimes precipitate m the process of gelation, and this may dimnush optical transmission. It is the authors experience that the rate of warming of the gel may Influence this Also, it IS the authors’ experience that Vitrogen 100 from Collagen is less likely to have this problem, compared to collagen type I from other sources 6 CD1 modification of agarose gels Modification of agarose gels wtth CD1 chemistry, and subsequent bmding of proteins (e g., lammm) or lammm peptides, may decrease the solubihty of agarose m PBS. If problems with solubulization occur, the authors suggest that the concentratton of CD1 solution m acetone be decreased m step Subheading 3.8, step 2., appropriately Also, the time of exposure of the above CD1 solutton to acetone-washed agarose may be decreased appropriately This would impact the yield (amount of protein bound to the gel), but CD1 has a high protem-bmdmg efficiency, and the decreased yields may still be effective m promotmg neurtte growth from neural cells m vitro and m viva 7 Benzophenone (BP) modification of agarose gels BP chemistry yields of protein conJugation to agarose gels are generally lower than that of CDI chemtstry However, the big advantage with BP chemistry is the possibility of spatially controlling peptide conlugatton m different areas of the agarose gel This may be useful for 3-D axon guidance m vitro and m vivo Care should be taken at the UV exposure step, so that the agarose polymeric chains are not broken because of prolonged exposure. This may result m loss of gelation properties of agarose gels

Gel-Filled Nerve Guidance Channels

117

8 In vitro vlsuahzatlon of neurlte extension In 3-D polymeric scaffolds (e g , agarose gels with embryomc dorsal root gangha). For optimal visuahzatlon (via optlcal microscopy) of neurlte extension of cells entrapped m 3-D matrices, the authors recommend that Nomarskl optics (for glass culture dishes) or Hoffman optics (for tissue-culture plastic dishes) be used. Phase contrast mlcroscopy does not yield the best results because of a halo of light obscurmg neurlte extension Imaging close to the neurons Also, it should be remembered that the neurite length measure m Subheading 3.11., step 2. is a 2-D projectlon of 3-D neurlte extension, and hence underestimates the actual neurlte length Measurement of neurlte extension in the ‘Z’ direction requires the use of specially design culture dishes (30) 9 Nerve guide placement. Care should be taken to perform all procedures aseptically and with the ald of an operating microscope. When lsolatmg the nerve, the nerve should be handled with caution; cutting of blood vessels, especially those near the knee, should be avoided. The sclatlc nerve bifurcates into the tibia1 and peroneal branches above the level of the knee, and the sural nerve branches off at midthigh, necessltatmg its sacrifice Care should be taken to remove any air bubbles from the nerve guide by rinsmg with syrmges contammg bubble-free sterde saline solution.

Acknowledgments The authors thank XlaoJun Yu and Laura Okun for their technical assistance m the preparation of this manuscript. The authors thank George P. Dillon for his technical assistanceand for assistance,in the preparation of figures used m this manuscript. References 1. Sanes, J R (1989) Extracellular matrix molecules Influence neural development Ann Rev Neuroscz

12,491-5

16

2 Purves, D and Llchtman, J (1985) Prmclples of Neural Development Smauer Associates, Sunderland, MA, pp 8 1-l 30 3 MedIcal Devices and Diagnostic Industry. (1985) August P. 3 4. De La Torre, J C (198 1) Spinal cord InJury, review of basic and applied research Spme 6,3 15 5. Puchala, E and Wmdle, W F (1977) The possibility of structural and functional restitution after spinal cord InJury, a review Exp Neural 55, 1 6 Beggs, J L and Waggener, J. D (1979) The acute mlcrovascular responses to spinal cord inJury Adv Neural 22, 179 7. Faden, A I , Jacobs, T P , and Holaday, J W (198 1) Endorphms m experlmental spmal InJury, therapeutic effect of naloxone Ann Neural 10,326 8 Hausmann, B., Slevers, J., Hermanns, J , and Berry, M (1989) Regeneration of axons from adult rat optic nerve. influence of fetal brain grafts, lammm and artlficlal basement membrane J Comp Neural 281,447-466

118


9 Guenard, V , Klettman, N , Morrtssey, T. K , Bunge, R P , and Aebtscher, P (1992) Syngenetc Schwann Cells dertved from adult nerves seeded m semtpermeable guidance channels enhance pertpheral nerve regeneratton J Neuroscz 12, 33 163320 10. Martmt, R (1994) Expression and functional roles of neural cell surface molecules and extracellular matrix components during development and regeneration of pertpheral nerves J Neurocytol 23, 1-28 11 Le Beau, J M , LIUZZI, F J., Depto, A S , and Vmtk, A I (1995) Up-regulation of lammm B2 gene expression m dorsal root ganglton neurons and nonneuronal cells durmg sciatic nerve regeneratton Exp Neurol. 134, 150-155 12 Combrooks, C J , Carey, D , McDonald, J A , Trtmpl, R , and Bunge, R P (1983) In vlvo and m vttro observattons on lammm productton by Schwann cells Proc Nat1 Acad Scl 80,385O 13. Bunge, R P and Bunge, M B (1983) Interelattonshtp between Schwann cell function and extracellular matrix productton Trends Neuroscz 6,499 14. Aguayo, A J , Vtdal-Sanz, M , Vlllegas-Perez, M P , Bray, G M (1987) Growth and connecttvtty of axotomtzed retmal neurons m adult rats with optic nerves substituted by PNS grafts lmking eye and the midbrain. Ann NY Acad Scz 495, l-9 15. Vlllegas-Perez, M P , Vtdal-Sanz, M , Bray, G M , Aguayo, A J (1988) Influences of peripheral nerve grafts to enhance neuronal survival, promote growth and permtt termmal reconnections m the central nervous system of adult rats [review] J Exp Blol 132, 5-19 16. Schwab, M. E (1990) Myelm associated mhtbttors of neurtte growth and regeneration m the CNS Trends Neuroscz 13,452-456 17 McCormack, M L , Goddard, M , Guenard, V , and Aeblscher, P , (1991) Comparrson of dorsal and ventral spmal root regeneration through semipermeable gutdante channels J Comp Neural 313,449456 18 Manthorpe, M., Engvall, E , Ruoslahtt, E , Longo, F M , Davts, G E , and Varon, S. (1983) Lammm promotes neurmc regeneration from cultured pertpheral and central neurons J Cell BIOI 97, 1882-l 890 19 Ctverchta-Perez, L , Fans, B , Lapomte, G , Beldekas, J , Letbowttz, H , and Franzblau, C (1980) Use of collagen hydroxyethylmethacrylate hydrogels for cell growth Proc Nat1 Acad Scz USA 77,2064-2068 20 Bergethon, P R , Trmkaus-Randall, V , and Franzblau, C (1989) Modtfied hydroxyethylmethacrylate hydrogels as a modeling tool for the study of cell-substratum mteractions J Cell Scz 92, 11 l-l 2 1 21 Woerly, S , Marchand, R , and Lavallee, C (1991) Interacttons of copolymertc poly (glyceryl methacrylate)-collagen hydrogels with neural tissue . effects of structure and polar groups Blomaterlals 12, 197-203 22 RefOJO, M F. (1965) Glyceryl methacrylate hydrogels J Appl Polymer Scz 9, 3161-3170 23. Elsdale, T. and Bard, J. (1972) Collagen substrate for studies on cell behavtor J Cell Blol 54,626-637. 24. Collagen Corporatton, Technical Product Notes on Vttrogen loo@‘, # 300 1- 1 l-0395.

Gel- Filled Nerve Guidance Channels

119

25 Hearn MTW (1987) 1,l - carbonyldumidazole- mediated unmobthzatton of enzymes and affinity hgands Methods. Enzymol 135, 102-l 17 26 Bellamkonda, R., Ramert, J P , and Aebtscher, P. (1995) Lammm ohgopeptide dertvattzed agarose gels allow three-dimensional neurrte extension m vitro J Neurosct Res 41,501-509. 27 Parker, J M R. and Hodges, R S (1985) Photoaffimty probes provide a general method to prepare peptide-coqugates from native protein fragments J Protezn Chem 3,479-489. 28 Bellamkonda, R , Dtllon, G P., and XiaoJun, Yu (1996) A hydrogel based 3D btopolymertc matrix for Nerve Regeneration Sot Neuroscl Abstracts 22,3 16 29 Lennard, P. R (1990) Image analysts for all, Nature 347, 103-l 04. 30 Bellamkonda, R , Ramen, J. P., Bouche, N , and Aebischer, P (1995) A hydrogel based three dlmenslonal matrix for neural cells J Blamed Mater Res 29,663%67 1 3 1 Shaw, E D , Salmon, and Quatrano, R S , (1995) Digital photography for the light microscope Results with a gated, video-rate CCD Camera and NIH-Zmage software Bzotechtuques 19,946955 32 Valentmt, R F , Aebtscher, P , Wmn, S R , and Gallettl, P M. (1987) Collagenand laminm-containing gels impede peripheral nerve regneratlon through semlpermeable nerve guidance channels Exp Neural 98, 350 33. Valentml, R. F , Sabatml, A M., Dana, P., Aeblscher. (1989) Polymer electret guldante channels enhance peripheral nerve regeneration in mice Brain Res 48,300

Small Animal Surgical and Histological Procedures for Characterizing the Performance of Tissue-Engineered Kenneth S. James, Mark C. Zimmerman,

Bone Grafts

and Joachim

Kohn

1. Introduction Developmg effective ttssue-engineered constructs for bone regeneration requires careful assessmentof the m viva bone response to novel biomaterials, scaffold architectures, and btologically augmented, tissue-engmeered constructs Both the implant material and scaffold architecture are known to sigmficantly effect the local tissue response (1-3). Consequently, m characterizmg the performance of new bone implants, it is prudent to establish materialdependent and scaffold-architecture-dependent bone-growth phenomena, m addition to the effect of biological augmentation, e.g., preseeded cells, growth factors, and cell-attachment proteins. Here we describe rabbit transcorttcal pm and trephme defect models, which, m combmation, yield a method to mvestigate such variables on bone regeneration. The necessary histological and histomorphometry procedures are also detailed. The transcortical bone-pm model is adapted from ASTM standard F98 l-9 1 Standard Practice for Assessment of Compatibility for Surgical Implants wath Respect to Effect of Materials on Muscle and Bone (American Society for Test-

ing of Materials, West Conshohocken, PA) Fabricated 2.0~mm diameter pms of the investigational material are implanted m the rabbit distal femur and proximal tibia to assess the cortrcal and cancellous bone response (2). The method is designed to specifically assessmaterial effects on bone. For example, the material may elicit a strong inflammatory response, osteoclastic bone resorption at the bone-implant interface, or a significant fibrous layer may form around the implant, preventmg direct bone apposition to the material. In the case of a degradable implant material, the effects of degradation products From

Methods Edlted

by

in Molecular J R Morgan

Me&me, and

Vol 18 Tmue

M L Yarmush

121

Engrneerrng

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

122

James, Zimmerman, and Kohn

and material resorption are also readily documented wtth this model Note that rabbits have life spans over 5 yr, allowing for short- and long-term experiments that can be performed relatively mexpenstvely The rabbit skull trephme defect model 1s designed to examine architectural and biological augmentation effects on bone growth. It can be used to analyze a number of quantitative variables, including the rate of bone formation and volume of tissue mgrowth (bone, cartilage, or soft tissue) (46). Two 8-mmdiameter defects are created m the rabbit skull, allowing for a paired compartson between implant configurattons or to a positive (autograft) or negative (empty) control defect. Note that thts is not a critical size defect model (defect size = 15 mm), but a delayed healing model (7). Kramer et al. (4) studied the normal healing process of unfilled 8.0~mm trephme defects m rabbit crania. Defects were observed at 1, 2, 4, 8, 16, 20, and 24 wk: 1 wk, areas of the origmal inner and outer tables were filled with fibrous tissue and spindle shaped cells; 2 wks, tongues of newly formed trabecular bone Jutted inward; 4 wk, new bone formed tapering tongues, which extended mto the center of the defect l-3 mm, along with a thin layer of cellular fibrous tissue between the advancmg bone fronts, 8 wk, increased bone formation into the center of the defect site, 1620 wks, bone bridging of the 8-mm defect was observed. In contrast, materials shown to be osteoconductive or osteomducttve have induced defect bone bridging by 24 wk (5,6). Consequently, time-points of 2,4, and 8 wk are typically Investigated, with longer time periods providmg msights mto possible material degradation effects and long term bone remodellmg.

2. Materials 2.7. Surgical Procedures 1 Implants* pins 2 mm m diameter x 2 cm, or porous scaffolds 8 mm m diameter x25mnl. 2. Male New Zealand white rabbits weighing approx 3 5 kg 3 Disinfecting solutions* 70% ethyl alcohol in dHzO, Betadme (Henry Schem, Port Washington, NY), and 0.2% mtrofurazone ointment 4 Anesthesia: ketamme hydrochloride (55 mg/kg), acepromazme maleate (1 .O mg/kg), xylazme (5 mg/kg), bupivacame hydrochloride (1 25 mg/site) 5. Antibiotics. enrofloxacm (10 mg/kg). 6. Analgesia: buprenorphme HCl (0 1 mg/kg). 7 General surgical supplies lubricant eye ointment (Vedco, St. Joseph, MO), disposable surgical drapes, sterile surgical gloves, sterile surgical gowns, scalpel handles, #lO and #15 blades, suture holders, scissors, small self-retaining retractor, small Sem retractors, towel clips, small battery operated cautery, forceps, hemostats, rongeur, osteotome, periosteal elevator, 3-O and 4-O resorbable sutures (Vicryl, Ethicon, Somerville, NJ), 3-O nylon sutures, 4x4 in and sterile gauze rolls, sterile saline, plastic spray sealant (Bard, Murray Hill, NJ)

Tissue-Engineered

Bone Grafts

723

8. Transcortical pin protocol surgical supplies manual hand held drill, 2-mm stamless steel drill btt, hammer, awl. 9 Trephme protocol surgical supplies pneumatic drill, universal T-handle chuck, two 8-mm stamless steel trephme bits (one with a center post and one without). 10 Euthamsa solution sodium pentobarbital

2.2. Histological

7

8 9.

10. 11

12

13

Processing

Dissection instruments. scalpel, forceps, sctssors, osteotome, rongeur, bone saw Medium square glass bottles (Fisher, Pittsburgh, PA, 03-325BB) 10% buffered formalin (Ftsher, SF1 00-4) 100% (200 proof), 95, 80, 70, and 40 ethyl alcohol soluttons m distilled water HEMO-DE (Fisher, 15-l 82-507A) clearing solution Dry benzoyl peroxide (Fisher, B-274-l). Benzoyl peroxide, shipped wet, is dried wtth the followmg procedure a Place the required amount m a beaker of 100% ethyl alcohol b Mix and let stand for 15 mm. c Filter solution to recover the benzoyl peroxide, usmg a vacuum filtering system d Place benzyol peroxtde m a shallow dish to allow ethyl alcohol to evaporate e When dry, wetgh out the amount of catalyst needed and store remaining benzoyl peroxtde m a closed metal container away from heat. Caution: Note that dry benzoyl peroxide 1shighly explosive Methylmethacrylate (Fisher, f&3629) PMMA solution I, PMMA solutton II (100 mL PMMA solution I and 1 0 g dry benzoyl peroxide), PMMA solution III (100 mL PMMA solutton I and 2.5 g dry benzoyl peroxide). PMMA solutton II and III are mtxed for at least 3 h on a stirrmg plate prior to use Plexiglas microscope slides and staining dishes. Cyanoacrylate glue, such as Permabond Industrial Grade 9 10 Adhesive (Permabond Internattonal, Englewood, NJ) Diamond-tipped electric engraver Stevenel’s blue stain, prepared by mixing solutions A (1 g of methylene blue [Fisher, M-2911 m 75 mL of dtstilled water) and B (1.5 g of potassium permanganate [Fisher, P279-5001) m a boilmg water bath, until the prectpatate dtssolves. Allow the stain to cool to room temperature and filter through medmm-grade filter paper Stam stored at room temperature is stable for many months and can be reused several times. Van Getson ptcro-fuschin stain prepared by first mixing 0.1 g acid fuschm (Fisher, F-97) wtth 10 mL of dHzO, followed by the addmon of 100 mL of saturated ptcric acid (Sigma, St Louis, MO, 925-40). Stam stored at room temperature is stable for many months and can be reused several times Specimen cutting, grinding, pohshmg stations, and accessortes equivalent to the Buehler Isomet (Lake Bluff, IL) 1 l-l 180 low-speed diamond saw and blade; Buehler Handimet II Roll Grinder, 240,320,400, and 600 grit Buehler Carbtmet abrastve paper rolls, Buehler Ecomet III Polishing Wheel, and 1 0 and 0 3 p Buehler alumma alpha micropolish II

124

James, Zimmerman, and Kohn

2.3. Histological

Interpretation

and Histomorphometry

1. Hardware* a stereomicroscope equipped with CCD camera and computer interface permtttmg digital image capturing We have found a system consisting of an Olympus BH-2 microscope, Apple Macmtosh IIci computer, 17-m color momtor, and a MT1 CCD72 video camera with adaptor box and video card (DageMTI, Michigan City, IN) works well 2. Software image analysis software comparable to NIH Image 1 60 (available for free download from the NIH at http.//rsb mfo.mh.gov/mh-image/download.html)

3. Methods 3.1. Surgical

Procedures

The transcortical pin and trephme defect models require approx 1-2 h to complete, mcludmg animal preparation, surgery, and recovery. A team of two (a surgeon and an anesthestologrst) can efficiently complete these procedures The rabbits tolerate the procedures very well and are typically found to be alert, werght bearmg, eating, urmatmg, defecatmg, and free of sigmfkant drscomfort 1 d postoperatron 3 1.1. Hard Tissue Transcortical Pin Biocompatlbility Model 1 Induce anesthesia with an mtramuscular (tm) injection of ketamme hydrochlortde and acepromazme maleate (doses given m Subheading 2.1.) 2 Administer preoperative antibiotics (enrofloxacm) and analgesia (buprenorphme HCI) im 3. Administer eye lubricant. 4 Shave the legs from the ankle to hip and bring ammal mto the surgical suite 5. Per normal sterile surgical procedure guidelmes, scrub the legs with betadme, rmse with 70% ethyl alcohol, wrap the feet, and drape the legs (see Note 1) 6 Administer an im mlection of xylazme lust prior to mitiatmg the procedure Supplemental mlections of xylazme at one-quarter the mitral dose is admmistered as needed to mamtam a surgical plane of anesthesia 7 Monitor and record the heart rate at 15-min intervals, either by manually palpatmg the pulse m an ear vessel or with an EKG machme 8 Create two l-cm medial mcisions along the knee to expose the proximal tibia and distal femur. Dtsplace the periosteum anteriorly and postertorly with the periosteal elevator 9 The implant sites are approx 7 mm from the dtstal end of the femur and the proximal tibia In practice, the tibia implant site is midway between the lomt space and tibia1 tuberosny. The femur implant site IS circa the proximal aspect of the medial condyle. To help guide the drill bit, create a small indention m the bone at the implant sites by lightly tapping on the awl with a hammer Drill 2 O-mm diameter holes m the ttbia and femur perpendicular to the long axis of the femoral-tibia1 shaft. Penetration of the medial and lateral cortices is easily felt with the manual hand drill

Tissue- Engineered Bone Grafts 10 Press-fit the implants mto place and carefully trim any excess material with scissors 11 Close the soft tissue m separate layers with resorbable (subcutaneous layer) and nylon interrupted sutures (skm) 12 Perform Implant procedure on the contralateral leg 13 Administer local anesthesia (buplvacame HCl) subcutaneously at each implant site. 14 Wash wounds of blood, dress with antlblotlc ointment, and coat with a plastic spray sealant 15 For pain management, administer, as needed, a subcutaneous mjectlon of buprenorphme HCl twice dally until the third postoperatlve day. 16 Antiblotlcs (such as enrofloxacm, 10 mg/kg) should be administered Irn dally until the third postoperative day 17 Remove the nylon skm sutures l&14 d postoperation

3.1.2. Trephine Model 1 Animal preparation, anesthesia, and monitoring are ldentlcal to steps l-7 m the Transcorrtlcal Pm Protocol, except for the surgical site Prepare a surgical site the size of a silver dollar on the rabbit skull, taking care to protect the eyes of the animal from the dlsmfectmg solutions. The rabbit skull can be stabilized with rolled towels and placed on either side of the rabbit skull Otherwise, an assistant can physically stabilize the head when necessary 2 Create a midline mclsion approx 2 5 cm m length through the skm along the saglttal suture of the skull, caudal to the eyes. Contmue the mclslon through the pellpheral muscles and perlosteum over the saglttal suture 3. With a small osteotome, carefully elevate the muscles and perlosteum Use small retractors to expose an area large enough for two 8 O-mm trephme holes to be respectively made on either side of the sagittal suture of the skull 4 Usmg the trephme bit with the post and the pneumatic drill, start the trephme hole and penetrate the other dlploe (see Note 2) Use the trephine bit without the post and universal hand chuck to slowly penetrate the mner dlploe Take care to contmually check for full penetration by pushing on the the trephme bone with forceps Once the trephme bone IS free, pry the circular bone fragment from the surrounding bone wrth forceps to expose the intact dura with viable blood vessels (see Note 3).

5 Place the test material mto the defect (see Note 4) 6 Repeat procedure for the second defect 7 Close the soft tissue in separate layers with resorbable (subcutaneous layer) and nylon interrupted sutures (skin) 8 Postoperatlve care IS m accordance with steps 13-16 m the Transcomcal Pw Protocol (see Note 5)

3.2. Histological

Processing

It IS critical that the fixation, dehydrating, clearing, and embeddmg procedures described below be performed on the pms and scaffolds prior to lmplantatton, to document the effects of these procedures on implant morphology.

126

James, Zimmerman,

and Kohn

Here we describe a tixatton, dehydration, clearing, and polymethylmethacrylateembedding technique that has been successfully employed m mvestigatmg the histological response of bone to degradable polymers (2,3). However, it is possible that some Implant materials may swell significantly or be soluable m one of these solutions. In such a case, it may be necessary to chose alternative procedures, and employ, for example, water soluable embeddmg media such as JB4 (Polysciences) (8). Caution: The fixation, dehydration, and embedding steps should all be performed under a fume hood. 3 2.1 Specimen Retrieval and fixation 1 Expose the implant sites through routine dissectton procedures 2 With a bone saw, grossly isolate the implant, taking care not to dtsturb the implant or surroundmg bone Dtssect away any remaining soft ttssues 3 Using a low-speed diamond saw with saline asthe lubricating fluid, make addlttonal bone cuts to isolate the Implant, whtle retammg approx 5 mm of pertphera1 bone 4 Place the Implant/bone m an approprtately labeled storage vial (medmm glass bottle) containing a volume of fixative (Formalm, see Note 6) equal to or greater than 10 times the specimen volume In most cases, allowmg the sample to remam m fixative overnight 1s adequate If necessary, change the flxattve solutton several times over the next few days, until the solutton remains clear

3.2 2. Dehydration, C/earing, and fmbeddmg 1 Sequentially place spectmens m 40% ethyl alcohol for at least 8 h, and m 70% ethyl alcohol for at least 16 h If required, the specrmen can be safely left m these alcohol solutions over the weekend 2 Replace dehydrating solution with 80% ethyl alcohol for 24 h Refresh the solution after the first 8 h. For this and the followmg dehydration steps, tt IS increasingly important that the spectmens soak m the specified alcohol solution for the stated duration 3 Replace dehydrating solution with 95% ethyl alcohol for 24 h Refresh the solution after the first 8 h 4 Replace dehydrating solution with 100% ethyl alcohol for 24 h Refresh the solutron after the first 8 h 5 Replace dehydratmg solution with Hemo-De clearing agent for 8 h Refresh the solution after 4 h 6 Replace the clearmg agent with PMMA solution I Allow to Infiltrate for 3-7 d 7. Replace embeddmg solution with PMMA solution II Allow to Infiltrate for 3-7 d 8 Replace embedding solutton with PMMA solutton III Allow to infiltrate for 3-7 d 9 Place bottle in 32°C water bath to polymerize methylmethacrylate solutron (see Notes 7 and 8) The solution should polymerize m l-3 d Check tf block 1s polymerized by removing the cap and touchmg the top of the block wrth forceps If it 1s soft (sticky), leave the bottle m the waterbath until the material has hardened

Tissue-Engineered

Bone Grafts

127

3.2.3. Preparing Hlstologlcal Slides 1 Break the embedded specimen out of glass bottle with a hammer. To prevent the glass from scattering, break the bottle within a thick polyethylene bag Placmg the bottle m the freezer for about an hour prior to breakmg helps the glass to fall away from the block Rmse the blocks m water to remove any small fragments of glass. Immedtately engrave necessary identification mformation mto the block. 2 Carefully orient the specimen and cut 0.5mm-thick secttons with the diamond saw Use pure ethyl alcohol as the lubricatmg fluid. For both the transcortical pm and trephme models, saggital and coronal sections, respectively, are routmely made through the implants For the implanted pins, cut three sequential sections, starting at either the medial or lateral cortex This yields corttcal and cancellous bone htstological sections For the embedded trephme defects, first cut the specimen m half, along the dtameter m the coronal plane, followed by 2-3 coronal serial sections. The remaming pieces of the embedded specimens are cut m the transverse plane to visuahze the implant sites from this perspecttve 3 Clean the section to be glued and the Plexiglas mtcroscope slide m a Petri dish filled with 100% ethyl alcohol and dry with a Kimwipe. Put a drop of cyanoacrylate on the slide. Place the specimen on the slide by holdmg the specimen with forceps at a 45degree angle and slowly dropping tt down on the glue and slide (similar to the way you would coverslip a slide). Put a piece of plastic wrap underneath the slide and fold it over to cover the slide. Clamp shde between two scrap Plexiglas slides Leave at room temperature overmght to make sure that the glue is set 4 Manually grmd the glued section on the Buehler Handtmet II Roll Grmder (or equivalent), using abrasive paper and water as a lubricant, until it becomes approximately one cell-layer thick (see Note 9). Begin by grmdmg on the lower grit papers and subsequently move to higher grit papers as the section becomes thinner. Take care to assure that equal pressure is applied to section, so that it grinds evenly. Turmng the slide occasionally helps to resolve thts problem Check shde frequently under the microscope to determine when the one cell layer thickness is reached, 1 e , only one dtstmct cell focal plane exists 5. Pohsh the ground slides on the Buehler Ecomet III Polishing Wheel (or eqmvalent) until all the scratches are no longer visible under the microscope Imtially use a suspenston of 1.O alumma polishing particles, followed by 0 3 alumma particles, for the final polished finish

3.2.4. Staining 1 Fill staining dashes with Stevenel’s blue and Van Geison’s ptcro-fuschin Place stammg dish with Stevenel’s blue and a beaker of dHzO m a 60°C water bath and bring up to temperature 2 Soak slides m Stevenel’s blue for 10 mm, rmse m dHzO, and blot dry. 3. Soak slides m Van Gelson’s picro-fuschm for 2 mm (room temperature), rinse m a beaker of 100% ethyl alcohol, and blot dry

128

James, Ztmmerman, and Kohn

3.3. Histological

Interpretation

and Histomorphometry

The Stevenel’s Blue and Van Getson ptcro-fuschm combmatton dtstmctively stams trssues over a wide color spectrum: extracellular structures and nonmmerahzed tissues, shades of blue; collagen fibers, green to green-blue; bone, red-orange or purple; osteord, yellow-green; muscle fibers, blue to bluegreen. In examining hrstologrcal sectrons of the transcortical pms (Fig. 1) or trephme defects (Fig. 2), we routmely scan the slides for changes m pm drmenstons, surface roughness and scaffold morphology, evidence of direct boneimplant contact, the presence of mtervenmg fibrous tissue between the Implant and bone, the quahty of the bone surroundmg and/or penetrating the Implant (is tt woven or lamellar), the number of osteoclasts at the bone-Implant mterface and whether significant osteoclastrc bone resorptron IS taking place, and the presence of inflammatory cells (Fig. 1). A calibrated hrstomorphometry system, i.e., calibrate the system, usmg a 1.O-mm cahbratton bar mtcroscope slide, IS used to accurately measure attributes such as the average thickness of any fibrous layer. Histomorphometric analysis of the the trephme defect slides also includes measuring the linear extent of bone growth across the defect and the area of the defect site occupied by bone. These values are normalized to the length or area of the orrgmal defect (6). 4. Notes We have found it convenient to place the rabbit’s feet withm sterile surgical gloves, which are subsequently wrapped with sterile gauze The trephme cutter with a post 1sa key component m starting the trephme defect We have also found that it 1ssometimes easier to start the trephme with the drill rotating m reverse The trephme teeth are large and by rotatmg the cutter m reverse, the cutter will simply start the trephme and cut very slowly. The droll is then reversed back once the trephme has been mitiated One possible major comphcatton 1s the drsruptton of the dura. If the trephme cutter is used too aggressively, the dura may be disrupted, resultmg m bleeding and local necrosis The implant material will reduce the bleedmg and the ammals should be carefully observed for neurological deficits and/or comphcations An autograft bone-control implant IS readily made by morsehzmg the trephme bone with a rongeur and carefully packing the defect Likewise, a negative control site is created by leaving the site empty. Some have found it useful to mcorporate florescent bone label mJections, i e , oxytetracyclme, 2-4-bzs(N,N’-dicarboxymethyl ammomethyl) fluorescem (DCAF), and so on, to identify areas of bone formatron at mtermedtate time pomts (6) If florescent bone markers are being used, e.g , oxytetracycline, the specimens should be fixed m 70% ethyl alcohol, because Formalm causes the florescent dyes to fade Close the cap ttghtly, because oxygen mhibits the PMMA solution polymertzanon reaction Do not shake or tilt the bottle. The PMMA solution will dissolve

Tissue-Engineered

Bone Grafts

129

Fig. 1. A representative sagittal section through a polymer pin (clear circular area) implanted in the rabbit distal femur for 9 mo. A ring of lamellar bone (red) surrounds the implant, which approximates the original shape and dimensions of the original 2-mm diameter pin (A). A very thin (2-3 cell layers) fibrous layer (blue) is found between the implant and bone, and very few osteoclasts are found at the boneimplant interface (B). The surrounding marrow cavities are free of inflammatory cells. In summary, the material elicited a very mild inflammatory response. However, it appears the implant does not promote direct bone bonding.

the bottle cap if it comes in contact with it. Place the bottles in a 32°C water bath under a fume hood. Besides warming the methacrylate to facilitate polymerization, the water serves as a heat sink for the energy generated by the polymerization reaction, that if not dissipated, will cause the the PMMA to bubble. Make sure that all four sides of each bottle are surrounded by water and that they are not touching other bottles. Check the water level and the temperature of the water bath often. Keep the water level as high as it can be without reaching the neck of the shortest bottle. The PMMA solution should polymerize in l-3 d. 8. If the PMMA bubbles significantly, the specimen can be re-embedded by first dissolving in PMMA I, and subsequently moving on to PMMA II and PMMA III

James, Zimmerman, and Kahn

Fig. 2. Coronal (A) and transverse (B) sections of negative control (empty) S-mm trephine defects at 12 wk. Bone (red) has successfully bridged the defect, though the inner and outer diploe have yet to fully form. Bone fragments are interspersed with fibrous tissue (blue) and osteoid (blue-green). Note that measurements of linear bone ingrowth are most conveniently made on the coronal sections, and bone area measurements on the transverse sections. per the normal embedding schedule. Recheck or lower the temperature of the water bath to prevent further bubbling. 9. To assist in gripping the slide during grinding, we machined a piece of Plexiglas that could be comfortably gripped and consisted of a number of 1.5-mm diameter holes in the bottom surface that spanned the width and length of a typical microscope slide. A vacuum is drawn through the holes by an attached vacuum line, thus gripping the slide.

References 1. Mikos, A. G., Sarakinos, G., Lyman, M. D., Ingber, D. E., Vacanti, J. P., and Langer, R. (1993) Prevascularization of porous biodegradable polymers. Biotechnol. Bioeng. 42,7 16723.

Tissue-Engineered

Bone Grafts

131

Ertel, S. I , Kohn, J , Ztmmerman, M. C , and Parsons, J. R (1995) Evaluation of poly(DTH carbonate), a tyrosme-derived degradable polymer, for orthopaedtc apphcattons. J Blamed Mater Res 29, 1337-1348 Choueka, J , Charvet, J L , Koval, K J , Alexander, H , James, K S , Hooper, K A., and Kohn, J (1996) Canme bone response to tyrosme-derived polycarbonates and poly(L-lactic acid) J Boomed Mater Res 31,35-41 Kramer, I., Kllley, H , and Wright, H. (1968) A htstologtcal and radtologtcal compartson of the healmg of defects m the rabbit calvarmm wtth and wtthout implanted heterogenous inorganic bone Arch Oral Bzol 1095-l 106 Damten, C J., Parsons, J R , Benedict, J J , and Wetsman, D S (1990) Investigation of hydroxyapatite and calcmm sulfate compostte supplemented wtth an osteomducttve factor J Blamed Mater Res 24,639-654 Damten, C J , Parsons, J R , Prewett, A B , Rtetveld, D C , and Zunmerman, M C (1994) Investtgatton of an organic dehvery system for demmerahzed bone matrtx m a delayed-healing cranial defect model J Blamed Mater Res 28,553-56 1 Frame, J. (1980) A convenient model for testmg bone substitute materials J Oral Surg 38, 176180 Burg, K J L , Jenkms, L , Powers, D L , and Shalaby, S W (1996) Spectal constderattons m embedding a lactide absorbable polymer J Hzstotechnol 19,3943

11 Preparation and Use of Porous Poly(a-Hydroxyester) Scaffolds for Bone Tissue Engineering Anna C. Jen, Susan J. Peter, and Antonios

G. Mikos

1. Introduction Skeletal defects resultmg from tumor resection, congenital abnormalities, or trauma often require surgical intervention to restore function. Current options for bone replacement include autografts, allografts, metals, ceramics, and polymeric bone cements. However, all of these materials have drawbacks, and their selection usually requires some degree of compromise. Autografts represent the ideal repair material, but are hmited by availabihty and donor site morbidity Allografts may be potential transmitters of disease, and also solicit immune response if not sufficiently pretreated. Ceramics suffer from slow integration and remodeling, and wear-debris from nondegradable polymeric implants may evoke chronic mflammation. Finally, metallic implants may cause atrophy of surrounding tissue through stress shielding, reqmrmg corrective procedures. Recent mvestigations mto tissue engineering implants offer alternatives to the current methods of skeletal repair. Synthetic biodegradable polymers are promismg as extracellular matrix analogs to facilitate tissue development and growth (2). Poly(a-hydroxyesters) are among the few synthetic biodegradable polymers approved by the Food and Drug Admmistration for human clmical use. These include poly(glycohc acid) (PGA), poly(L-lactic acid) (PLLA), poly(n,r-lactic acid) (PLA), and copolymers of poly(oL-lactic-co-glycohc acid) (PLGA). All are biocompatible, degrading to products that can be eliminated from the body through either metabolic pathways or by direct renal excretion (2) In addition, the PLGA copolymers have the distinct advantage of being capable of degradmg from weeks to years, as required for a specific apphcanon, by altering their copolymer ratio (3). From

Methods Edlted

by

m Molecular J R Morgan

Medicine, and

Vol 18 T/ssue Engmeenng

M L Yarmush

133

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

Jen, Peter, and Mlkos The ideal bone substitute should be osteoconductlve, osteomductlve, structurally sound, and, if possible, provide an osteogenic cell population (4). Cell presence 1sbelieved to enhance healing and promote bone formation A tissue engineering approach using polymer-cell constructs attempts to improve upon previous methods by satlsfymg all such requirements (5,. Advances m cell culture also allow for m vitro expansion of bone forming cells, which reduces harvest requirements We have studied the function of primary rat calvarlal osteoblasts on different poly(a-hydroxyesters) and showed that these polymers support cell adhesion, prohferatlon, and mlgratlon (6,7). In addition, we have successfully cultured rat marrow stromal osteoblasts on three-dlmenslonal (3-D) PLGA foam scaffolds and showed the formatlon of bone-like tissue m vitro as guided by the polymer (81. Moreover, we have demonstrated both ectoplc and orthotoptc bone formation m vlvo by transplantation of polymer-cell constructs m rat models (9,lU). In this chapter, we illustrate the fabrication of 3-D poly(a-hydroxyester) foam scaffolds by a solvent-casting and particulate-leaching method m comblnation with heat-compression molding We also discuss techniques for seeding and culturing osteoblasts in 3-D polymer scaffolds 2. Materials 2.1. Polymer Processing 2 1.1. Chemrcals 1 Polymers. poly(L-lactic acid) and poly(oL-lactic-co-glycohc acid) The polymers come sealed under vacuum Once the orlgmal packagmg has been opened, the polymers must be stored either under N2 m a freezer at -4”C, or under vacuum m a desiccator This precaution reduces premature degradation by hydrolysis caused by the moisture m the air The mherent vlscoslty of the polymer 1s provided by the company, but exact mol wt and polydlsperslty can be determmed using gel permeation chromatography The physlcal properties of the polymers, mcludmg the glass transItIon temperature (T,) and meltmg temperature for the semlcrystallme polymers (T,,), are listed m ref II and 22. The PLGA copolymers are usually available at lactic*glycohc acid ratios of 85.15, 75 25, 65.35, and 50.50 The polymers can be purchased from* Bummgham Polymers (Bnmmgham, AL), Boehringer Ingelhelm KG (Ingelhelm, Germany); Medlsorb@ (Cmcmnatl, OH), Purasorb@ (Lmcolnshlre, IL) 2. Solvents dlchloromethane (Sigma, St. Louis, MO), chloroform (Sigma) (see Note 1) 3 Sodmm chloride (Sigma) Store m a dry, cool place, especially after size separation

2.1.2. Laboratory Supplies 1 50-mm glass Petri dishes (with covers) (Fisher Scientific, Pittsburgh, PA) 2 Teflon-foil protective overlay (Cole-Parmer, Nlles, IL)

Porous Poly(a-Hydroxyester)

135

Scaffolds

3 USA standard testing sieves, ASTM E-l 1 Specification (Fisher Sclentlfic) 4 Heat-Jacketed Teflon molds (custom-deslgned) 5 Press (Carver Laboratory Eqmpment, Wabash, IN)

2.2. Cell Culture The culture solutions should be sterilized and kept refrigerated when not m use. To prevent temperature shock, heat up to 37°C m water bath before contact with cells. 1. Primary media a Dulbecco’s modified eagles medium (DMEM: Gibco-BRL, Galthersburg, MD) b Fetal bovme serum (FBS) (HyClone, Logan, UT). c 25 mg/mL Gentamlcm (GS) (Sigma) 2 Complete media a DMEM (Glbco BRL) b l&l 5% FBS (HyClone) or Calf Serum (Glbco-BRL) c 8 mg/mL GS (Sigma) d 10 mM Na P-glycerol phosphate (Sigma) e 50 mg/mL L-ascorbic acid (Sigma) f 1O-* A4 dexamethasone (Sigma) (see Note 2). 3 Phosphate buffered salme (PBS) (Sigma)* The buffer should be magnesium- and calcmm-free at pH 7.4 4 Trypsm solution The trypsm solution is kept m the freezer and can be thawedfrozen up to 3 x before activity 1s lost a 0 0625% Trypsm (Sigma) b 0 0125% Ethylene glycol-bzs(b-ammoethyl ether) N,N,N’,N’-tetraacetlc acid (EGTA) (Sigma) c 0 625% Polyvmylpyrrolldone (Sigma). d 26 mMN-(2-hydroxyethyl)plperazme-l\r-(2-ethanesulfomc acid) (HEPES) (Sigma) e 1 125% NaCl (Sigma)

3. Methods 3.7 - Polymer Processing The solvent-casting and particulate-leaching method described adapted from ref. 13. The ad&on of the heat-compression molding adapted from ref. 14.

here 1s step 1s

3 1.1. Solvent-Casting 1 Sieve the salt to desired size range Grinding, for instance, with a mlcromlll, may be helpful if small salt size is desired (see Note 3). 2. Remove polymer from storage If frozen, let equilibrate 1 h and wipe off condensation before opening container Meanwhile, cover Petri dish bottoms with Teflon foil This IS necessary for removal of the polymer-salt composite later

136

Jen, Peter,

and Mikos

3 In a glass vlal, dissolve polymer m 10x (g/mL) dlchloromethane for 30 mm, or until no polymer particles are vlslble 4. Fill Petri dish with desired wt% of NaCl (see Note 4) 5 In the fume hood, cast the polymer solution into the Petri dish with the salt Quickly cover the Petri dish to slow down evaporation of the solvent Fast evaporatlon may cause uneven distrlbutlon of the salt and polymer mixture, and result m nonumform foams Swirl around until salt appears wetted and 1s evenly dlstnbuted. Leave on a leveler platform overnight. 6 The following day, the polymer-salt composite should be further dried m the hood with the lids removed, then vacuum-dried overnight 3.1.2.

Heat-Compression

Moidrng

A Teflon mold of desired shape should be previously constructed, with a heating Jacket attached to a temperature control unit. A press IS also needed to apply the compression (see Note 5). 1 Remove the dry polymersalt composites from the Petn dishes using a strong spatula, and place mto the mold m small chunks Fill the mold as much as possible, but leave room for the end plugs and enough mdentlon for the pressmg bar to be properly aligned. 2 If possible, design a heatmg program so as not to overshoot the desired internal temperature of approx 20°C above the T, or T, of the polymer. 3 After heating cycle (1 h), allow coolmg Note that pressure will drop a httle with the temperature decrease If crystalline properties are important, one can adJust the cooling cycle to quench quickly or anneal slowly 4 Remove from mold 5 A diamond saw (Isomet 1 l-l 180, Buehler, Evanston, IL) can be used to cut the molded samples mto mdlvldual sizes We generally use foams 2-mm thick to mmlmlze diffusion hmltatlons of nutrients to the cells When cuttmg, avoid the use of cutting fluid, which can contaminate the samples Water IS the best lubncant, although one has to be careful not to leach the samples prematurely 3.1 3. Part/&ate

Leachmg

1 Both large casted foams and smaller precut foams are leached m deionized-dH,O for 24 h to remove the salt porogen. Use of excess water 1s recommended to prevent saturation between water changes (6 h intervals) 2 The leached foams should first be dried m air and then m vacuum. They can be stored m a desiccator until use 3.2. Cell Culture 3.2. I Prewetting

and Stenilzation

of Foams

1. Soak foams at least 30 mm m 100% EtOH Keep the foams m mdlvldual contamers (welled plates) during this step, because the ethanol softens the polymer, and aggregates may otherwise form (see Note 6)

Porous Poly(a-Hydroxyester)

Scaffolds

137

2. Mamtammg sterile condltlons, transfer foams into fresh wells Wash for 1 h m sterile PBS, changing the solution every 15-20 mm This rinses out the EtOH, which 1s harmful to cells 3. Immerse foams in media for 1 5 h Use of complete media is preferred, because this solution IS to be added after seeding 4. Change mto fresh media and begin preparmg cell suspension

3.2.2. Cell Seeding and Culture 1. After rmsmg primary culture with PBS, add enough trypsm solution to cover the bottom of the culture flask (about 1 5 mL for a T-75 flask) and incubate for 8-l 0 mm 2. Make sure cells are detached from the bottom of the flask If not, tap lightly on countertop to dislodge stubborn cells Add 5 mL of complete media, plpetmg up and down to create an uniform cell suspension 3. Remove a sample amount to determine cell number, either by a hemacytometer or Coulter counter (Hlaleah, FL) 4 Centrifuge for 10 mm at 400g and 4°C (see Note 7) 5. Aspirate the supernatant and resuspend cells m desired dilution, according to seeding density and cell counts (see Note 8). 6 Seed cells (see Note 9) 7. SIX hours after seeding, add more complete media to sufficiently cover sample (seeNote 10). 8 Media should be changed every l-3 d, depending on the seeding density and media volume Regular changeswill reduce the effects of polymer degradation products, reflected m a decreasem pH, aswell asensuresufficient nutrient supply.

4. Notes 1. PGA IS msoluble m both dlchloromethane and chloroform, It IS only soluble m hexafluorolsopropanol, which IShighly toxic (12) 2. This potent glucocortlcold is often added to induce osteoblastlc phenotype of marrow stromal cells 3. Salt size IS directly correlated to pore diameter (13,14). Pore diameters ranging from 150 to 700 ptn were found to have a neghglble effect on marrow stromal osteoblastproliferation and function m vitro (8) The size of the pores, however, 1simportant for vascular mgrowth mto implants/transplants(15,16) Larger pores promote faster vascular penetration into the foam interior 4. Porosity 1sdirectly related to the mltlal salt-weight fraction (13) The actual porosity can be determined with mercury poroslmetry, or calculated from total volume and properties of the polymer. Porosity of 90% 1srecommendedto create an interconnected pore network and facilitate nutrient transport 5 We have used cylmdrlcal molds to create templates of either 6 or 12 mm dlameter for cell culture Figure 1 presentsa diagram of the heat-compresslon-moldmg system used m our laboratory 6 Passingthe polymer foamsthrough alcohol servesasthe sterihzatlon step before cell seeding. This prewettmg step also enhances permeation of water (and

138

Jen, Peter,

and Mikos

Frg 1 Diagram of the custom-made Teflon mold with heating Jacket (1) compressron bar, (2) Teflon mold, (3) Teflon end plug, (4) heating Jacket, (5) center of mold, and (6) thermocouple insert. The polymer is first placed mto the center of the Teflon mold Then, with the end plugs m place, the compressron bar IS Inserted mto the top of the mold and the unit IS placed m a press An Inserted thermocouple IS connected to a temperature controller for the heatmg Jacket.

7

8

9

10

medium) into the core of the foam (17). Use of a shaker or rotor mixer 1s recommended to enhance this effect Ethylene oxide sterrhzatron IS another method we have used successfully Generally, to avoid leaving cells in suspension longer than necessary, adJust centrtfuge to 4°C m advance Also, do the cell counting whrle the rest of the suspensron IS being centrtfuged We generally seed cells on 3-D polymer foam scaffolds at approx 53,000 cells/cm2, based on the top surface area Thus corresponds to confluency of osteoblasts on an equrvalent two-drmensronal substrate. We seed by mrcroprpetmg the desired amount of cell suspension onto of the scaffold The seedmg volume should be adjusted according to the srze of the foams Avoid using too much volume, because excess cell suspensron will flow out of the foams and populate the well-plate instead of the scaffold For 6-mm disks, we use about 100-200 & We also suggest changing wells before analyses are performed, so results reflect populatron on the foam and not of the surroundmg well Addition of media less than 4 h after seedmg IS not recommended, because the cells require time to attach onto the scaffold so they will not be washed off during

Porous

Po/y(a-Hydroxyester)

Scaffolds

139

handlmg Wamng too long 1salso dtscouraged, because the culture may not have enough nutrtents for the seeded cells to properly attach and functton

5. Final Comments The methods presented m this chapter demonstrate one method of fabrtcatmg btodegradable polymer foam scaffolds for 3-D cell culture. Other methods of polymer processmg are available for the manufacture of scaffolds (I&29). The development of new polymer-cell constructs m tissue engineering 1smtegrally related to advances m cell culture engmeermg through the development of new bloreactors for enhanced nutrient supply to overcome diffusion hmitattons and form clnncally useful bone. References 1 Langer, R and Vacantt, J P. (1993) Tissue engineering Sczence 260, 92@926 2. Hollmger, J. 0 and Batttstone, G C (1986) Biodegradable bone repair materials synthetic polymers and ceramics Clan Orthop Rel Res 207,290-305 3 Suggs, L J and Mtkos, A G (1996) Synthettc biodegradable polymers for medtcal apphcattons, rn Physical Properties of Polymers Handbook (Mark, J E , ed ), AIP, Woodbury, NY, pp. 615624. 4 Gazdag, A R , Lane, J M , Glaser, D., and Forster, R. A (1995) Alternatives to autogenous bone graft efficacy and mdtcattons J Am Acad Orthop Surg 3, l-8 5 Crane, G M , Ishaug, S L., and Mtkos, A G (1995) Bone tissue engineering Nature Med 1, 1322-1324 6 Ishaug, S L , Payne, R G , Yaszemski, M. J., Aufdemorte, T B , Btztos, R , and Mtkos, A G (1996) Osteoblast migration on poly(a-hydroxy esters) Bzotechnol Bloeng 50,44345 1 7 Ishaug, S. L., Yaszemskt, M J , Btztos, R., and Mikes, A G (1994) Osteoblast functton on synthetic biodegradable polymers J Blamed Mater Res 28, 144%1453 8. Ishaug, S L , Crane, G M., Miller, M. J., Yasko, A W , Yaszemskt, M J , and Mtkos, A G (1997) Bone Formation by three-dtmenstonal stromal osteoblast culture m btodegradable polymer scaffolds J Blamed Mater Res ,36, 17-28 9 Ishaug-Riley, S L , Crane, G M., Gurlek, A., Mtller, M J , Yasko, A W , Yaszemskt, M. J , and Mikos, A G. (1997) Ectoptc bone formation by marrow stromal osteoblast transplantation using poly(DL-lactic-co-clycohc acid) foams implanted mto the rat mesentery J Blamed Mater Res ,36, l-8 10 Smith, M L , Miller, M J., Crane, E , Khoo, A K M , Gulek, A , and Mtkos, A G (1997) Cranial defect repair with osteoblast transplantation Abstracts of Plastic Surgery Research Council, Galveston, TX, p. 45. 11 Lu, L and Mtkos, A. G. Poly(lacttc acid), in Polymer Data Handbook (Mark, J E , ed ), Oxford Umversity Press, New York, m press 12 Lu, L and Mikes, A G Poly(glycohc acid), m Polymer Data Handbook (Mark, J E , ed ), Oxford University Press, New York, in press 13 Mikes, A G , Thorsen, A J , Czerwonka, L. A , Bao, Y , and Langer, R (1994) Preparation and charactenzatton of poly(L-lacttc acid) foams Polymer 35, 1068-1077.

140

Jen, Peter, and Mlkos

14 Thomson, R C , Yaszemskr, M J , Powers, J M , and Mrkos, A G (1995) Fabrrcation of biodegradable polymer scaffolds to engmeer trabecular bone J Blomater Scl -Polym Ed 7,23-38. 15 Mikos, A. G., Sarakinos, G , Lyman, M D , Ingber, D E , Vacanti, J P , and Langer, R (1993) Prevascularrzatron ofporous biodegradable polymers. Bzotechnol Bzoeng 42,7 16-723 16 Wake, M C , Patrick, C W , and Mrkos, A G (1994) Pore morphology effects on the fibrovascular tissue growth m porous polymer substrates Cell Transplant 3, l-5 17 Mrkos, A G., Lyman, M D , Freed, L. E , and Langer, R (1994) Wetting of poly(L-lactic acid) and poly(m-lactic-co-glycohc acid) foams for tissue culture Blomaterzals 15, 55-58 18 Thomson, R C , Yaszemskr, M J., and Mrkos, A G (1997) Polymer scaffold processmg, m Prmclples of Tzssue Engmeerlng (Lanza, R P , Langer, R , and Chick, W L , eds ), R G Landes, Austin, TX, pp 263-272 19 Lu, L. and Mrkos, A G (1996) The importance of new processmg techniques m tissue engineering Mater Res Sot Bull 21,28-32

12 Quantitative

Assessment

of Autocrine

Cell Loops

Gregory Oehrtman, Laura Walker, Birgit Will, Lee Opresko, H. Steven Wiley, and Douglas A. Lauffenburger 1. Introduction Regeneration of functtonmg trssue essentially mvolves recapitulating relevant aspects of organogenesrs, so that the starting composite of cells, matrix, and molecular factors develops mto the desu-ed structure and physrology. A crucial aspect of development IS local cell-cell commumcatton; that IS,molecular regulatory factors are more typically paracrme and autocrme than endocrine m nature. Autocrme loops were origmally thought of predommantly as being involved m pathological behavior, but it 1sbecoming increasingly clear that a large portron of normal physiologrcal behavior-and a tremendous portron of development-1s strongly regulated by autocrme factors (I). Thus, contmumg progress of the field of tissue engineering will require increased understanding of how autocrme loops operate, so that they can be designed or manipulated systematically We have made an effort m this dn-ectron, and some early experimental and modelmg results can be found m the literature (2-5). In this chapter, we describe the methods we have used for creating autocrme cell loops and quantrtatrvely assessingthen operation. A crucial assessment issue is how much of the hgand synthesized by autocrme cells IS captured by the secreting cells themselves, and how much IS permitted to escape mto the bulk extracellular medmm. A perhaps countermturttve prmcrple of autocrme loops 1sthat they might be provrdmg the most useful physrologrcal functron when one cannot expertmentally fmd any autocrme lrgand m the bulk medmm, because that IS when the system should be most senstttvely regulated by what IS captured; when one can experrmentally rind substantial levels of autocrme ligand m bulk medium, that may be when pathologrcal behavior arises. Another important prmcrple of autocrme From

Methods m Molecular Medrone, Vol 78 Ttssue Engmeermg Methods and Protocols E&ted by J A Morgan and M L Yarmush 0 Humana Press Inc , Totowa, NJ

143

144

Oehrtman et al

loops 1sthat proper regulation can only occur when the ligand IS synthesized m its transmembrane precursor form for proteolytlc cleavage at the cell surface. When a purely secreted form of the hgand 1ssynthesized, receptor binding can take place along the secretory pathway even before the hgand reaches the cell surface; this short-cu-cultmg probably compromises aspects of receptor traffickmg dynamics and mteractlons of hgand with extracellular matrix that affect cell regulation by autocrme loops. Our assessment methods are ELISA, for determination of the amount of autocrme hgand escaping mto bulk extracellular medium, and the Molecular Devices Cytosensor, for determmatlon of the amount of autocrine ligand captured by secretingcells.The Cytosensormeasuresextracellular acldlficatlon rate (ECAR) m real time, which can be correlated to receptor-mediated slgnalrng processes(6,7) Our model experimental system1sof B82 mouse L-cells, transfectedwith the genes for human epldermal growth factor receptor (EGFR) and transforming growth factor alpha (TGF-a) using the two-plasmld tetracycline-controlled gene-expresslon system This system allowed us to calibrate readouts from the Cytosensor with hgand bmdmg, becausewe could quantify Cytosensorreadouts from nonautocrme cells challenged with exogenous ligand, and to vary hgand synthesisrate by changmg the concentration of tetracyclme m the medium 2. Materials 2.1. Creation of Autocrine Cell Loops 2 1.1. Reagents for Construction of pUHD10.3-TGF-a

Plasm/d

pUHDl5 lneo and pUHD10 3 (first and second plasmlds of tetracycline controlled, inducible expression system) obtained from M Gossen and H BuJard (8) (Heidelberg University, Germany) pMTE4 (transmembrane TGF-a cleavable protem) obtamed from R Derynck (9) (Umverslty of Cahforma at San Francisco) Bluescript II KS+ (Strategene, La Jolla, CA) Restriction enzymes Hz&III, NcoI, RsaI, XhoI, EcoRI, BumHI, PstI (Boehrmger Mannhelm, Indlanapohs, IN) Modifying enzymes Klenow, T4 hgase (Boehrmger Mannhelm) Low melting point agarose (Sigma, St Louis, MO) and @agaraseI (New England Blolabs, Beverly, MA) Wizard Maxlprep (Promega, Madison, WI)

2. I 2. Reagents for Creation of Autocrme Clones 1 Hlstldmol reslstence plasmld (PREPS) (Invltrogen, San Diego, CA) PREP 1s modified mto pR8 with the removal of the EBNA- 1 and OrlP segments (sac1 to XbaI deletion and reltgatlon) to prevent eplsomal repllcatlon, and to allow for plasmld rncorporatlon into chromosomalDNA

Autocrine Cell Loops

145

2 CaPO, transfectton reagents: a Salt solutton dissolve 3 7 g KC1 (Sigma), 10 g D(+) glucose (Sigma), and 1 g Na,HPO, (Sigma) into 50 mL final volume ddH,O, filter sterilize with 0 2-pm filter, and store at -20°C b 2X HBS (HEPES buffered saline). Add 1 mL salt solution, 1 g HEPES (Sigma), and 1 g NaCl (Sigma) to 85 mL of ddH,O, pH to 7 1 with 1 NNaOH, and bring volume up to 100 mL wtth ddH,O, Filter sterilize with 0 2-pm filter, store rn 20-mL aliquots at 4°C. c 2 5 MCaCl, dissolve 36.8 g of CaCl, 2H,O (Srgma) into 100 mL final volume H,O Autoclave stertle Store at room temperature or 4°C 3 8 x 8 glass cylinders (Bellco, Vineland, N.J )

2.2. Determination of Autocrine Ligand in Bulk Extracellular Medium 2.2.1. Reagents for Measurement of Autocnne Llgand 1 Monoclonal arm-EGFR antibodies 528 and 225 (hybrtdomas, ATCC, Rockvtlle, MD) Antibodies purified from hybrtdoma supematant, 50% (NH&SO4 (Sigma)ddH,O cut at 4”C, spm at 7600g for 10 mm at 4°C Dialysis pellet overmght m PBS with 30,000 MWCO dialysis tubing (Spectrum, Houston, TX) 2 TGF-a ELISA kit (Calbtochem, San Dtego, CA). 3 1X PBS buffer, 0 2 g KC1 (Sigma), 0.2 g KH2P04 (Sigma), 8 g NaCl (Sigma), 1 15 g Na,HPO, (Stgma) dtssolved m 1 L ddH20, pH 7 0, sterile-filter through 0 2-pm filter and store at 4°C

2.2 2 TGF-a Cellular Processing 1 Protem A-comugated Sepharose beads (Srgma) 2 Amicon Concentrators (YM3000 membrane)/Centrrcons (3000 MWCO) (Amtcon, Beverly, MA) 3 Protem standards. albumm, carbonic anhydrase, cytochrome, aprotmm (Stgma) 4. Glycerol (Sigma) 5 G-50 Sephadex (Pharmacra, Ptscataway, NJ).

2.3. Determination of Autocrine Ligand Captured by Secreting Cells 2.3.1. Molecular Devices Cytosensor Material 1. 2 3. 4

Transwells. 12 mm dtameter, 3 pm pore stze (Coming, Cambridge, MA) Transwell mserts 8 mm diameter, 3 pm pore size (Molecular Devtces, Menlo Park, CA) Spacer 50 un-rheight/6 mm td (Molecular Devtces). DV-cyto buffer On the day of a Cytosensor expertment, make 1 L of Dulbecco’s volt modified Eagle’s media (DMEM, 4500 mg/L glucose, L-glutamme, Sigma) without sodium bicarbonate. Instead, add 2.59 g of NaCl (Stgma), pH to 7.4, and add 100 mg of BSA (Sigma) Filter-stertlize with 0.2~un-~ filter. Somettmes, DV-cyto can be stored at 4°C but re-pH to 7.4 and filter

146

Oehrtman et al

2.3 2. Cal/brat/on of Cytosensor Readout Reagents 1 hEGF (Gtbco BRL) EGF at 1 mg/mL m 50 mM sodium phosphate, stored at -20°C. 2 1X WHIPS buffer 1 g polyvmyl pyrrohdone mol wt 40,000 (Stgma), 7 6 g NaCl (Sigma), 0 373 g KC1 (Sigma), 0 102 g MgCl, 6H20 (Sigma), 0 147 g CaCl, 2H20 (Stgma), 4 76 g HEPES-actd (Sigma), dissolve m 1 L of ddH,O at a pH of 7 4, store at 4°C 3 I’25 EGF Add together 1 rodobead (Prerce, Rockport,

IL), 100 mA4 80 p.L Tns

(Sigma), 30 uL 100 mCr/mL I’25 (NEN, Boston, MA), 10 pL of 10 ng/pL EGF (Glbco-BRL) Wart 15 mm and remove protein to new tube Made fresh day of use, add 40 pL of 12 mg/mL sodmm metabtsultite (Sigma) m PBS to rodobead tube, remove, and add to EGF tube. Add 40 pL of BSA chase solution (20 mL ddHzO, 100 mg BSA-RIA grade (Sigma), 1 crystal KI (Sigma), sterile-filtered with 0 2-p filter, and store 1-mL aliquots at -20°C) Separate I’25-EGF from I’25 over a G- 10 Sepharose column (Pharmacra, Ptscataway, NJ) wtth PBS Store I’25-EGF at 4°C 4 D/H/B buffer make 1 L of Dulbecco-V-modified Eagle media (DMEM, 4500 mg/L glucose, 1 mM L-glutamme, Sigma) wtthout sodmm brcarbonate Add 5 95 g/L HEPES (Sigma) and 1 mg/mL BSA (Stgma), pH to 7 2

2 3 3 Cytosensor Readout of Antibody InhIbition Reagents 1 Polyclonal anti-TGF-a anttbody (R & D Systems, Mmneapolts, MN) 2 Monoclonal anti-EGFR antrbodres 528 and 225 (hybndomas - ATCC) (see Note 2) 3 Antrbodres used m Cytosensor experiments a Because of the sensmvlty of Cytosensor measurements, it has been noted that notse and baselme changes occur when addmg a hrgh concentratton of anttbodres m glycerol or a buffer with high salt concentratrons, such as PBS b To avotd thts problem, anttbodtes or any other addmves used at a hrgh concentratton need to be dtalyzed m DV-s (DV-cyto buffer without BSA added) overnight at 4°C Ahquot protem mto small volumes to prevent cychc freezethaw problems Suggestron for antrbodtes and small volumes 1s Slide-a-Lyzer

CO.5 mL capacity, 10K MWCO (Pierce, Rockford, IL) 2.4. B82 Fibroblast

Tissue

Culture

Medium

B82 cells are a mouse L-cell ltne that lacks the EGFR chromosomal segment and does not express detectable levels of eptderrnal growth factor (EGF) and TGF-a. B82 receptor minus and receptor positive cells were obtained from G. Grll (Untversity of Calrfornia at San Diego) (10) All B82 cells are grown at 37”C, 5% CO2 and 98% relative humrdrty. 1 R- media (Parental cells (B82R-) tissue-culture medium) a DMEM 4500 mg/L-glucose, 1 m/t4 L-glutamme, Sigma), add 3.7 g/L sodmm bicarbonate (Sigma), filter-stenhze with a 0.2~pm filter (Gelman Sciences, Ann Arbor, MI) under 12 pstg positive pressure (N2 gas), and store at 4°C untd use

Autocrine Cell Loops

147

b L-glutamme 1 mM (Stgma) 100 mM glutamine stock solution; dissolve 3 65 g mto 250 mL 1X PBS buffer, filter sterilize through 0 2-pm filter, and store at -20°C m IO-mL ahquots c Pemcillm-streptomycin 100 IUImL-100 pg/mL (Sigma) d Bovine calf serum 10% (Sigma) 2 R+ media (B82 with pXER (EGFR, B82R+) tissue culture media). a DMEM (Sigma) b L-glutamme 1 mM(Stgma). c. Penicillin-streptomycin 100 IU/mL-100 pg/mL (Sigma) d 10,000 MWCO dialyzed bovme calf serum 10% (Sigma) e 1 @4 methotrexate (Sigma) 1 mM 1000X stock solution, dissolve 25 mg methotrexate mto 55 mL of 0 1 N HCl, sterile-filter with 0 2-pm filter, and store at -20°C in lo-mL aliquots 3 R+/lst plasmtd media (B82R’cells with pUHDlS.lneo (B82R+/lst plasmid) tissue-culture media) a R+ media plus 600 pg active G4 18/mL (Sigma) 60 mg/mL- 100X genettcm sulfate (G418) stock solutton, 5 g (70% active ingredient) dissolved m DMEM, final volume 58 3 mL, sterile-filtered using a 0 2-pm filter, store at -20°C in IO-mL aliquots 4. Autocrme media (B82 EGFR cells with pUHD15 lneo, pUHD10 3ITGF-a, pR8 (B82R’/TGF-a) tissue-culture media) a R+/lst plasmid media plus 2 4 mM htsttdmol (Sigma) 2 4 A4 htstldmol stock solutton, 5 g dissolved m 9 7 mL ddH,O final volume, sterile-filtered using a 0 2+-n filter, store at 4°C b 1 pg/mL tetracycline (repression of TGF-a expression, Sigma) 20 mg/mL tetracycline superstock, dissolve tetracycline powder mto ddH20, and store at -20°C Dilute tetracycline 1 40 in ddH,O, sterile-filter using 0 2-pm filter, and store at -20°C m 1-mL ahquots

3. Methods 3.7. Creation of Autocrine Cell Loops 3.1.1 Construct/on of pUHDlO.3/TGf-a First, Bluescrlpt II KS+ (pBS)ITGF-a molecular clonmg because of Bluescrtpt’s

plasmtd multiple

1s constructed for ease of clonmg region.

1 The entire 800 bp TGF-a sequence is removed from pMTE4 with a HzndIII digest (37°C 2 h) and gel purified (1% low melting point agarose, 2.5 hours, 50 mAmps) Isolate the TGF-a gene fragment from the agarose using P-agarase1 and ligate into H&III digested Bluescrtpt II KS plasmtd (T4 hgase, overnight, 16°C). Minipreps can be checked via NcoI (only cuts if insert is present) digests and large scale preps of pBS/TGF-a purified with Wizard Maxiprep Orientation of TGF-a insert (need T7 to T3 orlentatlon for part 2) m Bluescript can be determined by RsaI digests (if T7 to T3, bands are 577, 1754, 1430 bps; if T3 to T7, bands are 348, 1754, 1659 bps) and sequencing.

148

Oehrtman et al.

2 The construction of pUHD10 3ITGF-a Remove TGF-cr sequence from pBS/ TGF-a by digesting the plasmid with XhoI and EcoRI (37°C 2 h each) Before digesting the plasmid with EcoRI, however, the lmear plasmid (cut with XhoI) must be blunted wtth Klenow enzyme. Prepare pUHD10 3 by digestmg it wtth BamHI (37°C 2 h), blunting with Klenow enzyme (same method as with pBS/ TGF-a), and dtgesting the linear plasmtd with EcoRI (37°C 2 h). Gel purify the DNA fragments with 1% low-melting-point agarose, 2 5 h, 50 mAmps, and isolate the fragments from the agarose using P-agarase1 Ligate the two DNA fragments together (T4 hgase, overnight, 16”C), check resulting plasmid preps for the correct pUHD 10 3/TGF-cz via PstI lmearization (&I only m TGF-cr insert, no sites m negative pUHD10.3 plasmids) and sequencing

3 1 2 Creation and Selection of Autocrine Cells The transfectron of the B82 cells were accomplished precipitation method (II).

using the CaPO,/DNA

1 Day 1: Seed one mllhon B82 R+/lst plasmid cells mto 60-mm tissue-culture dishes m R+/l st plasmid media 2 Day 2. Prepare a 500-mL ahquot of pUHD10 3ITGF-a and pR8-Cas(PO,), precipitate by mixing 30 pg of pUHD 10 3ITGF-a and 10 pg of pR8, up to 450 pL ddH,O (depending on plasmid concentrattons), 500 pL of 2X HBS, and 50 pL of 2 5 M CaCl* Add the precipitate to the cells’ media plated on d 1 On d 3, rinse the cells wrth PBS and add back fresh R+/lst plasmid media 3 Day 4. Dilute cells 1 4 mto autocrme media, keeping all of the cells Note Sometimes, if a large number of clones continue to grow m the selective media, and tsolatmg mdtvldual clones is difficult, dilute the cells serially 1 2, 1 10, 1 100, 1.1000, 1. 10,000 Refresh the selective media every third d until mdividual cell colonies appear (approx 200-1000 cells per colony), typically appearing l-2 wk after begmnmg selection. Isolate the colonies using autoclaved Bellco 8 x 8 mm glass clonmg cylinders dabbed into autoclaved Vaseline, and pass the colony mto mdividual well of a 24-well plate contaming selective media

3.2. Determination of Autocrine Ligand in Bulk Extracellular Medium 3.2 1. ELISA Measurement of Autocnne Ligand 1 Plate autocrme cells m four sets (one for induced TGF-a expression and one for unmduced expression, plus and minus antireceptor blocking antibodies) at a cell concentratton allowing for 2-3 d of growth before cell confluence is reached m normal autocrme growth media. For example, B82 cells double once a day, thus plate 125,000 cells mto a 35-mm tissue culture dish and confluence is 3 d later at 1 5 million cells 2. Two days before cell confluence, remove tetracycline containing autocrme media, rinse the cells wtth PBS, and, to half of the cell wells, add tetracyclme-free

Autocrine Cell loops

149

autocrme media, mducmg TGF-a expresston. To the other cell wells, add tetracycline contammg autocrme media, keepmg TGF-a expression off. NOTE Protem synthesis takes about 6 h before detection at the surface and m the media, thus TGF-a expression is Induced 24 h before beginning the experiment to ensure a steady-state expression of TGF-a 3 Next day, with TGF-a expression at steady state, rinse the cells with PBS and add fresh autocrme media [tetracycline-free for induced cells, and tetracyclinecontaining media for unmduced cells) To half of the induced and unmduced autocrme cells’ media, add 10 pg/mL anti-EGFR-blocking antibody 225 or 528 A compartson of TGF-a concentrations, plus and minus antibody, allows the analysis of TGF-a uptake by its receptor. 4 On the last day, remove 1 mL of conditioned media, spin at 13,OOOgfor 10 mm at 4°C to remove cell debris Store the supernatant at -20°C if measuring TGF-a concentration another day, otherwise on ice Measure TGF-a concentration via TGF-a ELISA Determine cell density by trypsimzmg the cells, and count cell number using a hemocytometer or Coulter counter (see Note 1)

3.2.2. Tetracycline Concentration Effect on TGF-a Secretion 1 Plate autocrme cells at a similar cell density at a cell concentration that allows 2-3 d before cell confluence is reached in normal autocrme growth media. Plate enough cell dishes for 6 or more different tetracycline concentrations, with replicates. 2 Two days before cell confluence is reached, rmse the cells with PBS and replace the media with fresh autocrme media contammg a gradient of tetracyclme concentrations For example, TGF-a expression IS mostly inhibited by 1 pg/mL tetracycline, thus a 10, I,0 1 ug/mL to 10, I, 0 I,0 ng/mL tetracycline concentrations is an appropriate range 3 Next day, with TGF-a expression at steady state, rinse cells with PBS and replace the media with a fresh autocrme media containing the tetracycline gradient. Also, mclude m the media, 10 pg/mL antireceptor blocking antibody 225 or 528, to prevent receptor TGF-a uptake and accurate measurement of TGF-a expression 4 On the last day, remove 1 mL of conditioned media, spin at 13,00g for 10 mm at 4°C to remove cell debris. Store the supernatant at -20°C if measuring TGF-a concentration another day, otherwise on ice Measure TGF-a concentration via TGF-a-ELISA. Determine cell density by trypsmizmg the cells and count cell number using a hemocytometer or Coulter Counter.

3.2.3 Determining

Cellular Processing of Secreted TGF-a Protein

1 Plate a large number of either transfected B82R-ITGF-a cells or with autocrme B82R+ITGF-a cells mto 100~mm dishes, T75 flasks, or roller bottles m tetracyclme-free growth media Once cells have grown confluent and expressing TGF-a, replace the media with serum/protem-free media Allow the media to condition for l-2 d

Oehrtman et al.

150

a If using autocrme cells, one must have a high TGF-a expressor, overcommg hgand uptake by EGFR, or use anttreceptor-blockmg anttbodtes 225 or 528 If using antrbodtes, they should be removed before adding the condtttoned media to the sizing column, because they mtght effect the mol wt standards elutton or mask protein detection b. The antibody can be removed usmg rabbtt anttmouse antlbodtes with protein A-conjugated Sepharose beads Incubate the condmoned media wtth rabbit anttmouse anttbody, protem A-beads for a mmtmum of 2 h, at 4°C on a rocker Spur beads down at 130g for 2 mm and continue wtth the supernatant A high concentratton of TGF-cr (~100 ng/mL) ~111 be requtred to run on the sizing column, thus, rt may be necessary to concentrate the condtttoned medta usmg Amtcon concentrators (large volumes, >lOO mL) or centrtcons (small volumes, < 10 mL) with 3000 mol wt cutoffs (MWCO) done at 4°C MIX 100 ngs of TGF-a wtth protein standards (albumm, 66 kDa, 2.5 mgs), carbonic anhydrase (29 kDa, 1 mg), cytochrome c (12 4 kDa, 1 mg), aprotrnm (6 5 kDa, 3 mg), and 50 pL of glycerol, to a total volume of 1 mL Overlay the TGF-a/protein standards mrxture on a 1-m G-50 fine Sephadex column (equaltbrated with PBS) with a small buffer head (99% HDMEC The vast maJorrty of magnetrc beads IS shed from the cell surface within 2 h after subculture

Some magnetic

beads are taken up by HDMEC,

however,

the mcorpo-

ration of magnetrc beads drd not influence the growth behavior of HDMEC. After immunomagnetrc purrficatron of first-passage cultures, secondpassage cultures regularly consist of 100% HDMEC, as evaluated by their characterrstrc morphology (Fig. 1) and then stammg with an anti-CD3 1 antibody. Staining for von Wtllebrand factor revealed the intracytoplasmrc granular stam-

Microvascular

Endothelial

Cells

265

Fig. 1. HDMEC, isolated by Dynabeads coupled with an anti-E-selectin monoclonal antibody (MAb), passage 5. Monolayer culture with characteristic cobblestone morphology.

Fig. 2. Passage 6 HDMECs that have been maintained for 6 mo without subculture: Monolayer culture with few tube-like structures and no evidence of contaminating fibroblasts or other cell populations.

ing pattern characteristic for endothelial cells. HDMEC also showed a rapid uptake of acetylated LDL. Even after 6 mo without passage, HDMEC cultures are characterized by the persistence of typical monolayers with some degree of capillary-like tube formation (Fig. 2). When confluent HDMEC cultures are

266

Richard, Velasco, and Detmar

Fig. 3. Spontaneous formation of tube-like structures 16 h after overlay of a collagen type I gel onto confluent HDMEC, passage 5.

overlaid with a collagen type I gel (500 ug/mL in endothelial growth medium), they form capillary-like tubes within 4-6 h (Fig. 3). The expression of the VEGF/VPF receptors, KDR (20) and Flt- 1 (21), has been widely considered as a specific feature of endothelial cells. HDMEC isolated by the described immunomagnetic protocol express both high-affinity receptors for VEGFWPF, KDR and Flt- 1, as demonstrated by Scatchard analysis of ‘251-VEGF binding and by immunohistochemistry, using specific monoclonal antibodies against KDR and Flt- 1. Accordingly, treatment with VEGF/VPF potently induced the proliferation of HDMEC in concentrations as low as 0.3 ng/mL. The mitogenic effect of VEGF was observed in HDMEC from passages 2-l 2. HDMEC isolated by the immunomagnetic protocol, using anti-E-selectin coupled Dynabeads, are highly sensitive to the induction of cellular adhesion molecules by cytokines such as IL- 1-p and TNF-a. Treatment of HDMEC with TNF-a, in passages 2-8, potently induced the expression of intercellular adhesion molecule-l (ICAM- 1; CD54), vascular cell adhesion molecule- 1 (VCAM- 1; CD1 06), and endothelialleukocyte adhesion molecule-l (E-selectin, ELAM-1; CD62E).

4. Notes 1. Tissue source and collection: The sources for dermal microvascular endothelial cell cultures are human neonatal foreskins obtained from routine circumcisions. Foreskins are easily obtained, and have a markedly higher density of dermal microvessels than adult human skin. Although it is possible to isolate and culture

HDMEC from adult skin, the endothehal cell yield is considerably lower, and the life span of HDMEC from adult skm is shorter, compared to HDMEC obtained from neonatal foreskins Foreskm-collecting medmm is usually kept m a refngerator near the surgical area, and specimens are collected over a period of 4 h, during which they are stored at 4°C. A storage period of 4 h does not affect the viability of HDMEC durmg later cultivation Sterlllzatlon of foreskins It IS important that the surgeon applies appropriate ant]septic protocols prior to circumcisions In our experience, a 2-h mcubatlon m collectmg medmm, contammg an increased concentration of antlblotlcs, has helped to keep the rate of bacterial or fungal infectlons at a minimum Trypsmlzatlon An overmght mcubatlon at 4°C is usually sufficient to allow for a separation of the epidermis from the dermis If the epldermls still remains attached, the trypsm solution should be replaced with fresh trypsm solution, and the specimens should be Incubated at 37°C for an additional 30 mm It IS our expellence that the success of the protocol depends on the correct techmque to mechanically isolate dermal endothehal cells by scrapmg the upper side of the dermis with a blunt scalpel blade, once the epldermal sheet has been removed Alternatively, after trypsm digestion, collagenase (5 mg/mL, Sigma C-7657) can be added to the specimens for l&20 mm to faclhtate tissue dlsmtegratlon However, when the collagenase dlgestlon 1s allowed to continue for too long, the derma1 tissue dlsmtegrates, rendermg the efficient mechanical release of mlcrovascular endothehal cells more difficult. Removal of epidermis from dermis After trypsm digestion of foreskms, the epldermis IS mechamcally removed, usmg two pairs of forceps. One forceps holds the epldermal sheet, and the other holds the lower dermis The epldermal sheet 1s then carefully removed and dlscarded Endothehal cell-growth medium Several authors have reported that the addition of up to 30% human serum to the growth medium was necessary to sustain HDMEC growth In our experience, pure HDMEC cultures can be mamtamed and efficiently propagated with only the addition of FCS This avolds the high costs and potential contammatlon risks associated with human serum In add]tion, the use of prepartum maternal serum has been suggested to enhance the growth of HDMEC m vitro (22). However, this has not been necessary m our experience, since cultlvatlon m 20% FCS enables the cultlvatlon of pure HDMECs, which express several endothellal cell markers, and are characterized by an extended hfe-span of more than 35 population doublmgs Moreover, preparturn maternal serum may contam a number of additional growth factors, possibly mcludmg VEGF, comphcatmg the study of isolated effects of growth factors and cytokines Epldermal growth factor has been used and recommended by most authors as an additional supplement to the endothellal growth medium for HDMEC cultlvatlon This seems to be rather unnecessary, since HDMECs m vitro do not express EGF receptors, do not react to the mitogemc effects of EGF, do not show any detectable expresslon of EGF receptors on Western blots, and do not show tyrosme phosphorylatlon after EGF stimulation (M Detmar et al., unpublished data)

268

R/chard, Velasco, and Detmar

6 Collagen coatmg of culture dashes We routmely coat culture dashes with 50 pg/mL collagen type I (Vttrogen) However, coating of culture dishes with fibronectm (50 pg/mL) may increase the seeding effictency of prtmary endothehal cell tsolates. Aside from primary endothelial cell cultures, we have been unable to detect stgmficant differences m the plating effictency of HDMEC on fibronectm vs collagen type I m later passages Instead of Vnrogen, some authors plate HDMEC on gelatin-coated dishes, wtthout any obvtous advantage or disadvantage over Vitrogen

References 1 Jaffe, E A , Nachmann, R. L , Becker, C. G., and Mmtck, C R (1973) Culture of human endothelial cells from umbtlmal veins Identtficatton by morphologtc and mununologtc criteria J Clan Invest 52,2745-2756 2 Maruyama, Y. (1973) The human endotheltal cell m tissue culture Z Zelljorschung 60,69-79 3 Kumar, A U , West, D C , and Ager, A (1987) Heterogeneity m endothehal cells from large vessels and mtcrovessels DEfferentlatlon 36, 57-70 4 Detmar, M , Imcke, E , Ruszczak, Z , and Orfanos, C E (1990) Effects of tumor necrosis factor-alpha (TNF) on cultured mtcrovascular endothehal cells derived from human dermis J Invest Dermatol 95,219S-222s 5 Detmar, M., Tenorto, S , Hettmannsperger, U , Ruszczak, Z , and Orfanos, C E (1992) Cytokme regulation of proliferation and ICAM- 1 expression of human dermal mtcrovascular endothehal cells m vttro J Invest Dermatol 98, 147-153 6 Lee, K H , Lawley, T J , Xu, Y., and Swerhck, R A (1992) VCAM-1, Elam- 1, and ICAM- l-Independent adhesion of melanoma cells to cultured human dermal mtcrovascular endothehal cells J Zmmunol 98, 79-85 7 Swerhck, R A , Lee, K. H , Li, L J , Sepp, N T , Caughman, S W , and Lawley, T J (1992) Regulation of vascular cell adhesion molecule 1 on human dermal microvascular endotheltal cells J Zmmunol 149, 698-705 8 Swerllck, R A , Lee, K. H , Wick, T M , and Lawley, T J (1992) Human dermal microvascular endothelial but not human umbiltcal vem endotheltal cells express CD36 m VIVO and m vitro J Immunol 148,78-83 9 Folkman, J , Haudenschtld, C C , and Zetter, B R. (1979) Long-term culture of capdlary endothehal cells Proc Natl Acad Sci USA 76,5217-5221 10 Davtson, P. M , Bensch, K , and Karasek, M. A. (1980) Isolatton and growth of endothehal cells from the microvessels of the newborn human foreskin m cell culture J Invest Dermatol 75,3 l&32 1 11 Davison, P M , Bensch, K , and Karasek, M A (1983) Isolation and long-term sertal culttvatton of endothehal cells from the microvessels of the adult human dermis In Vztro 19, 937-945 12 Kubota, Y., Klemman, H. K , Martin, G R , and Lawley, T J (1988) Role of lammm and basement membrane m the morphological dtfferenttatlon of human endothehal cells mto capillary-like structures J Cell l?zoZ 107, 1589-l 598

Microvascular

Endothelial Cells

269

13 Kraelmg, B. M., Jimenez, S A.. Sorger, T., and Maul, G. G (1994) Isolation and characterization of microvascular endothehal cells from the adult human dermis and from skm biopsies of patients with systemic sclerosis Lab Invest 71,745-754 14 Bevilacqua, M P , Stengelm, S , Gimbrone, M. A , Jr , and Seed, B (1989) Endothehal leukocyte adhesion molecule 1 An mducible receptor for neutrophils related to complement regulatory proteins and lectms. Sczence 243, 116&l 165 15 Groves, R W., Allen, M H , Barker, J N. W N , Haskard, D 0 , and MacDonald, D M (199 1) Endothehal leucocyte adhesion molecule-l (ELAM- 1) expression m cutaneous mflammation Br J Dermatol 124, 117-123 16 Jackson, C. J., Garbett, P. K , Nissen, B., and Schrieber, L (1990) Bmdmg of human endothelmm to Ulex europaeus-1 coated dynabeads:application to the isolation of microvascular endothehum J CelZScz 96, 257-262 17 Drake, B L. and Loke, Y W. (1991) Isolation of endothehalcells from humanfirst trimester decidua using nnmunomagneticbeads Human Reprod 6, 1156-l 159 18 Hewett, P W. and Murray, J C (1993) Immunomagnetic purification of human microvessel endothehal cells usmg Dynabeads coated with monoclonal antibodies to PECAM-1 Eur J Cell Blol 62,45 l-454 19 Sepp, N T , Gille, J , Li, L J , Caughman, S W , Lawley, T J , and Swerhck, R A (1994) A factor m human plasmapermits persistent expression of E-selectm by human endothehal cells J Invest Dermatol 102,445-50 20 Terman, B I., Dougher, V M , Carrion, M. E., Dimitrov, D , Armellmo, D C , Gospodarowicz, D , and Bohlen, P (1992) Identification of the KDR tyrosme kmaseasa receptor for vascular endothehal cell growth factor Blochem Bzophys Res Commun 187,1579-1586 21 deVries, C , Escobedo, J , Ueno, H , Houck, K , Ferrara, N , and Willlams, L T (1992) The fms-like tyrosme kmase, a receptor for vascular endothelial growth factor Science 255,989-99 1 22 Karasek, M. A. (1989) Microvascular endothelial cell culture J Znvest Dermatol 33S-38s

22 Initiation, Maintenance, and Quantification of Human Hematopoietic Cell Cultures Paul C. Collins, Sanjay D. Patel, William M. Miller, and E. Terry Papoutsakis 1. Introduction The culture of hematopolettc cells for cell and gene theraptes IS a raptdly growmg area wtthm the field of applied hematology and ttssue engmeermg As evidenced by recent clmlcal trials (I), ex vtvo expanded hematopoiettc cells offer great promise for the reconstttutlon of rn VIVO hematopotests m tmmunocompromtsed patients who have undergone chemotherapy Other potential apphcatlons for ex vtvo expansion mclude productton of cycling stem and progenitor cells for gene therapy, expansion of dendrmc cells for unmunotherapy, and productton of red blood cells (RBCs) and platelets for transfustons (2) The heterogeneous cell populatton contained m a hematopotettc culture 1s always changing as a result of the delicate balance between prohferatton of certain cell types, their dtfferenttatton mto other cell types, and the death of vartous cell populations. Through the combmatton of different cytokmes, media, and physlcochemlcal parameters, hematopotetrc cultures can be drrected toward lineages of interest. Complete evaluation of the performance of hematopotettc cultures requires the use of long-term assays such as the 2-wk methylcellulose assay, to detect progenitor or colony-forming cells (CFC), and the very primtttve long-term culture-nnttatmg cell (LTC-IC) assay (7 wk) Flow cytometry can be utthzed to quantify cells bearing antigens, such as CD34 (prtmtttve progenitors), CD 15 and CD 11 b (granulocyte and monocyte postprogemtors), and glycophorm A (maturing erythrocytes) In this chapter, we describe methods for the mittatton and mamtenance of hematopoletlc cultures, the quantification of a variety of cell types, and the setup of mixed suspension cultures m spinner flasks. From

Methods m Molecular Medune, Vol 18 Tmue Engmeenng Methods and Protocols Edlted by J R Morgan and M L Yarmush 0 Humana Press Inc , Totowa, NJ

271

Collins et al

272 2. Materials 2.1. lsola tion 1 Fmoll (Sigma, St. Louis, MO) 2 Phosphate-buffered salme (PBS) (Sigma)

2.2. Culture Devices and Setup Human long-term medium (HLTM) a 300 mL McCoy’s 5A medium (Sigma) b 50 mL horse serum (HS, StemCell Technologies, Vancouver, BC) c 50 mL fetal bovine serum (FBS, StemCell Technologies) (see Note 1) d 4 mL 100X nonessential ammo acid solution (Irvme Scientific, Irvine, CA) e 4 mL 100 mM sodium pyruvate (Sigma) f 4 mL 100X L-glutamine (Sigma) g 4 mL 100X cr-thioglycerol (Sigma) solution (l@ Mm McCoy’s 5A medium) h 4 mL 100X MEM vitamm solution (Irvme Scientific) 1 4 mL hydrocortisone (Sigma) solution (1 OA A4 m McCoy’s 5A medium) (see Note 2) J 2 5 mL 50X essential ammo acid solution (Irvine Scientitic) k 0 4 mL gentamycm sulfate (Sigma) XVIVO-20 serum-free medium (BioWhittaker, Walkersville, MD) Hematopoietic cytokmes stem cell factor (SCF, Amgen, Thousand Oaks, CA), Interleukm 3 (IL-3, Nova&, East Hanover, NJ), Interleukm 6 (IL-6, Novartis), granulocyte-colony stimulating factor (G-CSF, Amgen), granulocyte/monocyte-stlmulatmg factor (GM-CSF, Immunex, Seattle, WA), erythropoietm (Epo, Amgen) PBS (Sigma) Trypsm (Sigma) 24-well tissue culture plate (Falcon, Lincoln Park, NJ) 25 cm2 tissue culture flask (Falcon) Model 1967- 100 spinner flask assembly (Bellco, Vmeland, NJ ). Do not sthcomze the flask (see Notes 3 and 4) Replace the standard agitator assembly with the agitator assembly designed for use with Bellco’s microcarrier spinner flasks (1965 series) This agitator assembly contams a flat blade impeller that improves cell suspension over the 1967 model’s magnetic stir bar (see Note 5) Remove the magnetic stir bar and flat blade from the agitator shaft assembly and look inside the shaft housing If there IS a small sihcone component mside the shaft housing, remove it (see Note 6) AdJust the agitator shaft so that the impeller assembly hangs freely above the flask bottom. The impeller assembly should rest l-2 mm above the flask bottom

2.3. Ghan tita tive Analysis 2 3 1 Total Nucleated Cells 1 Cetrimide solution Combme all components and filter mto bottles a 90 g cetrimide (Hexadecyltrimethylammonlum bromide) powder (Sigma)

Human Hema topoietic Cell Cultures

273

b 25 g NaCl (Sigma) c 1 1 g EDTA (Sigma) d 3 L deionized water (dH,O)

2 3.2. Colony-Forming

Cell Assay

2 3.2 1. COLONY ASSAY MEDIUM Complete colony assay medium (methylcellulose) StemCell Technologies (see Note 1) AlternatIvely, can be made from mdlvldual components.

can be purchased from the colony assay medmm

1 2X IMDM + 2-mercaptoethanol a Prepare IMDM solution with twice the required IMDM (Sigma) and sodmm bicarbonate, as specified m the manufacturer’s mstructlons b Add 25 4 p.L of 2-mercaptoethanol(2-ME) solution (Fisher Sclentlfic, Itasca, IL) to 1000 mL of 2X IMDM c Add 2 mL of gentamycm sulfate (Sigma) to the 1000 mL of 2X IMDM + 2 ME 2 Methylcellulose stock solution. a Autoclave 800 mL of ddHzO with a 4-m magnetic stir bar m a 2-L Erlenmeyer flask b Place the autoclaved flask on a heated stir plate and add 33 2 g of methylcellulose powder (Glbco, Grand Island, NY) while the water is still hot c Slowly Increase the heat level on the stir plate until the methylcellulose solution begms to boll The slurry has a tendency to boll over, so watch the bollmg process carefully. Remove the flask from the stir plate if necessary d Remove the flask from the stir plate to stop bollmg Repeat the bolllng process 2x e Allow the methylcellulose solution to cool to room temperature Slowly add 800 mL of the cold 2X IMDM solution prepared in step 1 The resultmg solution will retain some of the cloudmess of the pure methylcellulose solution f Place the methylcellulose-IMDM solution onto a stir plate inside a cold (4°C) room Allow the methylcellulose to stir over a 2-d period. The methylcellulose should now appear optically clear. g. Ahquot the methylcellulose stock solution mto sterile 50-mL tubes and freeze at -20°C until needed 3 Colony Assay Medmm (CAM) a Place the followmg components mto each of four 50-mL sterile centrifuge tubes 3 75 mL IMDM, 3 35 mL BSA (StemCell Technologies), 15 mL FBS, 27 5 mL methylcellulose stock solution, 0 4 mL sterile ddH,O b Shake the contents of the tube until well mixed. Centrifuge at 750g for 20 mm A pellet of undissolved methylcellulose powder will be at the bottom of the tube. Decant the CAM mto a sterile 200-mL contamer Add the followmg growth factors at the concentrations given for liquid culture: SCF, IL-3, IL-6, G-CSF, GM-CSF, and Epo.

274 2 3 3 LTC-IC

Collrns

et al.

Assay

1 24-well plates 2 75cm* T-flasks 3 LTBMC medmm. LTBMC 10% HS, 4 mMglutamme, 5 pmol/L hydrocortisone

is formed from a base of 1X IMDM with 10% FBS, 100 pg/mL pemcillm, 10 U/mL streptomycin, and

2.3 4 Flow Cytometry 1 PBS 2 PAB. The followmg reagents are required. 4 L PBS, 20 g BSA fraction V (Sigma), 4 g sodium azide (Sigma) a. Dtssolve BSA and sodmm aztde m PBS under stnrmg and low heat (overnight, if needed) b Filter and store at 4°C Expires m 3 mo 3 NH&l lysmg solution The followmg reagents will be required 1 L dH,O, 8 26 g NH,Cl (Sigma), 1 0 g potassium bicarbonate (Sigma), 0 0037 g tetra sodium EDTA (Sigma) Stir until all the salts are dissolved Store at room temperature Expires m 3 mo 4 Proptdmm iodide The followmg reagents are required 1 mL ethyl alcohol, 0 002 g propidmm iodide (Sigma); 9 mL PBS a Dissolve propidmm Iodide m ethyl alcohol b Dilute the solution with PBS, wrap the container m aluminum foil, and store at 4°C. Expires m 1 yr 5 Fluorescent antibodres FITC IgG (Becton-Dickinson, San Jose, CA), PE IgG (Becton-Dickinson), FITC CD45 (Becton-Dickinson), PE CD45 (Coulter, Hialeah, FL), FITC CD15 (Becton-Dickmson), PE CD1 lb (Becton-Dickmson) 6 Polystyrene round-bottom tubes, 12 x 75 mm (Becton-Dmkmson Labware, Franklin Lakes, NJ) 7 1% paraformaldehyde fixative solution The followmg reagents are required 2 1 4 g cacodylate acid sodium salt (Sigma), 2 L dH,O, 20 g paraformaldehyde (Sigma); 15 g NaCl (Sigma) a Dissolve cacodylate acid sodium salt m water b AdJust pH to 7 2 using 1 N NaOH c Add paraformaldehyde, stn-rmg on a hot plate m a fume hood until dissolved d Add and dissolve NaCl e Filter sterilize and store at 4°C Expires in 1 yr

3. Methods 3.1. /so/a tion Hematopotettc cells for culture are obtained from three maJor sources: bone marrow aspirates, umbtltcal cord blood, and moblllzed peripheral blood. Depending on the source and the nature of the intended experiment, separation

and isolation of particular hematopoletlc populations may be necessary The

Human Hematopoletic

275

Cell Cultures

two most commonly utiltzed separation techniques for hematopotetic cells are mononuclear cell (MNC) separation from whole blood (Ftcoll) and CD34” cell selection from mononuclear cells. 3.7.1. MNC Separation from Who/e Blood Of the three cell sources menttoned above, both bone marrow aspirates and umbtltcal

cord blood samples

require separation

of the MNC

fraction

because

of then high RBC content Mobthzed peripheral blood 1smore variable m rts RBC content, but can often be used as obtained from the apherests process (see Note 7). Frcoll 1sslightly more dense than water (1.077 g/mL). Cells will be separated based on relative density, with denser RBCs and mature neutrophtls passing through the Ficoll, and MNC retained on the Frcoll-cell interface. Place 15 mL of 37°C F~coll III a sterrle 50-mL polypropylene tube Wtthdraw 5 mL of sample mto a 25-mL ptpet Using the same pipette, remove an addttronal 20 mL of 37°C PBS, brmgmg the total volume m the prpet to 25 mL Slowly prpet the cell suspensron on top of the Ficoll (see Note 8) Repeat step 1 with as many tubes as necessary to separate the entire sample Place the tubes mto a centrifuge, being careful not to mix the two layers Run the centrrfuge at 300g for 30 mm. When the tubes are removed, you should notice a red pellet at the bottom of each tube and an opaque layer of cells at the F~coll-PBS interface Carefully remove the interface layer from the tube with a lo-mL ptpet, being careful not to remove much of the Frcoll Combme all the interface layers into one or two 50-mL tubes and centrtfuge at 300g for 10 min to pellet the MNC (see Note 9) Aspirate the supernatant off the MNC pellet and wash the cells with the medium that will be used m the experiment. Count the cells and resuspend to a density that will facilitate the setup of the experiment.

3 1 2. CD34+ Cell Selection CD34 1san antigen expressed on more primmve hematopotetic cells. Cells at the colony-formmg cell stage, and those more prtmttive, express CD34. Colony-forming cells are capable of greater prolrferatton than more drfferentiated cells, and, as such, cultures rich m CD34+ cells (selected populattons range from -7O-95% CD34+, depending on technique) will undergo a larger total cell expansion than those with a lower CD34+ content (such as MNC cultures, which are typically O.l-10% CD34+, depending on the cell source). The relatively undifferentiated nature of CD34+ cells also allows for flexrbiltty in dnectmg the culture mto lineages of Interest. Several methods exist for separatton of MNCs mto CD34’ and CD34- populattons. All methods rely upon the use of an antibody that recognizes the CD34 antigen and subsequent recovery

276

Collins et al

of the cell-antibody complex The MmiMACS (Miltenyt Btotech, Sunnyvale, CA) CD34 separatton column provides excellent separation effictenctes (3) m 2-h time (see Note 10) In this procedure, the CD34 antibody 1sattached to a magnetic microsphere and the posittve population ts trapped within a magnetic field, but the negattve populatton 1swashed free Use the protocol supplied by the manufacturer (see Note 11). 3.2. Culture Devices and Setup Tissue-culture-treated polystyrene well plates and T-flasks provide an excellent growth envtronment for hematopoietic cells. Hematopotettc cultures m these systems are stmrlar to those of other animal cell cultures, except that hematopoiettc cells are largely nonadherent. The homogeneous envtronment of a stured system 1s excellent for experiments requiring frequent sampling and charactertzatton. The utthzatton of stirred systems presents some challenges, and we therefore present a method for the setup and culture of hematopoiettc cells in lOO-mL spinner flasks. 3.2.1 Cytokine Choice The cytokmes listed m Materials (Subheading 2.2.) will yield an acceptable expansion of total cells and all types of progenitor cells. The cytokmes listed can be classified into three groups: a group acting on primtttve hematopotetic cells (SCF), a group (IL-3, IL-6, GM-CSF) acting on a wide array of progemtor cells; and a group (G-CSF, Epo) acting on more mature, lineage-restricted cells This list of cytokmes 1sby no means exhaustive. There are several addtttonal early-acting (Flt3-hgand), multilmeage (PIXY321), and lineage-restrtcted (thrombopotetm, macrophage-colony stimulatmg factor) cytokmes that can also be used (4,5) It may not be posstble to obtain all the recommended cytokmes because of cost or avatlabthty. In this case, tt is generally acceptable to use a combmatton derived by choosing at least one cytokme from each of the three aforementtoned groups. Consult literature references to determine opttmal concentrations if their use 1sdesired. Additionally, if the production of a spectfic lineage of cells is desired m culture, consult literature references to determme which cytokines should be included m or excluded from the medium. 32.2 Brocompathlity of Mater/a/s Tissue-culture-treated polystyrene, commonly uttltzed m the constructton of well plates and T-flasks, is brocompattble with hematoporetrc cells. However, other materials commonly used for the construction of culture devices for animal cells may not be compatible with hematopoiettc cells Serum can partially protect hematopoietrc cells from the negative effects of some construction materials. Material compatrbrhty 1stherefore especially important tf the

Human Hematopoietic Cell Cultures

277

culture is carried out m serum-free medium. Silicone, glass, and polycarbonate are a few of the materials that adversely affect hematopolettc culture performance, as tdentlfied in a recent publication (6). When designing custom culture systems,the relattve performance of hematopotetlc cells on these matertals should be examined. 3.2.3. Static Culture Initiation and Maintenance 1 Using the medium of choice (see Note 12), add the proper concentrations of cytokmes, and suspend the cells to the desired moculum density (see Note 13) Recommended cytokme concentrations are: SCF (50 ng/mL), IL-3 (5 ng/mL), IL-6 (10 ng/mL), G-CSF (150 U/mL), GM-CSF (200 U/mL), Epo (3 U/mL) 2 Pipet the proper volume of cell suspension (see Note 14) mto a 24-well plate or 25-cm2 tissue-culture flask (see Note 15) 3 Incubate the cultures m a fully humidified atmosphere of 5% CO2 and air, or 5% CO,, 5% 02, and balance N2 (see Note 16). 4. Feed the cultures periodically (see Note 17), as follows* For well plates a Add 1 mL of fresh medium and cytokmes to each well. If the volume m a well has reached its maximum (2 mL), then the contents of the well can be split mto two wells, with 1 mL fresh medium and cytokmes added to each well For T-flasks a Remove one-half of the cell suspenston from the flask, and place in a sterile 15-mL polypropylene tube Centrifuge this fraction for 10 mm at 3008 b Using a ptpet, gently remove the supernatant, and replace it with fresh, prewarmed (37°C) medium and cytokines c Gently resuspend the cell pellet, and return the cell suspension to the original T-flask. 5. To perform the desired cell assays on a given day for T-flask cultures, simply remove the required cell suspension volume. To perform the assays on well-plate cultures, the contents of the entire well must be harvested in the followmg manner: a Withdraw the entire cell suspension from the well using a pipet b Wash the well with 250 pL PBS c Add 250 pL of prewarmed trypsm (37°C) to the well, and Incubate the well plate at 37°C for 10 mm. This step (along with the next two) can be neglected if only the nonadherent cell population IS to be examined, as IS the case m T-flasks d Add 250 p.L of HLTM to the well, and scrape the well bottom using a ptpet (see Note 18)

e Withdraw the adherent cell fraction, and wash the well with 250 pL PBS

3.2.4. Spinner Flask Initiation and Maintenance 3.2.4.1

SPINNER CLEANING AND PREPARATION

1 Aspirate out the previous contents of the spinner If the spinner flask is new, proceed to step 2.

278

Collms

et al.

Fill the spmner flask with ddH,O and autoclave the flask for 20 mm Do not bleach the flask (see Note 19) Discard the sterilized water and flush the flask well with water Remove the flask side-arm caps and agitator assembly for separate washing F111the flask with water and a detergent suitable for culture glassware. Scrub the inner walls of the spmner flask and rmse well with water Remove the agttator stir bar and flat blade from the shaft assembly Wash the bar and blade well Scrub the shaft housmg and rmse the Inside portion of the shaft well Wash the side-arm caps Reassemble the agitator assembly and the spinner flask assembly Rinse the entire spinner flask assembly well with deionized H,O and then fill with 10 mL of ddH,O Sterilize the spinner flask for 20 mm m an autoclave Aspirate out the sterile water m the spinner flask and rinse the inside of the flask with 20 mL sterile PBS Swirl the flask to contact as much of the spinner surface as possible with PBS. Aspirate out the PBS The flask is now ready to use 3 2 4.2 SPINNER FLASK CULTURE INITIATION 1 For a serum-containing medmm, the mmimum acceptable seedmg density for cord blood or peripheral blood mononuclear cultures is 2 x lo5 cells/ml, serum-free cultures requne an moculum density of 3 x lo5 cells/ml For CD34+ cultures, 5 x lo4 cells/ml IS the mmimum density for either type of medmm We have Initiated MNC cultures as high as 1 5 x lo6 cells/ml with good results Inoculum densities for CD34+ cultures should not exceed 1 2 x 1O5cells/ml (see Note 20) A good intermediate seeding density for both serum-contaimng and serum-free MNC cultures is 5 x 1O5cells/ml and 7 5 x 1O4cells/ml for CD34+ cell cultures 2 Inoculate the culture with 3&50 mL of the cell suspension 3 Place the spmner flask on a magnetic stir plate at 30-40 RPM m a 5% CO, mcubator (see Notes 21 and 22) Spmner cultures will grow well m either a 5 or 20% oxygen environment 3 2 4.3. SPINNER MAINTENANCE Depending of the purpose of the experiment being conducted in the spinner, the culture duration will vary. We have carried out cultures as long as 28 d, although 10-14 d are usually sufficient to observe large changes in many hematopoietrc cell populations. Regardless of the length, most cultures will requrre medium exchange. Many protocols can be used successfully to mamtam sufficient nutrient levels and low byproduct levels The protocol presented here is the most commonly utthzed m our spmner cultures, and serves the majority of cultures well. 1 On d 4, use a pipet to remove 50% of the culture volume and place m a sterile 50-mL polypropylene tube. Pellet the cells by centrifugation at 300g for 10 mm Place the spmner flask, with its remammg cells, back mto the mcubator durmg this time (see Note 23)


279

2 Aspirate the medium supernatant off the cell pellet and replace with an equlvalent amount of fresh, prewarmed (37T) medrum plus cytokmes Plpet the medium up and down to resuspend the cell pellet. Return the cells to the spmner flask 3 Repeat the feeding procedure every other day until the end of the culture

(see Note 24)

3.3. Quantitative Analysis 3 3. I Total Nucleated Cells The increase or decrease m the number of total nucleated cells present m the culture IS a good measure of the effectiveness of the particular experlmental condltlon m stlmulatmg hematopoletlc expansion. Nucleated cells are of pnmary interest, because they are capable of further prohferatlon or dlfferentlatlon Mature RBCs do not have a nucleus, and therefore are of less interest m hematopoletlc cell culture. For this reason, cell lysls IS performed and nuclei are counted. The slow growth kmetics of most hematopoletlc cultures makes a once-a-day analysis adequate Use a Coulter Counter or slmllar apparatus to count the cells (see Note 25) 1 A total of 10 mL of 37T cetrlmlde (see Note 26) and culture sample should be mixed m a coulter vial, I e , 9 9 mL cetrlmlde and 100 pL of sample Other dllutlon ratios can be used (see Note 27) 2 Set the gates on the Coulter counter to measure particles (cell nuclei) between 2 5 and 7 5 ~1 3 Perform a count on cell-free cetrlmlde to determme the background level within the sl.ze range of Interest 4 Count the nuclei m the prepared Coulter vial and subtract the blank value Based on dllutlon ratios, calculate the nucleated cell density present m the culture

3 3.2. Colony-Forming Cells (CFC) Counting total nucleated cells only measures overall prohferatlon of the culture The heterogeneous nature of hematopoletlc cell culture requires the assay of specific cell types within the nucIeated population Colony-forming cells (CFC) give rise to the mature, dlfferentlated cells of their lineage The measurement of CFC therefore provides insight as to the ability of the experimental condltlon tested to expand cells of a particular hematopoletlc lmeage Assays for CFC rely on immoblhzatlon of the cultured cells in a semisolid medrum, until the dlfferentlated progeny of the CFC accumulate to a degree that they can be dlstrngulshed under a microscope The CFC assay should be performed along with the nucleated cell count. 1 Place 3 mL of colony assay medium (CAM) into a sterile 17 x loo-mm polypropylene tube Allow the medium to warm to 37’C (see Note 28)

Collins et al

280

Calculate the volume of cells necessary to Inoculate the colony assay medium at a sultable density (see Note 29), such as 10,000 cells/ml Multiple seedmg densltles are recommended, untd experience with the particular parameters allows for the optlmal choice to be made Plpet the culture sample mto the CAM, and vortex well The CAM should be removed from the tube with a positive-displacement repeat plpeter (see Note 30), and ahquoted mto 354nm Petri dishes (1 mL per dish) The

35mm dishes should be placed inside a larger Petri dish (150 mm) contammg 15 mL of sterile water, which helps control local evaporation. Culture for 13-l 5 d m a 5% 02,5% CO,, balance N, incubator at 37°C (see Note 31) Examine the cultures under a dark-field

stereo microscope at x25-35 magmfica-

tlon Colonies of granulocytes, which arise from the colony-forming

umt-granu-

locyte (CFU-G), will appear as highly refractlle, white colonies of greater than 50 cells Colonies of monocytes/macrophages, which arise from the CFU-M, are also white colonies of greater than 50 cells They are usually dlstmgulshable from

granulocyte colonies through exammatlon of mdlvldual cells of the colony Indlvldual cells of a monocyte colony are generally larger and less refractlle than

mdlvtdual members of a granulocyte colony Monocyte colonies are often less tightly packed than granulocyte colomes CFU-GM

(granulocyte/monocyte)

give

rise to colonies that contam cells from both the G and M lineages Colonies ansmg from the burst-forming unit erythrold (BFU-E) are pmk to deep red m appearance and contain subunit clusters, which can be numerous (see Note 32) From the number of each CFC type counted and the plating density, calculate the number of each CFC type present m the sample For example, if the colony assay

was inoculated at 5,000 cells/ml

and 50 granulocyte colonies were counted after

mcubatlon, the frequency of occurrence of CFU-G m the sample 1s l/100

3 3 3. Long-term Culture-lnitlating The LTC-IC

Cell (L TC-IC)

assay detects cells close to the stem-cell

compartment

Many

variations of the LTC-IC assay exist (7-9), but most are centered around the concept of culturmg cells for a 5-wk time period, after which the cells are plated into a standard CFC assay Only very prlmltive cells will not have termmally differentiated in 5 wk, and hence will be able to form colontes in the CFC assay. Thus, these colonies represent LTC-IC. Because of the time Involved m carrymg out the LTC-IC assay, It IS not usually performed as frequently as the CFC assay We outline a modification of the method described by Koller et al. (10) 3 3 3 1 STROMA PREPARATION Most LTC-IC assays involve the use of accessory cells (known as stroma) from the bone marrow. Stromal cells produce growth factors and extracellular matrix molecules that support hematopoietuz cells. Bone marrow aspu-ates are the traditional source of stromal cells for this assay (see Note 33)


281

Frcoll the bone marrow aspnate using the method described above for MNC tsolatton Seed T-75 flasks with 1 x lo6 MNC/mL m a total of 10 mL of LTBMC medium Incubate at 33°C for 2 wk m a 5% CO, incubator Perform a 50% medium exchange wrth LTBMC medmm once a week The stromal cells should have formed a confluent layer on the bottom of the T-flask. Remove the medium from the flask and wash wtth 10 mL of PBS Remove the stromal layer by trypsnnzmg at 33°C for 10 mm Pellet the collected stromal cells and wash twice with LTBMC. Determine the nucleated density of stromal cells Resuspend the stromal cells so that the denstty IS 5 x 1O4cells/ml Irradiate the stromal cells at 20 cGy (see Note 34) Prepare 24-well plates for the assay The outer rmg of wells should be filled with 2 mL of sterrle water per well to reduce medium evaporation over the 5-wk culture period Prpet 1 mL of stromal cell suspensron mto each of the eight center wells Each well should therefore contam 5 x lo4 stromal cells Each LTC-IC assay will use three wells Use as many well plates as needed to carry out the desired number of LTC-IC assays. Allow the stromal cells to attach to the bottom of the well plate overnight The stromal layers are then ready for use m the LTC-IC assay 3 3.3 2. LTC-IC INITIATION Remove 3 x lo6 cells from the culture of interest Pellet the cells and remove the supernatant Resuspend m LTBMC medmm at a densrty of 1 x lo6 cells/ml Place I mL of the suspension onto one of the eight prepared stromal layers wtthm the well plate The assay should be done m triplicate (see Note 35) Each week, carefully remove I mL of medium from the top of the culture wtthout removmg cells at the bottom of the well Replace the dtscarded medium with fresh LTBMC medium The culture 1scarried out for a total of 5 wk At the end of the culture, remove all nonadherent cells, using a prpet. Trypsnnze the adherent layer and combme wtth the nonadherent cells Rinse the well with 250 pL of LTBMC to quench the trypsm. Determine the nucleated cell density of the suspension. Inmate standard CFC assays at seedmg densities of 2.5 x 1O4and 5 x 1O4cells/ml Enumerate the number of colomes at d 13-l 5 of culture m the same fashion as for the CFC assay.Most colonies will either be CFU-M or CFU-G. BFU-E appear more rarely An LTC-IC 1sgenerally thought to dtfferentrate mto 4 CFC The LTC-IC frequency of the culture 1stherefore Number of CFC counted/4 [Colony assay seeding density]

3.3.4. Flow Cytometry Flow cytometric techniques can be used to identify and quantrfy a specific populatron of cells m a heterogeneous mixture of other cells, based on cell size, granularity, and surface antigen expression. Because the expression of surface antigens on hematoporettc cells changes as the cells drfferentlate, flow cytometry can also be used to monrtor cell differentiatron in the cultures.

Collins et al.

282 3 3 4 1 CELL STAINING

1 A mmlmum of four polystyrene tubes will be required (three tubes to adjust the flow cytometer settmgs, and one analysis tube) Add 1 x IO5 - 1 x lo6 cells to each of the tubes 2 Centrifuge the tubes for 3 min at 35OOg and decant the supernatant 3 Vortex the pellet, and wash with 1 mL PAB (centrifuge and decant supernatant) 4 Add the antibodies to each pellet as described below (see Notes 36 and 37), vortex the pellet, and incubate the tubes m a 4°C refrigerator for 15 mm Tube 1 FITC lsotype controUPE lsotype control (see Note 38) Tube 2 FITC CD45 Tube 3 PE CD45 Tube 4 FITC CD 15/PE CD 11b (see Note 39) 5 Add 3 mL NH,Cl red cell lysmg solution, and shake each tube for 1 mm Centnfuge the tubes and decant the supernatant, blotting each tube 3x on a paper towel to remove all of the lysmg solutron 6 Wash the cell pellets twice with PAB 7 Resuspend the cell pellet m 1 mL PAB for immediate use, or 1% paraformaldehyde for later use (see Notes 40 and 41) 3 3 4.2

DATA ACQUISITION

1 Turn the flow cytometer on Depending on the manufacturer, different setup procedures (voltage adjustments, fluorescent bead callbratlon) will be required Follow the manufacturer’s mstructlons for preparmg the instrument for data acqulsltlon 2 Set the machme to acquire data contmuously (without saving to disk), place tube 1 on and display a dot plot of forward scatter (FSC) vs side scatter (SSC) AdJust the linear forward and side scatter values until the cell population(s) are centered on the plot (see Fig. 1 and Note 42) Adjust the threshold on forward scatter to ellmmate any debris 3 St111usmg tube 1, display fluorescence channel 1 (FL 1) vs fluorescence channel 2 (FL2) The lsotype control should be uniformly negative for both FL 1 and FL2 Any positive events are the result of nonspecific staining AdJust the FL1 and FL2 voltages until the cell population sits m the lower left portion of the plot (negative for FL 1 and FL2; see Fig. 2 and Note 43) Place a quadrant around this population 4. Still acquiring the data continuously, place tube 2 on, and display FL1 vs FL2 FITC CD45 should only be positive m FLl, but because of shght spectral emlsslon overlap between FITC and PE, some fluorescence is detected by FL2 Therefore, adjust the FL2-%FLl compensation until the cell population lies m the lower right quadrant (see Fig. 3 and Note 44) 5 Still acqulrmg the data contmuously, place tube 3 on, and display FL1 vs FL2 PE CD45 should only be posltlve m FL2, but because of slight spectral emlsslon overlap between PE and FITC, and PE and propldmm lodlde, some fluorescence is detected by FL 1 and FL3 Therefore, adjust the FL l-O/nFL2 compensation until the cell population lies in the upper left quadrant (see Fig. 3).

283

Human Hematopotetic Cell Cultures

0

200

400

600

600

1000

FSC

Fig. 1 Dot plots of forward scatter vs side scatter for uncultured peripheral blood cells (apheresls product) (top) and cultured cells (bottom) The lymphocytes (low forward and side scatter) and monocytes (mtermedlate forward scatter and low side scatter) are clearly separated Note the absence of granulocytes (intermediate forward scatter and high side scatter) m the apheresls product 6 St111usmg tube 3, display FL2 vs FL3. Adjust the FL3-%FL2 compensation until the cell population hes m the lower right quadrant. 7 Save the instrument settmgs for future use (see Note 45) 8 Add 10 pL of propldmm lodlde solution to tubes 1 and 4 (see Note 46) 9 The samples are now ready to be run. Place tube 1 on and acquire no less than 10,000 events Save the data on a floppy disk. Repeat this step for each of the remammg tubes

284

Collins et al

Ftg 2 Dot plot of pertpheral blood cells stamed with IgG control antibodies 3 3 4 3 DATA ANALYSIS 1. Recall the saved file, and dtsplay a dot plot of side scatter vs FL3 Using the computer mouse, draw a region around the FL3 negative cells These are the viable cells 2 Display a plot of FL 1 vs FL2, using only the viable cells determmed m step 1 Using the quadrants estabhshed in tube 1, the percentage of vtable cells that are CD15CD1 lb-, CD 15+CDl lb-, CDlS-CD1 1b+, and CD1 5+CDl lb+ can now be determined The amount of nonspecific stainmg in each quadrant (determmed from tube 1) should be subtracted from each of these values to obtam more accurate results 4. Notes Both horse serum and fetal bovine serum may be obtained from alternate vendors such as HyClone (Logan, UT), Gibco, Sigma, and BioWhittaker StemCell Technologtes prescreens then serum, as well as serum components (such as BSA), for use m hematopoiettc culture, which ensures good results Serum lots from other vendors must be screened to determme tf they are suttable for hematopotettc culture Hydrocortisone is only needed for the growth of adherent cell populattons m bone marrow If peripheral blood or cord blood is the source of the culture moculum, hydrocortisone may be omitted from HLTM StemCell Technologies also sells an equivalent flask designed to culture hematopoletic cells Sihcomzation is not necessary because the cells will not adhere to the wall. Monocytes will adhere to each other and form small clumps The addition of the sthconization compound to the flask can only harm culture performance vta mtroductton of toxic materials to the culture environment The 1965 model flasks designed for microcarrier cultures will also work well for hematopotettc cells, but require slightly more volume than the flat-bottom I967 model.

285


0'

4

Fig. 3 Dot plot displays for compensatton tubes #2 (top) and #3 (bottom) of FL2 vs FL3 for tube #3 IS omrtted

Drsplay

5 The addrtron of the flat-blade impeller 1scrrttcal to obtaining culture expansion. WrthJust the magnetic stir bar, cells will clump under the agitator and dre Agitator flat-blade impeller to vessel diameter ratios of O&O.8 work well 6 The sillcone component IS intended to serve as a cushion between the glass shaft and the soft teflon flat-blade impeller This component IS difficult to clean, though, and can harbor leachables that negatively affect cell proliferation The cell death associated with this component 1smuch more pronounced m serum-free culture 7 Bone marrow and umbilical cord blood can be cultured wtthout MNC separation, however, the high RBC content can make visual observatron of culture growth quite drfficult. Peripheral blood samples should be low m red blood cell content because of the harvest procedure utilized (leukapherests), which separates the MNC from the RBCs However, variations in the leukapheresis protocol will alter the number of RBC m your sample

Co//rns et al. 8 It IS crltlcal that the cells be plpeted slowly onto the Flcoll layer If the rate of plpetmg IS too rapld, cells will nnx with the Flcoll and make separation Impossable When all of the cells are deposited on the F~coll, you should note a clear demarcation between the Flcoll and blood 9 Do not leave the cells m the Flcoll for any longer than necessary Extended exposure of cells to the Flcoll can reduce vlablhty 10 For apphcatlons that require extremely high purity (>95%) of CD34+ cells, we would recommend fluorescence-activated cell sorting (FACS) Additionally, the MmlMACS system does not remove the antibody-magnetic bead complex from the cell If this presents a concern, other methods exist that ~111 remove the antlbody (3). 11 The protocol from Mlltenyl is relatively straightforward and easy to use successfully We caution that utlhzatlon of the supplied reagents beyond their expu-atlon date will result m a low yield of CD34+ cells The protocol does require the use of degassed buffer for one step This can be accomphshed by retrofittmg a bottle-top filter, so that the filter IS replaced by a rubber cork The buffer IS poured mto the bottle and the modified filter top 1splaced onto the bottle Attach a vacuum hose to the filter nozzle and mltlate vacuum Gas bubbles ~111be drawn out of the buffer 12 Either serum-contammg or serum-free medium can be used Serum-contammg medium generally gives higher progenttor cell and total-cell expansion, however, If a more defined medium IS desired (e g , for chmcal apphcatlons), acceptable expansion can be attained using XVIVO-20. Equlhbratmg the medium with the mcubatlon atmosphere and temperature 1s recommended pnor to suspending the cells 13 The chosen moculum dens&y for MNC cultures should not exceed 500,000 cells/n& because higher values will deplete key nutrients quickly, necessltatmg tedious feeding schedules Additionally, to ensure reproducible cell expansion, the moculum density should not be below 50,000 cells/ml In general, lower-density cultures will exhibit a greater expansion of total cells and progenitor cells than higher-density cultures However, if a large number of cells are required for specific assays (such as flow cytometry), high-density cultures are recommended For CD34+ cultures, an appropriate moculum density 1s 20,000-50,000 cells/ml 14 In the 24-well plate, maculate 1 mL of the cell suspension mto each of the eight center wells The 16 outer wells should be filled with 1 5-2 0 mL sterile water to mmimlze media evaporation from the center wells T-25 flasks should be moculated with enough of the cell suspension to wet the surface (at least 5 mL), but not too much (more than 10 mL), m order to prevent oxygen transport hmltatlons to the cells caused by excessive hquld height 15 6-well culture plates should be avolded, because they are prone to uneven evaporation patterns. Larger tissue-culture flasks can be used if a large culture volume IS desired. 16 For low- and mtermedlate-density cultures, either atmosphere will allow acceptable total-cell expansions Studies suggest that culturing at a reduced-oxygen tension may be beneficial for progenitor cell expansion (11-13) For high cell density MNC cultures (>400,000 cells/ml), a 5% O2 atmosphere should be avoided to preclude the posslblhty of creating a hypoxlc mtcroenvlronment for the cells restrng at the bottom of the cultures

Human Hematopoietlc

Cell Cultures

287

17 In general, the higher the seedmg density of the culture, the more often feeding ~111 be required Typically, low-seeding-density cultures (50,00@100,000 cells/ml) will not need to be fed more than once over 15 d, medium-density cultures (100,000-400,000 cells/ml) should be fed once or twice, and high-density cultures (400,000-500,000 cells/ml) should be fed at least twice. CD34+ cultures are usually fed once over 15 d A good guldelme for determmmg the proper time of feeding 1s depressed pH (~7 1, pale-orange to bright-yellow medrum) or a high concentration of lactate (above 15 mM, mdlcatmg a feeding requirement) In all cases, it 1s important to remember to equilibrate the medium with the mcubatlon atmosphere and temperature prior to feeding. 18 Generally, the adherent cell population will cover approx 3@-40% of the well bottom Scraping of the well bottom should focus on the perimeter, as most of the adherent fraction will be found there Visual exammatlon of each well using light microscopy will give a good idea as to when most of the adherent fraction has been removed 19 This addition of water 1s not absolutely necessary, but highly recommended In addition to klllmg cells m the flask, the hot water cleans the agitator assembly of any cell clumps that have adhered over the course of the prior culture The removal of bleach from some flask components is extremely difficult, even with copious flushing, hence the use of steam sterlllzatlon as a decontammatlon step 20 Occasionally, MNC cultures seeded as low as 1 x lo5 cells/mL will proliferate, but 2 x lo5 cells/ml 1s a more reliable density Although low-seeding-density cultures may not exhibit expansion of total cells, they quite often will realize limited expansion of CFC CD34+ cell cultures carried out above 1.2 x lo5 cells/ml will die within 6-8 d unless they are fed at least once a day begmnmg at d\ 4 Our seeding densities are designed for short-duration cultures (under 2 1 d) Longer-term spinner cultures may require seeding densities above 1 x lo6 cells/ml (24) 21 Agitation rates above 50 RPM will often negatively affect cell growth Suspension 1s adequate at 30 RPM, and, at volumes of 30-50 mL, the agitator blade breaks the llquld surface, provldmg sufficient oxygen transport 22 The standard animal-cell-culture magnetic stir plates sold by Bellco are intended for higher agltatlon rates As such, the agltatlon motion at RPMs less than 100 1s not smooth. We prefer the use of the Cellgro magnetic stir plate (Thermolyne), which 1s designed for agltatlon rates between 1 and 100 RPM 23 If the serum-free medium XVIVO-20 1s left outside of a CO, incubator too long, the pH will rise to > 8 0, and often will not return to normal (pH -7 4) when placed back into a CO2 incubator We therefore suggest that cultures carried out m this medium not be left out of the incubator for more than 5-10 mm 24 This feeding protocol retains all cells throughout the culture. Cultures can be diluted as m fed-batch, or can be reduced m density by replacmg medmm without cell retention It 1s our personal preference to retam the cells at constant volume 25 A hemacytometer may be used to make whole cell counts A difficulty m usmg this apparatus 1s that intact enucleated RBCs can be difficult to dlstmgulsh from

Collins

26.

27

28

29

30

31 32,

et al.

nucleated cells Since the point of the nucleated cell count IS to determme the density of nucleated cells, use of the hemacytometer can potentially result m inaccurate measurements, depending on the number of RBCs m the culture We have not had much success m counting nuclei with the hemacytometer Prewarmmg cetrtmtde is recommended prior to cell countmg. Dependmg on room temperature, cetrtmtde crystalhzatton can occur The crystal size is of the same order of magmtude as the nuclet stze, therefore, crystalltzed cetrtmtde can arttfictally raise the measured cell density The choice of dilution ratto 1s important to the accuracy of the cell count When the number of cells m the cetrtmtde is too high, clumpmg can occur These clumps may be counted as a single nucleus, thereby reducmg the measured cell density When the number of cells 1stoo low, the cell count may be of the same order of magnitude as the cetnmtde blank. Total counts of 5,00&40,000 work well (see ref IS) We have not nottced that unwarmed colony assay medium has an adverse effect on CFC viabtllty, however, cool medium will not mix well and can increase vartability between replicate plates Plating density 1s the most critical parameter for achieving accurate colony assays The goal m choosmg plating density IS to space CFC far enough apart that their resultant colomes are easily distmguishable from each other. High platmg densities result m overcrowdmg of the medium with CFC This overcrowdmg increases the difficulty m determmmg dtstmct colomes Low plating densmes often do not detect less frequently occurrmg CFC, such as the CFU-MIX A good number of total colonies to atm for 1s 100 per plate Generally, we find that the CFC content of the culture rises for the first 4-6 d and then declines We therefore adJust our densities m an attempt to mamtam a constant number of colonres per plate each day A plating density of 3,000 cells/ml works well until d 6 for mononuclear cultures A good rule of thumb ford 7-14 1sto multiply the day by 1,000 to get the plating density CD34+ cell cultures should be plated at 500 cells/ml for the first 5 d of culture, because of then high clonogemcity The plating denstty should be slowly increased to 5,000 cells/ml until d 10. After d 10, the gutdelmes for MNC plating apply If a growth factor combination other than the one suggested m the CFC protocol is used, these guidelmes may not apply For example, removal of SCF from the CAM will reduce the number of colomes detected at a given moculum density. The posmve-displacement ptpeter must be used to assure quantitative transfer into the Petri dishes A standard ptpeter with a standard ptpet tip will leave CAM on the walls of the pipet We use the Brmkmann (Westbury, NY) Eppendorf Repeater Ptpetter (Brinkmann #22 26 000-6). A standard CO1 mcubator can be used tf a controlled O2 incubator 1s not avatlable CFC detectton is improved at low 0, tensions (11-13) The descriptton of the difference m granulocyte and monocytelmacrophage colonies applies to most cultured cells However, the only accepted method to dtstmgutsh these cell types 1s to pluck the colonies from the plate and perform morphologtcal analyses. For this reason, these two distinct colony types are often lumped into a single category, referred to as CFU-GM (G + M + GM), and are

Human Hema topoie t/c Cell Cultures

33 34

35

36

37 38

39

40

41

42

289

mdlcatlve of total white CFC This CFC deslgnatlon IS not to be confused with the true CFU-GM, which IS a progemtor of both the CFU-G and CFU-M Murme stromal cell lmes (I 6,Z 7) and genetically engmeered varlatlons of these stromal cell lmes (18) have also been used successfully for the LTC-IC assay Other methods for stopping stromal cell prohferatlon exist (19), but Irradiation 1s preferred Irradiation prevents overgrowth of the stroma, while mamtammg (and even enhancmg) then ability to produce growth factors Accessory cells m the culture affect the outcome of the LTC-IC assay (20) The assay 1stherefore moculum-density-dependent For this reason, the assay 1scarried out m trlphcate at the one density of 1 x IO6 cells/ml The high seedmg density IS used to ensure that at least some LTC-IC are present m the sample of cells used in each assay well (LTC-IC exist at a very low frequency) The amount of antibody to add depends on the source of the antibody Generally, antlbody solutions are added at 20 pL per 1 x lo6 cells Consult manufacturer or antlbody data sheet to determine the appropriate amount Add the antlbodles to the cell pellet, with the tubes on ice to mmlmlze receptor internalization Because most nonspecific stammg occurs via bmdmg of the Fc region of the antibody to the cell, lsotype controls should be of same class as other antIbodIes (generally IgG) This antlbody combmatlon IS used to identify developing granulocytes and monocytes If quantlficatlon of other lmeages IS desired, different surface antigens must be investigated A brief summary of surface antigens characterlstlc to specific hematopoletlc lineages 1s given below. Monocytes. CD14, CD15, CD1 lb Granulocytes CD16, CD15, CDllb Megakaryocytes CD4 1a Erythrocytes Glycophorm A, CD7 1 Lymphocytes CD3, CD 19 Flxmg m 1% paraformaldehye preserves the cells for up to 1 wk However, Identlficatlon of dead cells usmg propidmm iodide is not possible after fixing the cells Therefore, whenever possible, run cells on the flow cytometer hve, because dead cell accumulation can become important m later stages of the culture If cells are not to be run nnmedlately, wrap the tubes m alummum for1 and store at 4’C m the dark to mnnmlze receptor mternallzatlon and maxlmlze the fluorescence signal Uncultured cells are clearly separated mto lymphocyte, granulocyte, and monocyte populations on dot plots of forward scatter vs side scatter Peripheral blood apheresls products will not contain any granulocytes, since they will have been removed m the apheresls process In contrast, cultured cells are seen as one dlffuse population possessing a higher forward scatter (because they are larger than uncultured cells). Accordingly, forward and side scatter values will have to be changed to properly view this population As a general rule, the cells to be analyzed (cultured or uncultured) can be located by looking for the population with

Collins et al

290

43

44

45

46

the highest forward-scatter Anythmg with a lower forward scatter represents debris or mature red blood cells, which can be removed from view using the electronic threshold Often, a tail-hke proJectton, which extends 45 degrees up and to the right, can be seen on this double-negative populatron. This tail 1s the result of two factors dead cells that nonspecifically bmd the antlbodtes, and autofluorescmg monocytes As a result, it may be difficult to move the entire population mto the lower left quadrant without cutting off any of the dots (events) Focus only on the mam cell population (excludmg the tall) when adJusting the voltages, because much of the tat1 ~111 disappear when the dead cells are removed m the data analysis Also, make sure the axes for FLI, FL2, and FL3 are logarithmrc Because of the presence of dead cells, sometrmes two cell populations can be seen m these plots a dense, compact cluster of cells, and a second, more diffuse smear of cells The compact population is usually the viable cell population, and admstment of compensation values should focus on this population The dead cells will be gated out m the data analysis Generally, the flow cytometer settmgs for uncultured cells will remam constant between cell samples. Therefore, these settings can be recalled for future samples without needmg to use tubes 1-3 Pertodtcally, however, the flow cytometer setup procedure should be repeated to ensure the settings are still appropriate This does not apply to cultured cells, whose settings often vary between samples Additionally, the settings for cultured cells can change as the culture time progresses Therefore, new settmgs should be determined for each day the culture 1s to be analyzed Add the proprdmm iodide no less than 1 mm and no more than 1 h before the data are to be acquired Propidmm iodide cannot be used with fixed cells

Acknowledgments Supported by National Science Foundation Grant BES-941075 1 and a predoctoral fellowshlp to P C Collms through the Natlonal Institutes of Health Predoctoral Biotechnology Training Grant (T32 GM08449). References 1 Brugger, W , Heimfeld, S., Berenson, R J , Mertelsmann, R , and Kanz, L ( 1995) Reconstttutton of hematopotests after high-dose chemotherapy by autologous progenitor cells generated ex viva N Engl J Med 333, 283-287 2 McAdams, T A., Winter, J N., Miller, W M , and Papoutsakts, E T (1996) Hematopoiettc cell culture therapies (Part II) chmcal aspects and applications Trends Blotechnol 14, 388-396 3 de Wynter, E A , Coutmho, L. H., Pet, X., Mars, J C W , Hows, J , Luft, T , and Testa, N G. (1995) Compartson of purity and enrichment of CD34+ cells from bone marrow, umbthcal cord and peripheral blood (primed for apherests) usmg five separation systems Stem Cells 13, 524-532

Human Hema topoletic Cell Cultures

291

4 McKenna, H J , devrtes, P , Brasel, K., Lyman, S D., and Williams, D. E. (1995) Effect of$fld hgand on the ex vzvo expansion of human CD34+ hematopotettc progenitor cells Blood 86,3413-3420 5 Debtlt, N , Wendlmg, F , Katz, A., Gutchard, J., Breton-Germs, J., Hunt, P , and Vamchecker, W (1995) The Mpl-ltgand or thrombopoietm or megakaryocyte growth and dtfferenttatton factor has both direct proliferative and differenttattve activities on human megakaryocyte progenitors. Blood 86,25 16-2525 6 LaIuppa, J A , McAdams, T A , Papoutsakts, E T , and Miller, W M (1997) Culture materials affect ex VIVO expansion of hematopotettc progenitor cells J Blamed Mater Res 36, m press 7 Quito, F L., Beh, J , Bashayan, 0 , Basthco, C , Basch, R. S (1996) Effects of fibroblast growth factor-4 (k-FGF) on long-term cultures of human bone marrow cells Blood 87, 1282-l 29 1. 8 Gordon, M. Y (1994) Plastic-adherent cells m human bone marrow generate longterm hematopotests m vitro Leukemza 8, 865-870 9 Traycoff, C M , Kosak, S T , Grtgsby, S , and Srour, E F (1995) Evaluation of ex VIVO expanston potenttal of cord blood and bone marrow hematopotettc progenitor cells using cell tracking and ltmitmg dtlution analysts Blood 85, 2059-2068 10 Koller, M R , Manchel, I , Palsson, M A , Maher, R J , and Palsson, B 0 (1996) Different measures of ex vzvo human hematopotetic culture performance are opttmtzed under vastly different condtttons. Bzotechnol Bzoeng 50,505-5 13 11 Koller, M. R., Bender, J G , Papoutsakts, E. T., and Miller, W M (1992) Beneficial effects of reduced oxygen tension and perfusion m long-term hematopotettc cultures. Ann NYAcad Scz 665, 105-I 16. 12 Koller, M R , Bender, J G , Miller, W, M., and Papoutsakrs, E T (1992) Reduced oxygen tension increases hematopotesrs m long-term culture of human stem and progenitor cells from cord blood and bone marrow. Exp Hematol 20,264270 13 Koller, M R , Bender, J G , Papoutsakis, E T , and Miller, W M (1992) Effects of synergtsttc cytokme combmattons, low oxygen, and irradiated stroma on the expansion of human cord blood progenitors Blood 80,403-411 14 Zandstra, P W , Eaves, C J , and Ptret, J M (1994) Expansions of hematopotettc progenitor cell populattons in stirred suspension bioreactors of normal human bone marrow cells Bzorechnology 12,909-914 15 Lm, A A , Nguyen, T , and Miller, W M (1991) A Rapid Method for Countmg Cell Nuclei Using a Particle Stzer/Counter Bzotechnol Techniques 5, 153-I 56 16 Sutherland, H J., Hogge, D E , Cook, D , and Eaves, C J (1993) Alternative mechantsms with and without steel factor support primtttve human hematopoieSIS. Blood 81, 1465-1470 17 Croistlle, L , Auffray, I , Katz, A., Izac, B , Vamchenker, W., and Coulombel, L (1994) Hydrocorttsone differentially affects the ability of murme stromal cells

and human marrow-derived CD34++/CD38-

adherent cells to promote the dlfferentlatlon

long-term culture-u-nttatmg

cells Blood 84, 41164 124

of

292

Collins et al.

18 Hogge, D E , Lansdorp, P M , Retd, D , Gerhard, B , and Eaves, C J (1996) Enhanced detectton, maintenance, and dtfferenttatton of prtmtttve human hematopoiettc cells m cultures contammg murme fibroblasts engineered to produce human steel factor, mterleukm-3, and granulocyte colony-sttmulatmg factor Blood 88,3765-3113. 19 Cmuttmt, F M , Martm, M , Salvarts, E , Ashman, L , Begley, C. G , Novotny, J , Maher, D , and Boyd, A W. (1992) Support of human cord blood progenitor cells on human stromal cell lmes transformed by SV40 large T antigen under the mfluence of an inducible (metallothtonein) promoter Blood 80, 102-I 12 20 Koller, M R., Palsson, M A , Manchel, I , and Palsson, B 0 (1995) Long-term culture-mttratmg cell expanston 1s dependent on frequent medium exchange combined with stromal and other accessory cell effects Blood 86, 1784-l 793

23 Methods to Isolate, Culture, and Study Osteoblasts Mechteld V. Hillsley 1. Introduction Primary cells are often desirable over clonal cell lmes, because they are more likely to retam the presence and actlvtty of certain enzymes and proteins that are often lost m clones. In additton, primary cells more closely resemble the cell m the actual animal than clones do. An often-cited drawback of prtmary cells is the heterogeneity inherent in them. Primary osteoblast cultures, for example, constst of preosteoblasts, osteoblasts, and osteocytes (if the culture is old enough), and possibly even a few Iibroblasts Such a mix of osteoblasttc cells, however, could be deemed beneficial and more realistic. The most detrimental cell contaminant 1sthe fibroblast. Ftbroblasts, if present m the culture, will eventually take over. We have never had this problem with our cultures Osteoblasttc cultures isolated by the method described m this chapter have been shown to be osteoblastic m nature by exhibiting extensive alkaline phosphatase acttvtty (j-3), by formmg nodules that stain posittve for mineral by a Von Kossa stain (1-51, and by showing a characteristic increase (three-fold) tn mtracellular CAMP m response to PTH stimulation (2 7 x 1O-gM for 15 mm) (1) The following method 1sa procedure, utilized successfully in our lab, for harvesting rat calvarial osteoblasts. The culturmg techniques employed are simple and cause little trauma to the cells The main drawback is the long turnaround time from start of isolation to the time when enough cells have grown for use m an experiment (2-4 wk). The technique employed is one based on a technique described by EcarotCharrler (6) and refined by Retch et al. (1) It is a mechamcal isolation procedure ltmmng the use of digestive enzymes, thereby hmitmg potential damage to the cells. The technique mvolves the selective migration of osteoblasts from calvarias onto glass chips For reasons that are not totally clear, osteoblasts From

Methods E&ted

by


Medme, and

Vo/ 18 Tissue Engrneenng

M L Yarmush

293

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

294

H~llsley

~111migrate from dissected, perlosteum free calvarlas onto glass chips Other cell types, such as fibroblasts, do not partake m this migration and are thereby kept separate from the osteoblasts. After 46 d of mcubatlon with the glass chips, the calvarlas are removed from the chips, and the cells that have migrated onto the chips are given time to grow and multiply. Only one enzymatic dlgestlon IS employed near the end of the lsolatlon and culturmg procedure, to remove the osteoblasts from the glass chips. 2. Materials 1 Sodmm bicarbonate washmg solution. Approx 2 g (the exact amount IS not critlcal) of sodium bicarbonate 1s dissolved m 1 L of dlstllled water This solution is made up Just prior to use 2 Basal medium Medium 199 (HyClone Laboratories, Logan, UT) (see Note 1) Other media are made usmg this basal medium Without antlblotlcs and other addltlves, It 1s only used for short time-periods and 1s not appropriate for cell growth 3 Complete growth medium Complete growth medium consists of basal medium supplemented with 10% fetal bovine serum (FBS) (HyClone) (see Note 2), 1% L-glutamme (200 m&Q, and 1% antlblotlcs (5000 U pemclllm and 5 mg/mL streptomycm, both from Sigma, St. LOUIS, MO) The FBS must be heat-mactlvated for 30 mm at 57’C prior to use All batches of serum should be tested for their ability to support osteoblast growth before use. 4 Buffers. Buffers used Include calcmm- and magnesium-free Hanks’ balanced salt solution (HBSS) and Dulbecco’s phosphate-buffered salme (DPBS) 5 Collagenase A* a A 1 g/L collagenase stock solution is made by dlssolvmg 0 5 g of collagenase b A (Boehrmger Mannhelm, Indlanapohs, IN) m 500 mL of DPBS, and addmg 1% antlblotlcs (5000 U pemclllm and 5 mg/mL streptomycin) The collagenase IS nonsterile and contams quite a few partlculates The solution IS thus sterilized by pressure-driven filtration (approx 20 psi) through a set of filters The first filter 1s a 0 45-p prefilter (Mllhpore Stenvex-HV), which IS followed by the final 0 22-p filter (Mllhpore Stenvex-GS, Bedford, MA) The 0 22-p filter usually lasts for an entire batch Often, however, the 0 45-p prefilter gets clogged and must be replaced l-2 times over 500 mL of collagenase solution The sterile stock solution 1s ahquoted and frozen Individual ahquots may be thawed and kept m the refrigerator for approx 1 mo (see Note 3) The stock solution should be sterile tested m combmatlon with growth medium before use b. The collagenase working solution of 0.2 mg/mL 1s made by dllutmg the stock solution fivefold with HBSS or any other calcium-free buffer This should be done Just prior to use 6. Flbronectm. Human fibronectm (1 mg) (Blomedlcal Technologies, Stoughton, MA) (see Note 4) 1s reconstituted by adding 1 mL of a sterile 2 M urea, 0 05 M

295

Osteoblasts

Tris-HCL, pH 8.0, solution at room temperature, and lettmg tt sit for at least 1 5 h. This solutton is then diluted 20-fold wtth HBSS. Ahquots are stored m the freezer. 7 Metofane is the drug used to euthamze the rat pups prior to the cell-harvesting procedure This, or other drug, IS obtained through a veterinarian working with the mstitution’s animal care and use committee (IACUC)

3. Methods 3.1. Preparation

of Glass Chips

Glass chips are prepared in advance and may be stored tn a sterrle container

mdetimtely (see Note 5). 1. Cover shps (No 1) are cleaned by placmg them, preferably, m a shde rack to prevent sttckmg, m a sodmm bicarbonate washing solution, which is then brought to a boil When botlmg commences, the heat 1sturned off The cover shps should be rmsed thoroughly soon after botlmg has stopped Rmsing should include extensive flushing wtth dtsttlled water, followed by a soak m clean dtsttlled water (at least 15 mm, and up to overnight), followed by another cycle of flushmg rmses The cover slips are then dried in a drier until completely dry. 2. Glass chips are made from the washed cover slips by wrapping the clean cover shps m several clean paper towels and crushing them by pounding the wrapped bundle wtth a hammer or other blunt object Care must be taken not to overcrush the chtps and thereby form only glass dust Caution: Safety glasses and gloves should be worn during thts step 3 Glass chips m the 250-500-pm size range are separated from larger and smaller chips by stevmg through appropriately stzed sieves (see Note 6) This size range was chosen because smaller size chips will float, thereby becommg a problem m cell culturing; and larger chips tend to cover too large an area of the calvarta, thereby limiting the number of cells m contact with glass chips Again, Caution: Safety glasses and gloves should be worn for this step 4 Sieved glass chips are placed m small autoclavable contamers (glass Petri dishes, for example) and steam autoclaved. They are now ready for use

3.2. Harvesting of Rat Calvarias Calvartas are harvested from 4- to 6-d old Sprague-Dawley rat pups. Younger rat pups are so small that the calvarras are tmy and paper thin, making them very difficult to handle and very easy to tear. Older rats provide fewer osteoblasts per calvana, probably because the cells are dtvidmg more slowly. The entire harvesting procedure (startmg with step 2) is performed aseptically in a lammar flow hood. The instruments used are autoclaved before the procedure. Smce generally a Inter of rat pups ~111 be handled at one sitting, the instruments are resterrhzed between pups by dipping in 70% ethanol after each use, and again dipping in 70% ethanol and flammg the remainmg ethanol m a propane flame mnnedrately before each use. The mstruments used Include sharp

Mlsley dissecting scwors, curved on the flat (for decapltatlon); a hemostat, blunt-tipped dissecting scissors (for cuttmg shn); delicate, sharp, spring-handled scissors (for cuttmg the calvartas), and four sets of forceps. All work mvolvmg the pups m the lamlnar flow hood IS performed on or over autoclaved paper towels These paper towels catch any spilled blood and waste scraps from the dtssectton The time involved for this step, not countrng settmg up, IS about 5 to 10 mm per pup. 1 The rat pups are euthamzed using Metofane Llqutd Metofane on a Q-ttp IS placed m a closed container with the pups The pups breathe the vaportzed metofane and dte wtthm l&15 mm The following steps are performed sequentially on one rat pup at a ttme 2 The rat pup IS dtpped m 70% ethanol (balance dtsttlled water) to dtsmfect It, and IS brought mto the lammar flow hood From this point on, all instruments used are dipped m 70% ethanol and flamed m a propane flame to burn off the ethanol 3 Using the curved sctssors, decapitate the rat so that the head falls onto the sterile paper towel, and dtspose of the body. Use the hemostat to grasp the head by the nose (see Note 7). Then, using the blunt-tipped scissors, cut the skm open along the center back of the head to the forehead. Cut down toward the neck along the top of the head to form two squares of skm Using blunt forceps, pull the skm off the skull (see Note 8) These forceps touch the outstde of the skm and should not be used thereafter to touch anything cleaner than the animal skm 4 The calvarta IS now exposed and needs to be cut and removed Usmg the delicate, sharp, sprmg-handled sctssors, cut out the frontal and partetal bones m the same pattern as the skm was cut, but staying away from the edge of the skm Usmg a different pair of forceps than those used m step 3, remove the pteces of calvarta (there should be two--one from each side of the skull) and place them m a Petrt dish of warmed basal medium (see Note 9) The Petri dash IS then placed m the mcubator while the remammg pups are processed

3.3. Stripping and Covering

the Periosteum from the Calvaria It with Glass Chips

This step must be done nnmedlately following the steps of Subheading 3.2. The cells are sitting m a mednrm that does not support growth (see Note 10) The followmg procedure should be done one dish at a time (i.e., one set of two calvarlas at once). Again, between sets of calvartas, all mstruments should be resterihzed by dlppmg m 70% ethanol and flaming over a propane flame, 1 Over sterlhzed paper towels, usmg two sets of forceps, carefully scrape off the perrosteal layer over the calvarta. This IS a tedrous process, and rt IS very easy to tear the calvarla As pieces of perlosteum are removed, they can be wiped off the forceps onto the paper towel 2 The pertosteum free calvarias are placed concave side up m clean Petrt dashes (6-cm diameter) (see Note 11) A scoop of glass chrps IS carefully placed into each calvarta The calvartas are then covered wtth warmed complete medmm by

0s teoblas ts

297

Fig. 1. Glass chip with osteoblasts that have migrated onto it. The chip has just been removed from the calvaria. x 100. placing several drops on the calvarias with chips directly before filling the remainder of the dish. This is to prevent spilling the chips off the calvarias. The dish is then placed in the incubator for 4-6 d.

3.4. Removal of Glass Chips from the Calvarias 1. After 4-6 d of incubation with the glass chips, the calvarias are removed from the glass chips. This is done by carefully transferring each set of calvarias, with the glass chips, to a new tissue-culture dish (one set to one 6-cm diameter dish). 2. Add 4 mL of warmed complete growth media to each plate while trying to flush the chips off the calvarias. Repeat the flushing of the calvarias by repipeting the same media over the calvarias several times. Try to get all the chips dislodged in this manner, while taking care not to splash media onto the rim of the culture dish. When all the chips have been dislodged, remove the empty calvarias from the dish and discard. Place the Petri dish in the incubator. 3. At this point some of the glass chips are covered with osteoblasts (see Fig. 1). These osteoblasts will grow onto the tissue culture dish bottom. Over time there will be enough cells for an experiment. Cells should be fed every 34 d, more frequently as the density increases. Generally, cells are ready for use during the period l-3 wk after removal of the chips from the calvarias.

3.5. Removing Osteoblasts from the Glass Chips and Seeding onto the Experimental Surface When cells are ready to be used in an experiment, they generally need to be removed from the glass chips still in the Petri dish. This is accomplished by collagenase digestion (see Note 12).

Hillsley 1 Surface preparation The surface onto whtch the cells will be plated generally needs to be pretreated with an attachment factor such as fibronectm Glass mlcroscope slides (75 mm x 38 mm) are cleaned following the same procedure as that used for cleaning the cover shps used m makmg glass chips The cleaned slides are coated with 1 mL of 20 pg/mL fibronectm at least 1 h prtor to seedmg (see Note 13) Just prior to seeding, the slides with the tibronectm solution are rinsed using a buffer such as HBSS, DPBS, or basal medium (see Note 14) 2 Cells are first rinsed with a warmed calcium-free buffer such as HBSS to remove the serum and calcium, which will hinder collagenase activity and cell detachment 3 The cells are incubated wtth 2-3 mL warmed collagenase working solution for 30 mm (see Note 15) After this time, a solid tap to the Petri dish should dislodge a good number of the cells The remainder can be dislodged using media trriganon First, the collagenase IS deactivated from further digestion by addmg 4-6 mL complete growth media (approx 2 times as much growth media as mttlal collagenase solution) The cells are then dislodged from the Petri plate and from each other by vigorous irrigation By trrlgatlon IS meant passing the media up and down a pipet and squntmg it vigorously against the side or bottom of the Petri dish (see Note 16) If the cells are not adequately separated from each other, the seeded surface will result m large clumps of multilayered cells, rather than a smooth monolayer 4 The cell suspension IS next centrifuged for 10 mm at 1OOOg The cell pellet IS then resuspended m the appropriate amount of complete media It IS critical that resuspension results m a suspension of mdivrdual cells (see Note 17) Dislodged cells are plated onto the desired surface m a small amount of media (1 mL cell suspension onto a 75 x 38 mm slide) The platmg density varies and IS dependent on the desned final density and the attachment efficiency In my experience, seeding at 9 x lo4 cells/cm* onto fibronectm-covered glass shdes (75 x 38 mm) produced a confluent monolayer culture 34 d after seeding (I,4) For other experiments, m which a much tighter monolayer was required and where the attachment efficiency was lower, cells were plated onto 4 2 cm* Mtlhpore PCF membrane transwell inserts at a seeding densrty of 2 1 x lo5 cells/cm* (5) (see Note 18). 5 The seeded cells are placed m the Incubator and are allowed to settle and attach to the surface Attachment occurs fairly quickly, so that the cells may be fed a larger amount of complete media from 2 to 24 h after seeding This time to settle and attach in limited media IS only necessary if there are attachment boundaries, such as keepmg cells on the microscope slide instead of attaching to the entire Petri dish If cells are desired over the entire contained area, no attachment time is required before adding additional growth media to the cells

4. Notes 1 Medium 199 was used mainly m our lab However, we have also used mmtmum essential medium (MEM) and other home-made base media successfully Drfferent media should be tested on the cells prior to use

Osteoblasts

299

2 Osteoblasts isolated as descrtbed here grew well m FBS They were less sensmve to serum quality than human umbthcal vem endotheltal cells (HUVEC) Our osteoblasts, however, died when bovine serum albumm (BSA) was used instead of FBS BSA can be used m combmatton with FBS (1% BSA, 2% FBS) when necessary (5), however, cells grow better and faster m 10% FBS 3 Collagenase ~111dtgest Itself, especrally at warmer temperatures Care should be taken to warm the refrigerated aliquot only when necessary, and to return the altquot to the refrigerator as qmckly as possible Repeated freeze-thawmg cycles of collagenase have been shown to be more damaging than stormg the altquot in the refrtgerator 4 Human fibronectm was used m our studies, because rt happened to be the cheapest Other sources of fibronectm have also worked as an attachment factor and could be used after testmg 5 Depending on the quality of the autoclave and steam used, repeated autoclavmg may over time form depostts on the glass Therefore, excessive autoclavmg of glass chips should be avoided 6 The size range of the chips 1s approximate. It was determined based on the steves avatlable The size range may probably be shifted 2&50 pm m either dtrectron, If needed, to accommodate Steve sizes 7 When holding the pup head with the hemostat, be careful not to clamp too hard This may result m cutting through the snout by the end of the procedure Holding tt too loose, however, ~111result m dropping the head and compromtsing sterrhty Some people find rt easrer to rest the neck on the paper towel whtle performing the calvartal dtssectton 8 It takes some force to pull the skm loose It IS best to wedge the skm between the two long sides of the forceps rather than by trymg to use the tips of the forceps Using only the ttps ~111often result m pulling off small pieces of skm rather than the entire square 9. Medium 199 IS used because of convenience. Any buffer should also work, since the calvanas will only be there for a short period whrle the rest of the Inter IS processed 10 Prolonged exposure of the calvarta to any buffer or medium before the removal of the pertosteum softens the calvarta and makes tt more difficult to remove all of the pertosteal layer without rtppmg the calvarta Itself 11 The ttssue-to-medta ratio (or dash size) IS important m the effectrveness of cell mrgratron onto the chips A &cm dtameter Petrt dish with 4 mL media works well for one set of calvartas (from a smgle rat pup) A larger dish (lo-cm drameter) has also been used effectively for 2-3 sets of calvarras with 8-10 mL medta. However, the smaller dishes seem to get slightly better results 12 Trypsm dtgestion 1s commonly used m seedmg cells However, we have found m our work that collagenase IS much less damaging to the cell surface than trypsm For example, osteoblasts seeded using trypsm lost vitamin D, (l,25[OH]2D3) responsrveness (measured by osteocalcm actrvrty in the media) for at least 8 d postseedmg Cells seeded usmg collagenase dtgestron, however, showed no such loss m vitamin D3 responstveness (2).

300

H~llsley

13 Osteoblasts attach better to fibronectm-covered mlcroscope slides than to uncoated shdes We have also seeded osteoblasts onto filters (Mllhpore PCF membrane transwell Inserts) These filters require fibronectm pretreatment as well We again used 1 mL of 20 pg/mL fibronectm per 4 2 cm2 Insert Tissue culture plates do not require pretreatment. 14 When rmsmg the shdes, take care not to spill solution off the slide itself mto the Petri dish This ~111 then provide a path for the cell suspension to follow, resulting m a fair number of the cells plating the Petri dish surface, rather than the shde surface 15 The amount of collagenase used may need to be adJusted, dependmg on the age of the collagenase aliquot Fresh collagenase should Just cover the cell layer and may not reqmre all of the 30 mm Older collagenase may require a larger quantity to be added to the cells and may also reqmre up to 45 mm of exposure If more time or a large quantity IS required, a new ahquot of collagenase should be opened 16 When plpetmg vigorously, take care not to place the plpet directly against the side or bottom of the Petri dish, since this will damage the cells. Instead, hold the plpet l-2 cm away from the surface, and plpet, at high speed, a stream of fluld and cells against this surface This impact will be enough to dislodge and separate cells, but the impact angle and force will not damage the cells slgmficantly 17. A seedmg suspension conslstmg of clumps of cells will cause the cells to grow m isolated, multllayered clumps, rather than a monolayer covering the entlre surface area. 18 Whether the osteoblasts grow into a monolayer or form large clumps of multllayered cells IS very sensltlve to the seeding density (as well as to the effectiveness of cell separation during the passmg procedure, as discussed m Note 17). A seedmg density that IS too high ~111result m osteoblasts growmg m clumps and forming nodules A seedmg density that IS too low will also result m osteoblasts mlgratmg toward each other and forming Isolated clumps of cells When a monolayer or smooth layer of cells IS desired, the cells must be well dispersed m the medium and seeded at the correct seeding density You may need to experiment with the seeding density to get the desired result

References 1 Reich, K M., Gay, C V., and Frangos, J A. (1990) Fluid shear stress as a medlator of osteoblast cychc adenosme monophosphate productlon J Cell Physlol 143, IO&l04 2 Hlllsley, M V (1990) The effects of fluid shear stress and 1,25-dlhydroxy-vltamm D, on collagen and osteocalcm productlon by osteoblasts MS Thesis Pennsylvama State University 3 Hlllsley, M V and Frangos, J A (1997) Alkalme phosphatase m osteoblasts 1s down-regulated by pulsatlle fluid flow C&c Issue Int ,60(l), 48-53 4 Hdlsley, M. V. and Frangos, J A (1994) Bone tissue engineering. the role of mterstltlal fluld flow Bzotechnol Bzoeng 43, 573-58 1 (review)

Osteoblasts

301

5 Hrllsley, M V and Frangos, J A. (1996) Osteoblast hydrauhc conductwty IS regulated by calcltomn and parathyrold hormone J Bone Mzner Res 11, 114-124 6 Ecarot-Chamer, B , Gloneux, H., van der Rest, M , and Perelra, G (1983) Osteoblasts isolated from mouse calvarla mltlate matrix mmerahzation m culture J Cell Blol 96,639-643

24 Cryopreservation of Rat Hepatocytes in a ThreeDimensional Culture Configuration Using a Controlled-Rate Freezing Device Michael J. Russo and Mehmet Toner 1. Introduction From engineered tissues to transfected cell Imes, the long term storage of hvmg blologlcals 1s desirable for a variety of medical, scientific, economic, and regulatory concerns, including transport, the expense of development, repeatability issues, and the point of use. Currently, the best option 1s cryogenic storage, placmg the biomaterlals m suspended ammatlon at very low temperatures (-196”C), halting all chemical reactions, llmltmg genetic drift, and ensuring the maintenance of cell vlabllity and function upon thawing (2). Obtaining such an advantageous state, however, can be a difficult achievement. This problem becomes further comphcated as we move toward next generation multicomponent products such as engineered skm and cartilage substitutes, composed of multiple cell types oriented in complicated three-dnnenslonal geometries wlthm an extracellular matrix scaffold (2-4). In order to survive freezing and warming protocols, cells must overcome extremes m solute concentration, volume, and viscosity, and even the formation of intracellular ice crystals (5-7). These protocols must therefore be carefully designed m a cell specific fashion to mmlmlze the damage that will be inflicted by these mhospltable condltlons There are many parameters that can be varied toward the optlmlzatlon of a cryoblologlcal protocol. the cooling rate, the type and concentration of cryoprotective agents (CPAs, additions made to the cytosol to aid the cell m overcommg freezing stresses, see ref. 8), the temperature at which extracellular ice is seeded, the temperature at which the cells are transferred to lquld mtrogen for storage, and the warming rate (see Note 1). From

Methods

m Molecular

Edited by J R Morgan

Me&one,

Vol

18 Tmue

and M L Yarmush

303

Engmeenng

0 Humana

Methods

and

Press Inc , Totowa,

Pfotocols

NJ

Russo and Toner Here, we outline the construction of a controlled rate freezmg device capable of executing multistep freeze-thaw protocols, with coolmg rates of up to -1 OO”C/mm (for temperatures as low as -8O’C) and warmmg rates m excess of +400”C/mm. This design operates by discharging cooled nitrogen (N) gas (-160°C) mto a channel underlying a heating device (9). The sample platform IS located on top of the heating umt, and the sample temperature IS thereby determined by a precise balancmg between the heat generated by the heating unit and the coolmg power of the nitrogen gas The control system employs a feedback system which can be programmed to follow a variety of coolmg and warming protocols by dynamically determmmg the power output and thus the temperature to the heating unit Though this system IS capable of freezing a variety of cultured and suspended cells, as well as tissue samples, the method described here utilizes the device for the freezing of cultured hepatocytes, sandwlched between two layers of collagen. From a tissue-engineering perspective, this IS an interesting system because it involves the freezing of a multiphase three-dimensional umt composed of extracellular matrix (collagen), hvmg ceils (hepatocytes), and culture media, each with its own transport, mechanical, and freezing characteristics (IO). Medically, hepatocyte cryopreservatlon IS necessary for ensurmg a contmuous supply of cells for use in bloartlfical hepatlc support devices, as well as for cell transplantation purposes (11,12). Previously, most attempts at hepatocyte cryopreservatlon have been performed on cell suspensions, yleldmg low vlablhty and function (13,14). The collagen-sandwich culture system, by partially recreating the hepatocytes’ zn sztu environment, results m longterm stable cultures, with hepatocytes expressing hepatospecific functions, such as albumin secretion and urea synthesis, for over 10 wk. By freezing hepatocytes m this culture configuration after they begin demonstrating phenotyplc stablhty (d 7 of culture), we can demonstrate a stable, full recovery of albumm secretion following a 2-3 d postthaw period (15).

2. Materials 2.1. Hepatocyte Isolation and Culture 2.1.7 Isolation of Hepatocytes from Rat Liver The materials necessary for the lsolatlon are covered in Chapter 33 of this volume.

and punficatlon

of rat hepatocytes

2.1.2 Long Term Hepatocyte Culture In addition to those materials for hepatocyte sandwich cultures listed m Chapter 33, this preservation method requires that the followmg additional materials be used for hepatocyte culture.

Cryopreservation

of Rat Hepatocytes

305 r------------

Power Leads PfeSSUfe

Reaulalor

--!I

I

LN, Dewar

N2 Gas Cylmder

Fig 1 Schematrc overvtew

of the controlled rate freezer umt (9)

1 No 1 cover glasses (24 x 60 mm, Baxter Healthcare, McGaw 2 Srhcone rubber (GE Srhcones, Waterford, NY) 3 Scalpel

2.2. Hepatocyfe

Park, IL)

Freezing

1 Freezing unit A detailed descrtptton of the design and charactertzatton of the freezmg umt IS gtven m ref. 9, and IS shown m Fig. 1 a Nitrogen gas tank with pressure regulator and flow meter b Lrqurd nitrogen c Liquid nitrogen dewar (3-5-m diameter) d Copper coil (3/S m OD, l/8 m ID, at least 10 turns of 3 m diameter) e Tygon tubing (l/2 m OD, 3/8 m ID) f Heating gun or hair dryer g Refrigerant Flow Channel (see Fig. 2 for multilayered design) 1 Bottom channel layer Polypenco acetal polymer rectangular solid (Delrm, DuPont, Wilmington, DE), 33 7 x 9.5 % 2.0 cm3. 2 Top channel layer: polypenco acetal polymer rectangular solid 33 7 x 9 5 x05cm3 3 Tubing connectors (1 -cm inlet diameter). 4 Stltcone RTV (GE Srhcones) 5 Cork,40Ox 12Ox20cm3 6 Styrofoam, 60 0 x 30.0 x 1 0 cm3 h Heater layer (see Fig. 3 for destgn)

306

Russo and Toner TOP VIEW ___------

-__-----

----

__-_-_

\

--

I

-_ --

_I_

-I

---------a________

----_____

4:------

D- L-e--

+I I

15 -

\

\-

--

-

,‘-

I I

\

-

---

/

-

/ -1

A-A CROSS SECTION

efriaerant Oilel

Relrloera

Get

BollorC Layer

Ftg 2 Flow channel schematic, wtth heater layer locatton indicated are m mm (9)

Dtmenstons

Tissue Bag Alummum (300pm) Epoxy (300pm) Heater (~oopm) Epoxy (300pm) Alummum Ooopm) fV2 Stream

Fig 3 Heater layer constructron, with sample bag (9) 1 Two alummum plates, 7 5 x 4 7 cm2 and 300-pm thick 2 Kapton heater layer, 7 5 x 4 7 cm2 and loo-pm thick (230 0, 10 W, HK-TBA, Mmco, Mmneapohs, MN) 3. Copper-constantan for1 thermocouple (C02-T, Omega, Stamford, CT) 4 Sthcone RTV (GE Slhcones) 1 Proporttonal integral controller (Interface Techniques, Cambridge, MA) and IBMTM compattble computer The bmldmg and destgn of a controller IS outsrde the scope of this method Many commerctally avarlable control systems are designed to provide the temperature-reference srgnal and control, and the power amplifier necessary for the prectston temperature control of the freezing unit

Cryopreservation

of Rat Hepa tocytes

307

2 Hepatocyte sample preparation a Coverslip wtth cultured hepatocytes as prepared m Subheading 3.1.2 b Freezmg solution. 2 Mdtmethylsulfoxide (99% pure, Sigma, St Louts, MO) m Dulbecco’s modified Eagle’s medium (DMEM, Grbco-BRL, Gaithersburg, MD) Prepared fresh c Scotchpak pouch sealer and 50-pm thick, heat sealable polyester pouches (Kapak, Mmneapohs, MN) d. Foil thermocouple (SA 1-T, Omega). e Strtpchart recorder (Hewlett Packard, Temperature Module, model 17502A)

3. Method 3.1. Hepatocyte Isolation and Culture 3.1 7 isolation of Hepatocytes from Rat her Hepatocytes were isolated from 2-3-mo-old female Lewis rats by a mod16 catton of the method originally reported by Seglen (16), and covered m detatl m Chapter 33

3 7.2. Long-term Hepatocyte Culture This cell-culture technique, conststmg of two layers of collagen, between which hepatocytes are sandwiched, IS a modlficatlon of that covered m Chapter 33 of this volume. The primary difference IS that here hepatocytes are cultured on coverslips which must then be cut out of the surrounding collagen gel. Prepare the coverslips with an edge of sthcone several mm htgh to prevent the slippage of the collagen sandwich Smce these will be mcluded m the culture system, they must be sterilized prior to use. Place two coverslips m a loo-mm plasttc ttssue-culture dish At least 60 mm prior to cell seedmg, coat the dishes contammg the covershps wtth 3 mL of type 1 collagen (1 11 mg/mL) and place them mto a 37°C mcubator for gelatton. Care should be taken to ensure even coatmg of the coverslips, to prevent collagen from gettmg under the glass, and to keep the glasses away from the edge of the dish To each dish, gently add 9 mL of hepatocyte culture medta contammg 1 x lo6 hepatocytes/mL Distribute the solutton while ptpetmg, and gently tap the dish to completely cover the collagen layer Evenly distribute the cells by gentle shaking and ttltmg of the dish Incubate the dishes for 24 h at 37°C and 10% CO2 Aspirate the supernatant media and unattached cells and evenly apply another 3 mL of collagen over the cell layer. Incubate for 60 mm at 37°C. Add 9 mL of hepatocyte culture media and Incubate for 24 h at 37°C and 10% co, Aspirate off the culture media and carefully cut out the coverslips from the collagen gel and transfer each to its own tissue-culture dash Use only sterthzed

308

Russo and Toner

mstruments and avoid ttltmg the coverslip to prevent gel slippage. Add 9 mL of culture media and incubate 9 Change culture media dally and collect samples for analysts as outlmed m Chapter 33. Cells on d 7-9 of culture are to be used for freezing

3.2. Hepatocyte Freezing This subheading has two parts: the first dealing with the constructlon of the freezing unit and the second addressing hepatocyte freezing techniques 3.2.1. Freezing Unit Thts gives an overvlew for the construction of a controlled-rate freezing device for small samples. Flow Channel This work should be performed by someone expertenced with the machining of plastics and resins Machme the bottom of the flow channel from the Polypenco acetal polymer 33.7 x 9 5 x 2 0 cm3-rectangular solid. Machme a 1 4-cm-deep entrance channel with a l-cm-wide and l-cm-long nntial entrance, followed by a gradual lo-degree widening from 4 to 7 cm wtthm a 21 5-cm length; mamtam 7-cm wtdth for next 8 cm, qutckly narrow channel to the exit port diameter of 1 cm (Fig. 2, see Note 2) Machme the top of the flow channel from a Polypenco acetal polymer 33.7 x 9 5 x 0 5-cm3 rectangular solid Machine 7 5 x 4 7-cm2 openmg, with a 0 5-cm bevel to hold the heater unit, as indicated m Fig. 2. Seal top layer to bottom layer with sthcone RTV (GE Silicones), carefully mamtaming then proper ortentatton. Allow stltcone to cure for 24 h, and check carefully for leaks Securely attach the tubing connectors at the entrance and exit ports with stllcone RTV Heater Layer As shown m Fig. 3, seal a single alummum plate to the top and a single aluminum plate to the bottom of the Kapton heater layer, each with a 300~pm layer of RTV Attach a copper-constantan thermocouple to the top alumtnum layer, with heat conductive epoxy Place the heater layer mto the specially machined openmg in the flow channel Plug the thermocouple and power leads mto the appropriate ports on the control system computer card Place the flow chamber onto the cork msulatton and make a small Styrofoam box to place over the setup Make small holes m the Styrofoam to allow for the inlet and outlet of nitrogen gas and for the thermocouple lead. Coolant Source Half fill the dewar with hqutd nitrogen, and allow the temperature of the dewar to eqmhbrate until botlmg stops Caution: Always be very careful when handling hqutd nitrogen to avoid severe burn injury Wear gloves and protective clothing Attach tubing from the pressure regulator to the mlet end of the copper cot1 Set the regulator at 5-10 psi and the flow meter at 100 cm3/mm Open the nitrogen tank valve and allow mtrogen to run through the copper cot1 for several minutes

Cryopreserva tion of Rat Hepa tocytes

9

10 11

12

309

It IS essentral that this purge of the cot1 and tubmg be done to avoid the accumulation of condensed water and, hence, me, when the cot1 IS placed into the dewar. Such an me blockage will limit the performance of your cooling system, and could prove dangerous tf pressure butlds withm the tubing or cot1 Place the copper cot1 into the dewar containing liqutd nitrogen Caution: the ltqurd nitrogen will splatter as it bolls. After the nitrogen has stopped borlmg, fill the remainder of the dewar with liquid mtrogen. Never turn off the mtrogen gas while the co11 1s m the dewar; always mamtam a mnumum flow rate of 100 cm3/mm Program the control system with the appropriate freezing protocol, startmg at ambient temperature. Attach tubing from the outlet of the copper coil to the refrigerant mlet port of the flow channel Allow the flow channel to equilibrate with the cooled mtrogen gas (5 mm), to reduce thermal gradients that will introduce large errors and ltmtt the control systems ability to properly execute the freezing protocol The system 1snow ready to freeze the sample placed on its heatmg umt. During a freezing protocol mamtam a high liquid nitrogen level m the dewar

3.2.2. Freezmg of Hepatocytes in Sandwich Culture (Fig. 4) 1 On d 7-9 of culture, aspirate the media from the dish and replace tt wtth 9 mL DMEM contamrng 2 M DMSO. Allow these dishes to stt m ambtent condtttons for 1 h After this mcubatton time, place a single coverslip and 3 mL of DMEM with 2 M DMSO mto a polyester pouch and seal the pouch with the ScotchPac pouch sealer Remove as much au as possible from the pouch and muumtze Its stze 2 Place a foil thermocouple (attached to a strtp chart reader) to the top of the pouch for the momtormg of vertical temperature gradients. 3. Place the sample onto the heater umt and use a weighted piece of Styrofoam to umformly press on the pouch, thus ensuring maximum thermal contact 4 Increase the flow meter to 5000 cm3/mm, and start the freezing protocol 5 Freezing Protocol A typical freezmg protocol IS shown m Fig. 5. The method outlmed here IS the protocol that best preserved hepatocytes under the condmons tested (15, see Note 3) Cool the pouch from 20” to -12°C at -lO”C/mm. Hold at -12°C for 3 mm to allow for the pouch to reach thermal equtlibratton Turn the heatmg unit off, allowmg for a rapid drop m temperature and the spontaneous seeding of extracellular Ice (at about -28’C). Upon the latent heat release mdtcated by an abrupt rise m temperature, turn the heater back on, and hold the temperature at -12°C for 15 mm to permit adequate cell dehydration and complete formatton of Ice Decrease the temperature from -12°C to -40°C at -lO’C/ mm These cells are now ready for storage at -196°C m hqutd nitrogen 6 Warming Warming 1s best accomplished by qutckly ptckmg up the pouch with precooled forceps and nnmersmg it m a 37°C sterile water bath. 7 Remove the covershp from the pouch and transfer tt to a tissue-culture dish Add 9 mL of 37°C culture media and incubate for 30 mm. Aspirate off media and replace tt with fresh medta to remove the DMSO, whtch has dtffused out of the cells.

Russoand

Toner

Day 0 . Day 1 37°C 10% co2 Day 2. Day 7

L

/Add xlhr

h(g)

-

Stir tn 37-C Hz0

2 OM Mq SO at22’C

bath

Day7‘\,

2xWash

un 37”C, 10% COz

Day 7 . Day 22

Fig 4 Outline of hepatocyte cryopreservatlon method showmg culture on coverslips, samplepreparation, freezmg, and thawmg (15)

2oL

2 4

1

OS

_

h -2o-

-4o-

-I

I

I

I

0

10

20

30

4b-----

50

MINUTES Fig 5 Typlcal freeze thaw protocol with temperature asa function of time (15) (l2) Cooling at -1 O”C/mm to -12°C (2-3) Eqmhbratlon at -12°C (3) Power off (3-4) Temperature decreaseuntil Ice IS seededat 4. (4-5) Latent heat release and power back on (5-6) Equilibrate at -12°C (6-7) Cool at prescribed coolmg rate to final temperature. (7-8) Hold at final temperature. (8-9) Thaw at various rates

Cryopreservation

of Rat Hepatocytes

311

8 Lower the N flow to purge levels and remove the cot1 from the hqutd nitrogen Defrost the system with the hair dryer until the cot1 and the tubing are at room temperature Turn off N gas flow 9 Charactertzatton of Cultured Hepatocytes Cultures should be followed for 2 wk with media changes and samples taken dally. Methods to characterize the mtegrtty and functton of the postthaw hepatocytes are covered m Chapter 33 4.

Notes

1 The effect of freezing parameters can be elucidated m the context of the two hit hypothesis of cell damage Upon the coolmg of cells, ice wdl preferentially form outside of the cells m the extracellular solutton. As the me IS formed m the extracellular solutton, the remaining hqutd fraction will become mcreasmgly concentrated, mducmg the cells to dehydrate against the imposed osmottc gradient If the coolmg rate 1s very rapid, a large amount of mtracellular water will remam, as cells pass through low temperatures, resulting m lethal intracellular ice formation If, however, the coolmg rate IS very slow, greater water efflux will occur, resultmg m extreme dehydratron and deletertously high solute concentrattons One goal, therefore, IS to find an appropriate intermediate coolmg rate to maxrmaze cell survival The choosing of this coolmg rate, however, IS somewhat conflated by the ice-seeding temperature, the temperature at which cellular dehydratmn begins Generally, the higher the seeding temperature, the greater the relative degree of cell dehydration caused by a higher level of water energettcs For most cell types, the optimal coolmg protocol still results m a low yield Therefore, tmprovements can be made by cryoprotecttve agents, which are added to Increase mtracellular vrscostty, depress the formation of mtracellular me, and reduce the degree of water efflux. Srmrlarly, because of damaging osmotic fluxes and ice recrystalltzatlon encountered during thawing, the warming protocol must ltkewrse be optlmtzed 2 The flow channel IS machined as described, because the gradual increase m channel width will reduce backflow and flow separations, thus mcreasmg channel efficiency This freezing apparatus can easily be scaled to efficiently accommodate other sample sizes Srmllarly, other covershp sizes can be used to reduce sample size and thermal mass, If necessary. Further, other gasses could be used as the coolant gas, though the channel may need a different design to accommodate the variation m flow Helium, for example, can be used to achieve much higher cooling rates (-500 to -1OOO”C/mm) 3. We explored several different coolmg rates (4 25, -1 0, -5 0, -10.0, -16.0, and -35 O”C/mm), final freezing temperatures (-4O’C and-80”C), and warming rates (+5, + 150, and +400”C/mm), with varying results (25) As evaluated by albumin secretion, cells performed best when cooled between -5 and -1 O”C/mm When cooled to -4O”C, full postthaw recovery was independent of warmmg rate, but cells cooled to -80°C required rapid thawing to attam even 75% functional recovery The extracellular DMSO concentratton can be varied to result m other intracellular DMSO concentratrons followmg uptake durmg the mcubatlon period Here, 2 M

312

Russo and Toner DMSO m DMEM results m 1 33 M mtracellular DMSO after 1 h of mcubatlon. Similarly, 0 5, 1 0, and 5 0 Mextracellular DMSO results m 0 33,O 66, and 3.33 A4 mtracellular DMSO, respectively, after 1 h, with 3 33 A4 causing irreversible toxic damage (17). Likewise, various other CPAs (ethylene glycol, propanediol, glycerol, and so on) can be used Finally, since this freezer design allows easy access to the sample, one could seed extracellular ice by touching the bag with a spatula or forceps previously cooled m liquid nitrogen

References 1. Coger, R and Toner, M. ( 1995) Preservation techniques for btomatenals tn The Bzomedlcal Engzneerzng Handbook (Bronzmo, J D., ed ) CRC, Boca Raton, pp 1557-1567 2 Karlsson, J 0 M and Toner, M (1996) Long-term storage of tissues by cryopreservation critical issues Bzomaterzals 17,243-256 3 Morgan, J and Yarmush, M. (1997) Bloengineered skm subsmutes Scz Med 4,5-l 5 4 Langer, R. and Vtcantt, J P (1993) Ttssue engineering Sczence 260,92&926 5 Mazur, P (1984) Freezmg of hvmg cells mechanisms and implications Am J Physzol 143, Cl255Cl42 6 Karlsson, J 0 M , Cravalho, E G , and Toner, M (1993) Intracellular ice formation* causes and consequences Cryo-Lett 14,323-334 7 Schwartz, G J and Duller, K R (1983) Osmotic response of mdtvldual cells during freezing Cryobzology 20, 6 l-77 8 Arakawa, T , Carpenter, J F , Lita, Y A , and Crowe, J H (1990) The basis of toxicity of certain cryoprotectants a hypothesis Cryobzology 27,40 14 15 9 Toner, M and Borel-Rinkes, I H M (1993) A controlled rate freezing device for cryopreservatlon of biological tissues. Cryo-Lett 7,43-56 10 Dunn, J C Y, Tompkms, R G , and Yarmush, M L (199 1) Long-term m vitro function of adult hepatocytes m a collagen sandwich configuration Blotechnol Prog 7,237-245 11 Yarmush, M L , Dunn, J C , and Tompkins, R G (1992) Assessment of artifictal liver support technology Cell Transplant 1, 323-341. 12 Sussman, N L , Chang, M G , Koussayer, T, He, D , Shong, T , Whisennand, H H , and Kelly, J H (1992) Reversal of fulmmant hepatic failure using an extracorporeal liver assist device Hepatology 16, 60-65 13 Chesne, C. and Gu~llouzo, A (1988) Cryopreservatlon of isolated rat hepatocytes a critical evaluation of freezing and thawing conditions Cryobzology 25,323-330 14 Fuller, B J , Morris, G J , Nutt, L H , and Attenburrow, V. D (1980) Functional recovery of isolated rat hepatocytes upon thawing from-l 96°C Cryo-Lett 1,139146 15 Borel-Rmkes, I H M , Toner, M , Ezzell, R. M , Tompkms, R G , andyarmush, M L. (1992) Long-term functional recovery of hepatocytes after cryopreservation m a three-drmensional culture configuration Cell Transplant 1,281-292 16 Selgen, P 0 (1976) Preparation of isolated rat liver cells Methods Bzol 13,29-83 17 Borel-Rmkes, I H M , Toner, M , Ezzell, R M , Tompkins, R G , andyarmush, M L. (1992) Effects of dtmethyl sulfoxlde on cultured rat hepatocytes m a sandwich configuration Cryobzology 29,443%453

25 Microencapsulation of Enzymes, Cells, and Genetically Engineered Microorganisms Thomas M. S. Chang 1. Introduction Microencapsulation of biologically active material m the form of artificial cell was reported as early as 1964 (l-4). However, it is only in the past 10 yr that many centers have extensively developed this (5). More recently, we have concentrated on three areas of artificial cells for blood substitutes, enzyme therapy, and cell therapy Space allows only a few examples to be given here HIV has stimulated extensive development in the past 10 yr. The early idea of crosslmked hemoglobin (1,4) has been developed as first generation blood substitute, now m phase III clmlcal trials by a number of groups (6,7) Secondgeneration blood substitutes Include the microencapsulation of hemoglobin m lipid vesicles. We are now developing a further generation of blood substitutes based on the use of nanotechnology to prepare 150~nmdiameter, biodegradable, polymeric membrane nanocapsules containing hemoglobm and enzymes (8,9). InJection of enzyme artificial cells is effective m enzyme therapy for inborn errors of metabolism (IO), and for cancer therapy (11). However, accumulation of the implanted artificial cells has been a problem. We have solved this problem recently by giving the enzyme artificial cells orally, as m the following example. Lesch-Nyhan disease is a very rare inborn error of metabolism, with accumulation of hypoxanthme. Being lipid-soluble, it can diffuse rapidly mto the intestine. By giving the enzyme microcapsules orally to act m the mtestme we can avoid the need for mJection. Dally oral admmistration of microencapsulated xanthme oxidase resulted in the lowering of hypoxanthine m the plasma and cerebral spinal fluid m a patient with Lesch-Nyhan disease (11,12). Phenylketonuria (PKU) is the most common inborn error of metabolism m humans with enzyme defects resulting m the elevation of the amino acid pheFrom

Methods Edlted

in Molecular

by J R Morgan

Medmne, and

Vol 18 Tmue

M L Yarmush

315

Engmeenng

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

316

Chang

nylalanine. However, amino actds from the body do not diffuse rapidly into the mtestme. We can now solve this problem, based on our new findings of extensive enterorecnculatton of ammo acids. We showed that there IS an extensive enterorecirculatron of amino acids between the body and the intestine. Large volumes of drgesttve Jurce containing enzymes and other proteins are secreted mto the mtestme These are digested by tryptrc enzymes m the mtestme mto ammo acids. The ammo acids formed in this way are reabsorbed mto the body. We showed for the first time that ammo acids formed from this source are higher m amount than those resulting from protein m Ingested food. This means that when we give orally mtcrocapsules contammg a specific enzyme, tt can remove the correspondmg specific amino acid in the intestine and therefore prevent tt from returnmg to the body. The net result 1s a depletton of this spectfic ammo acid from the body. For example, mtcroencapsulated phenylalanine ammonia lyase given orally once a day can selectively remove phenylalanme from the enterorecirculatmg amino acids This explains our earher observatton of the effectiveness of this approach for lowering the elevated systemic phenylalanme m the PKU rats (Id), and more recently m the ENU2 phenylketonuric mice (17). This approach can also be used to remove any of the other 26 ammo acids m the body, with potential for treating other mbom errors of metabolisms and for ammo acid dependent tumors This 1s now being developed for clmtcal trials. This has resulted m renewed interest in the mrcroencapsulation of enzymes. We have also studied mtcrocapsules contammg multrenzyme systems with cofactor recyclmg for multistep enzyme conversions (I&19) The third major area of our present interest IS based on the mrcroencapsulanon of cells and microorganisms. As early as 1965, this author successfully encapsulated cells and wrote (20). microencapsulation of Intact cells or tissue fragments the enclosed material mrght be protected from destructron and from particlpatlon m lmmunologlcal processes, whtle the enclosmg membrane would be permeable to small molecules of specrtic cellular product whrch could then enter the general extracellular compartment of the reclplent. For instance, encapsulated endocrme cells might survive and mamtam an effect supply of hormone ,. The srtuatlon would then be comparable to that of a graft placed m an lmmunologlcally favourable site ” “There would be the further advantage that implantation could be accomphshed by a simple Injection procedure rather than by a surgical operation ” “Mlcroencapsulatlon of intact cells . The erythrocytes were suspended m hemolysate rather than m the dlamme solution; and a slllcone 011[Dow Commg 200 flmd] was substituted for the stock organic liquid The mlcroencapsulatlon was then carried out by the prmclple of mterfaclal polymerlzatlon for membranes of crosslinked proteins A large number of human erythrocytes suspended m hemolysate within a microcapsule of about 500 p diameter was prepared by the syrmge [drop] method ”

Microencapsulation

317

The author also described this in other publtcations (2,4). However, it is only with the more recent interest m biotechnology that many groups around the world have extended this approach of cell encapsulation (5,21). Some of our recent research mterest 1sas follows. We studied the use of encapsulated hepatocytes as a model for cell and gene therapy Implantatton of encapsulated hepatocytes Increases the survival of fulmmant hepatic-failure rats (22). Encapsulated rat hepatocytes are not rejected after bemg implanted into mice (23). Instead, there is an increase in vtabtlny after intraperitoneal implantatton (23). Thts is a result of the retention of hepatostimulating factors inside the microcapsules as they are secreted by the hepatocytes (24). Gunn rat 1sthe model for the Crlgkr-NaJJar syndrome in humans caused by defects of the liver enzyme UDP-glucuronosyltransferase (UDPGT). Intraperttoneal implantation of artificial cells containing hepatocytes lowered the high systemic btlirubm levels (25,26). Kinetic analysis shows that this is because of the hepatocytic UDPGT in arttfkial cells that conjugated bilirubin to the monoconjugated and dtconjugated form for excretron m the urine as m normal animals (27). For encapsulation of high concentratton of small cells we have to develop a two-step method for cell encapsulation m artificial cells. This is because we observe that some cells are exposed on the surface of artificial cells prepared by the standard method (28) We have therefore developed a new method to prevent this problem (28,291. This will be parttcularly useful for the encapsulation of hepatocytes and genetically engineered cells Another area is the study of arttficial cells contaimng nonpathogemc genetttally engineered microorgamsms. However, they cannot be injected. We are studying the possibility of using oral mtcrocapsules containing these cells We started with basic research using Escherzchia coli with K aerogenes gene for urea removal (30). For more than 30 yr mvestigators have been unable to find an oral treatment for uremia, mainly because of the inability to remove the large amount of urea. We have just found a new approach that combines arttficial cells with gene expression genetic engineering technology (30). Oral admmistration of a small amount of artificial cells containing genetically engtneered E coli DHS cells once a day resulted m the decrease to normal of high urea level m uremic rats to normal (31). This was maintained during the 2 1 d of treatment. Uremic rats survived during thts period when treated with thts approach. On the other hand, 50% of untreated control rats died (31) We have also studied the microencapsulation of erythropoietin (EPO)secretmg renal cells (32), and we microencapsulated two other microorganisms, one that removed cholesterol (33), another that converted substrates to L-DOPA (34).

Chang

318

2. Materials 2. I. Microencapsulation of Enzymes 2.1.1. Cellulose Nitrate Membrane Microcapsules. 1 Hemoglobm solution contammg enzymes. Fifteen grams of hemoglobin (bovine hemoglobm type 1,2X crystalhzed, dialyzed, and lyophlllzed Sigma, St. LOUIS, MO) was dissolved m 100 mL of distilled water and filtered (Whatman No 42) 2 Water-saturated ether Analytical grade ether IS shaken with dHzO m a separating funnel, then left standing for the two phases to separate, so that the water can be discarded 3 Cellulose-nitrate solution This solution 1sprepared by spreading 100 mL of USP collodlon (Fisher Scientific, Montreal) m a evaporating disk m a well-ventilated hood overnight This allowed the complete evaporation of its organic solvents, leaving a dry, thm sheet The thm sheet of polymer IS then cut into small pieces and dissolved m a 100-mL mixture containing 82 5 mL analytical grade ether and 17 5 mL analytical grade absolute alcohol 4 Tween-20 (atlas powder) solution The 50% (v/v) concentration solution 1s prepared by mixing equal volumes of Tween-20 and dH*O, then adJusting the pH to 7 0 The 1% (v/v) concentration solution was prepared by mixing 1% of Tween20 to the buffer solution used as the suspending media for the final mlcroencapsulated enzyme system.

2.1.2. Polyamide Membrane M/crocapsules 2.1 2.1 Two SOLUTIONS 1 The terepthaloyl organic solution is prepared Just before use 100 mg of terephthaloyl chloride (ICN K+K) IS added to 30 mL organic solution (chloroform cyclohexane, 1.4) kept m an ice bath This was covered and stirred with a magnehc stirrer for 4 h and then filtered with Whatman no 7 paper 2 Dlamme-polyethylenelmme solution is preparedJust before use 0 378 g NaHCOj and 0 464 g 1 6-hexadlamme (J T. Baker) IS dissolved m 5 mL dH20 that contams the material to be encapsulated Then the pH IS adJusted to 9 0 Two mL 50% polyethylenelmme (ICN) IS then added to the dlamme solution and the pH readjusted to 9 0 and the final volume was made up to 10 mL with dHzO

2 1.2 2. 10 ~A00

ML HEMOGLOBIN SOLUTION

Prepared as described above for cellulose nitrate mlcrocapsules. Material to be encapsulated was dissolved m 5 mL of the hemoglobm solution, Instead of Hz0 as described above. The final pH IS adjusted to 9 0. 2.2. Materials for the Encapsulation of Cells and Microorganisms 2.2 1. Mater/a/s for the Standard Method 1 Hepatocytes Rat liver cells were isolated from 125-150 g male Wlstar rats from Charles River Breedmg, Montreal, QC, Canada,

Microencapsula tion

319

2 Genetically engmeered E colz* From nonpathogemc bacteria (E co/z DH5) 3 Reagents for encapsulation* Low-vlscostty sodmm algmate (Keltone LV), mol wt 12,00&80,000, was obtained from Kelco, Clark, NJ The reagents poly-Llysme mol wt 15,000-30,000, type IV collagenase; and type I-S trypsm mhtbttor were purchased from Sigma HEPES (4-[2-hydroxyethyll-1-piperazme ethane sulfomc acid) buffer was purchased from Boehrmger Mannhetm (Montreal) Analytical grade reagents included sodium chloride, sodium hydroxide, calcium chloride dehydrate, trtsodium citrate dihydrate, and d-fructose William’s E medmm (Gtbco-BRL; Burlmgton, On), supplemented with streptomycm and pemctllm (Gtbco-BRL), was used for hepatocyte.

2.2.2 Two-Step Method for High Concentration or Microorganisms

of Cells

Two differently sized droplet generators were used for mlcroencapsulatlon. 1 The first droplet generator conststed of 2 co-axially arranged Jets’ the central Jet consisted of a 26-gage stainless steel needle (Perfektum, Popper, New Hyde Park, NY), and a 16-gage surrounding an Jet, through which the sample and au was passed, respectively To prevent the extruding sample from occludmg the outlet of the surrounding airlet, the tip of the sample Jet was constructed so that the ttp proJected 0 5 mm beyond the end of the air Jet 2 The second droplet generator was a larger and a slightly modified variant of the first droplet generator described above. The second droplet generator was constructed with a 13-gage sample Jet, and a g-gage surrounding an Jet The ends of the lets were cut flush to each other A 1 7 x 1 l-mm PTFE capillary tube (Pharmacia, Montreal, PQ) was inserted mto the sample Jet until it protruded approx 15 mm from the outlet of the sample Jet. The end of the capillary tubing was tapered to facilitate shearmg by the flow of passing air from the air Jet The capillary tubing was approx 3.2 m m length, and had the capacity of be filled with microspheres suspended m 2 5 mL of sodium algmate

2.2.3. Materials for Preparation of Macroporous Microencapsules 1. Mtcroorgamsm* Pseudomonaspzctorum (ATCC #23328) was used because of its ability to degrade cholesterol It was cultured first m nutrient broth (Difco) at 25°C followed by harvesting and resuspension m a cholesterol medium After culturing this suspenston for 15 d at 25°C it was used as an inoculum for btomass production The culture was grown m bovine calf serum (Sigma) at 37°C for 36 11, and then harvested This was used to prepare bacterial suspensions for mnnobtllzation The concentration was about 0.4 mg of dry cell/mL 2. Inoculum medmm Bovine calf serum was from Sigma It was used m all experiments unless otherwise specified. Nutrient agar plates were prepared by dissolvmg 8 g of nutrrent agar (Difco) m 100 mL of water The solution was autoclaved for 15 mm at 12 1“C, allowed to cool at 50°C and poured mto plastic Petri dishes (Fisher Screntific) The plates were stored at 4°C for up to 2 mo.

Chang 3 Cholesterol medium* The composition of the medium was ammonium mtrate (0 l%), potassmm phosphate (0 025%), magnesium sulfate (0 025%), ferric sulfate (0 OOOl%), yeast extract (0 5%), and cholesterol (0 l%), all dissolved m water, with pH adJusted to 7 0 and autoclaved for 15 mm

3. Methods 3.1. Microencapsulation of Enzymes 3.7.1 Cellulose N&ate Membrane Mrcrocapsules Cellulose nitrate membrane mtcrocapsules are prepared using an updated procedure based on earlier pubhcattons (l-4,18) 1 Enzymes and other materials to be microencapsulated were dissolved or suspended m 2 5 mL of the hemoglobm solution The final pH was adJusted to 8 5 with Tris buffer, and hemoglobm concentration adJusted to 10 g/dL 2 2 5 mL of this solution was added to a 150-mL glass beaker, and 25 mL of water saturated ether was added 3 The mixture was unmediately stirred with a Fisher Jambo magnetic stirrer at 1200 i-pm (setting of 5) for 5 s 4 Whtle stirring was continued, 25 mL of a cellulose-nitrate solutton was added Stirring was continued for another 60 s. 5 The beaker was covered and allowed to stand unstirred at 4°C for 45 mm 6 The supernatant was decanted and 30 mL of n-butyl benzoate added The mixture was stirred for 30 s at the same magnetic stirrer setting 7 The beaker was allowed to stand uncovered and unstirred at 4°C for 30 mm Then the butyl benzoate was completely removed after centrifugation at 350g for 5 mm 8 Twenty-five mL of the Tween solutton at 50% (v/v) concentration, pH 7 0 was added Stirring was started at a setting of 10 for 30 s 25 mL of water was added, and sttrrmg continued at a setting of 5 for 30 s, then 200 mL of water added 9 The supematant was removed and the microcapsules were washed 3X with 200 mL of a 1% Tween-20 solution, pH 7.0. The microcapsules were then suspended m a suitable buffer In properly prepared microcapsules, there should not be leakage of hemoglobm after the preparation

3.1.2. Preparation of Polyamide Membrane Microcapsules by In terfacial Polymerization Polyamide membrane microcapsules of 100 pm mean diameter pared using an updated method based on earlter methods (1-4,18).

were pre-

1 Enzyme is added to 2 5 mL of the hemoglobm solution with pH and concentrations adJusted as described in Subheading 2.1.1., item 1. 2 2 5 mL of the diammepolyethylenelmlne solution was mixed for 10 s m a 150-mL beaker placed m an ice bath

Microencapsulatlon

321

3 25 mL of an 0 5% (v/v) Span 85 (Atlas Powder) organic solution (chloroform: cyclohexane, 1 4) was added and stirred in the Fisher Jambo magnetic stirrer at speed setting of 2 5 for 60 s 4 25 mL of the terephthaloyl-ehloride solution prepared earlier was added, and the reaction was allowed to proceed for 3 min with the same stirring speed 5 The supernatant was discarded and another 25 mL of the terephthaloyl-chloride solution was added 6 The reaction wascarned out with strmng for another 3 nun The supematant wasdiscarded 7 Then 50 mL of the 0.5% Span 85 chloroforrn-cyclohexane solutron was added and stirred for 30 s The supernatant was drscarded 8 After this, the procedure of the use of Tween-20 as described for cellulose-nitrate microcapsules was used here for the transfer of the microcapsules into the buffer solution

3.1.3. Lip&Polymer

Membrane Microcapsules that Retain Cofactors

Cofactors covalently linked to macromolecules like dextran or polyethylenetme can be retamed wtthm semrpermeable microcapsules to be recycled enzymatically. However, lmkage of cofactors to macromolecules increases steric hmdrance and reduces their rate of reactions with enzymes. In btological cells like erythrocytes, free cofactors and multienzyme systems are all retained wtthm the cells m free solution. Thus studies were carried out here to tmmobtlize free cofactors mstde mtcrocapsules, with membranes impermeable to cofactors, but permeable to the mittal substrates. This way, the free cofactor can function without steric hindrance m close proximity to the enzymes Furthermore, all enzymes and cofactors inside the microcapsules are m free solutton Lipid-polyamide membrane mtcrocapsules have been prepared (4). These are permeable to ltpophrlic molecules, but with little or no permeability to hydrophilic molecules as small as K+ and Na+ Lipid-polyamide microcapsules of 100 pm mean diameter, containing multienzyme systems, cofactors and a-ketoglutarate, were prepared (35). The first part is similar to the procedure described earlier m this chapter under basic procedure for the preparation of polyamide microcapsules. 1 To 2 0 mL of the hemoglobin solution was added glutamic dehydrogenase, 12 5 mg (bovine liver, type III, 40 U/mg, Sigma); alcohol dehydrogenase, 6 25 mg (yeast, 330 U/mg, Sigma); Urease, 0 5 mg (51 U/mg Mrllipore), ADP, 1 18 mg; and either NAD+ (0 52, 105, 2 11, or 21.13 mg) or NADH (21 13 mg) dissolved m 0.25 mL water Finally, 0 25 mL a-ketoglutarate (56 5 mg), MgC12 (2 5 mg), KC1 (0 93 mg) were added 2 2 5 mL of the hemoglobin-enzyme solution so prepared was added to 2 5 mL of the diamine-polyethyleneime solution. 3 The remaining steps were the same as described above, except that the Tween-20 steps were omitted here Instead, after washing with the 0 5% Span 85 organic

322

Chang

solutton, as described, the followmg steps were carrted out to apply the ltptds to the polyamide membranes 4 The mrcrocapsules were rinsed twice with 10 mL of a llptd-orgamc hqurd This was prepared earlier as follows: 1 4 g lectthm and 0 86 g cholesterol were added to 100 mL tetradecane and s&red for 4 h at room temperature If a more permeable lipid membrane is required to allow urea to diffuse across, then the lrprd compostttons should be 0 43 g cholesterol and 0 7 g lectthm 5 Then another 10 mL of the ltptd-orgamc hqutd was added and the suspension was slowly rotated for 1 h at 4°C on a multipurpose rotator 6 After this, the supernatant was decanted and the ltpld-polyamide membrane mtcrocapsules were recovered and left m this form at 4°C without being suspended m aqueous solutton until tt was added to the substrate solutionJust before the reaction The procedure takes practice and the microcapsules prepared must be tested for the absence of leakage of enzymes or cofactors before being used in expertmental studies

3.2. Microencapsulation of Cells and Microorganisms 3.2 1. Standard Method (Fig. I) 3 2 1 1. PREPARATION OF RAT HEPATOCYTES 1 Each rat was anesthetized with sodtum pentobarbttal, and cannulated via the portal vein 2 The thoractc vena cava was cut and the liver was perfused with a calcmm-free perfuston buffer (142 mMNaC1,6 7 n-&4 KCl, 10 mM HEPES, pH 7 4) for 10 mm at 40 mL/mm 3 Afterwards, the liver was perfused with a collagenase perfusion buffer (67 mA4 NaCl, 6 7 mM KCl, 100 n-J4 HEPES, 5 mA4 CaC12, 0 05% collagenase, pH 7 5) for an addmonal 15 mm at 25 mL/mm 4 The liver was then excised, placed m Willtams E medmm supplemented with 100 ug/mL streptomycin and pemcrllm, and gently shaken to free loose liver cells from the liver tissue 5 The cells were collected, filtered through a 74-pm nylon monofilament mesh (Clstron, Elmford, NY), and centrifuged to remove connective tissue debris, cell clumps, nonparenchymal cells, and damaged cells 6 Isolated hepatocytes were prepared for encapsulation by first washing and suspending the cells with buffered salme (0 85% NaCl, 10 m1!4 HEPES, 20 mA4 o-fructose, pH 7 4) 7. The cells were then mixed wtth a 4% stock solutton of sodmm algmate (4% sodium algmate, 0 45% NaCl), to make a cell suspension conststmg of 20 x 1O6 cells/ml of 2% sodmm algmate 3 2 1.2 GENETICALLY ENGINEERED E cm DH5 CELLS 1 E colz DH5 cells were grown m L B medmm 2 Log phase bacterial cells were harvested by centrtfugmg at 10,OOOgfor 20 mm at 4 0°C The supematant was discarded

Microencapsulation

AGLINATE GEL

SPHERE

323

FORM

MEMBRANE

DISSOLVE

GEL

Fig 1 Standard method for btoencapsulatton of cells and mtcroorgamsms Step 1 formatton of gelled algmate spheres contaming cells Some cells are present on the surface of the sphere Step 2, formatton of membrane around gelled algmate spheres Those cells on the surface of the spheres would be entrapped mto the membrane. Step 3 dtssolutton of content of spheres to form mtcrocapsules with hqutd contents Membrane would be weakened by cells entrapped or exposed on the membrane When used to encapsulate a htgh concentration of small cells or mmroorgamsms, some cells may become entrapped m the membrane matrrx. This can result m membrane weakness or perforatmn Reproduced with permtsston from ref 28 3 The cell mass was then washed and centrtfuged at 10,OOOg for 10 mm at 40°C for five times with sterrle cold water to remove medta components 4 Bacterral cells were suspended m an autoclaved sodmm algmate m ice-cold 0 90% sodium chloride solutton 3 2 1 3 ENCAPSULATION

USING THE STANDARD METHOD

1 Hepatocytes or bacterial cells were suspendedm an autoclaved sodium algmate m me-cold 0 90% sodmmchloride solutron 2 The vtscous algmate suspenstonwas pressedthrough a 23-gage stamlesssteel needleusmg a syrmge pump (Compact mfusion pump model-975 Harvard Apparatus, M111,MA) Sterile compressedan, through a 16-gagecoaxtal stamlesssteel needle,was usedto shearthe droplets coming out of the tip of the 23-gageneedle 3 The droplets were allowed to gel for 15 mm m a gently stirred, heat-sterthzed and me-cold solutton of calcmm chlortde (1.4%, pH 7 2) Upon contact wtth the calctum chloride buffer, algmate gelatton 1sImmediate 4. After gelatron m the calcmm chloride solution, algmate gel beadswere reacted with polylysme (PLL), mol wt 16,100(0.05% in HEPES buffer saline,pH 7 2) for 10mm The posmvely chargedPLL forms a complex of semtpermeablemembrane. 5 The beadswere then washed with HEPES (pH 7 2) and coated wtth an algmate solution (0.1%) for 4 mm 6 The algmate-poly+lysine-algmatecapsulessoformedwerethenwashedm a 3% cm-ate bath (3% m 1.1HEPES-buffersalme,pH 7.2) to liquefy the gel m the mtcrocapsules

324

Chang

7 The APA microcapsules formed, which contams entrapped hepatocytes or genetreally engineered bacteria E colt, were stored at 40°C and used for experiments The conditions were kept sterile during the process of microencapsulation

3.2.2. Two-Step Method for High Concentration or Microorganisms (Fig. 2)

of Cells

The standard method described above is not optimal for encapsulating high concentrations of cells or mtcroorganrsms. As shown schematrcally in Fig. 1, cells or mrcroorgamsms may be trapped m the membrane matrix This can weaken the membrane. If cells are exposed to the surface, this may also resulted m loss of tmmunotsolation and rejectton As a result, we have developed a twostep method that prevented this problem (28,29). 1 First, the hepatocytes suspended m sodium algmate were entrapped wtthm solid calcium algmate microspheres This was done by filling a 5 0-mL syrmge (Becton Dickinson, Rutherford, NJ) with the cell suspension, and extruding the sample with a syringe mfusion pump (Harvard Apparatus) through the sample jet of the first droplet generator The droplets formed at the end of the sample Jet were allowed to fall dropwlse mto a Pyrex dish (125 65 mm) contammg 300 mL 100 mA4 CaCl* (100 mM CaQ, 10 mM HEPES, 20 mM o-fructose, pH 7 4) Every 5 mm, the cells m the syringe were resuspended by gentle mverslon of the syringe, to mimmtze the effect of cells sedimentmg m the alginate solutron The air-flow and mfuslon rates through the droplet generator were 2.0-3 0 L/mm and 0 28-O 39 mL/min respectively; and the clearance height between the end of the sample Jet and the surface of the calcium solution was set at approx 20 cm. A strainer cup was fitted mslde the dash to collect the droplets, and to facllltate the removal of the formed microspheres 2 The mrcrospheres were allowed to cure for approx 15 mm, after whtch they were removed and temporarily stored rn Hanks’ balanced salt solution (Glbco-BRL) supplemented with 1O%, 100 mM CaC12. 3 In the second step, 1 0 mL of formed microspheres were collected and washed three ttmes with buffered saline (0.85% NaCl, 10 mMHEPES, 20 mMD-fructose, pH 7 4) 4 The final salme washing was aspirated, and 1 0 mL of 1.2-l .6% sodium alginate was added to the 1 0 mL of washed microspheres The sodium algmate was prepared by diluting the 4% stock solution with buffered salme. With a 5-mL syringe, the length of the PTFE capillary tubing was filled with the sodium algmate and suspenston of microspheres The tapered end of the capillary tubmg was inserted through the top of the sample Jet of the second droplet generator until the tip of the tubmg extended approx 15 mm beyond the end of the sample jet The an flow and extruston rates through the modified droplet generator were 7.g9.0 L/mm and 0.28-O-39 mL/mm, respectively The tip of the capillary tubing was set approx 20 cm above the surface of the calctum solution. With the 5-mL syrmge stall attached to the other end of the tubing, the microsphere suspenston m the tubing was extruded with the Harvard mfusion

325

Microencapsulation

GEL

SPHERE

IN LARGER GEL SPHERE

FORM MEMBRANE

DISSOLVE INSIDE

GEL

Fig. 2. Two-step method for bioencapsulation of a high concentration of smaller cells or microorganisms. Step 1: formation of small gelled alginate spheres containing cells. Step 2: to include the small gelled alginate spheres containing cells into larger gelled alginate spheres. This way, no cells are present on the surface of the larger gelled alginate spheres. Step 3: formation of membrane around the large gelled alginate spheres. Step 4: dissolution of content of spheres, including the smaller spheres inside, to form microcapsules with liquid contents. This helps to prevent entrapment of cells in the membrane matrix. Reproduced with permission from ref. 28.

5. 6.

7.

8.

pump. Similarly, the drops formed at the end of the sample jet were allowed to fall dropwise into a Pyrex dish containing a strainer cup, and filled with 300 mL of 100 mM CaCI,. The spheres were allowed to cure in the calcium solution for approx 15 min, after which they were removed and washed with buffered saline. The alginic acid matrix on the surface of the sphere was stabilized with polyL-lysine by immersing 5 mL (settled volume) of macrospheres in 80 mL of 50 mg% poly-L-lysine (50 mg% poly-L-lysine, 0.85% NaCl, 10 mA4 HEPES, 20 mA4 o-fructose, pH 7.4) for 10 min. The spheres were then drained, washed with buffered saline, and immersed into 200 mL of 0.2% sodium alginate (0.2% sodium alginate, 0.85% NaCl, 10 mA4 HEPES, 20 mM o-fructose, pH 7.4) for 10 min. to apply an external layer of alginate. After 10 min, the spheres were collected and immersed in 200 mL 50 mA4 sodium citrate (50 mA4sodium citrate, 0.47% NaCl, 20 mMo-fructose, pH 7.4) to solubilize the intracapsular calcium alginate. This may require up to 30 min, with frequent changes of the sodium citrate solution.

3.3.3. Macroporous Microcapsules When using cells or microorganisms to act on macromolecules, the above methods cannot be used. Thus, in using microorganisms to act on cholesterol

Chang bound to hpoprotem, porous microcapsules

we have to encapsulate (33)

the mlcroorgamsms

m macro-

1 A 2% agar (Dlfco) and 2% sodium algmate (Kelco) solution was autoclaved for 15 mm and cooled to 45’C to 50°C 2 P p~~torurn suspended in 0 4 mL of 0 9% NaCl was added drop by drop to 3 6 mL of agar algmate solution at 45”C, while stmmg vigorously. 3 Three mL of the mixture obtained was kept at all times at 45°C while it was bemg extruded through the syringe The extruded drops were collected mto cold (4°C) 2% calcium chloride, and allowed to harden These agar-algmate beads were about 2 mm m diameter 4 After 15 mm, the supernatant was discarded and the beads were resuspended m 2% sodmm c&ate for 15 mm 5 Then they were washed and stored m 0 9% saline at 4°C

When testing

for lmmoblhzed

bacterial

activity,

capsules were placed m a sterile 50-mL Erlenmayer

1 mL of beads/mlcro5 mL of serum were added

and a foam plug was fitted. Samples were withdrawn at specified intervals. When empty beads or microcapsules were prepared, the bacterial suspension was replaced by salme, and all the other steps were kept the same.

4. Notes 4.7. Microencapsula

tion of Enzymes

1 Cellulose nitrate membrane mlcrocapsules Hemoglobm at a concentration of 10 g/dL 1s necessary for the successful preparation Furthermore, this high concentration of protein stablhzes the enzymes durmg the preparation and also durmg reaction and storage (36). When the material (e g , NADH) to be encapsulated IS sensltlve to the enzymes present m hemoglobm, highly purified hemoglobm IS used This requires the use of purification using affmlty chromatography on an NAD+ sepharose column 2 Crosslinkmg with glutaraldehyde The long-term stability of microencapsulated enzyme activity can be greatly Increased by crosshnkmg with glutaraldehyde (36) This 1s done at the expense of reduced mltlal enzyme actlvlty 3 Oral admmlstratlon. When using cellulose mtrate mlcrocapsules contammg enzymes for oral admmlstration, the permeabihty of the membrane may need to be decreased This 1s to prevent the entry of smaller tryptlc enzymes. PermeabllIty can be decreased by decreasmg the proportion of alcohol used m dissolvmg the evaporated cellulose tutrate polymer 4 Interfacial polymenzatlon: Failure 111preparing good microcapsules is frequently

caused by the use of dlamlne or dlacids that have been stored after they have been opened. A new unopened bottle ~111usually solve the problems. Unlike the cellulose mtrate mlcrocapsules, m mterfaclal polymerization, the hemoglobm solution can be replaced by a 10% polyethelenelmme solution adjusted to pH 9 0 However, the mlcrocapsules prepared wlthout hemoglobm may not be as sturdy

327

Mcroencapsulation

Crosslinking the mrcroencapsulated enzymes with glutaraldehyde after the preparation of the enzyme mtcrocapsules could also be carried out to Increase the long term stabtlrty of the enclosed enzymes (36), although this decreases the mmal enzyme activtty Multtenzyme reaction In multrenzyme reaction requiring cofactor recyclmg, the cofactor can be crosslmked to dextran-70 and then encapsulated together wrth the enzymes For example, NAD+-N6-[N-(6-ammohexyl)-acetamtde] was coupled to dextran T-70, polyethylenetmme, or albumin to form a water soluble NAD+ derivative, and then encapsulated together with the multienzyme systems m the mtcrocapsules (18,29). This way both the cellulose nitrate mtcrocapsules and polyamide microcapsules can be used. This allows for hrgh permeation to substrates and products However, lmkmg cofactor to soluble macromolecules resulted m significant increases m steric hindrance and diffusion restrtcttons of the cofactor Retention of cofactors Ltprd-polyamide membrane mrcrocapsules contammg multrenzyme systems, cofactors, and substrates can retam cofactors m the free form Thus, analogous to the intracellular environments of red blood cells, free NADH or NADPH m solutron mstde the mtcrocapsules 1seffectively recycled by the multrstep enzyme systems that are also m solutron However, only hpophtltc or very small hydrophrltc molecules like urea can cross the membrane For example, ammonia and urea equthbrate mto the mrcrocapsules to be converted mto ammo acids (18,19) However, external alcohol instead of glucose had to be used as substrate for the converston and recyclmg of NAD+ to NADH Some substrates, e g , a-ketoglutarate, had to be encapsulated, smce they cannot enter the hprd complexed mrcrocapsules (18,29).

4.2. Microencapsulation

of Cells and Microorganisms

7 Standard method Algmate concentratton m the tested range, l-2 25% (w/v), does not affect the bactertal cell vtabthty or cell growth Quality of mtcrocapsules improves with increasing algmate concentratton from 1-l 75% (w/v) The use of 2% (w/v) algmate resulted m perfectly sphertcal shape and sturdy mtcrocapsules, wrth maximum number of encapsulated bacterial cells. An increase m hqmd flow rate of the algmate-cell or bacterial suspensron through the syrmge pump, from 0 00264 to 0.0369 mL/mm, resulted m increase m mrcrocapsule diameter The flow rate m the range of 0.007240 278 mL/mm resulted m good spherrcal mrcrocapsules. At an au-flow rate of 2 L/mm, the microcapsules had an average diameter of 500 f 45 m diameter. At the au-flow rates above 3 L/min, mrcrocapsules were irregular m shape These results mdrcate that algmate concentratton, au-flow rate, and hqutdflow rate are crtttcal for obtaining mtcrocapsules of desired characteristtcs and permselecttvtty (30) Usmg the following* 2% (w/v) algmate, 0 0724 mL/mm hqutd-flow rate, 2 L/mm air-flow rate, which has enabled the development of suitable mtcroencapsulatron process for cells or mrcroorgamsms

328

Chang

Mtcrocapsules prepared thts way are permeable to albumm, but tmpermeable to molecules with higher mol wt (37) Thus, hepatosttmulatmg factors (24) and globulin (37) cannot cross the membrane of the standard mtcrocapsules 8 Two-step method Thts method prevents the entrapment of small cells m the membrane matrix Mtcrocapsules prepared thts way when Implanted are much more stable, wtth decrease m reJectton (28,29) 9 Macroporous microcapsules Temperature IS a very crttical parameter in the tmmobtltzatton of P pzctorum (33) A low temperature produces gelatton of the polymer m the syringe or conduits A htgh temperature prevents gelation, but Increases the mortahty rate of P plctorum Exposmg P pzctorum to 55°C for 10 mm or more can completely inhibit enzymattc acttvtty However, up to 20 mm exposure to 45°C does not stgmficantly mhtbtt cholesterol acttvtty Open pore agar beads stored at 4°C dtd not show any stgn of detertoratton The beads retam their enzymattc activtty even after 9 mo of storage

Acknowledgments This research has been carried out with support to TMSC from the Medical Research Counctl of Canada tn both operating grants and the career mvestigator award It IS also supported by the Quebec MESST Vu-age Award of Centre of Excellence m Btotechnology to TMSC.

References 1 Chang, T M S (1964) Semipermeable mtcrocapsules Sczence 146, 524,525 2 Chang, T M S , Macintosh, F C , and Mason, S G (1966) Semipermeable aqueous mtcrocapsules: I Preparation and properttes Can J Physzol Pharmacol 44, 115-128 3 Chang, T M S., Macintosh, F C , and Mason, S G (1971) Encapsulated hydrophtltc composmons and methods of makmg them. Canadtan Patent 873,s 15, 197 1 4 Chang, T M S (1972) Arttfictal Cells, Monograph, Charles C Thomas, Sprmgfield, IL 5 Chang, T. M S (1995) Arttfictal cells with emphasis on btoencapsulatton m biotechnology. Blotechnol Annu Rev 2,267-295 6 Chang, T, M S (1997) Recent and future developments m modtfied hemoglobm and mtcroencapsulated hemoglobm as red blood cell substmtes Artzf Cells Blood Substitutes Immoblllzatlon Blotechnol 25, l-24 7 Chang, T. M. S (1997) Blood Substitutes Prmclples, Methods, Products and Clmlcal Trzals Karger/Landes, Austm, TX. 8. Chang, T M S andYu, W. P. (1996) Btodegradable polymer membrane contammg hemoglobm for blood substitutes U S A. Patent approved m 1996 9 Yu, W P and Chang, T M S (1996) Submicron polymer membrane hemoglobm nanocapsules as potenttal blood substitutes. preparation and charactertzatton. Artzj” Cells Blood Substitutes Immobzl~zatzon Bzotechnol 24, 169-l 84 10. Chang, T M. S. and Poznansky, M J (1968) Semipermeable mtcrocapsules contammg catalase for enzyme replacement m acatalsaemtc mace Nature 218,242-245

Microencapsulation

329

11 Chang, T M S (1971) The m vtvo effects of semipermeable microcapsules contammg L-asparagmase on 6”C3HED lymphosarcoma Nature 229, 1 17,118 12 Chang, T M. S. (1989) Preparatton and characterization ofxanthme oxtdase tmmobthzed by mtcroencapsulatton m arttfictal cells for the removal of hypoxanthme. J Blomater Art$ Cells Artzf Organs 17, 611416 13 Palmour, R M , Goodyer, P , Reade, T , and Chang, T. M S (1989) Mtcroencapsulated xanthme oxldase as experimental therapy m Lesch-Nyhan Disease Lancet 2, 687,688 14 Chang, T. M S., Bourget, L , and Ltster, C. (1992) Orgal admmlstratton of mlcrocapsules for removal of ammo actds, US Patent No 5,147,64 I 15 Chang, T M S , Bourget, L , and Ltster, C (1995)A new theory of enterorecnculatton of ammo acids and Its use for depletmg unwanted ammo acids usmg oral enzymearttfictal cells, as m removing phenylalanme in phenylketonurta Artlf Cells Blood Substitutes Immoblllzatlon Blotechnol 25, l-23 16 Bourget, L and Chang, T M S (1986) Phenylalanme ammonia-lyase tmmobthzed m mtcrocapsules for the depletion of phenylalanine m plasma m phenylketonurtc rat mode1 Bzochzm Bzophys Acta 883,432438 17 Safos, S and Chang, T M S (1995) Enzyme replacement therapy m ENU2 phenylketonurtc mace using oral mtcroencapsulated phenylalanme ammoma-lyase a prehmmary report Artlf Cells Blood Substrtutes Immoblhzutlon Blotechnol 25,68 1-692 18 Chang, T. M S (1985) Arttficlal cells wrth regenerating multtenzyme systems Methods Enzymol 112, 195-203. 19 Gu, K F., Chang, T M S (1990) Productton of essential L-branched-chained ammo actds, m broreactors contammg arttficlal cells tmmobthzed multtenzyme systems and dextran-NAD” Appl Bzochem Bzotechnol 26,263-269 20 Chang, T. M S (1965) Semtpermeable aqueous microcapsules PhD Thesis McGill Umverstty 21 Llm, F and Sun, A M (1980) Mlcroencapsulated islets as bioartlfictal endocrme pancreas. Sczence 210,908-909 22 Wong, H. and Chang, T M S (1986) Btoartlfictal liver implanted arttfictal cells mlcroencapsulated hvmg hepatocytes increases survival of hver failure rats /nt J Artlf Organs 9, 335,336 23 Wong, H and Chang, T M S (1988) The vtabtlny and regeneration of arttficlal cell mtcroencapsulated rat hepatocyte xenograft transplants m mice J Blomater Artlf Cells Artif Organs 16, 73 1-740 24. Kasham, S and Chang, T M S (199 1) Effects of hepattc sttmulatory factor released from free or mlcroencapsulated hepatocytes on galactosamme induced fulmmant hepattc fatlure animal model J Bzomater Artzf Cells Immoblllzatlon Bzotechnol 19,579-598 25 Brum, S. and Chang, T M S (1989) Hepatocytes immobthzed by mtcroencapsulatton m arttfictal cells: effects on hyperbthrubmemla m Gunn Rats J Blomater Art/f Cells Artlf Organs 17, 403-I 2 26 Brum, S and Chang, T M S (1991) Encapsulated hepatocytes for controllmg hyper-brhrubmemta m Gunn Rats Znt J Artlf Organs 14,239-241

330

Chang

27 Brum, S and Chang, T M S (1995) Kmetrcs of UDP-glucuronosyl-transferase in brhrubm conJugatron by encnapsulated hepatocytes for transplantation mto Gunn rats J Artlf Organs 19, 449-457 28 Wong, H and Chang, T. M S (199 1)A novel two stepprocedure for rmmobthzmg hvmg cells m mrcrocapsulesfor rmprovmg xenograft survrval J Blomater Artzf Cells immobdlzatcon Blotechnol 19,687498 29 Chang, T M. S and Wong, H (1992) A novel method for cell encapsulatronrn artrficral cells USA Patent No 5,084,350 30 Prakash, S and Chang, T M S. (1995) Kmetrc studies of mrcroecnapsulated genettcally engmeeredE colt cells contammgK aerogenesgene for urea and ammoma removal. J Bzotechnol Bloeng 46,621-626 3 1 Prakash, S and Chang, T M S (1996) Mrcroencapsulated genetically engmeered hve E colz DH5 cells administeredorally to mamtam normal plasmaurea level m uremic rats Nature Med 2, 883-887 32 Koo, J. and Chang, T M S (1993) Secretion of erythroporetm from mrcroencapsulated rat krdney cells prehmmary results ht J Artf Organs 16, 557-560 33 Garofalo, F and Chang,T M S (1991) Effects ofmass transfer and reaction kmettcs on serum cholesterol depletion rates of free and rmmobrlrzed Pseudomonas pzctorum Appl Blochem Blotechnol 27,75-91 34 Lyold-George, I and Chang, T M S (1995) Characterrzatron of free and algmatepolylysme-algmate mrcroencapsulatedErwznla her&cola for the conversion of ammonia,pymvate and phenol mto L-tyrosme and L-DOPA J Bloeng Bzotechnol 48,70&7 14 35 Yu, Y T and Chang, T M S (198 1) Lipid-polymer membraneartrficral cells contaming multrenzyme systems,cofactorsand substratesfor the removal of ammoma and urea Trans Am Sot Artlf Intern Organs 27, 535-538 36 Chang, T M S. (1971) Stabrhzatron of enzyme by mrcroencapsulatronwrth a concentrated solutron o or by crosslmkmg with glutaraldehyde Blochem Bzophys Res Corn 44, 1531-1533 37 Coromrli, V and Chang, T. M S (1993) Polydrsperse dextran as a diffusing test soluteto study the membranepermeabrhty of algmate polylysme mlcrocapsules J Blomater Artlf Cells Immoblkzatlon Bcotechnol 21,323-335

26 Methods for Microencapsulation

with HEMA-MMA

Shahab Lahooti and Michael V. Sefton 1. Introduction Encapsulation of cells m a membrane prior to implantation holds potential for controllmg the adverse Immune response that may be generated agamst the transplanted cells, by physically isolating the cells from the host’s immune system.If successful, encapsulation eliminates or mnnmizes the adverse effects of immunosuppresstve therapy and permits the use of xenogeneic cells. Ideally, the capsule membrane holds permselective properties, so that the passage of nutrients, growth factors, and the therapeutic product secreted by the cells occur readily across the membrane, but mediators of the immune system do not penetrate the membrane (Fig. 1). The major types of immunoisolation devices include intravascular arteriovenous shunts, diffusion chambers of tubular or planar geometry, and microcapsules (I-6). Our current standard microencapsulation technique utilizes a stationary coaxial needle assembly submerged m a flowing stream of hexadecane or dodecane (Fig. 2; 7). Cells suspended in the appropriate tissue-culture medium (and extracellular matrtx components) are pumped through the inner needle, while the polymer solution is pumped through the outer needle. This leads to the formation of droplets of polymer solution surrounding a core of cell suspension at the tip of the coaxial needles. The droplets are sheared from the tip of the needle by the flowing hexadecane (Fig. 3). Subsequently, upon travelling through a hexadecane layer, the droplets enter a phosphate-buffered salme (PBS) solutton m a receiving dish. The hexadecane in the receivmg dish is recirculated to the needle assembly by a pump In the PBS, the polymer solvent 1sextracted to precipitate the polymer membrane, forming spherical microcapsules containing a core of cell suspension The capsules are maintained m suspension m the PBS prectpitatton bath by a magnetic stirrer. The PBS solution contains 100-150 From

Methods

m Molecular

Edlted by J R Morgan

Medicine,

Vol

18

and M L Yarmush

337

Ttssue

Engtneenng

0 Humana

Methods

and

Press Inc , Totowa,

Protocols NJ

cell-derived therapeutic agent

4

immunoisolating permselective membrane

lymphocytes, i.,, antibodies, ’ complement f

extracellular matrix cell

nutrients, growth factors, external stimulus Ftg 1 Schematic drawing of a model mtcrocapsule. Cells are encapsulated with or wlthout an extracellular matrix m a permselecttve membrane, whtch provides tsolatton from the mediators of the mnnune system, but allows the passage of nutrients, growth factors, and the therapeutic product secreted by the cells

cell suspension polymer solution pump

co-axial needle assembly

It

overhead stirring rod hexadecane

c,

l

PBS + surfactant

l

I

0

l e

l

ll *a a @o

l I

l

0

capsule

0-8 ,’

l 1

Ftg 2. Schematic drawmg of the stationary-needle assembly mtcroencapsulatton setup producing small-diameter (-3OMOO pm) capsules. Droplets of polymer solution surroundmg a core of cell suspension formed at the ttp of coaxial needles are sheared off by flowmg hexadecane into a PBS bath, where the capsule membrane 1s formed by interfactal prectpttatton. A peristaltic pump rectrculates the hexadecane and a magnetic overhead stirrer mamtams the capsules m suspensron

Microencapsulation

333

with HEMA-MMA

co-axial needle flowing shexadecane or dodecane convection nozzle

sheared droplet Fig. 3 Schematic drawing of the release of droplets from the stattonary co-axial needle assembly Cells suspended m the appropriate tissue-culture meduun (and extracellular matrix component) are pumped through the inner needle, the polymer solutton IS pumped through the outer needle of a co-axtal needle assembly The co-axral needles are submerged m a flowrng stream of hexadecane or dodecane As a result, droplets of polymer solution surroundmg a core ofcell suspension are sheared from the ttp of the needle assembly as they are formed. The rate of capsule production and the size of the capsules may be controlled by the flow rate of the hexadecane or dodecane and by the verttcal posltton of the needle assembly within the nozzle ppm of a noniomc surfactant, Pluronn? LlOl (BASF, Parstppany, NJ), which reduces the interfactal tenston at the hexadecane-PBS interface, and thus facrhtates the passage of the capsules through the interface. After washing the capsules m a fresh PBS solutton, they are Incubated m a Petri dish containing the appropriate tissue-culture medmm for the cells at 37°C and m 95% air/5% co*. In a previous setup, droplets were sheared from the coaxial needles by oscrllating the needle m the vertical plane by a motor and cam assembly at a hexadecane-an interface (Fig. 4; 8). This simpler setup resulted m formatron of mtcrocapsules wtth an od of -700-900 pm and a wall thickness of 50-150 pm (8). In the present setup, the higher shearmg force produced by the flow of hexadecane leads to smaller diameter capsules. Mlcrocapsules ranging in dtameter from 300 pm to 600 pm may be produced by adJusting the hexadecane flow

334

Lahoot~ and Sefton

polymer

suspension

syringe

pump

cam and motor

L-

:

extrusron

needle

hexadecane

preapctatlon

bath

PBS/L1 01

0

-

magnetic

stmer

Fig 4. Schematic drawmg of the osclllatmg needle assembly mlcroencapsulatlon setup producmg large-diameter (-70&900 pm) capsules Droplets of polymer solution surrounding a core of cell suspension are sheared from the tip of the coaxial needle assembly by the vertical osclllatlon of the needles at a hexadecane-air Interface The sheared droplets fall mto a PBS bath, where the capsule membrane IS formed by mterfacial preclpltatlon The vertical osclllatlon of the co-axial needles IS achieved by a cam and motor assembly The capsules are maintained m suspension by a magnetic stirrer Reproduced with permlsslon from ref 13

rate; the hrgher the flow rate* the smaller the capsules (9) The polymer membrane has a thrckness ranging from 20 p to 90 )SI-I(Fig. 5). The polymer utilized for the formation of our capsules 1s a thermoplastrc polyacrylate copolymer, namely, hydroxyethylmethacrylate-methylmethacrylate (HEMA-MMA, -75 mol % HEMA), prepared by solution polymerrzatton after careful monomer purtficatton to reduce the crosslmker content (5,IU). This copolymer IS hydrophilic, wtth a 25-30% (w/w) water uptake (&II) con-

Microencapsulation

with HEMA-MMA

335

Fig. 5. Scanning electron micrographs of cut-open small diameter microcapsules prepared by the stationary needle assembly setup: (A) capsule produced with a 15% (w/v) Ficoll400 core solution, (B) capsule produced with a 2-mg/mL collagen solution, in which the collagen gel can be seen in the core of the capsule.

sistent with the poly(HEMA) content, but has mechanical strength, toughness, and elasticity imparted by the poly(MMA) component. The water insolubility of the HEMA-MMA copolymer provides stability in the aqueous physiological environment, but necessitates the use of an organic solvent to prepare the polymer solution. Polyethylene glycol 200 (PEG 200) is used as the solvent, because cells tolerate PEG 200 without loss of viability, provided the direct contact between the cells and the solvent is limited, as is the case here; alternatively, triethylene glycol (TEG) may be used. HEMA-MMA is dissolved in PEG 200 at a concentration of 9-l 0% (w/v). The capsule structure can be modified by changing the ratio of the polymer solution to cells during co-extrusion. Also, the capsule membrane permeability and mol wt cut-off can be modulated by the polymer solution concentration or composition (12), or the composition of the precipitation bath (13). Normally, the cell suspension contains 1O-20% of Ficoll@ 400 (Pharmacia, Uppsala, Sweden), a neutral, highly branched, hydrophilic polymer of sucrose. Ficoll400 increases the viscosity of the core solution. As a result, less mixing of the cell suspension with the polymer solution occurs, a well defined core is produced, and cell viability is improved. Addition of a viscous extracellular matrix (ECM) solution, e.g., collagen, Matrigel@ (Collaborative Research,

336

Lahooti and Sefton

Bedford, MA), agarose, or chitosan to the cell suspension can also serve the same purpose. Collagen provrdes a matrix for attachment and better dtstrtbution of the cells within the core of the capsules. Matrigel, a reconstituted basement membrane, IS prepared from a urea extract of Engelbreth-Holm-Swarm (EHS) tumors. Simrlar to collagen, Matrigel provides an extracellular matrix within the core of the capsules to which the cells attach and grow and maximize their expression of differentiated functions. Collagen and Matrigel are ltquid at low temperatures (4”C), but are a gel at 37°C (gelling temperature -22-35°C) Thus, it is necessary to cool the syringe and tube delivermg the cell suspension to the needle assembly in order to prevent the soltdtlication of collagen or Matrigel prior to droplet formation Also, because of the lower cell-suspension flow rate in the stationary nozzle setup, the temperature surroundmg the coaxial needle assembly 1slowered by usmg a cooled flowing stream. In this case dodecane is used m place of hexadecane because of its lower melting point (-9.6”C vs 18.2”C). Also, the precipitation bath is warmed to about 30°C to promote gelation of the core collagen or Matrtgel solution, as soon as the capsules are formed In contrast, in coencapsulation with agarose, the temperature of the prectpitation bath is lowered to below 15°C and the core syringe is maintained at room temperature Typically, ultra-low gelling agarose, which IS a liquid at 37”C, but forms a gel below 17°C 1sused. Because of contact with the cooled precipttatton bath, dodecane 1sused m coencapsulation with agarose Agarose acts as an nnmobihzation matrix that provides better distrtbution of the cells withm the core, so that nutrient diffusion limitations are mmlmized Other unmobrllzation matrices may also be used. For example, m coencapsulation with chitosan, the deacylated form of chitm which is the mam constituent of shrimp and crab exoskeletons, the neutral pH of the precipttatton bath promotes preciprtation of chitosan, thus, temperature adjustments of the setup are not required A variety of cell types, mcludmg the Chinese hamster ovary (CHO) libroblast (24), rat pheochromocytoma, PC12 (15), human hepatoma, HepG2 (16,Z 7), murme fibroblast, L929 (18) cell lures, primary rat hepatocytes (19), and rat islet tissue (20), have been encapsulated with HEMA-MMA microcapsules. These studies have shown that the cells survrve the encapsulation procedure, despite the exposure to shear forces and organic solvents/nonsolvents, and grow or function afterwards in vitro for periods from 2 to at least 6 wk. Also, mouse fibroblasts (2A-50), genetically engineered to secrete human growth hormone and p-glucuronidase (21), have been microencapsulated to evaluate the use of encapsulated cells for gene therapy. The performance of encapsulated cells IS generally evaluated by a calorimetric assaytesting the metabolic activity of cells accordmg to their ablhty to convert MTT to formazan (14,22), by direct cell counts and trypan blue exclusion to test proliferation and viability (24), by scanning

Microencapsulatlon

w/th HEMA-MMA

337

CELL SU: ‘ENSION 22g needle ___c pipette tiptubing adapter i 16g Luer-Lok --. I

I l----d7

J--

POLYMER SOLUTION

pipette tip Ftg. 6 Schemattc of the needle assembly used m the osclllatmg-needle assembly mrcroencapsulatron setup The coaxial needle assembly conststs of an outer 16-gage Luer-Lok needle to whrch IS added a side arm for the polymer solutron, and an inner 22-gage needle that the cell suspensron ISpumped through A plastic prpet tip IS placed on the bottom of the 16-gage needle to facilitate the shearing off of droplets Reproduced with permrssron from ref 8

electron mtcroscopy and htstologtcal sections for morphologtcal (Id), and by enzyme-linked tmmunosorbent assay techniques secretion (I 7).

assessment for protein

2. Materials 1 Needle assembly. The coextruston needle for the setup m which the assembly undergoes vertical osctllattons consrsts of a 16-gage Luer-Lok needle that has been modified wtth a stde-arm that acts as the inlet for the polymer solution (Fig. 6) A plastic ptpet tip (Ftsher Sctentlfic, Don Mills, Ontario, Canada) 1splaced on the bottom of the 16-gage needle to facilitate the shearing off of droplets. The inner needle IS a 22-gage needle fitted through a tubing adapter (male Luer-Lok tip to tubing, Becton Drckmson, Franklin Lakes, NJ) that is locked mto the Luer-Lok mechanism The inner needle extends through the outer needle, m a centered

Lahootl and Sefton

338 \

+-

Inner Needle

Hexadecane or Dodecane Inlet

Fig 7 Details of the needle assembly used m the stationary-needle assembly mtcroencapsulatton setup The coaxial needle assembly conststs of an outer 16-gage needle and an inner 22-gage needle The core cell suspenston 1s pumped through the Inner needle and the polymer solution IS pumped through the outer needle via a side arm The needles are enclosed m a Lucite casing that forms a nozzle, provtdmg a uniform flow of hexadecane around the needle tip The needles have been tapered to factlltate droplet formation and shearing off Also, the inner needle is equipped with three knobs m its mtdsectton that provide centering within the outer needle without tmpedmg the flow of the polymer solutton Adapted from ref 21. position, and protrudes -100 pm past the end of the ptpet tip The ptpet ttp and the 22-gage needle are blunt cut In the current setup, m which the coaxral needle IS stationary wtthm a flowing stream of hexadecane, modtfications have been made to the needle assembly (Fig. 7) The ptpet tip at the bottom of the 16-gage needle has been removed Instead, the needle itself has been machined so that it is tapered on the mstde and outside, thus facthtatmg the shearing off of droplets The inner needle has also been tapered on the outside to facthtate better flow of the polymer solutton around the core solution. Also, the inner needle is equipped with three knobs m its mtdsectton that facilitate its centering within the outer needle without impedmg the flow of the polymer solution In addmon, the needle assembly 1s enclosed m a Lucite@ casmg, which forms a convectton nozzle that creates a umform flow of hexadecane around the needle tip The Lucite casing also has two inlets for the hexadecane

Microencapsulation

with HEMA-MMA

339

flow Thus, the hexadecane flows through the inlets mto the casing, around the coaxial needle, and exits, along with the sheared droplet, through the convectton nozzle, mto the recervmg dash 2 Syringes B-D, l-cc, regular Luer-Lok tip syringes (Becton Dickinson) are used for both the polymer and core solutrons If desired, a Multifit, ~-CC, glass syringe with Luer-Lok trp (Becton Dtckmson) that has a micro stir bar (7 mm L x 2 mm od, Cole-Parmer Instrument Co, Vernon Hills, IL) mstde its barrel may be used to mamtam the cells m suspension and prevent them from settling to the side of the syrmge. The microstir bar IS stirred by a mtcroelectromagnettc stirrer (ColeParmer) that 1s suspended above the syringe. The sttrrmg of the core solutton, however, is not performed m coencapsulatrons with collagen or Matrtgel, since the syringe IS immersed m an ice-water bath. The core and polymer syringes are fitted with 22-gage and 18-gage blunt needles (MonoJect@, Sherwood Medical, St Louts, MO), respecttvely 3. Pumps The pump for hexadecane or dodecane IS a Masterflex@ L/S system with a standard drive (MR-7553-70) and a Quick Load pump head (Cole-Parmer). Also, a pulse dampener with a dead volume of 190 mL (MR-07596-20, ColeParmer) IS used to attam a steady flow of hexadecane around the needle assembly. Separate syringe pumps are used for the core and polymer syringes (model A-99, Raze1 Sctentrfic Instruments, Stamford, CT) Tubmgs The syringe contammg the core solution is connected to the needle assembly with polyethylene tubing (PE-50, 0.58 mm id x 0 965 mm od, Clay Adams, Becton Dtckmson, Parsrppany, NJ). The syringe contammg the polymer solution is connected to the needle assembly with Srlastlc@ tubing (0 04 m Id x 0 085 m od, Dow-Cormng, Midland, MI) For hexadecane flow, the needle assembly, pump, and recelvmg dish are connected with Tygon@ fuel and lubricant tubing (size 16, Masterflex, Cole-Parmer) Recervmg dish and stirrer For the oscillating needle assembly setup, the receiving dish is a 250-mL volumetric flask (Pyrex, Corning, Cornmg, NY) A 8 mm od x 32 mm L (Fisher Scientific) stir bar IS used to maintain the capsules m suspension In the setup m which a flowing stream of hexadecane is used, a receivmg dish with a larger openmg IS utilized In thus case, the receiving dish (155 mm L x 95 mm id) 1s custom-made from glass. An overhead stlrrmg rod 1s used to prevent damaging the small capsules against the bottom of the dish. The strrrmg rod 1scomposed of a metal rod that is placed m a bearing with a Teflon-coated magnetic strrrmg bar of 40 mm L x 10 mm od at the end HEMA-MMA solution The appropriate mass of polymer IS dissolved m PEG 200 (BDH, Poole, England) to make a 9-10% (w/v; the volume of the solvent rather than the volume of the final solution IS considered) polymer solution Complete dtssolutton of the powder requires about 3 d, and a transparent solutron IS produced. Precrpltation bath Pluromc@ LlOl (BASF) 1s added to precooled (4°C) PBS (Dulbecco’s calcmm/magnesmm-free, tissue-culture medium preparation, Umverstty of Toronto) to produce a 1500 ppm (10X) surfactant solution. The bottle

Lahooti and Sefton is mamtamed at 4°C overmght (and thereafter), m order for the LlOl to go mto solutton. This stock solution is diluted 1:9 with PBS, prior to encapsulation 8. Core solutions The cells used for encapsulation are suspended m one of the following soluttons to form the core mixture. In all cases, the type of tissue-culture medmm used is dictated by the cells a Ftcoll solution Ficoll 400 is dissolved m an appropriate volume of tissueculture medium by vigorous mrxmg, and the solution 1s sterile-filtered through a 0 2-pm filter (Mtlhpore, Bedford, MA). Subsequently, the appropriate amounts of serum and antibiotics are added to produce a l&20% (w/v) Ficoll solution In addition, HEPES buffer (Gibco-BRL, Grand Island, NY) may be added to mamtam the neutral pH of the medium b Collagen solutron A Vltrogen lOO@(Bovine dermal type I collagen, Collagen Corp , Fremont, CA) solution IS prepared according to the manufacturer’s mstructtons Eight mL of chilled 3 mg/mL Vttrogen 100 IS mixed wtth 1 mL of 10X solution of buffered cell-culture medium containing phenol red Then, 1 mL of 0 1 MNaOH (endotoxm-free, Stgma, St Louis, MO) is added, and the solution is mixed The pH of the solution is adjusted to 7 4 & 0 2 by dropwtse addition of 0 1 A4 HCl or 0 1 MNaOH The solution is mixed with an appropriate amount of tissue-culture medium to produce a l-2 mg/mL collagen solutton The collagen solution IS mamtamed at 4°C and used within hours for encapsulation c Matrtgel solutton Matrigel is mixed m a 1.1 (v/v) ratio with tissue-culture medium at a temperature of 4’C d Agarose solutton A 2% (w/v) agarose solution is prepared according to the manufacturer’s mstructions SeaPrep@ agarose (FMC BioProducts, Rockland, Maine), an ultra-low gelling temperature (~17°C) agarose, IS weighed and added to a beaker containing stnrmg PBS (Dulbecco’s calcmm/magnesmm free) at room temperature The beaker is 2-4 times the volume of the solution The beaker and solution are weighed and the beaker IS covered with plastic wrap, which has a small hole that allows for ventilation The beaker is transferred to a heater/stirrer umt and the solution is brought to a boll The gentle boilmg and stirring are contmued until all of the agarose powder is dissolved (10 mm) Sufficient hot PBS is added to the beaker to replace that which ts lost, and the solutton 1s mixed thoroughly The agarose solutton 1s maintained at 37°C and IS used for encapsulatton withm 24 h For encapsulation, the solution is mixed at a 1.1 (v/v) ratio with tissue-culture medmm to obtam a 1% (w/v) agarose concentration 9 Temperature control components For coencapsulation with collagen or Matrigel m the stationary needle assembly setup, additional apparatus is required to cool the cell suspension to prevent gelation prior to formatron of droplets at the ttp of the needle assembly (Fig. 8) An ice-water bath is used for the core mixture syringe and the polyethylene tubing dehvermg the cell suspensionto the needle assembly The polyethylene tubmg is also inserted m a water-Jacket tube (sillcone tubing, size 17, with size 16 side-arm, Masterflex, Cole-Parmer), which is

Microencapsulation

341

with HEMA-MMA water-jacket tube

I/ synnge

Ice-water bath

0

0 000

0

0 000 0 0

O

PBS o ’ 0

0

0

heater/stirrer n

Fig 8 Modifications to the stationary needle assembly microencapsulation setup for coencapsulation with collagen or Matrigel. Collagen and Matrigel solutions have a gelling temperature of -22-35’C, thus, they must be maintained at a low temperature prior to droplet formation The syringe containing the collagen or Matrtgel solution is immersed m an me-water bath The polyethylene tubing deltvermg the cell suspension to the needle assembly is cooled by msertton m a water-Jacket tubing, where cold water is recirculated by a pump The fIowmg stream of dodecane is cooled by immersion of the dampener of the pertstalttc pump used to recirculate the dodecane m dry ice Also, once the capsules are formed, the prectpttatton bath is warmed to facilitate the gelling of the core solution

fed by a pump (same model as the pump for dodecane) Also, the stream of dodecane surroundmg the coaxtal needle IS cooled by immersmg the dampener for the peristaltic pump m dry ice In order to increase the exposure time of a given volume of circulatmg dodecane to the dry me, a 500-mL bottle, which has inlet and outlet ports is used m place of the smaller 190 mL dampener To faciletate the gelling of the collagen-Matrigel solutions m the core of the capsules, a magnetic stirrer/heater is used to warm the precipitation bath to approx 30°C. In the osctllatmg needle assembly setup, the flow of cell suspension is higher (see

Lahooti and Sefton Subheading 3.3.), therefore, tmmersion of the core syringe m me-water bath IS sufficient to prevent gelatton prior to droplet formatton For coencapsulatton with agarose, the recetvmg dish 1s placed m a cool water bath to reduce the temperature of the prectpttatton bath to below 15’C

3. Methods 3.1. Stationary

2

3 4.

9 10

Needle Assembly, Small Diameter Microcapsules

To ensure sterility, all operations are performed m a stertle flow hood Glassware and the Stlasttc tubmg are steam-sterdtzed, the polyethylene and Tygon tubmg are gas-stertbzed All chemtcals are dtspensed msrde the flow hood The needle assembly IS cleaned by aspirating water and then 95% ethanol through the core needle The polymer needle 1s cleaned by aspiration of 95% ethanol only; ethanol 1s a solvent for HEMA-MMA The needles are subsequently andried The Lucite casing and nozzle are also cleaned with water and 95% ethanol, but overexposure to ethanol IS avotded To ensure sterrltty of the needle assembly, tt IS placed m an environment saturated with alcohol vapor 1 h prior to encapsulatton Prior to cleanmg, the needle assembly IS exammed under a stereoscoplc mtcroscope to ensure the appropriate protrusion (-100 pm) and centermg of the core needle with respect to the outer polymer needle, and the appropriate verttcal position of the coaxial needles wtthm the convectton nozzle (4 mm from the tip of a nozzle that has a 20-degree angle and a 2 2 mm opening) The magnetic stirrer, pertstalttc pump, and syringe pumps are postttoned m the flow hood The syringe pumps are typically placed on stands The recetvmg dish IS filled with 555 mL of PBS containing 150 ppm LlOl The dish IS placed on the stirrer and the overhead sttrrmg rod IS posittoned m a centered, vertical position, so that the stnrmg rod 1s 2-3 cm from the bottom of the dash The dish 1s subsequently tilled completely with 500 mL of hexadecane or dodecane The polyethylene, Stlasttc, and Tygon tubmgs are connected to the needle assembly The needle assembly is posmoned above the receiving dish The other end of the Tygon tubing IS connected to the dampener, and the dampener 1s connected to the recetvmg dish with a tube that IS threaded through the peristaltic pump The pump IS used to fill the Tygon tubing, dampener, and the needle assembly wrth hexadecane The cells are prepared and suspended m one of the core solutions mentioned above, typically at a concentration of 5 x lo6 cells/ml The core syringe IS tilled with the cell suspension, ensuring that au bubbles are removed Subsequently, the syringe 1s fitted with the appropriate blunt-ended needle, connected to the polyethylene tubmg, and placed wtthm the syringe pump The mtcrostnrer IS posmoned above the syringe and IS set to the lowest stnrmg speed Slmtlarly, the polymer syringe IS tilled and installed m the syringe pump The pertstalttc pump IS started and the overhead sturer IS operated at approx 300 rpm The speed of snrrmg should be such that the capsules are mamtamed m suspension

Microencapsulation

11

12

13

14

15

with HEMA-MMA

343

without creating undue mrxmg between the top hexadecane layer and the bottom PBS layer Ptlor to stnrmg, the correct posrtton of the nozzle over the recetvmg dish IS checked by the locatton that the hexadecane Jet impacts the hexadecane-PBS mterface It 1s desuable for the droplets to land at the hexadecane-PBS interface, at a positton between the sttrrmg rod and the wall of the recetvmg dish. The momentum of the droplets, the low mterfactal tension at the Interface, and the vortex created by the overhead stirrer facrhtate the entry of the droplets into the prectprtatton bath To commence the encapsulatron process, the syringe pumps are turned on The core mixture and polymer solution are advanced manually to the ttp of the needle assembly while the syringe pumps are on First, the polymer 1s advanced to the tip of the needle assembly, and then the core solutron IS advanced, manual operation of the syringe pumps 1s stopped once the solutions have reached the ttp of the needles. Typtcally, the mtttal contact between the polymer solutton and the core solutton produces a polymer prectpttate at the ttp of the needle Thus, It IS advisable to set the mittal hexadecane flow rate sltghtly higher, m order to dislodge the precipitate from the tip of the needles Once the mltlal dtsturbance IS relieved, and good core and polymer droplet formation IS achieved, the hexadecane flow rate IS reduced to the operattonal level The approprtate hexadecane flow rate may be checked by a measurement of the rate of capsule production Typical flow rates for the cell suspenston, polymer solutton, and hexadecane are 0 0055 mL/mm, 0 0 125 mL/mm, and 118 mL/mm, respectively. These flow rates result m the productton of -60 capsules/mm with a 20-degree nozzle angle, 2 2 mm nozzle opening, and a 4-mm distance between the ttp of the nozzle and the coaxtal needles These capsules have a final diameter of -400 pm and a membrane thtckness of -50 pm However, these flow rates may be adjusted to alter the srze of the capsules and the membrane thickness In order to stop encapsulation, the core pump 1s turned off, then the polymer pump, and then the perrstalttc pump The needle assembly IS removed from the recetvmg dish, all tubmgs are disconnected, and the hexadecane layer IS removed for reuse The capsules are subsequently transferred to a spinner flask that contains fresh PBS (wtthout Pluromc LlOl) and maintained in suspension for an addrttonal 30 mm. Once the sttrrmg rod IS removed from the receiving dish, the capsules ~111 accumulate at the bottom of the dish They should be transferred immediately to the spinner flask with a transfer ptpet Subsequently, the capsules are washed m a Petri dish with tissue-culture medium, to remove any remammg traces of hexadecane or the prectpttatton bath. They are then transferred to a Petrt dish contammg the appropriate tissue-culture medium and maintained at 37°C m 95% au/5% COZ. Alternatively, m order to prevent the capsules from adhering to one another, the capsules may be kept m suspension by a magnetic stirrer m the incubator. Grossly defective capsules are removed under a stereoscoprc mrcroscope

Lahootl and Sefton

344 3.2. Oscillating

Needle Assembly

Large-Diameter

Microcapsules

The encapsulation process for the stationary needle assembly setup follows the same prmctpal steps as outlined above, with the followmg modtficattons. 1 The mner needle IS Inserted manually through the Luer-Lok end of the assembly mto the outer needle, and its centermg is adjusted The plastic tip is Inserted onto its sleeve on the outer needle, and the protrusion (-100 pm) of the mner needle from the plastic ttp IS adJusted wrth the aid of a stereoscopic microscope 2 The 250-mL volumetric flask IS filled with the precipitation bath, except for the top 7 cm m the neck, which IS tilled with a hexadecane overlayer 3 After attachmg the polyethylene and Silastic tubing, the needle assembly is mstalled on the arm of the cam and motor assembly, above the openmg of the volumetrtc flask 4 Once the syrmge pumps and the magnetic sturer (low settmg) are started, the coextruslon needle is oscillated m the vertical plane by the motor and cam assembly at a frequency of 30/mm, with one droplet released each stroke The height of the assembly IS adjusted so that the droplet formed and the ttp of the needle assembly are completely immersed m the hexadecane layer Because the needles are not enclosed, any initial precipitate can be easily drslodged from the tip of the needles 5 The cell suspension is pumped to the needle tip at a flow rate of -0 02 mL/min, and the ratio of polymer solutton to cell flow rate varies from 1 5 1 to 2 67.1, dependmg on the cell suspension composition (viscosity)

3.3. Coencapsulation

with Extracellular/lmmobilization

Matrices

1 For coencapsulatton with agarose, the receiving dish is placed in a water bath to maintain Its temperature below 17°C The core syringe IS mamtamed at room temperature. Also, dodecane is used, rather than hexadecane 2 For coencapsulatron with collagen or Matrtgel, cooling of the core solution IS required In the oscillatmg needle assembly setup, after attaching the polyethylene tubmg to the core syringe, the syrmge and a portion of the polyethylene tubmg are immersed m an ice-water bath 3 In the stationary needle setup, the core flow rate is less than the flow rate m the oscrllatmg needle setup (0 0055 mL/mm vs 0 02 mL/mm) Therefore, addmonal provisions are needed to cool the core suspenston for coencapsulation wtth collagen or Matrlgel (Fig. 8) a After attachmg the polyethylene, Stlastrc, and Tygon tubmgs to the needle assembly, the polyethylene tubing 1sinserted completely inside another tubing, so that the end of the polyethylene tubmg that is not attached to the needle assembly protrudes to the outstde This outer tubing acts as a cold water-Jacket tubing around the polyethylene tubing The water-Jacket tubmg is press-fitted onto the tubing adapter of the needle assembly, thus sealing the cold water within the tube The water-Jacket tubing is transparent, to allow for the visual momtormg of the advancement of the core solution m the polyethylene tube

M/croencapsulation

with HEMA-MMA

b The end of the polyethylene tubing that protrudes from the water-Jacket tubmg IS attached to the core syringe Subsequently, the filled core syrmge, the portion of the polyethylene tubing that protrudes from the water-Jacket tubing, and the tip of the water-jacket tubing are immersed m an Ice-water bath c The water-Jacket tubing IS equipped with a sidearm, near the needle assembly tubing adapter, that IS also connected to the Ice-water bath by a pump. Once the core syringe IS tmmersed m the water bath, this pump IS started, to clrculate the cold water at a flow rate of -100 mL/mm, around the polyethylene tubing and the portion of the core needle wlthm the water-Jacket tubing d Once the perlstaltlc pump IS started to circulate the dodecane (hexadecane cannot be used here), the 500-mL bottle used as dampener IS Immersed m dry Ice Simultaneously, the heater of the magnetic stirrer IS set to a low setting By allowmg the system to reach a steady state for approx 1 h, the temperature of the dodecane wlthm the nozzle may be maintained below lO”C, the preclpltatlon bath IS at -30°C e When the peristaltic pump IS stopped at the end of encapsulation, the dampener IS removed from the dry ice m order to prevent freezing of the dodecane After removing the needle assembly from the receivmg dish, the water-Jacket tube IS emptied pnor to dlsconnectlon of all tubings from the needle assembly.

4. Notes 1. Some of the modlficatlons made to the needle assembly in our current standard mlcroencapsulatlon setup may also be applied to the setup that utlllzes osclllatmg coaxial needles A metalhc, tapered outer needle, and a tapered inner needle that has centering knobs, may be used. 2 The vertical posltlonmg of the coaxial needles withm the nozzle affects the shear rate that the droplets experience, and thus the rate of production and size of the capsules The closer the needles are to the tip of the nozzle, the smaller the capsules and the higher the productlon rate Care must be exercised m attaching the tubmgs to the needle assembly, m order to avoid altermg the protrusion of the core needle and vertical posltlon of the needle assembly (Fig. 7) A key feature IS keeping the needles centered relative to each other Also, nozzles with varying angles may be used to alter the shear force experienced by the droplets. The typlcal nozzle angle is 20 degrees 3 For a thorough cleanmg of the coaxial needles, the core needle IS removed from the assembly through the Luer-Lok tubing adapter, and a mdd soap solution IS aspirated through the needles Any partlculates that may have adhered to the 2mner wall of the core needle are dislodged with a 24-gage needle that IS -6 m. m length. The needles are then rmsed with water and ethanol, and air-dried 4. The polyethylene and Sllastlc tubings must be of adequate length to provide flexlblhty m maneuvermg the needle assembly durmg the setup 5 The mterfaclal tension of a 0 0 1% aqueous solution of Pluromc L 10 1 vs mmeral 011(Nujol) is 3.5 mJ/m2 (BASF). Pluromc LlOl attams its highest surface actlvity at its cloud pomt The cloud point for a 0 0 1% aqueous solution IS - 18°C The

346

6

7

8

9

10

Lahooti and Sefton preclpltatlon bath may be mamtamed at the LlOl cloud point by lmmersmg the recelvmg dish m a cool-water bath; however, the overlying hexadecane layer may freeze If the PBS preclpitatlon bath IS excessively cooled To prevent the presence of air bubbles m the hexadecane portlon of the needle assembly, only the dampener 1s filled mltlally with hexadecane from the recelvmg dish. Then, the pump 1srun m the reverse duechon to fill the needle assembly casing with hexadecane directly from the receiving dish If the mltlal polymer precipitate 1s not dislodged by the hexadecane flow rate, intervention with a bent needle that 1sinserted inside the nozzle 1srequired However, contact with the coaxial needles should be avoided Immediately after preclpltatlon, the polymer tends to be sticky Any prolonged contact between the capsules at this point ~111mevltably lead to agglomeration, whtch IS extremely difficult to reverse The tendency of the capsules to adhere to one another continues for days after precipitation, although with a reduced affinity Capsules mcubated with serum-free medium m a Petri dish gently attach to the dish and hence do not agglomerate In the first week postencapsulation, the capsules experience a 2030% reduction m overall size, become more elastic and tougher, and also, the uutlal white, opaque appearance of the capsule membrane becomes more transparent The tubes used for preparation of the collagen solution should be maintained on ice, m order to prevent premature preclpltatlon of the collagen Any precipitates that are formed should be removed, because they may disrupt the flow of the core solution These precipitates have a tendency to rise to the top of the collagen solution, and may be aspirated easily Depending on the ambient temperature and the concentration of the extracellular matrix component used, It may be sufficient to immerse the dampener for the dodecane pump in an ice-water bath, rather than m dry rce Because of the heat exchange between the dodecane overlayer and the PBS preclpltatlon bath m the receiving dish, the setting on the heater may need to be adJusted accordingly

Acknowledgments Financial support was provided by the Natural Sciences and Engineering Research Council and the Medical Research Council of Canada The technical assistance of Vlad Horvath m deslgnmg and fabrlcatmg the needle assembly and the protocol 1s greatly appreciated.

References 1. Lanza, R. P , Kuhtrelber, W M , and Chick, W L. (1995) Encapsulation technologies. Essue Eng 1, 181-196. 2 Colton, C K (1995) Implantable blohybrld artificial organs Cell Transplantatzon 4,415-436

3. Lanza, R P , Hayes, J. L , and Chick, W. L (1996) Encapsulated cell technology Nature Bzotechnol 14,1107-l 111

Microencapsulation

with HEMA-MMA

347

4 Babensee, J E. and Sefton, M V (1996) Protein delivery by mtcroencapsulated cells, m Controlled Drug Delwery Challenges and Strategzes (Park, K , ed ), Amerrcan Chemrcal Socrety, Washmgton, DC, pp 3 1l-332 5 Sefton, M V and Stevenson, W T. K (1993) Mtcroencapsulatron of hve animal cells using polyacrylates, Adv Poly Scz 107, 143-197 6 Uludag, H , Kharhp, L , and Sefton, M V (1993) Protein delivery by mrcroencapsulated cells. Adv Drug Del Rev 10, 115-130 7 Uludag, H , Horvath, V, Black, J P , and Sefton, M. V (1994) Vrabrhty and protein secretion from human hepatoma (HepG2) cells encapsulated m 400 pm polyacrylate mlcrocapsules by submerged nozzle-liquid Jet extrusron, Blotechnol Bloeng 44, 11991204 8 Sefton, M V, Uludag, H , Babensee, J , Roberts, T, Horvath, V, and De Born, U. (1994) Mrcroencapsulatron of cells m a thermoplasttc copolymer (hydroxyethyl meythacrylate-methyl methacrylate) Methods Neuroscz 2 1, 3 7 1-3 86 9 Hwang, J R and Sefton, M V (1995) Effect of mrcrocapsule diameter on the permeabtltty to horseradish peroxtdase of mdtvtdual HEMA-MMA mlcrocapsules .Journal of Controlled Release 33,273-283 10 Stevenson, W T K , Evangelista, R A , Broughton, R L , and Sefton, M. V (1987) Preparation and characterlzatron of thermoplastic polymers from hydroxyalkyl methacrylates Journal of Applied Polymer Science 34, 65-83 11 Stevenson, W T K and Sefton, M V (1988) The equthbrrum water content of some thermoplastrc hydroxyalkyl methacrylate polymers Jurnal ofApplzedPolymer Science 36,1541-1553 12 Hwang, J R and Sefton, M V (1995) The effects of polymer concentratron and a pore formm agent (PVP) on HEMA-MMA microcapsule structure and permeabdrty. J Membrane Scl 108,257-268 13 Crooks, C A , Douglas, J A , Broughton, R L , and Sefton, M V (1990) Mlcroencapsulatron of mammahan cells m a HEMA-MMA copolymer effects on capsule morphology and permeabrhty J Blamed Mater Res 24, 1241-1262 14 Uludag, H and Sefion, M V (1993) Metabohc actrvrty and proliferation of CHO cells m hydroxyethylmethacrylate-methylmethacrylate (HEMA-MMA) mrcrocapsules Cell Transpluntatzon 2, 175-l 82 15. Roberts, T , De Born, U , and Sefton, M V (1996) Dopamme secretron by PC12 cells mtcroencapsulated m a hydroxyethyl methacrylate-methyl methacrylate copolymer Blomaterlals 17,267-276 16 Babensee, J. E , De Born, U , and Sefton, M V (1992) Morphologrcal assessment of hepatoma cells (HepG2) mlcroencapsulated m a HEMA-MMA copolymer wrth and without Matrtgel J Boomed Mater Res 26, 1401-1408. 17 Uludag, H and Sefton, M V (1993) Mtcroencapsulated human hepatoma cells m vtro growth and protein release J Blamed Mater Res 27, 12 13-1224 18. Ung, D Y-P (1993) The effect of cytokmes on mrcroencapsulated mammalian model cells, Bachelor of Applied Science Thesis, Umversity of Toronto 19 Wells, G D. M ) Fisher, M M , and SefIon, M. V. (1993) Mlcroencapsulatton of viable hepatocytes III HEMA-MMA micmcapsules. a prehrmnaty study. Biomaterzals 14,615-620

348

Lahooti

and Sefion

20 Sefton, M V and Kharlip, L (1994) Insulm release from rat pancreatic islets mlcroencapsulated m a HEMA-MMA polyacrylate, m Pancreatzc Islet Transplantation, Vol III, lmmunolsolatlon of Pancreatic Islets (Lanza, R P and Chick, W. L , eds.), R G Landes, Austin, TX, pp. 107-l 17. 2 1 Tse, M , Uludag, H , Sefton, M. V , and Chang, P. L (1996) Secretion of recombinant proteins from hydroxyethyl methacrylate-methyl methacrylate capsules, Blotechnol Bloeng 51,27 l-280. 22 Uludag, H and Sefton, M V (1990) A calorimetric assay for cellular activity m microcapsules Blomaterlals 11,708-7 12 23 Rodriguez, F R (1994) Engmeermg of a microencapsulation system, Master of Applied Science Thesis, University of Toronto

27 Micropatterning

Cells in Tissue Engineering

Sangeeta N. Bhatia, Martin L. Yarmush, and Mehmet Toner 1. Introduction Recent advances m tissue engineering have been facihtated by the ability to control the environment of cells in vitro Modulation of cell-extracellular matrix, cell-substrate, and cell-cytokine interacttons have proven to be useful m organotypic cultures of many kmds. In some cases,such as skin, bone marrow, and liver, the recovery of physiologic tissue function is enhanced by co-cultivation of two or more cell types together. The m vitro dependence of tissue function on cell-cell interactions is reminiscent of cellular cues during embryogenesis, or adult mteractions between parenchymal cells and supporting stroma. In vitro attempts to study co-cultures of two cell types have typically assessedthe influence of nonparenchymal cell populations on parenchymal cells by variations in cell seeding density or addition of excised tissue or confluent coverslips to existing cultures Alternatively, physical separation of cell cultures through use of conditioned media (1,2) or porous filter inserts (3) has been utilized. In addition, dynamic cell-cell mteraction has been studied in monolayers of a primary cell type m the presence of a shearing fluid containing a secondary cell type (45) One hmitation of these co-culture systems is the inability to vary local cell seeding density independently of the cell number. Micropatterning technology, or the ability to spatially control cell placement at the single-cell level, allows the precise manipulation of cell-cell interactions of interest. Microfabrication techniques have been widely utilized for the spatial control of cells in culture (6-15). Many strategies have employed variations in charge (16,17), hydrophiliclty (11,18), and topology (11,19,20) to mediate From

Methods

III Molecular

Edlted by J R Morgan

Medmne,

Vol

18

and M L Yarmush

349

Tk%sue Engmeenng

0 Humana

Methods

and

Press Inc , Totowa,

Protocols

NJ

350 selective adhesion of one cell type by differential serum-protein adsorption, or variations m surface free energy In addition, specrtic use of biomolecules alone (21,22), or m conJunctton with aminosilanes (IO,I3) or self-assembled monolayers (23), have been used to micropattern cells. A variety of cell types have been exammed with micropatternmg techniques, such as neuroblastoma cells (18), BHK epitheltal cells (II), hepatocytes (23-25), and myocytes (9), wtth spatial resolution on the micron scale. These studies examined a wide array of physiologic functions, such as neuronal growth cone guidance, effects of cell shape on functron, and electrical couplmg through gap Junctions. However, these methods have not been adapted to the simultaneous co-cultivation of more than one cell type. In this chapter, we describe a versatile method for generating two-dimensional, anisotropic, model surfaces capable of orgamzmg a single cell type or two different cell types m discrete spatial locations. We have chosen a primary rat hepatocyte/3T3 fibroblast cell system because of its potential sigmficance m both baste sctence and technology development, and based on widely reported interactions observed m this co-culture model (1,2,26,27) We have used photolithography, existmg strategies for surface modification of glass substrates with ammosilanes linked to biomolecules, and mampulatton of serum content of cell culture media to pattern cells The first cell type, hepatocytes, attach and spread on an immobihzed collagen I pattern; the second cell type, 3T3 fibroblasts, undergoes nonspecific, serum-mediated attachment to the remammg unmodified areas This co-culture technique allows the mampulation of the mitral cellular mtcroenvnonment wtthout vartatton of adhered cell number, thereby allowing measurement of the influence of local variations m cell-cell mteraction on bulk tissue function.

2. Materials 2.1. Microfabrication

of Substrates

1 Borosihcate substrates* 2 in diameter x 0 02 in thickness float glass wafers are custom made (Erie Scientific, Portsmouth, NH) to fit a standard 60-mm diameter (P-60) tissue-culture dish 2 Chrome masks: 5 in x 5 in chrome masks are fabricated by a high precision photolithographic process from artwork Artwork is converted by computer-aided design packages (1 e., KIK, Berkeley, CA) to machine language (GDS) to generate an optical pattern (pattern generator) on a chrome-evaporated-photoresist coated glass or quartz plate (Advanced Reproducttons, N. Andover, MA) Alternatively, some university microfabrication facilities complement their wafer microfabrication with mask-making facilities (Microsystems Technology Laboratory, Massachusetts Institute of Technology, Cambridge, MA) Another mexpensive option for relatively large feature size (> 100 pm) IS reduction of artwork

357

Mkzropatterning Cells

onto microfiche usmg standard library services Use 2 mIcrofiche masks on top of one another to prevent imaging of microfiche defects. These microfiche masks are fairly fragile and are primarily useful for exploratory studies 3 Cleanmg solution (piranha solution) 3: 1 mixture of H$O, 30% H,O, prepared at time of experiment, Caution: Extreme caution must be observed during chemlcal handling because of the corrosive nature of the chemicals m use Pyrex contamers must be utlhzed and chemical removal should be performed via asplrahon and dllutlon of acid mixture 4. PhotoresIst. Theoretically, any photoresist can be utlhzed that will adhere sufficlently to untreated boroslhcate, remam mtact during subsequent processing, and yet be removed with relative ease after surface modification. We use a positive photoreslst, OCG 825-835 centlstokes (Olin-Cuba-Geigy, West Paterson, NJ) 5 Developer* Use developer appropriate for selected photoreslst. We use OCG 934 1 1 MIF (Olin-Cuba-Geigy) 2.2. Surface

Modification

of Substrates

1 3-[2-(ammoethyl)ammo] propyl trlmethoxysilane (AS) AS solution IS 2% AS (Aldrich, Milwaukee, WI) m water, prepared Immediately prior to use AS should be stored m a nitrogen box, or bottle should be gassed with mtrogen prior to closure and storage 2 Phosphate buffered salme (PBS) Can be purchased at lo-fold concentration (Blofluids, Rockvllle, MD). 3 Glutaraldehyde. Stock solution 1s 25% glutaraldehyde m water (Fisher Chemlcal, Fair Lawn, NJ) Glutaraldehyde solution is prepared by dllutmg stock solution in PBS to a final concentration of 2 5% v/v, pH 7 4. 4 Collagen solution* Stock collagen I solution 1sprepared from rat tad tendons to a concentration of approx 1 mg/mL m 1 mM HCl (described m Chapter 20 of this volume) Collagen I solution 1sprepared by 1: 1 dilution of stock solution (final concentration is approx 500 pg/mL) with water, pH 5.0, prepared unmedlately prior to use Alternatively, commercial preparations of collagen (or other desired adhesive proteins) can be substituted for the collagen solution described here 5 Acetone: General laboratory grade (Baxter Dlagnostlcs, Deerfield, IL) 6 Ethanol. 70% in dHzO.

2.3. Cell Culture 2 3.7. Fibroblast Culture 1 Flbroblast feeder layer 3T3-52 mouse fibroblast cell lme (originally provided by H. Green, Department of Physiology and Biophysics, Harvard Medlcal School, Boston, MA) 2 Flbroblast medium (FM)* Dulbecco’s mochfied Eagle’s medium (DMEM, high glucose, L-glutamme, 110 mg/L sodium pyruvate, Gibco-BRL, Grand Island, NY), bovine calf serum 10% (BCS, HyClone, Logan, UT), peniclllm-streptomycm (Boehrmger, Indlanapohs, IN) 100 IU/mL-100 &mL

352

Bhatla, Yarmush, and Toner

3 Mltomycm C (Boehrmger, Indlanopohs, IN)* 10 pg/mL m fibroblast medium 4 Trypsm. 0 1% trypsm (ICN Biomedlcals) m PBS 5 Versene. 0 1% EDTA (Boehrmger) m PBS

2.3.2. Hepatocyte isolation and Culture 1 Isolation of hepatocytes from rat liver Isolation of rat hepatocytes by collagenase perfusion 1s described m detail m Chapter 20 of this volume Alternatively, isolated, purified rat hepatocytes may be obtained from another laboratory 2 Serum-free hepatocyte culture medium (SFM) DMEM with high glucose, without sodium pyruvate, without sodium blcarbonate (Glbco-BRL, # 12 100) Glucagon (Lilly, Indlanapohs, IN). Add to final concentration of 7 ng/mL Insulm (Squibb, Princeton, NJ) Add to final concentration of 0 5 U/mL Epldermal growth factor (Becton Dlckmson, Bedford, MA) Add to a final concentration of 20 ng/mL Hydrocortlsone (UpJohn, Kalamazoo, MI) Add to a final concentrataion of 7 5 E,irnL Pemclllm-streptomycm, respectively, 5000 U/mL and 5 mg/mL, m 0.9% NaCl, (Sigma, St Louis, MO). Add 1% v/v to medmm L-prolme Add to final concentration of 40 &mL (Sigma) (16) Filter-stenhze medium through a 0 2-p pore membrane.

2.4. Cell Culture on Modified

Substrates

1 Culture medmm (CM) Identical to SFM, except for exclusion of L-proline and mcluslon of 10% fetal bovine serum (FBS, JRH Bioscience, Lenexa, KS) 2 Bovine serum albumin (BSA): BSA solution of 0.05% w/w m water (Sigma) Stenhze by filtration through a 0 45-p filter, and store at 4’C 3 Sterilized deionized water Prepare two autoclavable containers of 500 mL of deionized water Autoclave to sterilize 4 Stenlized vessels. Prepare two glass beakers by covering with autoclave wrap Autoclave

3. Methods Mlcrofabrlcation techniques are used to modify glass substrates with blomolecules. These modified substrates are utilized to pattern a single cell type, or to mtcropattern two cell types to control their level of mteractlon. Figure 1 schematically depicts the overall process for one representative pattern, although this procedure can be utilized for a variety of spatial configurations. 3.1. Micro fabrication of Substrates We produced experimental substrates utlhzmg standard mlcrofabrlcatlon techniques at Microsystems Technology Lab, MIT, Cambridge, MA. This mlcrofabrication facility, like many others of its kind, IS dedicated primarily to

Micropa tterning Cells

on Glass Wafer

353

Photoresist

AS + G + protein

Removal Photi;esist

D

F + Serum

Remove Unattached Cell A

Cell A

Fig. 1. Schematic of process for generating micropatterned cultures of a single cell type or two distinct cell types.

silicon wafer processing for electronic applications. As a result, the presence of potential contaminants, such as class III-V compounds, precious metals, alkali, and transition metals, are restricted. Therefore, the process described here utilizes separate facilities for surface fabrication and surface modification with biomolecules. In general, almost any integrated circuits manufacturing facilities could be used to perform these procedures. 1. Borosilicate wafers are cleaned by placement into wafer carriers. Place into a Pyrex vat, pour piranha solution over wafers, and wait 10 min. Rinse wafers 3X in a dump-to-resistivity tank, and nitrogen dry using a spin-dryer. In cases in which a spin-dryer is not available, manual drying is done by using a Nz gas stream. This gas stream can either be accessed through a house N2 gun, or a portable N2 gas tank connected to tubing. 2. Dehydrate wafers promote adheszion of photoresist by baking for 60 min at 200°C. 3. Mount wafers on vacuum of a spin-coater chuck and coat with positive photoresist to a uniform layer of approx 1 pm. In our laboratory, this is accomplished as follows: Dispense photoresist at 500 rpm for 2 s, spread photoresist at 750 rpm for 6 s, and spin at 4000 rpm for 30 s (step A, Fig. 1). 4. Soft-bake for 30 min at 90°C to drive out excess solvent and anneal any stress in the film.

Bhatla, Yarmush, and Toner 5 In order to create a latent image of the desired pattern m the resist layer, expose wafers to ultravlolet hght m a defined pattern We expose coated substrates to 365 nm UV hght m a Bottom Side Mask Ahgner (Karl Suss, Munich, Germany) through a 5 m. x 5 tn. patterned chrome mask under vacuum-enhanced contact for 3 s, at a dose of 10W/cm2 6 To produce the final three-dlmenslonal rehef Image, lmmerslon develop the exposed photoreslst m the appropriate developer Complete removal of photoreslst m exposed areas 1scntlcal to subsequent surface modlficatlon Furthermore, residual photoreslst IS not easily detected on transparent (boroslhcate) wafers, therefore, care should be taken to optlmlze this step of the protocol Presence of residual photoresist can be assessedby mspectlon under light interference or fluorescent mlcroscopy We develop by lmmerslon and agltatlon m a bath of developer for 70 s Surfaces should then be rinsed three times under running delomzed water and cascade rmsed (If possible) for 2 mm (step B, Fig. 1) 7 Postbake patterned disks for 30 mm at 120°C to drive off residual solvent and promote film adhesion 8 Wafers can be stored m closed contamers (preferably wlthm a nitrogen box) at room temperature for at least 1 mo 9 Fmally, 24 h prior to surface modification, substrates can be exposed to oxygen plasma m order to dry-etch (remove) a small layer of photoreslst This ensures complete removal of photoresist from exposed boroslhcate, however, if substrates are well developed and pattern dlmenslons are larger than 10 w, this step may be skipped We use a parallel-plate, plasma day etcher at a base vacuum of 50 mTorr m an 0, atmosphere, and pressure of 100 mTorr at a power of 1OOW for 2-4 mm, which corresponds to an etch rate of approx 0.1 pm/mm

3.2. Surface Modification

of Substrates

There are a number of techniques for lmmoblhzmg adhesive proteins on solid substrates (see ref. 28). Here we describe a technique modified from experimental methods developed by Lom et al. (13) and Brltland et al. (10) (step C, Fig. 1). This sllane-based coupling has advantages over thlol-based techniques (19) for patterning more than one cell type, because of the availablllty of the unmodified glass for attachment of a second cell type. In contrast, m some thiol-based techniques, outlymg regions are modified with a nonadheslve compound (polyethylene glycol [PEG]), preventmg attachment of a second cell type. 1 Patterned substrates are handled with wafer tweezers (Fluoroware, Chaska, MN) Pay special attention to the onentatlon ofthe patterned surface of the wafer, transparent substrates can be eastly Inverted without any obvious differences rn appearance. Wafers are rinsed by nnmersion m delomzed dH20 m a glass lOOmm Petri dish Repeat 2 Silane munoblhzatlon of AS IS performed by lmmerslon of samples m AS solutlon for 30 s at room temperature, followed by two rinses m deionized water

Micropa tterning Cells

Fig. 2. Differential

355

hydrophilicity

of patterned glass surface.

3. Dry wafers with a stream of nitrogen gas to avoid drying artifacts. 4. Bake wafers in a closed container for 10 min at 120°C. Note: Temperature greater than 150°C will cause hardening of many photoresists and cause difficulty during removal. 5. Next, soak disks in a covered container of 2.5% v/v solution of glutaraldehyde in PBS (pH 7.4) for 1 h at 25’C, followed by two rinses in fresh PBS. Visually inspect wafers every 15 min to evaluate the integrity of the photoresist. In some instances, the glutaraldehyde solution can cause peeling of the photoresist and the process may need to be abbreviated. 6. To immobilize collagen, immerse wafers in 4 mL of collagen solution in a P-60 Petri dish for 30 min at 37°C. 7. To remove photoresist and expose underlying unmodified glass, float each wafer in acetone in a glass container and sonicate the container in a bath sonicator for l-l 5 min. The duration of sonication is empirically determined by observation of the first wafer in each batch. Examine wafers for complete removal of photoresist (previously pink wafers will appear clear). Treat all wafers in a experimental batch identically to ensure comparability of immobilized protein layers on all substrates. We use 10 mL of acetone in a loo-mm glass Petri dish (step D, Fig. 1). 8. Rinse wafers twice by immersion into deionized water. As shown by Lom et al. (13) and others, modified areasshould display differential wetting upon removal of substratefrom water, thereby indicating successfulpatterned-surface modification (Fig. 2). 9. Wafers can be stored dry, in a covered container, at 4°C for at least 2 wk. We store wafers on a piece of filter paper (to absorbresidual water and prevent sticking) in

356

Bhatia, Yarmush, and Toner 60-mm Petri dishes. Storage of wafers m solution (i e , PBS or ethanol) results m transfer ofpatterned protein to unmodified areas, presumably via desorptlon from modified areas and adsorption to unmodified areas, therefore, If the lmmoblllzed protein will tolerate unhydrated storage conditions, wafers should be stored dry If munoblllzed protein requires hydration to mamtam its bloactlvlty, storage time must be empIrIcally determined-m our case, less than 48 h

3.3. Cell Culture Co-cultivation of two cell types has been proven to be useful m organotyplc cultures of many kinds. skin, vasculature, bone marrow, as well as liver. Hepatocytes have been co-cultured with many secondary cell types, ranging from isolated blliary ductal eplthiha (29) to cell hnes such as murme 3T3 fibroblasts (30) These co-cultures have been reported to sustain hepatocyte morphology and viability, form functlonal homotypic gap Junctions, and produce elevated levels of various liver-specific markers, mcludmg albumin synthesis and vanous detoxification functions (29,30) We chose co-culture of primary rat hepatocytes with 3T3-52 fibroblasts, because of their easeof culture and avallablllty, although genetlcally modified fibroblasts or alternative cell types could also be Incorporated into this methodology 3 3 1 Fibroblast Culture 1 Grow 3T3-52 cells to preconfluence m 150-cm2 flasks or other sultable dishes, 10% COZ, balance au 2 Passage cells by washing with 8 mL of versene, and incubating with 2 mL of trypsm solution and 2 mL of versene, for -5 min at 37°C Resuspend cells m 25 mL me&a to neutralize trypsm Approximately 10% of the cells may be inoculated mto a fresh tissue-culture flask contammg a total of 50 mL of media. Cells may be passaged no more than 12 times. 3 A smgle 150-cm2 flask will yield 4-8 x lo6 cells at confluence Time the passage and number of flasks so that a sufficient cell number will be available at preconfluence on the day of the expenment-in our case, 1 d after hepatocyte lsolatlon

3 3.2. Hepatocyte Isolation and Culture 1 Isolation of hepatocytes from rat hver Isolation of rat hepatocytes by collagenase perfusion IS described m detail m Chapter 20 of this volume. Hepatocytes are isolated from 2- to 3-mo-old adult female Lewis rats welghmg 180-200 g, by a modified procedure of Seglen (1976). Routmely, 200-300 mllllon cells can be isolated, with vlablhty between 85 and 95%, as Judged by trypan blue exclusion Nonparenchymal cells, as Judged by their size (~10 pm m diameter) and morphology (nonpolygonal or stellate), are typlcally less than 1%. 2 Assess cell concentration of stock suspension by hemocytometer and record 3 Store hepatocytes on ice prior to use In our hands, rat hepatocytes are viable and retam the capacity to attach and spread for at least 6 h on ice

Micropatterning

Cells

357

4. Dtlute cell stock suspension m SFM to appropriate concentration (see Note 4) In our experiments, this concentration IS l-2 x lo6 hepatocytes/mL. Prepare sufficient volume to perform multiple seedings for each experimental condition. In our experiments, we prepare 6-8 mL of cell suspension for each dish, which corresponds to 3-4 sequential seedings of 2 mL on each dish

3.4. Cell Culture on Modified Substrates Coordmatton of surface preparation, culture of prtmary cells, and culture of a cell lme requires precise timing. Smce isolated cells, such as hepatocytes, must be seeded wrthm a few hours of isolatron, surface preparation requu-ed on the day of cell seeding must be completed prior to receipt of cells from isolation. Stmilarly, the tibroblast populatton should have been expanded to suffrcient cell numbers for trypsmizatron the following day. 1. Soak premodified wafer m 70% ethanol for sterilization. Use wafer tweezers to place wafer m P-60 with 5 mL of ethanol solution for at least 1 h, but not more than 24 h, at room temperature m a sterile hood (step E, Fig. 1). 2 In older to remove residual ethanol from surface of wafer, first pour autoclaved water into sterilized beaker. Flame sterilize wafer tweezers, allow to cool, and remove wafer from 70% ethanol under sterile conditions. Immerse wafer m water and agitate gently for approx 10 s, being sure to preserve orientation of the wafer 3 In order to deter nonspecific attachment of hepatocytes on unmodified areas, the sterilized wafers will be coated with bovine serum albumm (BSA), a large, globular, negatively-charged protein thought to deter cell adhesion by nonspecific charge mteractions for many cell types, however, the usefulness of this coatmg IS cell-type dependent, and may vary (14) For rat hepatocytes, BSA coating reduces nonspecific cell attachment to glass from 30% to negligible levels (25) Place sterilized wafers m sterile P-60 dishes, add 4 mL of BSA solution to each dish, and place m incubator for 45 min at 37°C. 4 To remove residual BSA solution, use sterile tweezers to remove wafers from dishes under sterile conditions, and immerse m autoclaved water m sterile beaker, gently agitating for 10 s Note. At this point, if one needs to extend the duration of the sample preparation period, substrates may be left under sterile conditions m dry Petri dishes 5 Substrates are then rinsed with SFM to remove residual water Add 3 mL of media to each dish, gently swirl solution, and aspirate. Repeat 6. Seed hepatocytes by placing 2 mL of hepatocyte solution (m SFM) on each wafer The underside of the wafers should still contam fluid from the SFM rinse, if not, add 0 2 mL of SFM to the dish to wet the underside of the wafer, to prevent wickmg of cell suspension underneath the wafer Agitate solution to disperse cell suspension, and place in incubator for l-l.5 h at 37°C 10% C02, balance air (step E, Fig. 1). Wafers should be periodically agitated (1-e , every 15 mm) to promote maximal cell attachment. 7 At this pomt, pattern features should be visible because of selective cell adhesion Typically, to ensure 100% confluence on patterned areas, hepatocyte seed-

Bhatia, Yarmosh, and Toner

Fig. 3. Phase-contrast micrographs of micropatterned hepatocytes. Hepatocytes attached to linear strips of width (A) 20 pm, (B) 50 pm, and (C) 200 pm, and circular patterns of diameter (D) 50 pm, (E) 100 pm, and (F) 250 pm. ing is repeated 2-3X. This is especially critical in cases of low cell viability, since dead cells render some modified sites inaccessible to live cells. Surfaces should be rinsed twice by pipeting and then aspirating 4 mL of serum-free media, reseeded with hepatocytes for 1.5 h, and rinsed again. Repeat as necessary (step F, Fig. 1). After the final rinse, incubate substrates in 2 mL of CM. 8. Hepatocytes are then allowed to spread over the remaining modified sites. Rat hepatocytes spread over greater than 10 h (32); therefore, incubate patterned hepatocytes overnight. Cells (hepatocytes, endothelial cells) on small pattern features will confirm to the edges of the pattern (Fig. 3; 33,349. Experimental studies can be performed using these micropatterned hepatocytes; otherwise, proceed with addition of second cell type. 9. The day after hepatocyte seeding, trypsinize the appropriate number of flasks of 3T3-J2s, centrifuge cells at 1000 ‘pm for 5 min, and resuspend in a small volume

Micropatterning

Ceils

Fig. 4. Phase-contrast micrograph of micropatterned co-cultures with varying spatial configurations. (A) 20091rn strips of hepatocytes with 500~unr lane of intervening fibroblasts. (B) 100~l.nn islands of hepatocytes with 250~pm center-to-center spacing and intervening Iibroblasts.

of flbroblast media (FM). Count with a hemocytometer and plate appropriate number of fibroblasts in 3 mL of FM per dish. We typically utilize 750,000 fibroblasts per dish; however, in some cases it is necessary to plate growth-arrested fibroblasts in greater numbers (step G, Fig. 1). (Growth-arrest fibroblasts by incubating each 150-cm* fibroblast flask with 15 mL of mitomycin C solution for 2 h, rinse with media, and trypsinize as usual. Incubate for at least 6 h to allow fibroblast spreading, aspirate media, and replace with 2 mL of CM. 10. Resultant micropatterned co-cultures of two distinct patterns are shown in Fig. 4. Proceed with experimental investigation. Pattern preservation is dependent on cell types, as well as pattern dimensions; for example, our patterns are stable on the order of weeks for collagen-modified areas of larger than a few hundred microns.

4. Notes 1. Microfabrication process: Facility regulations will largely determine the process utilized for preparation of patterned surfaces. Here, we have detailed the most versatile technique, whereby all biological materials are utilized subsequently in a separate facility. Others have pretreated wafers with organic materials in their own facilities, to allow for greater versatility (1.3). This approach should also be considered where applicable. Facility specifications regarding dedicated equipment will also affect problems that may be encountered via contaminants, and so on (see Note 6). In addition, the overall process will vary with the characteristics of the desired protein, In our case, collagen I retained its bioactivity for hepatocyte attachment and spreading, despite treatment with acetone, ethanol, and dehydration, Other proteins may require modifications of this protocol to retain their bioactivity.

Bhatla, Yarmush, and Toner 2

Surface moditicatton The method described here was Intended to be robust In many instances, users may be able to eliminate certam elements of this process Exposure to plasma oxygen may be unnecessary, tf patterns are well-developed and not contaminated durmg storage Stmtlarly, bakmg of wafers after AS modificatton may be unnecessary m some protocols In some cases, covalent bmdmg of protein to glass may be elimmated entirely by simply adsorbing protein to exposed glass For example, collagen and some other protems are known to irreversibly adsorb to certain substrates, however, tt should be noted that reliance on adsorption alone is likely to result m different immobtltzed protem conformattons than covalently bound proteins 3 Cell-culture issues Cell-culture issues involved wtth mtcropatterned cultures melude cell-seeding protocols for hepatocytes, cell seeding protocols for fibroblasts, media formulation for co-cultures, and control of cell populatton Our seeding protocol for rat hepatocytes (2-4 x lo6 per dish, with agttatton every 15 mm for l-l .5 h, followed by 2 repeated seedmgs, which yields approx 250,000 attached hepatocytes) may be modttied, depending on cell source, duratton of vtabthty m suspension, and cell aggregation at htgh denstttes It is important to note that the photoresist pattern determines available wafer surface area for hepatocyte attachment Since hepatocytes are seeding m excess, the number of attached hepatocytes IS ultimately determined by the mask design Ftbroblast seeding protocols include spectficattons for cell number, passage restrtcttons, and ttmmg of seeding (m our case, 750,000 per dish fibroblasts are seeded, before the twelfth passage, m a single seeding, and allowed to attach for 6 h m fibroblast media). Ftbroblasts attach to unmodified glass, and, therefore, m contrast to hepatocyte number, fibroblast number IS dictated by the number of viable cells seeded per substrate Finally, selection of media for two cell types must be considered Because of the rapid mttotic rate of fibroblasts, our media formulatton contains high levels of hydrocorttsone as a growth mhtbitor, each type of culture will requtre some degree of media formulation 4 Probing tissue function of mtcropatterned cultures Cultures of two different cell types often require umque analysis For example, selective markers of cell function (m our case, albumm or urea syntheses as markers of liver-specttic function of hepatocytes), methods to quantttate relative growth of both cell populattons (m our case, growth-arrestmg fibroblasts allowed measurement of total DNA as a vehicle for trackmg both cell populattons), spatial trackmg of both cell populattons (m our case, either tmmunofluorescent stammg of cytokeratm, an mtermedtate filament m hepatocytes as compared to F-actm, a cytoskeletal polymer present m both cell type, or labeling of each cell population with a dtfferent longhved Intracellular fluorescent probe, such as Molecular Probes [Eugene, OR] Cell Tracker dyes) 5. Strategies for achieving selective cell adheston of different cell types For patterning of two different cell types, the general approach 1s to pattern an adhesive substrate on a background that IS both nonadhesive to one cell type, and adhestve to a second cell type The sequence of seeding two distmct cell populattons can be

Micropatterning

Cells

crttrcal to this process For example, hepatocytes must be patterned prror to tibroblast seeding, hepatocytes would attach and spread umformly on the surface of a fibroblast patterned wafer Other variables that can influence selecttve cell adhesion include the kmettcs of attachment (we allow 1 5 h hepatocyte attachment), the influence of hgand density (we use saturated levels), the affinity of different hgands (we use collagen I), the influence of charged coatings (BSA), and denaturatton of adsorbed proteins to remove nnmobrhzed protein (8 A4 urea [Brrtland]) 6 Troubleshootmg Common problems include mn-nmal cell adhesion to substrate, uniform cell adhesion to substrate, difficulty with lift-off of photorestst, and peelmg of photoresrst during processmg. Surface wetting (Fig. 2) IS often an tmportant clue m targeting the source of difficulty, since a lack of pattern at that point indicates a problem upstream of sterthzatton and cell seeding. a Lack of cell adhesion to the substrate can indicate many problems. The most common, however, ISunderdevelopment of the exposed photorestst, resulting m a lack of exposed borosthcate for protein nnmobtlizatton. This can be allevtated by mcreasmg development tune, or mcreasmg exposure to oxygen plasma pnor to surfacemod& canon This effect 1soften exacerbated with small (6 pm) pattern dtmenstons Alternatrvely, contaminants that coat the glass and prevent protem ammobtltzation, or commonly utilized undercoatings for promotmg photorestst adhesion, could produce this effect b. Umform cell adhesion A lack of discermble pattern because of exuberant cell adhesion IS frequently caused by a defect m the ongmal photorestst coating Photoresist can degrade and crack over time, allowing exposure of all areas of the substrate to surface modrficatron Alternatively, excessive exposure to oxygen plasma will strip the photorestst from the surface of the wafer completely c Peeling of photo during surface modrficatton The mtegnty of photoresist coatings varies with solvents In some instances, photoresrst ~111peel away from the wafer surface prematurely This mdtcates either msufftctent adhesion of the photoresist to glass (often because of an msufftcrent dehydration bake pnor to photoresist coating) or insufficient baking after development to harden the photoresist d Dtfficulty wrth photorestst removal Extended somcatton in acetone for photoresist removal often mdrcates exposure of photoresist to elevated temperatures

Acknowledgments This work was supported m part by the Shriners Hosprtal for Children and the American Association for University Women (SNB). The authors would like to thank Octavto Hurtado and the Microsystems Technology Laboratory, Massachusetts Instttute ofTechnology, Cambndge, MA for help wtth microfabrrcatton.

References 1 Shtmaoka, S , Nakamura, T , and Ichthara, A (1987) Strmulatton of growth of pnmary cultured adult rat hepatocytes wtthout growth factors by coculture with nonparenchymal liver cells Exp Cell Res 172, 228-242

362

Bhatia, Yarmush, and Toner

2 Goulet, F , Normand, C , and Morm, 0 (1988) Cellular mteracttons promote tissue-specific function, biomatrlx deposmon and Junctional communication of primary cultured hepatocytes Hepatology 8, 10 l&l 0 18 3 Mot-m, O., Goulet, F , and Normand, C , (1988) Liver smusoidal endothehal cells tsolatton, purtficatton, charactertzatton, and mteraction with hepatocytes Revwones Sobre Blologia 15, 146. 4 Lawrence, M B , McIntire, L V , and Eskm, S G (1987) Effect of flow on polymorphonuclear leukocyteiendotheltal cell adhesion Blood 70, 1284-l 290 5 Lawrence, M. B , Smith, C W, Eskm, S G., and McIntne, L V (1990) Effect on venous shear stress on CD 18-mediated neutrophrl adhesion to cultured endohelmm Blood 75,227-237 6 Hammarback, J. A and Letourneau, P. C (1986) Neurite extension across regions of low cell-substratum adhesivity tmphcations for the guidepost hypothesis of axonal pathfindmg Devel Blol 117,65.5-671 7 Matsuda, T , moue, K , and Sugawara, T Development of mlcropattetnmg technology for cultured cells ASAIO Transactions, 36 M5.59-M562 (1990) 8 Corey, J M , Wheeler, B C , and Brewer, G J (199 1) Compliance of hippocampal neurons to patterned substrate networks J Neurosct Res 30, 3OG307 9 Rohr, S , Scholly, D. M , and Kleber, A G (1991) Patterned growth of neonatal rat heart cells m culture morphological and electrophysiological charactenzatton Cwc Res 68, 114-130 10 Bntland, S , Perez-Amaud, E , Clark, P, McGmn, B , Connolly, P, and Moores, G (1992) Micropatternmg proteins and synthetic peptides on solid supports a novel application for microelectronics fabrrcation technology Bzotechnol Prog 8, 155-l 60 11 Stenger, D A , Georger, J H , Dulcey, C S , Hickman, J J , Rudolph,A S , Nielsen, T B , McCort, S M , and Calve& J M. (1992) Coplanar molecular assemblies of ammo- and pefluormated alkylsilanes characterization and gemoetrtc definmon of mammalian cell adhesion and growth J Am Chem Sot 114,8345-8442 12 Clark, P , Brttland, S , and Connolly, P (1993) Growth cone guidance and neuron morphology on micropatterned lammm J Cell Scz 105,203-2 12 13 Lom, B , Healy, K E , and Hockberger, P E (1993) A versatile technique for patterning btomolecules onto glass coverslips J Neuroscl Methods 50, 385-397 14. Ramen, J P , Bellamkonda, R , Jacob, J., Vargo, T G , Gardella, J A , and Aebischer, P (1993) Selective neuronal cell attachment to a covalently patterned monoamine on fluorinated ethylene propylene films J Boomed Mater Res 27,917-925 1.5 den Braber, E T, de RuiJter, J E , Smuts, H. T J , Gmsel, L A, von Recum, A F, and Jansen, J A (1995) Effect of parallel surface microgrooves and surface energy on cell growth J Blamed Mater: Res 29,5 1 l-5 18 16 Lee, J , Morgan, J R , Tomkms, R G , and Yarmush, M. L (1993) Prolme-mediated enhancement of hepatocyte function m a collagen gel sandwich culture configuration FASEB J 7, 586-591 17 Soekamo, A , Lom, B , and Hockberger, P E (1993) Pathtindmg by neuroblastoma cells m culture is directed by preferential adhesion to positively charged surfaces Neurolmage I (2), 129-144

Mcropatternmg

Cells

363

18 Matsuda, T., Sugawara, T , and moue, K (1992) Two-dtmenstonal cell mampulanon technology ASAIO J 38, M243-M247. 19 Smghvt, R , Stephanopoulos, G , Wang, D I. C (1994a) Review effects of substratum morphology on cell phystology Blotechnol Bloeng 43, 764-771 20 Oakley, C and Brunette, D M (1995) Topographic compensatton Guidance and dtrected locomotton of fibroblasts on grooved mrcromachmed substrata m the absence of mtcrotubules Cell Mot11 Cytoskeleton 31,45-58 21 Gundersen, R W (1987) Response of sensory neurttes and growth cones to patterned substrata of lammm and tibronectm m vitro. Dev Bzol 121,423-43 1 22. Hammarback, J A., McCarthy, J. B , Palm, S. L , Furcht, L T , and Letourneau, P C (1988) Growth cone guidance by substrate-bound laminm pathways IS correlated with neuron-to-pathway adhesivny. Dev B~ol 126,29-39 23 Smghvi, R , Kumar, A , Lopez, G P., Stephanopoulos, G N , Wang, D I C , Whitesides, G M., Ingber, D E (1994b) Engineering cell shape and function Sczence 264,696-698 24 Mtyamoto, S , Ohasht, A , Kimura, J., Tobe, S , and Akatke, T (1993) A novel approach for toxtctty sensing using hepatocytes on a collagen-patterned plate Sensors Actuators B 13-14, 196-199 2.5 Bhatia, S N , Toner, M , Tompkms, R G , and Yarmush, M L (1994) Selective adhesion of hepatocytes on patterned surfaces Ann NY Acad Scl 745, 187-209 26 Langenbach, R , Mahck, L , Tompa, A , Kuszynski, C , Freed, H , and Huberman, E (1979) Maintenance of adult rat hepatocytes on c3H/lOT1/2 cells Cancer Res 39,3509-35 14 21 Kurt-Harcuch, W and Mendoza-Ftgueroa, T. (1989) Cultivation of adult rat hepatocytes on 3T3 cells* expression of vartous liver dtfferenttated functions Dzfferentzatzon 41, 148-157 28 Drumheller, P D and Hubbell, J A (1995) Surface munobihzation of adhesion hgands for investigations of cell-substrate interactions, m Bzomedzcal Engzneering Handbook (Bronzmo, J D , ed ), CRC, pp. 1583-1596 29 Guguen-Guillouzo, C , Clement, B , Baffet, G , Beaumont, C , Morel-Chany, E , Glaise, D , and Guillouzo, A (1983) Mamtenance and reversibthty of active albumm secretion by adult rat hepatocytes co-cultured wtth another hver eprthehal cell type. Exp Cell Res 143,47-54. 30 Donato, M T, Gmez-Lechn, M J , and C&tell, J V (1990) Drug metabohzmg enzymes m rat hepatocytes co-cultured with cell hnes. In Mtro Cell Dev Bzol 26, 1057-1062 31 Seglen,P 0. (1976) Preparation of isolated rat liver cells Methods Bzol 13,29-83. 32 Rotem, A , Toner, M , Bhatia, S , Foy, B D , Tompkms, R G , and Yarmush, M L (1994) Oxygen is a factor determining m vitro tissue assembly* effects on attachment and spreading of hepatocytes Blotechnol Bzoeng 43,654-660 33 Bhatia, S N , Yarmush, M L., and Toner, M (1997) Controllmg homotypic vs heterotypic mteracttons by mtcropatternmg* co-cultures of hepatocytes and 3T3 fibroblasts J Boomed Mater Res 34, 189-199. 34 Chen, C , Mrksich, M , Huang, S , Whitestdes, G M , and Ingber, D E. (1997) Geometric control of cell life and death. Science 276, 1425-1428

28 Methods for the Serum-Free Culture of Keratinocytes and Transplantation of CollagetwGAG-Based Skin Substitutes Steven T. Boyce 1. Introduction ObJectIves for dermal-epldermal skm substitutes for treatment of acute and chronic wounds Include, but are not limited to. increased avallablhty, stlmulatlon of wound healing by transplantation of parenchymal cells; regulation of wound healing responses, and, predictable composmon and efficacy to reduce mortality and morbldlty Particularly, the importance of including a cellular component has been demonstrated in experimental grafts of cells and blopolymers (I-7) Transplanted cells may include normal, nontransformed populations isolated for primary culture, or genetically modified cells to deliver specific gene products of therapeutic interest (8-10) Nontransformed cells may include autologous, allogenelc, or chimerlc populations within composite grafts. A variety of approaches are directed toward repair of skm wounds by restoration of the functional anatomy and physiology of skin. An hypothesis shared by several laboratories presumes that duphcatlon of native anatomy and physiology of skm will provide a level of efficacy comparable to split-thickness skm graft. This common rationale has generated models that differ mostly m then selection of blopolymer for delivery of skin cells Acellular dermis has been populated with keratmocytes and fibroblasts m preclmlcal models (11,12). Biosynthetic analogs of skm have combined cultured skm cells with polylactlcipolyglycolic (PLA/PGA) fabric (13,14), polyethylene oxlde/polybutylene terephthalate (PEO/PBT) (3,151, collagen gels (I&28), and collagen-glycosammoglycan (GAG) sponges (7,19,20). This chapter describes specific techniques for preparation and grafting to surgical wounds in athymlc mice of cultured skm substitutes (CSS) from collagen-GAG substrates From

Methods III Molecular Edited by J R Morgan

Medrone, Vol 18 Tmue Engmeenng Methods and Protocols and M L Yarmush 0 Humana Press Inc , Totowa. NJ

365

366

Boyce

populated wtth normal human keratmocytes, melanocytes, and fibroblasts grown in serum-free or low-serum conditions. Principles for CSS based on collagen-GAG substrates that duplicate native skm Include separation of epidermal and dermal cells mto respective histologic compartments, reformatton of eptdermal barrier and basement membrane before grafting, regeneration of skin pigmentation, and restoration of a vascular plexus. CSS before grafting possess these properties, except for cellular components of the vasculature. However, addition of endothellal cells to analogs of cornea (6) has been reported to stimulate more raptd and complete morphogenesrs of the cornea1 epithehum. Conversely, absence of vascular cells and precursors confers limttations to clmical efficacy of CSS, including increased time to healing, and graft loss from mrcrobial contamination and nutrient deprivation (21-24). Furthermore, any functtonal phenotypes of CSS that develop m vttro are not stable, and eventually deteriorate unless they are transplanted and engraft to a viable wound. Anatomic and physiologic deficiencies of CSS impose additional considerations to their chmcal use Therefore, chmcal protocols to compensate for deficiencies of CSS must be designed to optimize their efficacy Conformrty with local protocols for surgical and nursing standards may expedtte comphante of the staff with modalities of wound care Wrth the model of CSS described here, its surgical apphcatron m a single procedure permtts tts admmistratron according to prevailmg standards of burn treatment (25-27) These standards have greatly influenced the development and modification of preclmical protocols described below. But, ultimately, experimental CSS must provide comparable efficacy to the prevailmg standards of wound care, which depend on ObJective assessment of outcome (28-31). After demonstration of efficacy, factors of cost-effectiveness must also be considered (32). If all these requirements are satisfied, then reduction of mortality and morbidity with CSS can be realized. In the followmg section, materials and methods are described for culture of human skin cells, preparation of CSS from cells and collagenGAG substrates, and closure of surgical wounds m athymlc mice for preclnncal studies with CSS

2. Materials Below are summarized the solutrons and other materials used for isolation and culture of human epidermal keratmocytes and melanocytes, and dermal fibroblasts; preparation of CSS; and grafting of CSS to athymic mice These materials provide consistent preparation and engraftment of CSS. Unless stated otherwise, all organic reagents were obtamed from Sigma (St Louis, MO); and morgamc reagents were obtained from Fisher Chemical (Fan-lawn, NJ).

Transplantahon

of Skin Substitutes

367

2.7. Solutions 1. 2 3 4

5 6 7 8 9 10 11

12

13

14

15

Isotonic HEPES buffered saline (HBS). 5% v/v Dettol (Reckitt and Colemen, Hull, UK). 2 4 U/mL Dispase II (Boehrmger Mannhelm, Indianapolis, IN) m HBS 625 U/mL collagenase (Worthington Biochemrcals, Freehold, NJ) m MCDB 153 to activate the enzyme with calcium and magnesium ions, plus 5% v/v bovine pituitary extract (BPE) to neutralize trypsm activities 0 025% trypsn-0 0 1% ethylenediammetetraacetic acid (EDTA) w/v 10% v/v fetal bovine serum (FBS, Grbco-BRL, Grand Island, NY) m keratmocyte growth medium to neutralize trypsm Keratmocyte growth medium Ftbroblast growth medium Melanocyte growth medium CSS maturation medium, as described below Basal medium for culture of human epidermal keratmocytes and melanocytes IS MCDB 153 (33-35), plus increased concentratrons of six ammo acids (histidme, tsoleucme, methronme, phenylalamne, tryptophan, tyrosme), as described by Pittelkow (23) Human tibroblasts are propagated in Dulbecco’s modrfied Eagle’s medium (DMEM, Gibco-BRL) Formulations of these media are presented in Table 1 These basal media are supplemented as described below to support selective cultures of each respective cell type Media supplements for selective cultures of skins cells. Supplements to basal nutrient media are added to stimulate mitosis of respective cells By understandmg the mitogemc responses of keratmocytes and melanocytes, basal medium MCDB 153 can be supplemented differently to prepare selective cultures of each cell type Colony-forming efficiency of cultured keratmocytes m serum-free or biochemically defined medium may be increased by addition of lethally rrradrated 3T3 feeder cells (36,37), if strictly defined condrtrons are not required DMEM 1s supplemented with 5% FBS, 10 ng/mL epidermal growth factor, 5 pg/mL msulm, and 0 5 pg/mL hydrocortisone. Media supplements for selective growth of these three cell types are summarized m Table 2 HEPES buffered salme (HBS). HBS 1s formulated as a lo-fold concentrated stock, filter-sterilized, and stored refrigerated at 4°C The concentrated stock IS prepared from 7 1 49 g HEPES buffer (Research Orgamcs; Cleveland, OH), 10 mL phenol red Stock 5 for MCDB 153 (see Table l), 18.02 g glucose, 2 236 g potassium chlorrde, 76 97 g sodium chloride, and 2.68 g disodium phosphate heptahydrate AdJust pH of the concentrate to 7 4 before filtration HBS working solution is prepared by diluting exactly 100 mL of the concentrated stock to 1 L, adJustmg the pH to 7 4, and filtering aseptically into a stertle bottle (34). Media for cryopreservatron of selective cell cultures Selective cultures of keratmocytes, melanocytes, or fibroblasts are cryopreserved m 70% v/v of then respective supplemented media, 20% v/v FBS, and 10% v/v dtmethyl sulfoxrde (DMSO) Collagen-GAG substrates Collagen-GAG substrates for transplantation of cultured cells are prepared, as described elsewhere (20), to generate a thm (~0 5

368

Boyce Table 1 Basal Media Composition

[M/L] MCDB

Ammo acids Alanme Arginme Asparagme Aspartic acid Cystme Cysteme Glutamic acid Glutamme Glycme Histidmea Isoleucinea Leucme Lysme Methiomnea Phenylalanmea Prolme Serme Threonine Tryptophana Tyrosmea Valme Vitamms and coenzymes Btotm Fohc acid Lipoic acid Niacmamide D-calcium pantothenate Pyridoxme Pyridoxal Riboflavm Thiamine Vitamm B 12 Organic compounds Adenme Choline chloride Glucose Inositol Myomositol

153

1 oe-04 1 Oe-03 1 oe-04 3 o&l5 2 Oe-04 2 4oe-04 1 Oe-04 6 Oe-03 1 Oe-04 2 5e-04 7 8Oe-04 5 Oe-04 1 Oe-04 9 2e-05 9 2e-05 3 Oe-04 6 Oe-04 1 O&4 4 6e-05 7.7e-05 3 Oe-04 6 Oe-08 18e-06 1 Oe-06 3 Oe-07 1 Oe-06 3 Oe-07 2 Oe-05 1 O&7 1 O&l6 3 Oe-07 18e-04 l.OeO4 6 Oe-03 4.Oe-05 1.Oe-O4

DME

4 044

4 Oe-03 4 Oe-04 2 Oe-04 8 Oe-04 8 Oe-04 8 Oe-04 2.0+04 4 Oe-04 4 8 8 5 8

oe-04 Oe-04 Oe-05 oe-04 Oe-04

9 l&6 3 3e-05 8 4e-06

1 le-06 12e-05

2 9e-05 5 6e-03

(contwwed)

Transplantation

of Skin Substitutes

Table 1 Basal Media

Composition

369 [M/L] MCDB

Putrescme Sodium acetate Sodmm pyntvate Thymidme Bulk Ions Calcmm chloride Potassmm chloride Magnesium chloride Magnesium sulfate Sodium chloride Sodium phosphate (dibasic) Sodmm phosphate (monob) Trace elements Cupric sulfate Ferrous nitrate Ferrous sulfate Manganese sulfate Molybdenum Nickel chloride Selenium Silicon Tm chloride Vanadium Zmc sulfate Buffers and mdicators HEPES NaOH NaHC03 Phenol red

153

1 Oe-06 3 7e-03 5 oe-04 3.Oe-06 3 Oe-05 15eXl3 6 Oe-04 8 2e-04 13e-01 2 oe-03 9 1+04

DME

1 Oe-03

18e-03 5.3e-03

1 le-01

1 lOe-08 2 5e-07 5 Oe-06 1 oe-09 1 oe-09 5.oe-10 3.Oe-08 5 oe-07 5 Oe-10 5.Oe-09 5 Oe-07 2 8e-Q2 1 O&2 14&2 3 3e-06

2.2e-02 3 8e-05

aConcentrations of SIXammo acids are increased as described by Plttelkow and Scott (23) mm), symmetrtc sponge that IS laminated on one side with a mtcroporous film of the same biopolymer mixture. Dry, acellular substrates are packaged, sterthzed by exposure to 2.5 mega-Rad gamma trradtatron, and stored at room temperature until they are populated with cultured cells Frames to contam cell suspensions over the surface of collagen-GAG substrates after maculation are cut from type-3 16 stainless steel to fit 1SO-mm dtameter Petri dishes Inoculatton frames are square and measure exactly 97 mm o d., and 85 mm 1.d. The bottom surface of the frame IS beveled from the outside edge

370 Table 2 Supplements

Boyce for Selective

Culture

of Cells from Human

Keratmocytes Basal medmm Epldermal growth factor Insulin Hydrocortlsone Bovine pltmtary extract Fetal bovme serum Penic~llin-streptomycmfunglzone Transferrm Basic fibroblast growth factor a-Melanocyte stlmulatmg hormone Endothelm- 1 Vitamin E (a-tocopherol)

MCDB 153O 1 ng/mL 5 0 pg/mL 0 5 ClglmL 0 5% v/v 0

Melanocytes MCDB

153” 0 5 0 &mL 0 0 4% v/v

Skin Flbroblasts DMEM 10 ng/mL 5 0 PglmL 0 5 pg/mL 0 5% v/v

1% v/v 0

1% v/v 0 5 &mL

1% v/v 0

0

0 6 ng/mL

0

0 0 0

17 ng/mL 2 5 ng/mL 10 Bg/mL

0 0 0

oPlus Increased concentrationsof hydrophobicammoacidsasdescribedm Table 1

to the msldeedge,to generatea blunt wedge at the bottom of the frame to Increase weight dlstrlbutlon where the frame contactsthe substrate,and to decreaseleakage of cells Inoculation framesusedm the author’s laboratory were fabricated by custom order In a local machmeshop.Framesare sterlhzed by steamautoclavmg 16 Maturation medium for cultured skin substitutes To stimulate development of eplthellal barrier m skm substitutes before grafting, it was observed that the keratmocyte growth medium was not satisfactory Poor eplthellal orgamzatlon was observed for at least two major reasons absenceof essentialfatty acids that are required for barrier formation, and an msuffictent calcium concentration Replacement of calcium alone was not sufficient to permit formation of eplderma1barrier m vitro However, studiesto supplementMCDB 153 with lipids (38) stimulated partial formation of barrier lipids and structures analogousto stratum corneum of epldermls This maturation medium IS formulated from lipid prestocks (olelc acid, palmltlc acid, arachldomc acid, vitamin E) carried in the medium with BSA, carmtme, and increasedserme Speclficatlons of this formulatlon are summarizedm Table 3 Supportsto hfi skm substitutesto the au-hquld Interfaceare usedto fit a 150~mm diameter Petri dish, and consist of three parts Square, seamlessframes measurmg 97 mm od and 87 mm id, and circular wire meshmeasurmg 12 5 cm m dlameter, both consistof type-3 16 stamlesssteel Both of thesedeviceswere fabricated m a local machmeshop The wire meshand lifting frames are sterihzed by steam autoclavmg The meshISoverlaid with a cotton filter pad (Schlelcher and Schuell,

Transplantation of Skin SubsMutes Table 3 Formulation Substitutes

of Lipid-Supplemented

371 Maturation

Pre-stock

Medium

for Skin

50X Stock

Final cone

Compound

mg/mL

WI

pL/mL

WI

[Ml

Lmoletc acid Palmmc acid Oleic acid Archrdomc acid a-Tocopherol-Ac BSA (FFA free)

21 05 32 05 35.30 10.65 47 40 27 60

75x10-2 125x10-l 1.25 x IO-’ 35x10-2 1.1 x10-1 42x1&“

10 10 10 10 10 950

75x 104 125x1e3 125x1Cr3 35x 104 1 1 x 10-s 40x 10-4

1.5 x 10-s 25x1c5 25x IO-5 7.0 x IO-6 2.2 x ICY-5 80x 1V’

100X Stock

Pre-stock Compound

mg/mL

WI

Set-me Carnitme

105 0 161

1 x lo-9A4 1 x lo-9A4

pL/mL 10 10

Final cone

[Ml

WI

10 1 0 x 10-3

10x lo-2 1.0 x l&s

Keene, NH) that 1s cut to a 12-cm diameter, and acts as a wick for umform distribution of medium under an-exposed skin substttutes 17 Culture vessels* Selective cell cultures of human keratmocytes or melanocytes are grown m 75-cm* polystyrene flasks (Corning, Corning, NY). Fibroblasts are grown m 150 cm* flasks (Cornmg) Cultured skm substitutes are inoculated wtth cells and Incubated m 150-mm diameter tissue-culture Petri dishes (Lux-Nunc, Gatthersburg, MD) All cells are cryopreserved m 2-mL screw-capped cryogemc vials (Cornmg). Assorted sterile tubes and ptpets are from Cornmg, or Falcon (Fisher Scientific, Itasca, IL). 18 Solution for nrigatton of wounds grafted wtth cultured skin substitutes: Solutions for topical nrtgatton of grafted wounds conststing of nutrients, mitogens, and antimtcrobtals have been shown to improve engraftment of avascular skin substitutes (39,40) For preclnncal studies the nrtgation solution consists of a Basal nutrtent medmm MCDB 153 (see Table l), supplemented wtth b Insulm at 5 pg/mL c Hydrocorttsone at 0 5 pg/mL d Ctprofloxacm (CtproTM, Miles, West Haven, CT) at 20 ug/mL e. Nystatm at 100 U/mL (Glbco-BRL)

2.2. Animal Surgery 1 Anesthesia Animals are anesthetized wtth Avertm, whtch consists of 25 mg/mL tribromoethanol dissolved m 1 25% ternary-amyl alcohol MI), Although this compound provides stable anesthesta for the 30-60 mm surgical procedure, any anesthettc with pharmacologic actions appropriate to the procedure may be substttuted

372

Boyce

2 Resusitatton* Resusitation solution after surgery consists of 3 mg/mL cephtazidime m salme 3 Surgtcal supplies* Stenle supplies of several varieties are used to graft skm substatutes to wounds m athymic mice, including BetadmeTM swabs (Purdue Frederick Company, Norwalk, CT) 70% isopropyl alcohol swabs (Kendall Healthcare, Mansfield, MA) Skm SkrtbeTM surgical markers (Hospital Marketing Services, Naugatuck, CT) 4 x 4 m , l&ply cotton gauze sponges (Johnson & Johnson Medical, Arlmgton, TX) EthtlonTM 6-O sutures (Ethicon, Sommervtlle, NJ) N-TerfaceTM (Wmfield Laboratories, Dallas, TX) XeroformTM gauze (Sherwood Medtcal, St Louis, MO) VaselmeTM gauze (Sherwood Medical) Steri-StripTM compound benzom tmcture ampules (3M Surgical Medtcal, St Paul, MN) OpSlteTM (Smith & Nephew Medtcal, Hull, UK) CobanTM bandage (3M, Surgical Medical, St Paul, MN) 4 Instruments for surgery’ a DumontTM fine forceps, #3C pointed-ttpped, and #2A flat-tipped (Ted Pella, Reddmg, CA) b Iris scissors, needle driver, bandage scissors, and mosqutto hemostat (Mtltex Instrument, Lake Success, NY) 5. Aseptic care and housing of athymic mice: a. Autoclavable, filter-topped cages, bedding, and chow b Water containing 0.2 mg/mL benedryl (as dtphenhydramme, Elkms-Smn, Cherry Hill, NJ), and 1% v/v SeptraTM (as trimethoprim and sulfamethoxizole, Burroughs-Wellcome, Research Triangle Park, NC) c Heatmg pads (non-sterile) d. Lammar flow cage tsolator (Forma Scientific, Marietta, OH) 3. Methods

3.1. Individual Cell Cultures in Selective Media 3.1.1 Primary Cultures of Skin Ceils Keratinocytes, melanocytes, and fibroblasts may be Isolated from a single skm btopsy by serial dtsaggregatton usmg enzymatrc and mechanical techniques. Split-thickness skm is preferable to full-thickness skin, because penetration of enzymes IS more rapid All procedures are performed aseptically, and all materials are sterile. Dlssectmg mstruments may be repeatedly stenlized during the procedures by dtppmg m 95% ethanol, followed by burnmg the ethanol m a gas flame and coolmg the mstruments before contacting the tissue. 1 Wash tissue m HBS, and trim away any subcutaneous tissue 2. Dtsmfect skm by submersion for 15-30 s m 5% v/v Dettol solution, followed by three washes m HBS

Transplantation

of Skin SubsMutes

373

Cut skm mto strips 2-3 mm wade, and transfer to Petri dishes containmg several layers of cotton gauze saturated with 2 4 U/mL Dtspase II m HBS Incubate skm dermis stde against gauze m Dtspase II for 30-120 mm, unttl eptdernus readily separates from the dermis with forceps. Do not allow skin strips to float After epidermis has loosened from dermis, transfer the skm strips to a fresh Petri dish containing HBS. Mechanically separate eptdermts from dermis, and place epidermis mto a fresh Petri dish containing HBS After all the eptdermal sheets have been collected, seal the Petri dish with ParadlmTM and Incubate for 3 h at 4°C. Because HBS contains no calcmm, this incubation depletes calctum and loosens desmosomal Juncttons between keratmocytes Mince dermts mto fine (0 5-l mm2) pteces, and transfer mto a 50-mL centifuge tube contammg 625 U/mL collagenase m culture medium, plus 5% v/v BPE, equthbrated to 37°C and 5% CO2 (collagenase activity is dependent on dtvalent cations, 1 e., calcmm and magnesium) Agitate the tissue pieces wtth a plugged Pasteur ptpet and Incubate at 37’C and 5% CO, for 30-90 mm Agitate the tissue pieces for l-2 min at 20-30 mm intervals during the digestion After the tissue has digested to mmtmal fragments, centrifuge the suspension at 250g for 5 mm at 4°C (excessive digestion will reduce the yield of viable cells) Aspirate the collagenase and resuspend the pellet m selective medium for fibroblasts, as described m Table 2 Repeat centrtfugatron and resuspenston, and maculate dermal cells and fragments mto 150-cm2 flasks contammg 10 mL selecttve medmm for fibroblasts, at an approximate ratio of 1 cm2 dermts.50 cm2 of culture surface. After 16-24 h, add 15 mL of medium per flask Refresh medium at 2-d intervals Retrteve eptdermal strips from 4°C mcubatton and collect HBS from dish Mince eptdermal strips mto fine pieces, as was done with dermis Transfer tissue fragments to a 50-mL centrifuge tube, add l-2 mL of 0 025% trypsm-0 01% EDTA per cm2 of eptdermal tissue, and agttate 4 min using a plugged Pasteur ptpet Let tissue fragments settle, collect the cell suspenston, and transfer tt to selective medium for keratmocytes with 10% FBS on tee Resuspend the eptdermal fragments m HBS, agitate for 4 min, collect the cell suspenston, and transfer to keratinocyte medium plus 10% FBS. Repeat resuspenston, agitation, and collection of cell suspension into keratmocyte medtum plus 10% FBS Resuspend the epidermal fragments in a small volume of keratmocyte medium contammg 10% FBS in a separate tube Centrifuge the suspension (s) and epidermal pieces at 250g for 5 mm at 4°C and aspirate the supernatants Resuspend the eptdermal cells and tissue fragments into a small volume of selective medium for keratmocytes Count the eptdermal cell suspenston. To grow keratinocytes, inoculate mto 75-cm2 flasks at 2-3 x lo4 cells/cm2 containing 15 mL of selective medium for keratmocytes, equthbrated to 37°C and 5% CO*. To grow melanocytes, inoculate mto 75-cm2 flasks at 1 x lo5 cells/cm* containing 15 mL of selective medium for melanocytes (42) (see Table 2), equthbrated to 37°C and 5% C02. Refresh keratinocyte medium at 2-d intervals Refresh melanocyte medmm three times per wk

374

Boyce

3.1.2. Cryopreservatron

of Cells by Controlled Rate Freezing

Before cells become confluent m prtmary culture, great efficiency is gamed If they are cryopreserved for later use. A representative procedure for cryopreservatton of keratinocytes 1s described below, but tt may be modified for other cell types by substttutron of respective media for cryoprotection 1 Prepare cryopreservation

medium by combmmg 7 parts selective medium for

keratmocytes, plus 2 parts FBS, and 1 part dtmethylsulfoxide tion through a 0.22~pm filter.

Stenhze by filtra-

2 Harvest keratmocytes by washing cells with HBS, aspiration of HBS, and addition of 1 mL/75-cm2 flask of cold 0 025% trypsm-0 01% EDTA Incubate for 34 mm at 37°C or 45 mm at room temperature. Gently tap the flask agamst the bench and examme m a phase-contrast contrast microscope After greater than 90% of the cells have detached, collect the cells by serial washing of multiple flasks with a single 10 mL aliquot of HBS, and transfer the cell suspension to selective medium for keratmocytes, plus 10% FBS on ice. Repeat the HBS wash and combme suspensions Centrifuge cells at 25Og, aspirate the supernatant, resuspend m a small volume of selective medium for keratmocytes, and count cells. Repeat centrifu-

gation

and aspiration

cryopreservatlon

of supernatant,

medium for keratmocytes

and resuspend

keratmocytes

at 0 5-3 0 x lo6 cells/ml

in

Dispense

into cryogenrc vials, and freeze at a controlled rate (1 e , 1“C/mm) m a CyroMed microprocessor-controlled freezer (Forma Scientific, Marietta, OH) with a program for suspensions of cultured cells, to retam best viability If a controlled-rate freezer 1s not available, place vials mto a sealed block of Styrofoam with at least 1 cm covering all sides of all vials, place into a -7O“C freezer overnight, and transfer to liquid nitrogen the next day. Store cells mdefimtely m liquid nitrogen

3 7.3 Recovery of Cryopreserved

Cells into Culture

Recover keratmocytes by thawing rapidly tn 70% tsopropyl alcohol, making sure that no alcohol leaks into the vials as they warm. To assure the seal on the vial, swab the vial with 70% isopropyl alcohol, allow it to dry, release pressure from the vial by loosemng and resealmg the cap, and place the vial into 70% tsopropyl alcohol m a 37°C water bath. As soon as keratmocytes are thawed, maculate them at 0 3-l 0 x 1O4cells/cm2 mto 75-cm2 flasks contammg 15 mL of selective medium for keratmocytes equthbrated to 37°C and 5% C02. Refresh medium after 12-24 h, and at 2-d intervals thereafter. Ftbroblasts and melanocytes are handled stmtlarly, but m their respective media. Incubate until sufftctent cells are available to prepare cultured skin substrtutes. Representative photomtcrographs of selective cultures of keratmocytes, melanocytes, and fibroblasts, m log-phase growth after recovery from cryopreservatton, are shown m Fig. 1.

3.2. Preparation

of Cultured

Skin Substitutes

A schemattc diagram for preparation of cultured skm substttutes 1s presented m Fig. 2 Selecttve cultures of skm cells are harvested on successtve days, and

Transplantation

of Skin Substitutes

.> d S’ ,

i f

_.,“L !

.>’ ,

a~.

-a&..

-:~~~,wx~~ iyi;“:?~

Fig. 1. Photomicrographs of normal human skin cells in log-phase, selective culture. (A) epidermal keratinocytes; (B) epidermal melanocytes; and, (C) dermal fibroblasts. Scale bar. 0.1 mm.

376

Boyce

maculated onto collagen-glycosammoglycan substrates m submerged culture. Eptthehum matures mto an analog of epidermis by an-exposed incubation m hptdsupplementedmedmm. Eptdermal melanocytesmay be added to the skm substttute by co-moculattonwith keratinocytesat a ratio of 1melanocyte/30keratinocytes. 3.2.1. Rehydration of Dry Collagen-GAG Substrates Rehydrate dry collagen-GAG substrates( 10 x 10cm) by removal from stertle packaging, and place into 150 mm tissue-culture Petri dishes contammg 75 mL HBS. Incubate 30 mm and invert substrate m dish. Aspirate and refresh HBS, gently express au from the substrate, and incubate 30 mm. Repeat procedure using selective medium for fibroblasts. Asptrate medium from dish, orrent the substrate with porous stde up, and install inoculatton frame (see Subheading 2.1.5.) over substrate. Add 20 mL of selecttve tibroblast medium only to outstde of frame, and equthbrate m cell culture mcubator at 37°C and 5% C02. 3.2 2. Harvest and Inoculation of Cultured Flbroblasts Harvest fibroblasts m log-phase growth, as described for keratmocytes m Subheading 3.1.2., step 2, except reduce the mcubatron time of fibroblasts m 0.025% trypsm-O.Ol% EDTA to 2-3 min. Collect cells m 10 mL HBS, transfer to a 50-mL centrifuge tube containing selecttve medium for fibroblasts, centrtfuge at 25Og, aspirate supernatant medium, resuspend fibroblasts m a small volume of their selective medium, and count cells. Adjust cell density to 4 x IO6 cells/ml Evenly Inoculate 10 mL of the fibroblast suspension over the area within the moculatton frame (5 x lo5 cells/cm*). After 24 h, remove the moculatton frame, aspirate medmm and unattached cells, and transfer the substrate with tibroblasts to a new 150-mm Petri dish Wash the substrate with attached fibroblasts m 35 mL selective medtum for keratmocytes, and carefully invert the dermal substitute, to orient tt with the mtcroporous side up, for maculation of keratmocytes. Wash dermal substitute with 35 mL of selective medium for keratmocytes, aspirate medium, and Install an moculatton frame, as before. Add 20 mL selective medium for keratmocytes only to the outside of the frame, and equthbrate m cell-culture mcubator at 37°C and 5% C02. 3.2 3. Harvest and Inoculation of Cultured Keratmocytes Harvest keratinocytes as described n-rSubheading 3.1.2., step 2, and, after centrrfugatton, resuspend the cell pellet u-ra small volume of selective medium for keratmocytes.Count the keratmocytes,and adjustthe density to 8 x 1O6cells/ml. Remove dermal substttutewith inoculatron frame from the incubator, and evenly inoculate 10 mL of the keratmocyte suspension over the area wtthm the moculatron frame (1 x 1O6cells/cm*). Inoculatton of keratmocytes onto dermal substitutes 1sdefined as day 0 of mcubatton of cultured skm substttutes

Transplantation

of Skin Substitutes

377

KSrStlnOcytSS Keratinocytes

3

Collagen-GAG

Substrate Fibroblasts

Cultured

Skln

Substitute

Flbroblasts

Fig. 2. Schematic diagram of preparation of cultured skin substitutes. Selective cultures of dermal fibroblasts are inoculated into the porous reticulations of a collagen-glycosaminoglycan (GAG) substrate. Selective cultures of epidermal keratinocytes are prepared to include or exclude selective cultures of epidermal melanocytes. If added, melanocytes are mixed with keratinocyte suspensions at a ratio of 1 melanocyte to 30 keratinocytes before inoculation onto the micro porous surface of a collagen-GAG substrate. The cell-biopolymer composite is incubated at the air-liquid interface to promote epithelial stratification and barrier formation.

3.2.4. Addition of Melanocytes to Cultured Skin Substitutes Add melanocytes to skin substitutes by simultaneous harvest of selective cultures of melanocytes. After each cell type is counted, and before inoculation of keratinocytes, add melanocytes into the keratinocyte suspension at a ratio of 1 melanocyte:30 keratinocytes. Inoculate dermal substitutes, as described in Subheading

3.2.3.

3.2.5. Incubation Day I After 24 h (incubation d l), add 10 mL selective medium for keratinocytes to the inside of the inoculation frame only. 3.2.6. Incubation Day 2 On incubation d 2, remove the inoculation frame and wash with maturation medium containing the lipid supplement described in Table 3. Increase the calcium concentration in the medium to 0.5 mM, as described in Table 4, and

378 discontinue m medium

Boyce BPE

Incubate the cultured

skm substitute submerged

3 2.7 lncubatlon Day 3 On mcubatlon d 3, prepare a new 150-mm Petri dish contammg a hftmg frame, mesh and cotton support (see Subheading 2.1., item 6). Add 60 mL of maturation medium to the dish according to the formulation and schedule m Table 4. Equrltbrate the assembly m a cell-culture incubator at 37°C and 5% C02. Aspirate medium from the dish contammg the skm substitute, and trim away edges of the collagen-GAG substrate that were outside of the maculation frame (no cells), and discard. Remove the Petri dish with lifting frame from the mcubator, and transfer the cultured skm substitute to the top of the cotton pad on the stamless steel mesh. Transfer the skm substitute by laytng a piece of Nterface of correct size (8.5 x 8 5 cm) on the epithelial surface Lift one corner of the skm substitute and N-terface with a flat-tipped forceps, and hold an adjacent corner with another identical forceps. Lift the skm substitute, transfer it to the dish containing the liftmg frame, and lay it on the cotton pad saturated with medium To reposition the skm substitute, lift it from one site and replace it m another site, but do not drag or pull it Return the lifted skm substitute to the incubator. 3 2.8 Incubation Day 4 On culture d 4, aspirate medium from the dashcontammg the skin substitute, and add 4.5 mL maturation medtum for d 4, as described m Table 4. Incubate the skm substitute m thts condition until use Refresh maturation medium daily. 3.2.9. Preparation of Cultured Skin Substitutes for Surgery On the day before animal surgery (ca mcubation d 14), prepare cultured skin substttutes for graftmg. Transfer skm substitutes to new 150-mm Petri dishes containing 10 mL maturation medium With a straight-edge razor blade (Week, Week, NC) held m a hemostat and a fine-tipped forceps, cut skm substitutes mto grafts exactly 2 x 2 cm Transfer the mdtvidual grafts to the dish contammg the lifting frame, and return to the incubator until the day of surgery A photomicrograph of the histology of a representative skm substitute ts presented m Fig. 3 3.3. Animal Surgery All care and use of ammals is approved by the Instttutional Animal Care and Use Committee of the University of Cmcmnati. All procedures are performed with aseptic techmque, and all materials are sterile. Surgical procedures are performed m a biological safety cabmet, and animals are housed m filter-topped cages in a laminar flow cage isolator

Transplantation Table 4 Schedule

of Skin SubsMutes

and Media for Maturation

Factor

Day 1

379 of Cultured

Day 2

Skin Substitutes Day 3

Day 4 and after

BPE (% v/v)

0.5

0

0

0

Lipid supplementa EGF (ng/mL) Calcium (mM) Au-exposure

No 1

Yes 1

02

05

No

No

Yes 1 10 Yes

Yes 0 1.5 Yes

aLlpld Supplement IS formulated as described m Table 3 All media are MCDB 153 as described m Table 1 supplemented with 5 pg/mL msulm and 0 5 pg/mL hydrocortlsone

3.3 I. Preparation of a Sterile F/e/d for Surgery Prepare materrals for surgery by packaging of multrple units of each item into sterrle specrmen jars Cover the inside of the brologrcal safety cabinet with a sterile drape. Place electric heating pads set on “low” beneath the operating field, and beneath a cage to be used for recovery, to prevent hypothermra durmg and after surgery.

3.3.2 Anesthesia Anesthetize an athymrc mouse by mtraperrtoneal mJectron of 350-500 p,L of avertm, as described m Subheading 2.2.1. Verify complete anesthesia by negatrve response to toe pmch.

3.3.3. Cleansing of the Surgical Site Cleanse skin on (right) flank of animal by serial swabbing with Betadme, followed by 70% rsopropanol Demarcate a 2 x 2-cm area on the flank with a square template by marking the corners and mrdpomts of the sides of the template.

3 3.4. Preparation of a Full-Thickness Skin Wound Carefully prepare a full-thickness skin wound to preserve the panmculus carnosus. With fine forceps and u-is scrssors, lift a corner of the demarcated site and incise through the eprdermrs and into the dermis. Continue the mcrsron around the entire wound perimeter by placement of the sctssors tip mto the dermis, application of gentle forward and upward pressure, and closing the scissors to advance the incision Apply gentle countertractron to the edges of the mcisron to promote its advancement. Periodically open the mcisron and verify visually that the panmculus carnosus remains intact. After the mcrsion IS complete, carefully hft one corner of the murine skin with a fine forceps, and dissect the panmculus carnosus from the dermis with the tip and edge of the scrssors An example of the completed dissection IS presented in Fig. 4A.

380

&WCC?

Fig. 3. Histology of a cultured skin substitute. Keratinocytes (HK) are restricted to the outer surface of the dermal substitute (C-GAG-HF) to form separate compartments within the skin substitute. Total thickness is less than 0.5 mm. Scale bar, 0.1 mm.

3.3.5. Suturing and Dressing of Cultured Skin Substitutes on to Wounds Place the cultured skin substitute with a backing of N-terface onto the wound. Secure the graft and dressing to the wound margin with stent-type suturing at eight points, as shown in Fig. 4B. Pack the graft with several layers of cotton gauze, and tie opposing sutures together, as shown in Fig. 4C. Apply benzoin adhesive around perimeter of wound, and cover the packing with OpSite to form a compartment over the graft that retains fluid, but is vapor permeable.

3.3.6. Resusita tion Resusitate the animals by intraperitoneal injection of 1.O cc saline containing 3 mg of cephtazidime. 3.3.7. Bandaging of Dressings Bandage the dressed graft with Coban, and prepare a small (3-5 mm) opening to inject irrigation solution. Inject 1.5 cc of irrigation solution (see Subheading 2.1.8.) through OpSite into gauze. Maintain mice on the heating pad until they recover from anesthesia. After animals recover from anesthesia, return them to their respective cages for housing. Keep cages on heating pads set on “low” until all dressings are removed. Supply water prepared with benedryl and Septra as described in Subheading 2.2.5.2. 3.4. Dressing and Care Protocol 3.4.1. Postoperative Days l-l 4 Inject 1.Occ/animal/d of irrigation solution into gauzepacking over the wound.

Transplantation

of Skin Substitutes

3 4.2 Post-operatrve

381

Day 14

Remove dressmgsand suturesand collect data (i.e., photographs, area tracings, nonmvasrve mstruments) (43,44). Replace dressingswith N-terface contacting the graft, followed by Xerofonn, cotton gauze,and Coban bandage until d 2 1 3 4.3. Post-operatrve

Day 21

Remove dressings, collect data and redress with N-Terface, cotton gauze, and Coban until d 28. Representative healing of wounds at d 2 1 after surgery IS shown m Fig. 5. Healed skm from skm substitutesdepleted of melanocytesISshown m Fig. SA. Co&matron of engraftment of humankeratmocytesis shown m Fig. 5B by nnmuno hrstochemrcalstaining for HLA-AE3C antigens (2,40,45). Skm substrtutes with added melanocytesproduce uniformly black skm, as representedm Fig. 5C Locahzattonof melanocyteswithin the basallayer of keratmocytesis verified by nnmunohrstochemical staining for the melanocyte marker, Mel-5 (46) 3 4.4. Postoperative Day 28 Remove all dressings and swab healed skin with Vaseline gauze Collect data as needed. 4. Notes

4.1. Media Preparation 1 Storage of media To increase reproducibility, prepare sufficient medra and supplements for 2-3 mo, ahquot, and freeze at -20°C or -7O’C Thus simple practice also decreases labor mtensrty of medium preparation. 2 Reconstrtutron and supplementatron of media Reconstnute and supplement media for use not less than weekly. Even serum-supplemented media deteriorate at 4°C Optrmum performance of serum-free media IS obtained rf all components are as fresh as possible Serum-free media are more subject to reduced performance if any component (s) lose (s) activity 3. Use of media a Warm medra to 37°C before use, mmimrze warm time b. Keep labile reagents (especially trypsin) on me during use c Avoid exposure of culture media to light, especially fluorescent light that degrades certain vrtamms (I e , rrboflavm, B2) d. Replace unused medra chemrcals no less than annually

4.2. Cell Culture 4 Safety Maintain universal precautions for laboratory staff Human cells can carry pathogens (1.e , human nnmunodeficency vmus, hepatms, cyto megalo virus) Certain sources (e g , surgical discard) of human tissues may not require pathogen testing. Other sources (e.g., tissue banks accredited by the American Association of Tissue Banks) perform tests for human pathogens, and requrre negative

382

Boyce

Fig. 4. Photographs of steps in the procedure to graft human cultured skin substitutes to athymic mice. (A) Murine skin removed with preservation of the panniculus carnosus. (B) Cultured skin substitute covered with N-TerfaceTM dressing is secured to the wound margin with stent-type sutures. (continued on oppositepage) tests before release of tissue. However, negative tests for pathogens do not provide absolute assurancethat pathogens are not present. Therefore, appropriate safety factors should be required to protect staff who handle human-derived materials. All human tissues and materials (media, culture vessels)that contact human cells and tissues should be treated asbiohazards, and be treated lethally (e.g., bleach, incineration) for disposal. 5. Tissue: Isolate cells from tissue of optimum viability. Discarded tissue from elective surgery (i.e., neonatal circumcision, reduction mammoplasty) may be

Transplantation

of Skin Substitutes

383

Fig. 4. (C) Grafted wound is packed with gauze and stent sutures are tied over to secure the packing. (D) Dressed wound is covered with OpSiteTM to retain moisture in the wound, and is bandaged with Coban TM. Irrigation solutions are injected through a small aperture in the CobanTM, through the OpSiteTM and into the gauze to modulate the healing wound. obtained according to guidelines and policies of the local institutional review board. Cadaveric donors who are young (18-30 yr), otherwise healthy individuals are usually sources of high proportions of proliferative cells. Samples of splitthickness skin are best. Minimize ischemic time between acquisition of tissue and isolation of cells to preserve optimum cellular viability.

384

Boyce

Fig. 5. Photographs of nonpigmented and pigmented human skin after engraftment of cultured skin substitutes. (A) By 21 d after grafting, the epithelium is fully functional and dry. (B) Persistence of human cells is verified by immuno histochemical staining of HLA-ABC antigens in healed epidermis (right of panel). Staining stops abruptly at the wound margin (center of panel), and murine epidermis (left of panel) is negative for human cell markers. (C) Cultured skin substitutes with added melanocytes generate black skin by 21 d after grafting. Localization of melanocytes to the basal layer of keratinocytes (arrowheads) is confirmed by immuno histochemical staining with the melanocyte-specific marker, Mel-5. The outer surface of the stratum corneum is identified with asterisks (*). Scale bar, 0.1 mm.

Transplantation

of Skin Substitutes

385

Fig. 5C and D.

6. Cryopreservation: Cryopreserve cells after primary culture in liquid nitrogen for long-term storage. If sufficiently large cell populations are propagated in primary culture, they may be distributed into aliquots and used for months or years. This simple step will assure continuous quality of results, and provide a performance standard for cellular responses to experimental conditions in preclinical studies. 7. Cell viability: Avoid repeated subculture before assay or surgery. All nontransformed cells have finite life spans, and cellular responses decline with time in culture.

386

Boyce

4.3. CSS Preparation 8 Inoculate cells from log-phase cultures. Avoid confluence and density mhlbttton of keratmocyte cultures, especially Density mhtbrtton IS not fully reverstble, and, If keratmocytes become arrested, the eptthehum ofthe skin substttute ~111be compronused 9 Determine kmetrcs of opttmum cellular vtabthty and dtfferenttatton in vnro to predtct the best surgtcal result. If skm substitutes are grafted too early, then strattticatton and barrter may not be opttmal; tf too late, then potenttal for cellular proltferatron may be reduced 10 Handle skin substitutes wtth a flexible, inert, nonadherent backmg (1 e , N-terface polypropylene dressmg) to reduce folding and mcrease mechanical strength durmg processmg m vitro and surgtcal procedures

4.4. Animal Surgery 11 Use a single gender of ammals to reduce phystologtc vartattons, and gender-gender mteracttons Female athymtc mace are very tolerant of handling and surgery 12 Use ammals 24 mo of age (young adults) to provrde greatest vigor after surgery, and an observatton pertod of more than 1 yr Use animals from the same breedmg lot for each experiment to avotd artifacts caused by age or other factors 13 Be very careful to preserve the Integrity and blood flow of the pannmulus carnosus to opttmrze the vascularity of the woundbed for grafts Although the panmculus carnosus IS not required for engraftment of skm substitutes (II), Its presence Increases the probability for engraftment 14 Place chow cakes m the beddmg of the cage for first 4 wk after surgery 15 Check ammals no less than dally while m dressings

4.5. Conclusion sktn subsmutes1sa complex process,with high requtrements for stringency to be successfulconsistently However, the htgh strmgencyand specialconstderattonsfor then routme use result from anatomtc and phystologtc deficienciesof the cultured grafts. Therefore, rt may be expectedthat the strmgencyofpreparattve andgrafting proceduresfor skmsubstttuteswill decreaseas graft cornpositron becomesmore homologous with name skm.Accomplishment of this goal ~111bestserve all of the btomedtcalapplications of cultured skrn substttutes Preparation

and grafting of cultured

Acknowledgments The author’s work 1ssupported by PHS grants GM50509 from the Nattonal lnstttutes of Health, and FD-R-672 from the United States Food and Drug Admmtstratton, and by grants from the Shrmers Hospitals for Chtldren References 1 Boyce, S T , Glafkides, M C , Foreman, T J , and Hansbrough, J F (1988) Reduced wound contractton after grafting of full-thtckness wounds with a collagen and chon-

Tw~~p/anfafion

2

3

4. 5

6

7

8

9

10

11

12

13

14

of Skin Substitufes

387

droltm-6-sulfate (GAG) dermal skm substitute and coverage with Biobrane J Burn Care Rehabll 94,3&I-370 Boyce, S T, Foreman, T J., English, K B , Stayner, N., Cooper, M. L , Sakabu, S , and Hansbrough, J F (1991) Skm wound closure m athymlc mice with cultured human cells, blopolymers, and growth factors Surgery 110,866-876. Beumer, G J , van Blitterswqk, C A., Bakker, D , and Ponec, M (1993) Cellseedmg and vitro blocompatlbdity evaluation of polymeric matrlces of PEOiPBT copolymers and PLLA Bzomaterzals 14(8), 598-604 Dolllon, C J., Sliver, F H., and Berg, R. A. (1987) Flbroblast growth on a porous sponge containing hyaluromc acid and fibronectm Bzomaterzals 8, 195-200 KreJci, N C , Cuono, C B , Langdon, R. C , and McGulre, J (1991) In vitro reconstitution of skm fibroblasts facilitate keratmocyte growth and dlfferentlatlon on acellular reticular dermis J Invest Dermatol 97(S), 843-848 Zieske, J D , Mason, V S , Wasson, M E , Meumer, S F , Nolte, C J , Fukal, N , Olsen, B R , and Parenteau, N L. (1994) Basement membrane assembly and dlfferentlatlon of cultured cornea1 cells importance of culture environment and endothehal cell mteractlon Exp Cell Res 214(2), 62 l-633 Boyce, S T and Hansbrough, J F (1988) Blologlc attachment, growth, and dlfferentlatlon of cultured human epldermal keratmocytes on a graftable collagen and chondroltm-6-sulfate substrate Surgery 103,421133 1 Morgan, J R , Barrandon, Y, Green, H , and Mulligan, R C. (1987) Expressron of an exogenous growth hormone gene m transplantable human epldermal cells Sczence 237, 1476-1479 Krueger, G. G , Morgan, J R , Jorgensen, C M , Schmidt, L , L1, H L , LI, L T, Boyce, S. T , Wiley, H S , Kaplan, J., and Peterson, M J (1994) Genetically modified skm to treat disease potentials and hmltations J Invest Dermatof 103(5), 76s-84s Emmg, S A, Lee, J , Snow, R G , Tompkins, R G ,Yarmush, M L , and Morgan, J R (1995) Genetically modified himan epidermis overexpressmg PDGF-A directs the development of a cellular and vascular connective tissue stroma when transplanted to athymlc mice-Impllcatlons for the use of genetically modified keratmocytes to modulate dermal degeneration J Invest Dermatol 105,756-763 Medalle, D A., Emmg, S A, Tompkins, R G ,Yarmush, M L , and Krueger, G G (1996) Evaluation of human skm reconstituted from composite grafts of cultured keratmocytes and human acellular dermis transplanted to athymlc mice J Invest Dermatol 107(l), 121-127 Griffey, E S , Hueneke, M., Sukkar, S , Wamwright, D , and Livesey, S A (1995) Production of a human m vitro reconstituted skin and grafting to a nude mouse model Wound Rep Reg 3(l), 92 (Abstract) Hansbrough, J F, Cooper, M L , Cohen, R , Splelvogel, R L , Greenleaf, G , Bartel, R L., and Naughton, G (1992) Evaluation of a biodegradable matrix containing cultured human fibroblasts as a dermal replacement beneath meshed sktn grafts on athymlc mice Surgery 11 l(4), 438-446. Cooper, M L , Hansbrough, J F, Spielvogel, R L , Cohen, R , Bartel, R L , and Naughton, G (1991) In VIVO optlmlzatlon of a hvmg dermal substitute employmg

388

15

16. 17 18 19

20

21

22

23

24

25 26. 27. 28

29

30

Boyce cultured human fibroblasts on a biodegradable polyglycohc or polygalactm mesh Blomaterlals 12,243-248 Beumer, G J , van BlitterswtJk, C A , and Ponec, M (1994) Btocompattbthty of a biodegradable matrtx used as a skm substitute. an m vtvo evaluation f Bzomed Mater Res 28(5), 545-552 Bell, E , Ehrhch, H. P , Buttle, D J , and Nakatsp, T (198 1) A living tissue formed m vitro and accepted as a full thtckness skm equtlvalent. Sczence 211, 1042-1054 Bell, E , Ehrhch, H P, Sher, S., Merrtll, C , Sarber, R , and Hull, B. (1981) Development and use of a hvmg skm equivalent Plast Reconstr Surg. 67, 386-392 Germam, L , Mtchel, M , and Auger, F A (1993)Anchored sktn equivalent cultured m vitro a new tool for percutaneous absotptton studtes In Kh Cell Dev Blol 29A, 834-837. Hansbrough, J F and Boyce, S. T (1984) What crtterta should be used for destgnmg arttlical skm replacement and how well do the current matertals meet these criteria J Trauma 24, 3 ls-35s. Boyce, S T, Chrtsttanson, D J , and Hansbrough, J F (1988) Structure of a collagen-GAG dermal skin substitute optimized for cultured human eptdermal keratmocytes J Blamed. Mater Res 22, 939-957 Gallico III, G. G., O’Connor, N E , Compton, C C , Kehmde, 0 , and Green, H (1984) Permanent coverage of large bum wounds with autologous cultured human eptthelmm New Engl J Med 311,448-45 1 Cuono, C., Langdon, R , Btrchall, N , Barttelbort, S , and McGmre, J (1987) Composite autologous-allogeneic skin replacement* development and clmtcal apphcatton. Plast Reconstr Surg , 80,626635 Ptttelkow, M R , and Scott, R E (1986) New techniques for the m vttro culture of human skm keratmocytes and perspectives on thier use for grafting of patients wtth extensive burns. Mayo Clm Proc 61, 771-777 Hansbrough, J F , Boyce, S T , Cooper, M L , and Foreman, T J (1989) Bum wound closure with cultured autologous keratmocytes and tibroblasts attached to a collagen-glycosammoglycan substrate JAMA 262,2 125-2 130 Tompkins, R G , and Burke, J F (1996) Altemattve wound covermgs, m Total Burn Care (Hemdon, D N , ed.), W B. Saunders, London, pp. 164-172 Warden, G D., Saffle, J. R., and Kravttz, M (1982) A two-stage techmque for exctston and graftmg of burn wounds J Trauma 22,98-103 Tanner, J C , Vandeput, J , and Olley, J F. (1964) The mesh skm autograft Plust Reconstr Surg 34,287-292. Boyce, S T., Goretsky, M J , Greenhalgh, D G , Kagan, R J , Rteman, M T, and Warden, G. D ( 1995) Comparative assessment of cultured skin substttutes and nattve skm autograft for treatment of full-thtckness burns. Ann Surg 222(6), 743-752 Helmbach, D., Luterman, A, Burke, J F , Cram, A , Hemdon, D , Hunt, J , Jordon, M , McManus, W, Solem, L., Warden, G., and Zawackt, B (1988) Artifical dermis for maJor bums, a multt-center randomtzed clmtcal trial Ann Strvg 208,3 13-320 American Medical Assoclatton (1993) The skin, m Guides to Evaluation of Permanent Impairment (Engelberg, A L , ed.), American Medtcal Assoclatton, Chicage, IL, pp 277-289

31 Herndon, D N. and Rutan, R L (1992) Comparison of cultured epldermal autograft and masstve excision with serial autografting plus homograft overlay J Burn Care Rehahll 13,154157 32 Gllpm, D A., Barrow, R E , Rutan, R. L., Broemelmg, L., and Herndon, D N (1994) Recombinant human growth hormone accelerates wound healmg m chlldren with large cutaneous burns Ann Surg 220(l), 19-24 33 Boyce, S T and Ham, R G (1983) Calcium-regulated differentiation of normal human epidermal keratmocytes m chemically defined clonal culture and serumfree serial culture J Invest Dermatol 81, suppl 1,33sAOs. 34 Boyce, S. T and Ham, R G (1985) Cultivatton, frozen storage, and clonal growth of normal human epldermal keratmocytes in serum-free media. J 75s Czdt Meth 9,83-93. 35 Boyce, S T and Ham, R G (1986) Normal human epidermal keratmocytes, m In Ktro Models for Cancer Research Carcmomas of the Mammary Gland, Uterus, and Skm, vol 3 (Webber, M M and Sekeley, L I , eds ), CRC, Boca Raton, FL, pp 245-274. 36 Rhemwald, J G and Green, H. (1975) Formation of a keratmizmg epithelmm m culture by a cloned cell lme dertved from a teratoma. Cell 6, 3 17-330 37 Rheinwald, J G and Green, H ( 1975) Serial cultivatton of strains of human epldermal keratmocytes the formation of keratmizmg colonies from smgle cells, Cell 6,33 l-343 38 Boyce, S T and Willlams, M L (1993) Lipid supplemented medmm induces lamellar bodies and precursors of barrier lipids in cultured analogues of human skm J Invest Dermatol 101, 180-184 39 Boyce, S T, Medrano, E E , Abdel-Malek, Z A., Supp, A P, Dodick, J M., Nordlund, J J , and Warden, G D (1993) Pigmentation and mhlbmon of wound contraction by cultured skm substitutes with adult melanocytes after transplantation to athymic mice. J Invest Dermatol 100, 360-365 40 Boyce, S T, Supp, A P, Harriger, M D , Greenhalgh, D G , and Warden, G D (1995) Topical nutrients promote engraftment and inhibit wound contractton of cultured skm substitutes m athymlc mice. J Invest Dermatol 104(3), 345-349 41 Cunhffe-Beamer, T L (1983) Biomethodology and surgical techniques, m The Mouse zn Bzomedzcal Research, vol III (Foster, H L , Small, J D , and Fox, J G , eds ), Academrc, New York, NY, pp 4 17-4 18 42 Swope, V B , Medrano, E E , Smalara, D , and Abdel-Malek, Z A (1995) Long-term proliferation ofhuman melanocytes is supportedby the physiologic nutogens a-melanotropm, endothehn- 1, and basic fibroblast growth factor Exp Cell Res 217,4X&459 43 Boyce, S T., Supp, A P, Harriger, M D , Plckens, W. L , Wickett, R R , and Hoath, S B. (1996) Surface electrical capacitance as a non-invasive index of epldermal barrier m cultured skm substitutes m athymtc mice. J Invest Dermatol 107(l), 82-87

44 Setup, J , and Jemec, G B. E (1995) Non-Invaszve Methods and the Skzn, CRC, Boca Raton, FL 45. Briggaman, R A (1985) Human skin grafts-nude mouse model* techniques and application, m Methods m Skm Research (Skerrow, D and Skerrow, C J , eds ), Wiley, New York, pp 25 l-276 46 Vgayasarahdl, S and Houghton, A N (1991) Purification of an autoantigenic 75kDa human melanosomal glycoprotem Intl J Cancer 47(2), 298-303

29-m Use of Skin Equivalent Healing Model

Technology

Michael A. Vaccariello, Ashkan Javaherian, and Jonathan A. Garlick

in a Wound Nancy Parenteau,

1. Introduction Re-epithehahzatton is defined as the reconstitution of cells mto an organized, stratified squamous epithelmm that permanently covers a wound defect and restores function (I). Followmg wounding, keratmocytes are activated to undergo a series of phenotypic changes that have been well-characterized m viva (2-4). However, m vitro studies of re-epithehahzation have often been limited by their mabihty to simulate the in vtvo tissue. Wound models using skin explants (5-s) or submerged keratmocyte cultures (9,10) demonstrate only partial differentiation and hyperprohferattve growth These systems have been useful for studying keratmoctye migration (II), but are hmrted m studying other aspects of re-epitheliahzation. The development of novel models to study wound healing and re-eptthehaltzatton relies on the abthty to engineer a tissue that mimics its m viva counterpart The skm equivalent (SE) 1san m vitro tissue that consists of a stratified squamous eptthelmm grown on a collagen matrix populated with dermal tibroblasts (Fig. 1). The ttssue-like character of this construct generates a threedimensional, organotypic culture, demonstrating epithehal differentiation, morphology, and proliferation rates similar to that found m skin (22,13) In this chapter, we descrtbe techniques for construction of a model for wound healing m vitro that has been generated by adapting SEs (14). Using this system, we have found that the chronology of events during re-epithehahzation is similar to that reported m skm during wound healing (15). Migration 1s untiated by keratmocytes at the wound margin, which form an epithehal tongue as progeny of proliferating cells move laterally to cover the wound defect (arrows, Fig. From


Medmne, Vol 18 Tmue Engmeenng Methods and Protocols and M L Yarmush 0 Humana Press Inc , Totowa, NJ

397

392

A

Vaccariello et al.

Keratinocytes on Plastic (RheinwaldlGreen)

C

Keratinocytes

on Collagen (E. Bell)

Raft

Fibroblasts in t:ollagen

matrix

Fig. 1. Comparison of cultivation in submerged keratinocyte culture versus skin equivalent culture.(A) keratinocytes are cultured while submerged in medium on supporting feeder layers of mitotically inactive, metabolically active 3T3 fibroblasts. The resultant cultures are partially stratified, yet do not undergo full biochemical and morphologic differentiation (B). Alternatively, keratinocytes can be cultured on a collagen matrix containing fibroblasts while being fed at the airliquid interface (C). Cultivation at this air-liquid interface results in skin equivalents demonstrating complete morphologic differentiation and cultures that more closely simulate the in vivo tissue (D).

2A). Proliferative activity continues to be high in the wound center as the epithelium stratifies to normal thickness (Fig. 2B). Using this model, the effect of growth factors, such as TGF-PI (Id), and expression of matrix metalloproteinases (17) have been characterized during re-epithelialization of skin and oral mucosal keratinocytes. These applications demonstrate the potential of this model in studying cell phenotypes characteristic of the switch from a normal to a regenerative tissue during wound healing. Since high-quality SEs provide the basis for the generation of the wound healing model, this chapter will first describe materials and methods needed in SE construction. Following this, the specific adaptation of these cultures to study wound healing will be detailed. It is hoped that by describing a model that recapitulates events occurring during re-epithelialization of wounds in vivo, further study of the nature and fate of keratinocytes mobilized during reepithelialization may be facilitated.

Skin Equivalent Wound Healing Model

393

Fig. 2. Histologic appearance of skin equivalents during re-epithelialization. Skin equivalent cultures were wounded 3 d after being raised to the air-liquid interface. Eight hours after wounding, an epithelial tongue is seen to extend onto the central wound surface (A). By 48 h after wounding, the wound is completely covered by a stratified epithelium (B). Arrows mark the wound margins.

2. Materials

2.1. Submerged Cultures 2.1.1. Submerged Kera tinocyte Culture Media (KCM) 1. Source of human keratinocytes: newborn foreskins (average size: l-2 cm2). 2. Keratinocyte tissue-culture medium: a. Dulbecco’s modified Eagle’s medium (DMEM):Ham’s F12 medium (GibcoBRL, Gaithersburg, MD) (3: 1). b. Fetal bovine serum (FBS): 5% (HyClone, Logan, UT). 1. Penicillin-streptomycin (Sigma, St. Louis, MO): Dissolve 2.42 g penicillin and 4.0 g streptomycin in 400 mL 2X dH20 to make a 100X stock, filter-sterilize, aliquot and store at -20°C. 2. HEPES (Sigma): Dissolve 47.24 g in 250 mL 2X dH20 to make a 100X stock. Filter-sterilize, aliquot, and store at -20°C. 3. Adenine (ICN): Dissolve 0.972 g in 2.4 mL of 4 N NaOH; bring vol to 400 mL with 2X dH20 to make a 100X stock (18 n&Z), filter-sterilize, aliquot, and store at -20°C.

394

Vaccanello et al 4 Cholera toxm (ICN, Costa Mesa, CA) Dtssolve 1 mg m 1 mL 2X dH,O, add this to 90 mL DMEM with 10 mL FBS to make a 1000X stock ( 1 2 x lop7 M), filter-sterilize, altquot, and store at -20°C 5 Eptdermal growth factor (#GF-010-9, Austral Btologtcals, San Ramon, CA) Dissolve 10 Clg/mL m 0 1% bovine serum albumm (BSA) to make a 1000X stock, filter-sterilize, altquot, and store at -20°C 6 Hydrocorttsone (Sigma) Dissolve 0 0538 g m 200 mL 2X dH,O to make a 500X stock (7 4 x lo-4 M), filter-stertltze, ahquot, and store at -20°C 7 Insulm (Sigma) Dissolve 50 mg m 10 mL of 0 005 N HCl to make a 1000X stock (5 mg/mL), filter-sterthze, ahquot, and store at -20°C

2 1 2 Fibroblasts for Submerged Keratinocyte Culture (3T3 Feeder Layers) 1 Irradiated 3T3-52 cells are used as feeder cells 2 3T3 media a DMEM (Gibco-BRL) b Bovme calf serum (BCS) 10% (HyClone) c Pemcillu-streptomycin (Sigma) as in Subheading 2.1.1., step 2.b.l d HEPES (Sigma) as m Subheading 2.1.1., step 2.b.2

2.2. Skin Equivalent Cultures 2.2.1. Flbroblasts for Skrn Equwalent Cultures 1 Source of human fibroblasts. newborn foreskin (average size l-2 cm*) 2 Isolation of foreskin fibroblasts a Collagenase buffer 130 mMNaC1, 10 mMCa acetate, 20 mM HEPES Solution adJusted to pH 7 2 and filter-sterilized b Dtspase Make 1OX stock by dissolvmg 5 g m 200 mL ddH20, filter-sterlthze, and store -20°C c Collagenase (Worthmgton Biochemtcal, Freehold, NJ) 3 mg/mL m collagenase buffer d Trypsm Prepared as a 1% stock by mixing 10X PBS (200 mL), 2 g glucose, 2 g trypsm, 0 2 g pemctllm, and 0 2 g streptomycin These components are dissolved m the PBS, then brought to 2 L with ddH*O, filter-sterilized Ahquot and store at -20°C e. EDTA. Prepared as a 0 2% stock by dtssolvmg 2 g m 1 L PBS and adJustmg the pH to 7 45 with 4N NaOH Autoclave and store at room temperature f. PBS. Prepared as a 10X stock by dtssolvmg 1 6 kg NaCL, 165 g Na2HPO,, 40 g KH2P04 and 40 g KC1 m 20 L ddH20 3 Ftbroblast culture medium (FCM) a DMEM (Gibco-BRL) b Fetal bovine serum 10% (HyClone) (Sigma) as m Subheading 2.1.1., step 2.b.l c Pemctllm-streptomycin d HEPES (Sigma) as m Subheading 2.1.1., step 2.b.2

Skin Equivalent

Wound Heahng Model

395

2 2.2. Collagen Matrix 1 A confluent culture of human foreskin fibroblasts (HFF) 2 6-well deep ttssue-culture tray contammg special resm ttssue-culture Inserts with a 3-m-n porous polycarbonate membrane (Organ0 Genesis, Canton, MA) 3 Sterile bovme tendon or rat tall acid-extracted collagen (1 &I 3 mg/mL) m 0 05% acetic actd 4 10X mmtmum essential medmm with Earle’s salts (#12-684F, BtoWhtttaker, Walkersvtlle, MD) 5 Newborn calf serum (HyClone) 6 L-glutamme (200 mM) (# 17-605E, BtoWhtttaker) 7 Sodium bicarbonate (7 1 2 mg/mL) 8 Ftbroblast culture medium

2.2.3 Skin Equivalent Culture Media

6

7 8 9 10 11

DME base modified (#56430-lOL, JRH Biosciences, Lenexa, KS) Ham’s F12 (Gtbco-BRL) L-glutamme (BtoWhtttaker) 200 mA4 1s a 50X stock Hydrocorttsone (Sigma) ITT (Stgma) 500X stock a Bovine insulin Dissolve tn 0 0001 NHCl to a ftnal concentration of 5 pg/mL. b Human transferrm Dissolve m ddH20 to a final concentratton of 5 pg/mL c Tritodothyronme dtssolve m actdtfied ethanol, dilute with double-dtsttlled dH20 to a final concentratton of 20 pM. EOP (Stgma) 500X stock a Ethanolamme* Reconstitute with ddHzO to 1OA M b 0-phosphorylethanolamme. Dilute with ddH20 to IO4 M Adenme (Sigma) Dissolve 0 18 &(500X stock) m actdtfied water warmed m a 37°C water bath Selemous acid (Aldrich, Milwaukee, WI) Dissolve 5 3 x 10-s M (500X stock) wtth ddHzO Calcium chloride Dissolve 132 5 pg/mL (500X stock) m ddH,O Progesterone 2 nM solutton 1s a 500X stock Serum (HyClone) a Chelated newborn calf serum (cNBCS) b Newborn calf serum (NBCS)

2.3. Wounding 1 Sterile acettc 2 Stertle 3 Stertle

of Skin Equivalents

bovine tendon or rat tall acid-extracted collagen (1.0-I acid scalpel (gage 22), long forceps, and dental mirror 100 mm tissue-culture dash

3 mg/mL) m 0 05%

396

Vaccanello et al.

3. Methods 3.1. Submerged Cultures 3. I. 1. Submerged Keratinocyte Cultures Preparation of primary cultures of keratmocytes from foreskins can be followed, as previously described by Emmg and Morgan (IS).

3.1.2 Fibroblasts for Submerged Keratinocyte Culture (Feeder Layers) The clonal growth of keratmocytes requires co-cultivation with a metabohtally active, nonprohferatmg feeder layer of 3T3-52 fibroblasts Grow 3T3-52 fibroblasts in DMEM containing 10% bovine calf serum, HEPES, and pemcllIn-streptomycm. When cells are 90% confluent, they are trypsmlzed, pelleted with centrlfugation at 7508 for 5 mm, and irradiated by a gamma source of 2,000 Cl (Cs- 137, 100% = 1,2 15 R/mm) for 6.5 mm. Irradiated cells are then plated at a density of 2 x 1 06/p 100 m KCM before keratmocytes are added

3.2. Culture of Skin Equivalents 3 2.3 Isolation of Dermal Flbroblasts Flbroblasts are first isolated from human foreskins by using the connective tissue remnant after dispase separation of the eplthelium used for submerged culture (see Subheading 3.1.1.) The connective tissue 1s then rinsed twice m PBS and placed m a 15-mL conical tube with 1 mL of collagenase m collagenase buffer. This mixture IS incubated at 37°C for 30 mm and IS agitated every 5 mm Trypsm EDTA mixed m a 1.1 ratio IS added for 10 mm at 37”C, at which time 1 mL of FCM IS added to inactivate the trypsm, and cells are counted These fibroblasts are grown so that they are densely confluent 1 d before the collagen matrix IS to be cast. At this time, passage the cells at high density, so they will regrow to full confluence the next day, when they will be added to the collagen matrix. This extra passage ensures that a high fraction of fibroblasts are proliferating at the time of mltlatlon of matrix construction (see Note 2).

3.2.2. Construction of the Collagen Matrix (Fig. 3, 16) MIX the followmg components on ice, to generate an acellular and cellular layer for the SE. The goal IS to create a thm layer of acellular collagen that ~111 act as a substrate for the thicker layer of cellular collagen. This will prevent the cellular collagen from contracting completely from the insert 1 Keeping all components on ice, mix the acellular matrix components m the order listed m Table 1 The color of the solution should be from straw-yellow to hght pink, and any extreme varlatlons in color may indicate a pH at which the collagen may not gel If the final solution 1sbright yellow, slowly titrate sodmm blcarbon-

Sk/n Equivalent Wound Healing Model Castmg the dermal culture msert In deep-well

tray

acellular

collagen

397 layer layer

cellular collagen

layer

Lattice contractlon

Epitheltakation

Differentiation

of the EmdermIs

Ftg 3 Dtagrammattc outlme of skm equivalent protocol 1 Culture insert m special deep well tray. 2. A 1-mL layer of acellular collagen is cast over the polycarbonate membrane of the insert 3 A 3-mL cellular collagen layer IS cast onto the acellular gel 4,5 The collagen gel contracts away from the sides of the insert, to form a contracted collagen lattice having a central raised area. 6 A suspension of eptdermal keratmocytes 1s seeded onto the central area 7 The epithelial sheet 1s allowed to develop submerged for 4 d 8,9 The developmg culture 1s raised to the au-ltqutd interface and cultured for an additional 3 d until wounded

ate drop by drop, unttl the appropriate color is noted Add 1 mL to each Insert, making sure the mixture coats the entire bottom of the Insert Once the gel has been poured, It should stand at room temperature, without being disturbed, until tt polymerizes (l&l 5 mm) As the gel polymerizes, the color of the matrix will change to a deeper pmk color (Fig. 3,2) 2 Whrle the acellular matrix layer IS polymertzmg, trypsmtze and resuspend fibroblasts m FCM to a final concentration of 3 0 x lo5 cells/ml Resuspend these cells fully, smce tt will be harder to resuspend them properly once mtxed with collagen

398 Table 1 Components

Vaccarlello et al. for the Collagen

Matrix

Acellular matrix for 6 mL (1 mL/msert) 10X DMEM L-Glutamme Fetal Bovme serum Sodturn Bicarbonate Collagen Ftbroblasts

Cellular matrrx for 18 mL (3 ml/Insert) 165mL 015mL 185mL 0.52 mL 14mL 4 5 x lo5 cells m 1 5 mL fibroblast media

0.59 mL 005mL 06mL 017mL 4.6 mL -

3 For the cellular matrix, agam keep all components on me, and mtx m the order Indicated m Table 1 Again, tf the color needs to be adjusted, carefully titrate m a small amount of sodium bicarbonate (the acellular collagen layers can thus serve as a useful test of gelatton condttlons) The fibroblasts should be added last, to ensure that the mtx has been neutralized by the addttton of collagen, so that the cells are not damaged by an alkaline pH MIX well and add 3 mL to the msert and allow tt to gel at room temperature without disturbing (30-45 mm) (Fig. 3, 3) When the gels are pmk and firm, they are covered wtth 12 mL of FCM (-10 mL m the well and 2 mL on top of the msert) and incubated for 4-7 d, until gel contractton IS stable and complete During thts interval, there 1s a 50-fold decrease m the volume of the matrix, and constderable vertical shrmkage of the matrix IS seen A ratsed, mesa-like area 1s seen m the center of the matrix, and tt 1s here that keratmocytes will be seeded (Fig. 3,3-5)

3.2.3. Growth of Keratinocytes on the Skin Equivalents Keratinocytes

are added to the collagen

matrix

as follows.

1 Use keratmocyte cultures when they are no more than 50% confluent This IS to mmtmtze the number of differentiated cells plated on the matrix (see Notes 1 and 5) Thoroughly remove the 3T3 feeder cells from the culture by mcubatron m PBS-EDTA (10 1 ratto) for 2 mm, and then remove 3T3s by gentle rmsmg Take care not to leave the culture too long m PBSEDTA, because small keratmocyte colonies may detach Rinse with PBS until the 3T3s have been removed 2. Trypsmtze the keratmocyte colonies with Trypsu-EDTA (10.1 ratio) for 5-10 mm and add cells to a 15-mL tube, so that 5 x lo5 keratmocytes will be plated m each insert Centrifuge the cells at 750g for 5 mm 3 Remove all fibroblast media from the inserts containing the contracted collagen matrices 4 Resuspend the keratmocytes so that they can be plated m a small volume, so that each insert will receive 50 pL of suspension containing 5 x lo5 keratmocytes

Skm Equwalent Wound Healing Model Table 2 Media Formulation

399

for growth of SEsa Epldermahzatlon

DME F12 L-glutamine Hydrocortlsone ITT EOP Adenme Selenium CaC12 Progesterone Serum

725 240 20 2 2 2 2 2 2 2 1 (cNBCS)

1

Epldermahzatlon

II

725 240 20 2 2 2 2 2 2 2 1 (NBCS)

uAll volumes are in mL

This can be done by ehmmatmg all residual supernatant above the pellet Use a sterile plastic I-mL plpet with the appropriate volume of epldermahzatlon I media to lift the pellet and transfer It to a sterile Eppendorf tube Then use a 200- or lOOO+L plpetman to gently resuspend the cell pellet m the Eppendorf tube The cell suspension can then be placed m the central, raised, mesa-like portion of the contracted collagen gel with a 200-G plpetman (Fig. 3, 6) 5 Do not touch the plate for 1 5-2 h, while the keratinocytes adhere At this point, add 12 mL of epldermahzatlon I media (-10 mL m the well and 2 mL on top of the keratmocytes), and incubate cultures at 37”C, 7% CO2 (Fig. 3, 7) 6 Cultures are fed for 7 d m the followmg way (media formulations seen m Table 2) a Epldermahzation I media* Submerged m 12 mL, first 2 d b Epldermahzatlon II media Submerged m 12 mL, next 2 d c Epldermahzatlon II media Grown at the air-liquid interface by feeding with 7 mL for 3 d (Fig. 3,8 and 9) The optimization of SE growth will vary with the strain of keratmocyte and fibroblast used (see Note 2)

3.3. Wounding

of Skin Equivalents

SEs are wounded 7 d after the keratinocytes are seeded onto the collagen matrix. The optimal time of wounding may need to be determined (see Note 3) SIX days before cultures are to be wounded, an additional collagen matrix IS fabricated according to Subheading 3.2.2. This will be used as the substrate onto which the wounded SE will be transferred. 1 Aspirate all media from the SE, remove the insert from the 6-well plate, and place it upside down onto a sterile 100 mm dish

400

Vaccariello et al.

2 Use the scalpel to cut out the entire polycarbonate membrane around the penphery of the insert Place the cut-out SE onto a 100 mm dish, so that It rests on its polycarbonate membrane 3 Trim the culture with the scalpel several mm from its raised, mesa-like region This removes parts of the SE without keratinocytes (since keratmocytes are mltlally only seeded on the center mesa of the matrix) and faclhtates removal of the culture from the insert membrane for transfer 4 Place the scalpel’s edge directly m the center of the culture and rock the blade back and forth, m order to create an mclslon 1 2 cm m length This mclslon completely penetrates the epidermis, collagen matrix, and membrane (Fig. 4A) The culture should not be completely cut m half, and the two sides are to be kept attached 5 At this point, place 2630 pL of unpolymerized collagen onto the second collagen matrix that the culture 1s to be transferred to (see Note 4) This 1s to promote adherence of the two matrices, and to prevent the keratmocytes from migrating between the two matrices 6 Use the forceps to gently lift the edge of the culture, so that it separates from the polycarbonate membrane At this time, bring the dental mirror, which will serve as a spatula, close to the culture and drag the culture onto the mirror with a long forceps, leaving the membrane behind The transfer will be facilitated if the mlrror IS slightly molstened with media 7 Unfold any wrinkles m the culture by moving it back and forth on the mirror, using the forceps Once the culture 1s smooth, pull one side of the culture slightly over the edge of the mirror, this 1s the site where the culture will be transferred onto the new matrix. 8 Bring the mirror directly over the second collagen matrix and lower it inside the insert, so that the edge of the mirror and the culture are m contact with the matrix (see Note 7) Slide the culture onto the collagen, using the forceps, as the mirror 1s pulled away slowly, leaving the culture on the collagen matrix (Fig. 4B) 9 Using the forceps, tease apart the mclslon m order to create an elllptlcal space between the two halves. This elliptical space will be 2-3 mm at its greatest width It 1s important that the transferred culture 1s kept completely free of any folds or wrinkles 10 Maintain the culture at air-liquid interface by adding 7 mL of epldermlhzatlon II media to the outer well (see Note 6) Incubate at 37°C with 7% CO*, and change the media every 2 d until the end of the experiment (Fig. 4C,D)

An example of a wounded culture 3 d after wounding

3.4. Processing

is seen in Fig. 5

of Skin Equivalents

Skin equivalents are fragile if handled improperly and require special care during processing. The followmg are steps that should ensure that samples can be cryosectloned without artlfactual damage.

Skin Equivalent

Wound Healing Model

401

Keratmocytes Loyer I Layer 1

I I

D

I I Kerotinocytes Layer II Layer I Membrane

Fig. 4 ConstructIon of composite organotyplc coculture woundmg model (A) Schematic of stratified keratmocyte sheetgrowmg on contracted collagen matrrx contammg fibroblasts (layer II) This skm eqmvalent has been cultured at the air-llquld mterface for 3 d. A wound IS formed by mcismg this culture (dotted lme). (B) The wounded culture IS then transferred onto a secondcontracted collagen matrix, which hasbeen prepared 6 d m advance (layer I) (C) The resultant wounded skm equivalent consistsof two layers of contracted matrix and one layer of eplthelium Wound margms are seenat the transition zone from layer I to layer II, and are noted with arrows (D) Followmg re-eplthellalrzatlon, stratified eplthellum covers the wound bed Reprmted with permlsslonfrom ref. I4 1 Cultures are rmsed twice m PBS, and SEs are separatedfrom the plastic culture Insert with a scalpel blade Inclslon should Include the polycarbonate membrane,

Vaccariello et al.

Fig. 5. Skin equivalent several days after wounding. A skin equivalent was wounded and transferred to a second collagen gel, as described. The original wound margins are marked by arrows.

because specimens will be easier to handle if still on the membrane. At this point, the wound can be sectioned to provide equal halves for formalin and frozen processing. 2. Immerse the SE for frozen section analysis in 1 M sucrose at room temperature for 1 h. This will remove some of the water from the collagen matrix and prevent fracture artifact upon freezing. After removal from sucrose, the SE should be noticeably firmer (should not flop over when standing on edge). As an alternative, SEs may be kept overnight at 4°C in 2 M sucrose, but incubation for any longer than this may result in the culture becoming somewhat brittle and difficult to cut. 3. A slow-freezing technique, in which the sample is frozen in liquid nitrogen vapor, is preferred, rather than snap-freezing in liquid nitrogen. This can be accomplished by using a three-quarter inch bottle cap as a template to make a mold with aluminum foil. Fill the mold with embedding compound (Tissue Freezing Medium [TBS]) to a thickness of one-half inch and place it on a test tube rack into a Styrofoam box filled with liquid nitrogen just below the upper surface of the rack. 4. In order to embed the tissue specimen on edge, the embedding compound must be cooled to increase its viscosity. Place the aluminum mold on the rack so that it starts to chill in the liquid nitrogen vapor. Within 30 s, the embedding material turns white near its edges, and should be sufficiently viscous in the center to allow the culture to be stood on edge (The TBS embedding compound is more viscous than most, and is well-suited for this purpose). If necessary, hold the SE upright with a forceps. The SE will freeze within several minutes and specimens can be placed in a pillbox for storage at -70°C.

Sk/n Equrvalent Wound Healing Model

403

4. Notes Opttmtzatton of SEs is essential in generatmg a responsive organotypic woundmg model, and can be accomphshed by assuring the presence of keratmocytes, which have a high growth potential m the SE Most keratinocytes plated onto the collagen matrix ~111 adhere well, but only cells capable of rephcatton will grow after plating Keratmocytes that have undergone termmal dtfferentiatton whtle m submerged culture wtll not generate a well-strattfied SE, and will not re-epithehahze after woundmg It IS therefore important to grow keratmocytes, so that a high growth fraction IS present m the cultures at the time of passage to SEs By growing keratmocytes as small colonies at high clonal density m submerged cultures on 3T3 feeder layers, terminal differenttatton will be mmtmtzed and the fractton of rephcatmg cells ~111 be optimized Alternatively, this can also be accomplished by growmg keratmocytes m a low calcmm media before passage to organotypic culture (13) We have found that opttmal growth of SEs varies wtth different keratmocyte and fibroblast strains used Keratinocyte strams should be tested to determine those provtdmg the best morphologic differenttatton of the SE It IS also apparent that the rate of re-epithehahzation varies wtth the keratmocyte strain used In addition, we have found great vartabthty m the degree to whtch fibroblast strains support keratmocyte growth This appears to be related to the degree to which fibroblasts are able to contract the collagen matrix. Fibroblast strains demonstrating more collagen gel contractton are better able to support keratmocyte growth, and may be used to determine fibroblast strains for optimal SE growth Another vartable that needs to be determined for each keratmocyte strain is the ttme after airlift at which SEs should be wounded. We find that culture at the anhqutd interface for longer than 10 d results m an overly thickened stratum corneum, a flattened basal cell layer, and a failure to respond to growth stimuli. It is important that cultures not demonstrate this morphology at the time of woundmg In our hands, cultures grown at the air-hquid interface for 3-4 d provtde the most responsive tissue for woundmg At this pomt m its maturation, the epithehum is well-stratified, demonstrates a thm stratum corneum and cuboidal basal cells, and shows indices of basal cell prohferation that approximate those m skin The small amount (20-30 pL) of nonpolymertzed collagen added to the lower collagen matrix at the ttme of transfer of the wounded construct serves two purposes Ftrst of all, it acts as a biological glue that promotes adherence of the two collagen layers Secondly, it prevents the vertical migration of the epithelmm between the two layers of collagen after transfer. Only when keratmocyte migration is limited to the wound surface do assays for keratmocyte phenotype during re-epithehalization become meanmgful. For example, if keratmocytes mtgrated between the two layers of collagen, it would be tmposstble to measure prohferanon of keratmocytes at the wound margin. Several points regarding some subtlettes of keratmocyte behavior m SEs are worth mentionmg The first concerns keratmocyte growth potential. Although SEs demonstrate a basal level of prohferatton that is similar to skin, these cul-

404

Vaccariello et al

tures have tremendous potenttal for prohferatton, and are very responsive to external growth sttmuh Dtsruptton of the cultures by wounding can result m a IO-fold Increase m basal cell prohferatton (26), and addition of growth factors can modify prohferatton, as well (I 7) Secondly, it should be kept m mmd that, although keratmocytes grown m SEs share many features m common with skm keratmocytes, they do differ somewhat m phenotype. For example, mtegrm receptors not normally expressed m skm are constttutttvely expressed m keratmocytes In addmon, SEs are somewhat deficient m barrier function The phenotype of keratmocytes grown m SEs have been compared to a newly re-eptthehaltzed wound m vtvo, m which morphologtc dtfferentiation 1s complete, but cells are still m a somewhat activated state 6 An advantage of the SE wounding mode1 1s the ability to add soluble reagents directly to the cultures during re-eptthehahzatton This has been useful m followmg the effects of soluble mediators, such as TGF-P, on re-eptthehahzation, as well as on tibroblast and keratmocyte phenotype Addition of reagents that assist m analysts of cell phenotype during re-epithehahzatton, such as a pulse of BrdU to determine proliferation mdtces, can also be added directly to media and mcubated with cultures. 7 Perhaps the most difficult technical aspect of the model 1s the skill m transferring the wounded organotypic construct to the second collagen bed An optimal SE construct has good tensile properties and should not tear during the transfer The key lies m the manipulation of the dental mtrror, which 1s used as a spatula to pick up and deliver the construct This mirror has proved useful because of the size of its face and the angle of the bend m us handle When the mirror IS slightly moistened, the wounded construct can be pulled directly onto the mirror from its membrane To deliver the construct, hold the face of the mirror at a 45degree angle to the surface of the second matrix in tts insert First, allow the edge that is slightly overhanging from the mirror to touch the second matrix Then, while holdmg the culture with a forceps at this area of contact, the mirror can be slowly removed from under the construct

References 1 Clark, R A F. (1996) Wound repair. overview and general considerattons, m The Molecular and Cellular Bzology of Wound Repaw (Clark, R A F , ed ) Plenum, New York, pp l-50 2 Hertle, M D., Kubler, M -D , Leigh, I M , and Watt, F M (1992)Aberrant mtegrm expression during eptdermal wound healing and m psoriatic epidermis J Clzn Invest 89, 1892-l 901 3 Viziam, C B , Maltotsy, A. G , and Mescon, H (1964) Eptthehahzation of small wounds J Invest Dermatol 43,499-507 4 Mansbrtdge, J N and Knapp, A. M. (1987) Changes m keratmocyte maturation during wound healing. J Invest Dermatol 89,253-263. 5. Freeman, A E , Eigel, H J , Herman, B. J , and Klemfeld, K L (1976) Growth and charactertzation of human skm epithehal cell cultures In Ktro 2, 352-358

Sk/n Equivalent

Wound Healing Model

405

6 Stenn, K. S. (1978) The role of serum in the eptthelial outgrowth of mouse skm explants Br J Dermatol 98,411-416 7 Marks, S and Nrshikawa, T (1973) Active epidermal movement m human skm m vitro Br J Dermatol S&245-248 8 Hmtner, H , Fritsch, P 0 , Foldart, J-M, Stmgl, G , Schuler, G , and Katz, S I (1980) Expression of basement membrane zone antigens at the dermo-epibohcJunctron m organ cultures of human skm J invest Dermatol 14,200-204 9 Stenn, K S , Madri, J A , Tmghitella, T., and Terranova, V P (1983) Multiple mechanisms of dissociated epidermal cell spreading. J Cell Bzol 96,63S67 10 Stenn, K S and Milstone, L M (1984) Eptdermal cell confluence and tmphcattons for a two-step mechamms of wound closure. J Invest. Dermatol 83,445-447. 11 Woodley, D T , O’Keefe, E J , and Prumeras M (1985) Cutaneous wound healmg a model for cell-matrix mteractions J Am Acad Dermatol 12,42CL433 12 Bell, E , Ehrhch, H P, Buttle, D J , and NakatsuJi, T (1981) Living tissue formed m vitro and accepted as &m-equivalent tissue of full thickness Sczence 211, 1052-1054. 13 Parenteau, N (1994) Skm equivalents, m The Keratznocyte Handbook, vol 11) (Leigh, I. M and Watt, F W, eds ), Cambridge University Press, pp 45-56 14 Garlick, J A and Taichman, L B (1994) The fate of human keratmocytes durmg re-eptthehallzatton m an organotypic culture model Lab Invest 70, 9 16924 15 Winter, G D (1972) Epidermal regeneration studted m the domestic pig, m Epzdermal Wound Healzng (Maibach, H I and Rovee, D T , eds ), Year Book, Chtcage, pp 7 l-l 112 16 Garhck, J A and Taichman, L B (1994b) Effect ofTGF-l3 1 on re-epithehahzatton of human keratmocytes m vitro an organotypic model J Invest Dermatol 103, 554-559 17 Garhck, J A , Parks, W C , Welgus, H. G , and Taichman, L T (1996) Re-epithehalization of oral keratmocytes m vitro J Dental Res 75, 912-9 18 18 Emmg, S A and Morgan, J R (1996) Methods for the use of genetically modtfed keratmocytes m gene therapy, m Methods m Molecular Medxzne, Gene Therapy Protocols (Robbms, P , ed ), Humana, Totowa, NJ

30 Preparation and Transplantation of a Composite Graft of Epidermal Keratinocytes on Acellular Dermis Daniel A. Medalie and Jeffrey

R. Morgan

1. Introduction Loss of skm because of burns or ulcers is a major medtcal problem, and IS the Impetus for the development of skm substitutes and skin replacement technologies Efforts in this area have focused on developing suitable substitutes for the epidermis or for the dermis, as well as ways to combme both technologies m a composrte skin. In thuschapter, we describe the constructton of a skm substitute of cultured human keratmocytes on acellular human dermis. Among the first efforts m this field was the development of a replacement for the epidermis by culturing keratmocytes, the primary cells of the epidermis (I). When grown to confluence, these cells form multicell layered eptthelral sheets which can be detached from the culture dish, secured to gauze and grafted to skin defects (1). Several chmcal trials have tested the efficacy of these epttheltal sheets, as eptdermal replacements for burns and ulcers (2,3). Despite scattered reports of the success of cultured eptthelium (both autografts and allografts), their use is not standard care m the vast majortty of burn centers (4,5). When compared to the success of splttthickness grafts, cultured eprthehal sheets have defrcrencies m several areas, mcludmg the mttral graft take, adherence to the graft bed, susceptlbthty to infection, prevention of wound contracture and fragility of the graft sheets (6-S). Other efforts have focused on the development of a substitute for the dermis m addmon to the epidermis, and several types of dermal matrices have been developed. Addmon of a dermal matrtx to epithehal replacements adds the theoretical advantage of a thicker, more durable graft that more closely resembles the “gold standard” of split-thickness skin grafts. One of the first From

Methods E&ted

m Molecular

by J R Morgan

Medmne,

Vol 18 T/sue

and M L Yarmush

407

Engineering

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

408

Medalie and Morgan

successeswas a porous collagen andglycosarmnoglycan(GAG) matrix (9,IO) which was destgned to be biodegradable and to serve as a template for the mgrowth of fibrovascular components.This collageflAG matrix has been modified by others so that cultured fibroblasts can be incorporated mto the matrix, and cultured keratlnocytescan be seededonto the surface to form an epldermls (11-14). Another early successwas the development of a skin equivalent based on a fibroblast contracted collagen lattice (15). Fibroblasts, seeded m gelled collagen, contract the lattice into a dermal analog, and keratmocytescan be seededonto this structure to make a skm equivalent (15,16). Other shn substitutesbemg developed include fibroblast seeded meshes of biodegradable fibers of polyglycohc acid (PGA) and polyglactm (PGL), or nondegradable nylon fibers. The biodegradable matrices have been seededwith keratmocytesand transplanted to athymlc mice, m which the fibers are absorbed m several weeks (I 7,18). We and others have had successwith a dermal analog of acellular human dermis that IS de-eplthehahzed and rendered completely acellular (29-24). Acellular dermis IS relatively nonnnmunogemc and retains many of Its structural elements after processmg The material IS also durable, and can be lyophlhzed and stored at room temperature (22,25-27). Composite grafts using the acellular dermis can be formed m vitro (2428-30) and subsequently transplanted to athymlc mice, generating a well-differentiated and fully pigmented epidermis with many of the same characteristics as normal skm (1%21,23,31) In this chapter, we describe methods for the preparation and storage of acellular human dermis, and the generation of composite tissue constructs by seedmg cultured keratmocytes onto the acellular dermis. We then discuss how to successfully transplant these grafts to open wounds on athymlc mice, and subsequently harvest grafts for analysis. 2. Materials 2.1. Cell Culture 2.1. I. Fibroblast Feeder Layer 1 Flbroblast feeder layer 3T3-52 mouse fibroblast cell line (ongmally provided by H Green, Harvard Medical School, Boston, MA) 2 Flbroblast tissue-culture medmm: Dulbecco’s modified Eagle medium (DMEM, high glucose, L-glutamme, 110 mg/L sodium pymvate, Glbco-BRL), bovme calf serum 10% (HyClone, Logan, UT), pen~cllhlrstreptomycm (Boehrmger) 100 IU/mL-1 00 pg/mL 3 Mltomycm C (Boehrmger, Mannhlem, Indlanopohs, IN) 1.5 pg/mL m serumfree DMEM

2 7 2. Keratinocytes I Source for human keratmocytes newborn foreskms 2 Keratmocyte tissue-culture medium (KCM)

(average size, l-2 cm2)

Epldermal Keratinocyte Composite Graft

409

a b c d

3 4 5.

6

DMEM/Ham’s F12 medium (Grbco-BRL, Gatthersburg, MD) (3 1) Fetal bovme serum 10% (JRH Bioscience). Pemcillm-streptomycin 100 IU/mL-100 pg/mL Adenme (6-ammopurme hydrochloride, Sigma, St LOUIS, MO) make up fresh at time medium IS prepared, prepare stock (50X) of 1 2 mg/rnL m DMEMiF 12 (3 l), adJust pH to 7 5 with 1 N NaOH, sterihze by filtration using a 0 45-m filter, add 2 mL of stock to 100 mL of KCM; final concentration 1 8 x 1tY4 M e Cholera toxin (Vtbrio cholerae, type Inaba 569 B, Calbrochem, La Jolla, CA) Piepare concentrated stock of 1O-5 M m dHzO, store at 4”C, take 0.1 mL of concentrated stock and make up to 10 mL with DMEM (10% fetal calf serum), sterrhze by filtration usmg a 0 45-pm filter, and ahquot m 1-mL portions (lO-’ M), store at -20°C; add 0 1 to 100 mL of KCM; final concentratton 1O-lo M f Epidermal growth factor (mouse, Collaborattve Btomedtcal, Bedford, MA) Resuspend lyophihzed maternal m dH20 to prepare stock of 10 Clg/mL, sterilize by filtration usmg a 0 45-pm filter, and ahquot m 1-mL portions, store at -20°C add 0 1 mL to 100 mL KCM wrth first medium change, final concentration 10 ng/mL g. Hydrocortisone (chromatographrc standard, Calbrochem). Prepare stock of 5 mg/mL m 95% ethanol, store at 4”C, take 0 4 mL of stock and make up to 10 mL with serum-free DMEM, sterrhze by filtration using a 0 45-mn filter, and allquot m 1-mL portions, add 0 2 mL to 100 mL of KCM; final concentrahon’ 0 4 pgg/mL h Insulm (pork, 100 UimL [3.8 mg/mL], Novo Nordisk, Danbury, CT). Add 0 13 mL of stock to 100 mL of KCM, final concentration 5 Clg/mL 1 T/T3 stock (transferrm/triiodo-L-thyronme stock) T stock (transferrm, human, partially n-on-saturated, Boehrmger) 5 mg/mL in PBS; T3 stock (3,3’,5-trnodo-L-thyronme, sodium salt, Sigma). Dissolve 13.6 mg m the mmimum amount of 0 02 N NaOH, and make volume up to 100 mL with dHz0, sterilize by filtration using a 0 45-pm filter, and store at -20°C (2 x lOA M), add 0 1 mL T3 stock to 9 9 mL of T stock, sterilize by filtration using a 0.45~pm filter, and ahquot in I-mL portions, store at -20°C add 0 1 mL of T/T3 stock to 100 mL of KCM; final concentration transferrm 5 pg/ mL, trnodo-L-thyromne 2 x IO-9 M PBS (phosphate buffered saline) 138 mM NaCl, 2 7 A4 KC], 8 1 mA4 Na,HPO,, 1 5 mA4 KH2P04, sterilize by filtration using a 0.45~pm filter EDTA solution ([ethylenedmitriolo]tetraacetic acid dtsodmm salt, Boehrmger). 5 mM m PBS; sterilize by filtration using a 0 45-pm filter Trypsm solution (trypsm l-300, ICN Biochemicals, Costa Mesa, CA). n-dextrose 0 1% (w/v), trypsm 0 1% (w/v) m PBS, pH 7 5; sterilize by filtration using a 0.45-pm filter, store at -20°C avoid repeated thawing-freezmg Trypsmtzmg flask (25 mL, Wheaton Scientific, Millville, NJ)

2.2. Harvesting

of Human Cadaver Skin

1 Donor skm preparation 70% isopropyl alcohol, Betadine surgical scrub and prep solution, topical mineral 011 (sterile), 0.9% sodium chloride irrigation solution, 0 R scrub brush, disposable shaving razors.

410 2 3 4 5 6 7

Medalie and Morgan Padget or Ztmmer dermatome Dermatome blades. Stertle drapes, stertle 4 x 4 sponges Sterrle specimen contamers. Sterile gowns, gloves, mask, and protective goggles Sterrle forceps and #I 5 sterile drsposable scalpels (Feather Safety Razor Co , Japan)

2.3. Storage of Human Cadaver Skin 1 2 3 4 5. 6 7

Stertle nylon gauze Sterrle drapes Cryopreservatron fluid (15% glycerol m Ringer’s lactate) Surgical mstruments. sterile scrssors, tissue forceps, sterrle basms Sealable plastrc or aluminum pouches Cryomed programmable freezer Llqurd nitrogen

2.4. Preparation

of Acellular

Human Dermis

1 Human cadaver dermts or freshly obtained skin (see Note 1) 2 Sterile PBS or DMEM (Grbco-BRL) 3 Anttbtotrcs Gentamycm (100 ug/mL), ctprofloxacm (10 pg/mL), amphoteracm B (2 5 pg/mL), pemcrllu-streptomycin (100 IU/mL-100 pg/mL) 4 Surgical mstruments Two pairs of fine forceps (tweezers)without teeth, # 10 sterile drsposable scalpels (Feather) 5 Sterile glass pipets

2.5. Preparation 1 2 3 4

of Composite

Grafts in vitro

35-mm trssue-culture dishes Circular steel mesh screens (Fisher, Putsburgh, PA) Surgical mstruments two pans of fine forceps (tweezers) Sterile glass pipets

2.6. Grafting

and Harvesting

of Composite

without teeth

Grafts

Surgical instruments Small, sharp scrssors, stertle disposable scalpels (Feather), two pairs of tine tissue forceps (tweezers) wrthout teeth, small needle driver, sterrle glass ptpets, l-cc tuberculin syringes, razor blades 6-O nylon sutures (Ethrcon, Somervrlle, NJ), with a small (p- 1) cutting needle Skin prep 70% Isopropyl alcohol, Betadme surgical prep solutron, tmcture of benzom skin adhestve, Neosporyn or Bacrtracm antrblotlc ointment Skm dressing Telfa no-stick gauze (Kendall, Mansfield, MA), Tegaderm polyurethane occlusrve dressmg (3M) (5 x 6 5-cm size), 3M flexible Sports Band-Atd (must be at least 718 to 1 m wide), OS-in wade waterproof adhesive tape (Johnson and Johnson) Protractor or template that can be used to draw squares on the back of the mouse

Epidermal Keratinocyte Composite Graft

411

6 Anesthesia (2,2,2-tnbromoethanol, Aldrich, Milwaukee, WI)* Prepare concentrated stock of 1 6 g/mL m 2 methyl-2 butanol, and store at 4°C Before use, prepare a working solution by diluting 12 pL (concentrated stock)/mL 0 9% NaCl at 40°C Sterilize by filtration using a 0 45-w filter 7 35-mm tissue-culture dishes 8 Formaldehyde (Sigma 37%) Dilute stock 1 10 (v/v) m PBS. 9 Dry ice (block form) 10. Camera for photo documentation (with macro lens) 11 Transparency film for tracing the grafts 12 Athymlc nude mice, NIH Swiss nu, 6-8 wk, average weight 20-30 g (outbred, TAC*N;NIFS-nuDF, Tacomc, Germantown, NY). 13 Dental wax (Byte ryte, Mlzzy Inc , Cherry Hill, NJ)

3. Methods 3.1. Harvest of Human Cadaver

Skin

Cadaver skin with an average thickness of 0.0 I5 m can be obtained fresh or from any local skin bank. If it is obtained from a skin bank, it should have been cryopreserved according to the bank’s protocols, using glycerol and controlled rate freezing and should be CMV-, hepatitis B- and HIV-negative Described below IS the method used for harvesting skm suitable for use in clmical and laboratory sltu-

atlons. (see Notes 1 and 2). 1. Prospective donor sites are completely shaved of all body hair with disposable razors, and then the area to be harvested 1swashed with regular tap water The person doing the prep and harvest does a full surgical scrub and then gowns m sterile attire, including mask, cap, and goggles Prepping IS performed m a circular fashion, starting m the center and working out First the Betadme surgical scrub is applied to the skin and the skm IS scrubbed gently for several minutes with an OR scrub brush The area 1s washed with sterile salme and then pamted three times, using 4 x 4 sterile gauze sponges m a circular motion with the Betadme skm-prep solution The Betadme is allowed to dry for 5 min, and then the skin 1swashed with 4 x 4 gauze soaked m 70% Isopropyl alcohol prep solution The area to be harvested 1s now draped and all unprepped areas of the body are covered with sterile sheets or towels The person who prepped the area now regowns and regloves, or a second person harvests the grafts. A Padget or Zlmmer dermatome is used for the skm harvest Lock a fresh dermatome blade mto the apparatus, and then screw the maximum-width guard tightly down onto the blade To test the cahbratlon of the dermatome, set the skm harvest thickness at 0.0 15 m and slide a # 15 scalpel blade between the dermatome blade and guard The cutting portion of the scalpel blade should just fit between the two This indicates that the cahbratlon setting of 0.0 15 m 1scorrect.

412

Medal/e and Morgan

10 Place the bevel of the dermatome on the donor site with the rtght hand The angle of the dermatome to the skm is approx 45 degrees The left hand rests firmly on the skm behind the Instrument and provides firm countertraction Depress the control button or lever and apply a steady and firm forward and downward pressure to begin the skm removal To complete the cut, tilt the dermatome gradually, with the bevel upwards, and release the control 11 Remove the skin from the dermatome with sterile forceps and place m a sterile container with a small quantity of salme These containers can be stored at 4°C for 48 h or placed directly mto a deep freezer. 12 In a lammar hood, pour cryopreservatton medium mto specimen containers and transfer the cadaver skin mto the containers These are then returned to the refrigerator for 2 h 13 The skin is then removed from the contamers and spread flat onto glycerohzed sterile mesh gauze and placed mto sealable plastic or alummum pouches 14 The skm IS now ready to be cryopreserved and can be frozen m the Cryomed freezer, according to local skm banking protocols We recommend a freezing rate of l”C/mm (see Notes 2)

3.2. Preparation

of Ace/Mar

Human Dermis

Several authors have recommended ways to prepare de-eprdermallzed, lular dermis (22,27) We have followed a similar protocol.

acel-

1 The cryopreserved skm is warmed and then subjected to three rapid freeze-thaw cycles m liquid nitrogen to devltaltze the cells, washed three times m sterile PBS, and then incubated at 37°C for 1 wk in sterile PBS with antibiotics (gentamycm at 100 pg/mL, crprofloxacm at 10 ug/mL, amphoteracm B at 2 5 pg/mL, pemctlIn-streptomycin at 100 IU/mL-100 ug/mL) 2 At the end of 1 wk, the epidermis can be gently stripped or scraped from the dermis with forceps m a laminar flow hood under strict sterile condmons Eprdermis that remains adherent ~111 detach spontaneously m the followmg weeks 3 The dermis 1smamtamed m antibiotic solutton at 4°C for 4 more wk to remove any remaining cells Again, stenle technique at all times is essenttal (see Note 2) 4 Prior to use, the acellular dermis should be washed three times with DMEM to remove residual antibiotics (see Note 3)

3.3. Culture of Human Keratinocytes The description of how to culture human keratmocytes using a fibroblast feeder layer 1s beyond the scope of this chapter, but 1s described m excellent detail elsewhere (I). Included m Subheading 2.1. of this chapter are the specolic culture media that we have used and found successful for cell culture,

3.4. Preparation

of Composite

Skin Grafts

1 The dermts 1scut into 1.25-l 5 cm2 pieces m a sterile fashion using forceps and a # 10 scalpel blade, and then each piece is placed mto a 35-mm tissue-culture dish,


413

papillary side up The paptllary side of the dermis can be dtstmgmshed from the reticular side by a rougher feel and duller sheen (see Fig. 1). To facilitate adherence of the dermis to the plastic, crosshatch scratches should be made on the bottom of the dish with a glass ptpet, and then the dermis is allowed to Incubate in the dish for 1 h at 37°C. 2 Keratmocytes that have been grown m culture are seeded onto the dermts by combmmg 500,000 freshly trypsnnzed keratmocytes m 0 25 mL of keratmocyte cultute medium, and then applymg the mixture to the surface of each ptece of dermrs If posstble, try to use the surface tensron of the culture medium to mamtam the cells on the dernns, rather than allowing the medium to run over the edge of the graft (see Note 4) 3 After 2 h of mcubatton at 37°C to allow prelimmary cell attachment, an additional 1 75 mL of medium is gently added to the dish. Culture medium 1schanged every 3 d 4 The composite grafts are maintained submerged m culture for a mimmum of 5 d Durmg thts ttme, a thm layer of keratmocytes (3-5-cells thick) will cover the papillary surface of the acellular dermts The composite grafts can then be transplanted onto mice Unlike grafts of sheets of keratmocytes, the compostte grafts can be maintamed submerged tn culture for up to 2 wk wtthout apparent detriment (and even longer if raised to the an-hqutd interface) (see Note 4)

3.5. Preparation at the Air-Liquid

of Composite Interface

Skin Grafts Grown

Compostte grafts that are maintained m the submerged state form a very thin eptthelmm. To generate grafts wrth a thicker, more stratified eptthebum, the composites can be prepared m the exact same way as described above, but after being submerged for 3 d, they should be ratsed to the air-liquid interface (24). We use cncular steel mesh screens that are approximately the stze of a quarter, and are bent at the edges to allow them to sit approx 3 mm off the bottom of a tissue-culture dish The graft should be gently lifted at two comers with two pairs of forceps, and draped over the screen Smooth out any wrinkles m the graft, but avotd scraping the surface of the graft The culture medium (KCM) should then be added to the dash, unttl tt 1s level wtth, but not submerging, the graft In this way, the dermis is exposed to the medmm, and the keratmocytes are exposed to an The keratmocytes survtve by diffusion of nutrients through the dermis, and ~111 begin to form a thicker, more stratified eptthelmm (Fig. 2, see Note 5). Grafts should be left at the an-hqutd interface a munmum of 7 d, and can be left for up to 3 wk

3.6. Grafting

of Composite

Skin Grafts

Several methods have been descrrbed for the transplantatron of human skm or skm equivalents to open wounds on athymrc mice. The method described m

414

Medal/e

and Morgan

this subheading was found by our group to maxtmize graft survtval. The entire surgical procedure must be performed on anesthetized ammals, usmg sterile conditions All surgical procedures are performed in a lamtnar flow hood. The graftmg procedure, when mastered, requires about 30 mm/mouse 1 Male nude mice can be rendered passive with a very brief CO2 or ether treatment and then defimttvely anesthettzed with approx 0 7-0.8 mL of 2,2,2-trtbromoethanol (mtraperttoneal inJectton with 25gage 314 m. needle, 30 p.L workmg solution/g of mouse). 2 When the animal is completely unconsctous, it is transferred to a sterile operating environment 3 The dorsum of the mouse 1s washed first with 70% alcohol and then Betadme The Betadme is allowed to dry, and then a 1 25-l 5 cm* square is marked out on the mouse’s side (the square should be no larger than the graft to be transplanted) The medial cephalad-caudad mark should be made Just to the right or left of the midline, to avoid grafting over the spine. The cephalad medial-to-lateral mark should be made lust inferior to the shoulder blade of either frontlimb. 4 Sharp scissors are used to create a full-thickness (including panmculus carnosus) defect down to the fascia of the dorsal musculature, and then the enttre piece of skm 1sremoved by sctssor dissection Bleeding may be encountered when coming across the anterior and postertor fat pads, and can be controlled by l-2 mm of gentle pressure (see Note 6) 5 Prior to removing the composite graft from the culture dish, a glass pipet or needle should be used to cncumferentially scrape away any keratmocytes on the dish directly adjacent to the dermal composite During culture, keratmocytes will grow off the edges of the dermis and attach to the plastic tissue-culture dish. If not removed, these keratmocytes may act as a tether that pulls the eptthebum from of the dermal graft when tt IS lifted from the dish 6 To transfer the graft to the wound, a square piece of Telfa no-stick gauze, which IS roughly the same size as the graft, is laid over the graft m the culture dish Two forceps are used to grab an edge of graft, along with Telfa, and lift tt from the dish The graft adheres lightly to the Telfa and thus does not bunch up The graft is then laid mto place over the wound, and the Telfa IS gently removed (see Fig. 3A) 7 Once on the wound, etther the graft or the wound edgesare tnmmed to generate a precise fit of the graft mto the defect (a startmg grafi slightly bigger than the defect is ideal) 8 A 6-O nylon stitch IS used to anchor the graft at each corner of the wound It is important to include mouse skin, mouse underlying musculature, and the graft with each stitch 9 Antibiotic ointment is applied to a square piece of Telfa, which is then applied to the graft One more piece of Telfa is placed on top of the first piece (to act as a compressive dressmg), and then both pieces are secured with several 6-O nylon sutures (tacked to the surroundmg mouse skin, not the graft). 10 Tmcture of benzom (which makes the mouse skm sticky) is applied to the mouse skm on all sides of the wound


415

Fig. 1. Acellular human dermis that is oriented with the papillary side up. The top half of the dermis (arrow) has been flipped over to show the reticular side.

Fig. 2. Keratinocytes are seeded onto the dermis and allowed to grow submerged for 1 wk (data not shown). When subsequently raised to the air-liquid interface, they form a thicker, more stratified epithelium. Note the completely acellular dermal matrix.

11. A 5 by 6.5 centimeter piece of Tegaderm brand polyurethane occlusive dressing is cut in half and one half is applied over the Telfa dressing. The Tegaderm, which is adhesive, should form a tight seal for several millimeters around the wound and adhere tightly to the mouse skin (aided by the Benzoin) (see Fig. 3B). 12. At this time, approx 0.7 mL of keratinocyte culture medium is injected with a tuberculin syringe through the Tegaderm into the Telfa pads to keep the grafts moist for the first several days (see Fig. 3B). 13. Next, a 3-M flexible Sports Band-Aid is trimmed so that only the pad and a thin (5mm) adhesive strip around the pad remains. This Band-Aid is then placed over the Tegaderm on top of the wound. A 6-O nylon suture is used to sew the Band-Aid to the surrounding mouse skin (Tegaderm included). This is a running “baseball” stitch that travels around the entire circumference of the Band-Aid (see Fig. 3C). 14. Finally, l/2 in. waterproof tape is wrapped circumferentially around the mouse to cover the band-aid. Care must be taken not to wrap too closely to the mouse’s hind legs, which might get caught in the tape as the mouse walks around after the surgery.

416

Medalie and Morgan

Fig 3. A typical graft after 1 wk of culture (A) is placed onto a full-thickness defect on the dorsum of an athymic mouse. The graft is dressed with Telfa and Tegaderm (B) and then a flexible band-aid (C). 15. If the whole procedure is performed correctly, the mouse retains its mobility and the dressing remains occlusive for at least 4-7 d (enough to allow epithelialization of the gaps between the mouse skin and the graft). The dressing should not be changed for at least 1 wk, and can, in fact, be left on indefinitely (see Note 7).

3.7. Graft Harvest In as early as 7 d, the grafts integrate with the surrounding mouse skin. The epithelium undergoes further stratification and the dermis is infiltrated by mouse fibroblasts and capillaries (19,21). By 6 wk, a fully differentiated, pigmented epidermis is generated that is quite durable and demonstrates a normal protective barrier, as evaluated by transepidermal water loss (Fig. 4A) (20,23). The dermis is well populated by host fibroblasts and blood vessels, and passenger melanocytes have multiplied and repopulated the basal layers of the epidermis (Fig. 4B; see Note 8) 1. To harvest the graft, the animal should be sacrificed by CO2 asphyxiation, rather than cervical dislocation (this prevents separation of the wound margins).


417

Fig. 4. By 6 wk, the graft is well incorporated into the surrounding mouse skin and has almost completely repigmented (A). Histology reveals a fully stratified epithelium with abundant keratin, an interdigitating dermal-epidermal junction, and a revascularized dermal matrix that is well populated with host fibroblasts (B).

2. The dressing is carefUy removed so as not to dislodge the underlyjng graft epidermis. At this time, it is our practice to photograph and trace the graft while it is still on the mouse. 3. The entire graft is then harvested by making a full thickness incision down to, and through, dorsal thoracic musculature of the mouse, with sharp scissors. The harvest should include the muscle underlying the graft, but not the mouse ribs. Make sure to include a segment of normal mouse skin around the edges of the graft. 4. Depending on the experiment, the graft can be handled in several different ways. We like to divide the graft into two pieces, and save one-half for frozen sections and the other half for permanent histologic sections. To do this, spread the dissected graft on a piece of dental wax (Byte ryte, Mizzy) and pin the four corners to the wax. A razor blade can then be used to evenly divide the graft into two or more pieces. 5. One piece (still pinned to the wax) is then immediately placed in a solution of 10% formalin. The other piece is spread flat on a piece of aluminum foil, and then snap-frozen on a flat block of dry ice, labeled, and transferred to liquid nitrogen or a deep freezer. The formalin fixed pieces can be paraffin embedded, sectioned, and stained with hematoxylin and eosin to reveal the structure of the epidermis and dermis. The frozen pieces can be processed for cryosections and immunostained to reveal specific structures or proteins of the skin.

4. Notes 1. An alternative source of skin, which is readily available, is skin harvested in reduction mammoplasty, abdominoplasty, or panniculectomy procedures. Each of these procedures yields skin attached to underlying subcutaneous tissue. We do not recommend using full-thickness skin for the composite grafts; thus, this skin will have to be harvested from the subcutaneous tissue using a dermatome (Padget or Zimmer are

418

Medalie and Morgan recommended), set at an average thickness of 0 0 15 in It IS much easier to harvest the skm whtle it IS still attached If the purpose of the harvested skm 1sonly to generate an acellular dermal matrix, then normal cryopreservatton methods, designed to protect the cells by controlled rate freezing may not not necessary However, our laboratory used only skm from a skm bank that had been cryopreserved according to strtct chmcal use protocols, and thus we never experimented with the use of uncryopreserved skm Freshly harvested skin can either be frozen according to the protocol outlined above or immediately subJected to the freezethaw cycles, and then placed into the anttbtottc containing PBS As with all human skin, tt should be remembered that disease transmission 1sstill a posstbrhty, and full precautions should be taken It cannot be overemphasized how important sterile technique 1s at all stages of the preparatton of the acellular dermis Even the slightest contamination can blossom mto fulmmant bacterial overgrowth once the dermis 1s placed mto an incubator When human skm 1s prepared as described above, the dermis 1s rendered completely acellular, but tt retains key architectural elements, such as papillary proIectrons, elastm fibers, and many of the basement membrane proteins (22,24) In the routme harvest and culture of keratmocytes from neonatal foreskins, passenger melanocytes have been observed to survive at about 10% of then normal skm number (2) These melanocytes will adhere, along with the keratmocytes, to the supportmg dern-ns, proliferate, and resume function (29,22,32) It should be noted that we were unable to find increased m vtvo survival from raising the grafts to the an-liquid interface Overall graft take for both graft types 1sexcellent, and thus the added time delay of growmg tt at the Interface 1snot necessary, unless the expernnent depends upon varying thickness and maturity of the graft Many other researchers have descrtbed preserving the panmculus camosus to ard m the survival of the transplanted skm graft. We do not find this to be necessary and have found that tt serves to increase the length and difficulty of the procedure We do not use systemic antrbtottcs for the mice durmg or after the procedure Good sterile technique results m few-to-no mfecttons The cells used m our experiments were derived almost exclustvely from darkly pigmented donor foreskins This was done m order to more easily study the survival and subsequent functton of passenger melanocytes We have observed that melanocytes proliferate and resume function once transplanted to athymlc mice The composite grafts develop foci of dark pigment after several weeks m VIVO, and these foci increase m size and number as the graft matures, generating a fully pigmented skm equivalent (19,21,31)

References 1 Green, H , Kehmde, O., and Thomas, J (1979) Growth of cultured human eptderma1 cells into multiple eprtheha suttable for grafting Proc Nat1 Acad Scz USA X,5665-5668 2 Galhco, G G , O’Connor, N. E , Compton, C C , Kehmde, 0, and Green, H (1984) Permanent coverage of large burn wounds with autologous cultured human epithehum N Engl J Med. 311,448451.


419

3 Hefton, J M , M R Madden, J L Finkelstem, and Shires, Cl. T (1983) Grafting of bum patients with allografts of cultured eptdermal cells. Lancer 2,428-430 4 De Luca, M ,Albanese, E , Bondanza, S., Megna, M , Ugozzoh, L , Molma, F, et al (1989) Multtcentre experience m the treatment of bums with autologous and allogemc cultured eptthelmm, fresh or preserved m a frozen state Burns. 15, 303-309 5 Madden, M R , Fmkelstem, J. L , Starano-Cotco, L , Goodwm, C W, Shires, G T , Nolan, E E., and Hefton, J M (1986) Grafting of cultured allogenetc epidermis on second- and third-degree bum wounds on 26 patients. J Trauma 26,955-962 6 Compton, C C , Htckerson, W, Nadne, K., and Press, W (1993) Acceleration of skin regeneration from cultured eptthelial autografts by transplantatton to homograft dermis J Invest Dermatol. 14, 653462 7 Cooper, M L , Andree, C , Hansbrough, J F , Zapata, S R , and Spielvogel, R L (1993) Direct compartson of a cultured composite skm substitute contammg human keratmocytes and libroblasts to an eptdermal sheet graft contammg human keratmocytes on athymlc mice. J. Invest Dermatol. 101, 81 l-819 7a Cuono, C , Langdon, R , and McGuue, J. (1986) Use of cultured eptdermal autografts and dermal allografis as skm replacement after bum mJury Lancet 1, 1123,1124 8 Kangesu, T, Navsaria, H A, Manek, S , Fryer, P R , Leigh, I M , and Green, C J (1993) Kerato-dermal grafts. the importance of dermis for the m vlvo growth of cultured keratmocytes Br J Plast Surg. 46,401409. 9 Yannas, I V and Burke, J F (1980) Design of an artlfictal skin I. Basic design prmctples J Blamed Mater Res 14,65-81. 10 Yannas, I V, Burke, J F , Gordon, P L , Huang, C , and Rubenstem, R. H. (1980) Design of an arttfictal skin II Control of chemical composttion J Bzomed Mater Res 14,65-81 11 Boyce, S T and Hansbrough, J F (1988) Biologic attachment, growth, and dtfferenttatton of cultured human eptdermal keratmocytes on a graftable collagen and chondrottm-6-sulfate substrate Surgery 103,42 l-43 1 12 Boyce, S T, Chrtstianson D J , and Hansbrough, J F (1988) Structure of a collagen-GAG dermal skin substitute opttmtzed for cultured human epldermal keratmocytes J Blamed Mater Res 22,939-957. 13. Hansbrough, J F, Boyce, S. T, Cooper, M L., and Foreman, T. J. (1989) Burn wound closure with cultured autologous keratinocytes and fibroblasts attached to a collagen-glycosammoglycan substrate JAMA 262,2 125-2 130 14 Tmots, E , Ttolher, J , Gaucherand, M , Dumas, H , Tardy, M , and Thtvolet, J (199 1) In vitro and posttransplantatton dtfferenttatton of human keratmocytes grown on the human type IV collagen film of a bdayered dermal substitute Exp Cell Res. 193, 31&319 15 Bell, E., Ehrlch, H , Buttle, D. J., and NakatsuJi, T. (198 1) A Living tissue formed m vitro and accepted as skin-equivalent tissue of full thickness Sczence 211, 1052-I 054 16 Parenteau, N L , Nolte, C M , Btlbo, P, Rosenberg, M , Wtlkms, L M , Johnson, E W , et al (1991) Eptdernns generated m vitro* practical constderattons and applications J Cell Blochem 45,245-25 1

420

Medalie

and Morgan

17 Hansbrough, J , Morgan, J , Greenleaf, G , Underwood, J (1994) Development of a temporary hvmg skm replacement composed of human neonatal tibroblasts cultured m Brobrane, a synthetic dressing material Surgery 115, 633436 18 Hansbrough, J F , Morgan, J , Greenleaf, G , Parikh, M , Nolte, C , and Wrlkms, L (1994) Evaluatton of Graftskm composite grafts on full-thickness wounds on athymlc mice J Burn Care Rehab 15,346-353 19 Medahe, D , Emmg, S E , Tompkms, R G ,Yarmush, M L , Kreuger, G G , and Morgan, J R (1996) Evaluation of human skm reconstituted from composite grafts of cultured keratinocytes and human acellular dermis transplanted to athymic mice J Invest Dermatol 106, 121-127 20 Medahe, D A , Emmg, S E , Collms, M. E , Tompkins, R. G , Yarmush, M. L , and Morgan, J R (1997) Dtfferences m dermal analogs mfluence subsequent pigmentation, epidermal differentiation, basement membrane and rete ridge formatron of transplanted composite skm grafts. Transplantation, 64, 454-465 21 Medahe, D A , Tompkins, R G , and Morgan, J R (1996) Evaluation of acellular human dermis as a dermal analog m a composite skm graft ASAZO J 42, M455-M462 22 KreJcI, N C , Cuono, C. B , Langdon, R C , and McGmre, J. (1991) In vitro reconstitutton of skm fibroblasts facilitate keratmocyte growth and differentiation on acellular reticular dermis J Invest Dermatol 97, 843-848 22a Langdon, R C , Cuono, C B , Birchall, N., Madrr, J A , Kuklmska, E , McGmre, J , and Moellman, G. E. (1988) Reconstitution of structure and cell function m human skin grafts dertved from cryopreserved allogenerc dermis and autologous cultured keratinocytes J Invest Dermatol 91,478&485 23 Choate, K A , Medahe, D A , Morgan, J R , and Khavarr, P A (1996) Corrective gene transfer m the human skin disorder lamellar ichthyosrs Nature Med 2, 1263-l 267 24 Prumeras, M , Regmer, M , and Woodley, D (1983) Methods for cultivation of keratmocytes with an an-lrqmd mterface (Review) J Invest Dermatol 81, 285-289 25 Kraut, J D , Eckhardt, A J , Patton, M. L., Antomades, K , Haith, L. R. J., and Shotwell, B S. (1995) Combined simultanous apphcatlon of cultured epithelial autografts and alloderm@ WOUNDS 7, 137-142 26 Lrvesey, S A , Herndon, D. N , Hollyoak, M A , Atkinson, Y H , and Nag, A (1995) Transplanted acellular allograft dermal matrix. Transplantatzon 60, l-9 26a McKay, I , Woodward, B , Wood, K , Navsarra, H A , Hoekstra, H , and Green, C (1994) Reconstruction of human skm from glycerol-preserved allodermis and cultured keratmocyte sheets. Burns 20, S 19-S22 27 Matouskova, E , Vogtova, D., and Komgova, R (1993) A recombined skm composed of human keratinocytes cultured on cell-free pig dermis. Burns 19, 118-123 28. Guo, M. and Grinnell, F. (1989) Basement membrane and human epidermal dlfferenttation m vitro. J Invest. Dermatol. 93,372-378 29 Heenen, M., Graef, C D , Parent, D , Dobbeleer, G D , and Galand, P (1992) Renewal and dtfferenttation of keratmocytes cultured on dead de-epidermahzed dermis Cell Prollferatzon 25, 3 1l-3 19

Epidermal Keratinocyte Composite Graft 30

421

Shakespeare, V A and Shakespeare, P G (1987) Growth of cultured human keratmocytes on fibrous dermal collagen. a scanning electron mlcroscope study. Burns Includzng Thermal Inpry 13,343-348 3 1 Hlgounenc, I , Demarchez, M., Regmer, M , Schmidt, R , Ponec, M , and Shroot, B. (1994) Improvement of epldermal differentiation and barrier function m reconstructed human skm after graftmg onto athymlc mice Arch Dermatol Res 286,107-l 14 32 Stalano-Colco, L , Hefton, J M , Amadeo, C , Pagan, Charry, I, Madden, M R , and Cardon, C (1990) Growth of melanocytes m human epldermal cell cultures J Trauma 30, 1037-l 042

31 Development

of a Bioartificial

Liver Device

Linda K. Hansen, Julie R. Friend, Rory Remmel, Frank B. Cerra, and Wei-Shou Hu 1. Introduction Liver dtsease contmues to be a challenge clmrcally, wrth 30,000 patrents dying each year from liver failure (I). Although liver transplantatron can successfully treat many patrents undergoing liver farlure, the scarcity of donor organs severely ltmtts thts treatment’s apphcatron. For thus reason, many mvesttgators are pursuing alternatives to total organ transplantatton, from livmg donors to cell transplantatron. One additional approach 1sthe development of a hybrid, broartrtictal ltver as an extracorporeal devrce for the temporary treatment of acute lrver farlure This approach has demonstrated early success, and may provide an important clmrcal treatment m the near future. In addrtton, a broartrfictal liver reactor 1suseful for prolonged m vitro studres of hepatocyte function. This chapter ~111provide mformatton on the design and use of such a reactor for m vitro applicatrons. The functrons of the liver are numerous and varied, mvolvmg many complex brochemrcal reactions wrthm hepatocytes that cannot be easily mrmtcked m an acellular environment. The development of a device to support and/or mtmrc normal liver function must therefore include ltvmg hepatocytes m an environment that maintains viability and promotes differentiated function. In addrtton, mmrunoprotection must be maintained for m vtvo applications, and mass transfer hmrtatrons must also be addressed. Several models have been designed to accomplrsh this task, which utilize rmmobrhzed hvmg hepatocytes m a hollow-fiber reactor Many reviews have been published descrtbmg dtfferent devices developed m several laboratories (2-5); this chapter ~111focus on the method of producmg the Mmnesota Broartrficral Liver (BAL) In this BAL, hepatocytes are suspended m a type I collagen solutton and placed msrde holFrom

Methods in Molecular Me&me, Vol 18 Edited by J R Morgan and M L Yarmush

423

Jtssue Engmeenng 0 Humana

Methods and Protocols Press Inc , Totowa, NJ

424

Hansen

et al.

WA-STE

C HE

Fig 1 Schemattc Dragram of Minnesota BAL Two perfusion streams are estabhshed wlthm the hollow-fiber cartridge Medmm 1sperfused through the extracaptllary compartment, or shell space (“shell medmm”) Durmg mmal reactor operation, this medium may be recirculated. Collagenihepatocyte suspension is InJected into the mtracaptllary space On collagen/cell contractron, a second perfusion stream IS estabhshed (hollow-fiber lumen) Medium supplymg hepatocytes with nutrrents and growth factors (“lumen medium”) IS perfused through this lumenal space, allowing nutrient flow directly adjacent to nnmobrhzed cells End caps keep the flow path separate Shell medium 1sreplaced with test media for detoxrficatron during m vitro testing, or blood for m vlvo apphcattons (without recrrculatton) Inset demonstrates contracted collagen, entrapped hepatocytes, and the mtercaprllary lumen wtthm a single hollow fiber

low fibers Incubation at 37°C promotes contractron of the collagen-cell suspensron, creating a lumen within the hollow fiber. Media IS perfused through the inner lumen, while media, or patient’s blood for m vtvo applmattons, can be perfused through the space outside the hollow fibers (Fig. 1) Thus desrgn differs from others that place the hepatocytes outside the hollow fibers. Because of the hybrid nature of the BAL, mvolvmg both ltvmg cells and a mechamcal device, there are several steps mvolved n-r its assembly. The Method subheading will thus be divided into four subsectrons, each descrrbmg a separate step m the process (reactor setup, cell rsolatlon, reactor loading, and reac-

Bloartificlal Liver Device

425

tor operation). Thus information should provide sufficient mstruction for producing an m vitro bioreactor for the study of living hepatocytes. 2. Materials 2.7. Reactor Hollow fiber cartridge was obtained from Amicon, Danvers, MA, Model HlPlOO. 2.2. Media 1 Shell medium.Williams Medium E (Gibco-BRL, Grand Island,NY), 2 2 mg/mL NaHC03, 50 pg/mL gentimycm, 0 2 U/mL msulrn Store at 4°C 2 Lumen medium Willtams’ Medium E (Gtbco-BRL), 2.2 mg/mL NaHCOs, 50 pg/mL genttmycm, 0 2 U/mL msulin, 500 pg/mL lmoletc acid, 4 ng/mL glucagon, 20 ng/mL liver growth factor, 6 25 pg/mL transfernn, 5 ng/mL eptdennal growth factor (EGF), 1 @Z dexamethasone, 6.25 ng/mL selenium Store at 4°C 3 4X Williams E medium for collagen solution. (For 250 mL) one 1-L package of powdered Williams’ E medium (Gibco-BRL), 2 2 g NaHCOs, 50 mg gentamycm, 2 mg/mL lmoleic acid, 16 ng/mL glucagon, 80 ng/mL liver growth factor (Sigma, St Louts, MD), 25 pg/mL transferrm, 20 ng/mL EGF, 4 Mdexamethasone, 25 ng/mL selenmm. Store at 4°C

2.3. Cell Harvest and Preparation 1 Perfuston buffer I (Per) 10X stock solutton: 1 43 MNaCl, 67 mMKC1, 100 mA4 HEPES, pH 7 5; store at 4°C 2 Perfusion buffer II (PerII) 1X stock solution. 1% bovine serum albumin, 67 mMNaC1, 6 7 mMKCI,4 8 mMCaC12, 10 mA4HEPES, adJust pH to 7.6, stole at 4°C 3 Collagen solution Type I collagen solution (Vttrogen, Collagen, Santa Clara, CA) mixed 3’ 1 with 4X Wtlhams’ E medium.

3. Methods 3.1. Reactor and Media Preparation The design of the hollow fiber reactor is shown in Fig. 1. The reactor consists of polysulfone hollow fibers sealed wnhm a cartridge (Amicon) End caps are also available through Amicon, wtth ports allowmg perfusion of media through both the inner hollow fiber lumen and the shell space outside the fibers. In addition, a loading cap was manufactured (University of Minnesota machine shop), which, like the end cap, slips over the end of the btoreactor cartridge, but has only one m-port mto the fibers connected on the outside to tubing with a stopcock. The reactor is hooked up to reservotrs of media with standard tubmg and pumps Additional monitors can be installed m line, if desired For example, shell medium pH can be mamtamed using a pH electrode and microprocessor-

426

Hansen et al

controlled CO2 switch (6). In addttton, the cartridges can be maintained tn an automated cell culture device, such as the Maxlmlzer, Jr. (Cellex Blosclences, Coon Rapids, MN). However, a cruder, more economical setup that maintains medium ctrculatton, pH, and 37°C temperature IS sattsfactory

3.2. Rat Hepatocyfe

Harvest

Several procedures have been pubhshed for the tsolatton of rat hepatocytes, most commonly denved from that of Seglen (7), using a two-step m sztu perfUsion of the hver to loosen cells from the surroundmg connecttve ttssue. Sm-nlar procedures have also been published for hepatocyte harvest fi-om other species (8’, which will not be discussed here 4-6-wk-old Sprague-Dawley rats, weighing 200-250 g, are used

3.2 1 Preoperatrve

Procedure

1 Prepare the first perfusate buffer, Per I* a Dilute 10X Per I stock up to 300 mL with dH*O, to make 1X b Add 0 285 g EGTA c AdJust pH to 7 4 with 1 MNaOH b Sterile-filter, and place m 37°C water bath 2 Prepare the second perfusate buffer, Per II* a Add 0 1 g collagenase D to 200 mL Per II 1X stock d Sterile-filter, and place m 37°C water bath 3 Aseptically prepare operatmg surface 4 Run each perfusate through tubmg of a multichannel pump to bleed lines free of au 5 Anesthetize rat with ip qection of Nembutal stock (50 mg/mL) with 0 1 mL/ 100 g body weight, then shave and sterlhze rat belly

3.2 2. Operative Procedure 1 Make a transverse mclsion through the skm approx 1 cm below the xyphoid process, and make a similar mclsion through the abdommal wall 2 Eviscerate the rat onto sterile gauze, exposmg the mferior vena cava (IVC) and portal vein 3. InJect 0.6 mL heparm mto the IVC, using a 1-mL syrmge, and allow heparm to circulate for 1 min. 4 Place two sutures loosely around the portal vem 5 Insert an 1&gage iv catheter mto the portal vein, and secure with suture ligatures 6 Carefully remove the IV catheter needle, and allow the portal blood to back up to the IV catheter hub 7 Connect the perfusion line to the catheter hub, turn on pump to start the Per I perfusion (25 mL/mm), and cut the IVC 8 Contmue Per I perfusion until gone, or until the effluent from the IVC is clear 9 Switch to Per II perfusion and reduce pump speed to 20 mL/mm (see Note 1) 10 When Per II is gone, cut the liver free and place mto a 60-mm Petri dish with lo-15 mL cold Williams’ E media, cover, and take to sterile hood

Bloartifmal Liver Device

427

3.2.3. Cell /so/a tion 1 Usmg forceps and scissors, mclse the liver capsule on all lobes, and gently agitate liver m the media to free the cells 2 Gently plpet the cell suspension onto nylon gauze, and funnel mto 50-mL centnfuge tube on ice, to remove clumps, Rinse gauze with addItIona cold Wllllams’ E medmm, using a second tube, if necessary Cells should remam on ice throughout the remamder of the procedure (see Note 2). 3 Spm tubes at 50 g for 2 mm at 4°C 4 Resuspend pellet m cold Wdhams’ E medmm, and wash two more times, combmrng pellets from all tubes during last resuspension (see Note 3) 5 Dilute a small volume of the cell suspension 1,100 with trypan blue dye Determme cell concentration and vlablllty, using a hemocytometer (see Note 4)

3.3. Loading Reactor 3 3.1. Reactor Rinsing The Amlcon hollow-fiber reactor is stored in phosphate buffered saline (PBS), and thus requires no rmsmg Other hollow-fiber reactor models, however, may require rmsmg m order to remove any potential residue from the hollow fibers that can harm the cells. This is done as follows: 1 In a sterile hood, vertically place the reactors, with end caps m place, m a rmgstand with the m-ports at the bottom 2 Hook up the out port from shell flow to the m-port of lumen flow, to create a single loop 3 Turn on the pump at 1 L/mm while checkmg for leaks 4 When all PBS (e.g ,4 L) has been pumped through the system, contmue to blow air through system for about 1 mm, to remove PBS

3.3 2. Preparabon of Cell Suspension 1 MIX together a 3.1 solution of type I collagen (Vltrogen) and 4X WIlllams E medmm The Amlcon Model H 1PI 00 hollow-fiber bloreactor can hold about 10 mL 2 Use 1 N NaOH to bring up the pH gradually Add a few drops of NaOH to the collagen-4X solution, and mix Keep adding NaOH dropwlse until color of 4X Wllllams’ E mixture turns red, mdlcatmg pH -7 2-7 4 3 Pellet cells by spmmng at 50g for 1 mm at 4”C, to form a soft pellet Remove supernatant and resuspend cells m the collagen-Williams E solution, to yield a cell concentration of about 3-5 x lo6 cells/ml. Since one Amlcon reactor holds about 10 mL, approx 50 x IO6 cells should be pelleted and brought up m 10 mL collagen-Willlams E solution for each reactor

3.3 3. Reactor Loading 1. Attach a loading cap onto one end of the reactor, leavmg the other end open Reactor should be m a vertical posltion m a rmgstand, with the loading cap posltloned at the bottom end.

428

Hansen et al.

2 Load the collagen-cell mixture mto a 35-cc syrmge, using a 14-gage needle 3 Remove the needle and attach the syrmge to the lumen m-port of the reactor 4 Slowly inject collagen-cell mixture until tt fills the fibers from bottom to top Close the stopcock of the loading cap 5 Remove the bioreactor from the rmgstand, remove the loading cap, and put endcaps m place

3.4. Reactor Operation 1 Incubate reactor at 37°C 2 About 10 mm followmg incubatton, put cartridge m line with medium, and begin perfusion of recirculated shell medium at about 30 mL/mm 3 Change shell perfusate after first 24 h, and every other day thereafter 4 24 h after shell perfusion began, during which time collagen gel should have contracted, begin perfusion of lumen medium at about 9 mL/hr 5 Cell function and vtability can be assessed over time by numerous methods (Refer to Note 5 and mdtcated references for further details )

4. Notes 1. Watch the liver whtle the buffers perfuse through tt One should see an tmmedtate blanching of the liver as the blood 1s replaced with perfuston buffer If this does not occur, or only occurs m part of the liver, there IS probably a bubble or other obstructton m the system This will lead to lower yteld and viabihty 2 Hepatocytes will lose viabihty very quickly if they remain in suspension for more than a few hours, parttcularly if the solution approaches room temperature Therefore, care must be taken to work quickly and keep the cell suspension on ice 3 Collagenase perfusion yields a heterogeneous cell population Although hepatocytes comprise the largest population of cell m the liver, there will also be other cells types, including hpocytes, Kuppfer cells, and others The low-speed spins (5Og, 2 mm) preferentially separate the hepatocytes from other cell types, as well as dead cells because of thetr greater density, but an additional purification step, such as Percoll gradient centrifugation (9), may be added if a homogeneous hepatocyte populatton is desired 4 Most rat harvests should yield around 5 x lo8 - 1 x lo9 cells Ideally, viabihty should be 90% or greater It IS not recommended to use cells from a harvest that yielded an mtttal viability below 80% 5 Hepatocytes are responsible for many metabolic, synthetic, and biotransformation functions m vtvo Quantttattve analysts of these activmes m the bioreactor will mdtcate the level of differentiated functton mamtamed by the entrapped hepatocytes Clearly, it is impossible to assay each of the many functtons of the liver, but there are several quantitative assays that have been successfully employed to measure a variety of hepatocyte functions One of the simplest assays to employ 1s the measurement of albumin secretion Media drawn from the lumen and/or shell space out-port can be analyzed for albumin protein concentration using enzyme-lurked immunosorbent assay

Bioartificial Liver Device (ELISA) (10) Although this is a common assay, it 1s a measurement only of the synthesis of one protem that may be regulated differently than other functions that may be of more interest, depending on the application Furthermore, albumin may also be released mto the medium from dead, lysed cells, leading to an overestimation of de now albumm synthesis and secretion Substrate metabolism by specific hepatlc enzymes provides another useful parameter of hepatocyte function Cytochrome P450 enzymes are mvolved m the phase I metabohc response, and this can be assessed by measuring lldocame metabolism (11) In addltlon, glucuronidatlon and sulfation (phase II) can be assessed by 4-methylumbelhferone conjugation by UDP-glucuronosyltransferase and sulfotransferase Addition of either of these compounds to the shell medium prior to input, and analysis of the compound and Its metabohtes m the shell medium output, gwes a good measure of the level of biotransformation activity (10,12). Because of the dlverslty of hepatocyte functions and the potential for these functions to be regulated differently, the performance of a bloartlficial liver device 1s best assessed using a variety of tests, as described here, to ascertain a spectrum of functions

4.7. General Design Considerations Several different hepatocyte bloreactor designs have been developed (13). The reactor described m this chapter was designed to achieve high cultlvatlon density and reduced mass transfer ltmitatlons, which are crucial m scaling up the reactor for clinical applications. Many reactor designs place the hepatocytes outside the hollow fibers (shell space), and perfuse medium only through the hollow fiber lumen. However, the mterfiber distances and geometry are variable, resulting in mass transfer hmttations m those areas farthest from the oxygen and nutrient source. Furthermore, in designs utilizing hepatocyte attachment to mlcrocarrlers, the space taken up by the mlcrocarrlers 1s essentially empty space. Placement of cells within the hollow fibers, and perfusing medium through the lumen created by collagen contraction, allows a much closer apposltlon between cells and medium, and can allow for a more efficient compaction of cells. In addltlon, the design described here creates space for a second perfusion stream, which, in clinical appllcatlons, can be utilized for patient blood, while still maintaining a nutrient stream m direct contact with the cells to which growth factors and other soluble mediators may be added. Immobilization of hepatocytes within the collagen gel allows for self-association of hepatocytes. It 1s clear that cell-cell mteractron and formation of three-dimensional structure that mimics that seen m vlvo can enhance cellular function. Immoblllzation of hepatocytes onto a solid surface, such as a mlcrocarrier, may inhibit these cell-cell interactions and formation of m vivolike structures; suspension and contraction in a collagen gel can promote such interactions (see Subheading 4.2.). Additional msoluble matrix molecules may

430

Hansen et al,

also be mcorporated mto the gel, which may enhance cell function and/or vtablhty (14). 4.2. Spheroids Primary hepatocytes exhibit loss of dtfferentrated function when removed from then native envtronment and placed mto tradmonal tissue-culture condttrans. Many studies have demonstrated that certain extracellular matrtx substrates, and other condrttons that mtmrc the native environment, can promote restoration of differentiated function. Under certain supportive condltlons, hepatocytes will self-aggregate mto multtcellular structures, or spheroids, that express htghly dtfferenttated function. Cells wtthm these spheroids possesscell polartty and cell-cell mteracttons very similar to those seen m IWO. Use of spheroids wtthm the broreactor may provide enhanced reactor function (15,16) This topic IS pursued further in Chapter 19 References 1 National Center for Health Statistics (1993) Vital Statzstzcs of the Unzted States 2. Yarmush, M L , Dunn, J C , and Tompkins, R. G. (1992) Assessment of artificial hver support technology. Cell Transplantatzon 1,323-341 3 Kasai, S K , Sawa, M , and Mtto, M. (1994) Is the biological artificial liver clmitally apphcabale? A historic review of biologtcal arttficial liver support systems. Artlf Organs 18,348-354 4 Dtxit, V. (1994) Development of a btoarttfictal liver using isolated hepatocytes. Artzf Organs 18,371-384 5 Jauregut, H 0 , Chowdhury, N R , and Chowdhur, J R (1996) Use of mammalian cells for artificial liver support Cell Transplantatzon 5, 353-367 6 Shatford, R A , Nyberg, S L , Meter, S. J , White, J G , Payne, W. D , Hu, W -S , and Cerra, F B (1992) Hepatocyte function m a hollow fiber bioreactor A potential bioartificial liver J Surg Res 53, 549-557 7 Seglen, P 0. (1976) Preparation of isoiated rat hver cells Methods Cell Bzol 13,2%33 8 Sielaff, T , Hu, M Y, Rao, S , Groehler, K , Olson, D , Mann, H J , et al. (1995) A technique for porcme hepatocyte harvest and description of differentiated metabolic functions m static culture. Transplantatzon 59, 1459-1463 9 Kreamer, B L , Staecker, J L , Sawada, N , Sattler, G L , Hsta, M T S , and Pitot, H C (1986) Use of a low-speed, iso-density Percoll centrtfugatton method to increase the vrabrhty of isolated rat hepatocytepreparations,In Jkro Cell Dev Bzol 22,201-211 10 Nyberg, S L , Shatford, R A., Peshwa, M V, White, J G , Cerra, F B , and Hu, W -S (1993) Evaluation of a hepatocyte entrapment hollow fiber btoreactor a potential bioartifictal liver Bzotech Bzoeng 41, 194-203 11 Schroeder, T J., Gremse, D A., Mansour, M E , Theuerlmg, A W, Brunson, M E , Ryckman, F. C , et al. (1989) Lidocame metabolism as an index of liver function m hepattc transplant donors and recipients Transplantation Proc 21,2299-2301

Bioartiflcial 11ver De vice

431

12 Nyberg, S L , Mann, H J , Remmel, R P, Hu, W S , and Cerra, F B (1993) Pharmacokmetlc analysis verfies P450 function during m vitro and m VIVO apphcatlon of a bloartlficlal liver ASAIU J 39, M252-M256 13 Rozga, J., Wllhams, F, Ro, M -S , Enuzil, D F., Glorglo, T D , Backfisch, G , Mosclom, A D , Haklm, R , and Demetriou, A A (1993) Development of a bloartlficlal liver properties and function of a hollow-fiber module inoculated with hver cells Hepatology 17,258-265 14 Hu, M Y, Slelaff, T. D , and Cerra, F B. (1994) Enhancement of cytochrome P450 tinctlon of collagen-entrapped hepatocytes by the addition of liver extracellular matrix components Tramp1 Proc 26,3293. 15 Wu, F J , Peshwa, M V, Cerra, F B , and Hu, W -S (1995) Entrapment of hepatocyte spherords m a hollow fiber bloreactor as a potential bloartlficlal liver Cell Trmsplantat~on 1, 29-40 16 Wu, F , Friend, J R , Hslao, C. C , Ztlhox, M. J , Ko, W J , Cerra, F B , and Hu, W S (1996) Efficient assembly of rat hepatocyte spheroids for tissue engmeermg apphcatlons Bzotechnol Bzoeng 50,404-4 15

32 Methods for the Implantation Stephen S. Kim, Hirofumi

Utsunomiya,

of Liver Cells and Joseph P. Vacanti

1. Introduction I. 1. His torical Perspectives There have been many major advances in the field of hver transplantatton m the past 30 yr. Orthotoprc liver transplantation IS currently the only established successful treatment for end-stage liver disease, with over 3000 liver transplantattons being performed each year in more than 120 lrver transplantation centers m the United States (I). There are several maJor challenges, however, that impede the wrdespread practice and applicabrhty of organ transplantatton. These Include the crtttcal shortage of donor organs, the high cost and technical dtfficulty of the procedures, and the intensive postoperative care involved, mcludmg those associated wtth life-long mununosuppression. Among these obstacles, the critrcal scarctty of donor organs, especially m the pedratrrc population, 1s perhaps the most sigmticant Each year, end-stage lrver disease accounts for 26,000 deaths in the United States (2) Although the supply of donor organs has increased only slightly over the last 5 yr, the number of patients on the wartmg list and the number of patients who die each year whtle on the waiting list have continued to grow at a dtsproportronate rate (I). These shortcommgs have sttmulated mvestrgatron mto selective cell transplantation and the emergence of the field of tissue engineering as an alternatrve approach to the treatment of end-stage hver drsease (3-5). Tissue engmeermg IS an mterdrscrplmary field that applies the prmcrples of engineering and the hfe sciences toward the development of btological substitutes that restore, maintain, or improve tissue function (4,5). Conceptually, the transplantation of only the essential tissue elements-in the case of the liver, the hepatocytehas many advantages. These include: From

Methods VI Molecular Edited by J R Morgan

MedIcme, Vol 18 T/ssue Engmeenng Methods and Protocols and M L Yarmush 0 Humana Press Inc , Totowa, NJ

433

K/m, Utsunomiya, and Vacant/

434 1 The potential for the allevlatlon

of the donor organ shortage by utlllzmg cells from a small amount of donor tissue, and expandmg them m vitro to create a potentially limltless supply 2 The decrease in the rrsk and expense associated with major surgical procedures and protracted hospltahzatlons 3 The potential of using autologous cells for transplantation, obvratmg the need for immunosuppression 4 The capacity for liver-dlrected gene therapy to treat Inborn errors of metabolism caused by smgle gene defects (6) Since the development of crucial techmques for the high-yield lsolatlon of viable hepatocytes (7,8), there has been a tremendous amount of Interest and sclentlfic mvestlgatlon mto hepatocyte transplantation Hepatocyte transplantation has been performed m nearly every organ system, mcludmg the liver (P-11), portal venous system (12-15), spleen (24,1618), pentoneal cavity (22--14,29), small

bowel mesentety(20-22) omentum (20), lung (23,24), pancreas (25,26), renal capsule (27,28), and subcutaneous tissue (20,21,29,30), m a variety of animal models, mcludmg the mouse, rat, rabbit, pig, dog, and monkey. A number of

studies have examined the use of hepatocyte transplantation m models of acute liver failure (13,14) and metabohc

deficiency states (22,15,32,32).

Hepatocytes

have been injected as cell suspensions or surgically implanted after bemg mlcroencapsulated (32,33), attached to microcarrter beads (3638), or seeded on polymer

matrlces (20,34,35).

Desprte this vast spectrum of mvestlgatlons,

the

opttmal method and site of hepatocyte tmplantatton has yet to be determined 1.2. The Properties of Implantation

of the Optimal

There are many Important

properties

Method

and Site

that should be Incorporated

mto the

development of an optrmal method of rmplantatton of liver cells, based on estabhshed characterlsttcs of the hepatocytes and the theoretrcal functtons that the cells must fulfill after tmplantatton. These mclude the followmg: 1 Hepatocytes are anchorage-dependent,

and requrre an msoluble extracellular matrrx for survival, orgamzatton, prohferatron, and functton The extracellular

matrices not only provide a surface for cell adherence, but they have profound mfluences on modulatmg cell shape and gene expressron related to cell growth and liver-specrfic function (39)

2 A large number of hepatocytes must be delivered and engrafted An esttmated lO-20%

of the liver mass needs to be Implanted to successfully

replace lrver

function (40). 3 Hepatocytes are highly metabohcally active, and require rapid access to oxygen and nutrient supply 4 The hver has a tremendous regenerative capacity m vrvo Liver cells are highly

responsive to hepatotrophtc stlmulatlon

(41-44)

lmplantatlon of Liver Cells 1.3. Overview

435

of Methods for the Implantation

of Liver Cells

Hepatocyte transplantation has been investigated m nearly every organ system m the body. For the purposes of this chapter, the methods that have demonstrated the most promrsmg results in the rat model will be addressed: 1 Direct mJectlon method a Into the portal venous system b Into the spleen 2 Implantation associated with an extracellular matrix a Attached to mlcrocamers. b After mlcroencapsulation c Seeded on three-dlmenslonal, biodegradable polymer scaffolds

1.4. Advantages and Disadvantages of Imp/an ta tion

of Different

Methods

1 4 1. Inject/on of Hepatocyte Suspension into Portal Venous System The major advantages of hepatocyte Implantation tem, and, subsequently, mto the host liver, Include: 1 The utlllzatlon architecture of 2 The Interaction trophlc factors 3 The lmmedlate waste removal

mto the portal venous sys-

of the intact host hepatlc extracellular matrix and hlerarchlcal the host liver for cellular reorgamzatlon with other nonparenchymal hver cells and exposure to hepatoreleased locally, and present m the portal system access to the blood supply for oxygen and m&lent delivery and

The mam disadvantages

assoctated

with direct qectton

of hepatocyte

sus-

pension mto the portal venous system include: 1 The comphcatlons related to portal hypertension, portal vem thrombosis, systemic embohzatlon, and host hepatlc embohzatlon and infarction 2 The great difficulty of locatmg and dlfferentlatmg the transplanted cells from the host cells after implantation 3 The llmlted mass of cells that can be transplanted

7.42 lnlection of Hepatocyte Suspension Into the Spleen The major include.

advantages

of direct injection

of hepatocytes

mto the spleen

1 The immediate access to the blood supply for oxygen and nutrient delivery and waste removal 2. The slmpllctty of the procedure, the ease of ldentlfymg transplanted hepatocytes m the spleen, and the relatively low risk of comphcatlons associated with splemc injection and temporary occlusion of the splemc vessels during mJectlon (45)

436

Kim, Utsunomiya, and Vacanti

3 The translocatron of a large fractron of hepatocytes Into the lrver after splemc mJectron, if performed

without

concomttant

occluston

of the splemc vessels

(10,46,47)

The mam disadvantages the spleen include:

associated with direct mjectron of hepatocytes

mto

1 The limited massof cells that can be implanted mto the spleen 2 The translocatton of hepatocytes mto the portal system and the host ltver after

splenic mjectron, with the associated risks of portal hypertension,

systemrc

embohzatton, and hepattc emboltzatton. 3 The question of apphcabthty m larger animal and human models

1.4.3. implantation of Hepatocytes Attached to Collagen-Coated Mxrocamers into the Peritoneum The major advantages of mrcrocarrrer-attached the peritoneum mclude.

hepatocyte

implantation

m

1 The easy anatomic accessand the mmimally mvastve procedure requtred 2 The provtslon of an extracellular matrix to enhancecell attachment and functton 3 The large surface area available for the potenttal delivery of a large massof hepatocytes

The mam disadvantages include:

of microcarrier-attached

hepatocyte

rmplantatton

1 The absenceof a hierarchical ttssue orgamzatton and structure of transplanted hepatocytes. 2 The lack of mrmedtate accessto a vascular supply 3 The lack of evidence for long-term cell survtval and function m vtvo

I 4 4. Implantation of Microencapsulated Into the Peritoneum

Hepatocytes

The major advantages of mrcroencapsulated peritoneum include: 1 The easy anatomic access, the mrnimally

hepatocyte rmplantatton

m the

rnvasrve procedure required, and the

large surfaceareaavailable for the potential dehvery of a largemassof hepatocytes 2 The provtston of an extracellular matrtx to enhancecell attachment and functton 3 The potential for tmmunorsolatton of the mtcroencapsulated hepatocytes, precluding the need for tmmunosuppresston

The main disadvantages of mrcroencapsulated

hepatocyte implantation

include

1 The absenceof a hterarchtcal tissue orgamzatton and structure of transplanted hepatocytes

Implantation of Liver Ceils

437

2 The lack of munedtate access to a vascular supply 3 The lack of evtdence for long-term cell survival and function m vtvo

1.45 Implantation of Hepatocytes Seeded on Biodegradable Scaffolds

Polymer

The major advantages of hepatocyte transplantation on biodegradable polymer scaffolds include. 1 The three dtmenstonal scaffold structure and the extracellular matrtx components that can be Incorporated mto the polymer may enhance cell attachment, function, and tissue reorgamzatton 2 The use of a biodegradable polymer scaffold obviates concerns regardmg foreign materials and long-term biocompattbility 3 The potential to customize the scaffold size, design, and mtrmstc properties for opttmtzatton of cell attachment, survival, and long-term function

The mam disadvantages of hepatocyte transpiantatlon on biodegradable polymer scaffolds include: 1. The lack of immedtate access to a vascular supply. 2 The lack of evidence for long-term cell survival and function m VIVO

2. Materials 2.1. Materials Universally Used in the Various Methods of Hepatocyte ln7plantation 1 Surgical instruments scapel, hemostats, tissue forceps, metzenbaum scissors, sterile cotton applicators (Ethtcon, Somerville, NJ) 2 Methoxyflurane (Ptttman-Moore, Mundelein, IL) 3 70% isopropyl alcohol prep pads (Baxter Health Care, Deerfield, IL) 4 Cluudme solution, povtdone todme (Climpad, Gutlford, CT) 5 Wtlhams’ E media (Stgma, St Louts, MO) supplemented wtth 1 g sodmm pyruvate (Sigma) and 5 mL pemctllm G sodium (10,000 U/mL)/streptomycm sulfate (10,000 pg/mL) and L-glutamme (2.92 mg/mL) (Gtbco-BRL, Grand Island, NY) per 500 mL of media 6 3-O Prolene, polypropylene monofilament suture with a cutting needle (Ethtcon) 7 Clippers to shave rat fur

2.2. Direct Injection

into the Portal Venous System

1 26-gage needle (Becton Dtckmson, Rutherford, NJ). 2. l-cc sterile syringe (Becton Dtckmson).

2.3. Direct Injection into the Spleen 1 26-gageneedle (Becton Dickinson). 2 l-cc sterile syrmge (Becton Dtckmson)

Kim, Utsunomlya, and Vacant/

438

3 Small vascular clamp (Btomedical Research Instruments, 4 4-O silk suture (Ethtcon)

2.4. Imp/an ta tion of Hepatocytes Microcarriers

Attached

Rockvtlle,

MD)

to Collagen-Coated

1 Cytodex 3 type 1 collagen-coated dextran mtcrocarrters (Pharmacta, Plscataway, NJ) 2 Phosphate buffered salme (PBS) without calcium and magnesium (Srgma) 3 Dulbecco’s modified Eagle’s medium (DMEM; Glbco-BRL) with 10% fetal calf serum (FCS, Gtbco-BRL) 4 lo-mL sterile syringe (Becton Dtckmson) 5 13-gage needle (Becton Dtckmson)

2.5. implantation

of Microencapsulated

Hepatocytes

1 Mtcroencapsulatton material. a mixture of 2% (vtscostty = 266 cps) sodmm alganate (Kelco Gel LV, Kelco, San Diego, CA) and 1 7 mmol/L Matrtgel (Collaborative Research Inc , Bedford, MA) (see Note 1) 2 Poly-L-lysine (Sigma) 3 13-gage needle (Becton Dtckmson) 4 20-mL sterile syringe (Becton Dtckmson)

2.6. Imp/an ta tion of Hepa tocytes on Biodegradable Scaffolds 1 2 3 4 5 6 7 8

Polymer

Poly-L-lactic acid (PLLA) (Boehringer Ingelhelm, Germany) Chloroform (EM Science, Gtbbstown, NJ) Sodium chlortde particles sieved to a size of 25&500 pm Teflon cylmders (Cole Palmer, Chicago, IL) 1% aqueous poly-vinyl alcohol (PVA) (Aldrich, Milwaukee, WI) solution Ethylene oxide (H.W Anderson, Chapel Hill, NC) warm gas sterlhzatlon 5-O Prolene, polypropylene monofilament suture (Ethicon) 8-O Monofilament nylon suture (Ethtcon)

3. Methods

3.1. Preparation

of Isolated Rat Liver Cells

The most commonly used methods for the lsolatlon and preparation of rat liver cell suspensions are based on an enzymatic digestion procedure prevlously described by Berry and Friend (7) and Seglen (8) The concentration of hepatocyte suspensions can be vaned, depending on the need 3.2. injection Sys tern

of the Hepatocyte

Suspension

into Portal Venous

1 The hepatocyte pellet obtained after the isolation procedure IS resuspended Wtllrams’ E medta to make a 2 x lo7 cells/ml suspension

m

Implantation

of Liver Cells

439

2. The recipient rat IS anesthetized usmg methoxyflurane mhalatronal anesthesia, and 1s weighed 3 After the abdomen IS shaved, it IS sterilely prepped with 70% isopropyl alcohol and clmtdme solution. 4 A mtdlme mctston 1smade, and the small bowel IS evtscerated to the animal’s left side, exposing the portal vem 5 Using a l-cc syringe, 1 mL of the hepatocyte suspension IS inJected mto the portal vem usmg a 26-gage needle slowly over l-2 mm (see Note 2) 6 After the mjectton IS completed, the needle IS withdrawn and hemostasrs obtained by applying pressure with a sterile cotton applicator for several mmutes (see Note 3) 7 The abdommal contents are returned to their anatomic postttons and the abdomen 1s closed m two layers, using 3-O prolene suture 8 The rat IS monitored postoperattvely for any signs of complications

3.3. Injection

of Hepatocyte

Suspension

into the Spleen

1 The hepatocyte pellet obtained after the isolation procedure is resuspended m Wllhams’ E media, to make a 2 x 10’ cells/ml suspension 2 The recipient rat IS anesthetized using methoxyflurane mhalatronal anesthesia, and IS werghed 3 After the abdomen IS shaved, It 1s sterilely prepped with 70% Isopropyl alcohol and clmrdme solutron 4 A mtdhne mctsron IS made and the small bowel is evtscerated to the ammal’s right stde, exposmg the spleen and the splemc hrlum 5 A small nontraumattc vascular clamp IS placed on the splemc htlum to occlude the splemc vessels wrthm the hrlum one-half mm before to 5 mm after the hepatocyte injection (4.5) Using a l-cc syringe, 1 mL of the hepatocyte suspension 1s injected mto the mfertor pole of the spleen using a 26-gage needle, slowly l-2 mm (see Notes 4). After the mjectton 1s completed, the needle 1swithdrawn and hemostasts obtained by ligating the mjectron site with a 4-O stlk suture. The abdommal contents are returned to their anatomic positions and the abdomen IS closed m two layers, using 3-O Prolene suture The rat IS momtored postoperattvely for any stgns of comphcatrons.

3.4. Implantation of Microcarrier-Attached in the Peritoneum

Hepatocytes

1 Mtcrocarrter preparatron and hepatocyte attachment (36,37) Cytodex 3 collagencoated dextran microcarrters are hydrated m 125 mL of PBS and incubated at 37°C for 90 mm The PBS IS removed and the microcarrters are washed with 50 mL of DMEM/IO% FCS Addmonal DMEM/lO% FCS IS added, to make a total volume of 125 mL The mrcrocarrrer suspension 1s then transferred mto 175-mL ttssueculture flasks and Incubated (5% COJhumidrfied au) at 37’C for 60 mm. The hepatocyte pellet obtamed after the tsolatton procedure IS resuspended m a small volume of DMEM/lO% FCS, and added to the mtcrocarrter suspenston m the

K/m, Utsunomlya, and Vacant/

440

2 3 4 5 6 7

tissue-culture flasks This cell-mlcrocamer suspension IS incubated (5% CO*/ humidified air) at 37°C for at least 2 5 h, to allow the cells to attach to the mlcrocarrlers The mlcrocarrler-attached cells are then washed with PBS and resuspended m DMEM/lO% FCS to make a 2 x IO6 cells/ml suspension The recipient rat IS anesthetized usmg methoxyflurane mhalatlonal anesthesia, and IS welghed. After the abdomen IS shaved, it IS sterilely prepped with 70% isopropyl alcohol and clmldme solution. A small mldlme mclslon IS made mto the peritoneum Using a lo-mL syringe, 5 mL of the mlcrocarrler-hepatocyte suspension IS injected into the peritoneal cavity, using a 13-gage needle After the injection IS completed, the abdomen IS closed m two layers, using 3-O Prolene suture The rat is monitored postoperatively for any signs of complications.

3.5. Implantation of Microencapsulated in the Peritoneum

Hepatocytes

1 MIcroencapsulatIon of hepatocytes (32,33) The hepatocyte pellet obtained after the lsolatlon procedure 1s resuspended at room temperature m a mixture of 2% sodium algmate and 1 7 mmol/L Matngel. A droplet-generatmg apparatus IS used to form mlcrodroplets of the mixture with a diameter of approx 300-700 p The algmate microdroplets are incubated at room temperature m 28 umol/L poly-Llysme for 15 mm, to form an outer envelop of polylysme. The polylysme-coated algmate mlcrodroplets are further reacted with a 0 2% sodium algmate solution, to complete the formation of the algmate-polylysme-algmate (APLA) membrane The mlcroencapsulated hepatocyte suspension contains approx 1 x 10’ cells/ml 2 The recipient rat IS anesthetized using methoxyflurane mhalatlonal anesthesia, and IS welghed 3 After the abdomen IS shaved, It IS sterilely prepped with 70% isopropyl alcohol and clmldme solution 4 A small mldlme mclslon IS made mto the peritoneum 5 Five mL of the mlcroencapsulated hepatocyte suspension IS suspended m 10 mL of media 6 Using a 20-cc syringe, the 15-mL mlcroencapsulated hepatocyte suspension IS injected into the peritoneal cavity, using a 13-gage needle 7 After the mJectlon IS completed, the abdomen IS closed m two layers, using 3-O Prolene suture 8. The rat IS monitored postoperatwely for any signs of comphcatlons

3.6. Implantation Scaffolds

of Hepatocytes

on Biodegradable

Polymer

1 Polymer scaffold preparation (see Note 5) Chloroform IS used to dissolve the PLLA polymer to yield a 5% solution. This solution 1s loaded into Teflon cylmders packed with sodium chloride particles (250-500 pm) After the chloroform

lmplantatlon of Liver Cells

2

3

4 5 6 7 8. 9

441

has evaporated, the polymer scaffolds, shaped like disks (18 mm x 1 mm), are placed m distilled water for 48 h to remove the salt particles The water 1schanged three times daily After the salt-leaching process, the polymer scaffolds are highly porous (95%), with pore sizes ranging from 250 to 500 ptn The polymer disks are then immersed in a 1% aqueous PVA solution for 16 h, to coat the surfaces with PVA The scaffolds are subsequently dried over 24 h, then sterlllzed with ethylene oxide One to 2 wk prior to hepatocyte implantation on polymer scaffolds, the recipient animals undergo an end-to-side portacaval shunt After the animal is anesthetized with methoxyflurane mhalatlonal anesthesia, the abdomen 1sshaved and sterilely prepped with 70% alcohol and clmldme solution A midline mclslon IS made and the portal vem and mfrahepatlc inferior vena cava (IVC) are identified The portal vem 1sclamped and divided The one end of the portal vem closest to the liver 1s ligated, and the other end closest to the bowel IS anastomosed end-to-side to the IVC, using 8-O monofilament nylon suture. The abdomen IS closed m two layers, using 3-O Prolene suture The animal IS monitored postoperatlvely for any signs of comphcatlons (see Note 6) The hepatocyte pellet obtained after the isolation procedure 1s resuspended m Wllhams’ E media to make a 5 x lo7 cells/ml suspension Four hundred pL of the hepatocyte suspension contammg 2 x lo7 cells are seeded on the polymer scaffold prior to lmplantatlon. The recipient rat 1s anesthetized using methoxyflurane mhalatlonal anesthesia, and 1sweighed. After the abdomen 1s shaved, It 1s sterilely prepped with 70% Isopropyl alcohol and chmdine soluhon A mldlme mclsion 1smade and the small bowel 1seviscerated. The cell-polymer constructs are implanted between the mesenterlc leaves, and secured in place using 5-O Prolene suture. The abdominal contents are returned to their anatomic posltlons and the abdomen is closed m two layers, using 3-O Prolene suture. The rat 1smonitored postoperatively for any signs of complications

4. Notes 1 7 mmol/L bovine dermal collagen (Vltrogen 100, Collagen, Palo Alto, CA) may be substrtuted for the Matngel; however, there IS evidence that Matrlgel may provide a better attachment substratum that may enhance the function of the mlcroencapsulated hepatocytes (33) There IS a wide variation m volumes and concentrations of hepatocyte suspensions infused mto the portal vein, m volumes ranging from 0.05 to 4 mL, and concentrations between 2.5 x lo6 to 3 x 107 cells. It is important to keep m mmd that the risk of comphcatlons from portal venous mjectlon is probably higher with mcreasing volume or concentration of suspensions infused. Other methods used to obtain hemostasis include the application of Surglcel (Ethicon) and Gelfoam (Upjohn, Kalamazoo, MI)

442

Kim,

fJt.sunom/ya,

and Vacant]

4 There ISa wide vanatton in the concentrattons of hepatocyte suspensions InJected into the spleen, rangmg between 4 x lo6 to 1 x IO* cells The volume injected IS usually 21 mL 5 Btodegradable polymer scaffolds can be fabricated from polyglycoltc actd (PGA) (Davis and Geck, Danbury, CT) and copolymers of polylacttc and polyglycohc acids (PLGA) (Medtsorb, Cincmnatt, OH) m addttion to PLLA The stze and shape, the pore stze and porosity, and the chemtcal composttton of the scaffolds can be mampulated, depending on the requirements Dtfferent extracellular matrices can be mcorporated or coated on the polymer scaffolds to enhance cell attachment, functton, and reorgamzatton. 6 Hepatotrophtc sttmulation is important for the engraftment and survtval of heterotoptcally transplanted hepatocytes (43,44) Since hepatotrophtc factors are known to exist m the portal blood (41,42), the creatton of a portacaval shunt (PCS) may allow these factors to bypass the liver and enter the systemtc cnculatton m htgher concentrations Partial hepatectomy has also been shown to have hepatotrophtc effects, and can be performed usmg the techmques prevtously described (48)

References 1 1996 Annual Report of the U S Sctenttfic Regtstry for Transplant Rectptents and Organ Procurement and Transplantation Network-Transplant Data 1988-l 995 UNOS, Richmond, VA, and the Divtston of Transplantatton, Bureau of Health Resources Development, Health Resources and Servtces Admmtstratton, U S Department of Health and Human Services, Rockvtlle, MD 2 Amerrcan Ltver Foundatton (1996) Fact Sheet Hepatms, Liver and Gallbladder Diseases in the United States Cedar Grove, NJ, Amencan Liver Foundatron Russell, P S (1985) Selective transplantatton an emergmg concept Antz Surg 201,255-262. Skalak, R and Fox, C F , eds (1988) Tzssue Engzneerzng, RISS, New York Langer, R and Vacantt, J. P (1993) Tissue engineering Sczence 260,920-926 Raper, S E and Wrlson, J M (1993) Cell transplantanon m hver-directed gene therapy. Cell Transplant 2, 38 1- 400 7 Berry, M N and Friend, D S (1969) Hugh yteld preparation of isolated rat liver parenchymal cells. J Cell Bzol 43,50&520 8 Seglen, P 0 (1976) Preparatton of isolated rat hver cells Methods Cell Bzol 13,2!L83 9 Zhang, H., Mtescher-Clemens, E , Drugas, G , Lee, S. M , and Colombam, P. (1992) J Pedzatr Surg 27, 3 12-3 16 10. Ponder, K P, Gupta, S., Leland, F , Darhngton, G , Fmegold, M , DeMayo, J , et al (1991) Mouse hepatocytes migrate to liver parenchyma and function mdefimtely after intrasplemc transplantation. Proc Natl Acad Scz USA 88, 12 17-122 1 11 Gupta, S., Aragona, E , Vemuru, R. P , Bhargava, K K., But-k, R. D , and Chowdhury, J R (1991) Permanent engraftment and mnctron of hepatocytes dehvered to the hver tmphcatrons for gene therapy and hver regeneratton HepatologV 14, 144-149 12 Matas, A J., Sutherland, D E. R , Steffes, M. W , Mauer, S M , Lowe, A , Stmmons, R L , and NaJartan, J. S (1976) Hepatocellular transplantation for metabolic de& ctenctes. decrease of plasma btlirubm m gunn rats Sczence 192, 892-894

Implantation of Liver Cells

443

13. Sutherland, D E R , Numata, M , Matas, A. J , Sunmons, R L., and NaJarian, J S (1977) Hepatocellular transplantation m acute liver failure Surgery 82, 124-132 14 Sommer, B G., Sutherland, D E. R , Matas,A J., Simmons, R L , andNajanan, J S. (1979) Hepatocellular transplantation for treatment of D-galactosamme-Induced acute hver failure m rats Transplantatzon Proc 11, 578-584 15 Holzman, M. D , Rozga, J , Neuzll, D. F , G&in, D , MoscIom, A D., and Demetnou, A A. (1993) Selective mtraportal hepatocyte transplantation m analbummemlc and gunn rats. Transplantatzon 55, 12 13-12 19 16. Mito, M , Kusano, M , Omshl, T., Saito, T., and Ebata, H (1978) Hepatocellular transplantation morphologlcal study on hepatocytes transplanted mto rat spleen Gastroenterol Jpn 13,48&490 17. Kusano, M and Mlto, M (1982) Observations on the fine structure of long-survlved isolated hepatocytes maculated mto rat spleen Gastmenterologv 82,616-628 18. Jiang, B , Sawa, M ,Yamamoto, T., and Kasal, S. (1997) Enhancement of proliferation of mtrasplemcally transplanted hepatocytes m clrrhotlc rats by hepatlc stlmulatory substance Transplantation 63, 131-135 19 Makowka, L , Rotstem, L E , Falk, R E., Falk, J. A., Zuk, R., Langer, B , Blendis, L M , and PhIllips, M J (198 1) Allogenelc and xenogenelc hepatocyte transplantation Transplantatzon Proc 13, 855-859 20 Vacantl, J P, Morse, M A , Saltzman, W M., Domb, A J , Perez-Atayde, A , and Langer, R (1988) Selective cell transplantation usmg bloabsorbable artificial polymers as matrices. J Pedzatr Surg 23,3-9 21 Uyama, S., Kaufmann, P M , Takeda, T , and Vacantl, J P (1993) Delivery of whole liver-equivalent hepatocyte mass usmg polymer devices and hepatotrophlc stlmulatlon. Transplantation 55,932-935 22 Johnson, L B , Alken, J , Mooney, D., Schoo, B L., Griffith-Coma, L , Langer, R , and Vacantl, J P (1994) The mesentery as a lammated vascular bed for hepatocyte transplantation. Cell Transplantation 3, 273-28 1. 23 Selden, C , Gupta, S , Johnstone, R , and Hodgson, H J F (1984) The pulmonary vascular bed as a site for unplantatlon of isolated liver cells m mbred rats Transplantation 38,8 l-83 24 Sandblchler, P, Then, P, Vogel, W, Erhart, R , Dletze, 0, Phlladelphy, H , et al (1992) Hepatocellular transplantation mto the lung for temporary support of acute hver failure m the rat Gastroenterology 102,605-609 2.5 Jaffe, V, Darby, H , Selden, C , and Hodgson, H J. F (1988) The growth of transplanted liver cells wlthm the pancreas. Transplantation 45,497,498 26 Vroemen, J P A M , Buurman, W A , van der Lmden, C. J , Visser, R , Henwegh, K P M , and Kootstra, G (1988) Transplantation of isolated hepatocytes mto the pancreas Eur Surg Res 20, l-1 1 27 Rlcordl, C , Lacy, P E , Callery, M. P., Park, P. W, and Flye, M. W. (1989) Trophlc factors from pancreatic Islets m combined hepatocyte-islet allografts enhance hepatocellular survival. Surgery 105,218-223. 28 Rlcorde, C., Callery, M P., Lacy, P E., and Flye, M. W (1989) Pancreatic Islets enhance hepatocellular survival m combmed hepatocyte-Islet-cell transplantation. Transplantation Proc 21, 2689-2690.

444

K/m, Utsunomrya, and Vacant!

29 Jntle, R L , Btles, C , and Michalopoulos, G (1980) Morphologtc and hutochemical analysis of hepatocytes transplanted into syngenen hosts. Am J Pathol 101, 115-126 30 Jn-tle, R L. and Mtchalopoulos, G (1982) Effects of partial hepatectomy on transplanted hepatocytes Cancer Res 42, 3000-3004 3 1 Takeda, T., Krm, T H , Lee, S K , Langer, R , and Vacantt, J P (1995) Hepatocyte transplantation m biodegradable polymer scaffolds using the dalmatian dog model of hyperurtcosurta Transplantation Proc 21,635S636 32 Dtxit, V, Darvast, R., Arthur, M., Brezma, M , Lewin, K , and Gttmck, G (1990) Restoratton of liver function m gunn rats without mnnunosuppression using transplanted microencapsulated hepatocytes Hepatologv 12, 1342-l 349 33 Dtxit, V , Arthur, M , Remhardt R , and Gitmck, G (1992) Improved function of mtcroencapsulated hepatocytes m a hybrid bioartificial liver support system Artzf Organs 16,33&341. 34 Mooney, D J , Kaufmann, P M , Sano, K , McNamara, K M , Vacanti, J P, and Langer, R (1994) Transplantatton of hepatocytes using porous, btodegradable sponges Transplant&on Proc 26,3425-3436 35 Mooney, D J , Park, S , Kaufmann, P M , Sano, K , McNamara, K , Vacanti, J P, and Langer, R (1995) Biodegradable sponges for hepatocyte transplantation J Blamed Mater Res 29, 959-965 36 Demetriou,A A , Whiting, J F., Feldman, D., Levenson, S M , Chowdhury, N R , Mosctoni, A D , Kram, M., and Chowdhury, J R (1986) Replacement of liver function m rats by transplantation of microcarrter-attached hepatocytes Sczence 233,1190-l 192. 37. Demetrtou, A A., Whiting, J., Levenson, S. M , Chowdhury, N R , Schechner, R , Michalski, S., Feldman, D , and Chowdhury, J R (1986) New method of hepatocyte transplantation and extracorporeal liver support Ann Surg 204,259-27 1 38 Demetriou, A A , Retsner, A , Sanchez, J , Levenson, S M , Mosctom, A D , and Chowdhury, J R (1988) Transplantation of microcarrier-attached hepatocytes mto 90% partially hepatectomtzed rats Hepatology 8, 1006-l 009 39 Mooney, D , Hansen, L , Vacantt, J P , Langer, R , Farmer, S., and Ingber, D (1992) Swnchmg from differentiation to growth m hepatocytes control by extracellular matrix J Cell Physzol 151,497-504 40 Asonuma K , Gilbert, J. C., Stem, J E , Takeda, T , and Vacantt, J P (1992) Quantttatton of transplanted hepattc mass necessary to cure the gunn rat model of hyperbilnubmemta J. Pedlatr Surg 27,298-301 4 1. Marchtoro, T. L , Porter, K. A., Brown, B I., Otte, J B , and Starzl, T E (1967) The Effect of Partial Portacaval Transposttton on the Canme Ltver Surgery 61,723-732 42 Starzl, T E , Francavilla, A , Halgrimson, C G , Francavilla, F R , Porter, K A , Brown, T. H , and Putnam, C. W (1973) The ongm, hormonal nature, and action of hepatotrophic substances m portal venous blood. Surg Gynecol Obstet 137,179-l 99 43 Sano, K , Cusick, R A., Lee, H , Pollok, J M., Kaufmann, P. M , Uyama, S., Mooney, D., Langer, R., andvacantt, J P (1996) Regenerative signals for heterotopic hepatocyte transplantation Transplantation Proc 28, 1859,186O

Implantation of her

Cells

445

44 Kaufmann, P M , Sano, K , Uyama, S , Takeda, T , and Vacant], J P (1994) Heterotoptc hepatocyte transplantatron* assessing the impact of hepatotrophtc strmulatton Transplantation Proc 26, 2240,2241 45 Nteto, J A , Escandon, J , Betancor, C , Ramos, J., Canton, T , and Cuervas-Mons, V (1989) Evidence that temporary complete occlusion of splemc vessels prevents masstve embohzatton and sudden death associated with mtrasplemc hepatocellular transplantatron. Transplantatzon 47,449,450 46 Gupta, S, Vemum, R P, Lee, C D , Yerneni, P R., Aragona, E , and Burk, R D (I 994) Hepatocytes exhibit superior transgene expression after transplantatton into hvet and spleen compared with peritoneal cavity or dorsal fat pad* tmphcations for hepattc gene therapy Hum Gene Ther 5,959-967 47 RaJvansht, P , Kerr, A , Bhargava, K K , But-k, R D , and Gupta, S (1996) Efficacy and safety of repeated hepatocyte transplantation for stgmficant liver repopulatton m rodents Gastroenterology 111, 1092-l 102. 48 Htggms, G M. and Anderson, R M (193 1) Experimental pathology of the liver. I. Restoratton of the liver of the white rat followmg partial surgical removal Arch Path01 12, 186202

33 Isolation and Long-Term Maintenance Hepatocytes in Culture

of Adult Rat

Fraqois Berthiaume, Ronald G. Tompkins, and Martin L. Yarmush 1. Introduction Long-ten-n and stable hepatocyte culture systemshave a wide variety of uses, both m basic science and m the development of hepatocyte-based applications In most cases,long-term cultures of hepatocytes are supertor to traditional cultures m collagen-coated dishes, which only transiently express a low level of liver-specific function during the first wk m culture (1,2). The collagen sandwich provides a system capable of maintaining long term and stable function of hepatocytes with which to study liver physiology (3,4). This system is now used along with several other long-term hepatocyte culture techniques that have been developed since the mtd 1970s. These other methods include the use of special extracellular matrix (ECM) materials, such as an extract from the Engelbreth-Holm-Swarm sarcoma grown in mice [5] (under the commercial appellations of Matrigel and Biomatrix, Biomedical Technologies, Stoughton, MA), co-culture with mesenchymal, endothehal, or epithehal cells (6--8), special culture media (e.g., dimethyl sulfoxide supplementation or argmine-free formulas), and culture at high seeding densities. In the context of studying liver physiology and morphogenests of the liver plate, the sandwich culture system appears to be particularly well-suited, since tt exhibits m vtvo-like ECM geometry, has relatrvely flexible medium requirements, and mdivrdual cell morphology and structure can be easily visualized. One disadvantage, however, is that, m current practice, the ECM layer on top of the cells may present a transport barrier that can slow down the exchange of nutrients, products, and chemical signals with the bulk of the medium. This could be a major disadvantage for kmetic studies requiring addition and removal of factors over a time-scale of From

Methods Edlted

in Molecular

by J R Morgan

Medicine, and

Vol 18 Tissue Engineering

M L Yarmush

447

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

448

Berthiaume, Tompkins, and Yarmush

less than 1 h (e g., amino acid transport studies). This problem could be overcome by developmg techniques for layering very thm films on top of the cells. Current apphcatlons that have used sandwiched rat hepatocyte cultures include. the mvestrgation of hepatrc tissue physiology and toxicology (9%II), the development and opttmtzation of hepatocyte and liver preservation techniques (12,13), and the development of clnucal apphcattons (e.g., bloartttictal hver) (14-16). The followmg IS a de&led protocol to isolate hepatocytes from rat hvers and place them m long-term culture, using the collagen sandwich technique

2. Materials All unspecified

2.1. Isolation

chemicals

were from Sigma (St Lotus, MO).

of Hepatocytes

from Rat Liver

1 Stock solutrons and media (can be made m large quantmes and well m advance) a N-(2-hydroxyethyl)piperazme-Ar-2-ethanesulfomc acid (HEPES) stock solution, 1000 mM b Calcmm chlorrde solutron, 11 mA4. c 10X HBSS Dissolve 4 0 g KCl, 0 6 g KH2P04, 80 0 g NaCl, 3 5 g NaHC03, 0 9 g Na2HP04 7H20, and 10 0 g o-glucose m 975 mL of distilled water. Adjust the pH to 7 4 Sterihze by filtratron through a 0 2 w pore membrane d 10X Dulbecco’s modified Eagle’s medmm (DMEM) dissolve powder with high glucose, without sodium pyruvate, wtthout sodmm btcarbonate (GtbcoBRL, Grand Island, NY, #12100) m one-tenth the recommended amount of water Filter-sterilize through a 0.2~pm porous membrane 2 Soluttons prepared the day of Isolation* a. Krebs-Ringer buffered salme (KRB) 154 n-J4 NaCl, 27 mM KCl, 5 5 mA4 o-glucose, and 5.4 mM NaHC03, 20 mM HEPES. Adjust pH to 7 4 b Ethylenedlamme tetraacetrc acrd solution (EDTA) 1 mMEDTA m KRB, stir at low heat until dissolved c. Collagenase solutron (prepare less than 3 h prior to tsolatton). Dissolve 0.05% w/v collagenase powder (type IV, Sigma) m KRB and add 5% v/v CaC&, solutton Stir for 30 mm at room temperature, filter through a 0 45-pm-pore membrane, and then a 0 2-pm membrane Keep at 37°C until use. d Percoll (Pharmacra, Milwaukee, WI) Mix 43 2 mL manufacturer’s stock wrth 48mLofIOXDMEM 3 Rats (Charles Rover, Wrlmmgton, MA) Female Lewrs; preferred srze IS 180-220 g 4 Surgrcal instruments, hardware a Perfusion apparatus (Fig. 1). b Autoclaved/sterile surgical mstruments (surgrcal blade, sctssors, toothed tweezers, fine curved forceps) c 6-O silk sutures d. 16-gage 2 in AngiocathTM iv catheter/needle unit (Becton Dtckmson Vascular Access, Sandy, Utah)

Adult Rat Hepatocytes In Culture

449

Fig 1 Collagenase perfusion system for hepatocyte lsolatlon e Surgical prep solution (Demson Pharmaceuticals, Pawtucket, RI) 5 Collagen 1 1% rat tail tendon collagen m I mMHC1. It is prepared as follows (17) Tendons are dissected from rat tans and stnred m 200 mL of 3% v/v acetrc acid overnight at 4°C The solution IS filtered through layers of cheesecloth and centrrfuged at 12,OOOg for 2 h The supernatant 1s precipitated with 40 mL 30% w/v sodtum chloride added m a dropwise fashion. The pellet IS collected by centrifugation at 4000g for 30 mm. The pellet is dissolved m 50 mL of 0 6% (v/v) acetic actd, and the solution dialyzed against 500 mL 1 mM HC15X For stenhzation, 0 15 mL chloroform IS added to the solution which IS then stirred for 2 d loosely capped to allow evaporation of the chloroform The yield is approx 100 mg/rat tail 6 Hepatocyte culture medium a DMEM with high glucose, without sodium pyruvate, without sodium btcarbonate (Gbco-BRL) b Defined fetal bovine serum (JRH Btosctences,Lenexa, KS) Add 10% v/v to DMEM c Glucagon (Ltlly, Indtanapolts, IN) Add to DMEM to a final concentratton of 7 ng/mL d Insulin (Sqmbb, Prmceton, NJ): Add to DMEM to a final concentration of 0 5 U/mL e Eptdermal growth factor (Becton Dtckmson, Bedford, MA) Add to DMEM to a final concentration of 20 ng/mL f. Hydrocortrsone (UpJohn, Kalamazoo, MI). Add to DMEM to a final concentration of 7 5 pg/mL

450


g Pemctlhn-streptomycm, respecttvely, 5000 U/mL and 5 mg/mL m 0 9% NaCl (Sigma). Add 1% (v/v) to medium h Falter-stertltze medmm through a 0 2-pm membrane 7 Charactertzatron of hepatocytes a Ftxattve 4% paraformaldehyde m phosphate buffered salme (PBS), freshly prepared For making 100 mL, first boll 50 mL of dH,O on a hot plate and add 4 g paraformadehyde whtle sttrrmg Add 2-3 drops of 6 A4 NaOH to cause the paraformaldehyde to drssolve. Add 40 mL of dH,O and 10 mL of 10X PBS (Btoflutds, Rockvrlle, MD), and allow to cool down on ice. Keep at 4°C until use b 0 1% Trtton X-100 m 1X PBS c 0 1% bovine serum albumm (BSA) m 1X PBS d Rhodamme-phallordm (Molecular Probes, Eugene, OR) Dtssolve 300 U m 1 5 mL methanol and dtlute loo-fold m 1X PBS e Btotm Blocking System X0590, Peroxtdase Blockmg Reagent S200 1, Rabbit Primary Umversal Peroxtdase Kit K0684, Dako, Carpintera, CA Prepare streptavrdtn reagent and substrate-chromogen reagent as descrtbed In manufacturer’s protocol f Anti-rat albumin, rabbtt IgG fraction (Cappel, Cochranvtlle, PA) Drlute 25 pL product m 10 mL 1XPBS

3. Methods 3.7. isolation

of Hepafocytes

from Rat Liver

The method for tsolatmg endothehal cells presented below 1s a modtficatton of the procedures ortgmally published by Seglen et al (28) Proper orgamzatron and the operator’s care in followmg stertle handling techniques are cnttcal for the success of the tsolations. Furthermore, tt IS extremely Important to handle collagenase-digested tissue and dissociated liver cells as gently as possible, because they are very sensmve to mechanical damage

3 1.1 PerfusIon and Collagenase DIgestion Fill the reservotr wtth 70% ethanol Turn on the pump and set flow rate to 50 mL/ min Flush the system with 70% ethanol for 15 mm (this requires a total amount of ethanol of 750 mL) When the ethanol level 1s almost at the bottom of the reservon, fill wtth dHzO and let the system flush with water Refill the reservoir with water and let the system flush again When the water level IS almost at the bottom of the reservoir, add 100 mL KRBEDTA to the reservotr When the hqurd level IS almost at the bottom of the reservoir, add 400 mL KRB-EDTA to the reservoir and place the outlet of the system m the reservoir to allow rectrculatton Turn on the gas Anesthetrze rat wtth ether and place tt on a perforated plate Sterrhze the ventral

area usmg surgtcal prep solutton Open the abdominal cavity and push the mtes-

Adult Rat Hepatocytes m Culfure tines to the side to expose the portal vem and inferior vena cava. Place a ligature around the portal vem proximal to the liver near the htlus and another one distally between the mesenterlc vems Tte ligatures loosely Insert the catheter into the portal vem and shde the cannula past the ligature proxtmal to the hver, wnhout gomg beyond the portal btfurcatton When blood starts to flow slowly out of the catheter, tte the distal ligature Connect the outlet of the perfusion system to the catheter. Immediately cut open the mfertor vena cava and tte the second hgature Open the thoracrc cage through the diaphragm, and cut open the mferior vena cava above the diaphragm When the flutd level m the reservoir 1snear the bottom, add the collagenase solunon to the reservoir Care should be taken to prevent bubbles reaching the hver If a bubble 1s seen to go past the bubble trap, temporarrly detach the perfusion system outlet from the catheter to let the bubble pass. When all the collagenase has gone through the liver, disconnect the catheter from the perfuston system and place the liver in a IOO-mm dash contammg - 1.5 mL cold KRB. Place on Ice and take to a laminar flow hood. The perfuston system should be extensively flushed with dHzO, and then stop the pump

3.12 Purification of Liver Cell Suspension All the followmg operations must be performed while keeping the cells at 04°C. 1 Hold the liver at the ttp of the vascular tree with tweezers, and, using a surgical rake, break open the ltver lobes Usmg a combmation of agttatton and raking, dtsperse the brown-colored material around the liver’s vascular tree (the latter 1s not digested by the collagenase) 2 Prewet a 250~pm mesh-size nylon filter (Small Parts, Mtamt Lakes, FL) with sterile KRB, and pass the cell suspenston through the mesh by using a gentle back and forth swnlmg motton Use additional sterile KRB to clear any remammg cells from the filter Repeat procedure usmg a 62-pm mesh-size nylon filter (Small Parts) 3. Altquot the cell suspenston m 50-mL conical centrifuge tubes and centrtfuge at 202g for 5 min. Resuspend all pellets in KRB to a final total volume of 50 mL. Centrifuge at 202g for 5 mm Resuspend all pellets in KRB to a final total volume of 50 mL and distrtbute m four 12 5-mL aliquots m 50-mL comcal tubes 4. Add 12 mL Percoll per tube Cap ttght and mix by fhppmg the tube up and down. Centrifuge at 60g for 5 mm Asptrate supernatants, which may contain many cells, and resuspend the pellet m KRB to a final volume of 50 mL 5 Centrrfuge cell suspension at 202g for 5 mm and resuspend m KRB to a final volume of 30 mL. Keep on ice until use 6. For cell counting, take a 100~pL abquot of cell suspension and mtx to 800 pL of 1X PBS and 100 pL of 0.4% trypan blue solution (Sigma). Yields are typtcally in

452

Berthlaume, Tompkms, and Yarmush the range of 1N-250 90%

Y lo6 vtable cells, and the fractton of vtable cells IS -8O-

3.2. Long-term Culture of Freshly Isolated Rat Hepatocytes The followtng procedureISfor culturestn standard60-mm dashes.Other stzedishes can be used aslong asthe seedingdensityis maintamedat around 1O5cells/dish 1 Place dashes on a small tray that fits m the incubator 2 MIX 9 parts of of 0 1% rat type I collagen m 1 mM HCl wtth 1 part 10X DMEM at &4”C Precoat 60-mm culture dashes (ttssue-culture-treated) by pourmg 1 mL of the reconstttuted collagen solutton u-reach dash, spreadmg it on the enttre surface of the dash before gellation occurs Incubate dashes at 37°C and 10% COz/ 90% air atmosphere to allow the collagen to gel and the pH to equthbrate to 7 4 3 Wet each collagen-coated dash with 2 mL of hepatocyte culture medium Asptrate medtum 4 Resuspend the cells to a final concentratton of lo6 cells/ml, using hepatocyte culture medmm Ptpet 2 mL of thts suspenston per 60 mm dish This yields cultures that are approximately half confluent The hepatocyte suspension settles very qutckly, and thus tt 1s recommended to load the ptpet wtth enough suspension to seed only three dashes at a ttme The suspenston should be swirled before loading the pipet again, m order to ensure reproductbthty m the seedmg 5 Just prtor to placmg the dishes m the Incubator, use a quick back and forth motton to spread the cells evenly over the entire collagen surface m the dishes Caution: Do not use a circular motton, since this would have the oppostte effect 6 Place overmght m an mcubator at 37°C m a 10% CO* atmosphere The next day, aspirate the medmm and place 1 mL/dtsh of reconstttuted collagen on top of the cells Place 30 mm m the Incubator to gel the collagen, and add 2 mL of fresh hepatocyte culture medium 7. Cultures are fed daily

3.3. Characterization of Cultured Hepatocytes The procedures outlined below are designed to assessthe expression of dlfferenttated, liver-specific characteristics m cultured hepatocytes. Because freshly isolated cells may require as much as 1 wk of culture before dtfferentratton, these assays are best done at 1 wk postseedmg, or later. 3.3.1. Stainrng for IntraceNular Albumin Albumm

1sexclusively synthesized m liver by hepatocytes. Thus, poslttve albu-

mm staining is unequivoqual evidence for differentiated hepatocytes.Lack of staming indicates lack of expression of differentiated functron or a contammatmg liver nonparenchymal cell. The followmg protocol was developed by S. Bhatta (19) 1 Rmse cultures with PBS and fix with 4% paraformaldehyde at 0-4”C for 30 mm Rinse with PBS again, and add 0.1% Trtton X- 100 for 10 mm

Adult Rat Hepatocytes in Culture

453

Fig. 2. Morphological appearance and hepatocellular-specific characteristics. (A) Hepatocyte monolayer as seen under the phase microscope. (B) Staining for actin microfilaments using rhodamine-phalloidin. (C) Intracellular albumin staining by immunofluorescence techniques.

Berthiaume, Tompkms, and Yarmush

454

2 Rmse wtth PBS, and incubate wrth avldm and blotm blocking reagents for 20 mm each. 3 Rinse with PBS, and treat with peroxidase blockmg reagent. Remove excess

blockmg reagent without rinsing 4 Incubate wtth anttrat albumm anttbody solution for 20 mm

5 Rinse with PBS Treat with antirabbit IgG peroxidase stammg krt 6 Observe by transmission microscopy Typical morphology is shown in Fig. 2

3 3.2 Staming for Actin Microfilaments Differenttated hepatocytes have been shown to form mtercellular bile canahcull, which are surrounded by a belt of actm mrcrofilaments (10). Actm mtcrotilaments are not normally found elsewhere m the cytoplasm and do not form stress fibers, except m poorly differentiated cultures, for example, on a smgle collagen gel (20-21). Thus, actin stammg, being extremely simple and specific, IS a useful way to assess the state of dtfferentiation of hepatocytes m cultures. 1 Rinse cultures with PBS and fix with 4% paraformaldehyde at @-4”C for 30 mm Rinse with PBS again, and add 0 1% Trtton X- 100 for 10 mm. 2 Rinse with PBS 3X, and incubate with rhodamme-phallotdm for 30 mm at 37°C. 3 Rinse with PBS 3X, and observe by fluorescence mtcroscopy, usmg a standard rhodamme filter set Typical morphology IS shown m Fig. 2

4. Notes 1 The expressron of drfferenttated functrons and markers by sandwiched hepatocytes may vary, depending on the chemical compositton of the substrate For example, mtercellular gap Junctions have not been detected m type I collagen sandwtch cultures, but are expressed when using Matrtgel(22) 2 Hepatocytes have relatively hrgh metabohc and oxygen-consumptron rates For thus reason, the oxygen tension near the cells IS usually constderably lower than at the medmm-gas phase Interface Thus, when seeding 2 Y lo6 cells/60 mm dash, tt IS not recommended to add more than a 1 mm thickness ofmedmm on top of the cells, to ensure adequate drffusron from the gas phase to the cells Furthermore, the oxygen tension near the cells IS a fimctton of seeding density, a factor that may have to be taken into account when comparmg cell function at drfferent seeding densmes A more complete dtscusston of thts Issue can be found m ref. 23

Acknowledgments Thts work was supported in part by the Shrmers Hospitals for Children. The authors would like to thank Erika Swmmch for the preparation of Fig. l., and Ulysses Balls for Fig. 2C.

References 1 Clayton, D F and Darnell, J E , Jr (1983) Changes m liver-specific compared to common gene transcription during primary culture of mouse hepatocytes. A401 Cell Btol 5,2623-2632

Adult Rat Hepatocytes in Culture

455

2 Clayton, D F , Harrelson, A L , and Damell, J E , Jr (1985) Dependence of ltverspectfic transcrtptton on ttssue organization Mel Cell Bzol 5,2623-2632 3 Dunn, J. C Y, Yarmush, M L , Koebe, H G , and Tompkins, R G (1989) Hepatocyte functton and extracellular matrtx geometry: long-term culture m a sandwtch configuration FASEB J 3, 174-l 77. 4 Berthtaume, F , Moghe, P V, Toner, M , and Yarmush, M L (1996) Effect of extracellular matrix topology on cell structure, function, and physiological responsiveness hepatocytes cultured m a sandwtch configuration. FASEB J 10, 1471-1484 5 Bissell, D M , Arenson, D M , Maher, J J , and Roll, F. J (1987) Support of cultured hepatocytes by a lammm-rtch gel Evidence of a functionally stgmficant subendothehal matrix m normal rat liver. J Clzn Zrzvest 79, 801-812 6 Kurt-Harcuch, W and Mendoza-Ftgueroa, T (1989) Cultivatton of adult rat hepatocytes on 3T3 cells: expressron of various hver differentiated functtons Dzfferentzatzon 41, 148-157 7 Gutllouzo, A., Delers, F , Clement, B , Bernard, N , and Engler, R. (1984) Long term production of acute-phase protems by adult rat hepatocytes co-cultured with another cell type m serum-free medium Bzochem Bzophys Res Commun 120,3 1 l-3 17 8 Conner, J , Vallet-Collom, I , Daveau, M , Delers, F , Htron, M , Lebreton, J -P, and Guillouzo, A (1990) Acute-phase-response mduction m rat hepatocytes co-cultured with rat liver epttheltal cells Bzochem J 266, 683-688 9 Bader,A , Rmkes, I H B , Closs, E I, Ryan, C M , Toner, M , Cunningham, J M , Tompkms, R G , and Yarmush, M L (1992) A stable long-term hepatocyte culture system for studies ofphysiologtc processes. cytokme sttmulatton of the acute phase response m rat and human hepatocytes Bzotechnol Prog 8,2 19-225 10 LeCluyse, E L , Audus, K L , and Hochman, J H (1994) Formation of extenstve canalicular networks by rat hepatocytes cultured m collagen-sandwich configuration Am. J Physzol 266, C1764-Cl774 11 Rotem,A , Matthew, H W T, Hsiao, P H , Toner, M., Tompkms, R G , andyarmush, M L, (1995) The acttvtty of cytochrome P450 IA1 in stable cultured rat hepatocytes Toxccol In Vitro 9, 139-149 12 Borel-Rmkes, I H M , Toner, M., Sheehan, S. J., Tompkms, R. J , and Yarmush, M L (1992) Long-term functtonal recovery of hepatocytes after cryopreservation m a three-dtmenstonal culture configuration. Cell Transplantatzon 1,281-292 13 Stefanovtch, P, Toner, M , Ezzell, R. M., Sheehan, S. J , Tompkms, R G , and Yarmush, M L (1995) Effects of hypothermia on the functton, membrane mtegrity, and cytoskeletal structure of hepatocytes C’ryobzology 23, 389-403 14 Bader, A , Knop, E , Boker, K , Fruhauf, N , Schuttler, W , Oldhafer, K , et al (1995) A novel btoreactor design for in vitro reconstructton of m viva liver charactertsttcs Artzf Organs 19,368-374 15 Tagucht, K , Matsushtta, M , Takahashr, M , and Uchino, J ( 1996) Development of a btoarttfictal liver with sandwiched-cultured hepatocytes between two collagen gels Artzf Organs 20, 178-I 85 16 Kotke, M , Matsushita, M , Taguchi, K., and Uchmo, J. (1996) Function of culturmg monolayer hepatocytes by collagen gel coattng and coculture with nonparenchymal cells. Artzf Organs 20, 186-192


456

17 Elsdale, T and Bard, J (1972) Collagen substrata for for studies on cell behavior J

Cell Bzol S&62&637. 18 Seglen, P 0 (1976) Preparation of isolated rat liver cells Methods Bzol 13,29183 19 Bhatla S (1997) Controllmg cell-cell mteractlons m hepatlc tissue engmeermg using mlcrofabrlcatlon PhD Thesis, Harvard-MIT 20. Dunn, J. C. Y, Tompkms, R G , and Yarmush, M L (1991) Long-term m vitro function of adult hepatocytes m a collagen sandwich configuratlon Blotecknol

Prog 7,237-245 21 Ezzell, R M , Toner, M., Hendricks, K , Dunn, J. C Y, Tompkms, R G , and Yarmush, M L (1993) Effect of collagen gel configuration on the cytoskeleton m cultured rat hepatocytes Exp Cell Res 208,442-452. 22 Moghe, P V, Berthlaume, F, Ezzell, R M , Toner, M , Tompkms, R G , and Yarmush, M L (I 996) Role of extracellular matrix cornposItIon and configuration m maintenance of hepatocyte polarity and function B~omatenals 17, 373-385 23. Yarmush, M L , Toner, M., Dunn, J C Y, Rote, A, Hubel, and Tompkins, R G (1992) Hepatlc tissue engmeermg. Development of cntlcal technologies Ann NY

Acad Scz 665.238-252

34 Design and Fabrication Arterial Bioprosthesis

of a Small Caliber Hybrid

Gilbert J. L’ltalien and William M. Abbott

1. Introduction Despite two generations of vascular prosthetic research, the opttmal arterial surgical replacement for aortocoronary, femoropopltteal, and femorodtstal reconstructton remains autogenous saphenous vein. It has become apparent from numerous mvestrgattons that the main reasons for the advantageous clmtcal performance of autogenous saphenous vem are its anttthrombogenic endotheltal surface and its elastic properties which resemble those of the host artery. The thromboresrstant surface of the autogenous vem assures satisfactory patency, and favorable btomechanical properttes mmlmize the development of occlustve neomttmal hyperplasia. Clearly, these factors must be given serious consideration m the design of an arterial prosthesis. The ideal arterial replacement ~111 possess a thromboreslstant surface, which can withstand arterral fluid shear rates, and btomechanical properties that are similar to those of the host artery. In thus chapter, we describe the design and fabrication of a hybrid blood vessel substttute that IS comprised of three components: a naturally compltant s111cone rubber matrix; a basement membrane composed of 0.1% fibronectm, and a lrvmg, confluent human saphenous vem endothelral cell monolayer that can withstand arterial fluid shear rates and arterial cyclrc strain. We will describe m detail the comparatrve bromechamcal properties of the hybrid prosthesis, and we will characterize the attachment, growth, and differentiation properties of the vascular wall cells comprtsmg the hybrid prosthesis

From

Methods m Molecular Medmne, Vol 18 7kme Engmeenng Methods and Protocols EdIted by J 13 Morgan and M L Yarmush 0 Humana Press Inc , Totowa, NJ

457

L’ltalren and Abbott

458 2. Materials 2.7. Fabrication of Silicone Rubber Tube Matrix 2 1 1 Preparation of the Silrcone Rubber Solutron

1 Silastic medical grade elastomer (SoluttonA MDX4-4210, Dow-Cornmg Medical, Arlmgton, TN) 2 Curmg agent (Solution B 1,I,1 trtchloroethane, Fisher Scientific, Pittsburgh, PA)

2 1.2. Fabrication of Silicone Rubber Tubes 1 Glass rods (Fisher Scientific) 2 1% gelatm (Difco, Detrott MI) 3 Harvard mfusion pump (Model 907, Harvard Apparatus, Natick, MA) 4 15-336-26 ultrasomc cleanmg solution (Fisher Scientific)

2.7.3. Applrcatlon of the FIbronectrn Basement Membrane Prior to Cell Culture 1 0 01% fibronectm (Collaborative

Biomedtcal)

2.2. Cell Culture 2.2 I Establishment of Vascular Wall Cell Cultures 2 3 4 5 6 7. 8 9 10 11 12

0 1% collagenase (CLS II, Worthmgton) 0 5% human serum albumm (Fraction V, Sigma,, St Louis, MO) 0 1% soy bean trypsm mhtbttor (Stgma). M I99 (Glbco-BRL, Grand Island, NY) Hanks’ salts (Stgma) HEPES (Sigma) 25 mmol/L Sodium btcarbonate (Stgma). 0.2%. Fetal calf serum (FCS, Hyclone defined): 20% Heparm (Sigma) 150 mg/mL L-glutamme (Gibco-BRL) 2 mmol/L Pemcillm and streptomycin (Gibco-BRL) 100 U/mL, 100 mg/mL, repectively EC growth supplement (ECGS, Collaborative Research) 50 pg/mL

2 2 2 Appkatlon

of Cells onto the Sdicone Rubber Tube Matrix

1 Blood-gas analyzer (Model 1306, Instrumentation 2 Ml 99 supplemented with 20% FCS 3 50 ug/mL ECGS, and 150 mg/mL heparm

Laboratories,

Lexmgton, MA)

2 2.3. Arterial Pulsatile Fluid Shear and Cyclic Strain Analysis 1 2 3 4.

Bellows pump (Model 14250-007, Gorman Rupp, Bellville, OH) Electromagnetic flowmeter (Statham SP2202, Statham Instruments, Oxnard, CA) P23 ID pressure transducer Strtp chart recorder (Hewlett-Packard, Model 7702B)

Hybrid Arterial Biopros thesis

459

5 Video-tracked motion analyzer(VMA; Motion Analysis, SantaRosa,CA) 2.2 4. Determination of Cell V/ability and Maintenance of Differentiated Products After Application of Physiological Pressure, Pulsatile Flow, and Wall Strain 1 Particle counter (Coulter) 2 Dl-I-acetylated LDL (BTI, Stoughton,MA) 3 4% paraformaldehyde(Fisher Scientific). 3. Methods 3.7. Fabrication of Silicone Rubber Tube Matrix 3.1.1. Preparation of the Silicone Rubber Solution Slhcone rubber solution IS prepared from a Silastlc medical grade elastomer and a curmg agent The elastomer consists of 30% Sllastlc medlcal grade elastomer and 70% of the solvent. The curmg agent 1smixed 1.2 5 with the solvent to generate solution B The final solution IS prepared by mlxmg solution A with solution B m a 10.1 ratio The obtained solution 1sthoroughly mlxed with a magnetic stu-rer, poured mto a 50-mL glass cylinder, degassed under vacuum for 30 mm, and than allowed to sit for 24 h before use. The entlre procedure 1s performed at room temperature m a chemical fume hood 3 1 2 Fabrication of Stlicone Rubber Tubes Before immersion, glass rods are coated with 1% gelatin at room temperature m distilled water by dipping, to facilitate the separation of the Sllastlc tube from the rod The gelatinized rods are dried in an oven at 70°C for 10 min prior to use. A vertically mounted Harvard infuslon pump 1sused to lower the glass rod mto the Silastlc solution at a constant rate (4 mm/s), thus achieving a umform wall thickness. After dlpping, the Sllastlc covered rods are placed honzontally m a drying oven a 70°C for 20 mm. By varying the number of dips and the diameter of the mandnl, a range of elastic moduh and calibers are achlevable that approximate those observed m human blood vessels The Slhcone rubber tubes (5 mm id x 115 mm m length) are then cleaned ultrasomcally for 10 min m a 1 20 dilution of 15-336-26 ultrasonic cleanmg solution m distilled water, followed by two 1O-mm ultrasomc washes m dlstllled water only. Tubes are sterilized under UV light for 48 h m a lammar flow hood, then autoclaved. The resultant tube matrlces are transparent and readily visible with conventional phase contrast and fluorescence microscopy. Tubes are cannulated with a Luer connector and fastened to the connector via 2.0 silk suture for delivery of blologlc materials. Cannulated tubes are mounted to a specially deslgned tube holder (see Fig. 1) comprised of a transparent plastic box and tube

460

L’ltal~en and Abbott

Pressure

tran8mitter

Fig 1 Schematicdiagram of the tube holder occluder. The box 1sdesigned to hold two tubes and to accommodate the axis of a rotation device (not shown). The rotation device 1sused to slowly rotate (10 revolutions/h) solutions and cell suspensions wlthm the Sllastlc tubes, to achieve even coating or dispersion. The tube occluder IS a rectangular, transparent segment of plastic mounted to a threaded rod, which permits both vertlcal and longitudinal movement. Thus, different posltlons along the length of the tube can be occluded for mlcroscoplc observation. 3 1.3 Application of the Fibronectm Basement Membrane Prior to Cell Culture: Solutions of 0.01% fibronectin m sterile dHz0 are prepared Just prior to use. Tubes are filled with the fibronectm solution, then closed (via three-way valve Inserted into the connector piece) and rotated at 10 revolutions/h for 1 h at room temperature The fibronectin was removed and replaced with complete medium, and the tubes are rotated for an addltlonal hour at room temperature. Prior to cell seeding, tubes are emptied of complete medium, and cell suspensions are added at a density of 1O5cells/cm2 surface area. 3.2. Cell Culture 3.21. Establishment of Vascular Wall Cell Cultures Human saphenous vem endothehal cells (HSVEC) are isolated by enzymatic treatment (0 1% collagenase, 0.5% Human serum albumm, and 0 1% soy bean trypsm mhlbltor) of the lumenal surface of the vessels Prtmary cultures are established, maintained, and passagedon gelatm (l%)-coated tissue-culture plastic Cells are plated to the fibrinectm-coated slhcone rubber tubes m passages36 HSVEC are cultured m M 199 contammg Hanks’ salts and supplemented with HEPES (25 mmol/L) to which are added sodium bicarbonate (0.2%), FCS (20%,), heparm (I 50 mg/mL,), L-glutamme (2 mmol/L), pemclllm and streptomycin (100 U/mL and 100 mg/mL, respectively), and ECGS (50 pg/mL)

Hybrid Arterial Bioprosthesis

461

Incubator 17

Tube holder

:OlltrOl

loop 3mllmln L

Frg 2 Schematic diagram of the pumpmg system

3.2.2. Application of Cells onto the Silicone Rubber Tube Matrix Suspensions of known cell denslttes are apphed to two parallel tubes and Incubated at 37°C m an atmosphere of 95% an/5% COZ, with rotation at 10 revoluttons/h. Routinely, plating 1sat a density of 1 x lo5 cells/cm2. The total internal surface area of each tube, including connecting pieces, 1s24 cm2. The porosity of the silicone rubber to gas exchange permits the pH of the culture medium to be maintained at neutrality, as determmed visually from the indicator color of the culture medium. This can be corroborated from direct measurements of pH and pOz m the culture medium, using a blood-gas analyzer Measurements can be made on ahquots of medium withdrawn from the tubes. After 24-48 h rotanonaltncubanon,cellscan be examrnedvrsually andphotographed with an revertedmicroscopeusrngphasecontrastopncs.Cells can be marntatnedon the tubeswith M 199supplementedwtth 20%FCS,50pg/mL ECGS,and 150rng&& hepann. 3 2 3. Arterial Pulsatile Fluid Shear and Cyclic Strain Analysis After the mmal24 h incubation penod, tubesmay be subjectedto physlologlcal conditrons of combined pressure,flow, and wall stram,equivalent to that observed m penpheral artenes(pressure*125/75mmHg, with meanof 100mmHg; pulsattleflow: 235/l 00 mL/mm, with mean of 120mJJmm; frequency: 1 Hz; compliance : 5 + l%/ mmHg 1e2) To accomplish this, a specially designed flow and pressure dehvery systemis required. The pulsatlle component of the systemis a bellows pump drrven by an adjustable eccentric cam that generatespulsatrle flow, pressure,and diameter waveforms (see Fig. 2). Pump outflow IS connected to a dampmg Inflow reser-

462

1 ‘ltalien and Abbott

votr, which serves to eliminate bubbles and to regulate pulse pressure by varymg the air volume. The inlet lme to the tube must be tsodtametric with the hybrid bioprosthesis, and long enough to ensure that velocity profiles under pulsatile flow are parabolic and fully developed. The mean pressure of the system is adjusted by external pressurization of the inflow reservoir. Pulsatile flow rates can be measured using a cannulatmg electromagnetrc flowmeter. Pressure can be measured using a pressure transducer coupled to a strrp chart recorder. Btomechamcal properties (compliance, elasttc modulus) can be measured using a video-tracked motton analyzer (VMA). The VMA uses a high-contrast tmagmg technique to measure the two-dimensional translational movement of an object. The image 1sconverted to a gray-scale pixel file, and the motion of the object can be computed using the commercially available VMA software 3.2.4. Determinatron of Cell Viabihty and Maintenance of DIfferentrated Products After Application of Physiological Pressure, Pulsatile Flow, and Wall Sham In addition to the use of trypan blue exclusion, vlabllrty of cells, after culture on silicone rubber tubes in the presence of pulsatrle flow, pressure, and strain, may be assessedby subsequent passage and growth determmatron. The presence of cell-specrfic proteins m these cultures may also be examined. Specifically, after harvest of cells from tube cultures, ahquots of the cell suspenston can be plated at known densities, on fibronectm-coated plastic culture wells. At 24 h and progressive time intervals followmg plating, the cells may be harvested and counted. Cell harvest from tubes 1s accomplished as follows: The medium is removed, tubes are rinsed m Hanks’ solution, and cells are enzymatrcally removed from the lumenal surface wtth trypsm-EDTA, then collected for counting. The drssociatton is stopped by the addition of complete medium or serum, and cells are pelleted by centrifugatron Pellets are resuspended m trypan blue m PBS. Cells may be examined and counted in a hemocytometer of particle counter. These determmatrons serve to demonstrate the maintenance of sterihty and viability m the tube cultures EC cultures can also be Incubated m dr-I-acetylated LDL, or fixed m fresh 4% paraformaldehyde and immunostamed with antibody to factor VIII-related antigen 3.3. Results of Analyses 3.3.1. Effects of Pulsatile Flow, Strain, and Pressure on the Adhesion and Orientation of HSVEC to Sihcone Rubber Matrix In our experience, 48-h apphcatton of physiological levels of pulsatile flow, pressure, and strain should not cause detachment of vascular wall cells from


463

Fig. 3. Phase contrast micrographs of HSVEC of the hybrid bioprosthesis. After 48 hr of pulsatile flow and cyclic strain, dense monolayers remained. Cells consistently demonstrate alignment in the direction of flow. the tube surface. Dense monolayers remained in the presence of 48 h of pulsatile flow, strain and pressure (Fig. 3). Cell alignment occurs consistently in the direction of flow, i.e., along the axis of the tube. To quantify the actual numbers of cells retained on control tube surfaces, cells were enzymatically harvested from the tubes and total cell numbers determined. The mean percentage of cells recovered from the bioprosthesis (i.e., adhesion) was 98 + 10%.

3.3.2. Cell Viability and Maintenance of Differentiated Silicone Rubber Tube Matrices

Phenotype on

Viability, as assessed by trypan blue exclusion of cells recovered from the bioprosthesis, was 89 f 2%. To further assess viability, cells isolated from tubes were replated at known densities onto fibronectin-coated wells of tissue-culture plastic, and grown for up to 12 d. Initial cell densities after 24 h incubation, as well as the final cell numbers, were determined. Although the absolute numbers varied between experiments, the final cell densities reached were similar for cells harvested from tubes as for cells grown on flat silicone rubber surfaces (Fig. 4). Immunostaining with antibodies to factor VIII-related confirmed that cells isolated from the tubes retained differentiated characteristics.

3.3.3. Biomechanical Properties of the Hybrid Bioprosthesis The stability of the elastic properties of the hybrid bioprosthesis was assessed over a period of 28 d of pulsatile flow and cyclic strain at a mean

L’ltahen and Abbott

464 15 12 9 6

0

4

8

12

16

20

24

Time (days) Fig 4 Growth curves of HSVEC on flat surfaces of silicone rubber (SIL) coated with fibronectm and cells isolated from the hybrid bioprosthesis after 48 h of pulsattle flow and cychc strain Cells were plated at known densities onto the indicated surfaces, and harvested and counted at regular mtervals over an approx 3-wk period A, Flat,

n , hybrid bloprosthesls

pressure of 100 mmHg and a mean flow rate of 150 mL/mm (mean shear -3dyns/cm2). No change was observed in the cucumferentral compliance or dtameter of the btoprosthesis durmg the 28 d period (Table 1). We also compared the elastic modulus (E&of the btoprosthesis to that of a human femoral artery (Table 2) At phystologic pressures (90-l 10 mmHg), the values E,,, were similar for both the artery and hybrid prosthesis A range of compliance values can be obtained for the bioprosthesis by varying the wall thickness (i.e., dtp frequency) of the silicone rubber matrix. As shown m Table 3, there 1s a lmear relationship between thrcknessradtus ratio and complrance (r = ,988); the range of comphance values shown encompasses that of a variety of canme and human blood vessels.

4. Notes 1, Blomechanical computations Values of dynamic compliance (independent of phase) were calculated from the systohc and dlastollc diameter(D), and pressure

(P) waves generated by the pumping system as follows Ccm

= @sys

- Dd,as)

/ Ddms

x cpsys

- Pd,as)

(1)

Compliance values were then averaged for 8-10 waves and reported as units of %dlmenslonal

change/mmHg

x lo”

Hybrid Arterial Biopros thesis

465

Table 1 Results of 28-d Stability Time (days)

Experiment

Diameter (*SD)

0 7 14 21 28

Compliance (LSD)

6 14+ 02 6.28 f 02 6.24 _+ 03 6 19 3~.02 6.21 21 03

Table 2 Comparison of Einc for the Hybrid and a Human Femoral Artery

3.58 3 64 3.76 3 67 3 54

Bioprosthesis

Pressure (mmHg)

Hybrid bioprosthesls

90 100 I10

13.3 x IO6 dyn cm-* 130 x 106dyncm-2 12 1 x lo6 dyn cm-2

Table 3 Comparative Compliance of Biologic for the Hybrid Bioprosthesis Type of vessel Canme Canme Human Canine Human

Vessel comphance

carotid a. femoral a femoral a femoral v saphenous v

_+ 33 3~.37 AZ.36 31 26 k .27

Human femoral artery 10 0 x 1O6dyn cmM2 11 Ox 106dyncm-2 14 6 x 10” dyn cm-*

Vessels and Matching

Values

Hybrid prosthesis compliance

T/R

141+12 11 5+0.6 60+1 1 54+07 4 6 -I- 0.5

06 07 09 10 11

139+08 115+09 59+10 55+06 45+07

An incremental elastic modulus was computed for both the clrcumferentlal and longltudmal displacements according to formulae described by Dobrm (5). For the clrcumferentlal modulus* E ,ncc,rc= ATI (AR 1 Rd,as)

(2)

where T = PRIh Here, T and R represent the difference between calculated values of T (stress) and R (radius) at systole and dlastole, respectively, and h IS wall thickness The mclemental longltudmal modulus was obtained as follows. E mc,ong

= ATI

AL

f Ldw

where T = PR* / [R2 - (R - h)*] + F / [nR* - n(R - h)2]

(3)

466

L’ltalren and Abbott

Here, T is the sum of stresses caused by pressure (P) and traction force (fl Estimates of F were obtamed from a series of m vitro experiments under pulsatile phystologtc pressures, as follows Length was measured m the pressurized vessel, which was subsequently coupled at one end to an isometrtc force transducer (Grass Instruments, FlO, Qumcy, MA) The mean longttudmal traction force was 30g All values of T and E,,, are expressed as dyns/cm2 2 Cell Culture The siltcone rubber tubes must be thoroughly cleaned ultrasonically for 10 mm m a 1.20 dilution of 15-336-26 ultrasonic cleanmg solution (Fisher Sctenttfic) m dtsttlled water, followed by two IO-mm ultrasomc washes m distilled water only, Exceedmgly low levels of contaminant (e.g., the curing agent) will result in cytotoxicity and/or loss of adherence of cells to the fibronectm basement membrane For the same reason, all tubmg and reservotr stoppers must be comprtsed of siltcone rubber

Acknowledgment We gratefully

acknowledge

the gift of slhcone

rubber

soluttons

from Tho-

mas W. Broadhagen and Linda M. Veresh of Dow Cormng, Midland, MI, and also the suggestions they provided. In additton, we wish to thank Patricia M. Joseph, Pulmonary Unit, MassachusettsGeneral Hospital, for her help with the use of the blood-gas instrumentation References 1 Sumpio, B E (1993) Role of HemodynamlcsUI Vascular Endothelzal Bzology R G Landes,Austm, TX. 2 Levesque, M J , Nerem, R M , and Sprague,E A (1990) Vascular endothehal cell prohferatton m culture and the Influence of flow Blomaterlals 11,702-707 3 Iba, T , Shm, T , Sonoda,T , Rosales,O., and Sumpto, B E (1991) Stimulatton of endothehal secretion of tissue-type plasmmogenactivator by repetitive stretch J Surg Res 50,457-460 4 Sterpetti, A V., Cucma, A , Santoro, L , Cardillo, B , and Cavallaro, A. (1992) Modulatton of arterial smoothmusclecell growth by haemodynamicforces Eur f Vast Surg 6, 16-20 5 Sutcliffe, M C and Davidson, J M (1990) Effect of stattc stretching on elastm production by porcme aorttc smooth muscle cells Matrzx 10, 148-153 6 Berguer, R , Higgms, R F , and Reddy, D J (1980) Inttmal hyperplasia An expenmental study. Arch Surg 115,332-335 7 Buck, R C ( 1983)Behavior of vascularsmoothmusclecellsduring repeatedstretchmg of the substratumm vttro Atherosclerosis46,2 17-223 8 Dtamond, S. L , Eskm, S. G , and McIntire, L V (1989) Fluid flow sttmulates tissue plasmmogenactivator secretion by cultured human endothehal cells Scrence 243, 1483 9 Diamond, S. L , Share&n, J B , Dieffenbach, C , Frasier Scott, K , McInttre, L V, and Eskm, S G. (1990)Tissueplasmmogenactivator messengerRNA levelsincrease


10 11 12

13

14 15 16

17

18

19 20 21 22

23 24.

25

26

467

m cultured human endothehal cells exposed to lammar shear stress.J Cell Physzol 143,364-371 Dobnn, P B., Ltttooy, F N , Golan, J , et al (1988) Mechanical and histologic changes m canme vem grafts J Surg Res. 44,259-265. Hume, W. R (1980) Prolme and thymtdme uptake m rabbit ear artery segments m vitro increased by chrome tangential load. Hypertension 2, 738-743 Leung, D Y M., Glagov, S., and Mathews, M. B. (1977) Elastm and collagen accumulation m rabbit ascending aorta and pulmonary trunk during postnatal growth Corrclatton of cellular synthetic response wtth medial tension Czrc Res. 41,3 16-323. Sumpio, B E., Banes, A J , Link, G W, and Johnson, G J (1988) Enhanced collagen productton by smooth muscle cells during repetittve mechanical stretchmg Arch Surg 123, 1233-1236. Sumpto, B E and Banes, A. J (1988) Response of porcine aorttc smooth muscle cells to cychc tensional deformations m culture J Surg Res 44, 696-701. Sumpio, B. E , Banes, A J., Levm, L G., and Johnson, G J (1987) Mechamcal stress stimulates aortic endothehal cells to proliferate J Vast Surg 6,252-256 Sumpto, B E , Banes, A J , Lmk, G. W , and Iba, T (1990) Modulation of endotheha1 cell phenotype by cychc stretch. mhibmon of collagen productton. J Surg Res 48,4 I 5-420 Sumplo, B E. and Banes, A J (1988) Prostacyclm synthetic activtty m cultured aorttc endothehal cells undergoing cychc mechamcal deformation Surgery 104, 383-389 Sumpio, B E and Wtdmann, M D (1990) Enhanced productton of endothelmmderived contracting factor by endothehal cells subJected to pulsatile stretch Surgery 108,277-282 Towne, J B , Qumn, K , Salles-Cunha, S , et al (1982) Effect of increased arterial blood flow on locahzation and progression of atherosclerosis Arch Surg 117, 1469-1474 Wolmsky, H (1970) Response of the rat aorttc medra to hypertension Morphological and chemical studies Czrc Res 26,507-522 Zwolak, R M., Adams, M C , and Clowes, A W (1987) Kmettcs of vem graft hyperplasta Association with tangential stress. J Jdzsc Swg 5, 126-36 Sharefkm, J. B , Diamond, S L , Eskin, S. G., McInttre, L V, and Dieffenbach, C W (1991) Fluid flow decreasespreproendothehnmRNA levels and suppressesendothelm1 pepttde release m cultured human endothelial cells. J J&SC Surg 14, l-9 Carost, J A , Eskm, S G , and McIntne, L V (1992) Cyclical stram effects on production of vasoactive matenals m cultured endothehal cells J Cell Physlol 151,29-36 Dewey, C F J , Bussolart, S R,, Gimbrone, M A Jr., and Davies, P F (1981) The dynamic response of vascular endothehal cells to fluid shear stress J Blomech Eng 103,177-185 Eskm, S G , Navarro, L T , O’Bannon, W., and DeBakey, M E (1983) Behavtor of endothehal cells cultured on Stlasttc and Dacron velour under flow condttions m vitro’ implications for prelmmg vascular grafts with cells Artzf Organs 7,3 l-37 Frangos, J A, Eskm, S G., Mdntne, L V, and Ives, C. L (1985) Flow effects on prostacychn productton by cultured human endothehal cells. Sczence227, 1477-1479

468

L’Italien and Abbott

27 Sterpeth,A V, Cucma,A , D’ Angelo, L S., Cardrllo, B., and Cavallaro,A (1992) Response of arterial smooth muscle cells to lamrnar flow. J. CU~&XXZ.SCSung 33,6 19624 28 Upchurch, G R , Jr, Banes, A J , Wagner, W H , et al. (1989) Differences m secretion of prostacyclm by venous and arterial endothehal cells grown m vitro m a stattc versus a mechanically active envrronment J Vast Surg 10,292-298 29 Benbrahim, A , L’Itahen, G J., Milmazzo, B B , Warnock, D F, Dhara, S., B S , Gertler, J P , Orkin, R. W., and Abbott, W M. (1994) A comphant tubular device to study the mfluences of wall stram and fluid shear stress on cells of the vascular wall J Vast Surg 20, 184-194 30 Hasson, J E., Wrebe, D. H , Shareikm, J B , and Abbott, W. M (1986) Mrgratron of adult human vascular endothehal cells effect of extracellular matrix proteins Surgery 100,38&91 3 1 McGurre, P G , Castellot, J J., and Orkm, R. W (1987) Srze-dependent hyaluronate degradation by cultured cells J Cell Physlol 133,267-276 32. Sambrook, J , Frrtsch, E F , and Mamatis, T. (1989) Molecular Clonzng, A Laborutov Munual, 2nd ed Cold Sprmg Harbor Laboratory, Cold Spring Harbor, NY

35 Methods for the lmmunoisolation and Transplantation of Pancreatic Cells Anthony

M. Sun

1. Introduction Although the admmistration of msulin by mjectton is clearly a life-savmg intervention for patients devoid of p-cells, this approach falls short of the remarkable titration of insulin delivery and consequent control of glucose levels achieved by normal, healthy mdivtduals. In the absence of the physiological control of the plasma glucose concentrations, the dally mlectton of msulm has not been able to prevent the common comphcations of the disease, namely nephropathy, retmopathy, and neuropathy, as well as vascular complicattons. This has been confirmed m the recent diabetes control and complications trial which has demonstrated that intensive treatment of pattents with msulmdependent diabetes melhtus (IDDM), wtth ttght glycemic control close to the control range, effectively delays the onset, and slows the progression, of the vartous dtabetic comphcattons (I). Therefore, it becomes mandatory to develop methods, applicable early m the course of the disease, and m any type 1 dtabetic patient, for obtammg perfect metabolic control without increasing the risk of severe hypoglycemia. Consequently, the transplantation of islet tissue, either as whole pancreas or as isolated islets, has been pursued, because these techniques can provide near-normal blood glucose control, and thus have the potential to prevent diabetic complications. Because of the many problems associated with the whole-pancreas transplantation, grafts of isolated pancreatic islets represent the most promtsmg approach to the restoration of the endocrine function of the pancreas m patients with IDDM However, the problem of immune rejection remains. This has meant that the transplantation of msulm-producing tissue has largely been confined to dtabettc recipients who already have a kidney transplant or who are From

Methods Edrted

by

in Molecular J R Morgan

Medmne, and

Vol 18 Tmue

M L Yarmush

469

Engrneerrng

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

470

Sun

receiving a kidney simultaneously with an islet transplant, because of the risks of long-term mununosuppression, which are considered to outweigh the potential benefits of transplantation m a newly diagnosed diabetic patient. To tackle the problem of immune rejection, a number of studies have mvestigated techniques to decrease the mununogemcity of transplanted islets, and thus reduce for immunosuppression. These approaches, known as immunoalteratton, are based on the hypothesis that islet tmmunogemcity is caused by passenger leukocytes (2) or dendritic cells (3), and not by the endocrine cells. Mampulations designed to destroy the dendrmc cells and decrease islet mununogemcity include: culturing the islets for prolonged periods at room temperature (4) or m 95% 02 (5), exposmg the islets to UV radiation (6), cryopreservatton (7) of islets, and pretreating the islets with antibody to Ia anttgen plus complement (8) Despite consrderable progress m the use of immunoalteration m past years, this approach contmues to have serious hmitations. There IS still no effective method of consistently protectmg islet transplants from rejection by means of tmmunoalteratton, with or without immunosuppression No successful transplantation of m-tmunoaltered islets mto large animals has been reported, and mdtcattons are that such transplants would require nnmunosuppresstve therapy m combmation with donor islet pretreatment Although depleting donor tissue of passenger leukocytes before transplantation may temporarily prevent rejection, it does not eliminate the possible vulnerabthty of the transplanted tissue to a recurrence of the ortgmal autotmmune attack. Because of the difficulties with overcommg the problem of tmmunorejection and autoimmune rejection, the concept of mununoisolation has been advanced Thts is achieved by enclosing the pancreatic islets by semipermeable and biocompatrble membrane (bioartificial pancreas). The enclosed islets would act m the host as they had in the donor, provided the surrounding membrane was impermeable to higher mol wt antibodies, but permeable to oxygen, glucose, other substances, and/or the mternally generated hormones Thus, the encapsulated cells would respond to external substrate concentrations (e g., blood glucose), and the required hormone (e.g., msulm) would be secreted mto the systemic circulation. The clmical benefit of this approach is that diabetic patients would be provided with normal pancreatic islets that not only would be protected from mnnunorqection, but would secrete, m addition to msulm, other hormones, such as glucagon, somatostatm, pancreatic polypeptides, and possibly other islet protems, in response to physiologtcal demand. This approach has the potential not only to allow allogemc transplantation without tmmunosuppresston, but also to allow the use of xenografts In the form of a vascular implant, the islets can be distributed m a chamber surroundmg the membrane, and the device 1simplanted as a shunt m the vascular system (9,lO).

Transplantation

of Pancreatic Cells

471

The limitations of the vascularrzed devrces are all associated with the creation of an arteriovenous shunt, i.e., requirement of major surgery, posstble vascular thrombosts, and a potentral for cardiac stresscaused by a stgmficant volume of shunted blood. Most important, an accrdental breakage of such a unit may result m a serrous hemorrhage. Alternatively, the islets can be immunoisolated within diffusion chambers and placed mtraperrtoneally, subcutaneously, or m other sites Numerous devices of this type have been evaluated during the past several decades (II) These include disk-shaped diffusion chambers, Millipore celluloslc membranes, hollow-fiber drffusron chambers, and wider-bore tubular membrane chambers Membrane materials used to make these devices Include polyvmylchloride, polyamides (nylon), polypropylene, cellular nitrate, and cellulose triacetate. The method of Islet transplantation usmg such devices ISfrequently referred to as macroencapsulatlon. The common feature of this approach IS that a great number of islets are placed within the chamber of the device. This severely hmrts the dtffusron dynamics of the system. Consequently, the problem of cell death or dysfunction as a result of oxygen supply hmttations, or accumulatton of wastes or other agents, IS likely to be severe. Chambers retrieved several weeks after tmplantatton often contam a central necrotic core. Another problem IS the frequent breakage of the dtffuston chambers, which can cause not only loss of islet function, but also mtraperttoneal inflammatory responses. The problems posed by the large volume of the implanted devices, as well as a possible fibrosis of the chamber surfaces, can further aggravate the sttuatton. Typically, the use of these devices can result only in a short-term amehoration of diabetic hyperglycemia. To overcome these problems, we developed semipermeable, algmatepolylysme-algmate (APA) btocompattble capsules to enclose mdtvtdual islets (mlcroencapsulatton). The encapsulation of individual islets results m removmg most problems associated with the macroencapsulatton concept as discussed above. Smce our ortgmal mventron in 1980 (12), many applicattons of this concept have been developed, with many centers now studying cell encapsulation. We have considerably improved the capsule constructron since its conception in 1980. The btocompattbility of the capsules was improved by replacing the crude alginate by a more homogenous preparation of lower VIScosity and high purity, and by replacing the outer polyethylenetmme layer by a second layer of low-vtscostty alginate. The size of the capsules was reduced from the original dtameter of 0.8 mm to the present 0.25-0.35 mm This was made possible by constructron of an electrostattc droplet generator (13). The smaller capsules are stronger, with tmproved spherrcrty and surface smoothness. It has been demonstrated that the size of the microcapsules, their sphertc-

472

Sun

ity, surface smoothness, and, most important, the strength of the capsules, all have an important bearing on the life-span of the graft. In our earlier experiments using rodents as both donors and recipients, we demonstrated that both allografts and xenografts of microencapsulated pancreatic Islets were protected from tmmunorelection, and that m both streptozotocm-induced and spontaneously diabetic mouse and rat recipients, diabetes could be reversed for the life-span of the animals (14-17). Recently, m an unprecedented preclinical study, we have demonstrated that mtraperitoneal xenografts of microencapsulated porcine islets mto naturally diabetic nonhuman primates result in normaltzatton of diabetic hyperglycemia for more than 1 yr, without recourse to exogenous msulm or mnnunosuppression (18). This is the first report ever on long-term discordant xenograft function resultmg m physiological glycemic control without recourse to immunosuppression m a large-animal model. The use of large animals and, eventually, humans m transplantation studies necessitatesthe development of a plentiful source of islet tissue. The shortage of human donor islets makes it obvious that tf the goal of islet transplantation is to succeed, it is necessary to develop methods making it possible to transplant nonhuman islets. Currently, the xenograft of porcine islets remains the most favorable solution. The choice of porcine islets is Justified, because it would represent an unlimited source of tissue, and because porcine insulin IS very similar to human msulm m its structure, and was used for decades m the treatment of human diabetic patients. In addition, m the pig the regulation of msulm secretton by glucose and other nutrients IS very stmtlar to that m humans. Porcine islet isolation is difftcult, however, because of the islets’ marked fragility, and because of the rapid dissociation of the pancreas mto single cells during the isolation procedure (19-24). In the pig pancreas, loose vascular channels are present within the islet, which makes it very easy for the islet to fragment during the isolation procedure. The method of porcme islet isolation developed in this laboratory (25,26) is highly reproducible m terms of both the quantity and quality of the isolated islets. In the procedure, porcme pancreata are perfused and digested with collagenase, and the islets are then purified on dextran density gradients In order to avoid any damage to the islets, no mechamcal devices nor any strenuous treatment is employed. A thorough separation of islets from the exocrme tissue is critical, since the dextran or Ficoll density purtficatton has proven to be meffectrve for those islets not completely free of acmar growth attachment. The viability and physiological competence of the isolated Islets were demonstrated m both m vitro and in viva studies (2526). The rate of msulm secretion of microencapsulated islets in culture studies was shown to be very similar

Transplantation

of Pancreatic Cells

473

to that of free Islets. Likewrse, the kinetics of insulin release m the perlfuslon study of encapsulated and free islets followed a similar pattern, The physlologrcal competence of porcine islets isolated by our new method m VIVOwas unequivocally demonstrated m our preclmtcal study (IQ, m which xenografts of microencapsulated porcine islets reversed diabetes m naturally diabetic monkeys for over a year, as mentioned earher in this chapter 2. Materials 2.1. Pancreatic Islets Isolation (as an Example of Endocrine Tissue lsola tion) Rat islets are isolated by the collagenase-digestron technique, and either handpicked or purified through discontinuous Ficoll gradients. Briefly, pancreatic tissue is dressed and perfused with Hanks’ balanced salt solution. The minced tissue IS then digested for 12-l 5 mm at 37°C with collagenase (Sigma, St. LOUIS,MO)( 12-l 5 mg/4 mL of Hanks’ solution) Islets can be handpicked from digest with the aid of a dtssecting mrcroscope, or the ttssue 1smixed with 25% Frcoll m a centrifuge tube. Three Frcoll concentrattons (23,20, and 11%) are then layered above this suspensron. Centrifugatlon 1scarried out for 10 mm at 8OOg.The islets are harvested from the interface of the 23 and 20% Frcoll layers. The cells can either be mrcroencapsulated nnmediately or cultured for l-3 d prior to encapsulatron. 2.2. Sodium Alginate Solutions To prepare sodmm algmate solutions, 3.0 g of sodium alginate (Kelco, San Drego, CA) is sprinkled or sifted slowly mto 100 mL of drstrlled water, with sttrrmg. Then 100 mL of 1.8% NaCl solutron is added and well-mixed. The mixture is then centrifuged at 3000g for 1 h at 4°C. The supernatant IS sterrlrzed by filtering through a filtration unit (0.2 pm) and stored at 4°C. The 0 15% sodium alginate solution IS prepared by diluting the 1.5% solution 1O-fold wrth physrologrcal saline. 2.3. Calcium Lactate Solutions To prepare 2.72% calcmm lactate solution (100 mA4), 68.0 g of calcmm lactate 1sdissolved in 2,500 mL (final volume) of distilled water. 2.4. Poly-r-lysine Solution (PLL) Fifteen mg of PLL (Sigma) IS dissolved in 30 mL of salme. 2.5. Sodium Citrate Solutions Trrsodmm crtrrc acid (3.23 g) is dissolved m 100 mL of dlstrlled water, to which 100 mL of salme IS added and mixed (55 n-J4 sodium citrate solutton).

474

Slln

2.6. Culture Medium (for Pancreatic Islets) Medium RPM1 1640 is supplemented wtth calf serum (7 5%), pemcillm (100 U/mL), and streptomycm (100 l.tg/mL). 3. Methods 3.1. Microencapsulation 3 1.1. A/r-jet Technique

Techniques

The an-let technique represented the ongmal procedure for cell mtcroencapsulatton, when it was first descrtbed m 1980 (12). A syrmge pump with a IO-mL syringe connected to a specialJet ts used for making sodmm algmate Islet droplets For each preparation, 200s3000 islets m 0.2 mL of saline is mixed gently with 1 mL of 1 5% sodium algmate, transferred to the IO-mL syringe, and connected to the an Jet. The distance from the tip of the an Jet 23-gage needle to the surface of the collecting fluid ISset precisely at 4 cm. The syringe pump and an flow are then turned on to extrude sodium algrnate droplets containmg islets mto 50 mL of the 1.1% CaC12solution in a beaker. Durmg extrusion, the islets are kept m suspension by gently rotatmg a small magnet instde the IO-mL syringe. After the extrusion process 1scompleted, the spherical calcium alginate gel droplets are transferred to a 50-mL polystyrene test tube with a conical bottom, and allowed to settle before withdrawing the supernatant down to 5 mL, using a vacuum aspirator. The gel droplets are washed once with 30 mL of 55% CaC12,once wtth 0.28% CaC12,and then suspended m 25 mL of 0.1% CHES solution for 3 mm. After aspirating the CHES solution, the capsuleswere washed with 1.1% CaC12and suspendedm 25 mL of 0.05% (w/v) poly-L-lysme for 6 mm. After further washing with CHES, CaCl,, and salme, the microcapsulesare incubated m 0.15% sodium algmate for 4 mm and washed with salme The capsulesare then suspendedm 10 mL of 55 mM sodium citrate solutton for 5 min. The final product is washed twice with salme and once with medium CMRL- 1969, and then transferred to culture flasks for mcubation at 37°C until required for m vitro and m viva studies. The au-Jet technique results m relatively large capsules of approx 0.8 mm m diameter. The capsule construction has improved considerably over the past years It has been demonstrated that the capsule stze 1san important factor m the kmetits of insulin release by encapsulated pancreatic islets, and m the access to nutrients and oxygen. Glucose tolerance m diabetic mice improved significantly after the mice received rat islets enclosed m capsules measurmg 0 3 mm m diameter. The smaller capsules allow for 1 Increasedstrength,with lesserchanceof rupture 2 Increasedcell viabrhty, becauseencapsulatedcells have easier accessto oxygen and nutrients.

Transplantation

475

of Pancreatic Cells

3 Faster cell response to glucose fluctuations, since the dead space tn the capsules IS reduced 4 Slgmficant reduction m the volume of capsules needed for transplants, smce the capsule volume IS dxectly proportlonal to the cube of the capsule radms Thus, If one mllhon mlcroencapsulated Islets should be necessary to normalize hyperglycemla m a type 1 dlabetlc patlent, the overall volume of the intraperltoneally implanted capsules (300 pm diameter) would only amount to 14 1 mL, dropping to Just 8 1 mL if the capsule diameter decreases to 250 ~1 5 Less susceptlblllty to cell overgrowth on capsular surfaces, because the smaller capsules have greater mobility

3 1.2 Electrostatic Droplet Generator The generatlon of smaller capsules was made possible by construction of an electrostatic droplet generator (13). The size of the capsules was reduced from the original diameter of 0.8 mm to the present 0.25-0.35 mm. The smaller capsules are stronger, wrth improved average spherlclty and surface smoothness. The size of the mlcrocapsules, their sphericlty, surface smoothness, and, most important, the strength of the capsules, all have an important bearing on the life-span of the graft. In an in vitro perlfuslon study, rat islets encapsulated m the new capsules showed a response to glucose challenges that was nearly comparable to that of free, unencapsulated islets. In VIVO,followmg transplantation with the smaller capsules, there 1smuch less contact lrrltatlon, which m turn leads to a considerably smaller probability of cell overgrowth on capsular surfaces In the process of droplet generation, the electrostatic droplet generator uses alternate current frequency of 60 Hz, at the pulse length of 0.1 ms. In the encapsulation procedure, pancreatic islets are suspended m 1 5% (w/v) purified sodium algmate at a concentration of approx 3-5 x lo3 islets/ml SpherIcal droplets are formed by an electrostatic

field interaction,

coupled with

syringe pump extrusion (Raze1A 99 syrmge pump, pump speed 55-65, Stamford, CT) and are collected m 100 mM calcium lactate solution. To achieve this, the negative pole 1sattached to the loop, which 1ssubmerged in the calcmm lactate solution, while the positive pole is attached to the needle. The tip of the needle 1spositioned about 1 cm from the surface of the calcium lactate solution. Calcium lactate solution temperature 1s 20°C (all solutions used m the encapsulation procedure are filter-stenllzed) The gelled droplets are washed with 0.9% salme prior to suspension in 0.05% poly+-lysine for 5 mm. The droplets are again washed with 0.9% salme and suspended in 0.15% sodium algmate for 4 mm. After another wash with 0.9% salme, the capsules are allowed to react with 55 n&’ sodmm citrate for 4 mm, and finally washed with 0.9% salme, and with culture medium. Most capsules contam one islet, and have a diameter

of 0 25-O 35 mm. The entire mlcroencapsulatlon

dure 1sIllustrated in Fig. lA,B.

proce-

476

Sun

A

Syringe containing sodium alginate and cell mixture + stirring magnet Needle Nat -alginat cell droW3ts

I

t

Ca+-alginategelled beads

B

Ca+-lactate solution steel loop L Stainless

- _ Gelled capsule ceils l ,*r - containing

t

Capsule covered with poljl-L-iysine (PLL) coating

I

PLL coating covered with second layer * of aiginate

Unbound aiginate is removed with citrate forming an inner cavity Fig. 1. Process of generation of alginate-polylysine-alginate the electrostatic droplet generator.

microcapsule using

Transplantation

of Pancreatic Cells

477

3 1.3. Capsule Improvement In an effort to further improve the strength of the capsule, we studied the parameters affecting the capsules strength (27). It can be postulated that the membrane strength is directly proportional to the thrckness of the capsule. Consequently, we studied quantitattvely different reaction conditions durmg the synthesis of the membrane. 1 Concentration of sodium algmate The algmate concentration was found to affect the membrane thickness within the range of l-2%. The osmotic pressure at thus concentratton IS not harmful to cells or tissues Similarly, the calcium concentration also has a bearing on the membrane thickness The thickness is at its maximum when the concentration of calcium lactate used in the encapsulation procedure is 100 mA4 2 Reaction time of sodium algmate with calcium lactate Another factor mfluencmg the membrane strength is the reaction time of sodium algmate with calcium lactate, 1 e , gelatmation time If the gelation time is shorter than 5 mm, the membrane strength will suffer The membrane strength Increases with time; however, if the reaction time exceeds 1 h, Instead of mcreasmg, the membrane thickness slightly decreases The washing time of calcmm algmate beads with sodium chloride solution is also critical for the capsule strength The membrane thickness decreases as the treatment time increases, especially during the first 15 mm 3 Parameters affectmg the membrane strength and membrane thickness The reaction of algmate beads with polylysme represents probably the most important step regardmg the strength of the APA capsules Four different parameters are essential in deterrnmmg the resultmg thickness of the capsular membrane* a Membrane thickness increases with mcreasmg time, during which algmate beads react with polylysme It is therefore possible to control membrane thtckness by regulatmg the reaction time b The membrane thickness increases with increasing polylysrne concentration c The membrane strength increases with mcreasmg mol wt of polylysme (Polylysmes with mol wt of 24,000 are now used for APA capsule preparation, because of the mol wt cutoffs of the resultmg membrane-around 60,000 kDa d. It can be demonstrated that the membrane thickness increases with a decrease of the pH of the polylysme solutton. 4 Microencapsulation of cells and tissues other than pancreatic islets It should be noted here that the capsules’ chemical composition can be altered for encapsulation of cells and tissues other than pancreatic islets, thus suitmg their special requirements For instance, microencapsulation of hepatocytes (artificial liver) requires a very strong capsule, m order to avoid capsule rupture when encapsulatmg large numbers of cells needed for a proper function of the artificial hver. To achieve thts, high sodium algmate concentrations have to be used and a long reaction time between algmate and polylysme IS required. In other applications, permeabihty of the APA membrane can be altered to suit the specrfic metabolic requirements of the encapsulated cells or tissues.

sun

478 3.2. Tramp/an ta tion of Microencapsula 3.2.1 Allotransplantation

ted Pancreatic

Islets

Transplants (both allografts and xenografts) were performed m dtabettc mace, rats, and monkeys (14-18). In the procedure, mtcroencapsulated islets are delivered mto the peritoneal cavity of the recipient by qection, using an 18-gage catheter (Insyte, Becton Dickmson, Sandy, UT). Single allografts of APA-microencapsulated rat islets, when implanted mtraperttoneally mto rats with streptozotocm-induced diabetes (4 5 x lo3 islets per rat), resulted m the reversion of diabetic hyperglycemia for 21 mo (14). In addition to normalizmg blood glucose concentrations, the encapsulated rat islet transplants reversed polyuria and polydypsia m the recipients. The recipients demonstrated a very rapid mcrease m body wt, and there was no evidence of cataracts. By contrast, untreated diabetic controls showed no significant weight increase during the same period, and developed cataractswtthm 2-3 mo Capsules containing intact viable islets, as revealed by htstologlcal and msuhn secretion studies, were recovered 156, 365, and 648 d postransplantation. The surfaces of most capsules were free of cell attachment and were phystcally intact, with the enclosed islets clearly visible. Scanmng electron microscopy revealed essentially smooth interior and exterior capsular surfaces. In the next stage, rat islet allografts microencapsulated m APA membranes were implanted mtraperitoneally m spontaneously diabetic BB rats (15) The BB rat represents a valuable experimental model of human type 1 diabetes The course of development of the diseasein this model resembles that of human type 1 diabetes, in that there IS an abrupt onset of insulin-dependent, ketosisprone diabetes and lymphocytic msulttis, with virtually complete destruction of the pancreatic p-cells. Therefore, the autoimmune etiology of this animal allows mvesttgatton of the effectiveness of pancreatic islets m an immunoisolated environment. The encapsulatedaltografts reversed the diabetic stateand mamtamed normoglycemia for up to 6 mo. Body wt and urine volumes were normal durmg thts period, and no cataractswere detected in the transplant recipients. Thus, encapsulation of islets provrded total protection from the host mune system,as well as from the autonnmune destruction, while ehmmatmg the need for nnmunosuppression. 32.2. Xenotransplantatron Following the successful allotransplantatron experiments, the first successful xenotransplants of microencapsulated islets were performed Rat islets mtrcoencapsulated m APA.mtcrocapsules were implanted mtraperitoneally mto BALB-c mice with streptozotocm-induced diabetes (16). As a result, normoglycemia was restored in the rectptent animals for up to 308 d.

Transplanta t/on of Pancrea t/c Cells

479

In another xenotransplantatron study, rat islets were Implanted mto spontaneously diabetic NOD mice, an ideal arnmal model of insuhn-dependent diabetes (27). Smgle intraperitoneal transplants of 800 islets per recrprent resulted m an effectrve restoratron of normoglycemia in all expenmental animals for up to 230 d. The development of a plenttful supply of pancreatic islets, as descrtbed earlier m this chapter, led to the mitlatron of a preclmmal study m which mrcroencapsulated porcine Islets were xenotransplanted mto nine spontaneously diabetic cynomologus monkeys (18). After one, two, or three transplants of 3-7 x lo6 islets per recipient, seven of the monkeys became msulm-independent for over 1 yr, with fasting blood glucose levels in the normoglycemic range. Glucose clearance rates m the transplant recipients were significantly higher than before the graft admimstratron, and the insulin secretion durmg glucose tolerance tests was significantly higher, compared to pretransplant tests. Porcme C-peptide was detected m all transplant recipients throughout their period of normoglycemra, but none was found before the graft admmlstratlon Hemoglobm A, c levels dropped stgmficantly within 2 mo after transplantation. Ketones were detected m the urme of all recipients before the graft admmlstratlon, but all experimental animals became ketone-free 2 wk after transplantation. Capsules recovered from two recipients 3 mo after the restoration ofnormoglycemta were found physically intact, with enclosed islets clearly vrstble. The capsules were free of cellular overgrowth. Exammatron of internal organs of two of the animals involved m our transplantatron studies for the duration of 2 yr revealed no untoward effect of the extended presence of the microcapsules.

References 1. The Drabetes Control and Comphcatlons Tnal Research Group (1993) The effect of mtenstve treatment of drabetes on the development and progression of long term complications m msulm-dependent diabetes melhtus IV Engl J A4ed 329,977-986. 2 Sollmger, H W., Stratta, R J , D’Alessando, A M , Kalayoglu, M, Prrsch, J D , and Belzer, F 0 (1988) Experience with srmultaneous pancreas-kidney transplantatron Ann Surg 208,475-483 3 Faustman, D., Hauptfeld, V., Davre, M , Lacy, P. E , and Shreffler, D C (1980) Mulme pancreatic beta cells express H2K and H2D but not Ia antigens. Exp Med 151,1563-l

568

4. Rlcordl, C., Kraus, C , and Lacy, P. E. (1988) Effect of low-temperature culture on the survival of mtratestrcular rat islet allografts. Trunsplantatzon45,465-468 5. Woehrle, M , Markmann, J. F., Silvers, W K., Barker, C F , and NaJi, A (1986) Transplantanon of cultured pancreatic Islets to BB rats Surgery 100, 334-340 6. Kenyon, N S., Strasser, S , and AleJandro, R. (1990) Ultraviolet light mununomodulatlon of canme islets for prolongatron of allograft survival Dzabetes 39,3os-311

480

Slln

7. Cattral, M., Wornock, G., Evans, M., and Rajotte, R. (1991) Transplantation of purified single-donor cryopreserved canine islet allografts with cyclosporine. Transplantation Proc. 23, 777,718.

8. Alejandro, R., Latif, Z., Noel, J., Shienvoid, F. L., and Mintz, D. H. (1987) Effect of anti-Ia antibodies,culture and cyclosporin on prolongation of canine islet allograft survival. Diabetes 36,269-273. 9. Sun, A. M., Parisius, W, Healy, G. M., Vacek, I., and Macmorine, H. G. (1997) The use, in diabetic rats and monkeys, of artificial capillary units containing cultured islets of Langerhans (artificial endocrine pancreas). Diabetes 26, 1136-l 139. 10. Sullivan, S. J., Maki, T., Borland, K. M., Mahoney, M. D., Solomon, B.A., Muller, T. E., Monaco, A. P., and Chick, W. L. (1991) Biohybrid artificial pancreas: long-term implantation studies in diabetic pancreatectomized dogs. Science 252,7 18-72 1. 11. Lanza, R. P., Sullivan, S. J., and Chick, W. L. (1992) Islet transplantation with immunoisolation. Diabetes 41, 1503-1510. 12. Lim, F. and Sun, A. M. (1980) Microencapsulated islets as bioartiflcial endocrine pancreas. Science210,908-910. 13. Hommel, M., Sun, A. M., and Goosen, M. F. A. (1984) Droplet generation. Canadian Patent 458605. 14. Sun, A. M., O’Shea, G. M., and Garapetian, H. (1985) Artificial cells containing islets as bioartificial pancreas. Prog. Art$ Organs 1985, 601. 15. Fan, M., Lum, Z., Fu, X., Levesque, L., and Sun, A. M. (1990) Long-term reversal of diabetes in BB rat by transplantation of microencapsulated pancreatic islets. Diabetes 39, 519-522.

16. Lum, Z. P., Tai, I. T., and Sun, A. M. (1992) Xenografts of microencapsulated rat islets result in prolonged reversal of the diabetic state in streptozotocin-induced diabetic BALB-c mice. Transplantation 53, 118&l 183. 17. Lum, Z. P., Krestow, M., Tai, I., Norton, J., Vacek, I., and Sun, A. M. (1991) Prolonged reversal of diabetic state in NOD mice by xenografts of microencapsulated rat islets. Diabetes 40, 151 l-1515. 18. Sun,Y., Ma, X., Zhou, D., Vacek, I., and Sun, A. M. (1996) Normalization of diabetes in spontaneously diabetic cynomologus monkeys by xenografts of microencapsulated porcine islets without immunosuppression. J. Clin. Invest. 98, 1417-1422. 19. Ricordi, C., Finke, E. H., and Lacy, P. E. (1986)A method for the mass isolation of islets from the adult pig pancreas. Diabetes 35, 649-653. 20. Crowther, N. J., Gotfredsen, C. F., Moody, A. J., and Green, I. C. (1989) Porcine islet isolation, cellular composition and secondary response. Horm. Metabol. Res. 21,59&595.

21. Marchetti, P., Zappello, A., Giannarell, R., Masiello, P., Masoni, A., Casanovi, E., et al. (1988) Isolation of islets of Langerhans from the adult pig pancreas. Trunsplantation Proc. 20, 707,708.

22. Marchetti, P., Socci, C., Davalli,A. M., Staudacher, C., Pierangelo, B.,Vertova, A., Sassi, I., Gavazzi, F., Pozza, G., and Di Carlo, V. (1991) Automated large-scale isolation, in vitro function and xenotransplantation of porcine islets of Langerhans. Transplantation 52, 209.

Transplantation

of Pancreatic Cells

487

23. Ricordi, C., et al. (1990) Isolation of the elusive pig islet. Surgery 107, 688694. 24. Basta, G., Falorni, A., Ostricioli, L., Brunetti, P., and Calafiore, R. (1955) Method for mass retrieval, morphologic, and functional characterization of adult porcine islets of Langerhans: a potential non-human pancreatic tissue resource for xenotransplantation in insulin-dependent diabetes mellitus. J. Invest. Med. 43,555. 25. Sun, Y., Ma, X., Zhou, D., Vacek, I., and Sun, A. M. (1993) Porcine pancreatic islets: isolation, microencapsulation and xenotransplantation. Artif: Organs 17,727-733. 26. Zhou, D., Yang, B., Sun, Y., Vacek, I., Sun, A. M. (1996) Effects of collagenase concentration on the puritry and viability of isolated porcine pancreatic islets for use in xenotransplantation studies. Xenotran.splantation 3, 1 l-l 7. 27. Ma, X., Vacek, I., and Sun, A. M. (1994) Generation of alginate-poly-Llysine-alginate (APA) biomicrocapsules the relationship between the membrane strength and the reaction conditions. Biomat. Artif: Cells Immobilization Technol. 22,43-69.

36 Methods for the Study of Nerve Cell Migration and Patterning Helen M. Buettner and Hsin-Chien

Tai

1. Introduction Recent advances in our ability to microfabricate tissue culture environments (I), and to store and retrieve microscopy data in a digital format, have made it increasingly possible to study fundamental aspects of neuronal migration and response to environmental cues, Neuronal migration refers to the migration of the growth cone at the tip of an extending neurite, a process that guides the neurite to its destination during neural development and regeneration (2). This chapter describes techniques for the observation of neuronal migration on both plain and patterned surfaces through the use of high-resolution, phase-contrast videomicroscopy. These techniques are particularly appropriate for analyzing the dynamics of single growth cone behavior in the presence of two-dimensional environmental microfeatures similar in scale to the growth cone dimensions (l-50 pm). Growth cone response to topographical features (3) or other three-dimensional environments are beyond the scope of this chapter. Substrate preparation procedures are derived from microlithography techniques first used in the microelectronics industry and adapted to cell culture systems (4-7). Glass cover slips are used as the substrate support to provide the required optical clarity for high-resolution microscopy, and patterns are created using laminin or collagen. The major procedures described in this chapter include: 1. Neuronal cell culture. This section focuses on one neuronal cell type, chick dorsal root ganglion (DRG), which represents an important model system for studies of neuronal migration, and complements the neuronal types described elsewhere in this volume. However, protocols for the culture of other types of neurons could be substituted here, and used with the procedures following this section. The From:

Methods in Molecular Edited by: J. R. Morgan

Medicine, Vol. 18: Tissue Engineering Methods and and M. L. Yarmush 0 Humana Press Inc., Totowa,

483

Protocols NJ

484

Buettner and Tar

advantages of the chick DRG system include the fact that it IS a relattvely troublefree source of primary cells, and produces growth cones that are somewhat larger than many other vertebrate neurons, correspondmg to more easily drstmgurshable features 2 Substrate preparatton Micropatterned cell culture surfaces are prepared usmg microlithography techniques. A photoresist pattern 1sfirst created on a glass cover sbp support, which is then further processed to yreld a final patterned, neuronalgrowth-promotmg, btologtcal substrate The use or services of a microfabricatron facility IS assumed for the photorestst patternmg step of the protocol. The method descrrbed can be used to create either lammm or collagen patterns It may be suitable for other neuronal growth substrates that have not been tested specifically with this protocol 3 Videonucroscopy Although an important feature of neuronal migration and patterning studies, vtdeomrcroscopy has become a relatively commonplace laboratory tool, and only the essential points of the techmque are described here The primary requirements are a microscope equipped with a good video camera, a personal computer with frame grabbing capabtlmes, and plenty of storage capacity for the image data

2. Materials 2.1, Tissue Culture of Chick Dorsal Root Ganglion Neurons 2.1.1. Dssection

(DRG)

1 Incubated fertile chrcken eggs. 2 Ca2+-free, Mg2+-free Hanks’ balanced salt solutron (HBSS). Prepare stock from 1 part HBSS 10X liquid (contammg no calcmm chloride, magnesium chloride, magnestum sulfate, sodium bicarbonate, or phenol red (Grbco-BRL, Gaithersburg, MD), and 9 parts dH,O Falter stenltze with a 0 2 pm filter 3 Ethanol* 70% m dHzO 4 Blunt end forceps, sterilized 5 Drssectmg forceps (Dumont No 5, Fme Science Tools, Foster City, CA), stenhzed 6 35-mm sterile plastic Petrr dishes

2 1.2. Cell Culture 1 Serum-free F 12 culture medmm (see Note 1): a. F12 medium: F12 nutrient mixture (146 mg/L L-glutamme, 1176 mg/L sodmm brcarbonate, Grbco-BRL, see Note 2) b. Glucose-glutamme-pen-strep (GGPS). MIX o-glucose 150 mg/mL, Lglutamme 50 mM, pemcrllm-streptomycin 2500 pg/mL-2500 U/mL (all components from Sigma, St Louis, MO) with F12 medium. Filter-sterthze usmg a 0 2-pm filter and ahquot mto 4-mL portions Stable for at least 3 mo at -20°C c Insulm-transferrin-selenmm (ITS) stock. Dissolve ITSTM Premix lyophilized powder (25 mg insulm, 25 mg transfemn, 25 pg selemous acid, Collaborative BiomedrCal, Bedford, MA) m 5 mL F 12 medium. Stable for 3 mo at -20°C (see Note 3)

Nerve Cell Migration and Patterning

485

d Putrescme stock* Dissolve 16.1 mg/mL putrescme powder (Srgma) u-r F12 medium Dilute 1 100 m F 12 medmm to make 100 mA4 stock solutron Frltersterihze with a 0 2-pm filter Aliquot into 5-mL portrons, and store at -20°C. e. Progesteronestock Dtssolve 0 63 mg/mL progesterone (Bgma) in 96% ethanol. Drlute 1.100 tn F12 medmm to make a 2 x 1O-5Mstock solutton Filter-sterilrze with a 0.2-pm filter Ahquot mto 5-mL porhons, and store,rightly wrapped, with Parafilm at -20°C f. Hormone stock. Combme 5 mL each of ITS, putrescine, and progesterone stock soluttons Aliquot into 300~p.L portrons. Stable for 3 mo at -20°C g Bovine pitmtary extract (BPE): Tnm bone, connective tissue, blood vessels, and membranes from 20-25 thawed bovine pituitaries (Pelfreeze Btologtcals, Rogers, AK, see Note 4). Wash wtth distilled water until the rinse water is free of blood, then rinse with Ca*+-free, Mg 2c-free HBSS, and dram Add the ptturtarres and 2.4 mL Ca2+-free, Mg*+-free HBSS/g pituitary to a blender and homogemze at low speed m 10 1-mm pulses with 2-3 mm rests on Ice between pulses (see Note 5) Pour homogenate mto a beaker. Rinse blender with additional HBSS and add to homogenate. Stir 90 mm m 4°C cold room. Centrifuge 40 mm at 10,OOOgand 4°C Dialyze supernatant against phosphate buffered saline (PBS, Dulbecco’s PBS, 0 2 g/L KCl, 0 2 g/L KH,PO,, 8 0 g/L NaCl, 1 15 g/L Na2HP0,, pH 7 4, Glbco-BRL) at 4’C, using Spectra/Par 4 dialysis tubing (mol wt cutoff 12,000-14,000, 16 mm diameter, 2 mL/cm, Fisher Sctentlfic, Sprmgfield, NJ) Dialyze twice, for at least 8 h each. Dialyze against F12 medium at 4’C, once, for at least 8 h (see Note 6). Centrifuge 1 h at 100,OOOgand 4°C. Falter twice through a 0 45-pm filter, then once through a 0 2-pm filter Measure BPE concentratton with a Bradford Assay (Bio-Rad, Cambridge, MA), using BSA as a standard Ahquot mto 0.15-mg portions. Stable for 3 mo at -4°C Procedure from Tsao et al (8), as described by Baird and Raper (9) h Nerve growth factor (NGF). Dissolve 100 pg 7s NGF (mouse submaxillary gland, Collaboratrve Bromedtcal) m 5 mL sterrle F12 medium Ahquot mto IOO-clr, portions

Stable for 1 mo at -2OT

I F12 culture medmm. Add 4 mL GGPS, 300 pJ. hormone stock, 0 1-O 2 mg/ mL BPE, and 100 pL NGF to 95 mL of F 12 medmm to make 100 mL of serum-free F 12 culture medium 2 Tissue Culture Substrate (see Subheading 2.2.)

2.2. Substrate Preparation 22.1. Cleaning 1 Cover slips, 22 mm, No 1 round (Fisher Screntific; see Note 7). 2. Alconox paste. MIX Alconox powder (Fisher Scientific) with a small amount of water to form a viscous slurry 3 Gloves, powder-free. 4. Forceps 5 Deronlzed, drsttlled water 6. Ultrasonic cleaner (55 kHz, 70 W, Fisher Sctentttic). 7. Absolute ethanol.

Buettner and Tar

486 2 2.2. Acid Etching

1 Ftltered au. 2 9 1 v/v sulfurtc acid hydrogen peroxide MIX 9 parts sulfurtc acid (98 wt%) to 1 part hydrogen peroxide (30 wt%) 3 Forceps, Teflon-coated 4 Detomzed, dtstrlled water 5 Petri dash, plasttc, 60 mm 6 Parafilm 7 Desrccator

2 2.3 Photoresist Patterning 1 2 3 4 5. 6 7 8 9 10 11 12 13 14 15

Hexamethyldrsrlazane (HMDS, PCR, Gamesvrlle, FL) Mtcroposrt S18 13 photorestst (Shipley, Marlboro, MA) Mtcropostt MF-3 19 developer (Shipley) 10-mL sterile plastic syringes Acrodtsc CR PFTE syringe filters (HPLC certtfied, 0 2 pm, Fisher Sctentrfic) Deionized water. Filtered mtrogen gas Parafilm Alummum foil Spin coater Contact prmter. UV source 8 mW, 365 nm Desiccator Forceps Photomask Oven

2 2.4. Protein Patterning 1 Use (a) or (b) a Collagen. Rat tall type I (Collaborattve Biomedical) As needed, dilute stock solutton to 150 pg/mL m 0 02 M acetic acid, and filter-sterilize with a 0 2-pm filter b Lammm: EHS mouse tumor (Collaborattve Biomedical) Ahquot stock solution mto single use portions and store at -70°C Stable for 9 mo As needed, dilute one aliquot to 50 ug/mL m Ca2+-free, Mg2+-free HBSS 2 Germrctdal UV light source 3 1% ammopropyltriethoxystlane (AMPS) Combine 1 part 3-ammopropyltrtethoxystlane (98%, Aldrich, Milwaukee, WI) wrth 99 parts dHzO. Make fresh as needed 4 Ultrasonic cleaner. 5 PBS. Dulbecco’s PBS (0.2 g/L KCl, 0 2 g/L KH2P04, 8 0 g/L NaCl, 1.15 g/L Na2HP04, pH 7 4, Gtbco-BRL). 6 Glutaraldehyde Drlute 1 part 25% glutaraldehyde, aqueous, with 9 parts PBS to make 2 5% glutaraldehyde


487

7 Acetone, electronic grade. 8 Bovine serum albumm (BSA). Bovine serum albumin, Fraction V (J. T Baker, PhillIpsburg, NJ) dissolved at 0 5 w/v% in PBS Make fresh as needed

2.3. Video Microscopy 2 3 1. Culture Chamber Preparation Microscope slides Alconox paste Ethanol 70% m dHzO Lens paper Parafilm Metal washer, < 15 mm Id, 225 mm od Valap: Melt together a I. 1 1 mixture of Vaseline, lanolin, and paraffin Cool to room temperature, and store covered until ready to use 8. Paintbrush 9 Wmdex solution. Dilute Wmdex window cleaner with dH20 m a 1 1 ratio 10 Q-tips 2 3. 4 5 6 7

2 3.2. Video Microscopy 1 2 3 4 5 6

MIcroscope equipped with a high-resolution video camera Time-lapse VCR Personal computer equipped with frame-grabbing capablhtles. Hard drive, >l Gb Tape drive, same storage capacity as hard drive Microscope stage incubator with temperature controller

2.3.3. Data Analysis 1 Image analysis software (e.g , NIH Image)

3. Methods 3.1. Tissue Culture of Chick DRG Neurons 3.1.1. DIssectIon 1. Remove egg from egg incubator and wipe the large end with 70% ethanol (see Note 8) 2 Crack the shell at the large end of the egg by tapping with blunt-ended forceps, and remove the shell to create an opening Scoop out the embryo and place m a 35-mm Petri dish containing 2 mL Ca2+-free, Mg2+-free HBSS. 3 Remove head and internal organs with dlssectmg forceps, and transfer embryo to fresh Petri dish of Ca2+-free, Mg2+-free HBSS. 4 Under a dlssectmg microscope, use dlssectmg forceps to remove the skin from the back and around the legs 5 Close the dissecting forceps, and insert mto the area between the spinal column and the leg Slowly open the forceps parallel to the spinal column, to tease tissue

488 apart Repeat until leg IS separated from body. Transfer the leg into a fresh Petri dish of Ca*+-free, Mg*+-free HBSS, and repeat for the other leg 6 Remove exposed DRGs from the interior side of the top of the leg, and transfer mto a fresh Petri dish of Ca*+-free, Mg2+-free HBSS (see Note 9)

3 1.1 Cell Culfure 1 Place prepared substrate cover slips into separate 35-mm Petri dish, and pipet 750 pL F12 culture medmm mto each dish 2 Cut DRG m half and place one-half DRG m the center of each substrate 3 Incubate 4 h at 37°C and 5% CO*, to allow explants to attach to the substrate 4. Add 1 25 mL of medium to each dish, and Incubate overnight (see Note 10)

3.2. Substrate

Preparation

This subheading describes the preparation of mrcropatterned substrates for patternmg nerve growth. It is divided rnto a series of procedures that can be

performed with some space of time between each, if desired, keeping m mind the tendency of the photoresist to degrade wtthm several weeks, followmg application m Subheading 3.2.3. To prepare a uniform substrate, the same procedure can be followed, omittmg the photomask m Subheading 3.2.3, step 4 Many other procedures exist for preparing a uniform substrate, and are readtly available in the general hterature on neuronal growth and migratron. However, this method provides a direct control for comparison with results obtained usmg the patterned substrates. 3.2. I. Cleanmg 1 Wearmg powder-free rubber gloves, rub each cover slip between fingers with Alconox paste for 30 s 2 Rinse cover slips mdivtdually m hot tap water until the water runs off cleanly from the cover slip, then place m a 400-mL beaker of deionized, distilled water 3 Somcate the cover slips m the beaker for 10 mm 4. Transfer the cover slips mdlvidually to a fresh beaker of deromzed, d&led water, and repeat step 3 5 Repeat step 4 6 Rmse wtth 2 L of running deiomzed, distilled water 7 Transfer the cover slips individually mto a beaker of absolute ethanol, and store covered with Parafilm until ready to acid etch.

3 2.2. Acid Etchrng 1 Remove cover shps mdlvidually from ethanol and blow dry with a stream of 0 22-pm filtered an. 2 Etch cover slips m 9: 1 v/v H#04/H202 for 20 mm 3. Transfer the cover slips mdividually mto a fresh beaker of deiomzed, dtstdled water


489

4 Rmse with five changes of deionized, distilled water, then blow dry with a stream of 0 22-pm filtered air 5 Seal cover slips m a Petri dish with Parafilm, and store m a desiccator until ready to apply photoresrst.

32.3

Photoresist Patterning (see Note 11)

1 Spm coat cover slips mdlvldually with 5-7 drops of HMDS (an adhesion promoter) for 30 s at 3000 rpm Apply HMDS with a lo-mL sterile plastic syringe, filtered through a 0 2-p Acrodlsk filter, while cover shp 1sspmmng (see Note 12) 2 Spm coat cover slips mdlvldually with 5-7 drops of MIcroposIt S 18 13 photoreslst for 30 s at 3000 rpm Apply photoreslst with a lo-mL sterile plastic syrmge, through a 0.2-pm Acrodlsk filter, while cover slip IS spinning 3 Bake cover slips at 110°C for 5 mm to cure the photoreslst 4 Using a contact prmter, expose each cover slip to UV light through a photomask for 30 s (see Note 13) 5. Develop exposed photoresist m Mlcroposlt MF-3 19 developer for 30 s, makmg sure to keep track of the coated side of the cover slip (see Note 14) 6 Rmse cover slip m delomzed water and dry with a stream of filtered nitrogen gas 7 Place cover slips m Petri dishes, seal with Parafilm, wrap m alummum foil, and store m a desiccator for up to 3 wk at room temperature, until ready to pattern with protein (see Note 15)

32.4. Protein Patternmg (10)

2

8

9. 10 11 12

Examme photoreslst-patterned cover shps under a microscope for pattern regularity, dlscardmg any with defective patterns. Sterdlze cover slips overnight (8-12 h) under the germlcldal UV lamp of a tlssue-culture hood Somcate cover shps m 1% AMPS for 2 mm Rinse cover slips m deionized, distllled water. Transfer the cover slips into a cover shp holder, and bake at 110°C for 10 mm Incubate photoresist-side-up for 1 h in 2.5% glutaraldehyde Wash 3x with fresh PBS Plpet 200 pL of 50 ,ug/mL lammm or 150 pg/mL collagen onto the cover shp, and mcubate for 1 h. Transfer the cover slips mto the cover shp holder again, making sure that the photoreslst side of the cover slips faces the same direction, and somcate m acetone for 8 mm Transfer each cover slip into its own 35-mm Petri dish containing PBS, and wash 3x with more PBS Incubate for 1 h m 0 5 mL of 0.5% BSA Rinse twice with Ca*+-free, Mg 2+-free HBSS, and once with F12 culture medium The cover slips are ready for use as cell culture substrates at this point

3.3. Video Microscopy The followmg method yields a sealed chamber with a volume of z30 & for mamtammg the neuronal culture durmg extended vldeomlcroscopy observa-

490

Buettner and Tai

non. The sealed chamber prevents evaporation of the culture medrum, IS suitable even for mrcroscopeswith little clearancebetweencondenserand objective, and can be adapted to any objectrve of interestby matching the cover slip thicknessm Subheading 2.2.1.to the workmg distanceofthe objectrve. DRG culturesremamhealthy, with no evidence of change or deterroratron, for 6-18 h followmg the preparation. 3.3 1. Culture Chamber Preparation 1 Clean a mtcroscope shde with Alconox paste,and rrnse thoroughly with derontzed, dtstrlled water 2 Spray 70% ethanol on both sides of the shde and wtpe dry m a cncular motion with crumpled lens paper, paying particular attention to the center of the slrde 3 Check that the slide looks clean with no streaks vrsrble to the eye Repeat steps 1 and 2, as necessary, untrl slide IS clean 4 Place four 3 x 11 mm Parafilm shims on the center of the microscope slide, so that then outer edges comctde with the edges of a 22-mm square, leaving gaps between the shims at the corners of the square 5 Place the slrde onto a heated metal washer, with the strips of Parafilm over the metal. Leave untrl the Parafilm melts m place onto the shde, approx 15 s 6 Prpet 200 pL of F 12 culture medium from the Petri dish into the area deltmrted by the Parafilm on the shde 7 Remove one cover shp culture from its 35mm Petn dash Place one edge of the cover slip, explant stde down, on an outside edge of the Parafilm Allow the rest of the cover slip to fall gently onto the hqurd, makmg sure that no au bubbles are trapped. Thus step should be performed smoothly, but rapidly, to ensure that the culture does not dry out 8. Blot excess lrquld from the mtcroscope shde outside the Parafilm boundaries, and allow to evaporate dry before contmumg to the next step. Briefly holdmg the slide up to a 60 W mcandescent light bulb ~111speed the drying, but care must be taken not to overheat the culture 9 Brush melted Valap across one edge of the cover slip m a single smooth stroke Repeat on the opposite side. Rest the slide on a cool surface, such as the surface of the lab bench, to help set the Valap 10 Very lightly press the cover shp with a blunt ObJect and blot any hqurd that escapes 11 Brush Valap onto the other two edges 12 Repeat step 10 and repair any leaks with Valap, until the cover slip IS completely sealed to the mrcroscope slide 13 Using a Q-up, clean the outside of the cover slip wnh a 1 1 solutron of Wmdex and water, then wipe wrth ethanol and allow to evaporate dry Repeat, rf necessary, to remove any streaks from the surface Do the same for the nucroscope slide on the opposite side

3 3 2 Video Microscopy 1 Place the sealed culture chamber on the incubated mtcroscope stage of an inverted mrcroscope maintained at 37”C, and bring mto focus, usmg a x63 or x 100 011 immerston objective for high-resolutron work, or other objective of choice


491

2 Allow the culture to eqmltbrate with its environment for at least 10 mm before recording microscopy data 3 Connect the vtdeo camera on the microscope to both a time-lapse VCR and to the video input of a personal computer equipped wtth frame-grabbmg capabrlmes (see Note 16) 4 Record mtgratton behavior at desired time intervals of desired duration (see Note 17)

3.3.3. Data Analysis It 1smuch easier to obtain visual data than to analyze it. Although sophtsticated techniques for analyzing dynamic neuronal migration characteristics are bemg developed (121, they are not generally available. However, some basic measurements can be made with readily available software. Using Image analysis software, such as NZH Image, open each digital image m sequence and make the desired measurements. Several examples using NIH image include: 1 Neurrte length. Using the ruler tool, measure the distance from the cell body to the neurtte tip. 2 Neurrte trajectory. Record the coordmates of the same neurrte tip through a serves of images 3 Growth-cone size and shape Using the polygon tool, outline the growth-cone, and record area and shape characterrsttcs provided by the software 4. Notes Alternatively, the less completely defined, but much simpler, serum-contammg formulatron may be used. Serum-containing F12 culture medmm F12 nutrtent mixture (146 mg/L L-glutamme, 1176 mg/L sodium bicarbonate), pemctllmstreptomycm 5000 pg/mL-5000 U/mL, fetal bovine serum lo%, chtck serum 10% (all components from Gtbco-BRL), 7s NGF 50 ng/mL (see Subheading 2.1.2., step lh , Collaboratrve Biomedical) Incluston of the sodmm bicarbonate provides a medium suitable for cell culture m a 5% CO, environment Subsmutton of 4 766 g/L HEPES for the sodmm bicarbonate yields a medium appropriate for incubation m a normal an envrronment Since the transferrm m the ITS Premix is derived from a human source, contact with It, or any solutions containing it, should be prevented by wearing gloves and protective clothing Thaw piturtarres m drstrlled water. They will comprise the reddish-brown and yellow tissue Caution: To avoid the risk of electrocutton, place a plastic bag over the bottom of the blender pitcher during the rests on Ice For efficient dralysrs, divide the supernatant into four batches of 20-25 mL each, and simultaneously dtalyze each batch against 1 L of the PBS or F12 medium Cover slrp spectficattons are dtctated by the followmg constderattons a Round cover slips tend to pattern better than square cover slips m Subheading 3.2.3

Buettner and Tai

8

9.

10

11

12

13 14

15

b. The dtameter of the cover slrps should be slightly less than the width of the mtcroscope slrde used m Subheading 33.1 c The thtckness of No 1 cover slips (~150 pm) provides the appropriate working distance for high-resolutton 011 nnmersron objectives m Subheading 3.3.2., although this must be determined for the specific obJecttve m use (II) This mformatron can typtcally be found m the mtcroscope manual, or obtamed from the technical representative for a gtven mtcroscope d. Fisher cover shps tend to produce the most uniform substrates m our own trials (unpublished observattons) Eggs are typrcally used at embryomc days 6-10 for neuronal mtgratton studies This represents a stgmficant window m the 21-d development of the chtck embryo, the qualttattve aspects of the dtssectton procedure change significantly with each day of development, although the specifics remam the same At day E6, features will be small and tissues very soft At day ElO, features ~111 be much larger and tissues somewhat resistant Depending on the precise locatton of the separatton, some DRGs may be recovered from the matching srte on the body. There are 8 DRGs per leg, although less than 100% recovery IS typtcal The DRGs look ltke tmy balloons attached by mdtvtdual strmgs to the leg. Another set of ganglia, the sympathetic chain ganglia, be closer to the spme, and can be dtstmgutshed from the DRGs by the slightly smaller size of the sympathetic chain gangha and then arrangement m a cham or series, rather than on mdtvtdual stalks Adding all of the medium at the time of platmg results m convectron currents that tend to move the DRG off the cover slip before rt becomes attached This IS not a problem rf the substrate comprises the total area available for attachment Incubation for 6-8 h provrdes the opportunity for neurttes to mrttate and extend past the dense portton of the neurtte halo, so that mdtvtdual growth cone mtgratton can be observed This procedure IS typically performed m a mtcrofabrtcatton factltty, m which a spin coater, contact printer, oven, fume hood, and adequate venttlatron are standard Lom et al (5) descrtbe a way to perform this procedure m the laboratory, although this requires the construction of some nonstandard apparatus Ltmtted commercral services are also avatlable (e g , Bhatla et al., ref. 7) Ftltermg the HMDS and photorestst IS important for preventing contammatton of the trssue-culture and for removing partrculates Resting the cover slip on a chuck m the spin coater that 1ssltghtly smaller than the cover shp (e.g , 16-18-mm dtameter) ~111help prevent transfer of photorestst to the undersrde of the cover slip Exposure time may vary, dependmg on the configuratton and condttton of the equipment. Cover slips should be developed mdtvtdually or m a holder that maintains space between them, to ensure good contact between the developer and the surface Swn-img the cover slips m the developer also ards m this respect. It may be posstble to extend the life of the photoresist pattern by storing m a desiccator at 4°C.


493

16 Stmultaneously recordmg the data on videotape provtdes a convement record of the dynamtc behavtor for later demonstratton purposes, and does not degrade the quality of the dtgttal images saved to the computer. The alternattve IS to record only dtgttal images online, and then to output those from the computer to the VCR at a later ttme 17 The ttme-lapse interval and the duratton of recording are determined by the specific behavior of interest, and frequency of data required to observe stgnificant events Typxal chotces for time-lapse recording of neuronal migration range from l/30 to l/300 of real time, or 1 frame every l-10 s, as required to provide a smooth video Digital Images, because of thetr high storage requirements (cl Mb/Image), are often recorded at more conservative intervals. Very dynamic behavior may require one image every S-10 s Other behavior may be recorded much less frequently The duratton of recording may be minutes or hours

References 1 Hoch, H C., Jelinskt, L W, and Cratghead, H C , eds (1996) Nanofabrrcatzon and Blosystems Integratmg Materials Science, Engmeermg, and Biology Cambridge Umverstty Press, New York, NY 2 Alberts, B., Bray, D., Lewts, J., Raff, M., Roberts, K., and Watson, J D (1994) Molecular Biology of the Cell, 3rd ed., Garland, New York, NY 3 Clark, P, Connolly, P, Curtis, A S G , Dow, J. A. T, and Wilkmson, C. D. W (199 1) Cell guidance by ultrafine topography m vitro .I Cell Scz 99,73-77 4 Clark, P., Brttland, S , and Connolly, P. (1993) Growth cone gutdance and neuron morphology on mtcropattemed lammm surfaces J Cell Scz 105,203-2 12 5 Lom, B , Healy, K H , and Hockberger, P. E. (1993) A versatile technique for patterning btomolecules onto glass coverslrps J Neuroscl Methods 50, 385-397 6 Klemfeld, D., Kahler, K. H., and Hockberger, P E. (1988) Controlled outgrowth of dtssocrated neurons on patterned substrates J Neurow 8,409%-4 120 7 Bhatta, S N , Toner, M., Tompkms, R G , andYarmush, M. L (1994) Selective adhesion of hepatocytes on mtcropattemed surfaces Ann NYAcad Scl 745,187-209 8 Tsao, M C , Walthall, B J , and Ham, R. G. (1982) Clonal growth of normal human epidermal keratmocytes in a defined medium J Cell Physzol 110,2 19-229 9 Baud, J L and Raper, J A (1995) A serme protemase mvolved in contact mediated repulsion of retinal growth cones by DRG neuntes J Neuroscl 15,6605-66 18 10 Bhatta, S. N , Toner, M , and Yarn-rush, M. L Personal communtcatton 11 Smtth, R F (1990) A4zcroscopy and Photomlcroscopy A Workzng Manual, CRC, Boca Raton, FL. 12 Gwydn, S S , Buettner, H M , and Dunn, S. (1994) Non-rtgtd motton analysts and feature labelmg of the growth cone Proc IEEE Workshop on Bzomedzcal Image Analysu, Seattle, WA

37 Estimating Number and Volume of Islets Transplanted Within a Planar lmmunobarrier Diffusion Chamber Kazuhisa Suzuki, Clark K. Colton, Susan Bonner-Weir, Jennifer Hollister, and Gordon C. Weir 1. Introduction

Cell therapy involving the transplantation of cells or tissues with specific differentiated functions has potential in the treatment of human disease. However, the need for lmmunosuppresslve drugs may lead to a variety of serious side effects. One approach to mimmlzmg or eliminating systemic immunosuppresslon is lmmunoisolatlon, m which the transplanted tissue 1senclosed in a semipermeable barrier or membrane in order to protect it from immune rejectlon, thereby creating what has been termed an lmplantable blohybrld artificial organ. Devices of this type are under study for the treatment of a wide variety of diseases, for example, diabetes In implanted devices, the implanted cells or tissues are separated from the body by an lmmunobarrler membrane. Cells can either be encapsulated at a high tissue-like density or dispersed In an extracellular matrix, such as agar, algmate, or chitosan. Ideally, the membrane prevents components of the cellular (and possibly humoral) immune responses from entering mto the vlcmlty of the transplanted tissue, but permits passage of the secreted product, for example, insulin. At the same time, the transport properties of the graft, membrane, and surrounding tissue must permit sufficient accessof nutrients, such as glucose and oxygen, and the removal of secreted metabolic waste products, such as lactic acid, carbon dioxide, and hydrogen ions. Transplanted cells must be supplied with nutrients by diffusion from the nearest blood supply, through surrounding host tissue, the immunolsolatlon membrane, and the graft tissue From

Mefhods m Molecular E&ted by J R Morgan

Medm?e, Voi 18 Jtssue fngmeenng Methods and and M L Yarmush 0 Humana Press Inc , Totowa,

497

Protocols NJ

498

Suzuki et al.

itself Oxygen supply is thought to be the hmltmg factor m determmmg how much tissue can be supported m a specific configuratlon. Transplanted msulm-secretmg pancreatic islets contained wlthm lmmunobarrier membrane devices may be partially protected from allo- or xenograft reJection and from autoimmune attack. Although there are some reports about the function and morphological appearance of islets contained wlthm such devices (1-111, httle 1s known about such parameters as chamber volume, islet volume, islet number, islet necrosis, and fibrosis under these condltlons. In order to better characterize the potential of this approach, it 1s important to quantify islet number and volume wlthm membrane devices. In this chapter, we describe a recently developed method to measure chamber volume, Islet number, and Islet volume (12) This method can be used to determine how much islet tissue can be supported by a planar diffusion device In previous studies with planar devices, sections have been taken perpendlcularly to the plane of the membrane. This generates extraordmarlly large numbers of sections and 1s lmpractlcal for analyzing chamber and islet parameters for an entire device. In this method, we use sectlons taken parallel to the nomlnal plane of the membrane, and a fraction of these are analyzed to estimate device parameters. This IS the first rigorous method for determining how much Islet volume must be contained wlthm an implanted device to normalize glucose levels m an animal model of diabetes (13).

2. Materials 2.1. Animals 1 2 3 4

Inbred male B6AFl mice, 25-30 g (Tacotuc, Germantown, NY) Streptozotocm, 180 mg/kg, tp (Sigma, St LOWS, MO) Portable glucometer (One Touch II, Lifescan, Mllpltas, CA) Methoxyflurane (Metofane@, Pltman-Moore, Mundelem, IL), used like ether Inhalant wrth gauze, until animal IS asleep 5 Veterinary tissue adhesive (Vetbond TM, 3MAmmal Care, St Paul, MN)

2.2. Islet /so/a tion 1 Collagenase solution, 1 mg/mL, 2 ml/mouse IndIanapolls, IN). 2 Hlstopaque, used pure (Hlstopague@- 1077, Sigma)

(Boehnnger

Mannhelm,

2.3. Islet Encapsulation 1 Sodium algmate (Protan, Copenhagen, Denmark) solution, 16 g/L m HEPESsaline buffer 2. BaC12 solution. 2 38 g/L HEPES, 7 01 g/L NaCl, 4 89 g/L Ba Cl2 2H20 3 HEPES-saline buffer (N - [2 hydroxyethyl] plperazme-N’-[2-ethane sulfomc acid]), 1 19 g/L, NaCl 8 48 g/L

Estimating Islet Number and Volume

433

2.4. Planar Diffusion Chamber Planar diffusion chamber (Baxter Healthcare, Round Lake, IL)* Tissue is placed mto a gap between two membrane laminates separated by a slhcone rubber washer (Id 0.70 cm, cross-sectional thickness 125 m), which seals the gap and defines its thickness (14,15). Each lammated membrane consists of three layers, an inner membrane of polytetrafluoroethylene (PTFE) with a hydrophilic surface (Blopore TM,Mlllipore, Bedford, MA) with a nommal pore size of 0.45 pm; a PTFE membrane (W L. Gore, Elkton, MD) with a nominal pore size of 5 pm and a thickness of about 15 m; this, m turn, is laminated to a highly open outer meshwork about 125-p thick of polyester fibers to provide support. The inner membrane excludes cells, but allows some penetration of all dissolved molecules. The second layer, with or without the third layer, promotes neovascularlzatlon at the interface with host tissue (16). The membrane-tissue sandwich 1sheld m place between a titanium rmg housing and a titanium sealing rmg, which IS pressed mto the housing by a hand press. When compressed by the tltamum rmgs, the slhcone rubber washer dimensions change The id becomes about 0 66 cm and the washer (gap) thickness about 100 pm (15). In an idealized device, m which the membranes maintain this geometry uniformly, the calculated contained volume would be 3.5 pL. 2.5. Sample Preparation 1 Tissue cassette(HlstoPrep@, Fisher Scientific, Pittsburgh, PA) 2. Bourn’s solution. 850 mL plcrlc acid, 50 mL glacial acetic acid, and 150 mL 37% (v/v) formaldehyde m ddHzO 3 Wann liquid agar (Bacto-Agar, Dlfco, Detroit, MI). 7% (w/w) m ddHzO 4 BIOPSY sponges (Fisher) 5 10% buffered formalm (Fisher)

2.6. Quantification

of Device Parameters

I Photographs (x 10 magmficatlon, Kodak Technical Pan Film TP 135, ASA 50) 2 Electronic plammeter (StgmaScan TM, Jandel Scientific, San Rafael, CA)

3. Methods 3.1. Animals, Islet Isolation, and Implantation Mice are used as donors and recipients, although other animals can be used. Recipients are made diabetic with streptozotocm at least 1 wk prior to transplantation. Blood glucose levels are measured with a portable glucometer before transplantation. Mice are only used as recipients If their blood glucose levels are over 300 mg/dL, m order to test the efficacy of the implant m normahzmg blood glucose concentration. Islets are isolated using a modlficatlon of the method of Gotoh et al (17), m which the pancreatic duct 1sdistended

500

Suzuh~ et al.

with a collagenase-containing solution. After purtfication on a Histopaque gradient, Islets with diameters between 75 and 250 pm are handpicked and encapsulated wtthm algmate microspheres. Islets are suspended m a 0.3% (w/w) alginate solutton, and BaCl* is used to crosslmk the algmate to form the mtcrospheres (18,19). Encapsulation is helpful, because it prevents clumping of islets withm the capsules. The method of analysis descrtbed in this chapter 1sapphcable when islets are encapsulated by any method, unencapsulated, or suspended in a matrix. Approximately 500 encapsulated islets are loaded mto a diffusion chamber. Recipient mice are anesthetized with the inhalant anesthesia methoxyflurane, and, followmg a midline mcision, the membrane device is placed m the eptdidymal fat pad and fixed m place with veterinary tissue adhesive. Two wk after implantation, the device is removed from the anesthettzed recipient and prepared for morphologtcal evaluation. 3.2. Sample Preparation The tmplanted device is removed from the recipient, placed m a tissue cassette, and fixed m Boum’s solutron overnight. The device is then washed m water and placed m 10% buffered formalm for a further 24 h Then the silicone spacer and titamum rings are gently removed. The membranes are held together with forceps and pushed back into the outer titamum rmg A drop of warm liquid agar 1splaced on a microscope slide, and the membranes with the metal rmg are placed upside down mto this pool of agar. Another drop of liquid agar is placed on the top membrane, and, on top of this, another microscope slide is placed to sandwich the membranes between the glass slides. Once the agar hardens at room temperature, the rmg is removed, along with excess agar. The membranes, embedded m agar, are then placed between biopsy sponges, returned to the tissue cassette, and agam placed into 10% buffered formalm, before being processed for paraffin embeddmg Serial sections (7 pm in thickness) are made m the plane of the membranes (see Note 1). The total number of cut sections is counted to determme the thtckness of the device. The sections are then stained with hematoxylin. 3.3. Quantification of Device Parameters In order to determine the total volume of the device chamber, Vc, the total volume of islets, Vt, and the total number of islets, M, photographs are taken of every evaluated sectton, using a light microscope (e g., Olympus BH-2). Several photographs are required of each sectton to mcorporate the entire area of that section. Prints (5 x 7 in.) are made mto montages of each section. The device chamber area, Ac, which is the space surrounded by the membrane (s), of each evaluated sectton, and the profile area, At, of each individual islet, are measured using an electrontc planimeter.


501

The device is assumed to be a shallow cylinder that is analyzed layer by layer. The sections are divided from top to bottom into N sequential zones. The top of each zone consists of an evaluated section, followed by a number of nonevaluated sections beneath it. The evaluated sectrons are selected for quality of section and stam at mtervals of no more than mne sections (see Note 2). The data obtained from the evaluated section at the top of each zone are assumed to be representative of all of the nonevaluated sections m that zone, and are used to estimate the chamber volume, islet volume, and islet number of that zone. Because each section 1s7-pm thick, the thickness of each zone 1s determined by the number of sections m that zone multiplied by 7-pm thickness. 3 3 1. Volume of Chamber The volume of the device chamber, Vc,, for a given zone, ~1,is calculated as the chamber area of the evaluated section, AC,, multiplied by the thickness of the zone, h,, (1) VC”= AC”h” The total volume of the device chamber, V& for the entire device can be then obtained by summation over all of the N mdividual zones (see Note 3):

vc=n=I 5 vc,

(2)

3.3.2. Volume of Islets The islet volume m the device can be obtained from similar equations, using the islet areas instead of device chamber areas. It IS assumed that the islet can be represented as a cylinder in each zone. The islet volume m zone IZ is calculated as the total islet profile area in the evaluated section, AI,, multiplied by the thickness of the zone, h, (3) VI” =&h, The total islet volume m the device, V,, is obtained by summatton over the N mdividual zones (see Note 4):

v,=?l=I 5 v,”

(4)

3.3.3. Number of Islets For transplantation, islets with a diameter between 75 and 250 pm are handpicked using a cahbrated eyepiece micrometer on a dissecting microscope. Becauseof this size,a portion of each islet ISfound on many more than one 7-pm

502

Suzuki et al.

section. The number of sectlons in which each islet IS observed depends on the size of the islet and the thickness between evaluated sections. When an islet profile m an evaluated section IS measured by planimetry, neither the size of the entn-e islet nor the plane of section IS known. Furthermore, portions of many islets are observed m each evaluated section. Different parts of an islet are included in different adjacent zones, the details of which depend on the diameter of the islet and the thickness of the zones To estimate the number of Islets, it IS assumed that the Islet profile area observed IS actually a section through a cylmdrlcal islet of equal diameter and height. The portlon of the mth islet contained within zone n, qnm, is the ratio of the volume of that Islet within zone n to the volume of the entire islet, which IS assumed to be a cylinder having a diameter (and height) equal to the equivalent diameter observed m the evaluated sectlon: 9nm

= Alnmhn

/ AInmDnm

= 4

/ &m

(5)

where Al,,,,,IS the profile area and D,, IS the equivalent diameter of the mth islet m zone ~1.Since qnm

A Inm-- nD2,,,, / 4

(6)

qnm = h, / 2(AIn, / TC)“~

(7)

can also be written as

The total cumulative portions, Q”, of all islets contained m zone n 1sobtained by summing mdlvldual islet portions over all of the 44, islets observed m the evaluated section of zone n:

The total number of islets m the device IS obtained by summing the portions of islets m all N zones (see Notes 5 and 6). N

4. Notes 1 To optimize the efflclency of this method, it IS desirable to make sectlons that are as parallel to the orlgmal flat surface of the membrane as possible, even though the device, when embedded m paraffin, IS usually not completely flat, but slightly askew. Some dlfflcultles can be experienced with the sectlomng Sometimes the edge of a membrane IS caught by the kmfe and pops out, making further analysis of that block impossible Even though the method of enrobing the membranes m

Estimating Islet Number and Volume agar and embeddmg the sample m paraffin usually works well, the membranes can still slip apart Obviously, as the number of zones and sections evaluated is Increased, the more accurate IS the estimation of device parameters, but the quanttlication process ~111 take much longer, and the method becomes impractical at some point. By embeddmg and sectioning the device as flat as possible, the number of serial sections needed to go through all of the device volume is mnnmized. To estimate the optimal number of sections for quantification (12), the device and islet volumes of one device have been measured by evaluating every other section throughout the serial sections (total of 13 evaluated sections and 13 zones). These results were compared with device and islet volumes obtained if only every fourth, sixth, eighth, or tenth section was evaluated There was no change m values until samples were taken only every 10 sections If every other section through only half of the device were used, the values were highly divergent Therefore, it IS recommended that every fourth to ninth section (each of which is selected because it IS nicely sectioned and stained) be evaluated through the whole device In an illustrative example (22), the total calculated volume of the chamber was 1 78 uL, roughly half of the 3 5 p.L calculated for an idealized device with planar membranes This difference IS not surprising, because it is assumed that the two membranes remam flat and separated from each other by about 100 pm in the idealized device Clearly, the membranes have some flexibility, and some bulgmg must occur when Islets with diameters of 75-250 lrrr~are placed mto the chamber With this flexibility, the height of the internal space can be expected to increase to 300 pm or more In the areas between the islets, the membranes approach each other, which leads to a loss of volume of the internal space. This lattei reduction more than counterbalances the increase from the bulging around the islets Although the islets handpicked for transplantation m the device are spherical or ovoid, considerable variation in shape is observed. To estimate islet volume and number, it is assumed that a circular islet profile is obtained in each zone as a shce through a cylmdrical islet m that zone. In determmmg the chamber and islet volume of each zone, this assumption provides a reasonably close estimate, because each zone IS relatively thin. If each zone is thm and all secttons are evaluated for estimating the volume of device and islets, the error would be mathematically negligible Therefore, this method of calculation should be suitable for an accurate estimate of chamber and islet volume Calculation of islet number is more complicated than total chamber or islet volume, because the volume of an islet within a zone is considered to be a portion of an Islet, that is, the volume within a zone divided by the total volume of the cylindrtcal solid. This calculational procedure is not based on theoretical considerations, but rather is empirically derived, and used on the basis that it works reasonably well To examme its validity, calculations have been carried out (22) by assuming that sections (each representing one zone) are cut through a spherical islet, and the calculations described here are applied to each section. The

504

Suzukr et al

result is that the portions sum to 1 38 islets Thus, this method gives an estimate of Islet number that 1s high by about 38% 6. This method can be used to estimate not only the volume of the device, volume of islets, and islet number, but also the volume of fibrosis, necrosis, and other possible condtttons that may decrease the function of a loaded device (13)

References 1 Altman, J J , Penforms, A , Botllot, I, and Malettt, M (1988) Btoarttfictal pancreas m autonnmune nonobese diabetic mtce. ASAIO Transact 34,247-249. 2 Brunette, P , Basta, G , Faloerm, A , Calcmaro, F , Petropaolo, M , and Calafiore, R (1991) Immunoprotectton of pancreatic islet grafts within artitictal microcapsules Int J ArtEf Organs 14,78!+-791. 3 Colton, C K and Avgoustuuatos, E S (199 1) Bloengmeermg m development of the hybrid arttfictal pancreas. ASMEJ. Blomech Eng 113, 152-170 4 Lacy, P E , Hegre, 0 D , Gerasimtdi-Vazeou, A , Gentile, F T , and Dianne, K E (199 1) Maintenance of normoglycemta m diabetic mice by SC xenografts of encapsulated islets Sczence 254, 1782-1784 5 Brauker, J., Martmson, L A , Loudovaris, T., Hill, R S , Carr-Brendel, V, Hodgson, R., et al. (1992) Immunotsolatron with large pore membranes. allografts are protected under condmons that result m destructton of xenografts. Cell Transplantatzon 1, 164

6. Hill, R S , Young, S K , Jacobs, S. A., Martmson, L A , and Johnson, R. C. (1992) Membrane encapsulated islets implanted m eptdidymal fat pads correct diabetes m rats Cell Transplantatron 1, 168 7. Lanza, R P, Borland, K M , Staruk, J E., Appel, M C , Solomon, B A , and Chick, W L (1992) Transplantatton of encapsulated canme islets mto spontaneously dtabettc BBrWor rats without immunosuppresston Endocrznology 131,637-642 8 Lanza, R. P, Borland, K M , Lodge, P, Carretta, M , Sullivan, S J , Muller, T E , et al (1992) Treatment of severely diabetic pancreatectomized dogs using a dtffusion-based hybrid pancreas Diabetes 41,886-888 9. Lanza, R P, Beyer, A M , Staruk, J. E., and Chick, W L (1993) Btohybrld arttficoal pancreas Transplantatzon 56, 1067-1072 10 Scharp, D. W, Swanson, C. J , Olack, B J , Latta, P P, Hegre, 0. D , Doherty, E J , et al (1994) Protection of encapsulated human islets implanted without immunosupression m patients with type I or type II diabetes and m nondiabetic control SubJects Dzabetes 43, 1167-l 170. 11 Colton, C K (1995) Implantable biohybrtd artificial organs Cell Transplantatzon 4,415-436 12 Suzuki, K , Bonner-Weir, S , Holhster, J , and Weir, G C. (1996) A method for esttmatmg number and mass of islets transplanted wtthm a membrane device Cell Transplantation

5,6 13-625

13 Suzuki, K , Bonner-Weir, S., Holhster, J , Colton, C K , and Weir, G C (1997) Number and volume of islets transplanted m mnnunobarrter devices Cell Transplantation, m press.


505

14 Dudek, R W , Lawrence, I. E , Jr., Hill, R. S., Johnson, R C (1991) Induction of

islet cytodifferentiation

by fetal mesenchyme in adult pancreatic ductal eptthelmm

Dzabetes 40, 1041-1048.

15 Brauker, J , Martmson, L A , Young, S K., and Johnson, R C (1996) Local inflammatory response around dtffuston chambers contannng xenografis. Transplantatl-on 61, 167 l-l 677 16 Brauker, J. H , Carr-Brendel, V E , Martmson, L. A., Crudele, J., Johnston, W D , and Johnson, R C. (1995) Neovascularizatron of synthetic membranes directed by membrane microarchitecture. J. homed. Mater Res 29, 1517-1524 17 Gotoh, M , Makt, T , Ktyotzumt, T , Satomr, S., and Monaco, A P. (1985) An tmproved method for isolatton of mouse pancreatic islets. Transplantation 40,437,438. 18 Zekorn, T, Entenmann, H., Horcher, A., Schnettler, R , Klock, G , Bretzel, R. G , Ztmmermann, U , and Federlm, K (1992) Banum-algmate beads for unmunoisolated transplantatron of islets of langerhans. Transplantation Proc 24,937-939 19 Schrezenmeir, J , Hermg, B J., Gero, L , Wiegnad-Dressler, J , Solhdju, M , Velten, F., et al (1993) Long-term functton of porcine islets and smgle cells embedded m barmm-algmate matrrx Harm Metab Res 25,204-209

38 Quantitative Measurement Under Flow Conditions

of Cell-Cell

Adhesion

Carroll L. Ramos and Michael 6. Lawrence 1. Introduction Cell-cell and cell-matrix contacts play an important role m immunological processes and tissue orgamzatlon. It has been shown that m many cases, cell adhesion IS mediated by specialized adhesion receptors that are typically anchored m the cell membrane with mol wt ranging from 50 to 200 kDa. Bmdmg of adhesion receptors anchored in the apposing cell membranes with their counterligands appears to be critical for many forms of cell-cell commumcation (1) Quantitative measurement of ceil-cell and cell-substratum adhesion can be used to compare the contrlbutlons of distinct receptors to an adhesive mteraction Fol instance, by measuring specific cell adhesion quantitatively, the modulation of receptor avidity or afflmty in response to biological stimuli can be tracked (2) Other apphcatlons of adhesion force assays are m the study of adhesion receptor ligation on cell slgnalmg processes and in the charactenzatlon of the strength of cell-blomatenals interactions. Table 1 contains a list of some of the commonly used methods for assessing cell-cell adhesion, with selected references. In this protocol, we will focus on the use of parallel-plate flow chambers, highlighting the strengths and limitations of the assay relative to other methods m the Notes section Assessment of adhesion requn-es that a mechanical force be applied to the cells to separate them from other cells or the substrate. In almost all forms of cell-cell adhesion assays,fluid shear, or washing, 1sused to distmgulsh bound from unbound cells. Parallel-plate flow chambers generate controlled, highly reproducible washing conditions (detachment forces), allow direct vlsuahzaFrom

Methods EdIted by


Medicine, Vol 18 Tmue Engmeenng Methods and frofocols and M L Yarmush 0 Humana Press Inc , Tofowa, NJ

507

508 Table 1 Cell-Cell

Ramos and Lawrence and Cell-Substrate

Adhesion

Assays

Assay

Principle

Stampler-Woodruff

Leukocyte suspension layered onto rotatmg trssue specimen (nonstattc) Leukocyte suspension allowed to settle m wells of an Ebsa plate coated wtth adhesive substrate (static) Cells mjected mto sealed chamber, allowed to settle, then chamber Inverted (static) Cells m sealed ELISA plates are centrtfuged at varymg speeds Suspensron of cells exposed to umform shear

Plate-bmdmg

assay

Smtth-Hollers chamber Centrrfugatton assay Cone and plate viscomter Couette viscometer Capillary tube

Radtal flow chamber Parallel-plate chamber

flow

Suspension of cells exposed to uniform shear Suspension of cells exposed to varymg shear forces, exposure of anchored cells to defined shear force, some designs employ a recirculatmg loop Anchored cells on adhestve substrate experience flurd shear force based on position from center of disk Anchored cells on adhesive substrate experience defined shear stress

Ref 13 14

15,16 17,18 9 19 20,21

22

23-25

interactions, have relatrvely low cost, and can be adapted to study a number of different cell and receptor types.

tton of cell-substrate

2. Materials 1 Flow chamber apparatus a Custom rectangular polycarbonate (Lexan) flow-chamber block machined from spectficatrons (two ports on each long side, vacuum port on one end, two flow slots and one vacuum hole on face; Godwin Machme Works [Houston, TX], spectficattons obtainable from authors), commerctal flow chambers are available from sources such as Glycotech [Rockvrlle, MD] or CytoDyne [San Diego, CA]. 1 Silicone rubber sheetmg for hand-cut gaskets that are applied over the face of the flow chamber (4 x 8 x 0.020 m.; Technical Products [Decatur, GA], #500-5, gasket specifications obtainable from authors) 2 Tygon tubing (1 6 mm id; Bto-Rad [Richmond, CA], #73 l-82 15). 3. Intramedic Luer stub adapters (15 gage; Becton Dtckinson [Sparks, MD], #427560).

Ceb-Cell Adhesion Under Flow Conditions

9

10 11. 12

509

4. 60-, IO-, and l-cc sterile, disposable syringes (Becton Dickinson [Franklin Lakes, NJ]) 5 Standard three-way stopcocks with male luer slip adapter. 6 Latex tubing (5132 x 3/64 m.) cut into one-half-m. lengths used as connectors 7 Vacuum grease (silicone) (Dow-Corning [Midland, MI]) Inverted phase-contrast microscope (e.g , Nikon Diaphot 300 [Melville, NY]), with a long working-distance condenser and x10, or x20 obJectives Video camera (e.g., Vicon [Melville, NY], WC24 10-24) Video recorder (consumer models are adequate), video monitor, and video timer (e g , Fora #VTG-33) Syringe pump allowing mfusion and withdrawal (Harvard Apparatus [South Natick, MA], #55-2226) Vacuum apparatus (side-arm flask connected to laboratory vacuum). Temperature-controlled water bath with 15 and 50 mL comcal tube racks. Polystyrene, bactertological (nontissue-culture) Petri dishes (150 x 15 mm, Falcon 1058, 100 x 15 mm, Falcon 1029; Becton Dickmson [Lincoln Park, NJ]) for cell or protein immobtlization Glass slides (70 x 40 mm; Commg [Commg, NY]), for cell or protem tmmobihzation, custom sizes at thicknesses of 1 or 2 can be made to order by Corning Glassworks Dtsposable, sterile, polypropylene conical tubes (15 mL; 50 mL, Falcon, Becton Dickinson, #2097 and #2098) Pyrex 10 x 15 x 2 in. rectangular dish (Corning). Buffers and assay media Hanks’ balanced salt solution (HBSS; Gibco-BRL [Grand Island, NY], #24020-059) supplemented with 20 nnt4 HEPES, pH 7 4, and 1% human serum albumm (HSA) (Sigma [St Louis, MO], #A1653). suitable for leukocytes isolated from blood HBSS without calcium and magnesium can be obtamed (Gibco-BRL #14170021) and supplemented with l-2 mM CaCl* or MgCl*, depending on the nature of the adhesive interaction studied (e.g , addition of calcium alone for assaying selectm-dependent rolbng mteractions of leukocytes may be useful to avoid mtegrm-mediated mteracttons that are magnesium-dependent). RPMI- 1640 (Gtbco-BRL, #11875-093) supplemented with 20 mM HEPES, pH 7 4, and 1% HSA suitable for cultured cell lmes of myeloid or lymphoid origin, and some transfected cell lines Ml99 (Gibco-BRL, #11150-026) supplemented with 20 mMHEPES, pH 7 4, and 1% HSA suitable for assays using cultured endothelmm as an adhesive substrate. 1X phosphate-buffered salme (PBS), pH 7.4. 10X PBS. 80 g NaCI, 2 g KCI, 11 5 g Na2HP04.7H20, 2 g KHtP04 added to sufficient deionized, distilled water to make 1 L, adjust to pH 7 4 at room temperature 50 mM Tris, pH 9 5, with 0.02% azide (adsorption buffer)

510

Ramos and Lawrence h 50 mM Tris, pH 9.5, with 0.02% aztde and 3% HSA (blocking buffer A), solution should be made fresh before use 1 1X PBS, pH 7 4, with 0 02% azide and 0 1% Tween-20 (Bio-Rad [Hercules, CA], #17&653 1) (blocking buffer B)

3. Methods 3.1. Assembly of Custom Polycarbonate

flow-Chamber

Apparatus

1, Using a disposable pipet tip, apply a thm film of vacuum grease to the face of the flow chamber block, being careful to avoid covering the slots or vacuum hole with grease. Reapply as necessary 2 Place the cut sthcone rubber gasket over the vacuum grease-coated face of the flow chamber Position the gasket using a pipet tip to gently press downward on the gasket, so that the vacuum grease is evenly spread below Excess grease that extends over the flow area can be removed by applying a small amount of liqutd detergent and gently washing with gloved fingers under runnmg water Durmg the first several usages, the gasket ~111 compress and stretch a httle, so it IS best to make gap-thickness measurements several times Once seated, several experiments may be performed before it is necessary to add more grease Do not rub harshly or dry the flow block with wipes, since the polycarbonate is easily scratched The gasket may be washed with mild detergent and water Ethanol will cause the gasket to stretch and thus should not be used for rmsmg 3 Add 200 mL 10X PBS to 1800 mL deionized, dtsttlled water, and dtspense this solution (2 L) mto the Pyrex rectangular dish Lower the flow-chamber block, with the gasket side up, mto the PBS Using a disposable Pasteur pipet filled with PBS, dtspense and fill several times over the flow slots, to remove air bubbles from the side-port channels. 4 Cut four lengths of Tygon tubing (two approx 12-15 m. and two l/4 m ) Using pliers, remove the blunt needles from two Intramedic Luer stub adapters Into one end of each piece of 12-l 5 m Tygon tubing, insert a blunt needle Over the needle and Tygon tubing, place a l/2 m piece of latex tubing, leavmg a small amount of needle exposed Insert the exposed needle mto a l/4 m length of Tygon tubing, and place this end of the tubmg system mto the Luer ends of the threeway stopcock Secure the fitting by pulling a portion of the latex tubing over the Luer outlet on the stopcock Attach Intramedic luer stub adapters to a 60-cc and lo-cc syringe, and insert the syringes into the open (opposite) ends of the two Tygon tubing preparations 5 Fill the tubing systems with PBS by drawing 5-10 mL PBS mto each syringe while the stopcock end is placed mto a container of buffer Immerse the filled tubing systems mto a PBS-tilled Pyrex dish containing the flow chamber Push the syringe plungers to dispel air from the tubing and stopcocks While immersed, insert the ends of the stopcocks mto each portal on one side of the flow chamber Insert l-cc syringes attached to stopcocks and filled with PBS mto the two por-

Cell-Cell Adhesron Under Flow Condltlons

511

tals opposite of the Inlets and outlet manifolds of the flow chamber. Insure that an bubbles are removed from the tubmg and syringes to facthtate an atrttght system The l-cc syringes can be used to remove an bubbles, or when only small numbers of cells are bemg analyzed, to inject a bolus of cells mto the chamber flow channel In addition, the syringes allow loading of I mL antibody or sttmulus soluttons onto the adhestve substrate m the flow chamber

3.2. Preparation of Adhesive Substrates 3 2.1. lmmunoaffinity Purified Natwe Adhesm

Receptor Proteins

Dilute the adhesion receptor or matrix protein m 50 mMTrts, pH 9.5, wtth 0 02% aztde, to a concentratton of 0 l-10 pg/mL To ensure efficient adsorption, proteins purified from cell membrane preparations that are m nomomc detergents (e.g , octylglucostde) need to be diluted, so that the protein-detergent solutton IS below the critical mtcelle concentratton Cut a 70 # 40-mm slide from a 150 Y 15-mm polystyrene Petri dash, using a surgical scalpel Latex gloves should be used to prevent coating the slides with oils from skm. We have found that nontreated polystyrene (Falcon 1058) is comparable to tissue-culture polystyrene for adsorption of protein and cells Alternatively, a 70 Y 40-mm glass shde may be used Wash the slide with a stream of detomzed, dtstilled water, followed by rmsmg with 100% ethanol Repeat washing two additional times, and place the sltdes upright to an-dry Scrubbing or drymg with tissues (Kimwipes) should be avoided, to prevent scratching of the slide surface Using a laboratory markmg pen, draw a 1O-mm diameter circle on the underside of the slide Mark a small dot m the center of the circle and place the slide markings side down m a 100 Y 15 Petri dash. Dispense 5&100 $ of the diluted protein preparation onto the sltde m the center of the marked circle To prevent evaporation and drying of the edges of the protein solution, place several 100-PL drops of Trrs-aztde or PBS on the slide surface Place the hd on the Petrt dish and incubate the slide for 1 5 h at room temperature, or overnight at 4°C. Blocking of the plate to eliminate nonspecific interactions with plastic or glass* Dispense 500 pL of the freshly prepared blockmg buffer A onto the protein solution on the plate Using a Pasteur ptpet connected to a vacuum source begm aspirating the solutton on the plate while simultaneously dtspensmg 3-5 mL of blockmg buffer A with a ptpeter Caution: The center area of the plate contammg the adsorbed protein should never be allowed to dry or be exposed to air, since thts may destroy the activity of the protem Cover most of the plate with blockmg buffer A Incubate 1 h at room temperature, or overnight at 4°C. The plate can be stored for several days at 4°C Before the expertment, wash the plate m the same manner just described except using 2-3 mL of 0.1% Tween-20 solutton (blocking buffer B), followed by rmsmg with 2-3 mL PBS, pH 7 4.

512

Ramos and Lawrence

3.2.2. Recombinant Adhesion Proteins Recombinant Adhesion Proteins are Frequently fused with the Fc domain of human IgG or IgM, and thus introduce the potential for Fc-receptor interactions. 1. Adsorb the recombinant fusion protein and block the plate as described for purified native proteins. 2. Before the experiment, incubate the plate with 1 mL of 10 pg/mL polyclonal goat antihuman IgG (Fc) F (ab’)2 (Biodesign International [Kennebunk, ME], #W99334G) or human IgA4 (Fc5p) (#W99340G) for 15 min at room temperature.

3.2.3. Monolayers of Vascular Endothelial Cells (3) 1. Characterized normal human umbilical vein endothelial cells (HUVEC) can be obtained commercially from Clonetics [San Diego, CA], or others, as proliferating cells in T-25 flasks, #CC-2617 or cryopreserved cells, #CC25 17. Culture media containing serum, antibiotics, and growth factors can also be obtained from Clonetics, #CC-3124. Alternatively, endothelial cells can be obtained from human umbilical veins, using collagenase digestion (an excellent protocol source: Freshney, R. I. (1994). Culture of specific cell types: endothelium, in Culture of Animal Cells, 3rd ed. Wiley-Liss, NewYork, pp. 333-334). Typically, primary or first passage endothelial cells are preferable for many types of studies because of the enhanced responsiveness to in vitro agonists relative to high-passage-number endothelial cells (p > 4). 2. Under aseptic conditions, collect cells from culture flasks, using 0.05% trypsin and 0.02% EDTA (Gibco-BRL, #25300-047) and wash into complete cell-culture media (with serum, antibiotics, and growth factors) to neutralize the trypsin. Resuspend cells in complete culture media and seed (0.5-l Y lo6 cells/ml) onto sterile (autoclaved) 70 Y 40-mm glass slides coated with tibronectin or gelatin (3) and placed in sterile Petri dishes. Place the slides in a 5% COZ, 37°C incubator. Slides can be used for adhesion assays upon reaching a confluent monolayer in l-3 d, depending on seeding density.

3.2.4. Platelet Monolayers (4) 1. 70 Y 40-mm glass slides should be soaked in 70% nitric acid chromerge and washed thoroughly with deionized, distilled water. Dry the slides with ethanol and allow to air-dry. A number of matrix proteins or peptides can be absorbed to the clean glass surface. Plastic slides will work, as well, for binding proteins such as fibronectin or fibrinogen, to which platelets will adhere. Alternatively, some investigators dip the clean slide twice for 2 min in 4% 3-aminopropyltriethoxysilane (Sigma, #A3648) in acetone (4). Rinse the slides with acetone, followed by deionized, distilled water, and place in sterile Petri dishes. Place the slides into a 37°C incubator overnight to dry. 2. Collect human peripheral venous blood into acid citrate dextrose (ACD) solution (Sigma, #C3821) at a ratio of 1 vol ACD anticoagulant to 8 vol whole blood.

Cell-CeN Adhesion Under Flow Conditions

573

Centrifuge at 600g for 5 min at room temperature and transfer the upper plateletrich plasma layer into a conical tube. Dilute to 2 Y lO*/mL in room temperature PBS, pH 7.4, with 0.3% bovine serum albumin (BSA), fraction V, essentially fatty-acid-free (Sigma, #A6003). 3. Dispense a sufficient volume of platelet suspension to cover the treated slides, and incubate for 30 min at room temperature to allow a platelet monolayer to form. The confluency of the monolayer can be controlled by the platelet seeding density and the addition of thrombin or ADP to promote platelet binding to the substratum.

3.3. Incorporation Apparatus

of Adhesive

Substrates

into the Flow-Chamber

1. Submerse the Petri dish containing the slide with the adhesive substrate in the PBS-containing rectangular Pyrex dish (see Subheading 3.1., step c). If a cellular substrate (i.e., endothelium or platelets) is being tested, then it is important to add Ca2+ (1 mM) and Mg2+ (1 mM) to the buffer to maintain cell-substrate adhesion. Carefully invert the substrate slide while under buffer, and place over (substrate side down) the upper face of the submerged polycarbonate flow apparatus. Lower the slide onto the gasket of the flow chamber, so that the adhesive substrate is in the center of the rectangular flow chamber block. Without applying downward pressure on the slide, connect the tubing from the vacuum source into the end portal of the flow chamber. This is best accomplished by letting the slide rest on the flow chamber, without using your hands to stabilize it, and connecting the vacuum tubing. The system is airtight and properly sealed if the buffer in the connected syringes is not drawn out when the vacuum is applied. If buffer is drawn out, disconnect the vacuum tubing and reassessthe position of the gasket, to determine whether it was displaced during incorporation of the substrate slide. Frequently, the gasket is slid to one side or the other during assembly. This is the most common problem that prevents sealing. If necessary, reapply the gasket to the face of the flow chamber, and avoid pressure on the gasket during placement of the slide on the chamber. 2. Remove the assembled flow chamber from the rectangular dish, and rinse the outer surface of the slide with a stream of deionized, distilled water to remove salts. Dried salts on the slide result in a poor microscopic image. Transport the entire flow-chamber system (syringes, tubing, and flow chamber) to the stage of the inverted phase contrast microscope. Place the flow chamber (slide down) on the stage over the objective opening, and secure it by placing stage clips over the tubing that exits from the polycarbonate block. 3. Remove the 1O-cc syringe from its tubing, and place the tubing end into a 50-mL conical tube containing a suitable assaybuffer (e.g., Ml99 supplemented with 10 mA4 HEPES, pH 7.4, and 1% HSA for endothelial monolayers, or HBSS supplemented with 10 n-J4 HEPES, pH 7.4, and 1% HSA for adsorbed-protein receptors and

514

Ramos and Lawrence platelet monolayers). Ehmmate an from the end of the tubmg before placmg into the assay medta, to avoid drawing an mto the flow-chamber system Mount the 60-cc syrmge (attached to the flow chamber by tubmg) onto the syringe pump Set the diameter accordmg to the syrmge chart provrded by the manufacturer Set the rate to mL/mm and the flow dtrectton to withdraw Inmate flow at 1-3 mL/ mm (for a 60-cc Becton-Dickinson syringe) and allow assay medta to be wtthdrawn through the flow chamber over the adhestve substrate for 2-3 mm, to eqmhbrate the system Assays can be performed at room temperature, or at 37’C, by placing the comcal tubes contammg assay medta or cell suspensrons m a temperature-controlled water bath Transfer the tubmg m the assay media to a conical tube contammg a cell suspenston Adhesron of leukocytes, such as neutrophtls, leukemta cell lures, or transfected cells, can be studted over a range of cell concentrations, typically 0 5-5 x lo6 cells/ml. Select the destred mtcroscope obJective (e g , x 10 or x20) and adjust the posttton of the flow chamber, usmg the stage mampulators If using adsorbed protem on plasttc slides, focus on the center dot m the marked circle outhmng the protein area (see Subheading 2., step a.3) Start the flow at 34 mL/mm, and watch for cells to enter the field of vtew Focus on the cells, so that they are phase bright and sharply spherical agamst the background

3.4. Scanning

for Adhesive

Interactions

and Recording

of Data

To begm scannmg for adhestve mteracttons, move the field of vtew to the edge of the marked circle, at the point where cells are begmmng to enter the area of the adsorbed protein In the case of endothehal or platelet monolayers, any starting point near the cell entry pomt can be selected, but tt IS best to analyze near the center of the channel, and midway along the length of the channel. When a bolus of cells enters the field at 3-4 mL/mm, decrease the pump speed m stages, until mteractrons become detectable Usmg the focus adjustment and the vtdeo momtor, begin scannmg enttre fields (wtdth of the momtor screen) at 5-30 s Intervals The scannmg sequence should Include 15-20 fields of vtew, and can proceed down the center of the substrate (with the flow of cells), followed by a shaft of view and scannmg up the plate (agamst the flow of cells). Adherence assay. The scanning process is repeated for a serves of flow rates at the same time point and the number of adherent cells counted from vtdeo tape recordings of the scans Attachment rate assay (rate of cell accumulatton on adhestve substrate at a gtven wall shear stress) The scanning process IS repeated over time at the same flow rate (applied wall shear stress) and the number of adherent cells accumulated at each time point IS quantttated from video tape recordmgs of the scans Detachment assay (resistance of cell adhesion to flurd shear stress) Whtle InJected cells (or cells infused through the inlet port) are settling onto the lower wall of the chamber (30-90 s), the tubing is transferred to assay media, m order to rinse away unbound cells (for example, l-2 mL/min for 2-3 mm) Flow rates

Cell--Cell Adhesion Under Flow Conditions

515

may be Increased at 30-40 s intervals, followed by scanning at each pomt to measure numbers of cells remammg bound at each flow rate (applied wall shear stress) 5 Tests for specttictty. Monoclonal antibodies agamst adsorbed adheston molecules, as well as those expressed on leukocytes, endothehal monolayers, or platelet monolayers, can be utthzed to assess whether flowmg cells are specttically bmdmg to adheston molecules of interest. Cell suspenstons to be perfused through the flow chamber can be treated with monoclonal anttbody at 5-20 ,ug/mL (1.100-1~500 for ascttes thud), followed by direct mtroduction mto the flowchamber system without washing to remove unbound antibody. To treat adsorbed protetns or cell monolayers with monoclonal anttbodtes, antibody solutions (5 20 ,ug/mL) can be infused over the adhesive substrate for 1-2 mitt, followed by stoppmg of the flow and incubation for 5-15 mm. Alternatively, I-mL antibody soluttons can be injected through the flow-chamber stde port using the 1-cc syrmge postttoned near the top or cell-entry area of the flow chamber To accomplish this, remove the withdrawal syringe from the pump and close off the tubmg from the cell/buffer reservotr, using a hinged paper clip as a clamp While slowly depressing the lower 1-cc side port syrmge, remove the upper 1-cc syrtnge-stopcock combmatton from Its flow chamber side port. Fill the syringe wtth anttbody solution, and tap to dispel an Depress the plunger to form a drop at the opentng of the attached stopcock Reattach the syringe to the flow chamber, usmg a dropto-drop connectton to avotd an bubbles Slowly inject the contents of the syringe mto the flow chamber

4. Notes Two general types of experiments are typically performed using parallelplate flow chambers. The most broadly applicable 1s the detachment, or shear resistance assay, which can be used to evaluate cell-cell, cell-matrix (with any number or combmation of mtegrm-ligand paus), or cell-btomaterial mteracttons. In this format, either proteins or cells, such as endothelial, epttheltal, or any anchorage-dependent cell, are unmobtl~zed or cultured on one wall of the chamber. The second format uses contmuous flow to create constramts of force and contact ttme on cell adheston, srmrlar to that imparted by blood flow on cells of the vasculature. This second method has recently proved useful for estrmatmg forces and bond ltfettmes of adhesion receptors (5-7) 1 Static-flow or detachment flow assay. Cells may be Introduced mto the flow chamber by mJection through the stde port (if only small numbers of cells are avatlable) and allowed to settle onto the lower wall of the flow chamber The sedtmentation veloctty for most mammalian cells 1s approx 1 pm/s, but can be estimated from Stokes’ Law vt=2R2 (p - p,)/9~ where p 1sthe density of the cell, ps IS the denstty of the suspendmg media, p 1sthe viscosny of the suspending medta, R IS the radius of the spherical particle or cell, and v, IS the sedtmentatton veloctty. Problems associated with the detachment assay Include inadequate ttme allowed for sedtmentatton and vartatton between fields of view m the number of

516

Ramos and Lawrence

cells because of inadequate mixing of the inJected cell suspenston Clumping of cells either m the stock vral or durmg sedimentation can srgmficantly affect the spattal dtstrtbutton of cells m the chamber For mteractrons m which cells form relatively stable mteracttons, such as those through mtegrm receptors with extracellular matrrx molecules such as fibronectin, It 1s helpful to begm wrth a relatively low flow (under 1 dyn/cm2 wall shear stress), to remove nonadherent cells before begmnmg scans of the number of bound cells To determme when the flow chamber’s solution has been exchanged (for example, if a drug is bemg Infused to modulate adhestveness), the nnsmg media can be doped wrth small beads lo-pdlameter (Duke Sctentific [Palo Alto, CA], #9010) or erythrocytes as flow markers, to indtcate when the new solutton has completely displaced the orrgmal solution Since most flow chambers are transparent to allow mtcroscoprc vtsuahzatton, the number of input cells can be drrectly counted and then fates determined Among the crmcal vartables that must be controlled are the magmtude of the flow and its duration at each level Over the time-scale of seconds, cells can deform m response to the shear, and possrbly increase then contact area and number of adhesive bonds For example, with seiectm-medrated leukocyte rollmg, the rapid bond-formatton rates appear to stabilize adhestons wrth mcreasmg sheat 2 Attachment of cells under steady flow conditions A second approach to quantrfymg cell-cell adhesron, using a flow chamber that has been widely applred to the study of blood cell adhesion to natural and artrficral surfaces, IS the attachment, or contmuous-flow, assay Adhesion of cells to surfaces m thts protocol takes place under flow condmons, which Imparts both constraints of force and contact ttme Higher flow rates exert higher levels of stress on the mrttal bonds formed, and, concurrently, hmrts the contact trme, because the flmd veloctty IS greater, Even if a bond creates a strong enough crossbrtdge to hold a cell under physrologtc levels of shear, the wmdow for the associatron of the hgand-receptor pan may be too brief for the dockmg event to take place The time-scale for a bond-formatron event, therefore, scales with the shear rate at approxtmately the rectprocal of the shear rate (l/G) (8,9) Mathematical analyses exist for derrvmg the actual forces imparted by the flow (6,10,11) The contmuous flow mmrmrzes nonspecific cellular interactions with the substrate, smce contact ttmes are short enough to remove the effect of matrtx protem secretton from the mechanisms of adhesion. 3 Data presentation. Data from flow chamber assays IS typtcally presented as number of cells bound per umt area of visualized substrate Data from detachment assays, using a parallel plate flow chamber, is typrcally presented as a percent remaining bound as a function of mcreasmg wall shear stress A force or shear resrstance titration can be generated that Indicates the average strength of adhesion of a populatron of cells. In contmuous-flow assays, the rate of accumulatton as a function of time and area can be determined and reported as number of cells bound per unit area. Calculatmg a cell flux IS posstble by dtvtdmg the number bound by the time, creating an adheston parameter of cells/area/time Cell accu-

Cel/-Ce// Adhesron Under Flow Conditions

517

mulatlon linear with time can be a useful parameter, but attachment of cells may create nonlmear effects because of coverage of binding sites Addmonally, rf flow rates are low enough, sedimentation will have some effect on the number of cells near the wall (12) 4 Hydrodynamic forces and relevant equations for wall shear stress The level of flow of the rmsmg media IS controlled by the syringe pump. The wall shear stress generated by the flow, which IS lammar under most condmons (Reynolds number on the order of 1 to 2), IS a descrrptron of the flow rate that allows compartson of the effect of flurd forces wrth other systems of drffermg dimensions and geometries. The Reynolds number (Re) IS defined as &p/p, where D is the chamber gap, v IS the mean velocity of the fluid, p IS the density, and p IS the vlscostty Since the fluid shear stress varies with distance from the wall of the flow chamber (parabolic velocity profile), the typical flow rate spectficatton is the wall shear stress. When Re IS low (where vrscous forces dominate interial forces), the shear stress correlates with the actual stresses imparted on the adherent cell by the flutd flow Calculatton of the magnitude of the stresses acting on an adherent cell depends m part on the geometry of the cell and its mechanical properties, which determine to what degree it deforms m response to the flutd shear stress However, to a reasonable approxtmatron m the case of comparatrve studies of a specific cell type, adherent cells experience nearly the same level of detachment force Of course, a rounded cell ~111probably experience a different force than a htghly spread one, and thetr contact areas may differ considerably, but over a population of cells, such effects may average out. Both time of flow and the magnitude can be varied easily by a syrmge pump, with wall shear stresses down to 0 1 dyn/cm2 attainable (though the step motor m many syringe pumps may mtroduce a noticeable degree of pertodtctty m the flow, which can be compensated by using smaller syringes to allow the motor to turn faster) At a wall shear stress of 0 1 dyn/cm2, the force acting on a cell can be estimated wrthm an order of magnitude by multiplying the wall shear stress (since the cell is very close to the wall) by the surface area of the cell. Cells such as leukocytes, whrch form transient tethers through some types of adhesive receptors, and pivot around these statronary points (6), exert a force on the bond estrmated by the pivot arm and the fluid force on a sphere in a shear flow (10) Solutions for the shear force generated by the flow, based on varying assumptions and srmplrficatron, have been reported for endothellal cells and platelet thrombr m flow chambers. Close to the wall, the parabolic veloctty profile can be approxtmated by a linear-velocity gradient based on the wall shear rate, which 1s useful for dtstmgutshmg adherent from nonadherent cells that are tumbling along m the flow The shear rate IS multiplied by the viscosity of the perfusion media, to give the shear stress The shear rate is the change m velocity with distance from the wall of the chamber or tube. The fluid shear stress 1sdefined by z = p (dv/dr), where dv/dr is the shear rate (s-t), p IS the viscosity (Poise), and r IS the shear stress (dyn/cm2). The wall shear stress IS given by rwall = 3 pQ/2ba2, where Q IS the volumetrtc flow rate (cm3/s), b IS the channel width (cm), p IS the vrscosrty

518

Ramos and Lawrence (Poise), and a is the half channel height (cm) For example, at a wall shear rate of 100 s-r, the fluid velocity is approx 100 pm/s at 1 p from the wall Although this 1s only an approxrmation of the velocity profile, whrch 1s actually described by a parabolic equatron, it does provide a scalmg factor If the vrscosity of the suspending media is increased, the shear stresses are increased proportionally This strategy has recently been applied to mvestrgate the force dependence of celladhesive crossbridges independent of the shear rate (9)

References

4 5

6

7

8

9

10 11

12 13

Springer, T A (1995) Traffic signals on endothehum for lymphocyte recrrculation and leukocyte emigration Annu Rev Physlol 57,827-872 Xrao, Y and Truskey, G A (1996) Effect of receptor-ligand afimty on the strength of endothehal cell adhesion Bzophys J 71, 2869-2884 Abbassr, 0 , Krshrmoto, T K , McIntn-e, L V, Anderson, D C , and Smith, C W (1993) E-selectin supports neutrophrl rollmg m vrtro under conditions of flow J. Clan Invest 92,27 19-2730 Buttrum, S M , Hatton, R , and Nash, G B (1993) Selectm-mediated rollmg of neutrophtls on unmobrhzed platelets Blood 82, 1165-l 174. Kaplanski, G., Farnarrer, C , Tissot, 0 , Pierres, A , Benohel, A M , Alessi, M. C , Kaplanski, S , and Bongrand, P (1993) Granulocyte-endothelmm mmal adhesion Analysis of transient bindmg events mediated by E-selectm m a lammar shear flow Bzophys J 64, 1922-1933. Alon, R , Hammer, D A , and Sprmger, T A (1995) Lrfetrme of the P-selectmcarbohydrate bond and its response to tensile force m hydrodynamic flow Nature (Lond) 374,539-542 Lawrence, M B., Kansas, G S , Kunkel, E J , and Ley, K (1997) Threshold levels of fluid shear promote leukocyte adhesion through selectms (CD62L,P,E) J Cell Blol 136,7 17-727 Capo, C , Garrouste, F , Benoliel, A , Bongrand, P , Ryter, A , and Bell, G I (1982) Concanavalm-A-mediated thymocyte agglutmation a model for a quantitatrve study of cell adheston. J Cell Set 56,21-48. Taylor, A. D , Neelamegham, S , Hellums, J D , Smith, C. W , and Simon, S I (1996) Molecular dynamics of the transmon from L-selectm to p2 mtegrm-dependent neutrophil adhesion under defined hydrodynamic shear Bzophys J 71,3488-3500 Goldman, A J , Cox, R G., and Brenner, H (1967) Slow VISCOUSmotion of a sphere parallel to a plane wall II Couette flow. Chem Eng Scz 22,653-660 Folre, B J. and McInttre, L V (1989) Mathematical analysis of mural thrombogenesrs Concentration profiles of platelet-activatmg agents and effects of viscous shear flow Bzophys J 56, 1!21-1141 Munn, L L , Melder, R. J , and Jain, R K (1994) Analysis of cell flux m the parallel plate flow chamber rmphcations for cell capture studies Bzophys J 67, 889-895 Stamper, H. B , Jr. and Woodruff, J J (1976) Lymphocyte hommg mto lymph nodes m vitro demonstration of the selectrve affimty of recrrculating lymphocytes for high-endothehal venules J Exp Med 144, 828-833

Cell-Cell Adhesion Under Flow Conditions

519

14 Dustm, M L and Sprmger, T. A (1989) T-cell receptor cross-lmkmg transiently stimulates adhesiveness through LFA-I Nature @lad ) 341,41-46 15 Smith, C W, Rothlem, R , Hughes, B J , Martscalco, M M , Schmalstteg, F C , and Anderson, D C (1988) Recognmon of an endothelial determmant for CD 18dependent neutrophtl adherence and transendothehal mtgratton J Clzn. Invest 82, 1746-l 756 16 Smtth, C W., Ktshtmoto, T K ,Abbasst, 0 , Hughes, B , Rothlem, R , McIntne, L V, Butcher, E. C , and Anderson, D C (199 1) Chemotacttc factors regulate lectm adheston molecule 1 (LECAM- 1)-dependent neutrophtl adhesion to cytokme-sttmulated endotheltal cells m vitro J Clrn Invest 87,609-618 17 McClay, D R , Wessel, G M , and Marchase, R B (1981) Intercellular recognmon quantttatton of mmal bmdmg events Proc Nat1 Acad Scz USA 78,4975--1979 18 Ward, M D , Dembo, M , and Hammer, D. A. (1995) Kmettcs of cell detachment. effect of &and density Ann Blamed Eng 23,322-33 1 19 Xta, Z and FroJmovic, M M (1994) Aggregation efficiency of activated normal or fixed platelets m a simple shear field. effect of shear and fibrinogen occupancy Bzophys J 66,2 190-220 1 20 Bell, D N , Spain, S., and Goldsmith, H L (1989) Adenosme dlsphosphate-Induced aggregation of human platelets in flow through tubes 1 Measurement of concentration and size of single platelets and aggregates Bzophys J 56, 8 17-828 21 Bargatze, R. F , Kurk, S , Butcher, E. C., and Jutila, M A (1994) Neutrophtls roll on adherent neutrophtls bound to cytokine-induced endotheltal cells via L-selectm on the rolling cells J Exp Med 180, 1785-1792. 22. Kuo, S C and Lauffenburger, D. A (1993) Relattonshtp between receptorihgand bmdmg affinity and adheston strength Bwphys J 65,2 191-2200 23. Sung, L A., Kabat, E. A , and Chien, S (1985) Interactton energies m lectm-medtated erythrocyte aggregatton J Cell Bzol 101,652-659 24 Palecek, S. P, Loftus, J C , Ginsberg, M. H., Lauffenburger, D A and Horwitz, A F. (I 997) Integrm-ltgand bmdmg properties govern cell mtgratton speed through cell-substratum adhesiveness Nature (Lond) 385, 537-540 25 Truskey, G A and Ptrone, J S. (1990) The effect of fluid shear stress upon cell adhesion to tibronectm-treated surfaces. J Blamed Mater Res 24, 1333-I 353

39 Quantitative Measurement of the Biological Response of Cartilage to Mechanical Deformation R. Gregory

Allen, Solomon

R. Eisenberg,

and Martha L. Gray

Introduction 1.1. General Introduction Cartilage functionality is defined, in part, m terms of the ability of the extracellular matrix to support a mechanical load It has been shown that such mecharncal loadmg can influence the biological response of the chondrocytes that are embedded m the extracellular matrix. Cultured tissue explants have served as useful models for studying such chondrocyte-medtated responses to mechanical deformation. The explant paradigm facilitates the control of mechanical and biological variables that may influence cellular behavior. This chapter presents the means for assessingthe biologic response of cartilage to controlled mechanical stimuli, and lays the foundation to further explore the response of cartilage to mechanical stimuli in the presence of other factors, such as cytokmes and growth factors (e.g , IL-l p, TGF-j3, IGF). A requirement for observing and reportmg the effects of mechamcal loading on cartilage metabolism is the abilrty to quantify the deformation or stress state of the tissue. This becomes a key consideration m all aspects of experimental design, mcludmg tissue harvest, culture, mechanical mtervention, and assessmg metabolic activity. In constructmg the experiment, one must carefully consider the geometry of the test sample and the macroscopic variables associated wrth the mechanical stimulus, because the choice of boundary condrtions and method of loading determine the extent to which the internal mechanical environment may be conceptualized. The macroscopic mechanical parameters of relevance are load and displacement (load can be more completely defined with compressive, tensile, and tangential forces, and hydrostatic and osmotrc 1.

From

Methods in Molecular Edrted by J R Morgan

Medicine, Vol 18 Tissue Engineering Methods and and M L Yarmush 0 Humana Press Inc , Totowa,

521

Protocols NJ

522

Allen, henberg,

and Gray

pressures), which relate to contmuum variables of stress and stram. These are dependent varrables The application of one specifies the other, and, therefore, necessitates methods for controllmg one and measuring the other. Ulttmately, the mechanical behavtor of the tissue m the vicmity of the cells must be considered, for therem lies the key to identlfymg the causal relatronship between mechamcal stimulus and biological response. Applied displacements, or forces, produce deformations of the extracellular matrix and the cells. For these deformatrons to occur, fluid must flow and redistribute. This flow IS induced by hydrostatic pressure gradients. Consolidation of the matrix alters the fixed-charge density, thereby affectmg the osmotic pressure of the mterstrtial flurd and altering pericellular chemical properties like pH and ionic strength Time-dependent material deformatrons and pressure gradients are present during transient mechanical processesand during dynamically applied drsplacements or forces. These, in turn, Induce flutd flows and effect physrcochemical changes, like electrokmetrc separation of charged metabolrtes It has been demonstrated repeatedly that cells are receptive to stattc and dynamic mechanical perturbations of their local environment, as assessedby the subsequent modulation of metabohc activity. However, the precrse nature of the transductron mechanisms have yet to be clearly elucidated This tissue-level view suggests that the mechanisms might include cell deformatton, fluid flow, pressure, electrical phenomenon, and physlcochemrcal effects-and data exists supportmg each of these mechanisms. The careful control of the macroscopic environment, combmed with models esttmatmg the tissue-level mechanical envnonment m which the cells resrde, has allowed mvestigators to begm to experimentally evaluate these transduction mechamsms, and 1sthe emphasis of this chapter In practice, the vrcrssrtudes of experimentatton become apparent, because our understanding of the mechanical envn-onment is lrmlted, m some cases,by constraints imposed by trssue-culture techmques, or by mcomplete mechanical descrrpttons of the tissue deformation state. For example, m our laboratory, one well-established long-term culture protocol entarls the statrc axial compression of radrally unconfined, cylmder-shaped cartilage explants. The expertmental system was designed to facilitate quantttative mvesttgation of the effects of compressron (matrix consohdatron and fluid exudation) on cell metabolism. However, the mterpretation of the results requires the understanding that matrix compaction is not the sole deformatronal result of axial compression The radially unconfined nature of the explant gives the sample the freedom to bulge, creatmg tensrle stressescaused by matrix extension m the radial direction. The application of trme-varymg loads or displacements Introduces addmonal complications. In this dynamic environment, the tissue experiences spatially complex pressure gradients and fluid flows, the magnitude and distri-

Response of Cart//age to Deformat/on

523

butron of which depends on frequency. Given that cells reside throughout the sample, and that the mechanics are comphcated functions of space and frequency, the mechanical stimulus seen by each cell can be different It IS also important to keep m mmd that many of the methods used to determine btologtcal acttvtty have ltmtted spatial resolutton. Most measurements based on radtotracer mcorporatron typrcally assess the bulk (spatially averaged) response, although autoradtography has the capabtltty to localize regrons of synthesis and assembly. Laboratory procedures detailed here discuss the formulatton of culture medium and the explantatton and preparatron of cartilage mto disks. Note that from an engmeermg perspective, the geometrically friendly cylinder, m some cases, stmphties the normaltzatron of loads and displacements mto stresses and strams. Presented are methods describing the radially unconfined, axially applied stattc and dynamic compression of cultured cartilage disks. The btologrcal response of cartilage, medtated by the resident chondrocytes, IS momtored via btochemtcal and radrorsotope measurements of syntheses and degradation of the extracellular constituents. Some common methods for quantifying btologtcal responses to mechanical stimuli are described at the end of this chapter, and references are provided for direction to experimental procedures routmely employed by others 7.2. Conceptual Framework Cartilage consists prtmartly of an aqueous electrolyte and a sparse population of cells that reside in a plentiful extracellular matrtx. There are two prmctpal components of the matrix. collagen, whtch forms an tsoelectric fibrillar network that IS thought to be the primary determinant of cartilage tensile strength; and proteoglycans, which, wtth the tissue fluid, constrtute a polyamomc gel that resrsts compresston. The collagenous network contams the aggregating proteoglycans and other noncollagenous proteins through chemrcal bmdmg reactrons and stertc hindrances. The negattvely charged components of the proteoglycans, glycosammoglycans, cause posrtlve ion (or counter-ton, e.g., Ca*+, H+ Na+) concentratrons to be elevated over then respective concentrattons m’the culture medium. Slmrlarly, concentrattons of negattvely charged ions (or co-ions, e.g., Cl-, SOd2-) are depressed m order to mamtam tissue electroneutrality. Mechanical compressron of cartilage dtsks alters the stress and deformatton state wtthm the extracellular matrix. Compresston also causes changes m hydrostattc pressure, flurd content, and osmolabty, and induces fluid flows and electric fields (streaming potentials). Chondrocytes may respond directly to applied stress, or to any of these compresston-Induced phystcal phenomena that occur m then envtronment. To factlrtate experimental destgn, included 1sa

524

Allen, Eisenberg, and Gray

brief discusston of the tissue-level mechamcal envuonment durmg commonly used methods of loadmg. The apphcation of a static compressron mmates a series of events that result in an equiltbrmm deformation. Durmg the mechamcal transtent (creep or stress relaxation [see Note 11) that follows compression, fluid convection and exudation ultimately leads to an equilibrium mechanical deformatron. During this transtent phase, counter tons are convecttvely separated from the fixed negative charges of the proteoglycans, generating electric fields collmear with the fluid flow The reduction m volume (caused by fluid loss) alters the phystcochemlcal environment by concentrating the negatively charged proteoglycans, which further partitions counter-tons mto the tissue, mcreasmg then concentratrons (and decreasing pH), and stmultaneously decreasing co-ion concentrations. Thus, from the perspective of the cell, many physical forces and flows occur during the transient phase that may trigger a cellular response. At equilibrmm, the cell may be affected directly by the applied deformation, or by the concomitant changes m the phystcochemrcal environment. Dynamic compression is the periodic apphcatton of a load or displacement, usually superimposed on a steady-state static compresston. In contrast to static compression, pressure gradients, fluid flows, and electric fields persist with the same frequency as the applied excitation. The mechamcal envn-onment during dynamic compression is spatially complicated and dependent upon boundary condttions, but modeling efforts have provided significant insight for certain lrmltmg conditions (1,2). Common to these models 1sthe findmg that fluid velocity and hydrostatic pressure in dynamically compressed, radially unconfined disks, exhibit both an axial and radtal dependence, which vary with the frequency of the applied excitation. Given a relatively low-excttation frequency (e g., 0.0001 Hz m a 3-mm diameter disk), the fluid velocity and pressure drstrrbutrons are expected to be spatially quasmmform As frequency is increased, however, pressure gradients develop and fluid flows arise, with flows becoming mcreasmgly restricted to the region adjacent to the radially unconfined surface of the dtsk. A cartilage disk mechanically loaded at a relatively high loading frequency (e.g , 0.1 Hz) would experience permeabihtyrelated fluid flow restrictions, elevating hydrostatic pressures m the bulk of the tissue. Fluid flows are expected at the radial surface (within 0.3 mm at 0.1 Hz for a 3-mm diameter sample), where the highest rates of fluid flow occur with correspondmgly steep pressure gradients (3). In the discussion above, and m the methods described below, our focus is the direct application of mechanical loads or displacements However, it is important to mention two alternative approaches used for assessingthe biologtcal effect of mechanical strmuh. These involve altering the hydrostatic or osmottc pressures of the bathing media. The mechanical effect of varymg hydrostatic

Response of Cartilage to Deformatron

525

pressure, vra a hydraulic actuator, is a spatially uniform change m pressure throughout the culture medra and cartrlage sample. Assuming that the solid and fluid phases are individually mcompresstble, ttssue deformations, fluid flows, and all of the secondary mechanical events are absent. The effects of hydrostatic pressure on the brologlcal response of cartilage have been studied, and the methods are descrrbed elsewhere (4). In contrast, varying osmottc pressure can Induce flurd flow, tissue deformanon, and so on. This IS achieved by altering the osmolality (osmotic pressure) of the bathing medium with a tissue impermeant solute, or by placing the tissue in dialysis tubmg and addmg a solute (e.g., polyethylene glycol) to the bathing solution outside the dialyses tube. When the osmotrc pressure of the bathing solution ts made hypertonic, fluid leaves the tissue, causing the tissue volume to decrease and the matrix to deform as though compressed. As with dn-ect mechanical compresston, osmotttally induced fluid loss has the capacity to mechanically deform cells and effect physlcochemtcal changes by altering the concentrattons of polyelectrolytes and the density of fixed charges. A more complete treatment of this toptc IS detatled elsewhere (5-8’. 2. Materials 2.1. Culture Medium 25 mM HEPES-buffered Dulbecco’s modified Eagle’s medium (DMEM) (Gibco-BRL, Gatthersburg, MD, 12320-032), DMEM (JRH Btosctences, Lenecia, KS, 5 1447-78P), 10 mM nonessential ammo acrds (Sigma P-7145), 100 mM L-proline (Sigma, St. Louts, MO, P-0830), 200 m&I L-glutamine (Sigma G-6392), 20 mg/mL L-ascorbate (Sigma A-4544), anttblottc/anttmycotrc (Srgma, A-7292), and heat-inacttvated fetal calf serum (FCS) (Note 2; HyClone, Logan, UT). 2.2. Dissection Autoclave or alcohol sterilize, as applicable: a vibratmg sawblade, dtssecnon tools, mcludmg scalpels, razor blades, forceps, hemostats, and wetghmg spatulas; several pairs of sterile gloves, a mask (if dissecting outside a biosafety cabinet) a topical antiseptic (e.g., Betadme); gauze; a drill press (configured to a rotational speed of 1200 rpm) and cormg bit (we use a custom bunt 3/X-mch ID bit), sterile specimen cups; and multtwell culture plates. Our rmsmg solution conststs of 1% by volume of antlbiotic/anttmycottc (Sigma, A-7292) m 500 mL of Hanks’ balanced salt solutlon (Gibco-BRL, 14025-092). We use a Sledge microtome (American Optrcal, Buffalo, NY) to prepare explants mto slices of umform thickness. For preparmg these shces into disks, assemble the followmg materials in a blosafety cabinet: forceps, rmsmg solutton, multrwell

Allen, Elsenberg, and Gray

526

culture plates with -0.5 mL/well complete media, dermal punch (Mtltex Instrument, Lake Success,NY), and a cutting block The function of the cuttmg block 1sto prevent damage to the punching tool; we use a block (approx 5 x 8 m. by 0.5-m. thick) of polysulfone, which 1sautoclavable. 2.3. Culture

Techniques

Multrwell culture plates, forceps, media, stertle Pasteur ptpets for exchanging media, and mrcrocentrifuge tubes for media cryopreservatton 2.4. Mechanical Stimulus Custom-butlt compresston devtces (see Subheading 2.5. Biochemical 2 5.1. DIgestion

3.4.1.).

Techniques

Papam (Sigma, P-3125), L-cysteme hydrochlortde (Sigma, C1276), Na2HP04, Na2EDTA, 0.20~pm syringe filter (Gelman Sciences, Ann Arbor, MI, #4 192), 10 mL syringe and a 25-gage needle, one 2-mL cryogemc veal per tissue sample (Corning), 70% ethanol sterile wipes, and a hot-water bath 2.5.2.

DMl3 Assay

2 5 2.1 DYE SOLUTION

DMB chloride (Polyscrences, Warrmgton, PA, #3610), NaCl (Mallmckrodt, Pans, KY, 7544), glycme (Sigma, G-7403), sodium aztde (Fluka, Swttzerland, #7 1290), 100% ethanol (reagent grade, Pharmco, Brookfield, CT), m HCL (Ftsher Sctenttfic, Hampton, NH, SA49- 100,50-mL centrifuge tube, flea magnetic stirrer, aluminum foil, mask, gloves, and demmerahzed water. 2.5 2.2. GLYCOSAMINOGLYCAN STANDARD SOLUTION

Chondroitm sulfate (Sigma, C4384) (Note 12), demmerahzed water, and several cryopreservation vials 2 5 2.3 SPECTROPHOTOMETRY

The spectrophotometershould be configured to measureabsorbanceat 525 nm Matenals should include severaldisposableptpet ups,500~mL pump-actrondrspenser tilled wtth DMB dye solutton, and po lystyrene cuvets or 4 5 mL mtcroplates 2 5.3. DNA Assay 2 5.3 1 10X TEN BUFFER

Trrs (Mallmckrodt H590), EDTA, and NaCl (Mallmckrodt 7544), demmerahzed water, HCl (-14 M HCl), and a 0.22-l-m filter

Response of Cartilage to Deformation

527

2.5 3.2. CONCENTRATED DYE STOCK SOLUTION

Hoechst 33258 (Polysctences, Warrington, PA, #09460), sterile demmeralrzed water, and a foil-wrapped glass or brown plastic bottle. 2 5 3.3. SPECTROFLUOROMETRY

Fluorometer, acryl cuvets (10 x 10 x 48 mm, Sarstedt, Newton, NC, 67755), several disposable ptpet ups, and a 500-mL pump-actton dispenser tilled with working dye solutton. 2.5.4. Radioactive Labeling 2.5.4.1,2

M BYANIDINE HCL SOLUTION

[35S]sulfafe (Na235S04, NEX-041, New England Nuclear, Boston, MA) and trttiated prolme (I-[5-3H]prolme, NET-573, New England Nuclear). 2.5.3 2. START COUNTING

Ltqutd scmttllatton counter, Omm veals and caps (Wheaton, Htllsboro, OR, 6602 l-23), Ecolume scmttllatron flutd (ICN, Costa Mesa, CA, HS 82470), guanidme hydrochlortde (Aldrich, Milwaukee, WI, G11705), sodmm acetate (Sigma S-8625), demmerahzed water, and 1 M HCl 3. Methods 3.1. Culture Medium Formulation Our complete media consists of low glucose DMEM buffered with 10 & N-2-hydroxyethylptperazine-IV’-2-ethanesulfomc acid (HEPES), and forttfied with 0.1 m&I nonessential ammo acids, 0.4 mM L-prolme, 2 nnI4 L-glutamme, 100 U/mL of penictllin, 0.10 mg/mL streptomycm, and 0 25 pg/mL amphotertcm-B, 1% by volume of heat-macttvated fetal calf serum (FCS), and 0 50 M/ mL L-ascorbate. We prepare culture medta m 500 mL batches and store rt between 2 and 6°C. Complete media IS comprtsed of: 1 2 3 4

200 mL 293 mL 5 mL of 2 mL of

of 25 mh4 HEPES-buffered DMEM of DMEM wtthout HEPES 10 rnh4nonessentlal amino acids 100 mML-proline

Twenty minutes prior to changing the medium, add 1% by volume of each of the followmg substances,which have limited stability at culture condmons, and warm m a hot water bath at 37°C 5 200 mM L-glutamine 6 20 mg/mL L-ascorbate 7 Antlblotx/antlmycotIc

528


8 Heat-rnactlvated FCS Example: For a desired total volume of 10 mL of complete media, add 100 pL of each of the four supplemental substances to 9.6 mL of base media, preequllibrate to culture temperature, and deliver to explants. 3.2. Dissection The ultimate goal is to prepare plane-parallel dtsks (i-e , cylmdrtcal disks of uniform thickness). The first step IS to harvest the cartilage so that it 1spossible to section the tissue using a mrcrotome. To mamtam a reasonable degree of samtatron, dtssect as much of the tissue as possible wrthm a brosafety cabmet. A vlbratmg saw has proven to be a useful device in harvesting tissue from larger joints (e.g., femoropatellar groove, humoral head). This tool, whrch 1s similar to a bone saw, IS commercrally available m hardware stores. Blades can be custom made to accommodate the larger joints of adult arumals For trrrgating cartilage exposed during the explantatton process, and for lubrrcatmg the cutting edges of power tools (e.g., bone-saw blade, drill bit), prepare a rmsmg solution of HBSS supplemented with an antrbrotrc/antrmycotic Combrne 500 mL HBSS and 5 mL anttbiotrc/antimycottc tnto a squirt bottle and store at 2-6”C. Use sterile spectmencups or multrwell culture plates for the bathing of tissue prior to secttoning. We use one 50-mL sterile specimen cup containing -20 mL rinsing solutron for each dtstal ulna joint. For tissue that requires cormg, we use a 24well culture plate containing -2 mL/well rinsing solution 3.2.1. Tmue Explantation Procedures The synovial joint from which the cartilage IS to be dissected should be delivered to the laboratory on ice, to preserve cellular viability. To prevent mfectron, the joint should be Intact and encapsulated with tissue If the dtssectron IS not immediate, store the tissue between 24°C. All tissue used m our studies is bovine (calf and adult) in origin: 1-2 wk-old distal ulna or hmd leg femoropatellar groove and humoral head. In the vernacular of the local abattorr, the termmology for the drstal ulna, humoral head, and femoropatellar groove are front knee, shoulder, and hind quarter, respectively Begin by applying the toprcal antrseptrc liberally to the external tissues of the joint (e.g., musculature); rt IS important that contact of the antiseptic with the cartilage be avorded 3 2 1 1. DISTAL ULNA Under quasisterile conditrons aided by the btosafety cabinet, relieve the distal ulna of its surroundmg musculature, and pertosteal and perichondral connective tissues (Fig. 1). Isolate the ulna from the radius (the ulna ISthe smaller


529

Calf Distal Ulna

Metaphym -

Region of hypertrophlc ceils (Plane of separation)

Fig 1 Harvesting

disks from the calf distal ulna

of the two and has a large -1 -cm-thtck epiphyseal cartrlage region that is easily vtsualtzed). Separate the eptphysts from the metaphysts. bending the ulna unttl tt snaps m half 1sa suitable means for accomphshmg this. Separatton occurs at the mterface of the hypertrophtc cells of the growth plate and calctfied cartilage region of the metaphysrs. Immerse the eptphyseal apparatus (which mcludes articular cartilage, the underlymg secondary center of ossrficatton, eptphyseal cartilage [a.k.a. progenitor carttlage], and growth plate) mto the rmsmg solutton contained m the sterile specimen cup, and drscard the remammg tissue

3 2 1 2 FEMOROPATELLAR GROOVE AND HUMORAL HEAD We dissect theseJoints on the quastsanitary benchtop With a scalpel or razor blade, begin by removmg as much connective tissue from the Joint and dtaphySISas possible, taking care to leave the Joint capsule intact (whrch IS the best possible means of thwarting an infection). Then, cut through the capsule and sever any ligaments, so that the carttlagmous surface of interest IS completely exposed. Transect the diaphysts with the bone saw, so that the shaft can be clamped m a vise (see Note 3), securing the Joint so that the arttcular cartilage faces up. With the Joint tmmobrhzed, drill as many cores as possible. Configure the drill press to have a slow rotational speed (1200 rpm), and ahgn the Joint so that the drill bit 1sperpendicular to the tissue surface (Fig. 2). While cuttmg, contmuously lubrtcate the drill btt with rmsmg solution. To free the cores from the underlying cancellous bone, transect the chondyle with the bone saw (severmg the cores 1O-15 mm beneath the surface), and expel the cores with a probe. Submerge the cartilagmous ends of the cores m rinsing solutton (we use a 24-well culture plate with -2 mL of rmsmg solutton m each well).


530

Artlcular Cartilage Compact Bone Cancellous Bone

Cut here to free

Cut here to clamp

Fig 2 Drllhng cores from the humerus

3 2.2 Preparation of Explant Disks We prepare the explanted tissue mto cylinder-shaped disks. The dlmenslons vary with the experiment, but typical diameters and thicknesses are 2-4 mm and 0.8-2.0 mm, respectively Choosmg the explant dlmenslons deserves some conslderatlon. Accurate determmatlon of biosynthetic activity may dictate large samples, because each specimen requires an adequate cell population Alternatively, small samples may be required m order to obtain all samples from a smgle animal, thereby ellmmatmg mterammal variations. Furthermore, the explant dlmenslons are limited by nutrient transport conslderatlons. Our studies utllzmg axially confining culture systems (I e , radral nutrient transport) have found biosynthetic rates for 2- and 3-mm diameter samples to be equlvalent, suggesting these dlmenslons are acceptable 3 2.2 1 DISTAL ULNA Clamp the eplphyseal bone mto the sledge mlcrotome, so that the eplphyseal cartilage IS unconstramed and poised for sectlonmg. The mltlal -8OO-j.m1thick slice creates a flat surface, removes the growth plate, and generally leaves -4 mm of eptphyseal cartilage for sectlonmg mto plane-parallel slices. With forceps, place the slices into a pool of rmsmg solution on the cutting block Excise disks from each slice with a dermal punch, place into the culture plate, and begin mcubatlon. 3 2.2 2 FEMOROPATELLARGROOVE AND HUMORALHEAD The drilled-out cores are generally comprised of 5 mm (young) or 0.6 mm (adult) of cartilage atop -1 cm of subchondral bone. Clamp a core into the

Response of Cart//age to Deformation

537

sledge mlcrotome for sectionmg. Face off the material with an initial section. Note that drlllmg perpendicularly into the chondyle minimizes the amount of material that must be discarded at this step. Also, it is possible to retam the artlcular surface, with careful attention being paid to makmg the core perpendicular to the surface. Section the remaining underlying cartilage into the desired thickness, place slices into pooled rinsing solution, excise disks from each slice with the punching tool, and place the disks into culture.

3.3. Culture Techniques The environmental condltlons of the incubator should be set to 37OC, 5% C02, and, to mmlmlze media evaporation, at least 95% relative humidity. Culture the free-swelling cartilage explants m 0.5 mL complete media for at least 2 d before mltlating any mechanical stlmulatlon (see Note 4). Media should be exchanged on a regular basis; the frequency of media collection depends on the time-points requn-ed for assessingblologlcal actlvlty. 3.4. Mechanical Stimulus In our laboratory, cartilage disks are cultured and mechamcally loaded m one of two custom built compression devices The simpler of the two, a static compression device, can be easily fabricated by any machme shop, and ISrelatively easy to use As suggested by the name, this device does not allow dynamic compresslon (i.e., no contmuous oscillatory loading), and typically controls either displacement or load. Alternatively, a mechanical spectrometer enables the apphcatlon of static or dynamic displacement-baseddeformations, and simultaneously allows for the momtormg of the associatedtransient or dynamic loads Both compression devices are designed to axially confine cartilage samplesbetween two surfaces,the bottom of the compression chamber and a compression post located superiorly (see Note 5) To ensure tissue vlablhty, the radial surfacesof the explants are unconstramed, thereby allowmg the tissue to exchange waste products for nutrtents with the culture medium. Uncompressed samples,which serve as untreated controls, are cultured m the same incubator as the mechanically challenged samples. Discard disks that have obvious visible damage incurred during the explantation process, do not exhibit plane-parallel faces, or do not retam cylmdrlcal shape. Note that all explanted plugs will swell by approx 10% of theuorlgmal prepared thickness. Use a mtcrometer to measure sample thicknesses before mechanical stlmulatlon. 3.4.1. Compression Devices

3.4 1 .l

STATIC COMPRESSION CHAMBERS

The static compression chamber, depicted schematically m Fig. 3, ISessentially a 24-well culture plate secured mto an anodized alummum frame (see

532

Allen, Osenberg, and Gray DisDlacement Control

//

Load Control

,Vetical Screw & Wing Nut Set Screw ,Compresnon Post

Fig 3 Static compression

devices

Note 6). A quartz post transfers axial loads or displacements through a rigid, anodized alummum guide bar and compresses the cartilage sample against the bottom of the culture well. In displacement control, the set screw is tightened down on the compression post so that the distance from the end of the screw to the bottom of the culture well equals the length of the compresslon post plus the desired cartilage thickness. Precision shims are used to prescribe the speclmen thickness. In load control, mechamcal stress1sapplied by placing weights (of known mass) on the compresslon posts (see Note 7). Media IS exchanged by msertmg a sterile Pasteur plpet under the Lexan cover shield. These studies can involve a large number of samples, smce each device can accommodate 10 cartilage disks and a single incubator can contam several compression devices. 3 4 1.2 MECHANICAL SPECTROMETER The mechanical spectrometer enables various kinds of mterventlon by controllmg static or perlodlc displacements The mechamcal spectrometer generally allows automated control of displacement or load (m some instruments, the user has a choice of which to control; others are designed to allow control of only one) These instruments also provide a means of measuring both load and displacement. These machines can be custom built and are also avallable comrnerclally (e g., Dynastat, Instron, MTS, Vitrodyne). Tradeoffs m choosmg mstrumentatlon include the number of samples for which independent measurements are desired and the duration of an experiment (how long ~111an incubator environment be needed). The sample chambers resemble those of the static compression devices, but the details depend on the spectrometer The methods will assume the cartilage sample 1spositioned between the bottom of a culture well and a platen.

Response of Cartilage to Deformation 3.4.2. Preparatton for D/splacement-Based

533

Mechanical Stimulation

3 4 2 1 STATIC CO~WPRESSION CHAMBERS (DISPLACEMENT CONTROL) 1 Autoclave-stenhze all components of the compresslon device, forceps, weighing spatulas, and several Pasteur plpets (one per sample for extracting media, and a single plpet for adding media) 2 Assemble the entire compression device in the blosafety cabinet 3 Place cartilage sample m the compression chamber and add culture media (ensure that the sample 1scompletely Immersed) 4 Impose the prescribed displacement Specific details for applying the mechanical stimulus depend on the apparatus being used. The displacement should not be apphed too rapldly, m order to mmlmlze the risk of a devltahzmg mJurlous compression (9) or sample expulsion. The macroscopic engmeermg strain to which the cartilage sample is confined IS the quotient of the applied displacement and the uncompressed thickness 5 Begin culture 3 4 2 2 MECHANICAL SPECTROMETER 1 Autoclave-stenllze the compression chambers, forceps, and weighing spatulas 2 Place samples mto the compresston chambers, lock them mto posltlon m the testmg system, and add culture media so that samples are immersed 3 To ensure that the compression posts contact the cartilage surface, gradually drive the compression posts down until the load cell mstrumentatlon indicates contact (see Note 8)

4 Close the incubator door and allow the interior environment to reach culture conditions. 5 Begin mechanical stlmulatlon (see Note 9).

3.4.3. Sample Protocols 3 4 3 1 COMPRESSION The most stralghtforward procedure for inducing mechanical stress IS the appllcatlon of a compressive displacement. The dose-dependent effects of compression can be elucidated by modulating the degree of static compression: Commonly used compression levels are 0, 10, 20% (depicted m Fig. 4), and 50% (see Note 10). As previously mentioned, after the displacement 1s imposed, It takes some time for the load to reach a new steady state (stress relaxation) This time can be estimated (see Note 1) or measured by the mechanical spectrometer. Studies of the metabolic response to static compression must be conslderably longer than this time For 3-mm diameter samples, the stress relaxation time IS typically on the order of 25 mm. For load control, the rule of thumb 1s that the creep time 1s four-times longer than the stress relaxation time. The blological response to static compression (observed via radtolabelmg periods rangmg from 6 to 12 h) 1spresumed to be dommated by the cell and tissue deformation,

Allen, Elsenberg, and Gray

534 Displacement Control

Controlled

IOOJ

Observed

,Compresswe

Mechamcal

Displacement

Response

Time m culture

Load Control

Controlled

Observed

Compresswe I

Mechamcal

Load

Response

Time m culture-

Fig 4 Static compression

and phystcochemtcal changes resulting from compresston. Data thus far have not revealed arole for the transrent physical phenomena in ehcttmg the response All subsequentmethods will focus on displacement control. The reader can extend from these dtscusstonshow to perform load control studies. 3.4 3 2. COMPRESSION RELEASE

Contrrved as an extension of the time-invariant stattc compresston paradigm, the compression-release cycle (Fig. 5) entails a provtstonal deformatton followed by a return to the original condmon. This technique has been used to examme recovery (studied by adding radrolabels subsequentto the release stroke), and to examme a low-frequency dynamic mechanical envnonment developed by constructmg a pulse-tram of concatenatedcompresstonrelease cycles. 3 4.3 3 OSCILLATORY COMPRESSION

The application of a contmuous osctllatory displacement superimposed upon an offset compresston (Fig. 6) induces addrttonal sttmulatory flow and deformattonal phenomena. Assuming that the btologlcal response mtrmsrc to the offset compresston has been characterized, the tmphcattons of dynamtc compresston can now be studied. By taking advantage of the complicated

Response of Cartilage to Deformation Mechanical Compression g

,;I:

Release

:

*I 3

535

T--&/&-T 100

b Trme m culture Fig 5 Compressron-release cycle

Static Offset g

O--

lO--

1

6

m-.-j

loo

Time m culture

b

Fig 6 Dynamrc compression

mechanical behavior of the tissue, modulatmg the frequency and magnitude of the dynamtc compressron creates regimes m which different physical sttmuh are pronounced For example, small-amplttude osctllations mmimize the exchange of ttssue fluid with the bathing medmm; therefore, fluctuatmg physrcochemtcal (fixed-charge density, osmolahty) propertres assocrated wrth ttssue hydration can be dismissed as possible sttmulatory factors. The static offset is required tf a symmetrtc osctllatron IS to occur. Furthermore, the loads required and the abthty of the tissue to reswell hmtt the amplitude of dynamic compressron to l-2%, values that are likely to be physrologrcal

3.5. Biochemical

Techniques

This sectton begms with a procedure for the drgestron of cartilage samples, a necessary step for further brochemrcal analysis. Included are protocols for the assessment of sulfated glycosammoglycan and DNA content, and the mtroductton of radtolabels for momtormg brosynthesrs. Measure sample wet and

536

Alien,

Eisenberg, and Gray

dry (lyophrlized) wt prior to any btochemrcal analysts Wetght measurements aid m the mterpretatron of brochemtcal data. 3.5.1. Dgestion of Cartilage Explants It IS necessary to solubihze cartrlage samples for btochemtcal analysts. Papam IS a sulfhydryl protease of wide spectficrty, whtch, when activated wtth 0 010 M cysteme, degrades most protemaceous substrates (see Note 11) Sufticrent drgestion occurs after incubation (-16 h at 60°C) with enzyme.substrate ratios of -150 (e.g., 0.125 mg papain per 6.25 mg tissue dry wt) (IO) Papam should be kept sterile at 2-6”C. Sterrlrzatron IS necessary rf DNA measurements are to be made on the drgests 3 5 1 1 1 L OF PHOSPHATE BUFFER WITH EDTA (PBE) 1 2 3 4 5 6

Weigh 14.20 g Na2HP04(0 100 h/r) Weigh336gNa,EDTA(OOlOM) Add to 1 0 L of demmerahzed water Adjust pH to 6 5 with 1 NNaOH Autoclave-sterilize Store at 2Z6”C

3 5.1 2 100 ML DIGESTION SOLUTION

Preparmg excessdigestion solutton ISnot recommended, because its acttvny 1s unstable (solutton ISactrve for about aweek, when stored at 2-6”C). This volume IS suffictent for digestmg 200 -20 mg (dry wt) cartilage samples (0.5 ml/sample). 1 2 3 4 5

Weigh 176 mg cysteme Dissolve m small volume of PBE (e.g , l-10% of final volume) Sterilize through 0 22-pm syrmge filter Brmg total volume to 100 mL with PBE Add papam to 0 125 mg/mL (Sigma’s suspension IS 25 mg/mL, so pipet 0 5 mL) Allow approx 5 mm for the stock papam solution to warm at room temperature, since the refrigerated suspension of papam is quite viscous Extractmg papam from the rubber-septum sealed container should be conducted m the aseptic biosafety cabmet clean the septum with a sterile 70% ethanol wipe, withdraw the desired volume with a syrmge, and rewipe the septum

3 5.1.3 1 2 3 4

DIGESTING THE CARTILAGE EXPLANTS

Measure and record the wet and dry (lyophihzed) explant wt. Deposit each cartilage sample within a 2-mL cryogemc vial Add 0.5 mL of digestion solution to each sample Incubate m a hot-water bath for 12-16 h at 60°C Samples should be fully submerged m the digestion solution and devoid of adherent air bubbles 5 If the samples are not completely dissolved, vortex and continue mcubation

Response of Cartilage to Deformation 3 5.2 DMB Assay for Glycosammoglycan

537

Content

The dtmethylmethylene blue assay 1sa well-estabhshed method for the raped determmation of the total sulfated glycosaminoglycan (GAG) content of carttlage samples (II). Although the assay cannot dnectly quantify synthesis or catabolism (i.e., the incremental change in the total amount of GAG), It is useful for measuring the amounts released to the media or retained within the explant 3 5 2 1 ONE LITER OF DMB CHLORIDE DYE STOCK SOLUTION

This quantity will accommodate 500 samples when apportioned at 2.0 mL per sample. Scale the followmg formulatron accordingly, if larger volumes are required. The requrred materials and then final concentrations are DMB chloride (46.0 m, NaCl (40 6 mA4), glycine (40.5 m&I), sodium azide, 3.1 mM 1 iV HCl (- 10 n-k!), demmerahzed water (-95% final volume). 1 Dtssolve 0 016 g of DMB mto 10 0 mL of 100% ethanol in a closed, flat-bottomed, 50-mL centrifuge tube Add a flea magnettc sttrrer, cap, wrap m alumrnum fotl (to block out light), and star at room temperature unttl dtssolved (4-16 h) Smce DMB 1smoderately soluble in low-pH water, this procedure will enhance solub&y m subsequent steps 2 Weigh out 2.37 g NaCl and 3 04 g glyctne and add both to 950 mL deminerahzed water 3 Wearing a mask and gloves, weigh out 0.2Og sodmm azrde and add to the solutron of step 2 (Caution: Sodium azrde 1stoxic and readily absorbable through skin). 4 Add the dtssolved DMB to the solutron Rmse the 50-r& centnmge tube wtth water to collect any prectpttate Trap the flea In the 50-mL centrifuge tube wtth a stir bar 5 Titrate the solutton to pH = 3 .O(rt0 05) with 1 N HCl Typically, -10 mL of HCl are required Caution: Wear a lab coat and safety glasses when handlmg HCl 6 Bring the volume to 1 L with water. 7 Verrfy that the pH remained at 3 0 f 0 05 8 Check the absorbance at 525 nm and write the observed value on the bottle The OD should be between 0 30 and 0 38 when read against an air reference m a plastic cuvet 9 Filter the solutron with large filter paper mto a brown/amber storage bottle The storage bottle must be cleaned prior to use, because restdual prectpttates may induce further precrprtatron. To minimize the prectpttatton of the dye, do not agrtate the solution. 10 Note the date of the solutton on the bottle The dye remams stable for approx 3 mo 11. Store the solutton at room temperature, preferably m a dark cabmet 3.522.

10 ML OF 1000 ~G/ML CHONDROITIN SULFATE STANDARD STOCK

SOLUTION

1 2. 3. 4

Weigh 10 mg chondroitm sulfate Add to 10 mL demmerahzed water. Vortex and aliquot 0.5 mL into cryogemc vials Freeze until needed at or below -20°C


538

3 5 2 3 ASSESSING GLYCOSAMINOGLYCAN CONTENT 1 Prepare samples m duphcate a For samples contammg large amounts of GAG, prepare standard soluttons of serral dduttons of chondrottm sulfate 1000, 500, 250, 125, 62 5, and 0 ug/ mL (a blank of demmerahzed water) For samples contammg dilute glycosammoglycan soluttons, prepare five serial drluttons from 125 pg/mL, mcludmg a blank Vortex and add 20 uL of each dtlutton to a cuvet or microplate b Vortex the drgested cartilage samples and add a small altquot mdenttcal to that of the standard (e g., 20 pL) of each sample to a cuvet or mtcroplate 2 Pump 2 mL of DMB solution to a cuvet and record the optical absorbance at 525 nm Repeat process for the remaining cuvets Altemattvely, add 200 pL to mtcroplate wells and measure the absorbance. DMB must be added unmedtately prior to measurement, because precrprtatlon IS likely. 3 The range of acceptable measurements span 0 3 (blank) to 8 (largest standard), and duplicates should be wtthm 5% 3.5.3.

Hoescht

33258

Dye Assay for DNA Content

This biochemtcal technique provtdes an Index for radtotracer mcorporatton potential, by measuring total DNA content (12). For a given explant, the mcorporation of radiolabeled metabolites IS mediated by its resident cells Since the quantity of DNA IS related to the abundance of cells, knowledge of DNA content (1 e., cell populatton densrty) may enable the compartson of mcorporatton rates among samples with drssrmtlar cell populations.

The followmg

protocols

describe the preparation of a concentrated buffer (1OX TEN) and a stock dye solutton (the combmatton of which furnishes a working dye solutron) 3 5 3. I. 1 OX TEN BUFFER

For the preparatton of 1 0 L of 1OX Trts-EDTA-NaCl stock buffer (1 OX TEN), the requrred materials and then final concentratrons are Trls (100 m&I), EDTA (10 mM), and NaCl(1 .OM) We typically prepare 3 0 L. 1 Mix 12 10 g Trts-base, 3 72 g EDTA, and 58.44 g NaCl mto 0.95 L sterihzed, demineralized water Allow 30 mm for all the salts to completely dtssolve 2 Adjust pH to 7 40 with concentrated HCl (-14 M HCl) 3 Add water to bring the volume to 1.O L 4. Filter sterthze with 500 mL capacity 0 22-mn filter 5 Store under sterrle condmons at 4°C. 3.5.3 2 CONCENTRATED DYE STOCK SOLUTION 1 Add 10 mg of Hoechst 33258 to 10 mL of sterile demmeralrzed water 2 Store m foil-wrapped glass or brown plastic bottle at 4°C in a dark place Stable for 6 mo Note preparation date and expiration date on the bottle

539

Response of Cartilage to Deformation 3 5 3 3 WORKING DYE SOLUTION 1 Prepare a 1X TEN (10 mMTrls-HCI, mg 1OX TEN 1.10 Take precautions 2 Add dye stock solution to 1X TEN First, shake the dye stock solutlon, 100mLof 1XTEN

1 WEDTA,

0. IMNaCl)

solution by dllut-

to ensurethe sterlhty of the stock 10X TEN to a final dye concentration of 0 1 pg/mL then combme 10 pL of dye solution and

3.5 3 4. ASSESSING DNA CONTENT 1 Vortex the dlgested samples and dispense a small ahquot (-50 pL) m duplicate mto cuvets 2 Pump 2 mL Hoescht dye solution mto each cuvet. This solution IS stable 3 Measure fluorescence and compare with standards

3.54. Radioactwe Labehng The blologlcal response of cartilage to mechamcal stress IS most commonly momtored by introducing radioactive precursors to the culture medium Presumably, the rate of mcorporatlon of radiolabeled metabolites into structural macromolecules IS related to the rate of biosynthesis. Since sulfate IS mcorporated mto glycosammoglycans and amino acids mto proteins, the uptake of [35S]sulfate and trltlated ammo acids from the culture medium allows for dual momtormg of newly synthesized glycosammoglycans and proteins, respectively (see Note 13). The mcorporatlon of both P-emitting radiolabels can be observed quantltatlvely through liquid scintillation counting, or spatially through autoradiography (3,9). Radiolabel concentrations m the culture media should be on the order of 5-50 @i/mL for both the [35S]sulfate (Na13%04) and tritlated ammo acids (l-[5-3H]prolme). To achieve at least 3000 counts/mm (which should be sufficiently greater than background) we typically use 20 @l/mL [35S]sulfate and 10 $i/mL [3H]prolme for 8 h labeling of 2-4mm diameter, I-2-mm thick disks. These concentrations can be varied accordmgly for substantially longer or shorter labeling times. Control studies should be performed to determine appropriate radiolabel concentrations. 35 1 2 3 4 5 6

4 1 2 M GUANIDINE HCL SOLUTION

19 1 06 g guamdme hydrochloride 68 04 g sodium acetate. MIX with 900 mL water Adjust pH to 6 8 with 1 M HCl Bring volume to 1 L Store at room temperature.

3.5.4.2. SCINTILLATION COUNTING 1 Allquot 40 p.L of digested cartilage or media samples mto vials m duphcate 2 Ahquot 10 rJL of the ongmal radloactlve media and 30 & of water into vlals m duplcate.

540

Allen,

Eisenberg,

and Gray

3 Add 460 ,uL of guamdme HCl and 2 mL of Ecolume to each vial 4 Cap and vortex all vials. 5 Place vials m scintillation counter Watt approx 1 h before measurmg the chemtlummescence to allow any residual fluorescence to extmgutsh, 6 Adjust altquot stzes to yteld counts of at least 3000 counts/mm

4.

Notes 1 Stress relaxation m cartilage 1s the transient phystcal process of matrix and flutd redtstrtbutton m response to applied displacements It 1s momtored mdtrectly by observing the load (Fig. 4) Immedtately after applying the displacement, the load IS maximum As flutd is redtstrtbuted, the load reduces, until leveling off at a new steady value needed to mamtam the tissue at the imposed displacement Stmtlarly, creep is the transtent response to applied loads Durmg stress relaxatton and creep, fluid transport effects will be present The ttme-course for stress relaxation, the process relevant to methods described here, as descrtbed by Armstrong’s model of unconfined compressron, 1s a function of the disk’s radms, a, the equthbrmm confined-compresston modulus, HA, and the hydraulic permeabtltty, k (I). The stress relaxation time constant, to first order approxtmattons, IS defined as: z = a2/{HAk} (as a rule of thumb, the creep time 1s approxtmately 4X longer) Typical values for HA and k m arttcular carttlage are 0 5 MPa, and 3 0 x IO-t5m4/(Ns), respectively For a 4-mm-dtameter sample, for example, one would expect a characteristic stress-relaxation ttme of -45 mm, note that halvmg the radius would decrease this time fourfold 2 To heat-macttvate serum, Incubate m a hot-water bath at 56’C for 30 mm Swtrl the serum at least twice during the mcubation period to avoid gelatton of serum proteins Dispense 0 5 mL of sera into cryopreservatton vials and freeze at -20°C Thaw at room temperature when needed 3 Our vise 1s essentially a stainless steel tube attached to a universal Joint that 1s mounted m a custom-built acrylic tray The shaft of the long bone 1s Inserted mto the tube and tmmobtltzed by tightening the three set screws The umversal Jomt allows for ortentmg the carttlage surface beneath the droll bit The tray serves as a reservoir for collecting spent rmsmg solution 4 GAG synthesis after explantation is highly variable (13-15), we have determmed that GAG synthesis is stable after 2 d of culture (m free-swelling condmons) (16) However, others have noted longer pertods, albeit for adult ovme ttssue (17). Wattmg for the stabthzatton of the GAG (and protein) synthesis IS important for ascribing any observed btologtcal vartattons m synthesis to mechamcal sttmulation 5 Compresston posts should be made of a material that is several orders of magmtude stiffer than cartilage (1 e., modulus of >lOO MPa). 6 This parttcular configuratton allows each carttlage sample to be cultured m Its own well A variety of other devices have been constructed by our lab and other labs The interested reader 1sencouraged to contact the authors tf they wash asststance in obtammg or designing a compresston chamber


9

10

11

12.

13.

541

For load-controlled static compression, mdividual weights can be fashioned from lead (or contamers of lead shot) so that they can rest on top of the compression post The stress applied to the tissue IS the product of the mass (as measured on a balance) and 9 8 1 m/s2 (acceleration because of gravity) divided by the surface area of the disk (e.g , to achieve 250 kPa-a stress that provides a nominal 20% compression-for a 3-mm-diameter disk, a mass of 180 g IS required) The size of the weights can make the approach difficult, especially for larger diameter samples. When the compression system estabhshes contact with the cartilage explant, an mitral displacement is imposed, which the load cell mstrumentatlon will Indicate as an offset or tare load The tare load should not exceed 5 g The effect of the tare load upon the slope of load-displacement relatronshrp (I e., sttffness) of the cartilage is assumed to be negligible Moreover, it is assumed that the tare load elicits a negligible mmal compressive displacement, therefore a 0% macroscopic strain may be defined at the pomt of mltial contact (this defimtion of strain assumes the reference length to be the swelled sample thickness as measured immediately prior to mechanical loading) All load-displacement behavior is attributed solely to the cartilage, because the stiffness of the compression system (m the absence of cartilage) ranges from 200 to 400 kN/m, whereas a 4-mm-diameter by 2-mm-thtck cartilage disk generally exhibits an equihbnum stiffness of approx 3 0 kN/m. The dynamic stiffness of cartilage increases dramatically with frequency in the range of 0 05 to 1 0 Hz, therefore, system compliance may introduce an artifact m determmmg the stiffness of cartilage Compression levels (1 e , macroscopic strain) of 0, 10, 20, and 50% for a tissue sample that was mitially 800~pm thick correspond to final compressed thicknesses of 800, 720, 640, and 400 pm, respectively. In stating the macroscopic strain, it is important to also state the reference thickness (e g , cut thickness, or swelled thickness, measured unmediately prior to mechanical testmg) Other digestion procedures can be used protemase K, hydrolysis, and so on We use papam, which is suitable for the most commonly used biochemical assays listed here Different chondroitm sulfate standards yield slightly different absorbances for the same concentration Therefore, it is important to use the same chondroitm sulfate standard (preferably the same lot) if it is desirable to compare across experiments. In order to interpret radiolabel mcorporation as reflecting synthesis, it is important to establish that mcorporation is linear with time, and varies appropriately as medium concentrations of cold label (nonradioactive precursor) are changed Usmg common media formulations does not appear to pose problems for sulfate mcorporatton m any cartilage system Others have shown that for the addition of 0.4 mM nonradioactive prolme to DMEM, the specific activity of prolme mcorporation appears to reflect the specific activity of the medium Control experiments should be performed, and data m the literature should be consulted for other ammo acids

542


References 1 Armstrong, C , Lat, W., and Mow, V (1984) An analysts of the unconfined compresston of arttcular cartilage J Bromech Eng 106, 165-173 2 Kim, Y., Bonasser, L , and Grodzmsky, A (1995) The role of carttlage streaming potential, fluid flow and pressure in the stimulation of chondrocyte biosynthesis during dynamic compression J Bzomech 28,1055-l 066 Kun,Y, Sah, R , Grodzmsky, A , Plaas, A., and Sandy, J. (1994) Mechamcal regulation of cartilage biosynthettc behavtor physical sttmuh. Arch Bzochem Bzophys 311, 1-12 Hall, A., Urban, J , and GehI, K. (1989) The effects of hydrostatic pressure on matrix synthesis in articular cartilage. J. Orthop Res 9, l-10 Maroudas, A and Bannon, C (198 1) Measurement of swelling pressure in cartilage and comparison with the osmottc pressure of constituent proteoglycans Bzorheology l&619--632 6 Maroudas, A , Mtzrahi, J , Ben Hatm, E and ZIV, I (1987) Swelling pressure m cartilage Adv Mzcroczrc 13,203-2 12 7 Maroudas, A , Wachtel, E , Grushko, G., Katz, F , and Wemberg, P (1991) The effect of osmottc and mechamcal pressures on water partmonmg m arttcular cartilage Bzochzm Bzophys Acta 1073,285-294 8 Schneiderman, R , Keret, D., and Maroudas, A (1986) Effects of mechamcal and osmotic pressure on the rate of glycosammoglycan synthesis m the human adult femoral head cartrlage an m vitro study. J Orthop Res 4, 393-408 I 9 Qumn, T (1996) Arttcular cartilage matrix assembly, mediation of chondrocyte metabolism, and response to compression PhD, MIT 10 Oegema, T , Carpenter, B , and Thompson, R (1984) Fluorometnc determmahon of DNA m carttlage of various species J Orthop Res 1,345-35 1 11. Farndale, R., Buttle, D., and Barrett, A. (1986) Improved quantitatton and dtscrimination of sulphated glycosammoglycans by use ofdimethylmethylene blue Bzochzm Bzophys Acta 173-177 12. Ktm, Y., Sah, R , Doong, J , and Grodzmsky, A ( 1988) Fluorometric assay of DNA m cartilage explants using Hoechst 33258 Anal Bzochem 174, 168-176 13 Hascall, V, Handley, C , McQuillan, D , Hascall, G , Robmson, H , and Lowther, D. (1983) The effect of serum on blosynthests of proteoglycans by bovine arttcular cartilage m culture. Arch Bzochem Bzophys 224,206-223 14 Lane, J. and Brtghton, C (1974) In vitro rabbit articular cartilage organ model I Morphology and glycosammoglycan metabolism Arthrztzs Rheum 17,235 15. McKenzie, L , Horsburgh, B., Ghosh, P., and Taylor, T (1977) Organ culture of human arttcular cartilage* studtes on sulphated glycosammoglycan synthesis In Vitro 13,423-428 16. Gray, M. L., Ptzzanelh,A M., Grodzmsky,A. J., and Lee, R C (1988) Mechanical and physicochemical determmants of the chondrocyte biosynthetic response J Orthop Res 6,771-792 17 Torzilh, P A , Grtgtene, R , Huang,C , Friedman, S M , Doty, S B , Boskey, A L , and Lust, G (1997) Characterization of the carttlage metabolic response to static and dynamic stress using a mechanical explant test system J Bzomech. 30, l-9

40 Measuring Receptor-Mediated Under Flow

Cell Adhesion

Cell- Free Sys terns Daniel A. Hammer and Debra K. Brunk 1. Introduction Leukocytes must bmd to vascular endothelium under condttions of flow to perform then appropriate phystologtcal functions, which require traffickmg mto and out of the ttssue space surrounding blood vessels Traffickmg mto tissues 1srequired of neutrophlls during the acute inflammatory response, and durmg trafficking of lymphocytes mto lymphotd tissue. Egress from blood lumen to tissue involves a series of adheston-dependent steps, each of which involve different leukocyte adheston receptors and counterreceptors on the endothelium. Transtent adhesion, or rollmg, of leukocytes over endotheltal cells IS a prerequisite to firm attachment and transendotheltal mtgration (2,2). Many of the molecules medlatmg rollmg are m the selectm adhesion molecule family, conststmg of three selectms, L-, P-, and E-selectm Selectms are selective lectms that bmd to carbohydrates A current research focus m cellular immunology is to determine which counterreceptors bind to these selectins. In static assays,tt has been shown that all three selectmsbmd stalyl Lew@ (sLeX), a stalyated, fucosylated carbohydrate, which is widely distributed on both glycoprotem and glycoliprd components of the neutrophil(3,4), and whtch 1sconstitutively expressed by human umbiltcal vem endothelial cells (5-s). For P-selectm, the phystologtcal hgand, P-selectm glycoprotem hgand- 1, has been tdentrfied This hgand bears sLex, but also contams other critical features that are required for recogmtton, including several tyrosmes that can be, and are, functional when sulfated Thus, the exact contrtbutron of sLeX to recogmtton 1s not clear. The lrgands for the other selectins, E- and L-selectin, have not been From

Methods Edlted

by


Me&me, and

Vol 18 Tmue

M L Yarmush

543

Engmeenng

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

544

Hammer and Brunk

conclusively demonstrated, and thus the importance of sulfated, fucosylated carbohydrates in their recognition is also not clear. Also, cells containing proteins bearing sLeX do not always support rolling interactions. Thus, although static assays may propose putative receptor-ligand pairs, they do not provide information on how sLeX interacts with the selectins under dynamic conditions (i.e., can these molecules mediate rolling?). An additional confounding aspect of rolling is the role of cellular features, such as clustering of receptors on microvilli and cell deformability and intracellular signaling. It would be interesting to know if sLe%+electin interactions have the requisite properties of recognition to mediate rolling in the absence of cellular features. Indeed, Pierres and coworkers have suggested that the initial formation of bonds may be a passive process, which could be studied using inert substrates as carriers for ligands (9). Therefore, we have developed a cell-free system to determine whether binding between sLe” and E-selectin can mediate rolling under flow (10,11). Such a clean system will allow us to identify receptorligand pairs that mediate rolling, and to better understand the functional properties of receptor recognition required for dynamic adhesion. This system can reproduce the dynamics of rolling adhesions seen with leukocytes over endothelium in vitro, both in average rolling velocity and in the fluctuations in velocity (10,11). Such a system allows us to unambiguously determine which carbohydrate ligands support rolling on which selectin adhesion molecules, and provides a clean method for establishing the blocking interaction of antagonists targeted to these molecules. This chapter explains how this cell-free system is produced, characterized, and tested. 2. Materials 2.1. Chemical reagents 1. NeutrAvidin (Pierce, Rockford, IL). 2. Biotin (Sigma, St. Louis, MO). 3. FITC-biotin (Molecular Probes, Eugene, OR).

4. Biotinylated-(sLex)4(Syntesome,Munich, Germany;Glycotech,Rockville, MD). 5. Monoclonal

antibody (MAb) to sLeX (KM93,

IgM, Kamiya Biomedical,

Thou-

sandOaks,CA). 6. FITC-labeled

mAb anti-IgM (Pharmingen, San Diego, CA).

7. Quantum26 calibration microspheres(Flow CytometryStandards,SanJuan,PR). 8. E-selectin-IgG chimera(E-selectin-IgG),a gift from Brian Brandley (Rush Medical Center, Chicago, IL) (12).

9. Silanatedglassmicroscopeslides (Sigma). 10. Flexiperm wells (Heraeus Instruments,

South Plainfield, NJ).

11. Anti-E-selectin (68-5H 11,IgGi, Pharmingen). 12. FITC antimouse IgG, directed at the Fc portion of IgG (Sigma).

Receptor-Mediated

Cell Adhesion

545

Fig. 1. Top view of flow channel, illustrating dimensions of the tapered channel. Units are in inches. Template is cut from reinforced Duralastic sheeting, 250 m thick. The width of the channel is dependent on the axial placement along the length of the channel. 13. Human IgG, (Sigma). 14. Duralastic sheeting (Allied Biomedical, Goose Creek, SC). 15. Carbodiimide [ 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide (Aldrich, Milwaukee, WI).

hydrochloride]

2.2. Equipment 1. 2. 3. 4. 5. 6.

Photoscan (Nikon, Garden City, NJ). Nikon Diaphot inverted microscope (Nikon). 67s video camera (Dage-MTI, Michigan City, IN). S-VHS video tape recorder (JVC, Elmwood Park, NJ). Infusion-withdrawal syringe pump (Harvard Apparatus, South Natick, MA). Computer for image analysis (Apple Macintosh, Cupertino CA, with Scion LG-3 frame grabber board, running NIH Image software).

2.3. Flow Chamber The tapered channel

design is based on the theory of Hele-Shaw

flow

between two parallel plates (13). In the theory of this design, the wall shear stress,‘I: is given 7:= (6pQ/h2wl) (1 -z/L), where p is viscosity, Q is flow rate, h is the channel height, w1 is width, L is length, and z is the axial position. The plates are separated by 250~pm Duralastic sheeting, which when compressed results in a thickness of h = 175 p (thus, h = 175 pm). The chamber is typically constructed in a machine shop from Lexan polymer. A design drawing for the Hele-Shaw template used for our (IO) is shown in Fig. 1. 2.4. Apparatus Experiments take place in a flow chamber on an inverted Nikon microscope. Two separate inlets lead to the chamber through a Y-connector, one from a syringe pump and the other from a hand-held syringe, from which beads and antibodies can be injected. The outlet of the chamber is equipped with a

546

Hammer and Brunk

Y-connector also, one exit leading to waste, the other to a buret for measuring flow rate The events m the flow chamber are captured by a video camera and recorded on videotape for later analysis. Image analysis 1s performed by an Apple Macmtosh computer with a video frame-grabber board and NIH Image software Ample hard-disk storage 1s needed (several Gbytes recommended) for storing Images

3. Methods 3.7. Microsphere Preparation 3.1 1. Attach NeutrAwdin to MIcrospheres NeutrAvldm was coupled to carboxylated polystyrene mlcrospheres usmg carbodumlde chemistry (24) The protocol was based on a protocol supphed by Polysclences (Warrington, PA). The detalled protocol is as follows. 1 Place 222 2 p.L of 2 5% suspension of carboxylated microspheres (Polysclences, Warrmgton, PA) mto a 1 5-mL lube tube. Add 1 mL carbonate buffer (0 1 M, pH 9 6) to bead suspension, and vortex Centrifuge for 3 mm at 6000 rpm (-2000g) Remove supernatant carefully usmg a plpet, and discard supernatant Repeat steps 2-4 Add 1 mL phosphate buffer (0 02 M, pH 4 5) and 2 5 pL Tween-20 (diluted 1 10 m phosphate buffer) to bead pellet and vortex 7 Centrifuge for 3 mm at 6000 rpm, remove supernatant, and discard 8 Repeat steps 6-7 twice 9 Add 0 3 mL phosphate buffer to bead pellet, and vortex 10 Mix 0 35 mL of a 2% solution of carbodllmlde [l-(3-dlmethylammopropyl)3-ethylcarbodumlde hydrochloride] (Aldrich, WI) Prepare carbodllmlde no sooner than 15 mm prior to use. I1 Add 0 3 mL carbodnmide solution dropwlse to bead suspension 12 Mix for 3.5 h at room temperature with end-to-end mixing 13. Centrifuge for 3 mm and discard supernatant (hazardous waste) 14 Add 1 mL borate buffer (0 2 M, pH 8 5) to bead pellet, vortex, and centrifuge for 3 mm at 6000 rpm (2OOOg) Discard supernatant 1.5 Repeat step 14 twice 16 Add 453 $ borate buffer and 3 pL Tween-20 (diluted 1 10 m borate buffer) to bead pellet and vortex 17 Add 144 p.L of 1 mg/mL NeutrAvidm to bead suspension 18 Leave overnight at room temperature with end-to-end mlxmg 19. Next day, centrifuge suspension for 10 mm 20 Remove supernatant and discard 21 Resuspend pellet m 0 6 mL 0 1 M ethanolamme m borate buffer and Incubate 30 mm at room temperature with end-to-end mixing

547

Receptor- Media ted Cell Adhesion

22. Centrifuge for 3 mm at 6000 rpm (ZOOOg), remove supernatant and discard 23 Resuspend bead pellet m 0.6 mL PBS+ (PBS, pH 7 4, 1% BSA, 1 mM CaQ, 1 m&I MgC&, sterile-filtered) and incubate for 1 h at room temperature with end-to-end nuxmg. 24. Wash twice with PBS+ 25 Resuspend bead pellet m 0 6 mL PBS+, vortex, and store m refrigerator at 4’C prior to use

3.7 2. Attaching B/otrny/ated-Carbohydrates Biotin-linked carbohydrates (Rockvllle, MD).

are commercially

to Spheres. available

from GlycoTech

1 Remove 1O6NeutrAvidm-coated microspheres for flow cytometry 2 MIX the required concentration of biotm-sLeX m PBS+ A concentration greater than 0 I mg/mL ~111 saturate the NeutraAvidm-coated microspheres 3 Centrifuge NeutrAvidin-coated microspheres for 3 mm, remove and discard supematant 4 Wash NeutrAvidm-coated microspheres once with PBS+ 5 Add 50 mL of biotm-sLeX solution to microshere pellet, vortex, and mcubate for 45 min, with occasional vortexmg 6 Centrifuge microspheres for 3 mm; remove and discard supernatant 7 Wash spheres once with PBS+

3.1.3. Preparing Spheres for Flow Cytometry After bmdmg sLeX to the spheres, the density of sLeX can be determined with flow cytometry.

4 5 6. 7 8 9

10

Prepare 30 ClgimL solution of anti-sLeX antibody, such as KM93, IgM (Kamiya Biomedical, Thousand Oaks, CA) m MES+ buffer Wash biotm-sLeX spheres twice m 50 n&f MES+ (pH 5.5, containing 1% BSA, 1 mM CaC12, 1 mA4 MgCl*) sterile filtered Add 50 & of KM93 solution to biotm-sLeX microspheres, and incubate for 30 mm, with occasional vortexmg. Centifuge for 3 mm, remove and discard supernatant MIX 60 pg/mL solution of FITC-labeled anti-IgM mAb (Pharmingen) m 50 n&fMES+. Add 50 mL of FITC-mAb to mtcropheres, and incubate 30 mm m dark, with occasional vortexmg Repeat step 4 Wash once m MES+, once m PBS +, and once m PBS+ (-BSA). Add 300 mL PBS+ (-BSA, +l% formaldehyde) to mtcrospheres, and store m refrigerator until flow cytometry is performed Determine the fluorescence of these beads m a flow cytometer. ConversIon of fluorescence peaks to site densities was done using Quantum 26 calibration microspheres (Flow Cytometry Standards) by comparing the fluorescence of the measured beads to that of the standards, with the additional mformation of the F*P ratio (the number of fluorescent probe/m01 of protein), and an assumption about the number of secondary antibodies that binds per primary antibody (usually 1 1)

Hammer and Brunk

548 3.2. Selectin Substrate

Preparation

1 Incubate slides m PBS overnight 2 Incubate E-selectm-IgG on silanated glass microscope slides (Sigma) in Flextperm wells (Heraeus Instruments, South Plamtield, NJ) at concentrations from 0 1 to 1 6 pg/mL for 2 h 3 Wash slides twice wtth PBS 4 Incubate m PBS+ (1% BSA, 1 mMCaC12, 1 mM MgCl2) for 1 h to block nonspecific adhesion

3.3. Determination

of surface density of E-selectin.

In addttron to steps 1-3 tn Subheading 3.2., perform the following: 1 Add 20 pg/mL anti-E-selectm (685H11, IgGl, Pharmmgen) m PBS+ to wells for 30 mm 2 Add 30 pg/mL Fc-specific FITC antimouse IgG (Sigma) m PBS+ to wells for 30 mm 3 Prepare a negative control using the same method, replacing the E-selectm-IgG with human IgG 1 4 Measure fluorescence m counts/s with a Photoscan (Ntkon) attached to a Ntkon Diaphot inverted mtcroscope, and then convert to site densities using a caltbranon curve constructed from known concentrattons of fluorescent IgG

3.4. Flow Chamber

Calibration

In this chamber, the width varies as a functron of axial posttton down the length of the chamber. I Assemble the chamber with a silanated glass slide m PBS 2 Attach a guide on the mtcroscope stage, consisting of equidistant ticks. 3 Place a field-finder microscope slide (Fisher) mto the chamber 4 Use the manual control of stage position to relate the distance m the slide to dtstance on the guide 5 At each axial posttion, use a rettcle or photo mask to measure the width of the chamber, and compare the width measurements to that theorettcally predicted by the design equation This requires scrolling across the field of view several muthples of the reticle dimension to span the entire width 6 Place chamber on microscope stage and measure the height of the channel, using the fine verticle focus 7 Perfuse PBS through the chamber, and measure the flow rate 8. Suspend uniform polystyrene mtcrospheres (10 pm in diameter) m PBS, and mJect microsphere suspension mto perfusion inlet 9. Focus ObJective on bottom plate 10 Record motion of microspheres usmg video camera and videotape recorder Results of flow chamber calibration are shown m Fig. 2

3.5. Adhesion Experiments E-selectm-IgG-coated slides were placed m the well of the flow chamber, the chamber was assembled m PBS, then placed on the microscope stage. For a

Receptor-Mediated

Cell Adhesion

8 0

0

549

B3xkiJcrty=omcm~sig*~~

I

I

1

I

0.5

I

1.5

2

S&XS,F?=O87 I I

2.5

3

I

I

3.5

4

wshearm(*d) Fig 2 Calibration of Hele-Shaw flow chamber, showing bead velocity near the plate surface as a function of wall shear stress. Wall shear stress was vaned either by changing the flow rate, Q, by observing at different axial positions, z Velocity is proportional to axial position, as predicted by theory n , z = 3 cm, varying Q; 0, z = 5, varying Q, A, Q = 134 &/mm, varying z channel height of 175 um, the perfuston buffer or microsphere suspension flow rate must be 128 &/mm to obtain a range m wall shear stress from 0 to 4 dyn/ cm2 down the length of the channel. The chamber 1s perfused with PBS (+l mA4 CaC12, 1 mM MgC12) for 15 mm, then the microsphere suspension (5 x 105/mL) IS Introduced. Data is collected by stepping down the chamber from inlet to outlet m 0 S-cm steps, allowing about 1 mm between steps. Mtcrosphere mteractton wtth the surface IS recorded at a total magmficatton of x200 for future analysts All expertments are done at 23°C

3.6. Antibody

Blocking

and Control Experiments

Antrbody blockmg and antibody control experiments are necessary to demonstrate spectfictty 1 E-selectin-IgG-coated slides are incubated for 45 mm in 20 pg/mL anti-Eselectm (BBA2, IgG2, R&D Systems, Minneapolis, MN), whrch should block

Hammer and Brunk

550

80

“E 2

70M Z 9 60-5 ,E 502 & 40e P 30< 10 200 1 0

I

ohs

. , I , 0’1 0 15 0’2 025

/ 03

, 035

( ( 0.4 045

, 05

[btotm-sLexlbUlk @g/ml) Fig 3 Average sLeX molecules per bead area as a function of solution concentration of biotm-sLeX, as achieved by varying the ratio of biotm-sLeX to blotm-LeX m solution, keeping the total concentration of biotm-carbohydrate constant at 0 5 pg/mL

binding, or with anti-vascular cell adhesion molecule- 1 (VCAM-1) (5 1 - lO”C9, IgG1, Pharmmgen) or anti-P-selectm (AK4, IgG,, Pharmmgen), which are positive controls. 2 After antibody mcubation, slide is inserted mto the chamber 3. Add 1 pg/mL of same antrbody to the perfusion buffer and the microsphere solution 4 Perform experiment as normal

3.6. Data Analysis 1 To determine the particle flux, the number of transient interacttons in a 0 03-mm2 area were manually counted as a function of time Transient attachment included all particles that interacted with the surface, 1e , rolhng and stopstart mteractlons Particles were considered firmly attached if they remamed stationary for > 10 s 2 For rollmg cells, velocity measurements are obtained from recorded data using a SCION LG-3 frame grabber m a Power Macmtosh 7100 runnmg NIH Image,

Receptor-Mediated

Cell Adhesion

551

pubhc domain Image analysts software. Particle coordmates were obtamed with 1 pixel accuracy (1 pm) by chckmg the mouse at the center of a rollmg microsphere Average velocity was calculated by divldmg the total displacement of a rollmg particle by the observation time Instantaneous velocity was determined by dividing the displacement of a rolling particle by the time between Incremental captured frames.

4. Notes 1 NeutrAvldm 1s an altered form of avldm that has had all of its surface carbohydrate groups removed, thus removing any nonspecific binding, and lowering the lsoelectrlc point to 7, compared to 9 for avldm It 1s thus easier to work with, and displays less nonspecific bmdmg induced by electrostatics 2 Bead lot varlablhty within single vendors 1s large, so stick with a lot that works In addltlon to the Polysclences beads, we also found that beads from Bangs laboratories (Fishers, IN) worked well 3. Make sure PBS+ is as fresh as possible 4 One can reduce the amount of blotm-slex on the bead surface by co-incubating it with a neutral blotmylated carbohydrate, such as blotm-LeX in different ratios m which the total concentration 1s kept at a value >O. 1 Clg/mL Figure 3 Illustrates changes m slex-bead surface density one can achieve with this method 5 The accuracy of our determmatlon of sLeX surface density was greatly increased by using an IgG, agamst sLeX (provlded by Am1 Singhal, Blomembrane Institute, Seattle, WA)

References 1 von Andnan, U. H , Chambers, J D., McEvoy, L M , Bargatze, R F ,Arfors, K -E , and Butcher, E C (199 1) Two-step model of leukocyte-endothehal cell interaction m mflammatlon. distinct roles for LECAM-1 and the leukocyte pZ mtegrms m VIVO. Proc Nat1 Acad Scz USA 88, 7538-7542. 2 Lawrence, M. B and Springer, T. A (1991) Leukocytes roll on a aelectm at physlologlc flow rates dlstmctlon from and prerequisite for adhesion through mtegrins Cell 65,859-873 3 Fukuda, M , Spooncer, E , Oates, J E , Dell, A , and Klod, J C (1984) Structure of slalylated fucosyl lactosammoglycans isolated from human granulocytes J Blol Chem 259,10,925-10,935 4. Symmgton, F. W., Hedges, R , and Hakomer, S I. (1985) Glycohpid antigens of human polymorphonuclear neutrophlls and the inducible HL-60 myelold cell lme J Immunol 134,2498-2506 5. MaJuri, M -L., Pmola, M., Nlemela, R , Tiisala, S , Natunen, J , Renkonen, O., and Renkonen, R (1994) 2,3-slalyl and 1,3-fucosyltransferase-dependent synthesis of slalyl Lewis x, an essential ollgosaccharlde present on L-selectm counterreceptors, m cultured endothehal cells Eur J Immunol. 24, 3205-3210 6 Polley, M. J , Phillips, M L , Wayner, E ,Nudelman, E , Smghal, A K , Hakomon, S. I , and Paulson, J C (199 1) CD62 and endothehal cell-leukocyte adhesion molecule 1 (ELAM-1) recognize the same carbohydrate ligand, slalyl-Lewis x. Proc Nat1 Acad Scl USA 88,6224-X5228

Hammer and Brunk

552

7 Foxall, C , Watson, S. R , Dowbenko, D , Fennte, C , Lasky, L A , KISO, M ,

Hasegawa, A , Asa, D , and Brandley, B IS (1992) The three members of the selectm receptor famtly recognize a common carbohydrate epttope, the stalyl Lew# ohgosacchande J Cell Btol 117,895-902 8 Berg, E. L , Magnam, J , Warnock, R A, Robmson, M K , and Butcher, E C (1992) Comparison of L-selectm and E-selectm hgand spectficmes the L-selectm can bind the E-selectin hgands Blochem Blophys Res Common 184,1048-1055 9 Pterres, A., Tlssot, 0 , Mahssen, B., and Bongrand, P. (1994) Dynamtc adhesion of CD8-posmve cells to anttbody-coated surfaces the mmal step IS mdependent of mtcrofilaments and mtracellular domains of cell bmdmg molecules J Cell Bzol 125,945-953

10 Brunk, D K , Goetz, D J , and Hammer, D A (1996) Stalyl LewtsVE-selectmmediated rolling m a cell free system Biophys J 71,2902-2907 11 Brunk, D K. and Hammer, D. A. (1997) Quanttfymg rolling adhesion with a cellfree assay E-selectm and its carbohydrate hgands Bzophys J. 72, 2820-2833 12 Watson, S. R , Imat,Y, Fenme, C , Geoffroy, J S , Rosen, S D , and Lasky, L A (1990) A horning receptor-IgG chimera as a probe for adhesive hgands of lymph node endothehal venules. J Cell B~ol 110,222 l-2229 13 Usamt, S., Chen, H.-H., Zhao, Y, Sheen, S., and Skalak, R. (1993) Destgn and constructton of a lmear shear stress flow chamber Ann Blamed Eng 21,77-83 14 Kuo, S C and Lauffenburger, D A (1993) Relationshtp between receptor/hgand bmdmg affinity and adheston strength Bzophys J 65,2 19 l-2200

In Vitro and In Vivo Quantification of Adhesion Between Leukocytes and Vascular Endothelium Rakesh K. Jain, Lance L. Munn, Dai Fukumura and Robert J. Melder 1. Introduction When a leukocyte enters a blood vessel, it may continue to move with flowmg blood, collide with the vessel wall, adhere transiently or stably, and finally extravasate (I). These interactlons are governed by both local hydrodynamic and adhesive forces. The former are determined by the vessel diameter, fluld velocity, viscosity, and hematocrit, and the latter by the number, strength and kinetics of bond formatlon between adhesion molecules, and by surface area of contact (I-6). Cellular deformablhty affects both types of forces (7-9) Two famdles of cell adhesion molecules (CAMS) are involved m leukocyte rolling and stable adhesion. In general, the selectms (P, L, and E) mediate rolling, while the IgG superfamily members (ICAM-I and VCAM-I) on endothelral cells, with their cognate receptors (p2 and /3, mtegrm receptors) on the leukocytes, mediate firm adhesion, with some overlap m these functions (10-12). The expression of CAMS on the endothellal cells and leukocytes can be modulated by cytokmes secreted by a variety of cells (e.g., cancer cells, fibroblasts, macrophages) (13,14). Cellular deformabllity can be modulated by altering the cytoskeleton, membrane, or cytoplasm, with the cytoskeleton playing the domlnant role (7,15,16) In this chapter, we describe methods to quantltate cellular deformablhty m vitro, CAM expression m vitro, leukocyte-endothehal mteraction (LEI) m vitro, and LEI m vlvo 2. Materials 2.1. Cellular Deformability In Vitro The materials required for this procedure depend on the culture requirements of the cells to be used m the procedure, and on the experimental objective of From


Me&one, and


M L Yarmush

553

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

554

Jam et al

the testmg procedure. materials are needed

In general,

aside from the equipment,

the followmg

1 Culture medium, such asRPM1 1640 for lymphocytes (Medtatech, Fisher Scientific, Pittsburgh, PA) 2 Cell population. 3 Glass capillary stock (Sutter Instrument, Novato, CA) 4 Inverted microscope (Nikon Diaphot, Garden City, NY) 5 Camera (Panasomc,Secaucus,NJ) 6 Video cassetterecorder (Panasomc). 7 Time code generator (WJ-8 10, Panasonic) 8 Video contrast enhancer(Model 605, Colorado Video, Boulder, CO) 9 Momtor (PM205A, Ikegami, Utsunomiya, Japan) 10 Pressureapphcation system damping chamber, three-way manifold, two water reservoir bottles, connecting tubes and a fast responsesolenoid valve (type 124, Burket, Orange, CA) 11 Mrcropipet aspiration system micropipet (4-7 mu od and 2-4 pm id), micromampulator (M0202, Narishige, Greenvale, NY), temperature-controlled observa-

tlon system (such as the MIcrowarm Plate, Schlueter Instrument, Boulder, CO) 12 Transducer (precision +l O%, Gould, Valley View, OH) 13 Digital display and chart recorder (or an A/D converter and computer).

2.2. Adhesion 1 2 3 4 5 6 7 8

Molecule Expression

Surface-dependent cells 2 x 4 microwell plate (Nunc, Labtek, Naperville, IL) Matrix component (s) for coatmg surface of plate (e g , Sigma, St LOUIS,MO) Paraformaldehyde. 2 mg/mL m PBS (Sigma) Bovme serum albumm 0 1% m PBS (Sigma, Hybrnech) Antibodres primary, secondary, control Propidmm iodide (Sigma) Mountmg buffer glycerol PBS (1 9) with 1 mg/mL paraphenylenediamme (Aldrich, Milwaukee, WI).

2.3. L-E Interactions 1 2. 3 4 5 6 7 8 9 10

In Vitro

In Vitro

Histopaque (Sigma) L-phenylalanme methyl ester (E I. du Pont de Nemours, Wilmmgton, DE) Antibodies CD19 (clone J4 119), CD16 (B73. l), CD1 5 (clone MMA) Paramagnetic beads(Advanced Magnetics, Cambridge, MA) Calcem acetoxymethyl ester (Molecular Probes,Portland, OR) Endothehal cells (ClonetIcs, San Diego, CA) EGM medium (Clonetics) Fibronectm (Sigma) Variable speedsyrmge pump (Harvard Apparatus, South Natick, MA) Hanks’ balanced salt solution (HBSS; Sigma)

Leukocyte-Endothelial

Adhesion

555

11. Parallel-plate flow chamber (Machine Shop, Department of Radiation Oncology, MGH, Boston, MA)

2.4. In Vivo L-E Observation Ammals. mice (30-35 g body wt), rats (200-250 g body wetght). Anesthesia (ketamme and xylazme mixture): For mice, mtx 1 mL of ketamme HCI (Ketalar 100 mg/mL, Parke-Davis, Morrts Plains, NJ) and 100 pL of xylazme (Xyla-Ject 100 mg/mL, Phoenix Pharmaceutical, St Joseph, MO), and make up to 10 mL wtth 0 9% sterile sodium chloride, for rats, mix 1 mL of ketamme and 100 pL of xylazme, final dose will be 10 mg ketamme/l mg xylazme/ 100 g body wt (mice) and 9 mg ketamme/O 9 mg xylazine/lOO g body wt (rats) Cannulatton tube. Cut polyethylene tubing (PE-IO, Becton Dtckmson, Sparks, MD) at 12 cm length, break off needle part of 30-gage needle (Becton Dtckmson) by folding several ttmes wtth needle holder, insert base end of this needle to PE10 tube, and insert regular 30-gage needle to the other end of PE-IO tube, Rhodamme 6G (Rho-6G, Molecular Probes): MIX 100 mg of Rho-6G powder with small amount of 95% EtOH to dtssolve and make up to 100 mL wtth 0 9% sterile sodmm chloride, sterthze by passmg through 0 22-pm filter and ahquot m 1-mL portions to syringes, store at -20 to -7OT; final concentratton IS 0.1% Fluorescent latex beads 0 8%, 1.O pm (Polysctences, Warrmgton, PA) Polyacrylate stage wtth rectangular hole (20 x 50 mm) covered with thin cover glass for mesentery Polyacrylate stage with mtcroscope slide glass observatton platform (4-6 mm height) for ear Symmetrical sandwich trtamum chamber (Machme shop, Department of Radtatton Oncology, MGH, Boston, MA), polyacrylate tube (Id 25 mm) and chamber holding stage (Machme shop) for dorsal skm chamber Ctrcular glass cover slip 8 mm in diameter (Asststent, Germany), and polyacrylate stage with stereotacttc apparatus for cranial wmdow

3. Methods 3.1. Cellular Deformability In Vitro There are several methods to measure the deformabthty of cells m vrtro, mcludmg filtratton (17,18), cell poker method (19), and the mtcroptpet asptratton technique (15,20-22). We prefer the latter, because It pertmts quantttattve esttmatton of mtrmsic biophysical parameters, and thus allows mterlaboratory compartson. In this method, a known negative pressure (suctton) 1sapplied on the cell membrane vta a mtcropipet, and the resulting dtsplacement of cell mto the ptpet 1smeasured as a functton of time (Fig. 1A). The resultmg deformation data can Itself be compared among different cell types, or the same cell type under different phystologtcal conditions. If necessary,these data can be analyzed usmg an appropriate mathematical model to yield viscoelastic parameters (22) Some

556

Jain et al.

B

we sure

transducer

amp and dIgItal

micromanl

display

lpulator

bath

video

VCR

enhancer

drain

monitor

Fig. 1. Micropipet aspiration for cell deformability:

(A) schematic, and (B) apparatus.

useful findings in leukocyte biology, using this technique, include demonstration that IL-2-mediated activation increases lymphocyte rigidity, and that thioglycolic acid reduces their rigidity (1.5,22).

L eukocy&Endothe/ia/

Adhesion

557

3.7.7. Equipment An inverted microscope (Nlkon Dlaphot) equipped with Hoffman modulation optics 1s connected to a televlslon camera (Panasonic), as shown m Fig. 1B A video cassette recorder (IAG-6500, Panasonic), with a time code generator (WJ-810, Panasonic), video contrast enhancer (Model 605, Colorado Video), and a television monitor (PM205A, Ikegami), 1s used to record and display the images of the cells during manipulation and deformation. The pressure application system consists of a damping chamber, a three-way mamfold, two water reservoir bottles, connectmg tubes, and a fast response solenoid valve (type 124, Btirket) A damping chamber 1s used to diminish any pulse in pressure that may be associated with the opening ofthe solenoid. This chamber (approx 10 mL vol) 1s half-filled with saline, and its height above the microscope stage 1s adjusted to give a slight negative pressure (CO.5 cm of water) to a mlcroplpet, enabhng the mvestlgator to pick up individual cells from the suspension. The mlcroplpet aspn-ation system consists of a microplpet and mlcromampulator (MO202, Narlshlge, Greenvale, NY), and should include a temperature controlled observation system (such as the MIcrowarm Plate, Schlueter Instrument), although equally satisfactory devices may be devised and manufactured by the investigator to meet their mdlvldual needs. Clean glass mlcroplpets, with 4-7 pm od and 2-4 pm rd, can be made using a manual mlcroplpet puller (Model 720, Kopf, TuJunga, CA) or an automated puller (Sutter). The pipets that are made with these devices must be selected according to the followmg criteria. 1 The opening must be umform around Its circumference, and wlthout chips or cracks 2 The opemng diameter should be no larger than one-third the diameter of the cells to be tested 3. The walls of the plpet near the opening should be nearly parallel 4 The opening must be perpendicular to the axis of the plpet. The selected pipets are generally filled with culture medium prior to mountmg on the mlcromampulator and connecting to the pressure system. A rapid pressure drop 1s applied by activating the solenoid valve, which simultaneously triggers the time code generator (Panasonic) through a relay circuit. In this way, the deformation history of the cells may be recorded, along with the time scale from the instant of the pressure drop. The pressure IS momtored by a transducer (preclslon fl .O%, Gould) connected to a digital display and chart recorder (or an A/D converter and computer).

3. I. 2 MicropIpe t Asp/t-a tion Procedure 1 A suspension of cells, approx 103-104/mL, is placed mto the observation dish on the microscope stage and permitted to settle for several mmutes

Jam et al. 2 Individual cells m the suspenston are selected and ptcked up with the ttp of the mtcroptpet with shght negative pressure and held wtthm 15-20 pm of the chamber bottom 3. A sustained pulse -25 mm water pressure is raptdly apphed by trtggermg the solenoid valve for 2 s, and the movement of the cell membrane wtthm the mtcroptpet 1srecorded on video tape. Observations are generally Itmtted to 2 s, m order to avoid the more complex mechamcal system assoctated with the asptratlon of whole cells mto the ptpet 4 Consecutive vtdeo frames at 0 03 s intervals for each measurement are observed from playback of the videotape, and measurements of deformatton are either taken dtrectly from the video monitor or the images may be dtgtttzed for measurement usmg software packages on PCs (such as the NIH Image 1 59 program, avatlable from the Nattonal Instttutes of Health, Bethesda, MD)

3.1 3. Data Analysis The cell diameter must be noted for each observation, as well as the Internal and external pipet diameter. The edge of the cell membrane within the micropipet is observed m sequential frames after the application of negative pressure The cellular deformation [d(t)] 1sdefined as the difference between the length of the displaced membrane at time, t, and its orrgmal location at t = 0 Each displacement value must then be normalized as d(t)/D,, where D,, = internal micropipet diameter, and the mean values for each test group of cells is plotted as a function of time after pressure mitiation The resultmg curves reflect the deformation history of the cells and may be used to model the viscoelastic behavior durmg deformation (15,22). Alternatively, the resistance factor (R$, which provides a useful Index of the cellular resistance to displacement m a microvascular network (2&22), may be calculated as the product of the cell diameter and the normahzed displacement at 1 s.

3.2. CAM Expression

In Vitro

The most widely used method to measure expression of molecules on the surface of cells is flow cytometry. Although this is suitable for blood-borne cells or cells grown m suspension culture, tt may not be appropriate for cells that are part of an organized tissue or that are grown as a monolayer In addition to possible alteration m CAM expression and loss of some cells because of chemicaVmechamca1 dissociation, required for flow cytometry, one loses the spatial mformatton on CAM expression. For these reasons, we have recently developed a new method-target samplmg fluorometry (TSF)--to measure CAM expression m intact monolayers of cells (14, Fig. 2) This method has yielded useful mformation about why lymphocytes bmd heterogeneously to an endothehal monolayer


Adhesion

559

Fig. 2: Targeted sampling fluorometry: (A) ROI selection criteria, (B) schematic of principle and (C) sample histograms for VCAM- 1 upregulation after TNF-a stimulation. Adapted with permission from ref. 14. 1. Maintain human umbilical vein endothelial cell (HUVEC) cultures (Clonetics) EGM medium (Clonetics) supplemented with 10% fetal calf serum.

in

560

Jain eta/.

20

VCAM-1

h e9 0” s g 10 z! u

0 60

70

Fluorescence

60

90

intensity

100

(relative

110

120

units)

Fig. 2C. 2. Form endothelial cell monolayers in the wells of a 2 x 4 microwell plate (Nunc, Labtek). These slides consist of a standard glass microscope slide with 8 tissueculture wells attached, to allow multiple-sample treatments on one slide. The wells are 0.9-cm square and can be removed after sample fixation for sample mounting. First, coat the glass surface with fibronectin (Sigma) at 6 mg/cm2 for 30 min, wash with PBS, and then seed with a suspension of HUVECs. Use a low seeding density (2 x IO4 cells/well) to produce an even monolayer that is confluent in 4-5 d. 3. Once confluent, activate the cells with the appropriate agent. 4. After activation, fix the monolayers with paraformaldehyde (2 mg/mL in PBS) for 15 min, wash four times with PBS, and then immunostain using a two-step procedure: First incubate with primary antibody directed against the adhesion molecule, and then with a goat F (ab’)2 antimouse IgG Ab conjugated with fluorescein (Tago, Burlingame, CA). All antibody incubations should be performed in the presence of 0.1% bovine serum albumin (BSA) to eliminate nonspecific protein binding. Primary antibodies used in this lab are: IgG controls (IgG*,, K and IgG,, K, Becton Dickinson), anti-E selectin (CD62E; clone HI 8/7, Becton Dickinson), antiP selectin (CD62P; clone GA6, Becton Dickinson), anti-ICAM- 1 (intercellular adhesion molecule-l, CD54; clone 84Hl0, AMAC, Westbrook, ME), and anti-VCAM- 1 (vascular cell adhesion molecule- 1, clone 1G 11, AMAC). Perform all incubations at 37°C for 1 h. To determine the appropriate amounts of antibodies to use, it is necessary to first perform a titration to obtain the level that saturates the system. This must be done with both the primary and secondary antibodies. 5. After the second incubation, wash the layer 4x, and then incubate with 1 mg/mL propidium iodide to stain cell nuclei. After 5 min at 27’C, wash the layers four times with PBS, and mount for observation. The mounting buffer consists of

1eukocyte-Endo thelial Adhesion

6.

7.

8.

9.

a.

b.

c.

d.

561

g1ycerol:PB.S (1:9), with 1 mg/mL paraphenylenediamine (Aldrich) added to retard photobleaching of the samples (23). All fixation, washing, and incubation solutions should contain the divalent cations Ca2’ and Mg2+, to minimize cell detachment from the surface. Gather the images for quantification using an epifluorescence microscope and a camera sensitive enough to image the fluorescence (e.g., a SIT or I-CCD camera). The linearity of the imaging system within the relevant intensity range must be verified with standard fluorescent samples (24). Digitize images with a frame grabber (DataTranslation, Marlboro, MA, or Scion, Frederick, MD) and store for later analysis. A x 10,0.3 NA objective gives a 1.2 x 0.9-mm field of view consisting of approx 700 cells. Record three fields for each well, with two images taken per field: record one image in the 488/530 fluorescein channel, and the other in the 488/600 channel (for nuclei location via propidium iodide staining). No overlap of PI fluorescence should be detected in the fluorescein channel. Record images of unstained HUVEC monolayers in one well of each slide in the 488/530 channel, to provide background fluorescence and lighting variation information. The image analysis is performed using a macro within NIH Image 1.61 (available via anonymous FTP at zippy.nih.nimh.gov) that automates the process. This macro is available via FTP at simon.mgh.harvard.edu. The following processing and analysis steps are performed: The images of cell nuclei stained with propidium iodide are first analyzed to locate the x,y coordinates of the cells. To identify individual cell nuclei, produce binary versions of the gray scale images by autothresholding each rectangular region of a 20 x 20 grid superimposed on the image. This procedure eliminates problems in the segmentation caused by shading variations, and results in an accurate binary representation of the HUVEC nuclei. The X, y coordinates of the centroid of each object in the binary image are recorded. The background fluorescence images are averaged and applied to the immunostain image to correct for lighting variation over the surface. Pixel-bypixel division using the averaged background image eliminates the shading effects and corrects for day to day lamp intensity variations. The X, y coordinates of the cell nuclei are then used to place Regions of Interest (ROls) over each cell in the corresponding processed gray scale images of adhesion molecule staining (Fig. 2A). The ROI diameters are adjusted in regions of higher cell density to minimize sampling overlap: The distances to the two nearest neighbors are determined and averaged to get the appropriate ROI diameter (Fig. 2B). For each ROI, the mean gray level is recorded. These values are then presented in histogram form (Fig. 2C).

3.3. Leukocyte-Endothelial Initial studies on L-E adhesion studies, leukocytes were allowed

Adhesion

In Vitro

were done under static conditions. In these to adhere to endothelial monolayers for a

Jain et al.

562 A

cells in

vacuum

entrance slit

cells

exit

OUI

silastic gasket cell ~2qolayer microscope slide

reservoir

with cells -

VCR

monitor

IBM PC

Fig 3. Leukocyte-endothehal mteractlons m vitro. (A) schematlc of parallel plate flow chamber, and (B) experimental setup Adapted with permission from ref. 3.5


Adhesion

563

fixed duration, and then nonadhered cells were washed off using a standard protocol (25,26) To simulate adheston under physiologtcally relevant flow condttions, a flow-chamber method, tmtially used for platelet adhesion studies (27), was introduced (28). In this method, cells are grown on the bottom plate of a parallel-plate flow cell (Fig. 3). The cell suspenston 1s flowed through the chamber at a known flow rate. The fraction of cells rolling and adhering to the EC layer are measured using video-microscopy and image analysis. We have made three key modifications in this method: determrned the duration of flow at each flow rate, so that the cell flux at the surface is equal at each flow rate; demonstrated that addition of RBCs to leukocytes is essenttal to mrmrc physrologtcal flow condttrons; and showed that keeptng the flow vertrcal, instead of horizontal, eliminates effects of densrty drfferences that influence cell delivery to the surface (29). Although these modifications are not widely used at the trme of thts wrrting, we recommend them.

3.3.1. Leukocyte isolation 1 Isolate human T-cells from leukopheresis products of normal platelet donors by centrifugation over a Histopaque gradient (Sigma) 2 Deplete the lymphocyte layer of monocytes by mcubation with PME (L-phenylalanme methyl ester, E I du Pont de Nemours) for 15 mm (30). 3 Deplete the remaining cells of the B-cell, NK cell and monocyte populations using antibodies directed against CD19 (clone 54 119, AMAC, Westbrook, ME), CD16 (B73.1, Becton-Dickenson) and CD15 (clone MMA, Becton-Dickenson) at 10 pg antibody/15 x lo6 cells, followed by adsorption to a paramagnetic substrate for 30 mm (Advanced Magnetics) and removal by a magnetic field The lymphocytes isolated from this procedure are typically 72% CD4+ and 28% CD8+ cells by flow cytometric analysis 4 If desired, label cells with the membrane-permeant form of calcem, calcein-AM (calcein acetoxymethyl ester, Molecular Probes) or similar dye Calcem is well retamed by the cells, has an extremely strong signal m the 4881530 channel, and causes no detectable change m the adhesron properties of these cells (32,32) Load the cells with the dye immediately prior to the flow chamber studies by Incubation at 24’C with 1 mM calcein-AM for 30 min.

3.3 2. Endotheiiai Monolayer Preparation 1. Maintain HUVEC cultures (Clonetics) in EGM medium (Clonetics) supplemented with 10% fetal calf serum. 2 Form endothelial cell monolayers m the wells of a 2 x 4 microwell plate (Nunc, Labtek) These slides consist of a standard glass microscope slide with 8 tissueculture wells attached, to allow multiple-sample treatments on one slide The wells are 0.9-cm square and can be removed after sample fixation for sample mounting First, coat the glass surface with fibronectm (Sigma) at 6 mg/cm* for 30 mm, wash with PBS, and then seed with a suspension of HUVECs Use a low

564

Jain et al.

seeding density (2 x 10“ cells/well) to produce an even monolayer confluent in 4-5 d. 3. Once confluent, activate the ceils with the appropriate agent.

that is

3.3.3. F/ow Chamber Experiments Adhesion studies require a parallel plate flow chamber machined in-house (Department of Radiation Oncology, MassachusettsGeneral Hospital, Fig. 3A). A silastic gasket forms a small gap between a plexiglass block and microscope slide, when the apparatusis assembledand held together with vacuum. Our chamber width is 1.25 cm, the length is 5.5 cm, and the gap size is 78 pm. The parallel plate flow chamber provides a controlled environment for determinations of the shear stressat which cells in suspension can bind to endothelial cell monolayers. It can be used in a number of different protocols, depending on the type of adhesion data required. 3.3.3.1. DETACHMENT ASSAYS 1. Begin perfusing flow chamber with the cell suspension at an intermediate flow rate, until the cell flux past the monolayer is steady. 2. Stop the flow and allow cells to settle onto the monolayer. The time of stoppage will depend on the cell concentration, chamber height, and cell type (sedimen-

tation rate). 3. After allowing cells to bind, start flow at a low flow rate, and then step up the velocity at regular intervals, until there are few cells left adherent, or the monolayer has detached. 4. From the videotape, quantify the number of cells remaining at each shear stress level.

For a more complete description of this type of experiment, see ref. 33. 3.3.3.2. SINGLE FLOW RATE CAPTURE ASSAYS The flow chamber can also be used to investigate how well cells adhere to the monolayer under dynamic conditions. The basic form of this assay is performed at a single flow rate. If all experiments are carried out at the same flow rate and sampling is performed at the same distance from the inlet to the chamber, then normalization of the data is simplified. 1. Start perfusion of cell suspension (lo6 cells/ml) at an intermediate flow rate. When the cell flux has stabilized, change the shear rate to the target value. 2. Allow adhesion to occur for 5-15 min. 3. While the suspension is still flowing, record five separate fields on videotape, moving horizontally across the chamber. 3. Quantify the number of cells per field that were stably adherent for 5 s or more, and, if desired, the number of cells that were rolling. Normalize to the cell flux passing through the chamber (flow rate multiplied by cell concentration).

Leukocytt+Endothelial

Adhesion

565

3.3.3.3. MULTIPLE FLOW RATE CAPTURE ASSAYS A more informative and, consequently, more complicated method for cellcapture assayinvolves the use of multiple flow rates (12,34,35). By decreasing the flow rate of cell-containing media over the monolayer and assessing the number of cells bound at each wall shear stress,the relationship between shear force and binding efficiency can be determined. However, the rate of binding depends on the delivery of cells to the surface, as well as the intrinsic cellsurface interactions; thus, only if the cell flux to the surface is known can the resulting binding curves be interpreted correctly. 1. Start flow at a high wall shear stress (e.g., 4.1 dyn/cm*), then decrease the flow rate in steps. 2. At each shear level, quantify the cumulative bound cell density at five fields sampled horizontally across the chamber. Monolayers in different runs should be sampled at approximately the same distance from the entrance slit, to avoid flux differences to the surface. 3. In order to facilitate normalization, the time interval at each flow step can be adjusted, so that the total number of cells delivered to the surface is constant for each shear-stress level (35). The wait time, twia between successive shear rates can be found by equalizing the number of cells passing at flow rate, i, according to: Gi = CX/~

(u,_~/u,

+ 2VX/uih

+1) + CU, (1 +

VXIU;h)(tM,;

-x/u~)

(1)

where Gj is the number of cells passing near the monolayer at step i, ui and u;-/ are the current and previous flow velocities, respectively, and t,i is the time spent at flow rate, i (including sampling time), C is the cell concentration, h is the characteristic distance of interaction, x is the distance from the entrance slit, v is the constant sedimentation velocity, and w is the chamber width (35). 4. Calculate the capture efficiencies. If the time interval at each flow step was adjusted so that the total number of cells delivered to the surface was constant for each shear stress level, then normalization is straightforward. The capture efficiency is defined as the number of cells that bind (Nh) at a given shear stress divided.by the total number of cells (GJ that have passed near the surface (E = Nd GJ. In order to extract the most information from the experiment, the cumulative binding curves can be fit to the equation Nb = E,’ e*s, where S is the shear stress, and E,’ and K are adjustable parameters. The value of E, is then calculated from E, = E,’ [ 1 - ePAs], where AS is the difference in shear stress values for successive steps, Thus, E = E,’ e@, with K giving the sensitivity of binding to shear stress, and E, representing the theoretical efficiency at zero shear. Differences in the resulting parameters, K and E,, can be compared between various treatment groups with an appropriate statistical test (e.g., t-test or Mann-Whitney).

3.3.3.4. FLOW CELL MEASUREMENTS WITH ERYTHROCYTES IN THE MEDIUM Additional complications occur if the suspending medium contains large numbers of cells or particles other than the binding species. For example, if

566

Jam et al

blood or RBC suspenstons are being used, sedimentation of the more dense RBCs to the surface of a horizontal flow chamber precludes delivery of the leukocytes to the monolayer. In thts case, the system may be modified by constructmg a vertical flow chamber, thus mmlmtzmg the influence of gravity on the flux at the surface. This has the added advantage of stmphfymg the normaltzatlon procedure, since Eq. 1 only applies m a horizontal chamber. In the vertical case, normalization can be accomplished by simple esttmates of the cell concentratton and the velocity near the surface. In non-Newtonian fluids such as blood, adhesion data should be expressed m terms of shear rate, to avoid complicattons of estimating the local shear force on a cell at the wall. Although the shear stressesat the wall can be estimated, the actual force on a cell rollmg or adhered to the wall may be much higher than this force would indicate, and will be transient m time because of stochastic RBC collisions (29,36) All data are therefore presented on the basis of equivalent shear rates, assuming parabolic flow The correspondmg range of the maxtmum flutd veloctties (assuming parabohc flow) was l&4.1 mm/s, and the average fluid velocittes ranged from 0.68 to 2.7 mm/s, respectively. The wall shear rates calculated as SW= 3/2 / Q/ 2b2w (with b the chamber half-height) were 53-210 s’. 3.4. L-E Adhesion In Vivo The prerequisites for LEI studies m viva are: avatlabthty of acute or chronic tissue preparations that permit momtormg of cells m viva; use of optical labels that are nontoxic and specific to the cell populatton of interest; and availability of equipment and techniques that permit measurement of leukocyte flux, numbers of rollmg and adherent cells, and flow parameters m blood vessels of choice. Acute preparations widely used include mesentery or cremaster muscle of rodents (37). Commonly used chronic preparations include the rabbit ear chamber (16,38-41), dorsal skin-fold chamber in rodents (12,41), and cranial windows in mice and rats (12,#2,43; Fig. 4). These tissue preparations can be used to monitor the normal tissue as a control or after treatment with appropriate chemical. The chronic windows permit transplantation of syngemc vs xenogemc tissues (normal or tumor) m immunocompetent vs nnmunodeficient hosts, respectively. The endogenous cell population is usually labeled with rhodamine 6G. The exogenously injected leukocytes can be labeled with a number of tracers (e.g., calcem), depending on the duration of experiment (shortterm vs long-term). The physiologtcal parameters (such as vessel diameter and RBC velocity to calculate the shear rate) are usually measured optoelectronically, using vtdeomrcroscopy and image analysts (Fig. 4). In recent years, the use of transgemc and knock-out mice has provided stunnmg insight mto the LEI (44).

Leukocyte-Endothelia!

Adhesion

567

Fig. 4. Various chronic window preparations (A) and experimental setup (B). Parameters measured include vessel diameter, red blood cell velocity, leukocyte flux, rolling velocity, and adhesion density.

3.4.1. Preparation of Acute Tissue Observation: Mesentery 1. Fast the animals for 24 h before the observation.

568

Jain et al.

2. Anesthetize the animals (SCinjection of ketamine and xylazine mixture, 10 p-L/g body wt for mice, 1 pL/g body wt for rats; final dose should be ketamine IO mgl xylazine 1 mg per 100 g body wt for mice, and ketamine 9 mg/xylazine 0.9 mg per 100 g body wt for rats). 3. Place the catheter (30-gage needle and PE-10 tubing) into tail vein for fluorescence tracer injection; fix the cannulation site by a cyanoacrylate adhesive. 4. Shave abdominal skin by electric shaver. Clean the skin with antimicrobial Betadine solution. Open abdomen via midline incision. 5. Expose ileocecal portion, gently develop intestinal loop onto thin glass part of the polyacrylate stage using Q-tip immersed in saline, avoiding direct touch and tension to mesentery. 6. Gently straighten the intestine and fix by cotton immersed in saline, so that mesentery will be unfolded. Keep mesentery moist and warm by superfusion with warm saline (37°C). Animal body temperature is also controlled by means of heating pad. 7. Place the stage on inverted/orthotopic intravital fluorescence microscope (Axioplan; Zeiss, Oberkochen, Germany); observe mesenteric microcirculation using x20 to x40 objective lens. 8. Visualize transilluminated image by a charge-coupled device (CCD) video camera (AVC-D7, Sony, Tokyo, Japan) and record the image for 60 s using S-VHS video cassette recorder (SVO-9500MD, Sony). 9. Inject 20 pL of 0.1% Rho-6G for in vivo labeling of whole leukocyte population, visualize leukocytes by epi-illumination, using specific filter set for Rho&G (excitation peak at 528 nm and emission peak at 550 nm) by intensified CCD video camera (C2400-08, Hamamatsu Photonics K.K., Hamamatsu, Japan) and record the image for 30 s. 10. In case of high-velocity vessel (>l mm/s), inject 10 pL of 0.8% FITC-fluorescence beads (1 .O pm), visualize, and record FITC fluorescence image for 60 s, using ICCD camera with high-speed electronic shutter (C5909 Hamamatsu Photonics K.K.). 11. Repeat steps 7-9 at each observation time-point or other area.

3.4.2. Preparation of Acute Tissue Observation: Ear Microcirculation 1. Remove the hair of ear skin using hair-removing cream (8 min) 24 h before the observation. 2. Anesthetize the mice. 3. Cannulate the tail vein. 4. Drop saline on the microscope slide glass on the observation platform, place the ear gently onto the slide glass, drop saline at the edge of the ear, put thin cover glass on the top, and allow to disperse saline between the ear surface and the slide/cover glass to hold the ear to the slide and to enhance image quality. 5. Observe ear skin microcirculation (see Subheading 3.4.1.).


Adhesion

569

3.4.3. Preparation of Chronic Tissue Observation: Dorsal Skin-Fold Chamber 1. Anesthetize the mice (SC injection of ketamine and xylazine mixture 10 pL/g body wt; final dose should be ketamine 10 mg/xylazine 1 mg/lOO g body wt). 2. Shave and deplete entire back of mouse using electric shaver and hair removing cream (8-12 min). 3. Assemble the sandwich titanium chamber: side 1, with three screws and nuts; side 2, close the window with circular cover slip and fix it by O-ring. 4. Suture on top and bottom of dorsal skin and pull up these sutures to make and extend double layer of the skin. 5. Suture at the top part of side 1 titanium chamber onto skin for temporal fixation. 6. Make two small holes in the double layer of skin, so that screws of the chamber can penetrate the skin. 7. Remove a circular area (15-mm diameter) of one layer of skin. 8. Remove membranous tissue carefully from the remaining layer as much as possible without tissue damage to get good optical quality. 9. Place side 2 titanium chamber to sandwich the extended double layer of the skin and fix the screws. 10. Remove the temporary sutures and suture the top part of the chamber to the skin. 11. Cover the remaining skin in the chamber with scotch tape, to protect from scratching. 12. Allow the mice to recover for 48 h after microsurgery and anesthesia. 13. Place a mouse in polyacrylate tube (id 25 mm), fix the chamber to the stage, and place the stage under the microscope. 14. Observe skin microcirculation (see Subheading 3.4.1.).

3.4.4. Preparation of Chronic Tissue Observation: Cranial Window 1. Anesthetize the animals (SC injection of ketamine and xylazine mixture 10 p.L/g body wt for mice, 1 &L/g body wt for rats; final dose will be ketamine 10 mgl xylazine 1 mg/lOO g body wt for mice, and ketamine 9 mg/xylazine 0.9 mg/lOO g body wt for rats. 2. Fix the animal’s head by a stereotactic apparatus and clean up the skin on top of the frontal and parietal regions of the skull with antimicrobial Betadine solution. 3. Make a longitudinal incision of the skin between the occiput and forehead, cut the skin in a circular manner on top of the skull, and scrape off the periosteum underneath to the temporal crests. 4. Draw a 6-mm circle over the frontal and parietal regions of the skul1 bilaterally, make a groove on the margin of the drawn circle using a high-speed air-turbine drill (CH420 1S; Champion Dental Products, Placentia, CA) with a burr tip size (0.5 mm in diameter), separate the bone flap using a blunt microblade from the dura mater underneath, and keep the dura mater moist with physiological saline. 5. Make a nick close to the sagital sinus, pass iris microscissors through the nick, and cut the dura and arachnoid membranes completely from the surface of both hemispheres, avoiding any damage to the sagital sinus.

570

Jam et al.

6 Seal the window with an 8-mm circular cover glass by adhering to the bone using a htstocompatrble cyanoacrylate glue 7 Allow the ammals to recover for 7 d after microsurgery and anesthesra 8 Anesthetize the animals, cannulate the tat1 vem, put the ammals on a polyacrylate plate, with the head fixed by plastic modeling compound, and put the plate under the mtcroscope. 9 Observe the ptal mtcrocuculation (see Subheading 3.4.1.)

3.4.5. Data Analysis 1 Vessel diameter In each obsetvatton, transtllummated tissue images should be recorded for 60 s and the video tapes wtll be analyzed off-line The vessel dlameter m mn (D) ~111 be measured using an image-shearing device (digital video Image shearing monitor, model 908, IPM, San Diego, CA) (45) 2 The RBC velocity (V~nc). Vauo can be measured using the four-sltt apparatus (Mtcroflow System, model 208°C video photometer version, IPM) equipped with a personal computer (IBM PS/2,4OSX, Computerland, Boston, MA) (45) In case of high-velocity vessel (~1 mm/s), the fluorescent bead method can be used to estimate maximum blood velocity (24) The distance a bead moves wtthm a vessel m a given time interval (defined by the frequency of the camera shutter) can be determined by measuring the distance between two posittons of the bead m the same track, using an image processing system (NIH Image Ver 1 58, Macintosh IIfx, Apple Computer, Cupertmo, CA) The maximum bead velocity can be calculated by dividing the maximum distance m more than 300 tracks during 1-mm recording by the given time interval 3 The mean blood flow rates of mdivldual vessels (Q) and shear rate Q can be calculated usmg D and the mean V,,, (V,,,,,,) as follows* Q = x/4 x If,,,,,, x D*, Vmean= VRncla (a = 1.3, for blood vessels < 10 pm, linear extrapolatton 1 3 < a < 1 6 for blood vessels 10 and 15 pm; and a = 1 6 for blood vessels > 15 pm) (45) Shear rate will be calculated for each vessel as Shear rate = 8 x V,,,,/D (42) 4. Leukocyte-endothelial mteracttons (LEI) LEI can be measured by off lme analysts of fluorescence images of leukocytes (42) The numbers of rolling (N,) and adhering (NJ leukocytes are counted for 30 s along a loo-pm segment of a vessel Rollmg 1s defined as a short-term mteractton and slow movement along vessel wall (~50% of VRBc), adhesion 1sdefined as stable bmdmg at a certain posttton for more than a 30 s pertod. The total flux of cells for 30 s is also measured (NJ The ratto of rollmg cells to total flux (Rolling count) is calculated as follows Rolling count (%) = 100 x N,/N, The density of adhering leukocytes (Density) is calculated as follows. Density (cells/mm*) = lo6 x NJ(7c x D x 100 pm)

4. Notes 1 Cellular deformabthty a It 1s absolutely crtttcal that the cells being tested do not adhere to the pipet tip during the testing procedure Cells that rapidly adhere to the glass ptpet, such as macrophages, can occlude the tip followmg the testing procedure, thus

Leukocyte-Endotheliai

Adhesion

preventing any additional measurements with the same pipet This can be avoided by using a suspending medium that is free of calcium and magnesium, or by mcludmg EDTA in the testing medium b. Nonviable cells and cell fragments can also be aspirated into the pipet tip, and block the fluid flow or rum the optical clarity of the pipet tip Thus, caution must be exercised when working with cell populations that have an overall low vrability c. The source of pressure m the hydraulic system must be free of oscillation Although this IS reduced with the in-line dampening chamber, it can be further suppressed by mcludmg a few drops of liquid soap mto the flasks of water that provide the driving pressure for the system (thus reducing the surface tension of the fluid in these flasks) 2 Adhesion molecule expression a. Even though the mounting medium contains a photobleach mhibitor, it IS good practice to mmlmize sample illummatton, so that photobleachmg does not sigmficantly affect the measurements b During the fixing and stammg steps, washmg should be performed as gently as possible (munmize fluid shear on the cells) to avoid damagmg the adherent cells We have found that the mclusion of divalent cations m the fixing and washing medrum helps mamtam an intact HUVEC monolayer c In some cases, the PI stammg IS nonuniform, with some nuclei stammg more brightly than others In this case, an additional PI application is necessary It is also acceptable to include a detergent wash (after the cell fixation) to permeabillze the membranes, allowing easier access for the PI 3 L-E mteractions m vitro a After the addition of PME to deplete the monocyte population, an aggregate of DNA may form, It is necessary to remove this from the centrifuge tube and wash out as many of the trapped lymphocytes as possible with gentle agitation b Isolated T-cells should generally be stored at 4°C and used within 24 h, but have been shown to retain adhesive capabdmes for as long as a week under these conditions c A VCR with Jog/shuttle capabilities helps m the quantification of the video tapes A dry-erase marker is also useful for making grids and for keeping track of bound -endothelial interactions m VIVO: a All surgical procedures should be done under aseptic conditions and all materials should be sterilized prior to use. If there is any sign of tissue mmry, mflammation, or compromised hemodynamtcs, the tissue preparations should be ehmmated from the observation b Ear setup ts suitable for the baseline observation, since surgical preparation is not required Mesentery set up has best opttcal quality and can be used for regional treatment by superfusion (46). c Chronic preparation can be used for long-term observation, regional treatment by contmuous (47) or one-time superfusion (42), growth factor incorporated gel assay (48), and tumor study (24,42,4.5,47) The procedure for the tumor implantation is as

Jain et al

572

follows. dorsal sktn chamber; remove the covershp and inoculate 2 pL of very dense tumor-cell suspension (approx 2 x lo5 cells) or 1-mm-diameter fragment of tumor tissue onto the center of the chamber and close the chamber again using a cover shp, cramal window* before sealing the window with glass cover shp, moculate a piece of tumor ttssue (1 mm in diameter) at the center of the window

References 1 Jam, R K , Koenig, G C., Delhan, M , Fukumura, D , Munn, L L , and Melder, R J (1996) Leukocyte-endothehal adhesion and anglogenesis in tumors Cancer Metastaszs Rev 15, 195-204 2 Bell, G I (1979) A theoretical model for adheston between cells mediated by multivalent hgands Cell Bzophys 1, 133-147 3 Dembo, M , Torney, D C , Saxman, K , and Hammer, D (1988) The reactionhmited kinetics of membrane-to-surface adhesion and detatchment Proc R Sot Lond. B 234,55-83. 4 Lawrence, M B and Springer, T A (1991) Leukocytes roll on a selectm at physi-

ological flow rates distmctton from and prerequisite for adheston through mtegrms Cell 65,85%873 5 Lipowsky, H. H , House, S D., and Fnrell, J. C (1988) Leukocyte endothelium adhesion and microvascular hemodynamics, m Vascular Endothellum In Health and Dzsease (Chlen, S , ed.), Plenum, New York, pp. 85-93.

Wattenbarger, M R , Graves, D J., and Lauffenburger, D A. (1990) Specific adhesion of glycophorm hposomes to a lectm surface in shear flow. Blophys J 57,76%777 Lipowsky, H. H , Riedel, D , and Shi, G S (199 1) In VIVOmechanical properttes of leukocytes during adhesion to venular endothehum Bzorheology 28, 5344 Sasakt, A., Jam, R K., Maghazachi, A. A , Goldfarb, R. H , and Herbennan, R B (1989) Low deformabthty of lymphokme-activated killer cells as a possible determinant of in viva dtstribution Cancer Res. 49,3742-3746 9 Schnnd-Schonbem, G. W (1990) Leukocyte biophysics Cell Biophys 17,107-135 10 von Andrian, U H., Hansell, P, Chambers, J. D , Berger, E M , Torres-F&o, I, Butcher, E C , and Arfors, K E (1992) L-selecnn function is required for beta 2-u&grin-mediated neutrophtl adhesion at physiological shearratesin VIVO.Am J Physzol 263, HI 034-H 1044 11 von Andnan, U H , Chambers, J D , McEvoy, L M , Bargatze, R F , Arfors, K -E , and Butcher, E C. (1991) Two-step model of leukocyte-endothehal cell mteraction m mflammation. distmct roles for LECAM-1 and the leukocyte p2 mtegrms in VIVO Proc Nat1 Acad Set USA 88,7538-7542

12 Melder, R J , Munn, L L ,Yamada, S , Ohkubo, C , and Jam, R. N (1995) Selectmand mtegrm-mediated T-lymphocyte rolling and arrest on TNF-a-activated endothelium augmentation by erythorcytes Bzophys J 69,2 13 l-2 138 13 Springer, T A , Anderson, D. A , Rosenthal, A S., and Rothlem, R , eds ( 1988) Leukocyte Adhesion Molecules Structure, Function, and Regulation Springer-Verlag, New York 14 Munn, L L , Koenig, G C , Jam, R K , and Melder, R J (1995) Kmetics of adhesion molecule expression and spatial organization using targeted sampling fluorometry Blotechnzques 19,622-63 1

Leukocyte-EndotheliaI

Adhesion

573

15 Melder, R J. and Jam, R K (1994) Reduction of rigidity m human activated natural killer cells by throglycollate treatment. J Zmmunol Methods 175,69-77. 16 Sasaki,A , Melder, R J., Whiteside, T L , Herberman, R. B., and Jam, R K (1991) Preferential locahzation of human adherent lymphokine-actrvated killer (A-LAK) cells m tumor mrcrocrrculatron J Nat1 Cancer Inst 83, 433-437. 17 Betticher, D C , Keller, H , Maly, F. E., and Remhart, W. H (1993) The effect of endotoxm and tumor necrosis factor on erythrocyte and leucocyte deformabrhty m vnro. Br J Haematol 83, 130-137 18 Welch, D R , Lobl, T J., Sefior, E A, Wack, P. J., Aeed, P A ,Yohem, K H , Seftor, R E., and Hendrrx, M. J. (1989) Use ofthe membrane mvasron culture system (MICS) as a screen for anti-mvasrve agents Int J Cancer 43,44!+457. 19 Downey, G P, Doherty, D. E. Schwab, B. D., Elson, E L , Henson, P M , and Worthen, G S (1990) Retention of leukocytes m capillaries role of cell size and deformability. J Appl Physzol 69, 1767-1778 20 Traykov, T T. and Jain, R. K. (1987) Effect of glucose and galactose on red blood cell membrane deformability. Int J Mzcrocwc Clm. Exp 6,35-44 21 Melder, R and Jam, R. (1992) Kinetics of interleukm 2 induced changes m rrgrdity of human natural killer cells. Cell Blophys 20, 16 l-1 76 22 Sasakr, A, Jam, R K., Maghazachr, A A, Goldfarb, R. H , and Heberman, R B (1989) Low deformabrhty of lymphokme-actrvated killer cells as a possible determmant of m VIVO distrrbutron Cancer Res 49,3742-3746 23 Kremk, K D , Kephart, G M , Offord, K. P., Dunnette, S L , and Gleich, G J. (1989) Comparrson of antifading agents used in immunofluorescence J Immunol. Methods 177,9 l-97 24 Yuan, F, Salehi, H. A, Boucher, Y., Vasthare, U. S , Tuma, R. F., and Jam, R. K. (1994) Vascular permeability and microcnculatton of ghomas and mammary carcrnomas transplanted m rat and mouse cranial wmdow. Cancer Res 54,4564-4568. 25 Gallik, S , Usami, S., Jan, K.-M , and Chren, S (1989) Shear stress-induced detachment of human polymorphonuclear leukocytes from endothehal cell monolayers Bzorheology 26, 823-834 26 Hochmuth, R. M , Mohandas, N , Spaeth, E. E., Wrllramson, J. R., Blackshear, P L and Johnson, D W (1972) Surface adhesion, deformation and detachment at low shear of red cells and white cells Trans Am Sot ArtiS Organs 18, 325-332 27 Hubbell, J A and McIntire, L V (1986) Visualrzatron and analysrs of mural thrombogenesrs on collegen, polyurethane and nylon Bzomaterzals 7, 354-363 28 Lawrence, M B , McIntne, L V, and Eskm, S G (1987) Effect of flow on polymorphonuclear leukocyte/endothelial cell adhesion Blood 70, 1284-l 290 29 Munn, L. L., Melder, R. J., and Jam, R. K. (1996) Role of erythrocytes m leukocyteendothehal interactions mathematrcal model and experimental vahdatton Biophys J 71,466478 30. Leung, K H (1989) Human lymphokme-activated killer (LAK) cells I Depletion of monocytes from peripheral blood mononuclear cells by L-phenylalamne methyl ester and optimization of LAK cell generation at high density Cancer Zmmunol Immunother 30,247-254

574

Jain et al.

31 Weston, S A and Parrsh, C R (1990) New fluorescent dyes for lymphocyte mtgratton studres analysis by flow cytometry and fluorescence microscopy J Immunol Methods 133,87-97 32. Weston, S. A and Partsh, C R. (1992) Calcem a novel marker for lymphocytes whtch enter lymph nodes Cytometry 13, 739-749 33 Menter, D G , Patton, J. T, Updyke,T V, Kerbel, R. S , Maamer, M , Mcinttre, L V, and Ntcolson, G L (1992) Transglutammase stabilizes melanoma adheston under lammar flow Cell Bzophys 18, 123-143 34 Melder, R J , Koemg, G , Munn, L L , and Jam, R. K (1997) Adheston of acttvated natural killer cells to TNFa-treated endothelmm under phystologtcal flow condmons Nat Immun l&154-163 35 Munn, L. L , Melder, R J., and Jam, R K (1994) Analysts of cell flux m the parallel plate flow chamber. tmphcattons for cell capture studies Bzophys J 67,889-895 36 Schmtd-Schonbem, G W, Fung,Y.-C., and Zwetfach, B W (1975) Vascular endothehum leukocyte mteractton Sttckmg shear force m venules Czrc Res 36, 173-184 37 Fukumura, D , Yuan, F , Monsky, W , Chen,Y , and Jam, R K (1997) Effect of host mtcroenvtronment on the mtcroctrculatton of human colon adenocarcmoma Am J Path01 150,679-688 38 Ohkubo, C., Brgos, D , and Jam, R K (1991) IL-2 induced leukocyte adhesion to the normal and tumor mtcrovascular endothelmm m vlvo and its mhlbmon by dextran sulfate. tmphcattons for vascular leak syndrome. Cancer Res 51, 1561-1563 39 Nugent, L J and Jain, R K (1984) Extravascular dtffuston m normal and neoplastic tissues Cancer Res 44,238-244 40 Zawtckt, D F , Jam, R K , Schrmd-Schonbem, G W., and Chten, S (198 1) Dynamics of neovasculartzatton m normal tissues Mzcrovasc Res 21, 37-47 41 Dudar, T. E. and Jam, R K (1983) Mrcrocnculatory flow changes during tissue growth Mzcrovasc Res 25, 1-21 42 Fukumura, D , Saleht, H A , Witwer, B , Tuma, R F , Melder, R J , and Jam, R K (1995) Tumor necrosis factor alpha-induced leukocyte adhesion m normal and tumor vessels effect of tumor type, transplantation site, and host strain Cancer Res 55,48244829 43 Melder, R J , Salehr, H A , and Jam, R. K (1995) Interactton of activated natunal killer cells with normal and tumor vessels in cranial windows m mice Mzcrovasc Res 50,35-44 44 Yamada, S , Mayadas, T N., Uan, F , Wagner, D D , Hynes, R 0 , Melder, R J , and Jam, R K (1995) Rolling m P-selectm-deficient mice IS reduced but not eltmtnated m the dorsal skin. Blood B&3487-3492. 45 Leumg, M Yuan, F, Menger, M D , Boucher, Y, Goetz, A E , Messmer, K , and Jam, R K (1992)Angtogenesq mlcrovascular archttechture, mtcrohemodynamtcs, and mterstmal fluid pressure during early growth of human adenocarcmoma LS 174T m SCID mice Cancer Res 52,6553-6560 46 Kurose, I., Fukumura, F, Mmra, S., Suematsu, M , Sekrzuka, E , Nagata, H , and Tsuchtya, M. (1993) Nitric oxide mediates vasoactrve effects of endothelm-3 on rat mesentertc mtcrovascular beds rn vzvo Angzology 44,483-490

Leukocyte-EndotheliaI

Adhesion

575

47 Fukumura, F, Yuan, F, Endo, M , and Jam, R. K. (1997) Role of mtrtc oxide m tumor microctrculatton blood flow, vascular permeabthty, and leukocyte-endotheha1 mteracttons Am. J Pathol 150,713-725 48 Delhan, M., Wttwer, B P, Saleht, H. A , Yuan, F , and Jam, R K (1996) Quantrtatron and phystologrcal charactertzatton of angtogemc vessels m mace effect of basic tibrooblast growth factor, vascular endothehal growth factor/vascular permeabthty factor, and host mtcroenvtronment Am J Path01 149, 59-72.

42 Quantitative Measurement on Endothelial Cells

of Shear-Stress

Effects

Maria Papadaki and Larry V. Mclntire

1. Introduction Over the past 20 yr, great strides have been made toward understanding the role of fluid hemodynamlc forces in the vascular wall homeostasls at the molecular level. In VIVOstudies have demonstrated that blood vessels are adaptlve to physlologlcal changes m blood flow, with vessels tending to enlarge m areas of high flow and tending to reduce their lumen diameter m low-flow regimes (1,2). Furthermore, altered hemodynamics have been lmpllcated m the pathogenesls of many cardiovascular disorders, such as thrombosis, atherosclerosis, and vessel wall injury. Vascular endothehal cells serve as a barrier between perfused tissues and flowing blood, and they are believed to act as a sensor of the local blomechamcal environment. The hemodynamlc forces generated m the vasculature Include frictional wall shear-stress, cychc strain, and hydrostatic pressure (3). For the purpose of this chapter, we will focus on methods for examining the lmk between fluid wall shear-stressand endothelial cell function. Advances m our understanding of the effects of shear-stress on endothellal cell function require that cell populations be exposed to controlled, well-defined, flow-induced shear-stress environments. Since in vwo studies have the inherent problem that they cannot quantltatlvely define the shearing forces or separate then effects from the other components of the hemodynamlc system, m vitro flow studies using cultured cells are extensively used. The objective of this chapter IS to review the methodologies and the flow chambers used to expose cultured endothehal cells to fluid shear-stress. There are two major ways of mtroducmg shear-stress m cultured cells: those that From

Methods III Molecular Meddne, Vol 18 Tissue Engrneenng Methods and Protocols Edlted by J R Morgan and M L Yarmush 0 Humana Press Inc , Totowa, NJ

577

578

Flow In

Papaciaki and Mclntrre

Flow Out

Fig. 1 (A) Rectangular parallel-plate flow chamber The glass slide, the silicon gasket, and the polycarbonate plate are held together by a vacuum that IS mamtamed at the periphery of the slide (B) Velocny and shear-stress profile for flow between parallel-plates The velocny profile IS parabohc and shear-stress changes linearly with y The wall shear-stress IS constant (except very near the lateral edges), w 1sthe channel wtdth, h IS the channel height, and L IS the channel length

produce uniform shear-stress throughout the fluid phase, and those m whtch shear-stress changes with the distance from the solid-liquid interphase (4-6) Some of the commonly

used instruments,

together with the advantages

and

experimental problems associated with each one of them, are described m the followmg pages In order to illustrate the fluid dynamics wtthm those devices, several simple designs are constdered m detail, devices with more complex geometries are briefly mentioned. 2. Methods 2.1. Parallel-Plate Flow Chamber The parallel-plate flow chamber IS the most widely used mstrument for studying the effects of fluid shear-stresson endothehal cells (or any other kmd of anchorage-dependent cells) (6,7). A schematic of the most common parallelplate flow chamber geometry 1sshown m Fig. 1 This parallel-plate flow chamber consists of a rectangular polycarbonate plate, a silicon gasket, and a glass slide (8). Endothehal cells are cultured on the glass slide, whtch can be coated with a variety of adhesive substrates A vacuum forms a seal to provide a channel of parallel geometry, and to ensure uniform channel height, h The polycarbonate plate has entrance and exit mamfolds

for the ttssue-culture

medium

machined mto the plastic block, as well as a vacuum-sealing ring. In other designs of the parallel-plate flow chamber the vacuum 1sreplaced by evenly torqued screws. For short-term experiments tissue-culture medium is perfused

Stress Effects on Endothelial Cells

579

through the chamber using a syringe pump; for long-term experiments, flow IS produced by a constant hydrostattc pressure head, which 1s created by the vertical distance between two reservoirs (8). Contmuous pumpmg of the fluid medmm from the lower to the upper reservoir is employed, so that there 1s an overflow of excess medium mto the lower reservoir. The overflow serves to mamtam a constant hydrostatic pressure head, to prevent air bubble entry mto the flow chamber, and to provide good mass transfer for medium oxygenation. The wall shear-stress, TV, on the cell monolayer for a Newtonian flutd 1s calculated using the momentum balance: z, = -6PQ wh2

(1)

where Q 1s volumetrtc flow rate, ~1s the viscosrty of the fluid/tissue-culture medium, and w 1s the channel width. The parallel-plate approxtmattons are only valid tf the height of the flow channel is much smaller than the width and length of the channel, L; h/w C-C 1, h/L C< 1 ($6) In practice, a heightwidth ratto value of at least 150 1s used. In order to ensure uniform veloctty across the channel wtdth, the reservoir dtameter, from which the medium 1s discharged, should be large enough so that the pressure drop along the parallel-plates 1s larger than the pressure drop m the reservotrs and the kmettc energy of the inlet stream. The entrance length, L,, must be small compared to the chamber length, in order for Eq. 1 to be valid for nearly all the cell monolayer except the extreme lateral edges of the flow chamber. The entrance length IS given by (6,s).

z h2p

L, = 0.04 h Re = 0 04 h (-“---)

6P2

(2)

where Re 1s the Reynolds number; and p 1s the fluid density For the low Reynolds numbers normally generated, the entrance length is approx 400x smaller than the length of the flow chamber. The advantages of the parallel-plate flow chamber are: 1, It makes possible study of the effects of constant shear-stress on endothehal cells over a defined time-period 2. The devtce IS simple in design, assembly, and operation 3 The endothehal cells can be grown under flow conditions, and can be observed under a microscope, or visuahzed in real time, utilizing vtdeo microscopy 4 This type of chamber IS preferable for studies on the metabolism of endothehal cells, in which a large number of cells needs to be harvested for measurement of rnRNA 5 It permits continuous sampling of the incubation medium for secreted metabolites (8). The parallel-plate flow chamber, m tts onginal design, 1s capable of producing well-defined wall shear-stress m the phystologtcal range of 0.01-30 dyn/cm2.

Papadahr and Mclnttre One disadvantage is that it ts dtffcult to dtscrimmate between effects caused by shear-stressand those caused by the hydrostatic pressure, since both of them increase m a manner proportional to the flow (6,9]. Furthermore, the parallelplate flow chamber cannot effectively be used to study effects of shear-stress on suspended cells, because shear-stress varies over the cross sectton of the chamber, and sedimentation and aggregation of suspended cells can take place. A variation of the rectangular parallel-plate flow chamber 1sthe use of dual parallel flow paths employmg a round tissue-culture dish as the substrate for endothehal cell growth (10). Those chambers are commonly used to study the vartous receptor-hgand interactions m cell-cell and cell-substrate adhesion under defined hydrodynamic condittons. The area of the flow section m these round chambers is approx 16x smaller than the rectangular chamber, which makes them ideal for studies when cell numbers are limited. In addmon, these round chambers are more convenient, since cells are cultured m plastic 35-mm tissue-culture dishes instead on glass slides. Other mvestigators have developed designs m which the parallel-plate approximations are vahd even at relatively high Reynolds numbers Koslow et al used a pair of parallel-plates with a divergent entranceand a convergent exit (21). At the divergent portion of the chamber,flow was fluctuatmg, but at the central area, where cells were placed (on a cover shp), flow was lamtnar The maxtmum shearrate studied was 1590/s(Re = 138). Vtggers et al. used a parallel channel that combmed bell-shaped entrancewith baffles, which helped to establisha constantfluid velocity at the begmnmg of the lamtnar flow section(12). Those sameresearchersused threedtmenstonalfinite element numerical analyststo calculatewall shear-stress,and they found that the cell cover shp can examine shear-stressesover 200 dyn/cm2,and that the veloctty gradient ts uniform acrossthe central 80% of the cover slip. Eskm et al. subjected bovine endothehal cells, seeded mside a square glass capillary tube, to flow, by connecting the tube into a closed loop consrstmg of two reservoirs and a pump (13). The analyttcal solutton for the velocity profile for pressure-driven low-Reynolds-number flow of a Newtonian fluid m a square tube is more comphcated than the one for the parallel-plate flow chamber (14). Srmilarly, Cooke et al. used SIX rectangular tissue-culture microcapptlary tubes per experiment to examine endothehal cell growth, and to perform adhesion studies under well-defined flow condmons (IS). 2.2. Cone and Plate The cone-and-plate device 1sa well-defined rheological system that was mlttally introduced for measuring fluid shear-stress effects on suspended cells (5,6,16) More recently, this assembly has been used to study flow effects on anchorage-dependent cells. As shown in Fig. 2, the bottom plate is not moving and can accommodate several cover slips (16). Shear-stress is produced on

581


Fig 2 Cone-and-plate device Endothehal cells or any other kmd of anchorage dependent cell are cultured on cover shps, which then are Inserted on the stationary plate For very small cone angles (a 1. It is also essential that the cone axis 1sperpendicular to the plate; otherwise, pulsatile flow will result. Furthermore, the fluid should fill only the space between the cone and plate, otherwise an addmonal torque from fluid shearing can be developed (6).

582

Papadaki and Mcln tire

The man-radvantages of the cone-and-plate are summarrzed here 1 There are no entrance and exit lengths, unlike the parallel-plate flow chamber 2 The lammar homogeneous flow field does not depend on a hydrostattc pressure gradient 3 Small changes m dtmenslons do not have a large effect on wall shear-stress 4 Shear-stress effects on several cover slips can be examined m the same expertment 5 The operating flmd volume IS small (16).

The cone-and-plate apparatus is capable of producing lammar shear-stresses between 0 01 and 100 dyn/cm*. As orrgmally designed, the cone and plate device has a small cell:volume ratio, does not perrmt contmuous

sampling

of the condtttoned

medtum,

has

stgmficant medium evaporatton, which requnes contmuous mfuston of fresh medmm, and does not easily allow continuous vtsualizatton of cells during shear-stress apphcatton ($68) To allow direct observation of anchorage-dependent cells, Schmttler et al. have developed a modtfted cone-and-plate vtscometer m which the rotating

cone IS transparent

(9). Cells are visualized

by phase-contrast

or

fluorescent optics, and the whole system IS connected to a computer-controlled linear stage, m order to allow automatic recording at any point of the cell cultures. This improved rheological device consists of 12 mdependent units (chambers), so that a number of different condtttons can be tested stmultaneously during a single experiment. Another modtfication of the cone-and-plate device 1s the rheoscope, m which the cone and plate counterrotate (17,18). The counter rotation of the cone and plate has the advantage

that a particle

at the center line between

the cone and plate

remams stationary to the observer and can be studied without the atd of high-speed cinematography. This device was developed to study rheologlcal properties of blood components. If a,,,, and c+,lateare the rotation speeds of the cone and plate, respectively, then the shear-stress m the apparatus, assuming

Newtonian

fluid,

z=z,=p

1s computed

as:

( 0 ‘one + w,tate ) 2a

(5)

By increasing the angular velocrty of the cone and mcreasmg the angle between the cone and the plate, turbulent flow can be achieved with the coneand-plate device (19). Turbulent shear-stresses,achteved with this device, are on the order of 2-200 dyn/cm*. Turbulence appears first at the outer boundary and spreads inward towards the cone apex, shear-stress is no longer constant across the gap. Davies et al. studied the effects of turbulent flow on endothehal cells, using a cone angle of 5 degrees. Shear-stress fluctuattons were hmtted to 25% of the mean value (19).


583

Ldum cdl monolayer bottom disk

Fig 3 Parallel-dtsk apparatus Cells are growing on the bottom disk, rotatton of the top disk causes the medmm to flow concentrtcally and to produce shear-stress on cells Shear-stress IS a functton of r and decreases from the perimeter to the center of the bottom disk

2.3. Parallel-Disk Apparatus Parallel-disk systemshave the advantage, over the aforementroned systems, that If the flow pattern 1swell defined, shear-stress gradients can be examined on anchorage-dependent cells (6,20,21). Because of the time-dependent nature of many btologrcal processes, stgnificant information can be obtained If cultured cells are submitted to a contmuous range of shear forces stmultaneously. A concurrent problem, however, is that, smce cells see different shear-stress hrstorres, evaluatton of shear-stress effects on secreted products or mtracellular messenger RNA levels is not possible. The operation of the parallel-disk system, a schemattc diagram of which is shown m Fig. 3, is very similar to the

cone-and-plate device (20). Rotation of the upper disk with a constant angular velocity, o, forces the medium between the two parallel dtsks to flow ctrcumferenttally, and produces a fluid shear-stress on the cells growing at the bottom disk. For a Newtonian fluid, the shear-stress 1scalculated as: z,=pWr

(6)

d where d IS the vertrcal distance between the two disks, and r is the distance from the center of the culture dish. The model approximatrons include neglectmg the effects of the centrifugal force in the rotating fluid, and, as a consequence, the radral and axial secondary velocity components are zero These assumptions are valid only when the fluid velocity, as well as the Reynolds number, are small (for most applications R, is less than 10):

R, =u)Ti V

(7)

Papadaki and Mcln tire

584

As indicated from Eq. 6, the level of shear-stress mcreases from 0 at the center of the bottom disk (r = 0), to a maximum at the perimeter (Y = R). The maximum reported wall shear-stress on endothehal cells growing at the bottom disk is in the range of 7-15 dyn/cm2 (20). Parallel-disk devices are not widely used, because they have the same experimental problems as the cone-and-plate devtces, and, in additton, the bulk phase shear-stress 1snot constant throughout the gap (d), even at low Reynolds numbers (6). Furthermore, this device cannot be effectively used to study the effects of pathophysiological levels of shear-stress on endotheltal cells. From Eqs. 6 and 7, tt 1spossible by reducing the gap, d, to elevate shear-stress wtthout increasing Reynolds number. However, with small gap size, nutrient supply is limited and vtabihty problems are observed (20) When the upper rotating cylinder is placed in a cylmdrical container, the mstrument is called dtsk-andcylinder device One modification of the parallel-dtsk device is the spmnmg disk apparatus (22-24). In this system, both disks are placed inside a cylmdrtcal chamber full of medium buffer. The solutions for the velocity field are derived usmg the boundary layer theory, and the shear-stress at any position on the lower disk surface 1sgiven as (23):

3 z,=[Tr2 +zm2 =0.8yr 1 1z=. [CT..1 l/2

112

V

(8)

where u is the veloctty component m the cn-cumferenttal dtrectton, and z,, z,+, are the radial and the circumferenttal shear-stresscomponents. Assumptions of the model include mfimte medium, mfinlte disk size, lammar flow, and steadystate shear field (22,231. It is necessary that the boundary layer thickness be much smaller than the disk radius, m order for the edge effects to be negligible. In most applications, the boundary layer thickness is l/100 the disk radius. Furthermore, tt has been calculated that the steady state flow regime is achieved withm a few revolutions of the disk. With the spmnmg disk system, changes m cell adhesion, cell attachment, and cell morphology in response to shear-stress have been studied (22-24). 2.4. Radial

Flow Device

The radtal flow chamber 1sdesigned to overcome some of the mechanical problems that were encountered with the spinning disk system (25). This device produces an axisymmetric flow field, providmg a large shear-stress range within a given experiment (25-27). As shown m Fig. 4, the radtal flow chamber conststs of two parallel stationary disks, the bottom one with a central mlet pipe, separated by a narrow spacing (25,28). Tissue-culture medium 1spumped at constant volumetrtc flow rate through the central pipe, and flows radially out


585

Fig 4 (A) Side view of a radial flow chamber Cells, bacteria or hgand are placed on the top disk and fluld IS pumped through a central pipe Shear-stress decreases with distance from the center of the disk h is the distance between the two parallel disks (B) Top view of the flow field Flow cross-sectional area Increases with radius

to a collectlon mamfold If disks of transparent material are used, the test surface can be observed mlcroscoplcally (26). Fluid IS restricted to flow only m the radial dire&Ion, which allows the derivation of simple equations for shearstress gradients across the test surface. Assuming lammar axlsymmetrlc flow and neglecting entry effects, the surface shear-stressand a local Reynolds number can be calculated as (25). 3QP z, = -

m-h’

As the cross-sectlonal

Q ’ R%cal = m-v

area for flow between the two disks increases radi-

ally, shear-stressdecreasesradially with distance from the center. Flow may be turbulent at some small radu (except at the stagnation pomt), showing local Reynolds numbers above 2000, so care must be taken in the selectton of the volumetric flow rates (25). Most radial flow chambers are operated m the laminar flow regime (RelocalC 2000, for r > 0.05 mm). One advantage of this system 1sthat operation IS achieved without moving parts, which decreases problems associated with stenhty, and greatly slmpllfies contmuous monitoring of the cell responses (25). This device has been mostly used to study cell adhesion/detachment to/from surfaces, and ligandreceptor bond affinity (26,28-31). The prmclple of operation 1s that particles (cells, bacteria, llgand) are swept away wlthm a circular zone around the pipe inlet, because of high shear-stress,but at the outer zone, where shear forces are lower, particles remam adherent to the test surface (26). A critical radius, rC,at

586

Papadaki

and Mcln tire

the boundary between those two zones, defines the posrtron where the parttcle-surface adhesive force is equal to the force exerted by the fluid to the parttcle. The shear-stress range examined with this devtce IS between 0 and 70 dyn/cm2 (26,28). A common problem of the radial flow chamber IS a central area of turbulence, around the Inlet, because of a sharp-edged boundary. The area and the pattern of turbulence can be mampulated by modrfymg the geometry of the inlet pipe/plate (25). An easy alternative IS to adjust the volumetric flow rate so that the critical radius IS well outside the turbulent zone 2.5. Linear Shear-Stress Flow Chamber The lmear shear-stress or tapered parallel-plate flow chamber 1sa modrftcatron of the parallel-plate flow chamber that IS designed to generate a linear vartatton of shear-stress wrthm the same flow field, without changmg the flow rate or the gap height (4,32) As shown n-r Fig. 5, the dtfference between the two flow systems 1sIn the gasket design, which, for the linear shear-stress chamber, gives a parallel-plate flow geometry with varying width (32) The idea for this design came from the fact that the velocity field of steady flow of a Newtonran fluid between two parallelplates (also known as Hele-Shaw flow) IS equivalent to the velocity field of a stagnation flow The properties of a two-drmensronal Hele-Shaw stagnation flow permit the design of a flow channel, so that the wall shearstress IS hnearly dtstrtbuted along the center lme of the channel The entrance end of the flow channel can be approxtmated by a straight line, the exit end can be formed by two perpendtcular stratght lmes through the center of the channel. Thus, the shear-stress along the center line IS calculated as (32) z,, = -6PQ 1-t (10) h2w, ( 1 where h IS the channel height, wl IS the entrance wtdth, and L 1sthe total length of the channel. Thus simple flow-channel design gives a constant shear-stress gradient over the entire length of the chamber, The maxtmum value of shear-stress IS at the entrance, and falls to zero at the extt. At the wall, and near the exit of the channel, shear-stressdoes not lmearly decreasewith distance from the entrance Usamr et al. used the linear shear-stress flow chamber to study cell-cell mteractions over a wade range of shear-stress (O-45 dyn/cm2) m a smgle expertmental run. Other forms of a flow channel can be designed, m which the shear-stress will not be zero at the exit, by selectmg a different part of the twodtmensional stagnation flow (32).


587

aseplate

Distance from entrance

Fig 5 (A) Linear shear-stress flow chamber. The only difference with the parallelplate flow chamber IS m the gasket geometry The width between the two parallelplates changes from the entrance to exit of the flow chamber w I IS the entrance width and L IS the length of the chamber (B) For this flow channel design, shear-stress changes from a maximum value at the entrance to zero at the exit along the center lme.

2.6. Cylindrical

Tube

Cylmdrical tube devices have been used to study fluid-surface mteracttons important in thrombosis (633-35). As indicated m Fig. 6, cylmdrical tubes have many srmilarmes with the parallel-plate flow chamber For a fully developed, lammar, viscous flow m tubes, the shear-stress along the tube wall and the Reynolds number are expressed as (6): z, = -4QP , Re =- 2Q (11) nRv 7CR3 where R IS the tube radius. The velocity profile IS parabolic and the shear-stress 1szero along the centerline and maximum along the walls From Eq. 11, for a 11Om radrus R, the critical Reynolds number above whrch turbulent flow occurs 1s 2300, and the correspondmg lammar shear-stress1s4600 dyn/cm2. In general, higher shear-stressescan be produced in a cyhndrrcal tube than m the parallel-plate flow chamber, but the tube must be longer than the parallel-plates (6). The cylmdrlcal tube systems have advantages and disadvantages similar to the parallel-plate devices, since the bulk fluid shear rate and stress vary over the cross-section. At the ends of the tube nonuniform flow areas exist, because fluid flows from a reservoir to the tube, and vise versa. One of the mam problems of the tube device IS that mtcroscoprc monitoring IS difficult because of the optrcal drstortion caused by the curved surfaces (6). This problem can be

588

Papadaki

and Mclntire

A’

\

cell monolayer

Fig. 6. Cylindrical tube. The wall velocity profile is parabolic; shear-stress linearly with r. Cells can be grown on the side walls of the cylinder.

varies

partially solved by embedding the tube in a water-filled chamber. In addition, generation of confluent cell monolayers on small-diameter, closed cylindrical surfaces can be difficult. Redmond et al. have developed a perfused transcapillary endothelial cellsmooth muscle cell co-culture system, marketed by Cellco, which mimics the architecture of the vascular wall (35). They used a bundle of 230 semipermeable capillaries (cylindrical tubes with diameter 330 pm), in which endothelial cells and smooth muscle cells were seeded in the luminal and extracapillary space of the tubes, respectively. Badimon et al., developed a tubular perfusion chamber to investigate the effects of blood flow on biological materials (33,34). This cylindrical chamber is flexible and can accept a variety of test surfaces (such as aortic subendothelium, collagen, and so on). The geometrical configuration of this device differs moderately from the geometry of the cylindrical tube.

2.7. Annular

Flow Chamber

The annular flow chamber has been used to expose endothelium denuded vessel segments to flowing blood (36-39). As shown in Fig. 7, inverted vessel segments (up to three) are mounted on the core of the chamber, and blood is passing through the annular spacing. Assuming laminar flow of a Newtonian fluid in an annulus, the shear rate at the vessel surface can be calculated as (39):

%(k)Q zw= 7c(d,-d,)’

’

‘Y(k) =

(1- k)2 2k(l+

1- k2 - 2k2 ln(1 / k)

k) (1 + k2)ln(l

/ k) + k2 - 1 I

(12)

where do, di are the outside and inside diameters, respectively, k = dildo, and g(k) is a constant. When k goes to zero, this equation reduces to the circular tube equation. As with the circular tube, the major disadvantage of the annular


589 Plexiglas

vessel sigment

tube

rod

Fig 7 Annular flow chamber Inverted vessel segments are drawn on a rod that IS placed mslde a Plexiglas tube Fluid (usually blood) enters the Plexiglas tube, and then streams along the rod, forming an annulus chamber 1s the difficulty m microscopically observmg the cellular processes Additionally, this device requires a preselection of vessels with certain dimensions, and 1snot easily adapted for the use of non vascular materials (33). Studies with the annular chamber have been very important m developing an understandmg of thrombosis, hemostasls, and the causes of various bleeding diseases.

2.8. Flow Chambers Expansion)

with Variable Degrees of Stenosis (or

Devices that model flow through a stenosis (or expansion) are designed to study the response of vascular cells to changes in local fluid-dynamic conditions (40-44). Such studies are very important, since disturbances m the blood flow (m areas of high shear-stress, flow stagnation, recirculation, and turbulence) can significantly contribute to processes such as platelet activation and fibrin deposition. Furthermore, near flow dividers and bifurcations, the onentatlon pattern of endothehal cells often changes. Flow chambers with stenosis (or expansion) are modlficatlons of the parallel-plate flow chamber or the cylindrical tube, and contain a recess m the flow channel (4). A major drawback of these devices is that they may require sophisticated computational fluid mechanics packages for the solution of the Navler-Stokes that describe fluid motion in the flow chamber. Truskey et al. studied the effects of flow separation and recn-culatlon on cultured endothehal cells, with the use of a sudden asymmetric expansion parallel-plate flow chamber (42). Shear-stresses on the order of 80 dyn/cm2 and shear-stress gradients as high as 2500 (dyn/cm*)/cm, were produced m the reclrculatlon zone of this chamber. Barstad et al. developed a parallel-plate

590

Papadakl and Mclntre

perfusion chamber with an eccentric stenosts, to investigate thrombus formatton at the apex of the stenosis (44) In order to allow a gradual, rather than a step, vartatton m cross sectional area of the flow chamber, those mvesttgators introduced the stenosisas an l%mm-long planar surface wtth a OS-mm, cosmeshaped step 3. Concluding Remarks In this chapter, we have described some of the more popular devices used to quantitatively measure shear-stress effects on endotheltal cells, and to address the experimental problems associated wtth them. Some of the flow chambers descrtbed here have been used princtpally to mvesttgate rheologtcal properties of blood cells, or adheston of flowing leukocytes and platelets to surfaces. However, with minor design modtficattons, these devices can also be effectively used to study flow effects on endothehal cells or any other type of anchorage dependent cell The exponenttally increasmg interest m understandmg how fluid shear-stressregulatesthe structureand functton ofvascular cellsmakes necessarythe development and design of tmproved flow-chamber technologtes. Acknowledgments This work was supported by NIH grants HL18672 and NS23326, NASA grant NAGW-5007, Welch Foundatton grant C-0938 and TATP grant 003604. References 1 Resmck, N and Gtmbrone, M A (1995) Hemodynamtc forces are complex regulators of endothehalgene expression.FASEB J 9, 876882 2 Davies, P F (1995) Flow-medtated endothehal mechanotransductton Physzol Rev 75,519-560. 3 Patrick, C W , Sampath, R , and McInttre, L V (1995) Fluid shear stress effects on vascular fimctton, m Bzomedzcal Engzneerzng Handbook (Bronzmo, J D , ed ), CRC, Boca Raton, FL, pp 1636-1655 4. Slack, S M and Turttto, V T (1994) Flow chambers and then standardlzatton for use m studies of tbrombosts on behalf of the subcommittee on Rheology of the Scienttfic and Standardrzatton Commlttee of the ISTH Thromb Haemost. 72, 777-78 1 5 Panaro, N J and McIntne, L V (1993) Flow and shear stress effects on endothehal cell function, m Hemodynamzc Forces and Vascular Cell Bzology (Sumpto, B E , ed ), R G Landers, Austin, TX, pp 47-65 6 Tran-Son-Tray, R (1993) Techniques for studymg the effects of phystcal forces on mammahan cells and measurmg cell mechanical properties, m Physzcal Forces and the Mammalzan Cell (Frangos, J A., ed ), Academic, San Diego, pp l-59 7 Frangos, J. A , E&m, S G , McIntne, L V, and Ives, C L (1985) Flow effects on prostacychn productton by cultured human endothehal cells Sczence 227, 1477-1479 8 Frangos, J. A., McIntire, L V., and Eskm, S. G (1988) Shear stress induced sttmulatton of mammahan cell metabolism. Bzotechnol Bzoeng 32, 1053-1060


591

9 Schmttler, H J , Franke, R P, Akbay, U , Mrowretz, C , and Drenckhahn, D (1993) Improved m vitro rheologtcal system for studymg the effect of fluid shear stress on cultured cells Am J Physzol 265, C289-C298 10 Gopalan, P K., Jones, D A , McInttre, L. V, and Wayne Smtth, C (1995) Cell adhesion under hydrodynamic flow conditions, m Current Protocols zn Immunology (Gotto, R , ed ), John Wiley, New York, pp 7 29 l-7 29 23 11 Koslow, A. R , Stromberg, R R , Frtedman, L I, Lutz, R J., Htlbert, S. L , and Schuster, P (1986) A flow system for the study of shear forces upon cultured endothehal cells. J Blomech Eng 108,338-341 12 Vtggers, R F , Wechezak, A R , and Sauvage, L R (1986) An apparatus to study the response of cultured endothehum to shear stress J Bzomech Eng 108,332-337 13 Eskm, S. G., Ives, C. L , McInttre, L. V, and Navarro, L T (1984) Response of cultured endothehal cells to steady flow Mzcrovasc Res 28, 87-94 14 Rosenhead, L (1963) Lamznar Boundary Layers. Oxford Umverstty, Oxford, UK 15 Cooke, B M , Usamt, S , Perry, I , and Nash, G B (1993) A stmphfied method for culture of endothehal cells and analysis ofadheston of blood cells under condmons of flow Mcrovasc Res 45,3345 16 Dewey, C F , Bussolarr, S R., Gtmbrone, M A , and Davies, P F (1981) The dynamtc response of vascular endothehal cells to flmd shear stress J Blomech Eng 103, 177-185 17 Shmtd-Schonbem, H , Gosen, J V, Hemtch, L , Klose, H J , and Volger, E (1973) A counter-rotating “rheoscope chamber” for the study of the mtcrorheology of blood cell aggregation by mtcroscoptc observatton and mtcrophotometry Mzcrovasc Res 6,36&376 18 Franke, R P, Grafe, M , Schnrttler, H , Setffge, D , Mtttermayer, C , and Drenckhahn, D. (1984) Inductton of human endothehal stress fibers by fluid shear stress Nature 307,648-649 19 Davtes, P F., Remuzzr, A , Gordon, E J , Dewey, C. F., Jr, and Gtmbrone, M A , Jr. (1986) Turbulent fluid shear stress induces vascular endothehal cell turnover m vttro Proc Nat1 Acad Scz USA 83,2114-2117 20. Nomura, H., Ishtkawa, C , Komatsuda, T , Ando, J , and Kamtya, A (1988) A dtsktype apparatus for applying fluid shear stress on cultured endotheltal cells Bzorheology 25,461-470 21 Ando, J , Nomura, H , and Kamtya, A (1987) The effect of fluid shear stress on the mtgratton and prohferatton of cultured endothehal cells. Mzcrovasc Res 33,62-70 22. Hochmuth, R M , Mohandas, N., Spaeth, E. E , Wtlhamson, J R , Blackshear, P L , and Johnson, D W (1972) Surface adhesion, deformatton and detachment at low shear of red cells and white cells. Amer Sot Artzf Int Organs 18, 325-332. 23. Horbett, T A , Waldburger, J. J , Ratner, B. D , and Hoffman, A S. (1988) Cell adhesion to a series of hydrophilic-hydrophobic copolymers studted with a spmnmg disc apparatus J Blom Mater Res 22, 383-404. 24 Weiss, L. (1961) The measurement of cell adhesion. Exper Cell Res Suppl. 8,141-153 25 Fowler, H W and McKay, A J (1980) The measurement of mtcrobtal adhesion, m Microbial adhescon to surfaces (Berkey, R C. W., Lynch, J M , Mellmg, J , Rutter, P R , and Vmcent, B., eds ), John Wiley, Chtchester, pp. 143-161.

592

Papadaki and Mclntlre

26 Cozens-Roberts, C., Qumn, J. A , and Lauffenberger, D A (1990) Receptor-medtated adhesion phenomena Model studies with the Radtcal-Flow Detachment Assay. Bzophys J 58, 107-25 27 Kuo, S C. and Lauffenburger, D A. (1993) Relattonship between receptorihgand bmdmg affinity and adhesion strength Bzophys J 65, 2 19 L-200 28 Duddrtdge, J E , Kent, C A , and Laws, J F (1982) Effect of surface shear stress on the attachment of Pseudomonasfluorescens to stamless steel under defined flow condrttons. Bzotechnol Bzoeng 24, 153-164 29 Cozens-Roberts, C , Lauffenburger, D A , and Quinn, J A (1990) Receptor-mediated cell attachment and detachment kmetics. I. Probabthsttc model and analysts Btophys J 58, 841-856. 30 Cozens-Roberts, C , Qumn, J A , and Lauffenburger, D A (1990) Receptor-medlated ceil attachment and detachment kinetics II Experimental model studies with the radral-flow detachment assay Bzophys J 58, 857-872 3 1 Groves, B J and Riley, P. A (1987) A mmraturtzed parallel-plate shearmg apparatus for the measurement of cell adhesion Cytobzos 52,4%62 32 Usami, S., Chen, H H., Zhao, Y, Chten, S , and Skalak, R (1993) Design and construction of a lmear shear stress flow chamber Ann Boomed Eng 21,77-83 33 Badtmon, L , Turttto, V, Rosemark, J A , Badtmon, J J , and Fuster, V (1987) Characterlzatton of a tubular flow chamber for studying platelet mteraction with btologic and prosthetic materials deposition of mdmm 1I l-labeled plateiets on collagen, subendothelmm, and expanded polytetrafluoroethylene [pubhshed erratum appears m J Lab Chn Med 111,5] J Lab Clm Med 110,70&718 34 Badimon, L , Badtmon, J J , Turnto, V T , and Fuster, V (1987) Thrombosts. studies under flow condmons Ann NYAcad Set 516,527-40 35 Redmond, E M , Cahill, P A , and Sttzmann, J V (1995) Perfused transcapillary smooth muscle and endothehal cell co-culture-a novel m vttro model, zn Vitro Cell Dev Blol Anzm 31,601&609 36 Baumgartner, H R. (1973) The role of blood flow m platelet adhesion, fibrin deposition, and formatton of mural thrombi Mcrovasc Res 5, 167-179 37 Turnto, V T and Baumgartner, H R (1975) Platelet mteraction with subendothelmm m a perfusion system. physical role of red blood cells Mzcrovasc Res 9,335-344 38. Turttto, V T., Weiss, H. J., and Baumgartner, H. R (1980) The effect of shear rate on platelet interaction wtth subendothelmm exposed to cltrated human blood Mlcrovasc Res 19,352-265 39. Turttto, V T. and Baumgartner, H. R. (1979) Platelet mteractlon with subendothehum m flowmg rabbit blood effect of blood shear rate Mzcrovasc Res 17, 38-54 40 Lasstla, R., Badimon, J J , VallabhaJosula, S., and Badtmon, L. (1990) Dynamic momtormg ofplatelet deposition on severely damaged vessel wall m flowmg blood Effects of different stenoses on thrombus growth. Arterloscleroszs 10,306-3 15 41 Badimon, L and Badtmon, J J (1989) Mechamsms of arterial thrombosis m nonparallel streamlmes: platelet thrombt grow on the apex of stenottc severely mJured vessel wall Experimental study m the pig model J Clan Invest 84, 11341144.


593

42 Truskey, G A., Barber, K M , Robey, T C., Ohvrer, L A , and Combs, M P. (1995) Charactertzatton of a sudden expansion flow chamber to study the response of endothehum to flow rectrculatton J Blomech Eng 117,203-2 10 43. Schoephoerster, R T , Oynes, F., Nunez, G , Kapadvaqwala, M , and Dewaqee, M K. (1993) Effects of local geometry and fluid dynamics on regtonal platelet deposmon on arttfictal surfaces Arterzoscler Thromb 13, 1806-I 813 44 Barstad, R M , Roald, H E , Cm,Y, Turitto, V T., and Sakartassen, K S (1994) A perfuston chamber developed to investigate thrombus formatton and shear profiles m flowmg nattve human blood at the apex of well-defined stenoses Arterroscler Thromb 14, 1984-1991

43 Quantitative by Diffusion Athanassios

Modeling

of Limitations

Caused

Sambanis and Sanda A.Tan

1. Introduction Transport of nutrients and metabohtes m many btoarttfictal tissue constructs rebes exclustvely on drffuston, I e., on the presence of a concentratton gradtent between the mstde of the construct and the surrounding milieu. A quantttattve evaluatton of the rate of dtffustonal processes 1s thus essential for properly destgnmg three-dtmenstonal cell-polymer systems, and for assessmg the chemical environment at varrous locales withm the construct Ordinary or bulk dtffusronal transport is generally assumed to follow Ftck’s law, which IS descrrbed by the equatron (l-3) NA=-D i3C,/iYx (1) where NA (mol/m2 s) IS the molar flux of species A in the x-dtrectton, D (m2/s) IS the dtffustvtty coeffictent, and C, (mol/m3) IS the concentration of species A The fraction dCAlax 1sthus the concentratton gradient of spectes A m the x-dn-ectton. Btoarttfictal tissue constructs do not comprtse a homogeneous medium for dtffuston, since they generally contain pores of complex geometry and solids of different structure (e.g., bropolymers and cells) Transport through such a medium 1stypically described by Eq. 1, using an aggregate effective dtffusivtty, Deff, based on the total cross-section of the construct (voids and solids) normal to the dtrectton of dtffuston (3) In certain constructs, dtffuston may not even follow Ftck’s law with an effecttve drffustvtty coeffictent; non-Ftcktan dtffuston will not be consrdered m the context of thuschapter Besides dtffusron, transport m constructs may also occur by convectton. Convecttve transport occurs as a result of a pressure gradient between the construct interior and the surrounding medium, and thus tt is dependent on the From

Methods Edtted

by


Medune, and


M L Yarmush

595

0 Humana

Press

Methods

and Protocols

Inc , Totowa,

NJ

596

Sambanis

and Tan

flow field around the construct One way to assessthe presence of convective fluxes IS to evaluate the transport of a solute from the medium to the construct (or vice versa), with the construct placed m different flow fields. If the rate of transport is dependent on the flow field, and not solely on the concentration gradlent, then this constitutes evidence for the presence of convective fluxes. In this chapter, we describe methods for measuring dlffusivltles for calcium algmate slabs and beads, and for synthetic membranes containing cells. Models of dlffuslon-reaction processes, which can be used to quantltatlvely assess the hmltations caused by diffusion, are discussed The preparation of the constructs themselves IS not described m detail, since this 1snot the focus of this chapter. Convective transport processes are also not examined. The methods presented may be adapted to characterize diffusional llmltatlons m other bloartlficlal tissue constructs, after performing any necessary modlficatlons 2. Materials 2.1. Construct

Preparation

1 Creating an algmate construct algmate (Keltone LV, Kelco, Chicago, IL), calcmm chloride (Fisher, Norcross, GA), delomzed water. 2 Creating alginate beads algmate (Keltone LV; Kelco), calcium chloride (Fisher), syrmge pump (Raze1 Sclentlfic Instruments, Stamford, CT), lo-mL syringe with 18-gage needle, air tank with gas regulator (Fisher), phosphate buffer salme (PBS, Sigma, St LOU, MO) 3 Measurement of bead sizes mlcroscope with grid on ocular lens (Nlkon, Tokyo, Japan) or connected to image-acquisition and processmg system (Perceptlcs, Knoxville, TN)

2.2. Specimen

Preparation

for Diffusion

Experiment

1 Glutaraldehyde treatment and mactlvation of cells 50% glutaraldehyde (Fisher), Hanks’ balanced salt solution (HBSS, Sigma) 2 Measurement of construct volume Dulbecco’s modified Eagle’s medium (DMEM, Sigma), a graduated cylinder

2.3. Diffusivity

Measurement

Across a Slab

1 Two reservoirs with stirrers (Fig. 1); copper-wire mesh with large pores (0 5 mm diameter), epoxy resin 2 Glucose-free DMEM (Sigma), fetal bovme serum (FBS, Sigma), glucose (Fisher), or an air and a nitrogen cylinder with gas regulators

2.4. Diffusivity

Measurement

Across a Hollow Fiber Membrane

1 Two reservoirs with stirrers (Fig. 2) 2 Hollow fiber, gas-impermeable tubmg (Fisher), Vernon Hills, IL) for medium cn-culatlon

perlstaltlc pump (Cole Palmer,

Limitations Caused by Diffusion

597

Reservoir B Construct

Stirrers

Fig 1 Experimental setup for steady-state measurement of diffuslvlty bloartlficlal construct of slab or sheet geometry

across a

Impermeable

probe or sample port

probe or sample port Pump

Ftber-

Reservoir I ’

Fig 2 Experimental low-fiber membrane

Stirrer ’

setup for steady-state measurement of dlffuslvlty across a hol-

598

Sambams and Tan

3 Glucose-free DMEM (Sigma), FBS (Sigma), glucose (Fisher), nitrogen cylinder with gas regulators

2.5. Diffusivity

Measurement

in Alginate

or an air and a

Beads

1 Reservoir with stirrer, e g., a spmner flask (Corning, Cornmg, NY) 2 Glucose-free DMEM (Sigma); FBS (Sigma); glucose (Fisher), or 1251-labeled msulm (Bmax, South Portland, ME)

2.6. Analytical

Techniques

1 Glucose Trmder reagent (Sigma), spectrophotometer, cuvets 2 Protem Blo-Rad protein assay (Blo-Rad, Richmond CA) or BCA protein assay reagent kit (Pierce, Rockford IL), spectrophotometer, cuvets 3 1251-labeled msulm 1251-labeled msulm (Bmax), gamma counter 4 L-glucose L-[~-~H] glucose (New England Nuclear, Boston, MA), scmtlllatlon vials and counter 5 Oxygen, polarographlc oxygen probes and amphfiers (Ingold, Wllmmgton, MA)

3. Methods 3.7. General Considerations The methods described m this section for the quantitative assessment of dlffuslonal limitations fall mto two general categories. steady-state and transient measurements. Steady-state measurements involve more sophisticated expenmental apparatl and pose specific requirements on the geometry of constructs, but the required data processmg IS simpler when compared to that of transient experiments On the other hand, transient measurements generally Involve slmpler expertmental systems, but the processmg of data for the determmatlon of dlffuslon coefficients IS complex. Computatlonally proficient investigators might find transient measurements preferable; those more experimentally mclmed might opt for the steady-state techniques, provided that the construct geometry allows for then Implementation. If dlffuslvltles are to be evaluated m constructs contammg cells, the cells should be included during the measurements, since they generally affect dlffuslvltles Reactions should be prevented from occurring at cell sites during the experiment. This can be accomplished by macttvatmg the cells either before they are Incorporated m the construct, or after the construct fabrlcatlon (#), or by using nonmetabohzable compounds for measurements (5). Since the construct geometry may affect Its pore structure, It IS advisable that, whenever possible, dlffuslvlty measurements be made on constructs havmg the same qualitative and quantitative geometric features as the actual ones Furthermore, smce medium components may affect the surface of constructs, dlffuslvlty measurements should be made m media stmllar to those to which

Limitations Caused by Diffusion the constructs will eventually be exposed. For instance, adsorbed serum proteins may have a stgntticant effect on the transport properties of the construct, so serum should be mcluded m the medium if tt will be present tn the actual working environment of the btoarttficlal tissue.

3.2. Construct

Preparation

1 Creating a slab-gel construct A polymer solutron or a mixture of polymer solution with cells is poured onto a flat surface to form a slab 1 O-5 0 mm m thickness Solutions to cause crosslmkmg are then added. For an algmate gel slab, m particular, a 2% algmate solution m PBS is exposed to 1 1% CaCl, solution for approx 2 h, excess CaC12 1swashed away with deromzed water 2 Creating alginate beads Sodium algmate is drssolved m water to form 2% solution Cells could be added to this solution if they are to be used m construct characterlzatron The solution IS transferred to a syrmge positioned on a syrmge pump and connected to an 1&gage needle. An IS blown parallel to the needle at a rate necessary to obtam beads of the desrred average diameter, as origmally described m ref. 6 Droplets fallmg m a 1 1% CaC12 solution produce calcium algmate beads Beads are washed wtth CaCl*, followed by two washes with 10 mM PBS Beads can be coated with poly-L-lysine, if this IS how the actual construct is prepared (4) 3 Analysis of bead sizes Beads are sized by bemg placed under an Inverted mrcroscope with a grid on the ocular lens, or connected to an image analyzer For each bead, two orthogonal diameters, Dr and DZ, are measured, and size is represented by their geometric mean The absolute difference of the two diameters IS also recorded and compared to the geometric mean, to estimate the degree to which each bead IS actually a sphere (7)

3.3. Specimen

Preparation

for Diffusion

Experimeni

1 Glutaraldehyde treatment Treat construct with 1% glutaraldehyde m HBSS for 1 h Alternatively, cells can be mactivated by bemg exposed to 1% glutaraldehyde solution m HBSS for 1 h prior to fabrication of the construct, followed by washing in HBSS 2 Measurement of construct volume. Make DMEM with 10% FBS Pre-eqmhbrate the construct wrth medium, so that It acqurres its final, constant volume. Add the construct to a graduated cylinder with medium, measure the total volume, asplrate the medium, measure the medium volume, the difference IS the volume of the construct

3.4. Steady-State Diffusivity or a Hollow Fiber

Measurement

Across a Slab

1. In experiments mvolvmg encapsulation, the cells are embedded wlthm a gel matrix and/or are surrounded by a semipermeable membrane For drffusrvrty measurements, a gel with entrapped cells can be fabricated m a slab or a sheet geom-

SambanisandTan

600

2.

3.

4.

5.

6.

etry. The semipermeable membranes used in the bioartificial constructs could be tested in sheet forms or in hollow cylindrical shapes. This section will present the experimental setup for both geometries. If cells are involved in the experiment, they are inactivated as described in Subheading 3.3. Alternatively, the diffusivity of a nonmetabolizable compound could be measured (5). For a gel slab, the construct is placed between two chambers with the help of copper wire mesh (Fig. 1). The wire mesh is supported using epoxy resin. A vertical position of the slab is preferred to a horizontal position, since it reduces the chances of deformation caused by a pressure difference between the reservoirs. The reservoirs should have a volume much higher than that of the construct. For a hollow fiber, a known length of the fiber is placed in reservoir B, with the ends sealed completely to impermeable tubes which lead to reservoir A. The medium from reservoir A is then pumped through the inner lumen of the hollow fiber continuously in a recirculation manner (Fig. 2). Reservoirs A and B are filled with known, constant volumes of medium containing a high and a low (or zero) concentration of the test molecule, respectively. For glucose, the two initial concentrations could be 25 mM and 0 mM, for oxygen, the medium in one reservoir could be equilibrated with air and in the other with nitrogen. Equilibration can be achieved by stirring the medium in the presence of the appropriate gas for about 30 min. Both reservoirs must be well mixed. Samples are withdrawn at known time-points for subsequent assays, or, for oxygen, dissolved oxygen probes are placed within each chamber to monitor the concentration changes (8). For the system of Fig. 1, assuming that there is negligible resistance to mass transfer from the liquid bulk to the surface of the construct, i.e., there are no significant boundary layer effects; that the partition coefficient of the test molecule between the liquid medium and the construct is 1; and that a steady-state concentration profIle has been established within the slab (a good approximation when the concentrations in the reservoirs change slowly with time), then the equation describing diffusion across the construct of slab geometry is: N = AD&C*

- CB)IL

(2)

In the above equation, N is the flux of the test molecule across the slab construct (mol/s); CA (Cn) are the concentrations of the test molecule in reservoir A (B) (mol/m3); A is the cross-sectional area of the construct across which diffusion occurs (m2); Deff is the effective diffisivity (m*/s); and L is the thickness of the slab (m). Equation 2 can be coupled with the equation describing conservation of mass of the test molecule: N = -d (V,C,)ldt

= d (V&,)Idt

(3)

where VA (Vn) is the (constant) medium volume in reservoir A (B) (m3). Equations 2 and 3 can be solved analytically yielding the following relationship between Cn and t:

Limitations Caused by Diffusion ln{[aC*, + G, --al

601

+ aw(c*,

- G,)l~ = [(I + ~wlw,d~~BP

(4)

where cx = V,/Va is the ratio of medium volumes in the two reservoirs, and C,, (Cn,) is the initial concentration of the test molecule in reservoir A (B) (mol/m3). Therefore, if measurements are made on Ca as a function of time, the values of the left-hand side of Eq. 4 can be plotted against time. The slope of the resulting straight line is calculated by the least squares method, and the effective diffusivity is evaluated as Deff = -(slope)(aLVn)l[(

1 + a)A]

(5)

7. In applying Eq. 5, initial points collected before the establishment of a steadystate concentration profile in the slab should be discarded; i.e., the initial points of the foregoing graph are expected to not fall on a straight line, and should not be taken into account. 8. An alternative approach is to start with a slab initially empty of test molecule and to calculate the effective diffusivity by fitting to the data the transient diffusion equation am

= D,~

(awax2)

with boundary conditions C(x=O)=C* C(x=L)=Ca and initial condition c (t = 0) = 0 In the above equations, C is the concentration of test molecule in the slab (mol/ m3) and x is the linear dimension along which diffusion occurs (m). This analysis is somewhat more involved mathematically than that of the steady-state system. Details can be found in ref. 9. 9. For the system of Fig. 2, if the same assumptions as in step 6 apply, and, additionally, the system does not deviate significantly from the slab geometry, i.e., R2 -RI is much smaller than RI, where R, and R2 are the inner and outer radii of the fiber, then the equations describing this system are the same as those listed for the slab construct. If measurements are made on Cn as a function of time, the values of the lefthand side of Eq. 4 can be plotted against time. The slope of the resulting straight line is calculated by the least squares method and the effective diffusivity is evaluated from Eq. 5.

3.6. Transient

Measurement

of Diffusivities

in Beads

1. The following protocol has been employed to evaluate difhtsivities of glucose and insulin in hydrogel beads containing insulin-secreting cells (4). However, the same protocol could be used, possibly with some modifications, to evaluate difisivities of various compounds in different types of three-dimensional tissue-engineered constructs. 2. Place the beads in a stirred vessel, such as a spinner flask (Corning) containing medium free of test molecule, and having a volume 5-10x the volume of the

Sambanrs

602

and Tan

construct Stir the medmm, and ensure that there IS movement of medium relattve to the beads Repeat this washing procedure twice to ensure a zero concentration of test molecule m the beads 3 Transfer the beads mto a spinner flask contammg medium wtth the test molecule The medium volume should not exceed 3x the volume of the beads Remove medium samples for subsequent assays Initially, samples should be removed frequently, 1.e , approximately every minute; later m the experiment, the sampling frequency may be reduced 4 Assay for the test molecule m collected samples 5. Use the followmg equations, derived from Eq. 1 and mass balances, to describe the dynamics of the system (4)

auat= Deff[aWW+ (2/r)(ac/ar)] V,, (dC,ldt) = -D,,A

@Cl&),

=R

(6)

(7)

Imtial condmons C (t = 0, r) = CO(r)

(8)

Cb (t = 0) = Cbo

(9)

Boundary condtttons dcldr (r = 0, t) = 0

(10)

C (r = R, t) = C, (t)

(11)

In the above equations, C is the concentratton of test molecule m beads (mol/m)), t is time (s), Deffls the effective dtffustvlty (m2/s), r 1sthe radial posttton m beads (m), R 1sthe bead radius (m), CO(r) is the mittal concentration of test molecule m

beads as a function of radius (mol/m3) (zero m the described experiment), C, is the concentration of test molecule m beads at the surface, Ct, IS the concentratton of test molecule m the hquid bulk (mol/m3), C,, 1sthe mmal concentratton of test molecule m the liquid bulk (mol/m3); Vb 1s the volume of medium surrounding the beads (m3), and A is the total surface area of beads (m2) If there are no boundary layer effects around the beads, and the partmon coefficient of the test molecule between beads and liquid medium is 1, then c, = c,

(12)

The system of Eqs. 7-12 may be solved numertcally, as such Alternatively, one may use its analytical solution for diffusion of solute from the ltqutd bulk into beads unttally empty of solute, which is given by (4,ZO) CI,/C,, = [a/( 1 + a)] [ 1 + C”,, = I 6( 1 + a) exp(-D,ffq2,,t/R2)l(9

+ 9a + q2na2)] (13)

where a 1sthe ratto of ltqmd volume to total bead volume and qn 1sthe nth root of the equation tanq, = 3qn/(3 + aq,)

(14)

Limitations

Caused

by Diffusion

603

6 Fit Eq. 13 to the experimental data by mimmizmg the sum of the squares of the differences between calculated and measured values, with the effective dlffuslvlty, Deff, bemg the fitted parameter Mimmization routines are available m a variety of commercially available software programs, including MIcrosoft Excel (Microsoft, Bothell, WA) The above procedure allows for the calculation of the effective dlffuslvlty for transport of the test molecule from the surrounding medium mto the construct 7. To measure the effective dlffuslvlty for transport from the construct mto the surrounding medium, the construct equilibrated with medium as m step 3 of this subheading 1stransferred mto medium without test molecule, the medium 1sagltated as in step 2, and medium samples are collected for subsequent assays as m step 3

8 Assay for the test molecule in collected samples 9 The dynamics of the system are described by Eqs. 6-12, which, for diffusion of solute from beads mto the liquid bulk, mltlally contammg no solute, have the following analytical solution* C,/C,, = [l/( 1 + a)] [ 1 - C”, = , [6a( 1 + a) exp(-D,fm2,t/R2]/9

+ 9a + q2,,a2)] (1.5)

with a and qn as defined above Fit Eq. 15 to the experimental data by mmlmlzmg the sum of the squares of the differences between calculated and measured values, with the effective diffuslvlty, DeK,being the fitted parameter

3.7. Analytical

Techniques

1. Glucose concentrations are measured spectrophotometrlcally, using an assay kit based on the Trmder reagent (Sigma), according to manufacturer’s protocols 2 Total protein concentration m a sample IS measured using a dye-binding assay (Blo-Rad) or BCA protein assay reagent kit (Pierce), according to manufacturer’s protocol 3 12sI-labeled msulm concentration 1s measured by placing a certain sample volume m a gamma counter and measuring the counts over a certain period of time (4) Radiolmmunoassays on nonlabeled, msulm-related peptldes yield standard errors too high for an accurate determination of diffuslvitles (4). 4 If L-glucose 1s used as a nonmetabolizable compound, the concentrahon of its trltlated form, L-[ 1-3H] glucose, can be measured by scmtillatlon counting (5) 5 Dissolved oxygen concentrations are measured with polarographlc dgssolvedoxygen probes connected to amplifiers

3.8. Quantitative Modeling of Reaction/Diffusion Processes The importance of diffusional resistances in the transport of nutrients and metabohtes to and from constructs should be evaluated m relation to the intrmsic reaction kmetlcs, rather than m an absolute fashion. In other words, the same extent of diffuslonal resistance may be limiting the observed overall rate of a process if the intrinsic kinetics are fast, and it may not be limiting if the mtrinslc kinetics are slow.

Sambanis and Tan The traditional engmeering approach in evaluatmg these effects mvolves calculatmg the value of the dimensionless Thiele modulus 4, which can be defined as (3) where Vis the volume of the construct(m3),A its outside surfacearea(m2) (so that V/ A is a characteristiclength), r, is the reaction rate per unit volume m the absenceof mass-transferInnnations (mol/m3s), K is a Michaelis constantfor the intrinsic reaction kmetics(mol/m3), and& is the effective diffusivity (m2/s).If t/iis much smaller than 1,the reaction rate IShrmtmg and diffusional resistancesare msignlficant; if 4 IS much greater than 1, the overall processis l~ted by difisional transport. The mtrmstc reaction kinetics, 1e , the kinetics of nutrient consumption and metabohte production by cells, can be evaluated m cultures that do not exhibit any significant mass transfer resistances, e.g., with cells cultured as monolayers A more accurate evaluation of diffusional limitations can be obtamed by combmmg the intrinsic kmetics with the mass transport equations m a mode1 of the three-dimensional construct (4,12). Such models can evaluate the concentration profiles of nutrients and metabohtes across the construct, thus allowmg an assessmentof the chemical environment to which cells are exposed m various locales; these models may also be used to simulate the overall behavior exhibited by the construct m different environments (411). 4. Notes There are several potential pitfalls associated with the experimental setups for the quantttattve measurement of diffustonal resistances. Some of these pitfalls will be discussed m this section. While making concentration measurements to calculate diffusivity across a construct, one must take mto account the possibility of adsorption of protems on the construct and on the walls of the contaming vessel. One method of overcommg this problem is to pretreat the construct and vessel with the protein solution, so as to saturate all the attachment sitesbefore the experiment. It is recommended that, prior to the experiment, a test be carried out to confirm the absence of protem loss by adsorption. The experimenter should also ensure that sample collectmg devices and storage vials be pretreated and washed thoroughly, to avoid additional errors, especially when low concentrations of molecules are involved. In experimental setups, a modified version of the actual construct may need to be used. When this is the case,care should be taken to mimic the geometric shape of the origmal construct, because the pore geometry may change when the three-dimensional architecture is altered. Furthermore, care slhould be taken to avoid deforming the construct when posittonmg it m the experimental apparatus and mstallmg the momtoring equipment.

Limitations Caused by Diffusion

605

Although most proteins can be stabthzed by lowering the temperature, any alteration of the environment should be avotded, and, if possible, the experiment should be carried out in an environment similar to that m which the construct will be used. This IS especially important for hydrogels, which could change their physical properties, such as their pore geometry, with temperature. This also raises the question of stab&y of the test molecule for the duration of the experiment, as proteins tend to degrade over a long period of time. Therefore, a control should always be run alongside to account for any degradation of protein, especially in longer-term, steady-state experiments. Last but not least,care should be taken to minnmze the mass-transferresistance causedby boundary layer effects. If increasingthe stnrmg rate m reservoirs increases the value of the measured effective dtfisivity, then this is indicative of boundary layers and/or convective transport within the construct.If boundary layer effects cannot be abolished by increasing the stirrmg rate, then the diffuslonal resistancetmposed by the constructis probably too small to be of any particular sigrnlicance References Bennett, C 0 and Myers, J E (1974) Momentum, Heat and Mass Transfer, McGraw-Hill, New York, pp 48 1-5 12 Bird, R B., Stewart, W E , and Ltghtfoot, E. N (1960) Transport Phenomena, John Wtley, New York, London and Sydney, pp 495-5 18. Blanch, H W and Clark, D S (1996) Bzochemzcal Engzneenng, Marcel Dekker, New York, pp 103-161 Tztampazts, E and Sambams, A (1995) Tissue engmeering of a btoarttfictal pancreas modeling the cell envn-onment and device function Bzotechnol Prog 11, 115-126 Casctan, J J , Sotirchos, S V, and Sutherland, R. M. (1988) Glucose dtlfusivtty m multtcellular tumor spheroids Cancer Res 48,3905-3909 Vorlop, K -D and Klem, J. (1983) New developments m the field of cell m-nnobtlizatton-formation of brocatalysts by tonotroprc gelatton, m Enzyme Technology III. Rotenburg Fermentation Symposwm 1982 (LeiYerty, R M , ed ), Sprmger-Verlag, New York, pp. 2 19-235. 7. Papas, K K , Constantmtdts, I , and Sambanis, A (1993) Culttvatton of recombtnant, msulm-secretmg AtT-20 cells as free and entrapped spheroids. Cytotechnology 13, 1-12. 8 Sun, Y, Furusakt S., Yamaucht, A , and Ichtmura, K (1988) Dtfhtstvrty of oxygen mto carrters entrapping whole cells Bzotech. Bzoeng 34,55-58 9 Hannoun, B. J M and Stephanopoulos, G. (1985) Dtffuston coeffictents of glucose and ethanol m cell-free and cell-occupied calcmm algmate membranes Blotechnol

Bloeng. 28, 829-835.

10 Crank, J (1975) The Mathematics of D@ks~on, 2nd ed , Clarendon, Oxford 11 Hannoun, B J M and Stephanopoulos, C. (1990) Growth and fermentatton model for algmate-entrapped Saccharomyces cerevwae Bzotechnol Prog 6, 349-356.

Tissue Engineering

Tissue Engineering

Tissue Engineering