Developmental Biology (9th Edition)

DEVELOPMENTAL BIOLOGY NINTH

EDITION

Companion Website www.devbio.com

devbio.com is an extensive Web companion to the textbook that is intended to supplement and enrich courses in developmental biology. Here you will find addi­ tional information for advanced students as well as historical, philosophical, and ethical perspectives on issues in developmental biology. Included are articles, movies, interviews, opinions (labeled as such), Web links, updates, and more. (There is even a developmental biology humor page!)

J

The site is comprised of Web topics relevant to all areas of the textbook, organized by textbook chapter. These topics are referenced throughout the textbook, both within the chapter, and at the end of the chapter. From the home page of the site, you can browse by chapter, search all topics, or quickly jump directly to any specif­ ic topic. The site also includes the complete Literature Cited for the textbook, most with links to the PubMed database.

DevBio Laboratory:

vade mecum3

REGISTRATION CODE

Scratch the area below to reveal your unique code.

Important Note: Each code is valid for only one registration. If this

code has been revealed, codes for DevBio laboratory: vade mecunr can be purchased online at http://labs.devbio.com.

it may no longer be valid. New registration

DevBio Laboratory: vade mecum3

An Interactive Guide to Developmental Biology Mary

Tyler and Ronald

Kozlowski

http://labs.devbio.com S.

N.

vade mecum3 is available online,

Designed to complement the textbook, this unique resource helps you understand the organisms discussed in lecture and prepares you for the laboratory. DevBio laboratory:

which allows you the flexibility to use the software from any computer with Internet access. Access to the site is included with each new textbook.

Over 140 interactive videos and 300 labeled pho­ tographs take you through the life cycles of model organisms used in developmental biology labora­ tories. The easy-to-use videos provide you with the concepts, vocabulary, and motivation to enter the laboratory fully prepared. A chapter on zebrafish addresses how to raise the organism and the effects of various teratogens on embryon­ ic development. The site also includes chapters on: the slime mold Dictyostelium discoideum; planarian; sea urchin; the fruit fly Drosophila melanogaster; chick; and amphibian.

ADDITIONAL FEATURES INCLUDE: •

Movie excerpts from Differential

Expressions2: Short video excerpts about key concepts in development.

• PowerPoint® slides of chick wholemounts and serial sections for self-quizzing and

creating tests. •

Full video instruction on histological techniques

• G l ossa ry: Every chapter of the Laboratory Manual includes an extensive glossary.

DEVBIO LABORATORY

VADE MECUM3

._-' _ ....-h_ -_. -� -� �� --

'..... &l.oO_ --"'-"

�-

---

-t.. . __ ._­ ---__ w _.-

::::.::.-

•

laboratory Manual: The Third Edition of Mary S. Tyler's laboratory manual ,

Developmental Biology: A Guide for Experimental Study, designed for use with the multimedia chapters of DevBio Laboratory: vade mecum3

• Website: www.devbio.net augments

DevBio laborator y, allowing convenient

access to recipes from the lab book, a searchable glossary, a module on Laboratory Safety, and more.

• Study Questions

• laboratory Skills Guides

'==:::::--. I Gene transa-iption

FIGURE 3.28 Wnt signal transduction pathways. (AI The canonical Wn t pathway. The Wnt protein binds to its receptor, a member of the Frizzled family of proteins. In the case of certain Wnt proteins, the Frizzled protei n then activates Disheveled, al lowi ng it to become an inhibitor of glycogen synthase kinase 3 (GSK3I. GSK3, if it were active, would prevent the dissociation of p catenin from the APe protein, So by i nhibiti ng GSK3, the Wnt signal frees r>-catenin to associate with an LEF or TCF protein and become an active tra n scri ption factor. (B) In a pathway that regul ates cell morphology, division, and move­ ment, certain Wnt proteins activate Frizzled i n a way that causes Frizzled to activate the Disheveled protein, which has been tethered to the plasma membrane (through the Prick­ le protein). Here, Disheveled acti vates Rae and RhoA protei ns which coordinate the cytoskeleton and which can a lso regu l ate gene expression. (q In a third pathway, certain Wnt proteins activate Frizzled receptors in a way that releases calcium ions and can cause Ca2+-dependent gene expression. -

,

�-catenin Transcription

I

pathway, the GSK3 protein is an inhibi tor that is itself repressed by the Wnt signal. THE "NONCANONICAL" WNT PATHWAYS The pathway described above is often called the "canonical" Wnt path­ way because it was the first one to be discovered. Howev­ er/ in addition to sending signals to the nucleus, Wnt can also affect the actin and microtubular cytoskeleton. Here, Wnt activates alternative, " noncanorucal," pathways. For instance, when Wnt activates Disheveled, the Disheveled protein can interact with a Rho GTPase. This GTPase can

activate the kinases that phosphorylate cytoskeletal pro­

teins and thereby alter cell shape, cell polarity (where the

upper and lower p or tions of the cell differ), and motility

(Fig u re 3.28B; Shulman et al. 1998; Winter et al. 2001). A third Wnt pathway diverges earlier than Disheveled. Here, the Frizzled receptor protein activates a phospholipase (PLC) that synthesizes a compound that releases calcium ions from the endoplasmic reticulum ( Figure 3.28C). The released calcium can activate enzymes, transcription fac­ tors, and translation factors. It is probab le that the Frizzled proteins (of which there are many) can be used to couple different signa l transduc­ tion cascades to the Wnt signal (see Chen et al. 2005) and that different cells have evolved to use Wnt factors in dif­ ferent ways.

94

CHAPTER 3

-It

BMP FAMI LY

The TGF-f3 superfamily There are more than 30 structurall y related members of the TGF-p superfamily: and they regulate some of the most important interactions in development (Figure 3.29). The proteins encoded by TGF-p sup erfa mily genes are processed such that the carboxy-terminal region contains the mature peptide. These p eptides are dimerized into homodimers (with themselves) or heterodimers (with other TGF-p pep tides) and are secreted from the cell. The TGF­ P superfamily includes the TGF-p family, the activin fam­ ily, the bone morphogenetic proteins (BMPs), the Vgl fam­ ily, and other proteins, including glial-derived neurotrophic factor (GDNF; necessary for kidney and enteric neuron dif­ ferentiation) and Mi.i.llerian inhibitory factor (which is involved in mammalian sex determination). TGF-p family members TGF-Pl, 2, 3, and 5 are impor­ tant in regulating the fomlation of the extracellular matrix between cells and for regulating cell d ivision (both posi­ tively and negatively). TGF-pl increases the amount of extracellular matrix e pitheli al cells make (both by stimu­ lating coliagen and fibronectin synthesis and by inhibiting matrix degradation). TGF-p proteins may be critical in con­ trolling where and when epithelia branch to form tile ducts of kidneys, lungs, and salivary glands (Daniel 1989; Hard­ man et a1. 1994; Ritvos et al. 1995). The effects of the indi­ vidual TG F- p family members are difficult to sort out, because members of the TGF- p family a ppea r to function similarly and can compensate for losses of the others when expressed together. The members of the BMP family can be d istinguished from other members of the TGF-p superfamily by having seven (rather than nine) conserved cysteines in the mature polypeptide. Because they were originally discovered by their ability to induce bone forma tion, they were given the name bone motphogenetic proteins. But it turns out that bone formation is only one of their many functions; the BMPs are ext remely multifunctional. ' They have been found to regulate cell diviSion, apoptosis (programmed cell death), cell migration, and differentiation (H"gan 1996). They include proteins such as BMP4 (which in some tis*TCF stands for "transforming growth factor. " The designation " superfamily" is often given when each of the different classes of molecules constitutes a "family." The members of a superfamily all have similar structures but are not as dose as the molecules within a family are to one another. tOne of the many reasons why humans don't seem to need an enor­ mous genome is that the gene products-proteins-hwolved in our construction and development often have many functions. Many of the proteins we are familiar with in aduJts (such ilS hemoglobin, keratins, insulin, and the like) do have only one function, which led to the erroneous conclusion that this is the norm. lndeed, the "one_ function-per-entity" concept is a longstanding one in science, hav­ ing been credited to Aristotle. Philosopher John Thorp has called this lJIol1olelism (Greek, "one end") "Aristotle's worsl idea."

GDFIO BMP3/osteogenin BMP9 Dorsalin 1 (chicken) BMPIO Vgr2/GDF3

�

GDFS (brachypodism) BMPI3/GDF6 BMPI2IGDF7

-C

BMPS (short ear) BMP6/Vgrl BMP7/0PI BMP8a/OP2 BMP8b (mouse) - 60A (Drosophila)

-[

-

�r

BMP2 BMP4 Dpp (Drosophila)

-c VgI (Xenopus)

rC L[

Univin (sea urchin) GDFI Screw (Drosophila) Nodal Activin PA Activin PB

ACTlVlNS

TGF-pi TGF-pS TGF-p2 TGF-p3

TGF-p FAMrLY

MrS GDF9 lnhibin GDNF

FIGURE 3.29 Relalionships among members of Ihe TGF-p su per­ family. (Aher H ogan 1 996.)

sues causes bone formation , in other tissues causes cell death, and in othe r instances specifies the epidermis) and BMP7 (which is m i portant n i neural tube polarity, kidney development, and sperm formation). As it (rather oddly) tUrns out, however, BMFl is not a member of the BMP family at all; it is a protease . The Drosophila Decapen taplegic (Dpp) protein is homol­ ogous to vertebrate BMP4, and human BMP4 can replace Opp and thus "rescue" dpp-deficient flies (Padgett et al. 1993). BMPs are thought to work by diffusion from the cells

CELL-CELL COMMU N I CATION IN DEVELOPMENT

TGF-� superfamily ligand

t

Receptor n

t

p

Receptor I

t

Two dimers bind ligands

Rece tor type !

Receptor typ e U

Smad 2,3

Smad activation

smad4

Smad dimerization

t

Activated (phosphorylated) Smad

New tran c ri ti on I

s p

;/

FIGURE 3.30 The Smad pathway activated by T G F � superfamily li gands (A) An activat ion complex is formed by the binding of the ligand by the pe I and type II receptors. This allows the type II receptor to phosphorylate the type I receptor on particular serine or threonine residues (of the "GS box"). The phosphorylated pe I receptor protein can now phosphorylate the Smad protei s (B) Those receptors that bind TGF�p family proteins or members of the activin family phosphorylate Smads 2 a d 3. Those receptors that bind to BMP family proteins phosphorylate Sm ads 1 and s . These Smads can co ple with Smad4 to form active transcrip­ tion factors. A si pl i fied version of the pathway is shown at the leh. ·

.

ty

m

m x

n

n

ty

.

/

t

t

Smad 1,5

�

/��== '

95

t

t

. � ) . �J Gene transcription or repression

phorylates a serine or threonine on the type I receptor, thereby activating it. The activated type I receptor can now phosphorylate the Smad' proteins (Figure 3.30A) . Smads 1 and 5 are activated by the BMP family of TGF-p factors, while the receptors binding activin, Nodal, and the TGF­ � family phosphorylate Smads 2 and 3. These phosphory­ lated Smads bind to Smad4 and form the transcription fac­ tor complex that wiU enter the nucleus (Figure 3.308). Other paracrine factors

producing them. Their range is determined by the amino acids in their N-terminal region, which determine whether the specific BMP will be bound by proteoglycans, thereby restricting its diffusion (Ohkawara et aJ. 2002). The Nodal and activin proteins are also members of the TGF-p superfamily. These proteins are extremely impor­ tant in specifying the different regions of the mesoderm and for distinguishing the left and right sides of the verte­ brate body axis.

Although most paracrine factors are members of one-of the four families described above, some of these proteins have few or no close relatives. Epidermal growth factor, hepa tocyte growth factor, neurotrophins, and stem cell factor are not included among these families, but each of these factors plays important roles during development. in addition, there are numerous paracrine factors involved almost exclusively with developing blood cells: erythro­ poietin, the cytokines, and the interleukins. These factors will be discussed when we detail blood cell formation in Chapter 12.

THE SMAO PATHWAY Members of the TGF-p superfamily

activate members of the Smad family of transcription fac­ tors (Heldin et aJ. 1997; Shi and Massague 2003). The TGF­ P ligand binds to a type II TGF-p receptor, which aUows that receptor to bind to a type I TGF-p receptor. Once the tvvo receptors are :in close contact, the type II receptor phos-

*Researchers named the Smad proteins by merging the names of the first identified members of this family: the C. eiegans SMA pro­ te-in and the Drosophila Mad protein.

CHAPTER 3

96

SIDEL IG H TS SPECULATIONS

O T

(A) c. degans

Cell Death Pathways be, or not to be: th at is the question " While we all are poised at life-or-death deci­ sions, this existenlial d ichotomy is exceptionally stark for embryon i c cells.

Ii

.

Programmed cell death, or apoptosis,'

is a normal part of development (see Baehrecke 2002). In the nematode C. e/egans, i n which we can count the number of cells, exactly B 1 cells die

according to the normal developmental pattern. Al l the cells of th i s nematode are "programmed" to die unless they are acti vely told not to u n dergo apop­ tosis. In humans, as many as 1 0 1 1 cells die in each adult each day and are replaced by other cells. ( I ndeed, the mass of cells we lose each year through normal cell death is close to our entire body weight!) Within th e uterus, we were constantly maki ng and destroying cells, and we generated about three times as many neurons as we eventual­ ly ended up � ith whe n we were born. lewis Thomas (1 992) has a ptl y noted, By the ti me I was born, more of me had died than survived. It was n o wonder I cannot remember; du ring that time I went th rough brain after brain for nine months, finally contriving the one model that could be human, equipped for language.

Apoptosis is necessary not o nly for the proper spacing and orientation of neurons, but also for generating the middle ear space, the vaginal opening, and the spaces between ou r fi ngers and

toes (Saunders and Fallon 1 966;

Roberts and M i l l er ·1 998; Rodriguez et al. 1 997). Apoptosis prunes unneeded structures (frog tails, male mammary tissue), control s the number of cel1s i n

*111e term apoplOsis (both ps are pro­ nounced) comes fro m the Greek word for the natural process of leaves fa l l in g from trees or peta ls falling from fl owers. Apoptosis is an active process that can be subject to evo l u tion ary selection. A second type of ce ll death, necrosis, is a pathological death caused by ex terna l factors such as i n flam mat i on or tox i c i nj ury .

particular tissues (neu­ rons in vertebrates and flies), and sculpts com­ plex orga n s (pa lat e ,

retina, digits, and heart). Different tissues use different s ignals for apoptosis. One of the Signals often used i n vertebrates is bone

morphogenetic protein 4 (BMP4). Some tis­ sues, such as connec­ tive tissue, respond to BMP4 by differentiat­ ing into bone. Others, s uc h as the frog gastru­ la ectoderm, respond to BMP4 by differenti­ ating into skin. Still others, such as neural crest cells and tooth pri mordia, respond by deg rading their DNA and dying. In the developing tooth, for i nstance, numerous growth and differentia­ tion factors are sec ret ed by the enamel knot ­

.

Aher the cusp has grown, the enamel knot synthes izes BMP4 and shuts itself down by apop tosis (see Cha pter 1 0; Vaahtokari et a l . 1 996). In other tissues, the

I EGL-I I

1 ICED-91 1

ICED-41

�

ICED-31

j

(B) Mammalian neurons

\

� f-- OO

�I Bax I

Intracellular membrane

�

I Ap.f1

ICaspase-9 I

t

IC.spase-31

+

Figure 3.31 Apoptosis pathways in nematodes and mam­ mals. (A) In C. elegans, the CEO-4 protein is a protease-acti ­

vating factor that can activate the CEO-] protease. The CEO­ ] p rotease initiates the cell destruction events. CEO-9 can inhibit CE04 (and CEO-9 can be in hibi ted upstream by EGl-l). (B) tn mammals, a similar pathway exists, and appears to function in a similar manner. I n th is hypothe tical scheme for the regu lation of apoptosis in mammalian neu· rons, BcI-XL (a member of the Bcl2 family) binds Apafl and prevents it from activating the precursor of caspase-9. The signal for apoptos is a l lows an other protein (here, Bik) to inhibit the binding of Apafl to BcI-Xc. A pafl is now able to b in d to the caspase-9 precursor and cleave it. Ca spase-9 dimerizes and activates caspase- 3 , which initiates apoptasis. The same colors are used to represent homologous proteins. (After Adams and Cory 1 99B.)

cells are " program­ med" to die, and will rem a in alive only if some growth or differ­ entiation factor is present to "rescue" them. This happens during the develop­ ment of mammalian red blood cells. The red blood cell precursors in the mouse liver need the hormone erythro­ poietin in order to survive. If they do not receive it, they undergo apoptosis. The erythropoietin receptor works through the JAK-STAT pathway, activat­ ing the StatS transcription factor. In this way, the amount of erythropoiet i n pres-

ent can determine how many red blood cells enter the c i rc ul ation One of the pathways for apoptosis was la rge ly deli neated th rough genetic studies of C. e/egans. Indeed, the importance of this pathway was recog­ nized by awarding a Nobel Prize to Sydney Bren n er, Bob Horvitz, and Jonathan Su lston i n 2002. tt was found that t he proteins encoded by the ced-3 and ced-4 genes were essential for .

CELL-CELL COMM U N I CATION I N DEVELOPMENT

97

SIDELIGHTS & SPECULATIONS (Continued) apoptosis, and that i n the cells that did not undergo apoptosis, those genes were turned off by the product of the ced-9 gene (Figure 3.31A; H e nga r tner et a l. 1 992). The CED-4 protein is a protease-activating factor that activates the gene for CED-3, a protease that ini­ tiates the destruction of the cell. The CEO-9 protei n can bind to and i nacti­ vate CED-4. Mutations that i nactivate the gene for CED-9 cause numerous cells that would normally survive to activate their ced-3 and ced-4 genes a nd die, leading to the death of the entire embryo. Conversely, ga i n -of­ function mutations in the ced-9 gene cause its protein to be made in cells that would n ormall y die, resulting i n those cel ls' survival. Thus, the ced-9 gene appears to be a b in ary switch that regu lates the choice between life and death on the cellular level. It is possi­ ble that every cell in the nematode emb ryo is poised to die, with those cells that survive being rescued by the activation of the ced-9 gene. The CED-3 and CED-4 proteins are at the center of the apoptosis pathway that is common to all animals studied. The trigger for apoptosis can be a developmental cue such as a particular molecule (e.g., BMP4 or glucocorti ­ caids) or the loss of adhesion to a matr ix . Either type of cue can activate CED-3 or CED-4 proteins or inactivate CED-9 molecules. I n mammals, the homologues of t he CED-9 protein are members of the Bel2 fami Iy (wh ich i neludes Bel2, Bel-X, and similar pro­ teins; Figure3.31B). The functi ona l simi­ larities are so strong that if an active human BCL2 gene is placed in C. e/e­ gans em b ryos, i t prevents no rrnally occu rri ng cell death (Vaux et al. 1 992). The mammalian homologue of CED-4 is Apaf"l (apoptot i c protease' activating factor 1 ), and it parti cipates

in the cytochrome c-dependent activa­ tion of the mammalian CEO-3 homo­ logues, the proteases caspase-9 and caspase-3 (see Figu re 3.31 B; Shaham and Horv itz 1 996; Cecconi et a l . 1 99B; Yoshida e t al . 1 998). Activation of the caspase proteins results in a utodigest ion-caspases are strong proteases that digest the cell from within, cleav i ng cellular proteins and fragmenting the DNA. While apoptos is-defi cient nema­ todes deficient for CED-4 are viable (despite hav i ng 1 5% more cells than w i ld-type worms), mice with 105s-of­ function m u ta tions for either caspase-J or caspase-9 d i e around birth from massive cell overg rowth i n the nervous system (Figure 3.32; Kuida et al. 1 996, 1 998; Jacobson et al. 1 997). Mice homozygous for targeted deletions of Apaf7 have simi larly severe craniofa­ cial abnormalities, brain overgrowth, and webbing betvveen their toes . There are instances where cell death is the normal state unless some

Figure 3.32 D isru ption of normal brain development by block­ ing apoptosis. In mice i n which the genes for caspase-9 have been knocked out, normal neural apoptosis fails to occur, a nd the over­ pro liferation of brain neu rons is obvious. (A) 6-day embryonic wild-type mouse. (B) A caspase-9 knockout mouse of the same age. The enlarged brain protrudes above the face, and the limbs are still webbed. ((,0) This effect is confirmed by cross sections through the forebrain at day 13.5. The knockout exhibits thick­ ened ventricle walls and the near-obliteration of the ventricles. (From Kuida et al. , 998.1

ligand "rescues" the cells. I n the chick neural tube, Patched protein (a Hedge­ h og receptor) wi II activate caspases. The binding of Sonic hedgehog (from the notochord and ventral neu ra l tube cells) suppresses Patched, and the cas­ pases are not activated to sta rt apopto­ sis (Thibert et al. 2003). Such "depend­ ence receptorsN probably prevent n eural cells from prol i ferati n g outside the proper tissue, and th e loss of such receptors is associated with cancers (Porter and Dhakshinamoorty 2004). Moreover, we will soon see that cer­ ta i n epithel ia l cells must be attached to the extracellular matrix i n order to function. If the cell is removed from the matrix, the apoptosis pathway is activated and the cell dies Gan et al. 2004). This, too, is probab ly a mecha­ nism tha t prevents cancers once cells have lost their adhesion to extracellu­ lar matrix proteins.

See WEBSITE 3.2 The uses of apoptosis

(A) caspase-9+1+ (Wild type)

(B) caspase-9-1- (Knockout)

(C) Wild type

(D) Knockout

98

CHAPTER 3

ergoes a confonnational change that enables a part of its cytoplasmic domain to be cut off by the presenilin-l pro­

und

J uxtacrine Signaling

In juxtacrine interactions, proteins from the indUCing cell

tease. The cleav ed portion enters the nucleus and binds to

interact with receptor proteins of adjacent responding cells

withou t d iffusing from the cell producing it, Two of the

bound to the Notch protein, the CSL transcription factors

most widely used families of juxtacrine factors are the Notch proteills (whlch bind to a family of ligands exempli­

Schweisguth 1998; Schroeder et al. 1998; Struhl and Adachl

a dormant transcription factor of the CSL family. When activate their target genes ( F i gu re 3.33; Lecourtois and

fied by the Delta protein) and the eph receptors and their

1998). Thls activation is thought to involve the recruitment

ephrin ligands, When the ephrin on one cell binds with

of histone acetyltransferases (Wallberg et a!. 2002). Thus,

the eph receptor on an adjacent cell, signals are sent to each

Notch can be considered as a transcription factor tethered

of the two cells (Davy et al. 2004; Davy and Soriano 2005), These signals are often those of either attraction or repul­ sion, and ephrins are often seen where cells are being told

to the cell membrane. When the attachment is broken, Notch (or a piece of it) can detach from the cell membrane

and enter the nucleus (Kopan 2002).

where to migrate or where boundaries are fonning. We will

Notch proteins are involved in the formation of numer­

see the ephrins and the ep h receptors functioning in the

ous vertebrate organs-kidney, pancreas, and heart-and

formation of blood vessels, neurons, and somites. For the

they are extremely important receptors in the nervous sys­

moment, we will look at the Notch proteins and their lig­

tem, In both the vertebrate and Drosophila nervous systems, the binding of Delta to Notch tells the receiving cell not to

become neural (Chitnis et al. 1995; Wang et al. 1998). In the

ands.

vertebrate eye, the interactions between Notch and its lig

The Notch pathway: juxtaposed ligands and receptors

­

ands seem to regulate which cells become optic neurons and which become glial cells (DoISky et a!. 1997; Wang et

r

While most known regulato s of induction are dillusible

a!. 1998). Notch proteins are also important in the pattern­

inducing cell surface. In one such pathway, cells express­

ing of the nematode vulva. The vulval precursor cell clos­

in their cell mem­

est to the anchor cell becomes the central vulva cell, and

branes activate neighboring cells that contain Notch pro­

tral vulval cells by signaling to them through its Notch

tein in their cell membranes. Notch extends through the

homologue, the LIN-12 receptor (BeISet et a!. 2001),

proteins, some inducing proteins remain bound to the ing the Delta, Jagged, or Serrate proteins

cell membrane, and its external surface contacts Delta,

Jagged, or Serrate proteins extending out from an adja cent

this cell is able to prevent its neighbors from becoming cen­

See WEBS ITE 3.3 Notch mutations

cell. When complexed to one of these ligands Notch ,

(B)

(A)

FIGURE 3.33 Mechan i sm of Notch activity. (A) Prior to Notch signaling, a CSL transcription factor (such as Suppres­ sor of hairless or CBF1 ) i s on the enhancer of Notch regulated genes. The CSl binds repressors of transcription. (B) Model for the activation of Notch. A ligand (Delta. Jagged. or Serrate protein) on one cell binds to the extracellular domain of the Notch protein on adjacent cel l. This binding causes 1\ shape change the intracellular domain of Notch, which acti­ vates a protease, The protease cleaves Notch and allows the i ntracel lular region of the Notch prolein to enter the nucleus and bind the CSL transcription factor. This intercellular regio of Notch displaces the repressor proteins and binds activators of transcription, including the histone acetyl­ transferase p300. The activated CSl can then transcribe its target genes. (After . Koziol·Dube. Pers. Comm.)

Signaling cell

-

ir

!e id ei

Ie e­

.'1'

0_

SA" 4� 5B

(A) Normal cleavage in 8- to 1 6-cell sea urchin embryos, seen from the animal pole (upper

s

115

J6-Ceil

8

6

pole

Side

Cleavage under pressnre 8-Ceil

J6-Cell

H--j--+-l - Hii5-*:-i;Tt--1 Animal

SPECIFICAT ION

monious equipotential system because a l l of its potentially independent parts inter­ acted together to form a single organism. D riesch's experiment implies that cell inter­ action is critical for normal development. Moreover, if each early blastomere can iorm all the e mbryon ic cells when Isolated, then it follows that, in normal develop­ ment, the comm u n i ty of cells must prevent it from doing so (Hamburger 1 997). "

Third, Driesch concluded that the fate of a cell depended solely on its location in the embryo. The interactions belWeen ,ells determined their fates. We now know (and

w i l l see in Chapters 5 and 7) t h at sea urch i n embryos and frog embryos use both autonomous and conditional ways of specifying their ea rl y embryonic cells. More-

�Th is idea of nuclear equivalence and the ability of cells to interact eventually caused Driesch to abandon science. Oriesch, who thought the embryo was l i ke a machi ne, could not explain how the embryo could make its missing parts or how a cell could change its fate to become another cell type. Harking back to Aristotle, he invoked a .it,,1 force, entelechy ("internal goal-di rected force") to explain how development pro­ ceeds. Essentially, he believed that the embryo was imbued with an internal psyche and the wisdom to accomplish its goals despite the obstacles embryologists placed in t5 path. However, others, especially Oscar Hertwig (1 894), were able to incorporate �riesch's experiments into a more sophisticated experimental embryology, which will be d iscussed i n the introduction to Part IV of this book.

IjB

116

PART I I over, they use a s i m i l ar strategy a n d even similar molecules. I n the 1 6-cell sea urchin embryo, a group of cells called the micromeres inherit a set of transcription factors from the egg cytoplasm. These transcription factors cause the m icromeres to devel­ op autonomously i nto the larval skeleton. These transcription factors also activate the genes for paracrine factors and juxtacrine factors that are employed by the m icromeres to conditionally specify the cells around them. I n the late b lastula frog embryo, the cel ls located opposite the point of sperm entry i n herit a set of transcription factors from the egg cytoplasm as well. These cells are autonomously i nstructed to form the head endodermal cells. They are also autonomously specified to produce and secrete paracrine factors that w i l l condition a l l y induce the cells around them to form the brain. So the head endoderm is specified autonomousl y, whereas the brain is speci­ fied conditional l y, in part, by the head endoderm. Those embryos (especially vertebrates) wherein most of the early b lastomeres are conditionally specified have traditiona l l y been called regulative embryos. But as we become more cognizant of the manner in which both autonomous and conditional specification are used in each embryo, the notions of IImosaic" and I'regulative" devel­ opment are appearing less tenable. Indeed, attempts to get rid of these distinctions were begun by no less an embryologist than Edmund B. Wi l son (1 894, 1 904). Wil­ son (a student of the above-mentioned W. K. Brooks) was one of the first scientists to

theorize that chromosomes i n the nucleus put forth cell-specifying factors into the cytoplasm. "Mosaic embryos," he wrote, received these factors from the cytoplasm of the egg during cleavage stages, w h i le the nuclei of "regulative embryos" were i nstructed by other cells to produce these factors later in development.

Morphogen G radients and Cell Specification Throughout this book, we will see many instances of cell fate specification that involve morphogen gradients. A morphogen (Greek, "form-giver") is a diffusable biochemical molecule that can determine the fate of a cell by its co ncentrati on . ' Morphogens can be transcription factors produced within cells (as in the Drosophila embryos described i n the following section). They can also be paracrine factors that are produced i n one group of cells and then travel to another population of cells, specifying the target cells differentially according to the concentration of morphogen. Uncommitted cells exposed

to high concentrations of the morphogen (nea rest its source of production) are speci­ fied as one cel l type; when the morphogen's concentration drops below a certain threshold, the cells are determined to another fate. When the concentration falls even lower, a cell of the same initial uncommitted type is spec ified in yet a third manner. Morphogen gradients provide a very important mechanism for conditional specifi­ cation. The existence of morphogen gradients as a force i n de lo pment and regener­ ation was predicted by Thomas Hunt Morgan (1 905, 1 906---before he became a geneti­

ve

cist), but it was many years before these gradient models were exten ded to explain how cells might be placed in specific positions along an embryonic axis (Horstadiu s 1 939; Toivonen and Saxen 1 955; Lawrence 1 966; Stumpf 1 966; Wolpert 1 968, 1 969). Lewis Wolpert i l lustrated such a gradient of positional information using the "French flag" analogy. Imagine a row of "flag cells," each of which is "mul tipotential"---

/

FIGURE 5.23

B

Spiral cleavage of the mol lusc Trochus v iewed from the animal pol e (AI and from one side (BI, The cells deri ved 'i-om the A blastomere are shown in color. The mitotic spindles, 5 'e tc hed in the ear ly stages, divide the cells unequally and at an angle to the vertical and horizontal axes. Each successive quartet of rnicromeres (indicated with lowercase letters) is d i sp l aced to the iight or to the left of its sister macromere (uppercase letters), creat­ ..ng the cha racteristic sp iral pattern.

placed to the right or to the left of its sister macromere, cre­ ating the characteristic spiral pattern. Looking down on the embryo from the animal pole, the upper ends of the mitotic spindles appear to alternate clockwise and coun­ :erclockwise (Figure 5.24) . This arrangement causes alter­ nate micromeres to form obliqu�ly to the left and to the right of their macromeres. At the third cleavage, the A macromere gives rise to hvo daughter cells, macromere 1A and micromere 1a. The B, C, Mid D cells behave sirnilarly, producing the first quartet of micro meres. In most species, these micromeres are to the right of their macromeres (looking down on the animal pole). At the fourth cleavage, macromere 1A divides to :orm macromere 2A and micromere 2a, and micromere 1a divides to form tvvo more micromeres, 1a1 and 1a2 (see fig­ ure 5.23). The micromeres of this second quartet are to the left of the macromeres. Further cleavage yields blastomeres 3_-'\ and 3a from macromere 2A, and micromere 1a2 divides :0 produce cells 1a21 and 1a22. In normal development, the first-quartet micromeres form the head structures, while ilie second-quartet micromeres form the statocyst (balance organ) and shell. These fates are specified both by cyto-

plasmic localization and by induction (Cather 1967; Clement 1967; Render 1991; Sweet 1998). The orientation of the cleavage plane to the left or to the right is controlled by cytoplasmic factors ir1 the oocyte. This was discovered by analyzing mutations of snail coiling. Some snails have their coils opening to the right of their shells (dextral coiling), whereas the coils of other snails open to the left (sinistral coiling). Usually the direction of coiling is the same for all members of a given species, but occasional mutants are found (i.e., in a population of right­ coiling snails, a few individuals will be found with coils that open on the left). Crampton (1894) analyzed the embryos of such aberrant snails and found that their early cleavage differed from the norm. The orientation of the cells after the second cleavage was different in the sinis­ trally coiling snails as a result of a different orientation of the mitotic apparatus (Figure 5.25). In some species (such as the pond snail Physa, an entirely sir1istral species), the sinistrally coiling cleavage patterns are mirror images of the dextrally coiling pattern of the right-handed species. In other instances (such as Lynmaea, where about 2% of the snails are lefties), sinistrality is the result of a two-step process: at each division, the initial cleavage is radial; how­ ever, as the cleavage furrow forms, the blastomeres shift to the left-hand spiral position (Shibazaki et a1. 2004). In Figure 5.25, one can see that the position of the 4d blas­ tomere (which is extremely important, as its progeny will form the mesodermal organs) iS,different in the two types of spiraling embryos.

See WEBSITE 5.2 Alfred Sturtevant and the genetics of snail coiling

1 80

CHAPTER 5

(A) FIGURE 5.24 Spiral cleavage in molluscs. (A) The spiral nature of third cleavage can be seen i n the confocal fluorescence micrograph of the 4-cell embryo of th e clam Acila cas­ trenis. Microtubules stain red, RNA stains green, and the DNA stain s yellow. Two cells and a portion of a third cell are visible; a polar body can be seen at th e top of the micro­ graph. IB-E) Cleavage in the mud snail lIyanassa obsoleta. The 0 blastomere is larger than the others, al lowing the identification of eac h cell. Cl eavage is dextral. (B) 8-cel l stage. PB, polar body. IC) Mid-fourth cleavage 11 2-cell em bryo) . The macro meres have already divided i n to large and small spirally oriented cells; 1 a-d have not divided yet. (D) 32-cell embryo. (A cou rtesy of G. von Dassow and the Center for Cell Dynamics; B-E from Craig and Morrill 1 986, courtesy of the authors.)

(B)

(D)

(C)

In snails such as Lymnaea, the direction of snail shell coil­ ing is controlled by a single pair of genes (Sturtevant 1923; Boycott et al. 1930). In Lymnaea peregra, rare mutants exhibiting sinistral coiling were found and mated with wild-type, dextrally coiling snails. These matings showed that the right-coiling allele, D, is dominant to the left-coil­ ing allele, d. However, the direction of.deavage is deter­ mined not by the genotype of the developing snail but by the genotype of the snail's mother. A dd female snail can produce only sinistrally coiling offspring, even if the off­ spring's genotype is Dd. A Dd individuaL will coil either left or right, depencling on the genotype of its mother. Such matings produce a chart like this: G e n otyp e

DD " x dd rJ

->

Dd

DD rJ x dd "

->

Dd

Dd x Dd

....

tDD:2Dd: ldd

Ph e notype

All eft oili ng All

All right-coiling

l -c

right-coiling

The genetic factors involved in snail coiling are brought to the embryo by the oocyte cytoplasm. It is the genotype

of the ovary in which the oocyte develops that determines which orientation cleavage will take. When Freeman and Lundelius (1982) injected a small amount of cytoplasm from dextrally COiling snails into the eggs of dd mothers, the resulting embryos coiled to the right. Cytoplasm from sinistrally coiling snails did not affect right-coiling embryos. These findings confirmed that the wild-type mothers were placing a factor into their eggs that was absent or defective in the dd mothers. Just as in sea urchins (and vertebrates), the right-left axis comes to be defined by the Nodal family of para crine fac­ tors. In the case of snails, Nodal activates genes on the right side of dextrally coiling embryos and on the left side of sinistrally coiling embryos. Changing the direction of cleav­ age (using glass needles) at the 8-ce11 stage changes the location of Nodal gene expresion (Grande and Patel 2009; Kuroda et aJ. 2009). Nodal appears to be expressed in the C-quadrant micromere lineages (which give rise to the ectoderm). This signal induces the expression of the gene encoding the Pitx transcription factor (a target of Nodal protein in vertebrate axis formation) in the neighboring D­ quadrant blastomeres.

FI o

., '(

7 T

EARLY DEVELOPMENT IN SELEOED INVERTEBRATES

( A)

Left-handed coiling

(E )

Right-handed coiling

A AB Zygote

B

CD D 4d

Looki ng down on the animal pole of leh-coiling and right-coiling WI snails. The origin of sinistral and dextral coiling can be traced to the orientation of the mitotic spindle at t he second cleavage. Left- and righ t-co i l i ng snails develop as mir­ ror images of each other. (Aher Morgan 1 927.1 FIGURE 5.25 AI

The snail fate map

The fa te maps of Ilyanassa obsoleta and CrepUlula Jornicafa were constructed by injecting specifi c micromeres with large polymers conj uga ted to fluorescent dyes (Render 1997; Hej no et a!. 2007). The fluorescence is maintained

l

La Left eye; left velum; apical plate 2a Left vdnm; left stomodeum; upper half left statocyst; lA mantle edge; upper left foot 3a Left velu left esophagus 3AB Velar retractor; digestive glands Ib Velum; apical plate 2b Velum; dorsal stomodeum; I B mantle edge; foot retractor muscle 3b Right vehlm; right esophagus VeJar retractor; digestive glands Ie Right eye; right velum; right tentacle; apical plate 2e Right velum; right stomodeum; upper half right statocyst Ie dorsal mantle edge; upper right foot; heart 3c velum; right statocyst; 2C Right right half of foot; mantle edge 3C Velar retractor; digestive glands; style sac Left velum near eye; apical plate 2d Mantle edge; tip of foot Left velum; left statocyst; ID 3d heart; lef! half of foot 2D MEl Left velar retractor; 4d part of intestine (ME) ME2 Right - retractor; � c cvelar 3D heart; kidney; part of intestine 4D Lumen of digest.ive glands; yolk 01;

C

4d

181

Id

32-

Fate map of ifyana" a 0650/eta. Beads containing Lucifer Yellow were i jected into individual blastomeres at the cell stage. When the embr os developed into larvae, their descen­ dants could be identified by their fluorescence. (After Render FIGURE 5.26

1 997.)

n

y

over the period of embryogenesis and can be seen in the

larval tissue derived from the n i jec ted ceUs. The results of the Tlyanassa studies, shown n F igure 5.26, indicated that the second-quartet rnicromeres (20-d) generally contribute to the shell-forming mantle, the velum, the mouth, and the heart. The third-quartet micromeres (3a-d) generate large

i

1 82

CHAPTER 5

(C)

(A)

FIGURE 5.27 Assoc iation of decapentap/egic (dpp) mRNA with specific centrosomes of lIyanassa. (A) In situ hybridization of the mRNA for the 8MP-like paracr ine factor Opp in the 4-cell snail embryo shows no Opp accumulalion. (81 At prophase of the 4- to 8 cell stage, dpp mRNA (black) accumulales at one centrosome of the pair forming the mitotic spindle. (C) As mitosis continues, -

regions of the foot, velum, esophagus, and heart. The 4d cell-the mesentoblast-contributes to the larval kidney, heart, retractor muscles, and intestine.

dpp mRNA is seen to attend the centrosome in the macromere rather than the centrosome in the micromere of each cell. The dpp message encodes a BMP-like paracrine factor critical to mol­ luscan development. (From Lambert and Nagy 2002, courtesy of L. Nagy.) rial but extrudes it again prior to second cleavage. After this division, the polar lobe is attached only to the D blastomere, which absorbs its material. From this point on, no polar lobe is formed. Crampton

The polar lobe: Cell determination and axis formation

(1896) showed that if one removes the polar

lobe at the trefoil stage, the remaining cells divide normal­ ly. However, instead of a normal trochophore larva/ the

Molluscs provide some of the most impressive examples of both mosaic development-in which the blastomeres

(intestine) and mesodermal organs (such as the heart and

are specified autonomously-and of cytoplasmic localiza­

retractor muscles), as well as some ectodermal organs (such

tion, wherein morphogenetic determinants are placed

as eyes; Figure

in a

result is an incomplete larva, wholly lacking its endoderm

5.29). Moreover, Crampton demonstrated

specific region of the oocyte (see the Part II opener).

that the same type of abnormal larva can be produced b y

Autonomous specification of early blastomeres is especial­

removing the D blastomere from the 4-ole cytop l sm (A), then m igr ti ng _p the presumptive posterior sur· ce of egg (B) ocalize in t e 84."1 l astomere C. ( rom

yt

a

in

a the and becoming d h b F Nishida and Sawada _001 , courtesy of H. Nishida and . Satoh.)

(A)

enchyme. However, FCF Signals from the endoderm prevent these mesenchyme precursors from developing into muscle celis, as we will see later. "'Macho-l mRNA is also localized in the cells that become the mes­

(B)

(e)

1 90

CHAPTER 5

(A)

(B)

(C)

staining of - a e n protein ho nd derm FIGURE 5.38 A ntib po e ,1 Ciona em b r . formation. (A) No p ca e n is seen in the nuclei i n he endoderm pre­ (6) the " O-cell stage. (e) not c or precursor bec m e endoderm and e p e s alkaline normal endoderm; the black show notochordal e pres in g endodermal enzymes. (From Imai et al. 2000, ourtesy H .

ody p c t ni s ws its involvement with e o - t ni in the animal l nuclei of a O-cell yo In contrast, nuclear p-catenin is readily seen t vegetal cursors at When p-catenin is expressed in o h dal cells, those cells will o x r s enclodermal markers such as phosphatase. The white arrows show arrows cells that are x s c of Nishi­ da and N. Satoh.) ically normal. Moreover, B4.1 blastomeres isolated from 11U1cho-I-depleted embryos failed to produce muscle tissue. Nishida and Sawada then injected Inacho-l mRNA into cells that would not normally fonn muscle, and found that these ectoderm or endoderm precursors did generate muscle cells when given macho-I mRNA. Macho-l turns out to be a transcription factor that is required for the activation of several mesodermal genes, including muscle acHl1, myosin. tbx6, an d snail (Sawada et aJ. 2005; Yagi et al. 200Sa). Of these gene products, only the Tbx6 protein produced muscle differentiation (as Macho-l did) when expressed in cells ectopically. Macho-1 thus appears to directly activate a set of tbx6 genes, and Tbx6 proteins activate the rest of muscle development (Yagi et al. 200Sb). Thus, the macho-I message is found at the right place and at the right time, and these experiments suggest that Macho-I protein is both necessary and sufficient to promote muscle differentiation in certai� ascidian cells. The Macho-1 and Tbx6 proteins also appear to activate the muscle-specific gene snail. Snail protein is important in preventing 8rachyury (T) exp ressio n in presumptive muscle cells, and is therefore needed to prevent the mus­ cle precu rsors from becoming notochord cells. It appears, then, that the Macho-l transcription factor is a critical com­ ponent of the tunicate yellow crescent, muscle-forming cytoplasm. Macho-l activates a transcription factor cas­ cade that promotes muscle differentiation while at the same time inhibiting notochord specification. AUTONOMOUS SPECI FICATION OF T H E E N DODERM: p­ CATENIN Presumptive endoderm originates from the veg­ etal A4.1 and B4.1 blastomeres. The specificati on of these cells coincides with the scription factor

localization of p-catenin, a tran­ discussed earlier in regard to sea urchin

endoderm specification. Inhibition of p-catenm results in the loss of endoderm and its replacement by ectoderm in the ascidian embryo ( Figure 5.38; Imai et aL 2000). Con­ versely, Increasing p-catenin synthesis causes an increase in the endoderm at the expense of the ectoderm (just as in sea urchins). The p-catenin transcription factor appears to function b y activating the synthesis of the homeobox tran­ scription factor Lhx-3. Inhlbition of the lhx-3 message pro­ hibits the differentiation of endoderm (Satou et al. 2001). CONDITIONAL SPECIFICATION OF T H E .MESENCHYME AND NOTOCHORD BY THE ENDODERM While most of the mus­

cles are specified autonomously from the yellow crescent cytoplasm, the most posterior muscle cells form through conditional specification by ceB interactions with the descendants of the A4.1 and b4.2 blastomeres (Nishida 1987, 1992a,b). Moreover, the notochord, brain, heart, and mesenchyme also form through inductive Interactions. In fact, the notochord and mesenchyme appear to be induced by the fibroblast growth factor that is secreted by the endo­ derm cells (Nakatani et aL 1996; Kim et aL 2000; Imai et al. 2002). The posterior cells that will become mesenchyme respond differently to the FGF signal due to the presence of Macho-1 in the posterior vegetal cytoplasm (Figure 5.39; Kobayashi et aL 2003). Macho-1 prevents notochord induc­ tion in the mesenchymal cell precursors by ac tiva ting the snnil gene (which will in turn suppress the activation of 8rachyllry). Thus, Macho-l is not only a muscle-activating determinant, it is also a factor that distinguishes cell response to the FGF Signal. These FGF-responding cells do not become muscle, because FGF also activates cascades that block muscle formation-another role that is con­ served in vertebrates. As can be seen in Figure 5.40, the

:

EARLY DEVELOPMENT I N SELECTED I N V E RTEB RATES

FIGURE 5.39 Th e two-step process for specifying the marginal cells of the tuni­ cate embryo. The first step involves th e acquis ition (or nonacquisition) by the cells of the Macho-1 transcripti on factor. The secon d step involves the reception or non reception) of the FGF signal from :he endoderm. (After Ko bayash i et al. 1003.)

Anterior ,---,---,_-,

en �� Macho-1 RNA"-erm fla nki ng Ihe venlral midline. (C) Dorsal view of a slightly older embryo, showing pole cells and posterior endoderm sinking into the embryo. (0) Dorsolateral view an embryo at fullest germ band extension, just prior to segmentation. The cephalic _trOW separates the future head region (procephalon) from Ihe germ band, which will m the thorax and abdomen. (E) lateral view, showing fu llest extension of the germ :.and and the beginnings of segmen tation . Subtle indentations mark the incipient seg· .....nts along the germ band. Ma, Mx, and Lb correspond to Ihe mandibular, max i l l ary, :-.l labial head segments; Tl-T3 are the thor.cic segment'; and A l-A8 are the axIominal segments. (F) Germ band reversing direction. The true segments are now .s.ible, as well as the other territories of the dorsal head, such as the c1ypeolabrum, ",ephalic region, optic ridge, and dorsal ridge. (G) Newly halched firsl-instar larva. otographs courtesy of F. R. Turner. D aher Campos-Ortega and Hartenstein t 98S.)

.

(G)

208

CHAPTER 6

FIGURE 6.5 Schematic representation of gastrulation in Drosophila. Anterio r is to the left; dorsal is facing upward. (A,B) Surfac e and cutaway views showing the fates of the tissues immediately prior to gastrulation. (C) The begi n ni ng of gastrul atio n as the ven tra l mesoderm i nvaginates into the embryo. (0) This view corresponds to Figure 6.4A, while (E) correspon ds to Figure 6.4B,C. I n (E), the neuroectoderm is largely differentiated into the nervous system and the epidermis. (After Campos� Ortega and Ha rtenstein 1 985.)

the anterior and posterior ends of furrow. The p ole cells are internalized a long with the endoderm ( Figure 6.4B,C) . At this time , the embryo b ends to form the cephalic furrow. The ectodermal cells on the surface and the mesoderm tmdergo convergence and extension, migrating toward the ventral midline to form the germ band, a collection of cells along the ventral midline that includes all the cells that will form the trunk of the emb ryo. The germ band extends pos­ teriorly and, p erha p s because of the egg case, w ra ps around the top (dorsal) surface of the embryo (Figure 6.40) . Thus, at the end of germ band formation, the cells destined to form the most posterior larval structures are located immediately behind the future head region (Figure 6.4E) . At this time, the b ody segments begin t o appear, dividing the ectoderm and mesoderm. The germ band then retracts, placing the presumptive posterior segments at the poste­ rior tip of the embryo ( Fi gure 6.4F) . At the dorsal su rface, the two sides of the epidermis are brought together in a process called dorsal closure. The amnioserosa, which had

Internal ectoderm

Amnioserosal covering of

( A)

to form two p ocke ts at

the ventral

(B)

(C)

(D)

been the most dorsal structure, interacts with the epider­

I t

cells to encourage their migration (reviewed in Pan­ filio 2007; Heisenbe rg 2009). While the germ band is in its extended position, sever­ al key morphogenetic processes occur: organogenesis, seg­ mentation, and the segregation of the imaginal discs.' In addition, the nervous system forms from mo regions of mal

�� ---'�

ventral ectoderm. Neuroblasts differentiate from this neu ­

rogenic ectoderm ",ithin each segment (and also from the nonsegmented region of the head ectoderm). Therefore in insects like Drosophila, the nervous system is located ven� tra lly, rather than be ing derived from a dorsal neural tube as in verteb rates (Figure 6.S) . The general body plan of Drosophila is the same in the embryo, the larva, and the a dul t each of which has a dis­ tinct head end and a distinct tail end, between which are repeating segmental units ( Figure 6.6). Three of these seg­ ments form the thorax, while another eight segments form the abdomen. Each segment of the adult fly has its own identity. The first thoracic segment, for exa mple, has only

Epidermis

,

,

*lmaginal discs are those cells set aside to produce the adult struc­

in Chapter 1 5 . For more information on Drosophila developmental

tures. The details of imaginal disc differenti(ltion will be discussed

(lnatomy, see Bate and Martinez�Arias 1993; Tyler and Schetzer 1996: and Schwatm 1997.

Nervous system

legs; the second thoracic segment has legs and wings; and the third thoracic segment has legs and halteres (balancing organs). Thoracic and abdominal segments can also be dis­ tinguished from each other by differences i.n the cu tiele of the newly hatched first-instar larvae.

GENES THAT PATTERN THE

D ROSOPH I LA BODY PLAN

Most of

the genes involved in shaping the· larval and adult 1990s u sing a powerful "forward genetics" approach. The basic strat­ egy was to randomly mutagenize flies and then screen for mutations that disrupted the normal formation of the body forms of Drosophila were identiiied in the early

THE G E N ETICS OF AXIS SPECIFICATION IN A)

DROSOPHILA

20 9

(B) Head

Metatho rax

Abdominal gments

FIGURE 6.6 Comparison of larval l eft) and adult (right) segmen­ tation in Drosophila. In th e adult, the three thoracic segments can be distinguished by their appendages: Tl (protherax) has legs o n l y; T2 m esoth e ra x) has wings and legs; T3 (metatherax has nalteres (not visible) and legs. (BI Segmen ts i n ad u lt tra nsgen ic Drosophila in which the gen e fo r green fl uorescen t protei n has n fused to the cis-regulatory region of the engrai/ed gene. Thus, GFP is produced in the areas of engraifed transcription, ... hich is active at the border of each segment and in the posterior l e es . compartment of the wing. (B courtesy of

(A)

(

(

)

A. K b 1

Primary Axis Formation during Oogenesis The processes of embryogenesis may "offiCially" begin at fertilization, but many of the molecular events critical for Drosophila embryogenesis actually occur during oogene­ sis. Each oocyte is descended from a single female germ cell-the oogonium-which is surrounded by an epithe­ lium of follicle cells. Before oogenesis begins, the oogonium

divides four times with incomplete cytokinesis, giving rise

to

16

interconnected cells:

15

nurse cells and the single

oocyte precursor. These 16 cells constitute the egg

cham­

ber (ovary) in which the oocyte will develop, and the oocyte will be the cell at the posterior end of the egg cham­ Ian. Some of these mutations were quite fantastic, and

ber (see Figure 16.4). As the oocyte precursor develops,

included embryos and adult flies in which specific body

numerous mRNAs made in the nurse cells are transport­

structures were either missing or in the wrong place.

These

:JIutant collechons were distributed to many different lab­

ed on microtubules through the cellular interconnections into the enlarging oocyte.

oratories. The genes involved in the mutant phenotypes were cloned and then characterized with respect to their expression patterns and their functions, This combined

Anterior-posterior polarity in the oocyte

effort has led to a molecular understanding of body plan

The follicular epithelium surrounding the developing

Drosophila that is unparaUeled in aU of biol­

oocyte is initially uniform with respect to cell fate, but this

y, and in 1995 the work resulted in a Nobel Prize for

uniformi ty is broken by two signals organized by the

evelopment in

E ward Lewis, Christiane Niisslein-Volhard, and Eric .. 'jeschaus,

The rest of this chapter details the genetics of Drosophi­

the same gene, gllrken. The gllrken message appears to be oocyte nucleus. Intereshngly, both of these signals involve synthesized in the nurse cells, but it becomes transported

development as we have come to upderstand it over the

specifically to the oocyte nucleus. Here it is localized

past two decades. First we will examine how the dorsal­

between the nucleus and the cell membrane and is trans­

'entral and anterior-posterior axes of the embryo are estab­

lated into Gurken protein (Caceres and Nilson

2005).

At

. hed by interactions between the developing oocyte and

this time the oocyte nucleus is very near the posterior tip

.- surrounding follicle cells. Next we will see how dorsal­

of the egg chamber, and the Gurken signal is received by

-entral patterning gradients are formed

within the embryo,

the follicle cells at that position through a receptor protein

d how these gradients specify cli!ferent tissue types. The

encoded by the torpedo gene< (Figure 6.7A). This signal

rmed along the anterior-posterior axis, and how the dif­

results in the "posteriorization" of these follicle cells ( Fig-

=w-d part of the discussion will examine how segments are

:erent segments become specialized. Finally, we will briefly !low how the positioning of embryOniC tissues along the ..'o primary axes specifies these tissues to become partic­

�,

":ar organs.

*Molecular analysis has established that gurken encodes a homo­ logue of the vertebrate epidermal growth factor (EGF), while torpe­ do encodes a homologue of the vertebrate EGF recep tor (Price et a1. 1989; Neuman-Silberberg and Sch upbach 1993).

CHAPTER 6

210

(A)

Posterior

Anteri or Nurse cells

Nucleus

o

T rpedo (Curken receptor )

e

T r min al follicle c lls

e

Uncommitted

polar fol l icle ce lls

o

j

o

o

(8)

Anterior follicle cells

Curken protein

Microtubules

o

Posterior follicle

(D)

bicoid mRNA

Anterior border cells

oskar mRNA in

associa tion with ki nesin I Nucleus

(E) Dorsal

I)

o o

Curken

0

Oskar protein (F) �-'-

An terior

FIGURE 6.7 The anterior- posteri o r axis is specified during ooge­ nesi s. (A) The oocyte moves into the posterior region of the egg chamber, while nurse cells fill the anterior portion. The oocyte nucleus moves toward the terminal follicle cells and synthesizes Gurken protein (green). The terminal follicle cells express Torpe­ do, the receptor for Gurken. (8) When Gurken binds to Torpedo, the terminal fol licle cells differentiate into posterior follicle cells and synthesize a molecule that activates protein kinase A in the egg. Protein kinase A orients the microtubules such that the grow­ ing end is at the posterior. (C) The Par-l protein (green) localizes to the cortical cytoplasm of nurse cells and to the posterior pole of the oocyte. (The Stauffen protein marking the posterior pole is labeled rcd; the red and green signals combine to Ouoresce yel­ low.) (D) The bicoid message binds to dynein, a motor protein associated with the non -growing end of microtubules. Dynein moves the bicoid message to the anterior end of the egg. The oskar message becomes complexed to kinesin I, a motor protein that moves it toward the growing end of the microtubules at the posterior region, where Oskar can bind the nanos message. (E) The nucleus (with its G u rken protein) migrates along the micro­ tubules, inducing the adjacent follicle cells to become the dorsal follicles. (F) Photomicrograph of bicoid mRNA (s tai ned black) passing from the nurse cells and localizing to the anterior end of the oocyte during oogenesis. Ie co u rtesy of H. Docrflinger; F from 5tephanscn et 01. 1 988, courtesy of the authors.) ure 6.7B). TI,e posterior follicle cells send a signal back into O,e oocyte. The identity of this signal is not yet known, but it recruits the par-l protein to t he po s ter i or edge of the

Posterior

oocyte cytoplasm (Figure 6.7C; Doerflinger et a1. 2006). Par­ I protein org aniz es microtubules specifically with their

T H E G E N ETICS OF AXIS SPECI FICATION I N DROSOPHILA

(8)

211

(C)

D)

nllilus (cap) and plus (growing) ends at the anterior and posterior ends of the oocyte, respectively (Gonzalez-Reyes

1995; Roth et a1. 1995; Januschke et a1. 2006). The orientation of the m.icrotubules is critical, because different microtubule motor proteins will transport their

et a1.

FIGURE 6.8 Expression of the gurken message and protein between the oocyte nucleus and the dorsal anterior cell mem­ brane. (A) The gurken mRNA is l ocal i zed between the oocyte nucleus and the dorsal follicle cells of the ovary. Anterior is to the left; do rsal faces upward. (8) The Gurken protein is si m ila rly locat­ ed (sh ow n here in a younger stage oocyte than A). (C) Cross sec­ tion of the egg through the reg ion of Gurken protein expressi on. (D) A more mature oocyte, showing Gurken protein (yellow) across the dorsal region. The actin is stained red, showi ng cell boundaries. As the oocyte grows, follicle cells migrate across the top of the oocyte, becomi n g exposed to Gurken. (A from Ray and Schupbach 1 996, courtesy of T. Schupbach; B,C from Peri et al.

1 999, courtesy of S. Roth; D courtesy of C. van Bu ski rk and T. Schupbach.)

A or protein cargoes in different directions. The motor

protein kinesin, for instance, is an ATPase that will use the energy of ATP to transport material to the plus end of the motor protein that will transport its cargo the opposite way.

ing event takes place. Here the gurken message becomes localized in a crescent between the oocyte nucleu::i and the

One of the messages transported by

oocyte cell membrane, and its protein product forms an

microtubule. Dynein, however, is a "minus-directed"

kinesin along the microtubules to the posterior end o f the oocyte is oskar A (Zimyanin et a1. 2008). The oskar mRNA is not able :.

be translated until it reaches the posterior cortex, at

which time it generates the Oskar protein. Oskar protein �ruits more par-l protein, thereby stabilizing the micro­ :ubule orientation and allowing more material to b e rECruited to the posterior pole o f

the oocyte (Doerflinger 2006; Zimyanin et a1. 2007). The posterior pole will :hereby have its own distinctive cytoplasm, called pol e

et .1.

?lasm, which contains the determinants for producing the �bdomen and the germ cells. This cytoskeletal rearrangement in the oocyte is accom­ :-anied by an increase in oocyte volllIT!e, owing to transfer i cytoplasmic components from

. icoid and

romponents

the nurse cells. These

include maternal messengers such as the

Hllnos

mRNAs. These mRNAs are carried by

=totor proteins along the micro tubules to the anterior and ""sterior ends of the oocyte, respectively ( Fi gure

6.7D-F) .

.\s we shall soon see, the protein products encoded by

-oid and 1IatJos are critical for establishing the anterior­

:'OSterior polarity of the eU1bryo.

( Fi gure 6.8; Neuman-Silberberg and Schupbach 1993). Since it can diffuse only a short distance, Gurken protein reaches only those follicle cells closest to the oocyte nucleus, and it signals those cells to become the more columnar dorsal follicle cells (Montell et al. 1 991; SchUp­ bach et a1. 1991; see Figure 6.7E). This establishes the dor­ anterior-posterior gradient along the dorsal surface of the

oocyte

sal-ventral polarity in the fullicle cell layer that surrounds

the growing oocyte.

Maternal deficiencies of either the gurkell or the torpedo

gene cause ventralization of the embryo. However, gurken

is active only in the oocyte, whereas torpedo is active only in the somatic follicle ce lls.

This fact was revealed by experi­

ments with germline/somatic chimeras. Tn one such exper­

iment, SchUpbach

(1987)

transplanted genm cell precursors

from wild-type embryos into embryos whose mothers car­

ried the torpedo mutahon. Cunversely, she transplanted the germ cell precursors from

embryos (Figure

torpedo mutants into wild-type

6.9). The wild-type eggs produced mutant,

ventralized embryos when they developed in a

torpedo

mutant mother's egg chamber. The torpedo mlltant eggs were able to produce normal embryos if they developed

•

:Jorsal-ventral patterning in the oocyte oocyte volume increases, the oocyte nucleus moves to m anterior dorsal position where a second major signal-

in a wild-type ovary. Thus, unlike Gurken, the Torpedo protein is needed in the follicle celis, not in the egg itself. TI,e GlUken-Torpedo signal that specifies dorsalized fol­ licle cells initiates a cascade of gene

activities that create

212

CHAPTER 6

n /-'" ;

Embryo from wild-type ! ! mother

torpedo.deficient germ cells in a wild-type female

\ \

,

,

\

)

n·- -.' V

Embryo from / mother i deficient in :. torpedo gene

(

\.

\

-+

torpedo-deficient oocyte in wild-type follicle

pOle cells ,--(germ ceU precursors)

.... .

----1.� ,

Wild-type germ ceJls ill a torpedo­ deficient female

\

- :;;­ E'--

i

V

-' .. . �

�:;s���el"ral aXIS

No dorsal-ventral -----.... axis (entire embryo ventral)

Wild-type germ cells ill a torpedodeficient follicle

FIGURE 6.9 Germline chimeras made by inlerchanging pole cells (germ cell precu rsors) between wild·lype embryos and emb ryos from mothers homozygous for a mutation of the torpedo gene. These IransplanlS produced wi ld-Iype females whose eggs came from mutant mothers, and torpedo-deficient females that laid wild-type eggs, The torpedo·defi cient eggs produced normal embryos when they developed in the wil d-type ovary, whereas the wild-type eggs produced ventra l ized embryos when they developed in the mutant mother's ovary.

and Snake is found throughout the perivitelline space sur­ rounding the embryo. Indeed,

this protein is very similar clot­

to the mammalian protease inhibitors that limit blood

ting protease cascades to the area of injury. In this way, the proteolytic cleavage of Easter and Spatzle is strictly limit­ ed to the area around the most ventral embryonic cells. The cleaved Spii tzle protein is now able to bind to its

the dorsal-ventral axis of the embryo (Figure 6.1 0) . The acti­ the pipe gene. As a result, Pipe protein i s made only in the ventral follicle cells (Sen et a1. 1998; Amiri and Stein 2(02). In some as yet unknown way (probably involVing sulfa­ tion), Pipe activates the Nudel protein, whlch is secreted

vated Torpedo receptor protein inhibits the expression o f

to the cell membrane of the neighboring ventral embryon­ ic cells (see Zhang et a!.

2009). A few hours later, activated

Nudel initiates the activation of three serine proteases that are secreted into the perivitelline fluiq (see Figure

6. 10B; 1995). These proteases are the prod­ ucts of the gastmlatiol1 defective (gd), snake (sl1k), and easter (en) genes. Like most extracellular proteases, these mole­ euJes are secre ted in an inacti ve form and. are subsequent­ ly activated by peptide cleavage. In a complex cascade of

Hong and Hashlmoto

events, activated Nudel activates the Gastrulation-defec­

tive protease, The Cd protease cleaves the Snake protein, activating the Snake protease, which in turn cleaves the Easter protein. This cleavage activates the Easter protease, whlch then cleaves the Spatzle protein (Chasan et a!.

1992; 1995; LeMosy et a!. 2001). It is obviously important that the cleavage of these three proteases be limited to the most ventral portion of the embryo. This is accomplished by the secretion of a protease Hong and Hashlmoto

inhibitor from the follicle cells of the ovary (Hashimoto et a!.

2003;

Ligoxygakis

et a!. 2003). Thls inhibitor of Easter

receptor in the oocyte cell membrane, the product of the toll gene . Toll protein is a maternal product that is evenly FIGURE 6.1 0 Generaling dorsal-ventral polarity in Drosophila. (A) The nucleus of Ihe oocyte travels 10 whal will become Ihe dor·

mRNA that becomes localized bet .... veen the oocyte nucleus and the cel l membrane, where it is translated into Gurken protei n . The sal side of I he embryo. The gurken genes of Ihe oocyte synlhesize

Gurken signal is received by the Torpedo receptor protein made by the follicle cells (see Figure 6.7). Given the short diffusibi l i ty of the signal, only the follicle cells cl osest 10 the oocyte nucleus (Le., the dorsal fol licle cells) receive Ihe Gu rken signal, which causes the fall icle cells 10 take on a characleristic dorsal foil icle morphol­ ogy and inhibits the synthesis of Pipe prolein. Therefore, Pipe pro­ tein is made only by Ihe ventral follicle cel l s. (8) The ventral region at a slightly later stage of development. Pipe modifies an unknown protei n (x) and allows it to be secreted from the ventral follicle cells. Nudel protein interacts with this modified factor to split the product of the gastrulation defective gene, which then splits the product of the snake gene to create an active enzyme that will split the inactive Easter zymogen into an active Easter protease. The Easter protease splits the Spatzle protei n into a form Ihal can bind 10 Ihe Tol l receplor (which is found Ihroughout Ihe embryonic cell membrane). This protease activity of Easter is striclly l i m ited by the protease inhib itor found in the perivitelline space. Thus. only the ventral cells receive the Tol l signal. This sig· nal separates the Cactus protein from the Dorsal protein, al lowing Dorsal to be translocated into the nuclei and ventralize the cells. (After van Eeden and st. Johnslon 1 999.)

T H E GENETICS OF AXIS SPECIFICATION I N DROSOPHILA

distributed throughout the cell membrane o f the egg ashlmoto et a1. 1988, 1991), but it becomes activated only by binding the Spatzle protein, which is produced only on the ventral side of the egg. Therefore, the Toll receptors on the ventral side of the egg are transducing a signal into the egg, while the Toll receptors on the dorsal side of the egg are not. This localized activation establishes the dorsal-ven­ tral polarity of the oocyte.

Generating the Dorsal-Ventral Pattern in the Embryo Dorsal, the ventral morphogen The protein that distinguishes dorsum (back) from ven­ trum (belly) in the fly embryo is the product 01 the dorsal gene. The mRNA transcript of the mother's dorsal gene is

(A)

(E)

\\ (porsa!) j} � �� c:/

Nucleus

cells

Oocyte

'\ Caclus ',� � .�eIastoderm margin is dependent on the epiboly of the YSL

cially in the dorsal region; see Cannany-Rampey and Schi­

can be demonstrated by severing the attacJunents between

er 2001) and some elements of involution (especially in the

278

CHAPTER 7

Endodermal behavior simulation Animal pole

Ventral

(�,J

::_::�.::�':':"�';'.,:�::::.:::.:,::.�:.:.::-:::��. Dorsal

Vegetal pole Onset of gastrulation

I

(/:::,: :� Mesoderm: AP-directed movement AP

, V : .

.

::..�:\

... . . .;:. .. .. . ... .. ..: . . ' ,, : \ . " .. . . . . " . . . .

'

VP Mid-gastrulation

:, D V

\

Endoderm: Random walk

.

AP

. " " ,. " ", '. . " " . : .'. . . :'. ' . . \ : :. : : - . : . . '" . ' . " 'i D . •

•

.

'

.

'

.

•

•

• t

VP Mid-gastrulation

FIGURE 7.43 Cel l migration of endodermal and mesodermal precursors. At the beginning of gastrulation, endoderma! and mesodermal cells are localized in the hypoblast. Once the hypoblast is internalized, the mesodermal cells migrate vegetally and then toward the animal pole. In contrast, endodermal cells spread th rough the internal regions of the embryo by a "random walk" over the yolk. (After Pezeron et af. 2008.)

future ventral regions; see Figure 7A2C). Thus, as the cells of the blastoderm undergo epiboly arpund the yolk, they are also internalizing cells at the blastoderm margin to form the hypoblast. As the hypoblast cells internalize, the future mesoderm cells (the majority of the hypoblast cells) initially migrate vegetally while proliferating to make new mesoderm cells. Later, they alter their direction and pro­ cede toward the animal pole. The endodermal precursors, however, appear to move randomly, walking over the yolk ( Figure 7-43; Pezeron et aL 2008). Once the hypoblast has formed, cells of the epiblast and hypoblast intercalate on the future dorsal side of the embryo to form a localized thickening, the embryonic shield ( Figu re 7-44A). Here, the cells converge and extend anteriorly, eventually narrowing along the dorsal midline ( Fig u re 7-44B). This convergent extension in the hypoblast forms the chordamesoderm, the precursor of the notochord (Trinkaus 1992; Figure 7A4C). As we will see, the embry-

onic shield is functionally equivalent to the dorsal blasto­ pore lip of amphibians, since it can organize a secondary embryonic axis when transplanted to a host embryo (Oppenheimer 1936; Ho 1992). The cells adjacent to the chordamesoderm-the paraxial mesoderm cells-are the precursors of the mesodermal somites. Concomitant con­ vergence and extension in the epiblast bring presumptive neural cells from all over the epiblast into the dorsal mid­ line, where they form the neural keel. The neural keel is usually quite a broad band of neural precursors, and it extends over the axial and paraxial mesoderm. Those cells remaining in the epiblast become the epidermis. On the ventral side (see Figure 7.44B), the hypoblast ring moves toward the vegetal pole, migrating directly beneath the epiblast that is epibolizing itself over the yolk celL Eventually the ring closes at the vegetal pole, completing the internalization of those cells that are going to become the mesoderm and endoderm (Keller et aL 2008). The zebrafish fate map, then, is not much different from tha t of the frog. If you can visualize opening a Xenopus blastula at the vegetal pole and then stretching that open­ ing into a marginal ring, the resulting fate map closely resembles that of the zebra fish embryo at the stage when half of the yolk has been covered by the blastoderm (see Figure 1.11; Langeland and KirnmeI 1997). Meanwhile, the endoderm arises from the most marginal blastomeres of the late blastula-stage embryo.

See WEBSITE 7.9 GFP zebrafish movies and photographs

Axis Formation in Zebrafish By different mechanisms, the Xenopus egg and zebrafish egg have reached the same basic state. They have become multicellular, they have undergone gastrulation, and they have positioned their germ layers such that the ectoderm is on the ou tside, the endoderm is on the inside, and the mesoderm is between them. As might be expected, zebrafish also form their axes in ways very similar to those of Xenopus, and using very similar molecules.

Dorsal-ventral axis formation The embryoniC shield of fish is, as mentioned above, homologous to the dorsal blastopore lip of amphibians, and it is critical in establishing the dorsal-ventral axis. Shield tissue can convert lateral and ventral mesoderm (blood and connective tissue precursors) into dorsal meso­ derm (notochord and s�mites), and it can cause the ecto­ derm to become neural rather than epidermaL This trans­ formative capacity was shown by transplantation experiments in which the embryonic shield of an early-gas­ trula embryo was transplanted to the ventral side of anoth­ er ( Figure 7-45; Oppenheimer 1936; Koshida et aL 1998). Two axes formed, sharing a cornmon yolk celL Although the prechordal plate and notochord were derived from the

AMPH I B IANS AND FISH (A)

Animal pole

279

Dorsal view

(D)

.JJr-:'Cfo--.hl�¥.I'!
Figure 8.19 Possible pathway initiating the distinction betvveen inner eel! mass and trophectoderm blastomeres. IAI The Tead4 transcription factor, when active, promotes transcription of the Cdxl gene. Together, the Tead4 and Cdx2 transcrip­ tion factors activate the that specify the outer cells to become trophectoderm. (B) Model for Tead4 activation. In the outer cells, the lack of ceUs surrounding the embryo sends a signa! (as yet unknown) that blocks the Hippo pathway from adivating the Lats protein. In the absence of functional Lats, the Yap tran­ scripti onal co-factor can bind with Tead4 to activate the Cdx2 gene. In the i nner cells, the Hippo path a is active and the Lats kinase phosphorylates the Yap transcriptional co-activator. The phospho­ rylated form of Yap is targeted for degra­ dation and does not enter the nucleus. (Mer Nishioka et al. 2009.1

genes

wy

.

sequent accumulation of Na+ draws in water osmotically,

Na' /H+

oviduct cells on which the embryo is traveling toward the

:DCoel. The membranes of trophoblast cells contain sodi­ pumps

.

name derives from the mythical Celtic land of perpetual youth.

Initially, the morula does not have an internal ca vit y Ho e ver, du ing a process called cavitation, the tro­ :"hoblast cells secrete fluid into the morula to create a blas­

=

Nanog* prevents the ICM blastomeres from becoming hypoblast cells, and stimu lates blastomere self-renewal i n the ep ib last The activated (phosp hory­ lated) form of Stat3 also stimulates self­ renewal of leM blastomeres (see Fig­ ure 8. 1 8; Pesce et al. 200"1; Chambers et al. 2003;· Mitsui et a l . 2003). If the inner cell mass blastomeres are removed in a manner that lets them retain t hei r express io n of N a n og, Oct4, and phosphorylated Stat3 proteins, these cells divide and become embry­ onic stem cells (ES cells). The plu ri po­ tency of ES cells (and the medical uses for them, which will be detailed later in t h is book) is dependent on their retaining the e p ress ion of these three transcription fa cto rs.

changer) that pump Na' into the central ca ity The sub-

1977; et al. 2004). Interestingly, this sodium pumping activi ty appears to be stimulated by the thus creating and enlarging the blastocoel (Borland Ekkert et aI. 2004; Kawagishi

uterus (Xu et a1. 2004).

304

CHAPTER 8

(A)

(8)

(C) 'o j

of uterus

FIGURE 8.20 H atch i ng from the zona and implantation of the mammalian blastocyst in the uterus. (A) Mouse b lastocyst hatch­ ing from the zona pellucida. (B) Mouse blastocysts entering the uterus. (C) Initial implantation of a rhesus monkey blastocyst. (A from Mark et al. 1 985, courtesy of E. Lacy; B from Rugh 1 967; CrCarnegie Institution of Washington, Chester Reather, photographer.)

Wassarman

1986; O'Sullivan et al . 2001). Once outside the

zona, the blastocyst can make direct contact with the uterus (Figure 8.20B,C). The endometrium-the epithelial lining of the uterus-"catehes" the blas tocys t on matrix made up of

c omp lex

an extracellular

sugars, collagen; l.aminin,

fibronectin, cadherins, hyaluron.ic acid, and heparan sul­

While the embryo is moving through the oviduct en route

fate receptors (see Wang and Dey 2006). As in so many intercellular adhesions during development, there appears to be a p er iod of labile at tachmen t, followed by a period of more stable attachment. The first attachment seems to be mediated by L-selectin on the trophoblast cells adher­ ing to sulfa ted polysaccharides on the uter ine cells (Gen­ bacev et al. 2003). These sulfated polysaccharides are syn­ thesized in response to estrogen and progesterone secreted by the corpus luteum (the remnant of the ruptured ovari­

to the uterus, the blastocyst expands within the zona pellu­

an follicle).

As the blastocoel expands, the inner cell mass becomes positioned on one side

of the ring of trophoblast cells (see

Figure 8.17F); the resulting type of blastula, called a blasto­ cyst, is another hallmark of mammalian cleavage >

Escape from the zona pel/ucida

cida (the extracellular matrix of the egg that was essential for sperm binding duting fertilization; see Chapter 4). Dur­ ing this time, the zona

pelluci d a prevents

the blastocyst

from adhering to the oviduct walls. (If this happens-as it

is called an ectopic, or "tubal," pregnancy, a dangerous condition because an embryo implanted in the oviduct can cause a life-threat­ ening hemo rrha ge when it begins to grow.) When the embryo reaches the uterus, it must "hatch" from the zona so that it can adhere to the uterine wall. The mouse blastocyst hatches from the zona pellucida by digesting a small hole in it and squeezing through the hole as the blastocyst expands ( F igu re 8.20A). A trypsin­ like protease secreted by the trophoblast seems responsi­ ble for hatching the blastocyst from the ZOna (Perona and

sometimes does

in

humans-it

*The interplay of my th and biology certainJy comes to the fore when describing mammalian development. Although the mam­ malian blastocyst was discovered by Rauber in 1881, its first public display WitS probably in Gustav Klimt's 1908 painting Danae, in which blastocyst-like patterns are featured on the heroine's robe as she becomes impregnated by Zeus.

After the initial binding, sever,,1 other adhesion systems appear to coor dinate tightly bound to the

their efforts to keep the blastocyst uterine lining. The trophoblast cells

synthesize integrins that bind to the uterine colJagen, fibroneetin, and laminin, and

they syntheSize heparan sul­

fate proteoglycan precisely prior to implantation (see Car­ son et al.

1993).

P-cadherins on the trophoblast and uter­

ine endometrium also help dock the embryo to the uterus (see Chapter

3).

Once

in

contact with the endometrium,

Wnt proteins (from the trophoblast, the endometrium, or

a set of pro­ strome lysin, and plasmino­ These p rotein-digesting enzymes digest the

from both) instruct the trophoblast to secrete

teases, including collagen ase, gen activator.

extracellular matrix of the uterine tissue, enabling the blas­

tocyst to bury itself within the uterine wall (Strickland et aI.

1976; Brenner et a1. 1 989; Pollheimer et a1. 2(06).

Mammalian Gastrulation Birds and mammals are both descendants of reptilian species (albeit different reptilian species). It is not surpris­ ing, therefore, that mammalian development parallels that

B I RDS A N D MAMMALS

of reptiles and birds. What is surprising is that the gastru­ lation movements of reptilian and avian embryos, which evolved as an adaptation to yolky eggs, are retained in the mammalian embryo even in the absence of large amounts of yolk. The mammalian inner cell mass can be envisioned as sitting atop an imaginary ball of yolk, following instruc­ tions that seem more appropriate to its reptilian ancestors. Modifications for development inside another organism The mammalian embryo obtains nutrients directly from its mother and does not rely on stored yolk. This adaptation has entailed a dramatic restructuring of the maternal anato­ my (such as expansion of the oviduct to form the uterus) as well as the development of a fetal organ capable of absorbing maternal nutrients. This fetal organ-the chori­ on-is derived primarily from embryonic trophoblast cells, supplemented with mesodermal cells derived from the inner cell mass. The chorion forms the fetal portion of the placenta. It also induces the uterine cells to form the mater­ nal portion of the placenta, the decidua. The decidua becomes rich in the blood vessels that will provide oxygen and nutrients to the embryo. The origins of early manunalian tissues are summarized in Figure 8.21 . The first segregation of cells within tl1e inner cell mass forms two layers. The lower layer is the hypoblast

(sometimes called the primitive endoderm or visceral endo­ demz); the remaining inner cell mass tissue above it is the epiblast (Figure 8.22A,B). Surprisingly, whether a cell

becomes epiblast or hypoblast does not depend on the pOSition of the cell within the [CM. Rather, the blastomeres of the [CM appear to be a mosaic of future epiblast cells (expressing Nanog transcription factor) and hypoblast cells (expressing Gata6 transcription factor) a full day before the layers segregate .at day 4.5 (Chazaud et al. 2006). The epi­ blast and hypoblast form a structure called the bilaminar germ disc. The hypoblast cells delaminate from the inner cell mass to line the blastocoel cavity, where they give rise to the extraembryonic endoderm, which forms the yolk sac. As in avian embryos, these cells do not produce any part of the newborn organism. The epiblast cell iayef is split by small clefts that eventually coalesce to separate the embryonic epiblast from the other epiblast cells that line the amnionic cavity (Figure 8.22C,0). Once the lining of the amnion is completed, the amnionic cavity fills with a secretion called amnionic fluid, which serves as a shock absorber for the developing embryo while preventing it from drying out. 111e embryonic epiblast is thought to con­ tain all the cells that will generate the actual embryo, and it is similar in many ways to the avian epiblast. By labeling individual cells of the epiblast with horse­ radish peroxidase, Kirstie Lawson and her colleagues (1991) were able to construct a detailed fate map of the

i TIEMSBURYEOS NIC I E m b r IT epi laysotnic r L streak rE l . I Amnionic mas Blastocyst L Hypublast Extraembryonic Yolksac L, Cytotrophoblast Syncytiotrophoblast k

305

FIGURE 8.21 Schem ati c diagram showing Ihe der vati on of tis­ sues in human and rhesus m on ey embryos. The dashed line indi­ cates a possible dual origi n of the exlraembryonic esoderm . (Ailer Luckett 1 978; Bianchi el aJ. 1 993.1

m

'-==='---J

p

ib ast

Primitive

ectoderm

-----+-

.

._."

.... . �

endoderm

�

----+-

----

-'----.-.

xr e by c TIEXSTRUAESMBRYONIC E t

a m r o

mesoderm

ni

306

CHAPTER 8

(A) Uter ine epitheli um _

(B)

Epiblast --i4M

Blastocyst, 7 days

Hypoblast

Uterine t issue (decidua)

(C) 8 Days

M"."nol capillary

/

Amnionic cavity

FIGURE B.22 Tissue formation in Ihe early mam­ malian embryo. (A) Mouse embryo at day 3.5, showing the ",ndom expression of Nanog (bloe, for the epiblast) and Gata6 (red, for the hypoblast) in the inner cell mass. In another 24 hours, the cells will sort out: the hypoblast cells will abut the blastocoel, and the epi­ blast cells will be between the hypoblast cells and the trophoblast (as in Figure 8:18B). (B-E) Human embryos between days 7 a nd 1 t . (B,q Human blastocyst imme­ diately prior to gastrulation. The inner cell mass delam­ inates hypoblast cells that line the blastocoel, forming the extraembryonic endoderm of the primitive yolk sac and a two-layered (epiblast and hypoblast) blastodisc. The trophoblast divides into the cytotrophoblast, which will form the villi, and the syncytiotrophoblas, which will ingress into the uterine tissue. (0) Meanwhil e the epiblast splits into the amnionk ectoderm (which encircles the amnionic cavity) and the embryonic epi­ blast. The adult mammal forms from the cells of the embryonic epiblast. (E) The extraembryonic endoderm forms the yolk sac. The actual size of the embryo at this stage is about that of the period at the end of this sen­ tence. (A courtesy of ). Rossant.)

,

--z:r.:---':::"�f-AR

'tGT----€=i-_ Syncyt.ioh'Ophoblast Cytotrophoblast

proliferating into uterine tissue

(D) 9 Days

Amnionic ec toderm

Amnionic --l-� cavity Emhryonic epiblast

!\=O�- Cytotrophoblast Syncytio­ trophoblast

embryonic endoderm

(E) 10-1 1 Days

extraembryonic mesoderm

BIRDS AND MAMMALS

(A)

(e)

Extraembryonic Syncytiotrophoblast

1 4 · 1 5 Days

307

Primitive

germ disc Primitive groove Yolk sac

Hypoblast

FIGURE 8.23 Amnion structure and cell movements during human gastrulation. (A,B) Human embryo and uterine connec­ tions at day 1 5 of gestation (A) Sagittal section through the mid­ line. (B) View looking down on the dorsal surface of the embryo. (e) Movements of the epiblast cells through the primitive streak and Hensen's node and underneath the epiblast are superimposed on the dorsal su rface view. At days 1 4 and 1 5, th e ingressing epi­ blast cells are thought to replace the hypoblast cells (which con­ tribute to the yolk sac lining), while at day 1 6, the ingressing cel/s fan out to form the m es odermal l ayer. (After Larsen 1 993.) .

mouse epiblast (see Figure 1.11). Gastrulation begins at the posterior end of the embryo, and this is where the cells of the node' arise (Figure 8.23). Like the chick epiblast cells, the mammalian mesoderm and endoderm migrate through a primitive streak; also like their avian counterparts, the migrating cells of the mammalian epiblast lose E-cadherin, detach from their neighbors, and migrate through the streak as individual cells (Burdsal et aJ. 1993). Those cells arising from the node give rise to the notochord. However, in contrast to notochord formation in the duck, the cells that form the mouse notochord are thought to become inte­ grated into the endoderm of the primitive gut Gurand 1974; Sulik et a l 1994). These cells can be seen as a band of small, ciliated cells extending rostraUy from the node. They form the notochord by converging medially and "budding" off in a dorsal direction from the roof of the gut. Cell migration and specification appear to be coordinat­ ed by fibroblast growth factors. The cells of the primitive streak appear to be capable of both synthesizing and responding to FGFs (Sun et al. 1999; Oruna and Rossant 2001). In embryos that are homozygous for the loss of the fgf8 gene, cells fail to emigrate from the primitive streak, and neither mesoderm nor endoderm are formed. PgfS (and per-

Extraembryonic mesoderm

Mesoderm

haps other FGFs) probably control cell movement into the primitive streak by dowmeguJating the E-cadherin that holds the epiblast cells together. Pgf8 may also control cell specifi­ cation by regulating snail, Brachyury (1), and Tbx6, three genes that are essential (as they are in the chick embryo) for mesodermal migration, specification, and patterning. The ectodermal precursors are located anterior and lat­ eral to the fully extended primitive streak, as in the chick epiblast; however, in some instances (also as in the chick embryo), a single cell gives rise to descendants in more than one germ layer, or to both embryonic and extraem­ bryonic derivatives. Thus, at the epiblast stage these lineag­ es have not become fully separate from one another. As in avian embryos, the cells migrating into the space between the hypoblast and epib,last la yers become coated with hyaluronic acid, which they synthesize as they leave the primitive streak. This substance keeps them separate while they migrate (Solursh and Morriss 1977). It is thought that the replacement of human hypoblast cells by endoderm precursors occurs on days 14-15 of gestation, while the migration of cells forming the mesoderm does not start until day 16 (see Figure 8.23C; Larsen 1993).

.

"'in mClrnrnalian development, Hensen's node is usually just called

J'the node/, despite the fact that Hensen discovered this structure in rabbit and guinea pig embryos.

Formation of the extraembryonic membranes While the embryonic epiblast is undergoing cell move­ ments reminiscent of those seen in reptilian or avian gas­ trulation, the extraembryonic cells are making the placen­ ta, a distinctly mammalian set of tissues that enable the fetus to survive within the maternal uterus. Although the initial trophoblast cells of mice and humans divide like most other cells of the body, they give rise to a population of celis in which nuclear division occurs in the absence of cytokinesis. The original trophoblast cells constitute a layer caTIed the cytotrophoblast, whereas the multinucleated cell type fonn, the syncytiotrophoblast. The cytotro­ phoblast initially adheres to the endometrium through a series of adhesion molecules, as we saw above. Moreover,

308

(A)

CHAPTER 8

(B) :�n-*,7 umDlllcal arteries (from fetus) vein (to fetus)

Chorionic villus

--�ilii��,: r��

Chorion portion of placenta

(fetal

Trophoblast cells

Maternal portion

o o h r e� : t�: :�:� m rr T IT From FIGURE 8.24 (A) Human embryo and placenta aher 50 days of gestation. The embryo lies within the amnion, and its blood ves­ sels can be secn extending into the chorionic villi. The small sphere to the right of the embryo is the yolk sac. (B) Relationship of the chorionic vi l li to the maternal blood supply in the primate uterus. In the umbilicus, there are two arteries and a single vein. (A from Carnegie Institution of Washington, courtesy of Chester F. Reather.)

of placenta (decidual ceUs)

aternal vein j�;;;;::;;;;��;::;;;:= MMaternal artery

cytotrophoblasts contain proteolytic enzymes tha t enable them to enter the uterine wall and remodel the uterine blood vessels so that the maternal blood bathes fetal blood vessels. The syncytiotrophoblast tissue is thought to fur­ ther the progression of the embryo into the uterine wall by digesting uterine tissue. The cytotrophoblast secretes paracrine factors that attract maternal blood vessels and gradually displace their vascular tissue such that the ves­ sels become lined with trophoblast cells (Fisher et al. 1989; Hemberger et al. 2003). Shortly thereafter, mesodermal tis­ sue extends outward from the gastrul ating embryo (see Figure B.22E). Studies of human and rhesus monkey embryos have suggested that the yolk sac (and hence the hypoblast) as we ll as primitive streak-derived cells con­ tribute this extraembryonic mesoderm (Bianchi et al. 1993). The extraembryonic mesoderm joins the trophoblastic extensions and gives rise to the blood vessels that carry nutrients from the mother to the embryo. The narrow con­ necting stalk of extraembryonic mesoderm that links the embryo to the trophoblast eventually forms the vessels of the umbilical cord. The fully developed extraembryonic organ, consisting of trophoblast tissue and blood vessekon­ taining mesoderm, is the chorion, and it fuses with the uter­ ine wall decidua to create the placenta. Thus, the placenta has both a maternal portion (the uterine endometrium, or decidua, which is modified during pregnancy) and a fetal

component (the chorion) . The chorion

may be very closely to maternal tissues while still being readily sepa­ rable from them (as in the contact placenta of the pig), or it may be so intimately integrated with maternal tissues that the two cannot be separated without dam a ge to both the mother and the developing fetus (as in the deciduous pla­ centa of most mammals, including hwnans).· Figure 8.24A shows the relationships between the em bryonic and extraembr-yonic tissues of a 6.S-week hwnan embryo. The embryo is seen encased in the amnion and is fur ther shielded by the chorion. The blood vessels extending to and from the chorion are readily observable, as a re the villi that project from the outer surface of the chorion. These villi contain the blood vessels and allow the chorion to have a large area exposed to the maternal blood. Although fetal and mater nal circulatory systems normal­ ly never merge, diffusion of soluble s ubsta nces can occur through the villi (Figure 8.24B). In this manner, the mother provides the fetus with nutrients and oxygen, and the fetus sends its waste products (mainly carbon dioxide and urea) into the maternal circulation. The maternal and fetal blood cells usually do not mix, although a small nwnber of fetal red blood cells are seen in the maternal blood circulation. apposed

See WEBSITE 8.4 Placental functions *The.re are numerous types of placentas, and the extraembryonic membranes form differently in different orders of marrunals (see Cruz and Pooersen 1991). Although mice and humans gastrulate and implanl in a similar fashion, their extraerubryonic structures are distinctive. It is very risky to extrapolate developmental phe­ nomena from one group of mammals to another. Even Leonardo da Vinci gol caught (Renfree 1982). His remarkable drawing of the human felus inside the placenta is stunning art, but poor science: the placenta is that of a cow.

a I

r

BIRDS AND MAMMALS

309

SIDEL IGHTS SPECULATIONS

Twins and Chimeras

T

he early cells of the mammalian embryo can replace each other and compensate for a missing cell. This regulative ability was first demonstrated in 1 952, when Seidel destroyed one cell of a 2-cell rabbit embryo and the remaining cell pro­ duced an entire embryo. Once the inner cell mass oeM) has become sep­ arate from the trophoblast, the leM cells constitute an equivalence group. In other words, each leM cell has the same potency (in this ease, each cell can give rise to all the cell types of the embryo, but not to the trophoblast), and their fates will be detenmined by interactions among their descendants. Gardner and Rossant ( 1 9 76) also showed that if cells of the leM (but not trophoblast cells) are injected into blas-

tocysts, they contribute to the new embryo. Since the leM blastomeres can generate any cell type in the body, the cells of the blastocyst are referred to as pluripotent. The regulative capacity of the leM blastomeres is also seen in humans. Human tvvins are classified into two major groups: monozygotic (one�egg, or identical) twins and dizygotic (two­ egg, or fraternal) twins. Fraternal twins are the resu lt of two separate fertiliza­ tion events, whereas identical twins See WEBSITE 8.5 Nonidentical monozygotic twins

See WEBSITE 8.6 Conjoined twins

2-cell embryo Figure 8.25 Diagram showing the timing of human monozygotic twi n­ ning with relation to extra.. embryonic membranes. (C) (A) Splitting occurs before formation of the t roph o­ blast, so each twi n has its own chorion and amnion. (B) Splitting occurs after trophoblast formation but before amnion for· �ation, r�su l ti ng in twi ns having individual amnionic sacs bu t shari ng one chorion. (C) Splitting alter ammon formation leads to twi ns in one amnionic sac and a si n gl e chorion. (After Langman t981.)

are formed from a single embryo whose cells somehow become dissoci­ ated from one another, Identical tvvi ns may be produced by the separation of early blastomeres, or even by the sepa­ ration of the inner cell mass into two regions within the same blastocyst. Identical twins occur in roughly 0.25% of human births. About 33% of identical twins have two complete and separate chorions, indicating that separation occ,urred before the forma­ tion of the trophoblast tissue at day 5 (Figure 8.25A). The remaining identical twins share a common chorion, sug­ gesting that the split occurred within (Continued on next page)

2 Chorions

1 Chorion

310

CHAPTER 8

SIDELIGHTS & SPECULATIONS (Cont i n u ed) the inner cel! mass after the tro­ ph ob l a st formed. By day 9, the human emb ryo has co m p l eted the co ns tr uc ­ tion of an other extraemb ryonic l ayer the l ining of the amnion. This tissue, derived from the ec toclerm and meso­ ,

derm, forms the amnionic sac, which surrounds the embryo with amnionic fluid and protects it from desiccation and ab rupt movement. If th e separa­ tion of the embryo were to come after the formation of the chorion on day 5 b u t before the formation of the amnion on day 9, then the res u l tin g

Figure 8.26 Producti on of chimeric mice. (A) Experimental procedu res used to produce chimeric mice Ea rly 8-cell embryos of gen eti ca l l y disti n c t .

mice (here, with (Oelt color diifercnces) are isolated from mouse oviducts and brought together after their zonae are removed by proteo lytic enzymes. The cells form a com posi te blastocyst, which is implanted into the uterus of a

foster mother. The photograph shows one of the actual chimeric mice produced i n this manner. (B) An adult female chimeric mouse (bottom) produced from the fus ion of th ree 4-cell embryos: one from two white-furred parents, one from two black-furred par­ ents, and one from two brown-furred parents. The resulting mouse has coal colors from a l l three embryos. More­ over, each embryo contributed germline cells, as is ev idenced by the three colors of offspring (above) pro­ duced when this chimeric female was mated with recessive (white-furred) mal es. (A courtesy of B. M intz; B from Markert and Petters 1 978, counesy of C. Markert.)

embryos should have one chorion and two a mnions (Figure 8.2SB). This hap­ pens in about two-thirds of human identical tw i ns A sma l l percentage of identical twins a re born within a s in ­ gle chorion and amnion (F igu re 8.2SC). T h is means the division of the embryo came after day 9. The abil ity to prod uce an en t i re embryo from cells that no rma l l y would have con tri b uted to on l y a portion of the embryo--called regulation-was d i scussed in the I ntroduction to Part II. Regulation is a l so seen in t h e abi l ity of ,

� � . .� . @ ,

.

,,-.

',. t ,

,

GG

Blastomeres

-

Blastocyst

Blastocysts implanted

two or more ear ly embryos to fo rm one ch i meric individual rather than twin s, tri p l ets or a multiheaded indi­ vidual. Chimeric mice can be pro­ duced by artificially aggregatin g two or more ea rl y c l eavage ( us ua l ly 4- or 8cell) embryos to form a com pos i te emb ryo As shown in Figure 8.26A, the zonae pellucidae of two genetically d ifferent em bryos can be artificially removed and the embryos brought together to form a common b l astocys t. These blastocysts are th en i mp l anted into the uterus of a foster mother. When t hey are born, the chimeric off­ spri ng have so me cells from each em b ryo. Th i s is readi ly seen when the agg rega ted blastomeres come from mouse strains that differ in their coat colors. When blastomeres from wh i te and black strains are aggregated, the result is commonly a mouse with black and wh i te bands. Markert and Petters (1 978) have shown that three ea r ly 8cel l embryos can uni te to form a com­ mon com pacted morula and that the resulting m o u se can have the coat col­ ors of th e three di ffere nt strains. More­ over, they showed that each of the three embryos gave rise to precursors of the gametes. When a chimeric (blacklbrown/white) female mouse was mated to a wh i te-fu rred (recessive) male, offspring of each of the three colors were produced (Figure 8.26B). ,

-

.

(B)

BIRDS AND MAMMALS

311

SIDEL IGHTS & SPECULATIONS (Continued) There is even evidence that h u ma n emb ryos can fo rm chimeras (de la Cha ppel le et al. 1 974; Mayr et al. 1 979). So me individuals have two genetically different cell ty pes (XX and XV) within the same body, each with its own set of geneti cal ly defined char­ acteristics. The simplest expl a natio n for such a p henomenon is that these individuals resulted from the aggrega­ tion of two embryos, one male and one female, that were developing at the same time. If this explanatio n is correct, then two fratern a l twins have fused to create the i ndividual'" (see Yu et al. 2002). According to these studi es of twin formation and chimeric mice, each

blastomere of the i n ner cell mass should be able to produce any cell of the body. This hypothesis has been confirmed, and it has i m po rtant con­ sequences for the study of mam­ ma l ian development. When ICM cells are isolated and grown under certain conditions, they remain undifferenti­ ated and continue to divide in culture (Evans and Kaufman 1 98 1 ; Martin

1 98 1 ) . Such celfs are embryonic (E5) cell s. When ES celfs are i n jected into a mouse b lastocyst, they ca n integrate i n to the host inner cel! mass. The resulting embryo has cells from both the host and th e donor tis­ sue. This techn i que has become extremel y important in determi ning the fu n cti on of genes du ri ng mam­ malian development. stem

*There are other expla nations, at least for some of the chimeras. Souter and co l leagu es (2007) have shown that i n at least one XXlXY ch i mera, the maternal alleles were i den­ tical and the paternal alleles differed. This would be expected if the egg undervvent pa rthenogen ic activation and each of the meiotic cells were fertilized by a different sperm (one bearing an X chromosome, one beari ng a V). The intermingling of the cells would produce the chimera, which in this case was a true herm aphrod i te. We still do not know the mechanisms through whi ch human twins an d ch i meras form.

Mammalian Axis Formation The formation of the anterior-posterior axis has been stud­ ied most extensively in mice. The structure of the mouse epiblast, however, differs from that of humans in that it is cup-shaped rather than disc-shaped. The dorsal surface of the epiblast (the embryonic ectoderm) contacts the amnion­ ic cavity, while the ventral surface of the epiblast contacts the newly formed mesoderm. In this cuplike arrangement, the endoderm covers the surface of the embryo on the "outside" of the cup (Figu re 8.27A).

vate the expression of posterior genes that are required for mesoderm formation; the AVE creates an anterior region where Nodal cannot act. The AVE also begins expressing the anterior markers Otx2 and Wnt-inhibitor Dickkopf. Studies of mutant mice indicate that the AVE promotes anterior specification by suppressing the formation of the primitive streak, a posterior structure, by Nodal and Wnt proteins, as in the chick (Bertocchini et a1 2002; Perea­ Gomez et a1 2002). However, the AVE alone calIDot ind uce

The mammalian embryo appears to have two signaling centers: one in the node (eqUivalent. to Hensen's node and the trunk portion of the amphibian organizer) and one in th e anterior visceral endoderm (AVE; equivalent to the

neural tissue, as the node can (Tam and Steiner 1999). Formafion of the node is dependent on the trophoblast (Episkopou et a1. 2001). As in the chick embryo, the place­ ment of the node and the primitive streak appears to be due to the blocking of Nodal signaling by the Cerberus and Lefty-l from the AVE (Perea-Gomez et a1. 2002). Once formed, the node will secrete Chordin; the head process and notochord will later add Noggin. These two BMP antagonists are not expressed in the AVE. Dickkopf is

chick hypoblast and similar to the head portion of the amphibian organizer) ( Figure 8.27B; Beddington and Robertson 1999; Foley et a1 2000). The node (at the "bot­ tom of the cup" in the mouse) appears to be responsible for the creation of all of the body, and the two signaling centers work together to form the anterior region of the embryo (Bachiller et a1. 2000). The notochord forms by the orsal infolding of the smali, ciliated cells of the node. The AVE originates from the visceral endoderm 'hypoblast) that migrates forward. As this region migrates, i secretes two antagonists of the Nodal protein, Lefty-l and Cerberus (Brennan et al. 2001; Perea-Gomez et a1. 2001; Yamamoto et a1. 2(04). (Lefty-l binds to the Nodal's recep­ :ors and blocks Nodal binding, and Cerberus binds to _ ·odal itself.) While the Nodal proteins in the epiblast acti-

expressed in both the AVE and in the node, but only the Dickkopf from the node is critical for head development (Mukhopadhyay et a1. 2001). While knockouts of either the chordin or the noggin gene do not affect development, mice missing both genes lack a forebrain, nose, and other facial structures (Figure 8.28). It is probable that the AVE func­ tions in the epiblast to restrict Nodal activity; thereby coop­ erating with the node-produced mesendoderrn to promote the head-forming genes to be expressed in the anterior por­ tion of the epiblast. This regulation can be altered by retinoic acid (RA), which appears to inhibit the migration of the AVE precur­ sors. As a result, more than one body axis can form (Fig­ ure 8.27C,D; Liao and C ol lins 2008). This effect has been proposed as a pOSSible cause of conjoined twins.

The anterior-posterior axis: Two signaling centers

312

CHAPTER 8

(A)

P r e s u m p t i v e e c o d r m Prechordal Al anrois cParvimityit ve sNodet eak Presumpttivhee node) (8)

AVE

dd

En o e rm

Amnionic

notochord (beneath

(C)

Tbx6

FIGURE 8.27 Axis and notochord formation in the mouse. (A) In the 7-day mouse embryo, the dorsal surface of the e p iblast (embryonic ectoderm) is in contact with the amnionic cavity. The ve ntral surface of the epiblast cont.acts the newly formed meso­ derm. In this cuplike arrangement, the endoderm covers the sur­

face of the e mbryo. The node is at the bottom of the cup, and it has generated chordamesoderm. The two signaling cente rs, the node and the anterior visceral endoderm (AVE), are located on

(B) By

Normal RA

An terior-posterior patterning by FGF and retinoic acid gradients The head region of the mammalian embryo is devoid of Nodal signaling, and IlMPs, FGFs, and Wnts are also inhib­ ited. The posterior region is characterized by Nodal, BMPs, Wnts, FGFs, and retinoic acid. There appears to be a. gradi­ the pos­ i ent of Wnt , 13M!>, and FGF proteins that is highest n terior and that drops off strongly near the anterior region.

opposite sides of the cup. Eventually, the notochord will link them. The caudal side of the embryo is marked by the presence of the allantois. embryonic day 8, the AVE l ines the foregut, and the prechordal mesoderm is now in contact with the fore­ brain ectoderm. The node is now farther caudal, largely as a result of the rapid growth of the anterior portion of the embryo . The cells in the midline of the epiblast migrate through the primitive streak (white arrows). (CD) Retinoic acid appears to i nhibit the migra­ tio n or the AVE precursors. In normal mouse embryos (left col­ umn), Tbx6 is expressed in the anterior primitive streak and epi­ blast surrounding the node (C), while Sonic h ed gehog is exp ressed in the notochord (D). In embryos treated with retinoic acid during early gastrulation (right column), there are often two axes. Here, two areas ofTbx6 are observed, and the notochord is bifurcated (arrows). (A,S courtesy of K. Sulik; C,D from Liao and Collins 2008, courtesy of K. Sulik and G. Schoenwolf.)

Moreover, in the anterior half of the embryo, starting at

the node, there is a high concentration of antagonists that pre­ vent IlMPs and Wnts from acting (Figure 8.29A). The Fgf8 gra dient is create d by the decay of mRNA:fgi8 is expressed at the growing posterior tip of the embryo, but thefgf8 mes­ sage is slowly degraded in the newly formed tissues. Thus there is a gradient offgi8 mRNAacross the poste rio of the embryo, which is then converted into an Fgf8 protein gra­ client (Figure 8.29B; Dubrulle and Pourquie 2004).

r

BIRDS AND MAMMALS

313

FIGURE 8.28 Expression o f BMP antagonists in the mammalian node. (A) Expression of chordin during mouse gastrulation is seen in the anterior prim­ itive streak, node, and axial mesoderm. It is not expressed in the anterior viscer­ al endoderm. (B-O) Phenotypes of 1 2 .S-day embryos. (B) Wi Id-type embryo. (C) Embryo with the chordin gene knocked out has a defective ear but an otherwise normal head. (D) Phe­ notype of a mouse deficient in both chordin and noggin. There is no jaw, and there i s a single centrally located eye, over which protrudes a large pro­ boscis (nose). (From Bachiller et 31. 2000, courtesy of E. M. De Robertis.)

In a dd ition to FGFs, the late gastrula has a gradient of retinoic acid. RA levels are hi gh in the posterior regions and low in the ante rior portions of the embryo. This gra­

dient (like that of chick, frog, and fish embryos) a ppears to be controlled by the expression of RA-synthesizing

(B)

(A) BMP, Wnt

Anterior

(C)

Wnts, BMPs, FGFs

�����§§§��

1

Hox genes streak

Organizer

\ fesoderm

enzymes in the embryo's p os terior and RA-degrading enzymes in the anterior parts of the embryo (Sa kai et al. 2001; Oosterveen et al. 2004). The FGF gradient patterns the posterior portion of the embryo by working through the Cdx fanuly of caudal­ related genes (Figure 8.29C; Lohnes 2003). The Cdx genes, in turn, integrate the various posteriorization signals and activate particula r Hox genes.

mesendoderm

fiGURE 8.29 Anterior·posterior patterning in the mouse embryo. (A) Concentration gradients of BMPs, Wnts, and FGFs i n :he late gastrula mouse embryo (depicted a s a flattened disc). The j)fimitive streak and other posterior tissues are the sources ofWnt .d!1d BMP proteins, whereas the organizer and its derivatives (such Ihe notochord) produce antagonists. FgfB is expressed in the posterior tip of the gastrula and continues to be made in the tail oud. Its mRNA decays, creating a gradient across the posterior ""rtion of the embryo. (8) Fgf8 gradient in the tailbud region of 3

Anterior·posterior patterning

9-day mouse embryo. The highest amount of Fgf8 (red) is found near the tip. The gradient was determined by ;n situ hybridization of an FgfB probe and staining for increasing amounts of time. (C) Retinoic acid, Wnt3a, and Fgf8 each contribute to posterior pat· terning, but they are integrated by the Cdx family of proteins that regulates the activity of the Hox genes. (A after Robb and Tam 2004; B from Dubrulle and Pourquie 2004, courtesy of O. Pourquic; C after Lohnes 2003.)

314

CHAPTER 8

are expressed posteriorly and later. As in Drosophila, a sep­ arate set of genes in mice encodes the transcription factors that regulate head formation. In Drosophila, these are the orthodenticle and empty spiracles genes. In mice, the midbrain and forebrain are made through the expreSSion of genes homologous to these-Otx2 and Emx (see Kurokawa et al. 2004; Simeone 2(04). The mammalian Hox/HOX genes are numbered from 1 to 13, starting from that end of each complex that is expressed most anteriorly. Figure 8.30 shows the relation­ ships between the Drosophila and mouse homeotic gene sets. The equivalent genes in each mouse complex (such as Haxal, Hoxbl, and Hoxdl) are called paralogues. It is thought that the four mammalian Hox complexes were formed by cllfomosome duplications. Because the corre­ spondence between the Drosophila Hom-C genes and mouse Hox genes is not one-to-one, it is likely that inde­ pendent gene duplications and d.eletions have occurred since these two allinlal groups diverged (Hunt and Krum­ lauf 1992). Indeed, the most posterior mouse Hox gene (equivalent to Dl'osaphiia AbdB) underwent its own set of duplications in some mammalian chromosomes (see Fig­ ure 8.30).

Anterior-posterior patterning: The Hox code hypothesis

In aJJ vertebrates, anterior-posterior polarity becomes spec­ ified by the expression of Hox genes. Hox genes are homologous to the homeotic selector genes (Hom-C genes) of the fruit fly (see Chapter 6). The Drosophila homeotic gene complex on chromosome 3 contains the Antennape­ din and bitlwmx clusters of homeotic genes (see Figure 6.35) and can be seen as a single functional unit. (Indeed, in some other insects, such as the flour beetle Tribolium, it is a single physical unit.) All of the known mammaHan genomes contain four copies of the Hox complex per hap­ loid set, located on four different chromosomes (Hoxa through Hoxd in the mouse, HOXA through HOXD in hwnans; see Boncinelli et a1. 1988; McGinnis and Krurn­ laul 1992; Scott 1992). The order of these genes on their respective chromo­ somes is remarkably similar between .insects and hlunans, as is the expression pattern of these genes. Those mam­ malian genes homOlogous to tile Drosophila labial, probosci­ pedia, and deformed genes are expressed anteriorly and early, while those genes homologous to the Drasaphila AbdB gene

(10 hours)

Drosophila embryo

Drosophila

Hom -C

.,."-n- ---c'-,---;-,-..-+ ...., -I f""r-rT--'-'..u .... """. --'-"_ """_=l. --'-" """fL../

-' !-=""""""''-__� . .. ... , ,-J----, .....

...

.. ..

............................

Mouse

a1 Hoxo 0 lJ

FIGURE 8.30 Evolutionary conservation of homeotic gene organization and tran­ scriptional expression in fruit flies and mice is seen in the similarity between the Hom-C cluster on Drosophila chromo­ some 3 and the four Hox gene clusters i n the mouse genome. The mouse genes of the higher numbered paralogous groups are those that are expressed later and more posteriorly. Genes having similar structures occupy the same relative posi­ tions on each of the four chromosomes, and paralogous gene groups display simi­ lar expression patterns. The comparison of the transcription patterns of the Hom-C and Hoxb gen es of Drosophila and mice are shown above and below the chromo­ somes, respectively. (After Carroll 1 995.)

[)lJ ) ) bl

Hoxb

) )".1 02

b2

LION bl W

03

b3

(12 days)

Mouse embryo

a6

b5

b6

liii ) ) W__

) Liiii d3

a5

.. ) 1 lJ b4

e4

Hoxc 0=::: d1 Hoxd O U

a4

d4

e5

e6

07

a9

U U � b7

b8

alO

all

at3

b9

J

U U lOOIl liOil lIOl l1Ii e8

e9

c10

ell

cJ2

cl3

LldOll � lJ l;l liIi d8

d9

d10

dll

d12

Neural tube Somites

d13

l 1

\,

tj

B I RDS AND MAMMALS

315

FIGURE 8.31 Schematic representation of the chick and mouse vertebral pattern along the anterior-posterior axis. (A) Axial skeletons stained with Aldan blue at comparable stages of devel opmen t The chick has twice as many cervical verte­ brae as the mouse. (8) The boundaries of expression of certai n Hex gene pa ra la gous groups (Hox516 and Hox9/l0) have been mapped onlo the vertebral type domains. (A from Kmila and Duboule 2003, cou rtesy of M. Kmil. and D. Duboule; B atter Burke el .1. 1 995.) .

­

Chick

Mouse

(B)

5 Chick

6 Sacral Coccygeal �� Vertebrae

Cervical

Somites Vertebrae

Mouse Sacral Hox Hox 5 6

• Occipital

l' Cervical

EXPRESSION OF HOX GENES ALONG THE DORSAL AXIS Hox gene expression can be seen along the mammalian dorsal axis (in the neuraJ tube, neural crest, paraxial mesoderm, and surface ectoderm) from the anterior boundary of the hindbrain through the tail. The regions of expression are not in register, but the 3' Hox genes (homologous to labilll, proboscopedia, and deformed of the fly) are expressed more anteriorly than the 5' Hox genes (homologous to Ubx, abdA, and AbdB). Thus, one generally finds the genes of paralo­ go us group 4 expressed anteriorly to those of paralogous group 5, and so forth (see Figure 8.30;.W!lkinson et a1. 1989; Keynes and Lumsden 1990). Mutations in the Hox genes suggest that the level of the body along the anterior-pos­ terior axis is determined primarily by the most posterior Hox gene expressed in that region.

Experimental analysis of the Hox code The expression patterns of mouse Hox genes suggest a code whereby certain combinations of Hox genes specify a par­ ticular region of the aoterior-posterior axis (Hunt and Krum­ lauf 1991). Particular sets of paralogous genes provide seg­ mental identity along the anterior-posterior axis of the body. Evidence for such a code comes from three sources: •

Comparative anatomy, in which the types of vertebrae in different vertebrate species are correlated with the con­

stellation of Hox gene expression

Caudal

10

• Thoracic

•

Lumbar

Sacral

• Caudal

•

Gene targeting or "knockout" experiments.in which mice are constructed that lack both copies of one or more Hox genes

•

Retinoic acid teratogenesis, in which mouse embryos exposed to RA show an atypical pattern of Hox gene expression along the anterior-posterior axis and abnor­ mal differentiation of their axial structures

COMPARATIVE ANATOMY A new type of comparative embryology is emerging based on the comparison of gene expression patterns among species. Gaunt (1994) and Burke and her collaborators (1995) have compared the ver­ tebrae of the mouse and the chick (Figure 8.31A). Although the mouse and chick have a similar number of vertebraef they apportion them differently. Mice (like all mammals, be they giraffes or whales) have 7 cervical (neck) vertebrae. These are followed by 13 thoracic (rib) vertebrae, 6 lumbar (abdominal) vertebrae, 4 sacral (hip) vertebrae, and a vari­ able (20+) number of caudal (tail) vertebrae. The chick, on the other hand, has 14 cervical vertebrae, 7 thoracic verte­ brae, 12 or 13 (depending on the strain) lumbosacral ver­ tebrae, and 5 coccygeal (fused tail) vertebrae. The researchers asked, Does the constellation of Hox gene expression correlate with the type of vertebra formed (e.g., cervical or thoracic) or with the relative position of the ver­ tebrae (e.g., number 8 or 9)? The answer is that the constellation of Hox gene expres­ sion predicts the type of vertebra formed . In the mouse,

CHAPTER 8

316

GENE TARGETING As noted above, there is a specific pat­

(A)

tern to the number and type of vertebrae in mice (Figure 8.32A), and the Hox gene expression pattern dictates which vertebral type will form. This was demonstrated when all six copies of the HoxlO paralogous group (i.e . , HoxalO,

dO,

and dlO in Figure 8.30) were knocked out and no lumbar vertebrae developed. Instead, the presumptive lumbar ver­ tebrae fom,ed ribs and other characteristics similar to those of thoracic vertebrae ( Figu re 8.32B). This was a homeotic transformation comparable to those seen in insects; how­ ever, the redundancy of genes in the mouse made it much

(B)

Tl3

more difficult to produce, because the existence of even one copy of the Hoxl0 group genes prevented the transfor­ mation (Wellik and Capecchi 2003; WeUik 2009). Similarly,

,

when all six copies of the Hoxl1 group were knocked out,

""

"8�

the thoracic and lumbar vertebrae were normal, but the

.g

sacral vertebrae failed to form and were replaced by lum­ bar vertebrae (Figu re 8.32C).

Lumbar

RETINOIC ACID TERATOGENESIS Homeotic changes are also seen when mouse embryos are exposed to teratogenic doses of reti.noic add, a derivative of vitamin A. As we saw

(e)

earlier, by day

TIJ

7 of development, an RA gradient has been in the posterior regions and low in

established that is high

the anterior portions of the embryo. This gradient appears

,

to be controlled by the differential synti,esis or degrada­

""

tion of RA in the different parts of the embryo. Hox genes

"8� '"

•

Lumbar

RA receptor sites in their enhancers or by being responsive

are responsive to retinoic acid either by virtue of having

to Cdx (the mammalian homologue of the Drosophila cau­ dal gene), which is activated by RA (Conlon and Rossant

1992; Kessel 1992; Sakai et a1. 2001; Lohnes 2003). Exogenous retinoic acid can mimic the RA concentra­

Axial s kelelon s of m i ce in gelle knockout experi­ ments. Each phOiograph is of an l 8 . S -day embryo, looking upward at the ventral region from the middle of the thorax toward the tail. (A) Wild-Iype mou se. (8) Complete knockout of Hox l O paralogues (Hox lOaaccdd) converts lumbar vertebrae (after the thirteenth thoracic vertebrae) into ribbed thoracic vertebrae, (C) Complete knockout of Hox 1 1 paralogues (H.ox l 1aaccdd) trans­ forms the sacral vertebrae into copi es of lumbar vertebrae. (After Wellik an d Capecchi 2003, courtesy of M. Capecchi.) FIGURE 8.32

tions normally encountered only by the posterior cells, and high doses of RA can activate Hox genes

in more anterior

locations along the anterior-posterior axis (Kessel and Gruss

1991; Allan et a1. 2001). Thus, when excess retinoic

acid is administered to mouse embryos on day 8 of gesta­ tion, shifts in Hox gene expression occur such that the last cervical vertebra is turned into a thoracic (ribbed) verte­ bra. Conversely, impairment of retinoic acid function caus­

es Hox gene expression to become more posterior, and the first thoracic vertebra becomes a copy of the cervical ver­ tebrae (Figure 8.33).

the transition between cervical and thoracic vertebrae is

14 and 15 (Figure 8.31 B). In both cases, the Hox5 par­

between vertebrae 7 and 8; in the chick, it is between ver­ tebrae

alogues are expressed in the last cervical vertebra, while

the anterior boundary of the Hox6 paralogues extends to the first thoracic vertebra. Similarly, in both aItimals, the thoracic-lumbar transition is seen at ti,e boundary between the Hox9 and

HoxlO paralogollS groups. It appears there is

a code of differing Hox gene expression along the anteri­

The Dorsal-Ventral and left-Right Axes The dorsal-ventral axis tral axis formation in mammals. After the fifth cell division in the mouse embryo, the blastocyst cavity begins to form, Very little is known about the mechanisms of dorsal-ven­

and the inner cell mass resides on one side of this cavity.

2007). In mice and humans,

This axis is probably created by the ellipSOidal shape of tl.J.e

or-posterior axis, and that code determines the type of ver­

zona pelludda (Kurotaki et al.

tebra formed.

the hypoblast forms on the side of the inner cell mass that

B I R DS AND MAMMALS

(A) Normal development

(B) Retinoic acid (RA)

(C) Hox genes Gain

Gain

Loss

Excess RA

RA receptor

317

Loss

Hox overexpressioll

Hox mutants

Cl

C7 C7

Tl

C6

Tl

Tl

Tl

Tl

T2

T2

T2

T2

fiGURE 8.33 Effec l of retinoic acid on mouse embryos. Addi­ tion of exogenous RA leads to differences in Hox gene expression and the transformation of vertebral form. Similarly, loss of RA function (through mutations in the RA receptors) can.lead to the opposite type of transformations. (Aher Houle et al. 2003.J

is exposed to the blastocyst fluid, while the dorsal axis forms from those ICM cells that are in contact with the tro­ phoblast and amnionic cavity. Thus, the dorsal-ventral axis of the embryo is defined, in part, by the embrYOnic-abem­ bryonic axis of the blastocyst. The embryonic region con­ tains the lCM, while the abembryonic region is that part of the blastocyst opposite the ICM. The first dorsal-ventral polarity is seen at the blastocyst stage, and as development proceeds, the primitive streak maintains this polarity by causing migration ventrally from the dorsal surface of the embryo (Goulding et a1. 1993).

The left-right axis The mammalian body is not symmetricaL Although the human heart begins its formation a t the midline of the embryo, it moves to the left side of the chest cavity and loops to the right. The spleen is found solely on the left side of the abdomen, the major lobe of the liver forms on the right side of the abdomen, the large intestine loops right to left as it traverses the abdominal cavity, and the right lung has one more lobe than the left lung (figure 8.34). Mutations in .mice have shown that there are two levels of regulation of the left-right axis: a global level and an organ-specific level. Mutation of the gene situs i,1versus vis­ "rllrn (iv) randomizes the left-right axis for each asymmet­ rical organ independently (Hummel and Chapman 1959;

� ��

�.

C7 T2

Layton 1976). This means that the heart may loop to the left in one homozygous animal but to the right in another. Moreover, the direction of heart looping is not coordinat­ ed with the placement of the spleen or stomach. This lack of coordination can cause serious problems, even death. A second gene, inversion of embryonic turning (inv), causes a more global phenotype. Mice homozygous for an insertion mutation at this locus had all their asymmetrical organs on the wrong side of the body (Yokoyama et a1. 1993) ' Since all the organs were reversed, this asymmetry did not have dire consequences for the mice. Several additional asymmetrically expressed genes have recently been discovered, and their influence on one another has enabled scientists to arrange them into a pos­ sible pathway. The end of this pathway-the activation of Nodal proteins and the Pitx2 transcription factor on the left side of the lateral plate mesoderm-appears to be the same in mammals as in other vertebrate embryos, although the path leading to this point differs among the species (see Figure 8. 14; Collignon et al. 1996; Lowe et aL 1996; Meno et al. 1996). In mammals, the distinction between left and right sides begins in the ciliary cells of the node (figure 8.3SA) . The cilia cause fluid in the node to flow from right to left. When Nonaka and colleagues (1998) knocked out a mOUSe gene encoding the ciliary motor protein dynein (see Chapter 4), the nodal cilia did

colleagues randomly inserted the tr :. .. . � � \':. . . ... : Enamel ·: : ::

Ameloblast progenitor cells

,

AmeIo blast s

(making enamel)

stem cell

T H E STEM CELL CONCEPT

(A)

(C)

(8)

GermJine stem cell

Hub ccUs

Cadherins "-'�""-Centrosome Hub cells

Germline stem cell

Gonialblast

FIGURE 11 1.6 Stem cell niche in Drosophila tes tes. (A) The apical hub consists of about 1 2 somatic cells, to which are attached 5-9 germ stem cells. The germ stem cells divide asym­ m etrica l ly to form another germ stem cell (which remains attached to the somatic hub cells) Jnd a gonialblast that will divide to form the sperm precursors (the spermatogonia and the

spermatocyte cysts where meiosis is i n i ti ated) . (B) Reporter � ga lac tosidase inserted into the gene for Unpaired reveals that this prote i n is transcribed in the somatic hub cells. (C) Cell division pattern of the germ l i ne stem cells, wherein one of the centrosomes remains in the cortical cytoplasm near the site of hub cell adhesion, while the other centrosome migrates to the opposite pole of the germl i ne stem cell. This results in one cell remaining attached to the hub and the other cell detaching from it and differentiating. (Arter Tul ina and Matunis 2001; -

B courtesy of E. Matunis.)

the stem cells are in close proximity to the bone cells (osteocytes) and the endothelial cells that l i ne the blood vessels. A complex cocktai l of paracrine factors, including Wnts, angiopoietin, and stem cel l factor, combines with cell s u rface signals from Notch and integrin to regulate stem cell prol iferation and differentiation. Hormonal signals and pressure from the blood vessels, as well as neurotransmitters from adja­ cent axons, also help regulate hematopoiesis (see Spiegel et a l . 2008; Malhotra and K i ncade 2009) . In many instances, germ cells are continuously p rod uced in stem cell niches. In h Drosophila tes tes, the stem cells for sperm reside in a regulatory microenvironment called the hub (Figure 1 11.6). The ub consists of about a dozen somatic testes cells and is surrounded by 5-9 germ stem cells. The division of the sperm stem cell is asym­ metric, always producing one cel l that remains attached to the hub and o n e unat­ tached cell. The daughter cell attached to the hub is maintained as a stem cell, while the cell that is not touching the hub becomes a gonialblast-a committed cell that will divide to become the precursors of the sperm cells. The somatic cells of the hub create this a symmetric proliferation by secreting the paracrine factor Unpaired onto the cells attached to them. Unpaired protein activates the JAK-STAT pathway in the adja­ cent germ stem cells to specify their self-renewal. Those cells that are distant from the paracrine factor cannot receive this signal, so th ey begin their differentiation into the s perm cell lineage (Kiger et al. 2001; Tul ina a n d Matunis 2001 ). Physically, this asymmetric division involves the i nteractions between the sperm stem cells and the somatic ce l l s I n the division of the stem cel l, one centrosome remains attached to the cortex at the contact site between the stem cell and the somat­ ic cells. The other centrosom e moves to the opposite side, thus establishing a mitot­ ic spindle that w i l l produce one daughter cell attached to the hub and one daughter .

329

330

PART I I I

cell away from it (Yamashita et a l . 2003). (We w i l l see a similar inheritance of cen­ trosomes in the division of mammalian neural stem cells.nhe cell adhesion mole­ cules linking the hub and stem cells together are probably i nvolved in retai n i ng one of the centrosomes in the region where the two cells touch. The stem cell niche is a criti ca l component of our phenotype, and it does nothing less than regulate the ratio of cell division to cell differentiation. This means that main­ tenance of such niches is critical for our health. Too much stem cell d ifferentiation depletes the stem cells and promotes the phenotypes of agi ng or decay. Too much stem cel l division can cause cancers to arise. The "graying" of mammalian hair (in mice as well as humans) can involve a breakdown in the maintenance of the stem cell niche such that both daughters of the divi di ng stem cells d ifferentiate, thereby leaving a smaller population of stem cel ls that cannot continue to make pigment­ forming melanocytes (Nishimura et a l . 2005; Steingrimson et a l . 2005). Conversely, myeloproliferative disease-a cancer of blood stem cells and their (non-lymphocytic) derivatives-results when the stem cell niche is unable to provide the signals need­ ed for proper blood cell differentiation (Walkley et a l . 2007a,b) Thus, stem cell niches provide microenvironments that regulate stem cell renew­ al, survival, and differentiation. Their paracrine factors, cell adhesion molecules, and architecture al low asymmetric cell divisions such that a stem cell divides i n a man­ ner that a l lows one of its daughter cells to have a high probability of leaving the niche and beginning to differentiate according to the new signals it encounters. .

Mesenchymal Stem Cells: Multipotent Adult Stem Cells

Adult stem cells are used in the adult body to replace worn-out somatic cells on a regu lar basis. Our epidermis, our intestinal epithelium, a n d our blood cells are con­ ti nual ly being replaced with cells generated by d i vidi ng a d u l t stem cells. Most (if not a l l) adult stem cells are restricted to forming only a few cell types (Wagers et al. 2002). When hematopoietic stem cells marked with green fluorescent protein were placed in mice, their labeled descendants were found throughout the animals' blood but not i n any other tissue. Some adult stem cells, however, appear to have a sur­ prisingly large degree of plasticity. These mul ti potent cells are called mesenchymal stem cells, or MSCs (sometimes called bone marrow·derived stem cells, or BMDCs), and their potency remains a controversial subject. Originally found in bone marrow (Friedenstein et al. 1 968; Caplan 1 99 1 ), m ultipo­ tent MSCs have also been found in adult tissue such as fat, muscle, thymus, and den­ tal pulp, as well as i n the umbilical cord (see Kuhn and Tuan 201 0). Indeed, the find­ ing that human umbilical cords and deciduous ("baby") teeth contain MSCs (Gronthos . et a l . 2000; H i rata et aL 2004; Traggiai et a l . 2004; Perry et a l . 2008) has led some physicians to propose that parents freeze cells from their chi ld's umbil ical cord or teeth so that these cells w i l l be available for transplantation later in life. However, the crucial test for pluripotency-the ability of a mo u se stem cell to generate cells of a l l germ layers when inserted into a blastocyst-has not yet been achieved. Mesenchymal stem cells are able to give rise to numerous bone, cartilage, mus­ cle, and fat l ineages !Pittenger et al. 1 999; Dezawa et al. 2005). The differentiation of MSCs is predicated on both paracrine factors and cel l matrix molecules in the stem cell niche. Certai n cell matrix components, especially lam in in, have been implicated in keeping MSCs in a state of undifferentiated "stemness" (Kuhn and Tua n 201 0). Cer­ tain paracrine factors appear to direct development into specific lineages. In one study (Ng et a l . 2008), platelet-derived growth factor was critical for chondrogenesis and fat formation, TGF-� signaling was important for chondrogenesis, and FGF Signaling was crucial for the d ifferentiation of MSCs into bone cel ls I n addition to paracrine factors, the repertoire of cell types from mesenchymal stem cells may also be enhanced by the surfaces on which the stem cells reside. Human .

THE STEM CELL CONCEPT

- p 3 tubulin (neural) - CBra! (bone) ..... MyoD (muscular)

..... + Dlebbistatin (no cytoskeleton)

• • •

Subslrare

elasticity (kPa) P3 Tubulin (neurogenic)

MyoD (myogenic)

CBFal (osteogonic)

MSCs can differentiate according to the elasticity of the surface on which they are placed. If p laced on soft matrices of collagen-coated polyacrilamide, these stem cells differentiate into neurons. A moderately elastic malrix of the same materials causes the same stem cells to become muscle cells, while harder matrices cause the MSCs to produce bone cells (Figure 111.7; Engler et al. 2006). It is not yet known whether this range of potency is found normally in the body. Mesenchymal stem cells have recently been linked to normal growth and repair conditions in the human body. Indeed, one premature aging syndrome, Hutchinson­ Gilford progeria, appears to be caused by the inability of MSCs to differentiate into certain cell types, such as fat cells (Scaffidi and Misteli 2008). These fi ndi n gs lead to speculation that the loss of MSC ab i l i ty to differentiate may be a component of the normal aging syndrome. Moreover, MSCs may work in ways other than differentiating into needed cel l types. They may produce paracrine factors that aid olher, more spe­ cific stem cells to divide and repair tissues (Gnecchi et al. 2009). '

A New Perspective on Organogenesis The ability to create, isolate, and manipu late stem cells offers a vision of regenera­ tive medicine, wherein a patient can h ave his Or her d iseased organs regrown and replaced by one's own stem cells. We wi l l deta i l the medical possibilities of stem cell therapy in Chapter 1 7 . But beyond their pOlential medical uses, stem cells tell us a great many facts about how the body is constructed and how it maintains its struc­ ture. Organs often form by the regulation of stem cells, and we will see that the skin, hair, blood, and parts of the nervous system rout i nely use stem cells in their construc­ tion. Slem cells certainly give credence to the view that "development never ends," and offer fascinating (if not frightening) potential ways to modify development.

331

FIGURE 111.7 Mesenchymal slem cell dirrerentialion is influenced by the elasticity of the matrices upon which the cells sit. {On collage-coated gels having elasticity similar to that of the brain (aboul 01-1 kPa), Human MS(s differentiated into cells containing neu­ ral markers (such as p3 lubulin) bul not into cells containing muscle cell mark­ ers (MyoD) Or bone cel l markers «(BFal ). As the gel s became stiiier, the MSCs generated cells exhibiting mus­ cle-specific proteins, and even stiffer matrices elicited the differentiation of cells with bone markers. Differentia­ tion of the MSC on Clny matrix could be abolished wilh blebbi­ statin. which inhibits micro­ filamenl assembly al the cell 'membrane. (A�er Engler et al. 2006; photographs courtesy of ). Shields.)

The Emergence of the Ectoderm Central Nervous System and Epidermis

"WHAT IS PERHAPS THE MOST INTRIGUING question of all is whether the brain is powerful enough to solve the problem of its own creation." So Gregor Eichele

( 1992) ended a review of research on mammalian brain development. The con­

struction of an organ that perceives, trunks, loves, hates, remembers, changes, deceives itsel( and coordinates all of our conscious and unconscious bodily processes is undoubtedly the most challenging of all developmental enigmas. A combination of genetic, cellular, and organismal approaches is now giving us a very preliminary understanding of how the basic anatomy of the brain becomes ordered.

The fates of the vertebrate ectoderm are shown in Figure 9.1 . In the past two chapters, we have seen how the ectoderm is instructed to form the vertebrate

nervous system and epidermis. A portion of the dorsal ectoderm is specified to become neural ectoderm, and its cells become distinguishable by their columnar appearance. This region of the embryo is called the neural plate. The process by which the neural plate tissue forms a neural tube-the rudiment of the cen­ tral nervous system-is called neurulation, and an embryo undergoing such changes is called a neurula (Figure 9.2). The neural tube forms the brain ante­ riorly and the spinal cord posteriorly. This chapter will look at the processes by which the neural tube and the epidermis arise and acquire their distinctive patterns.

Establishing the Neural Cells This chapter follows the story of the committed neural and epidermal precur­ sor cells. Neural cells become specified through their interactions with other cells. There are at least four stages through which the pluripotent cells of the epi­ blast or blastula become neural precursor cells, or neuroblasts (see Figure 111.4; WIlson and Edlund 2001): •

Competence, wherein multipotent cells can become neuroblasts if they are exposed to the appropriate combination of Signals.

•

Specification, wherein cells have received the appropriate signals to become neuroblasts, but progression along the neural c1ifferentiation pathway can still be repressed by other Signals.

o

Commihnent (determination), wherein neuroblasts enter the neural differen­ tiation pathway and will become neu rons even in the presence of i.nhibitory Signals.

•

Differentiation, wherein the neuroblasts leave the mitotic cycle and express those genes cl1aracteristic of neurons.

For the real a mazement, if you wish to be amazed, is this process. You start out as a single cell derived from the coupling of a sperm and an egg; this divides in two, then four, then eight, and so on, and at a certain stage there emerges a single cell which has as all its progen y the human brain. The mere existence of such a cell should be one of the great astonishments of the earth. People ought to be walking around all do)" all through rheir wak­ ing hours calling to each other in end­ less wonderment, talking of nothing except that cell.

LEWtS THOMAS ( 1 979)

CHAPTER 9

334

Epidern1is Hair Nails Sebaceous glands Olfactory epithelium Mouch epithelium Lens. cornea

Ectoderm

t

Anterior pituitary Tooth enamel Surface ectoderm Cheek ium (epidermis) epithe l Schwaun Peri p heral cells nervous system Neuroglial cells Adrenal Neural medulla Sympathetic crest -----1 Melanocytes nervous system ParasymFacial pathetic cartilage nervous Dentine system of teeth Brain Neural pituitary Neural Spinal tube ----t- cord Motor neurons Retina

Q

r-,.

Neural p��la�te�-2�� � ,��

'Primitive streak



o Qo 1il) 0 0 0 " "

FIGURE 9.1 Major derivatives of the ectoderm germ layer. The ectoderm is divided into three major domairw the surface ecto­ derm (primarily epidermis), the neural crest (peripheral neurons, p i gm ent fac i a l cartil age), and the neural tube (brain and spinal cord). ,

Neural folds

(E) FIGURE 9.2 Gastrulation an d neurulation in a ch ick embryo (dorsal view). (A) Flat neural plate. (B) Flat neural plate with under­ lying notochord (head process). (e) Neural groove. (D) Neura l folds begin closing at the dorsalmost region, forming the inc ipie nt neural tube. (E) Neura l tube, showing the three brain regions and the spinal cord. The neural tube rema i ns ope n at the a nterior end and the optic bulges (which become the retinas) have extended to the lateral margins of the head. (F) 24-hour chick embryo, as n (D). The cephal i c (head) region has undergone neurulation, w le the caudal (tail) regions are still undergoing gastrulation. (A-E cou tesy of G. C. Schoenwolf; F aher PaNen 1 971 .) ,

i hi

r

CONSTRUCTING THE CENTRAL NERVOUS SYSTEM Formation of the N eural Tube There are two major ways of converting the neural plate into a neural tube. In primary neurulation, the cells sur­ roundi.ng the neural plate direct the neural plate cells to proliferate, invaginate, clnd pinch off from the surface to form a hoUow tube. In secondary neurulation, the neural

tube arises from the coalescence of mesenchyme cells into a solid cord that subsequently forms cavities that coalesce to create a hollow tube. In general, the anterior portion of the neural tube is made by primary neurulation, while the posterior portion of the neural tube is made by secondary neurulation. The comp ete neural tube forms by joining these two separately formed tubes together (Ha rringt on et aJ. 2009).

l

T H E EMERGENCE O F T H E ECTODE RM

(F)

Anterior

Ectoderm

�

mesoderm

� ::;:;;O-=,,\NeuraJ �_ _

region of coelom Yolk adherent [0 endoderm groove

of blastoderm

r==-_---\-- Cnmial mesenchyme

Ecoxtelroaembryonic�;:�����

Lateral body fold

____Ect"""l.-/,' floor

plate

(E) D

Dbxl

Irx3

Dbu

�====.oJ - . -- - --

-

-

---

- -- -

Pax6

-

(F)

(G)

--

FIGURE 9.13 Dorsal-ventral specification of the neural tube. (A) The newly formed neural tube is i n fluenced by two signaling centers. The roof of the neural tube is exposed to BM P4 and BMP7 from the e pi derm i s, and the floor of the neural tube is exposed 10 Sonic hedgehog pro· tein (Shh) from the notochord. (8) Secondary signaling centers are established in the neural tube. BMP4 is ex pressed and sec reted from the roof plate cells; Sonic hedgehog is expressed and secret­ ed from the floor plate cells. (C) BMP4 establishes a nested cas­ cade ofTGF-p factors, spreading ve n trall y into the neural tube from the roof p late . Sonic hedgehog diffuses dorsall y as a gradient from the floor p late cells. (D) The neurons of the spinal cord are given their identities by their exposure to these gradie n ts of paracri ne factors. The amounts and types of paracri ne factors pres­ ent cause different transcription factors to be activated in the nuclei of these cells/ depending on their position in the neural

Sonic hedgehog, and this paracrine factor from the floor plate cells forms a gradient that is highest at the most ven­ tral portion of the neural tube (Roelink et al. 1995; Briscoe et al. 1999). The dorsal fates of the neural tube are established by proteins of the TGF-p superfamily, especially bone mor­ phogenetic proteins (BMPs) 4 and 7, dorsalin, and activin

tube. IE) Sonic hedgehog is confined to the ventral region of the neural tube by the TGF p factors, and the gradient of Sonic hedge­ hog specifies the ventral neural tube by activating and inhibiting the synthesis of particular transcription factors. (F) Chick neural tube, showing areas of Sonic hedgehog (green) and the expression domain of the TGF-p-family protein dorsalin (blue). Motor neu­ rons induced by a particular concentration of Sonic hedgehog are stained orange/yellow. (G) In situ hybridi zat io n for three tra ns crip­ tion factors: Pax7 (blue, characteristic of the dorsal neural tube cells), Pax6 Igreen), and Nkx6.1 (red). Where Nkx6.1 and Pax6 overlap (yellow), the motor neurons become spec ified (E, F cour­ tesy ofT. M. Jessell; F from Jesse1l 2000; G courtesy of J. B riscoe .) -

.

(Liem et al. 1995, 1997, 20(0). Initially, BMP4 and BMP7 are found in the epidermiS. Just as the notochord establishes a secondary signaling center-the floor plate cells-on the ventral side of the neural tube, the epidermis establishes a secondary signaling center b y inducing BMP4 expression in the roof plate cells of the neural tube. The BMP4 pro­ tein from the roof plate induces a casca de of TGF-p pro-

T H E EMERGENCE OF THE ECTO D E RM

-�

Secondary lioor plate

(C)

(B)

(A)

. . .

f--''-----i

__ - V3

--4---,,-L-- FP

0.5 1 2 3 4

�

[ShhJ, nM

FIGURE 9.14 Cascade of inductions in i tiated by the notochord in the ventral neural tube. (A) Rel ationship between Sonic hedge­ h og concentrations, the generation of pa rti cu lar neuronal Iypes in vitro, Clnd distance from the notochord. (8) Two cell types in the newly formed neural tube. Those c losest to the notochord become the floor plate neurons; motor neurons emerge on the ventrolater­ al sides. (C) If a second notoch o rd, floor plate, or any other Sonic hedgehog-secreting cell is placed adjacent to the neura l tube, it induces a second set of floor plate neurons, as well as two other sets of motor neurons. (A after Briscoe el al. 1 999; B/C after Placzek et al. , 990.)

Secondary set of motor neurons

, ::il,f::/ .

§ .-

-+-

Floor plate

345

I

Donor notochord, floor plate, or other Shh-secreting c('lls

Notochord

plate cells at its sides

(Figure 9.1 4B,C). These ectopic floor

plate cells then induce the formation of motor neurons on

either side of them. The same results can be obtained if the notochord fragments are replaced by pellets of cultured

mutants. Those mu tants deficient in certain BMPs lacked

ceUs secreting Sonic hedgehog (Echelard et aI. 1993). More­ over, if a piece of notochord is removed from an embryo, the neural tube adjacent to the deleted regjon will have no floor plate ceUs (Placzek et a1. 1990). By combining infoffiiation about the anter ior-posterio r axis with information establishing the dorsal-ventral axis, netuons can become uniquely specified for certain regions. For instance, one would expect the motor neurons inner­ vating the forelimb to recognize different targets than the motor neurons innerva ting hlndHmb targets. The Hox genes (and especially the HoxC paralogue group) play important roles in determining motor neuron targets

dorsal and intermediate types of neurons (Nguyen et a1.

(Dasen et a1. 2003, 2005).

teins n i adjacent cells (see Figure 9.13C). Dorsal sets of ceUs

are thus exposed to higher concentrations of TGF-p pro­

teins and at earlier times, when compared with the more

ventral neural cells. The importance of the TGF-p super­ family factors in patterning the dorsal portion of the neu­ ral tube was demonstrated by the phenotypes of zebrafish

2000). The paracrine factors interact to instruct the synthesis

of different transcription factors along the dorsal-ventral

See WEBSITE 9.3 Constructing the pituitary gland

axis of the neural tube. In the ca�e of the ventral neural

Differentiation of Neurons i n the Brain

tube, a gradient of Sorric hedgehog specifies different cell

The human brain consists of more than

identities as a function of its concentration. Cells adjacent

ciated with over 1 0

lOll neurons asso­ 12 glial cells. The neuroepithelial cells

to the floor plate that receive high concentrations of Sonic

of the neural tube give rise to three main types of cel1s.

hedgehog (and hardly any TGF-� signal) synthesize the

First, they become the

Nkx6.1 and Nkx2.2 transcription factors and become the

remain integral components of the neural tube lining and

ventral (V3) neurons. The cells dorsal to these, exposed to

that secrete the cerebrospinal fluid. Second, they generate

tors, produce Nkx6.1 and Pax6 transcription factors, and

and coordinate our bodily functions, our thoughts, and our

slightly less Sonic hedgehog and slightly more TGF-p fac­

they become the motor neurons. The next two groups of

ventricular (ependymal) cells

that

the precursors of the n.eurons that conduct electric potentia Is sensations of the world. Third, they give rise to the precur­

ceUs, receiving progressively less Sonic hedgehog, become

sors of the glial cells that aid in constructing the nervous sys­

2001; Muhr et a1. 2(01).

be important in memory storage. As we

the V2 and VI interneurons ( F ig ure 9.14A; Lee and Pfaff The importance of Sonic hedgehog in indUCing and pat­

terning the ventral portion of the neural tube can be shown

experimentally. If notodlOrd fragments are taken from one

embryo and transplanted to the lateral side of a host neu­

ral tube, the host neural tube will form another set of floor

tem, provide insulation around the neurons, and that may

have seen, the dif­ be large­

ferentiation of these precursor cells is believed to

ly determined by the environment they enter (Rakic and Goldman 1982). At least in some cases, a given neuroep­ itheHal cell can presumably give rise to both neurons and

glia (Turner and Cepko 1987).

346

CHAPTER 9

Input

Diagram of a motor neuron. Elec­ tric impulses (red arrows) are received by the den­ drites, and the stimulated neuron transmits impulses th rough its axon to its target tissue. The axon (which may be 2-3 feet long) is a cellular process through which the neuron sends its sig­ nals. The growth cone of the axon is both a loco­ motor and a sensory apparatus that actively explores the environment and picks up directional cues telling it where to go. Eventually the growth cone wi II form a connection , or synapse, with the axon's target tissue. FIGURE 9,15

The brain contains a wide variety of neuronal and glial types (as is evident from a comparison of the relatively small granule cell neuron with the enormous Purkinje neu­ ron). The fine, branclting extensions of the neuron that are used to pick up electric impulses from other cells are called dendrites (Figure 9. 1 5) . Some neurons develop only a few dendrites, whereas other cells (such as the Purkinje neu­ rons) develop extensive dendritic arbors. Very few dendrites are found on cortical neurons at birth, and one of the amaz­ ing events of the first year of human life is the increase in the number of these receptive processes. During this year, each cortical neuron develops enough dendritic surface to accommodate as many as 100,000 cOlmections (synapses) with other neurons. The average cortical neuron connects with 10,000 other neural cells. This pattern of synapses enables the human cortex to function as the center for learning, reasoning, and memoryi to develop the capacity for symbolic expressioni and to produce voltmtary respons­ es to interpreted stimuli. Another important feature of a developing neuron is its axon (sometimes called a neurite). Whereas dendrites are often numerous and do not extend far from the neuronal cell body, or soma, axons may extend 2-3 feet (see Figure 9.15). The pain receptors on your big toe" for example, must transmit their messages all the way to your spinal cord. One of the fundamental concepts of neurobiology is that the axon is a continuous extension of the nerve cell body. At the beginning of the twentieth century, there were many competing theories of axon formation. Theodor Schwann, one of the founders of the cell theory, believed that numer­ ous neural cells linked themselves together in a chain to form an axon. Viktor Hensen, the discoverer of the embry­ onic node, thought that the axop formed around preexist­ ing cytoplasmic threads between the cells. Wilhelm His (1886) and Santiago Ram6n y Cajal (1890) postulated that the axon was an outgrowth (albeit an extremely large one) of the neuronal soma. In 1907, Ross Harrison demonstrated the validity of the outgrowth theory in an elegant experiment that founded both the science of developmental neurobiology and the technique of tissue culture. Harrison isolated a portion of

/ other

via

neurons t -------;; Dendrites axons from

lpma L;. . �=I,F�== '(;:J' (cell body)

f

'>-

-"

.. .�

3

2/3 4

5 6

� ��

Subventricular zone (white matter)

60 40 20

�---,,------' Whma iter 2/3

5

4

6

)

Cortical layer s

Homigsrtanteiounral �

tO t

Vewnntericular Cel -autonomousfate ..h.. ena e transplanted after last S ph s

thre

FIGURE 9.24 Cortical neurons are generated from types of neural precursor cells: radial glia cells, short neural precursors, and intermediate pro­ genitor cells. RGCs and SNPs divide al the apical (luminal) surface of the ven· tricular layer. SNPs are committed neu­ ral precursors. IPes divide away from the luminal surface in the ventricular ?nd subventricular zones. Most IPes undergo neurogenic divisions, with a small fraction undergoing symmetrical proliferative divisions (dotted circular arrow). Through asymmetrical divisions, RGCs give rise to IPes that migrate to the subventricular layer. The ventricular zone generates lower-layer neurons; the subventricular zone generates upper­ layer neurons. (After De ay and Kennedy 2007.)

h

353

t Radial glial cell (RGC)

�

Short ncuronal precursor

(D)

Homi��gsrtanteiounral cel n e phfatse wh n _ Glial

Host (conditional) t ra spJam d in S

e

354

CHAPTER 9

SIDEL IGH TS SPECULATIONS

Adult Neural Stem Cells

O

(B) CA) ntil recently, i t was generally believed that once the mam­ malian nelVQUS system was mature, no new neurons were "born" -in other words, the neurons formed in utero and during the first few years of life were all we could ever expect to have. The good news from recent studies, however, is that the adult brain is capable of producing new neurons, and environmental stim­ ulntion can increase the n umber of these new neurons. In these experiments, researchers injected adult mice, rats, or mar­ mosets with bromodeoxyuridine (BrdU), a n ucleoside that resemb les thymidine. SrdU is incorporated into a ce l l 's DNA only if the cell is undergo­ i n g DNA replication; therefore, any cell labeled with BrdU must have been undergoing DNA synthesis dur­ ing the time it was exposed to BrdU. Figure 9.25 Evidence of adult neural stem This l abe l ing technique revealed that cells. The green s tai n i ng, which indicates newly divi ded cells, is from a fluo rescent thousands of new neurons are being antibody agai nst BrdU (a thym idi ne ana­ made each day in adult mice. More­ l ogue that is taken up only duri ng the 5 over, these new brain cells integrated phase of the cell cycle). (A) Newly gen erated with other cells of the brain, had nor­ adult mouse neurons (green cells) have a mal neuronal morphology, and exhib­ no rma l mo rpho logy and receive synaptic ited action potentials (Fi g u re 9.25A; inputs. The red spots are synaptophysin, a van Praag et a l . 2002). pro tei n found on the dendrites at the synaps· Injecting h u mans with SrdU is usu� es of axons from other neurons. (8) Newly ally unethical, since l a rge doses of generated neuron (arrow) in the adul t BrdU are often lethal. However, i n human brain. This cell is located in the den­ certain cancer patients, the progress of tate gyrus of the hippocampus. The red fluo­ chemotherapy is monitored by trans­ rescence is from an anti body that stains only fusing the patient with a small amount neural cells. Yellow indicates the overlap of of BrdU. Gage and colleagues (see red and green. Glial cells are stained purple. Erikkson et al. 1 998) took postmortem (A from van Praag et al. 2002; B from Erikks­ samples from the brains of five such son et al. 1 998, photograph courtesy of F. H. patients who died between 1 6 and Gage.) 781 days after the BrdU infusion. tn all five subjects, they saw labeled (new) neurons i n the granular cell layer of The existence of neural stem cells the h i ppocampal dentate gyrus (a part in adults is now well established for of the brain where memories may be the olfactory epithelium and the hip­ formed). The BrdU-labeled cells also pocampus {Kempermann et al. stained for neuron-specific markers 1 997a,b; Kornack and Rakic 1 999; (Figure 9.256). Thus, although the rate van Praag et al. 1 999; Kato et a l . of new neuron formation in adulthood 2001). These cells respond to Sonic may be relatively low, the human hedgehog and can proliferate to brain is not an a natomical fait accom­ become multiple ce l l types for at least pli at birth, or even after childhood. the first year of a mouse's l ife (Ahn

and Joyner 2005). I t appears that the

stem cells prodUCing these neurons are located in the ependyma (the for­ mer ventricular zone, i n which the embryonic neural stem cells once resided) or in the subventricular zone (SVZ) adjacent to it (Doetsch et al. 1 999; Johansson et al. 1 999; Cassidy and Frisen 2001). These adult neural stem cells represent only about 0.3% of the ventricle wall cell population, but they can be distinguished from more differentiated cells by their cell surface proteins* (Rietze et a l . 2001 ). In the adult mouse, thousands of new neuroblasts are generated each day, migrating from the lateral SVZ to the olfactory bulb, where they differenti­ ate into several different types of neu­ rons. Recent evidence suggests that *These neural stem cells may have partic� u l ar physiological roles as wel l. During

pregnancy, prolactin stimulates the pro� duction of neuronal progenitor celts in the subventricular zone of the adult mouse forebrain. These progenitor cells migrate to produce olfactory neurons that may be important for maternal behavior of rearing offspring (Shingo e t a l . 2003).

T H E EMERGENCE O F T H E ECTODERM

355

SIDELIGHTS & SPECULATIONS (Conti n u ed ) ste m ce l ls are n ot mult i poten t beco ming spec i fi ed only when th ey reach the o lfacto ry bulb) but i nstead are a pop u latio n of h eterogenous neu­ robl asts that a re already committed to becoming certain neuronal types ,V\e rk l e et al. 2007). These adult neu­ ral stem cells proliferate in respo nse to exercise, learni ng, and stress (Zhang et

th ese

al. 2008).

The existence of adult neural stem cells in the cortex is more controver­ sial (see Gould 2007). Some i nvestiga-

tors (Gould et al. 1 999a,b; Magavi et al. 2000) c laim to have identified them; other scie ntists (see Rakic 2002) q uestion the existence of these corti­ cal neural stem cells. The mechanisms by which neural stem cells are kept i n a state of ready qu iescence well into adulthood are c urrently being explo red . Before they become neu­ rons, neural stem cells are character­ ized by the expression of the NRSE translational inhibi tor that prevents neuronal differentiation by binding to

a si lencer region of DNA (see Chapter 2). When n eu ra l stem cells begin to differe nt i a te, they synthesize a small, double-stranded RNA that has the same sequence as the si lencer and which might bind NRSE and thereby perm it neuronal differentiation (Kuwabara et al. 2004). The use of c u l­ tured neuronal stem cells to regener­ ate or repair parts of the brain wi ll be considered in Chapter 1 7 .

CORTICAL CELL MIGRATION The first cortical neurons to be

times. One of these switches is the gene encoding the tran­

generated migrate out of the germinal zone to form the

When the mouse neuronal progen­ to form the first layer of cortical neurons, Foxgl is not exp ressed in the progenitor cells or in the first­ formed neurons. However, later, when the progenitor cells generate those neurons destined for layers 4 and 5, they express tlus gene. If tl.e FoxgI gene is conditionally knocked out of this In.eage, the neural precursor cells con­ tinually give rise to layer 1 neurons. Therefore, it seems

:ransient prep l ate (Kawauch.i and Hoshino 2008). Subse­ quently generated neurons migrate into the preplate and "'parate it into two layers: the Cajal-Retzius layer and the 5ubplate. The Cajal-Retzius layer becomes and remains the most superficial layer of the neocortex, and its cells express

the cell surface glycoprotein Reelin. The subplate remains .he deepest layer through which the successive waves of neuroblasts travel to form the cortical plate. The Reelin­ producing cells of the Cajal-Retzius layer are critical in the separation of the preplate. In Reelin-deficient mice, the pre­ plate fails to sp lit, and the neurons produced by the ger­ minal layers pile up behind the previously generated neu­

scription factor Foxg 1 . itor cells divide

that the Foxgl transcription factor is required to suppress

the "layer

1" neural fate.

Neither the vertical nor the horizontal organization of the cerebral cortex is clonally specified-that is, none of the functional units form from the progeny of a single cell.

(instead of migrating through them). By activating the

Rather, the developing cortex forms from the mixing of

'\latch pathway, Reelin on the surface of the Cajal-Retzius

cells derived from numerous stem cells. The early region­

rons

cells allows the neuronal stem cell to produce a long fiber

alization of the neocortex is

that extends through the cortical plate (Hashimoto-Torii et

paracrine factorS secreted by the epidermiS and neural crest

2008; Nomura et al. 2008). This (and the fact that it pro-

cells a t the margins of the developing brain (Rakic et al. 2009). The paracrine factors induce the expression of tran­ scription factors in the specific brain regi ons, which then mediate the survival, differentiation, proliferation, and migration of the newly generated neurons. For instance, Fgf8 protein is secreted by the anterior neu­ ral ridge and is important for specifying the telencephalon (Figure 9.26). If Fgf8 is overexpressed n. the ridge, specifi­ cation of the telencephalon is extended caudally, whereas if Fgf8 is ectopically added to the caudal region of the cortex, part of that caudal region will become anterior (Fukuchi­ Shimogori and Grove 2001, 2005). Sonic hedgehog is secret­ ed by the medial ganglioniC eminence and helps form the ventral neurons of the cortex, including those of the sub­

al.

uced some proteins thought to be glial-specific) caused the neural stem cell to be called the radial glial cell. The

process from this cell becomes cr iti.cal for

the migra tion of

the neural cells produced by the germinal zones. We also know that there are mutations that specifically affect the microtubular cytoskeleton of the migrating neu­

geJ:>e prevent neuronal with microtubule assembly; humans with such mutations have been seen to suffer from mental dysfunctions, among them autism, bipolar enchyme (Figure 1 1 .9). Ectodermal Signals appear to cause the peripheral somitic cells to undergo mesenchymal-to-epithelial transition by lowering the Cdc42 levels in these cells. Low Cdc42 levels alter the cytoskeleton, allowing epithelial cells to form a box around the remaining mesenchymal cells, which have a higher level of Cdc42. Another small GTPase, Rac1, must be at a certain level that allows it to activate Paraxi�, another tran­ scription factor involved in epithelialization (Burgess et a1. 1995; Barnes et aJ. 1997; Nakaya et aJ. 2004). The epithelialization of each somite is stabilized by syn­ thesis of the extracellular matrix protein fibronectin and the adhesion protein N-cadherin (Lash and Yamada 1986; Hatta et aJ. 1987; Saga et al. 1997; Linask et al. 1998). N-cad­ herin links the adjoining cells into an epithelium while the fibronectin matrix acts alongside the Ephrin and Eph to promote the separation of the somites from each other (Martins et a1. 2009).

Separation of somites from the unsegmented mesoderm Two proteins whose roles appear to be critical for fissure formation and somite sep aration are the Eph tyrosin e kinases and their ligands, the ephrin proteins. We saw in Chapter 10 that the Eph tyrosine kinase receptors and their ephrin ligands are able to elicit cell-cell repulsion between the posterior somite and migrating neural crest cells. The separation of the somite from the presomitic mesoderm occurs at the ephrin B2/Eph A4 border. In the zebra fish, the boundary between the most recently separated somite and the presomitic mesoderm forms between ephrin B2 in the posterior of the somite and Eph A4 in the most anteri­ or portion of the presomitic mesoderm (Figure 1 1 .8; Durbin et a1. 1998). Eph A4 is restricted to the boundary area in chick embryos as well. Interfering with this signallng (by injecting embryos with mRNA encoding dominant nega­ tive Ephs) leads to the formation of abnormal somite bOlllldaries. In addition to the posterior-ta-anterior induction of the fissure (from Eph proteins to ephrin proteins on their neighboring cells), a second signal originates from the ven­ tral posterior cells of the somite, putting all the ceUs in reg­ ister so that the cut is clean from the ventral to the doral aspects of the somite (Sato and Takahashi 200S).

,

Somite specification along the anterior­ posterior axis Although all somites look identical, they will form differ­ ent structures. For instance, the somites that form the cer­ vical vertebrae of the neck and lumbar vertebrae of the abdomen are not capable of forming ribs; ribs are generat­ ed only by the somites forming the thoracic vertebrae. Moreover, spedfication of the thoracic vertebrae occurs

Epithelialization of the somites Several studies in the chick have shown that epithelializa­ tion occurs immediately after somitic fission occurs. As

(A)

(B) Somites: Anterior Posterior

2

3

Somite number 5 4

7

-l ephriu 82 --+--""

EphA4 --+-�

Unsegmented paraxial mesoderm

FIGURE 1 1 .8 Ephrin and its receplor constitute a poss i ble fissure site for somite formation. (A) Expression pattern of the receptor tyrosine kinase Eph A4 (blue) and its ligand, ephrin 62 (red), as somites deve lop. The somite boundary forms at the junction between the region of ephrin expression on the posterior of the last somite formed and the region of Eph A4 expression on the anterior

of the next somi te to form. In the presomitic mesoderm, the pattern is created anew as each somite buds off. The posteriormosl (egion ofthe next somite to form does not express ephrin until that somite is ready to separale. (6) In situ hybridizalion showing Eph A4 (dark blue) expression as new somites form in the chick embryo. (A after Durbin el al. 1 998; 6 courtesy of J . Kaslner.)

421

PARAXIAL AND INTERMEDIATE MESODERM

(A)

Epithelial cel s ��t>1 r����������;��,! �e /�"':'J1�������.vIkQ Formed �;:..i • -..... h �

EMP,

i ��

: :

Wm

(sF P )

-2

h

RA, while the RA gradient p at terns the pharyngeal arch a graded manner. This is probably accom­ plished by activating and repressing particular sets of tran­ scri p tion factor genes. Highe levels of RA specify the Pdxl endoderm in

tube to develop their very different structmes. As the diges­ tive tube meets different mesenchymes, the mesenchyme cells instruct the endoderm to differentiate into esophagus, stomach, small intestine, and colon (Okada 1960; Gumpel­ Pinot et al. 1978; Fukumachi and Takayama 1980; Kedinger et a1. 1990). The endodenn and the splanchnic lateral plate mesoderm Wldergo a complicated set of interactions, and the signals for generating the different gut tissues appear to b e conserved throughout the vertebrate classes (see Wa l­ ace and Pack 2(03). The stabilization of the gut boundaries results from interaction with the mesoderm (Figure 12.28B) . As the gut tube begins to form at the anterior and p oste ior ends, it induces the splanchnic mesoderm to become regionally specific. Roberts and colleagues (1995, 1998) have impli­ cated Sonic hed gehog (Shh) in this specification. Shh is secreted in different concentrations at different sites, and its target appears to be the mesodenn surrounding the gut tube. The secretion of Shh by the hindgut endoderm induces a nested pattern of "posterior" Hox gene expres-

r

LATERAL PLATE MESODERM AND THE ENDODERM

Da y er bu Stomach a

(A)

i

Lv

Day 30

35

(B) Bile duct ep

d

H

(C)

� """G r\ Gal blad er

Gal blad er

e tr pbaundctrae tic pDobaundrcsraelatic Vpancreas rl

n

al

in the mesoderm. As in the vertebrae (see Chapter 8),

the anterior borders of Hox gene expression delineate the

sian

morphological boundaries of the regions tl,at will form the cloaca, large intestine, cecum, mid�cecum (at the midgut­ hindgut border), and posterior portion of the midgut (Roberts et at.

1995;

(AcpDan)ceDasreoayrtyi4c2duct

40

tic

•

Ven

Day

Yokouchi et at.

1995). When Hox­

(B)yBy

475

duct Ventral pancreatic d t

u

c

r c an i VentraJ ductanc e p

a ti

c reat c d uct

p

FIGURE 12.29 Pa n creat i c development in humans. (A) At 30 da s, the ventral pancreati c bud is close to the liver primordium. 35 days, it begi ns migrati ng posteriorly, and (C) comes i nto contact with the dorsal pa ncreati c bud during the sixth week of development. (D) In most individuals, the dorsal pancreatic bud loses its duct into the duodenum; however, in about 10% of the pop u lati on , the dual duct system p""ists.

expressing viruses cause the misexpression of these Hox genes in the mesodenn, the mesodermal celis alter the dif­ ferentiation of the adjacent endoderm ( Rober ts et al.

1998).

The Hox genes are thought to specify the mesoderm so that

muscle that normally lines the stomach and intes tine

it can interact with the endodermal tube and specify its

(TIleodosiou and rabin 2005). Thus, a signal from the small

are established, d ifferentiation can begin. The regional dif­

intestine instructs the mesoderm of the adjacent stomach

ferentiation of the mesoderm (into smooth muscle types)

Such communication allows the endoderm to form a

and the regional differentiation of the endoderm (into dif­

with very exact regional compartments.

regjons. Once the botmdaries of the transcription factors

tube

to tell its mesoderm and endoderm how to differentiate.

ferent functional units such as the stomach, duodenum, small intestine, etc.) are synchronized. The molecular signals by which the mesenchyme influ­ ences the gut tube are just becoming known (Figure

12.28C).

For instance, the mesenchyme lining the posterior (intes­ prote ins. 111ese Wnt proteins suppress genes such as Hhex tine-forming) region of tLte endodermal tube secretes Wnt

Liver, pancreas, and gallbladder organs that develop immediately caudal to the

The endoderm also forms the lining of three accessory

stomach: divertic­

the liver, pancreas, and gall bladder. The hepa tic

ulum

is a bud of endoderm that extends out from the

and prevent the stomach, liver, and pancreas from forming

foregut into the surrounding mesenchyme. The endoderm

(Bossard

of this bud comes from two populations of ceUs-a later­

and Zaret 2000; Mclin et a1. 2007). [n the anterior

regions of the gut tube (which form the thymus, pancreas,

al group that exclusively forms liver cells, and ventml­

stom ach, and liver), Wnt Signaling must be blocked. In the

medial endoderm cells that form several midgut regions,

stomach -forming region, the mesen chy m e lining the gut

including the liver (Tremblay and Zaret

tube expresses the transcription factor Barxl, which acti­

enchyme induces this endoderm to proliferate, branch, and

vates production of Frzb-Iike Wnt-blocking proteins (sFRPI and sFRP2) . These Wnt antagonists block Wnt signaling in the vicinity of the stomach but not around the intestine.

Kim et aJ. 2005.)

2005).

rhe mes­

form the glandular epithelium of the nver. A portion of the hepatic diverticulum (the region closest to the digesti ve

branch from this duct produces the gallbladder (Fig­ 12.29).

tube) continues to function as the drainage duct of the liver,

(Indeed, BanI-deficient mice do not develop stomachs and

and a

express intestinal markers in that tissue;

ure

In addition to FGF and Wnt signaling, BMPs also act in

TI,e pancreas develops from the fusion of distinct dorsal

regional specification. The intestinal mesenchyme secretes

and ventral diverticula. As they grow, they come closer

BMP4, which instructs the mesoderm anterior to it to express the Sox9 and Nkx2-5 transcription factors (see Fig­ lire 12.28C). These factors tell the mesoderm to become the muscles of the pyloric sphincter rather than the smooth

together and eventually fuse. In humans, only the ventral

to carry digestive enzymes into the intestine. In other species (such as the dog), both dorsal and ventral duct survives

ducts empty int'O the intestine.

476

CHAPTER 1 2

SIDEL I G H TS SPECULATIONS

Specification of Liver and Pancreas

T

here ;s an i ntimate relationship between the splanchnic lateral plate mesoderm and the foregut endoderm. Just as the foregut endo­ derm is critical in specifying the car­ diogenic mesoderm, the blood vessel endothelial cells induce the endoder­ mal tube to produce the liver primordi­ um and the pancreatiC rudiments.

liver formation The expression of liver-specific genes (such as the genes for a-fetoprotei n and albumin) can occur in any region of the gut tube that is exposed to car­ diogenic mesoderm. However, this induction can occur only if the noto­ chord is removed. If the notochord is placed by the portion of the endoderm normally i nduced by the cardiogenic mesoderm to become liver, the endo­ derm will not form liver (hepatic) tis­ sue. Therefore, the developi ng heart appears to induce the liver to form, while the presence of the notochord i n hi bits liver formation (Figure 12.30). This induction is probably due to FGFs secreted by the developing heart cells (Le Douarin 1 975; Gualdi et at. 1 996; Jung et a t . 1 999). However, Matsumoto and col­ leagues (200 1 ) found that the heart is not the only mesodermal derivative needed to form the l iver. Blood vessel endothel ial cells are also critical. If endothel ial cells are not present in the area around the hepatic region of the gut tube, the liver buds fail to form. This induction occurs even before the endothel ial cells have formed tubes, so it does not have anything to do with getting nutrients or oxygen into the region. Thus, the heart endothelial cells have a developmental function in addition to their circulatory roles: they induce the formation of the l iver bud by secreting FGFs. But in order to respond to the FGF signal, the endoderm has to become competent. This competence is given to the foregut endoderm by the fork­ head transcription factors. Forkhead

Figure 12.30 Pos itive and negative signal­ i ng in the formation of the hepatic (liver)

26-Day embryo

endoderm. The ectoderm and the notochord block the abil ity of the endoderm to express liver·specific genes. The cardiogenic meso· derm, probably through Fgf1 or Fgf2, pro­ motes liver·specific gene transcription by bl ocki ng the inhibitory fadors induced by the surrounding tissue. (After Gualdi et a l. 1 996.)

�5� 02 ": :�r.Ew ICE!2B =:::;:��

Dorsal endode rm hepatogenesis blocked

:

____- Hep'''i? region of ____

gut endoderm expressing a·fetal protein and albumin

mesoderm

transcription factors Foxa1 and Foxa2 are required to open the chromatin surrounding the liver-specific genes. These proteins d isplace nucleosomes from the regulatory regions surround­ ing these genes and are required before the FGF signal is given (Lee et a l . 2005). Mouse embryos lacking Foxa 1 and Foxa2 expression in their endoderm fai l to produce a liver bud or to express liver-specific enzymes. Once the Signal is given, other forkhead transcription factors, such as H N F4a, become critical. HNF4a is essential for the morphological and . biochemical differentiation of the hepatic bud i nto liver tissue (Parviz et al. 2003). When conditional (floxed) mutants of HNF4o. were made such that this factor was absent only in the developing l i ver, neither the tissue arch itecture, cell u lar structu re, or Jiverwspecifi c enzymes formed i n the liver bud cells. Meanwhile, Odom and colleagues (2004) found that fork­ head transcription factors were also critical for the differentiation of the endocrine islands of the pancreas.

H N F4a bound to the regulatory regions of almost half the actively transcribed pancreas-specific genes i n these tissues, including those involved in insulin secretion. A link between H N F4a mutations and late-onset type 2 diabetes has been observed (see Kulkarni and Kahn 2004), confi rming the importance of this transcription factor in pancreatic, as well as hepat­ ic, development.

Pancreas formation The formation of the pancreas may be the flip side of liver formation. Where­ as the heart cells promote and the notochord prevents liver formation, the notochord may actively promote pan­ creas formation, while the heart may block the pancreas from forming. I t seems that this particular region of the digestive tube has the ability to become either pancreas or l iver. One set of conditions (presence of heart, absence of notochord) induces the liver, while another set of conditions (presence of notochord, absence of heart) causes the pancreas to form.

LATERAL PLATE MESODERM AND THE E N DODERM

477

SIDELIGHTS & SPECULATIONS (Con i n u ed) The notochord activates pancreas development by repressing shh expres­ sion in the endoderm (Apelqvist et a l . 1 997; Hebrok e t a l . 1 998). (This was a surprising finding, si nce we saw i n Chapter 1 0 that the notochord is a source of Shh protein and an inducer of further shh gene expression i n ecto­ dermal tissues.) Sonic hedgehog is expressed throughout the gut endo­ derm, except in the region that wi l l form the pancreas. The notochord i n this region of the embryo secretes Fgf2 and activin, which are able to down­ regulate shh expression. If shh is experimental l y expressed i n this region, the tissue reverts to being intes­ tinal Uonnson et a l . 1 994; Ahlgren et a l . 1 996; Offield et al. 1 996). The lack of Shh i n this region of the gut seems to enable i t to respond to signals coming from the blood vessel endothelium. I ndeed, pancreatic development is initiated at precisely those three locations where the foregut endoderm contacts the endothelium of the major blood ves­ sels. It is at these points-where the endodermal tube meets the aorta and the vitel l i ne veins-that the transcrip­ tion factors Pdxl and Ptfl a are expressed (Figure 12.31A-C; Lammert et a l . 2001; Yoshitomi and Zaret 2004;). If the blood vessels are removed from this area, the Pdxl and Ptf1 a expression regions fai l to form, and the pancreatic endoderm fails to bud; if more blood vessels form i n this area, more of the endodermal tube becomes pancreatic tissues . The association of the pancreatic tissues with blood vessels is critical i n the formation of the insuli n-secreting cells of the pancreas. Pdx1 appears to act i n concert with other such tran­ scription factors to form the endocrine cells of the pancreas, the islets of Langerhans (Odom et al. 2004; Burli­ son et a l . 2008; Dong et a l . 2008). The exocrine cells (which produce diges­ tive enzymes such as chymotrypsin) and the endocrine cells appear to have the same progenitor (Fishman and Melton 2002), and the level of Ptf1 a appears to regulate the proportion of cells in these l ineages. The exocrine pancreatic cells have higher amounts of Ptf i a (Dong et a l . 2008). The islet

(A)

Vitelline veins (B)

� Gut

�----

tel�l� �1�R�ig�� h t�v�i� me�ve�i�n

�J

����J

__ __

Figure 12.31 Induction of pdx l gene expression in the gut epithelium. (A) In the chick embryo, pdx f (purple) is expressed in the gut tube and is induced by contact with the aorta and vitelline veins. The regions of pdx 7 gene expression create the dorsal and ventral rudi­ ments of the pancreas. (8) In the mouse embryo, only the right vitelline vein survives, and it contacts the gut endothelium. Pdx l gene expression is seen only on this side, and only one ventral pancreatic bud emerges. (e) In situ hybridization of pdx l mRNA in a section through the region of contact between the blood vessels and the gut tube of a mouse embryo. The regions of pdx 7 expression show as deep blue. (0) Blood vessels (stained red) direct islets (stained green with antibodies to insulin) of chick embryo to differentiate. The nuclei are stained deep blue.

cells secrete VEGF to attract blood ves­ sels, and these vessels surround the developing islet (Figure 12.31 D). Pdxl is exceptionally important i n pancreatic development. This tran­ scription factor el icits budding from the gut epithelium, represses the expression of genes that are character­ istic of other regions of the d igestive tube, maintains the repression of shh, initiates (but does not complete) islet cell d ifferentiation, and is necessary (but not sufficient) for i n s u l i n gene expression (Grapin-Botton et a l . 2 00 1 ). Using an inducible Cre-Lox system to mark the progeny of cells expressing either NGN3 or Pdx l , Gu

and col leagues (2002) demonstrated that the Ngn3-expressing cells are specifica lly islet cell progenitors, whereas the cells expressing Pdxl give rise to a l l three types of pancreat­ ic tissue (exocrine, endocrine, and duct cells) (Figure 12.32; see Chapter 2). Moreover, Horb and col leagues (2003) have shown that Pdx1 can respecify developing liver tissue into pancreas. When Xenopus tadpoles were given a pdx 1 gene attached to a promoter active in liver cells, Pdxl was made in the liver, and the liver was converted into a pancreas con­ taining both exocrine and endocrine (Continued on next page)

CHAPTER 1 2

478

SIDELIGHTS & SPECULATIONS (Cont i n ued)

Exocrine acinar cell

o

o

\ ;

Ptfra, Rpbj/

Dorsal endoderm

@ Foxal' Foxa2 Htxb9' Pltt'a . Activin

..

Pdxl, Ptfra @

0

(Wnt suppressed, Shh suppressed)

s

t

e

, y � polypep. �

� Pancreatic

Mafa, PdXI. Hixb9, Pa:x4, Pax6, lsI!, , 2, J:.::: n Nkx2. Ngn3, Nkx6' l g P NeuroD' Imml ax4 P a:x6 e .. FGFI O

HIlJ6/0C-l Isll,

.. :::; ==-'-----.... ---:;::::.;::-

"- " ' Ventral� �epatobl'st r\n\�e@ p. tocyte endoderm ''';8� � W Wnt notch '191 Hex, Tbx3, -X O Proxl, OC-2, � HIlj6/0C-l

'

Endocrine p ogenitors

Dor al and ventral Pancreatic progenitors

RA

@

Beta (insulin)

Pancreatic duct

r

Pancreatic lineages

Somatostatin

Foxa2, Nkx2.2, Pax6, Arx

AJpha (ghlCagon)

Foxa], Foxa2

c:/

cells. Thus, Pdxl appearS to be the c riti ca factor in distinguishing the liver from the pancreatic mode of development. The growth of the pancreas takes pl ace at the tips of the endodermal buds, Here, the cells are multipotent, giving rise to endocrine cells, exocrine cells, and duct cells (Zhou et al. 2007). However, after a pa rti c u l a r

l

is

and

n

n

n from

in- o

l

r

(Karnik et a1. 2007). I n mice, this lift­ ing of menin regulation promotes a twofold increase in the maternal � cells, al lowing for increased insulin synthesis to meet the metabolic stress of pregna n cy. After giving birth, the mother's p-cell mass returns to pre­ pregnancy levels.

r

though they serve no role in digestion. In the center of the p ha ry ngea l floor, between the fourth pair of pharyngeal p o uc he s, the aryngo tracheal gr oove ex tends ventra lly Figu re 1 2.33). This groove then bifurcates into the branch­ e that form the p aired bronchi and lungs. The laryngotra­ cheal endoderm becomes the lining of the trachea, the two bronchi, and the air sacs (alveoli) of the lungs. Sometimes this sepa ration is no t complete a nd a b aby is born with a connection between the two tubes. This digestive and res­ pira tory condition called a tracheal-esophageal fistula,

s

Lineage of

n

time, these multipotent cells lose the abil ity to produ c e endocrine cell types. There are no adult stem ce l s in the pancreas, but the insulin-secreting � cells do have the capacity to pro­ duce other insulin-producing cell s (Sta nge r et a1. 2007). Indeed, du i n g pregnancy the maternal hormone p rol act i n represses men in, a negative growth egu l a tor of pancreatic islets

The l ungs are a deriva tive of the dig estive tube, even

(

Figure 12.32 pancreatic liver cells. All pa crea tic cells express pdx 1, d isti gu ish i g them from the cells that will become the liver. Within the pancre­ atic li eage, the Ngn3-expressing endocrine proge itor cells give rise to the endocri ne l ineages. (A-C lammert et a1. 2001, courtesy of E. lam mert; D from Gra p B tton et a1. 2001 , courtesy of D. Melton; E after Zare! and Grompe 2008,)

n

@ Cholanglocyte (bile duel cell)

The Respiratory Tube

l

Liver lineages

ll

and must be su rgica y repaired so the b ab y can breathe and swallow proper y. The produ ction of the aryngotrachea groove is correlat­ ed with the appearance of retinoic add .in the ventral meso­ derm, and is p robab y induced by the s ame wave of RA that induces the posterior region of the heart. If RA is blocked, the foregut will not produce the lung bud (Desai et al. 2004). Retinoic acid probab ly induces the formation of FgflO by activating Tbx4 in the sp a nchnic mesoderm adjacent to the ventral foregut (Sa ki yama et a1. 2003). Wn t Signa ing is a so requ red to form the ungs from the digestive tract. Two Wnts-Wnt2 in the earliest stages of u ng deve opment, and the re a ted p rotein Wnt2b-

l

l

l

l

l

l

l l

l

i

l

l

479

LATERAL PLAT E MESODERM A N D THE ENDODERM

A)

(D)

�P h arynX

(e)

dgRlieavosrepyvnitregac)otuotlrruaymcheal

h

FIGURE 12.33

Esophagus

h

Partitioning of the foregut into t e esop agus and -espiratory diverticulum during the third and fourth weeks of human gestation. (A,B) Lateral and ventral views, end of week O Ventral view, week 4. (Aher Langman 1 981 .)

(Al Wild

3.

FIGURE 12.34

type

h

t

i

Wnt signal ng is critical for separation of the trachea and early differen i a­ lung. (A) I n normal mice, the trach ea separates from the gut tube between days and 1 2 . Lungs are visible 3 days later. (B) In mu lant mice that lack hnth Wnl2 and Wnl2b, t e trachea does not separate from the gut tube and no l ungs develop. (C) Mer em bryonic mouse lung epithelium had branched into IwO bronchi, the entire rudiment was excised and cul­ tured. The right bronchus was leh untouched while the tip of t e len bronchus was covered with trach eal mesenchyme. The tip of the right bronchus formed the branches characleristic of the lung, whereas hardly any branching has occurred in the len b ro nchu s. (A,B from Goss et al. 2009, cou rtesy of E . E. orrisey; C from Wessells 1 970, courtesy of Wessells.)

tion of the

11

__

cause the accumulation of �-catenin in the region of the gut tube that will become the lung and trachea. Without these signals, the separation from the gut tube of the trachea and its development into the lungs fail to happen (Figure 12.33A,B; Coss et al. 2009). Conversely, extra lungs can form if �-catenin is expressed ectopically in the gut tube (Har­ ris-Johnson et al. 2009). As in the digestive tube, the regional specificity of the mesenchyme determines the d i fferenti a tion of the devel­ oping respiratory tube. In the developing mammal, the res­ piratory epithelium responds in two distinct fashions. In the region of the neck, it grows straight, forming the tra­ chea. After entering the thorax, it branches, forming the two bronchi and then the lungs. The respiratory epitheli­ um of an embryonic mouse can be isolated soon after it has split into two bronchi, and the two sides can be treated dif­ ferently. Figure 1 2 .34C shows the result when the right bronchial epitheliwn was alloyved to retain its lung mes­ enchyme while the left bronchus was surrounded with tra­ cheal mesenchyme (Wessells 1970). The right bronchus pro­ liferated and branched under the influence of the lung

h

M

(B) Wnt2 and WI1t2b absent

(C)

--

Len

N.

Right

480

CHAPTER 1 2

FIGURE 12.35 The immune system rel ays a s ig nal from the em bryoni c lu ng. Surfac­ tant protein-A (SP-A) activates macrophages in the amnionic fluid to m igrate into the uterine muscles, where the macrophages secrete ILl p. ILl P stimulates production of cyclooxygenase-2, an enzyme that in turn triggers the production of the prostaglandin hormones resp on s i bl e for initiating uterine muscle contractions and birth.

;?\�� " t �;;�"'" .. •

·

•

• ·

:

. i .. . • • • •

•

.

•

•

•

•

•

•

•

'

•

Embryonic lung tissue

mesenchyme, whereas the left bronchus continued to grow in an unbranched manner. Moreover, the differentiation of the respiratory epithelia into trachea cells or lung cells depends on the mesenchyme it encounters (Shannon et aJ. 1998).

See WEBSITE 1 2.2 Induction of the lung The lungs are among the last of the mammalian organs to fully differentiate. The lungs must be able to draw in oxy­ gen at the newborn's first breath. To accomplish this, the alveolar cells secrete a surfactant into the fluid bathing the lungs. This surfactant, consisting of specific proteins and phospholipids such as sphingomyelin and lecithin, is secret­ ed very late in gestation, and it usually reaches phYSiolog­ ically useful levels at about week 34 of human gestation. The surfactant enables the alveolar cells to touch one anoth­ er without sticking together. Thus, infants born premature­ ly often have difficulty breathing and have to be placed on respirators until their suufactant-producing cells mature. MammaHan birth occurs very soon after lung matura­ tion. New evidence suggests that the embryoruc lung may actually signal the mother to start delivery. Condon and colleagues (2004) have shown that surfactant-A-one of the final products produced by the �mbryonic mouse lung-activates macrophages in the arnnioruc fluid. These macrophages migrate from the amnion into the uterine muscle, where they produce immune system proteins such as interleukin 1� (IL1 �). IL1� initiates the.contractions of labor, both by activating cyclooxygenase-2 (which stimu­ lates production of the prostaglandins that contract the uterine m uscle cells) and by antagonizing the progesterone receptor (Figure 1 2.35). Surfactant-stimulated macrophages injected into the uteri of female mice induce early labor.* Thus the Signal for birth may be transmitted to the mother via her immune system.

*ILlP is also produced by macrophages when they attack bacterial infections, which may explain how uterin e infections can cause pre­ mature labor.

Mac rophage mig rates

SP-A sec re ted into amnionic

@ \

'- .

--

@

/

Dormant macrophage Activated macrophage

I

Ammo nlc fluid

� t t

Interleukin-l�

Dt JIIlL

Cyclooxygenase :!

A ra C had ' acid Prostaglandins "'----Uterine mnsdes

f

Uterine muscle contraction, labor

The Extraembryonic Membranes In reptiles, birds, and mammals, embryonic development took a new evolution ary direction-the amniote egg. This remarkable adaptation, which allowed development to take place on dry land, evol ved in the reptile lineage, free­ ing them to explore ruches that were not intimately linked to water. This evolutionary adaptation is so significant and characteristic that reptiles, birds, and mammals are grouped together as the amniote vertebrates, or amruotes. To cope with the challenges of terrestrial development, the anutiote embryo produces four sets of extraembryonic membranes to mediate between it and the environment. The evolution of internal development and of the placen­ ta displaced the hard-shelled egg in mammals, but the basic pattern of extraembryonic membranes remains the same. In developing anmiotes, there initi all y is no distinction between the embryoniC and extraembryonic domains. However, as the body of the embryo takes shape, the epithelia at the border between the embryo and the extraembryoruc domain divide unequally to create body folds that isolate the embryo from the yolk and delineate which areas are to be embryonic and which extraemb ry­ onic (MiUer et aL 1994, 1999). These membranous folds are formed by the extension of ectodermal and endodermal epithelium underlain with lateral plate mesoderm. The combination of ectoderm and mesoderm, often referred to as the soma topleuue (see Figure 12.4A), forms the amnion and chorion; the combination of endoderm and meso­ derm-the splanchnoplenre--forms the yolk sac and aUan­ tois. The endodermal or ectodermal tissue supplies func­ tioning epithehal ceUs, and the mesoderm generates the essential blood supply to and from the epithelium. The for­ mation of these folds can be followed in Figure 12.36.

LATERAL PLATE MESODERM A N D T H E E N DODERM

(A) foldofamnion ofe r o coelom ea Embryo � Somatic mesoderm e e Splanchnic me:sorjerm Vitel ine enveYlooplke ni NeNouroal r a eEndnohderme :;�� Splancohfnopleure Al l a n t o i s p u s h i n g i n t o e x r e r y evagination oel oni Gu t AmnGuiont Amnioi n i Chorion ---�

Head

Head

Ectoderm

mb y

\

Extraembryonic

H d fold

( B)

coelom

Somatic m wrl nn

� _ _ _

481

__

Tail fold of amnion

----

tube -=�;=����ilii

(C)

Anlnionic cavity

Ectoderm

,Ch,orioa.ll

t cho d ----

°n,

ic cavi ty

Aor t

Mes

c ym

Mesoderm

yolk sac

--��:/" " "

Proctodeu111

c

(D)

Embryo ------,

Allantoic

t a mb om

c

(E)

Embryo -----,

_ _ _ _

----�

---�

-----

Amnionic ---1 cav ty

cavity

Chorion

lo u l n

l

membrane

FIGURE 12.36 Schematic drawings of the extraembryo ic mem­ branes of the chick. The embryo is cut ngit d ina y, and the albu­ men and shell coatings are nol shown. (A) 2·day embryo. (8) 3-day

embryo. (e) Detailed schematic d iagram of the ca uda (hind) region of the chick embryo, showing the formation of the allantois. (D) 5-day embryo. IE) 9-day embryo. IAher Carlson 1 981 .)

The amnion and chorion

The second problem of a terrestrial egg is gas exchange. This exchange is provided for by the chorion, the outer­ most exh

Aromatase

The human element Extrapolating from rodents to humans is a very risky business. No sex-specific behavior has yet been identified in humans, humans do not use a-fetopro­ tein to bind circulating estradiol, and humans do not use pheromones as a primary sexual attractant (sight and touch being far more critical). No "gay

. �

:

� :

l

Arachidonic acid

E s tr e og n --- ' ' C OX2 '_ _ "...__ _ I _ ceptor ; . L,:-- _,

The role of testosterone If testosterone's conversion to estrogen is involved in brain masculinization, then what exactly is going on? It appears that many of testosterone's orders are carried out by prostaglandin E2 (PGE2). PGE2 i s made from arachidonic acid b y the enzyme cyclooxygenase-2 (COX2); COX2 is induced by estrogen in the brain. So estrogen acts to produce more PGE2 (Figure 1 4.1SA). Studies by Amateau and McCarthy (20041 demonstrated that PGE2 is stim­ ulated by estradiol in the newborn rat brain, and that PGE2 induces the growth and differentiation of those regions of the brain involved in male sexual behaviors. They found that PGE2 was as effective as estrogen in mas­ culinizing these brain regions, and that PGE2 induces male-specific morpholo­ gy and behaviors when injected into female rat brains. Moreover, when male rats were given COX2 inhibitors, their brains became similar to those of females. Not only did their brain anato­ my change, but so did their behaviors. The PGE2-treated females acted like males, attempting to mount and copu­ late with other females, while male mice treated with COX2 inhibitors lost their sexual drive and did not exhibit male behaviors (Figure 14.1 S B). Studies comparing the structure of brain regions responding to pheromones in mice treated with PGE2 or estrogens indicate that the processing of pheromone signals is a sexually dimor­ phic phenomenon regulated by early exposure to PGE2 and estradiol.

Estradiol

COX2 gene

PGE2

!

Masculinization of rat brain regions

(B)

18

Mounts

Intromissions

16

G 12 14

� JO

isJ:

8 6

Figure 14.15 Masculinization of the brain by hormo nes. (A) Testosterone mediates its effect on the rodent brain by being converted to estradiol by aromatase. Estradiol adivates the gene for cyclooxygenase, which converts arachidonic acid to PGE2. (B) PGEl promotes male behavior, and its inhibition by COX2 inhibitors prevents that behavior. Males treated wilh indomethacin (COX2 inh ibitor) did not display the mounting behavior expected of males, nor did Ihey attempt intercourse (P < 0.001 between control and treated males). . Conversely, females given PGEl acted li ke control males in both cases and had sign ificantly more mounting and mating behavior than control < 0.001) . (After Amateau and McCarthy 2004.)

females (P

gene" has been discovered, and the concordance of gender identity between identical twins is only 30°/0far from the 1 00% expected if sexual orientation were strictly genetic (Bailey et al. 2000; CRC 2006). Moreover, behaviors that are seen as "masculine" i n one culture may be considered "femi n i ne" i n another, and vice versa

(see Jacklin 1 981 ; B l eier 1 984; Fausto­ Sterling 1 992; Kandel et a l . 1 995). How humans acquire their gendered behavior appears to involve a remark­ ably complex set of interactions between genes and environment. Like many other behaviora I phenotypes, how individuals acquire sexual behav­ iors has yet to be determined.

SEX DETERMINATION

529

determining, each developing according to its own aspi­

Chromosomal Sex Determination in Drosophila

ration/' and each sexual decision is "not interfered with

Although both mammals and fruit flies produce XX

the action of the gonads."

by the asp i rations of its neighbors, nor is it overruled by

females and XY males, their chromosomes achieve these

Although there are organs that are exceptions to this

ends using very different means. In mammals, the Y dU"o­

rule (notably the external genitalia), it remai11s a good gen­

mosome plays a pivotal role in determining the male sex.

eral principle of Drosophila sexual development. Moreover,

Thus, XO mammals are females, with ovaries, a uterus,

molecular data show that differential RNA spUci11g to cre­

and ovid ucts (but usually very few, if any, ova).

ate male- and female-specific transcription factors is criti­

In

Drosophila,

the Y chromosome is not involved in

cal in determining fly sex and can explain how XX indi­

determining sex. A fruit fly's sex is determined predomi­

viduals become females and XY i11dividuals become males.

nantly by the number of X chromosomes in each cell. If there is only one X chromosome in a diploid cell, the fly is male. If there are two X chromosomes in a diploid cell, the

The Sex-lethal gene

fly is female. Should a fly have two X chromosomes and

Although it had long been though t that a fruit fly's sex was

three sets of autosomes, it is a mosaic, where some of the

determined by the X-to-autosome (X:A) ratio (Bridges

cells are male and some of the ceUs are female. In flies, the

1925), this assessment was based largely on the fact that

Y chromosome appears to be important in sperm cell dif­

flies have aberrant numbers. of chromosomes. Recent

ferentiation, b u t it plays no role in sex de termination.

molecular analyses suggest that X chromosome number

Rather, it seems to be a collection of genes that are active

alone is the primary sex determinant in normal diplOid

in forming sperm in adults. Thus, XO

Drosophila are ster­

In

i11sects (Erickson and Quintero 2007). The ma in basis for this assertion is the fact that the X chromosome contains

ile males.

Drosophila, and in insects in general, one can observe

gynandromorphs-anlinals i11 which certain regions of the

genes encoding transcription factors that activate the crit­

Sex-lethal (Sxl).

ical gene in

Drosophila sex determination� the au tosomal

body are male and other regions are female (Figure 1 4.1 6;

locus

14.13). Gynandromorph fruit flies result X chromosome is lost from one embryonic nucle­ us. The cells descended from that cell, instead of being XX

ACTIVATING SEX-LETHAL The Sxl gene has two promoters.

see also Figure

when an

(female), are XO (male). The XO cells display male charac­ teristics, whereas the XX cells display female traits, sug­

gesting that,

in Drosophila, eC'lch

cell makes its own sexual

"decision." Indeed, in their classic discussion of gynandro­

The early promoter is active only in XX cells; the later one is

active in both XX and XY cells. The X dU"omosome appears

of Sxl. Three of these proteins are transcription fac­

to encode four protein factors that activate the early pro­

moter

tors-SisA, Scute, and Runt-which bind to the early pro­

morphs, Morgan and Bridges (1919) concluded, "Male and

moter to activa te transcription ( Figure 1 4.1 7). The fourth

female parts and their sex-linked characters are strictly self-

protein, Unpaired, is a secreted factor that reinforces the

(AJ

(B)

miniature

o

xo

Gynandromorph i n sects. (A) D. melanogaster in which the left side is femal e

(XX) and the right side is male (XO). The male side has lost an X chromosome heari ng the FIGURE 1 4.16

wild-type alleles of eye co lo r Jnd wing shape, thereby a llowing expression of the recessive alleles eosin eye and miniature wing on the remaining X chromosome. (8) Tige r swallowtail butlerfly Paplio glaucus. The len half (ye ll ow) is male, while Ihe right half (blue-black) is female. (A, drawing by Edith Wallace from Morgan and Bridges 1 9 1 9; B, photograph cour­ tesy of J. Adams ©

2005.)

530

CHAPTER 1 4

xx

XY

High concentrations of SisA, Scute, Runt, and Unpaired

Low concentrations of SisA Scnre, Rnnt, and Unpaired

+ Is;d'I

,

'. '.

��l protein) \ crra: � :J (Fllnctional

(Functional Tra protein)

fruitless

[];!ill t

� . 1 ��-,-----,

Female

Male transcription rate

Male phenotype

to

(XX)

repress

X

production.

X

�

Male differentiation genes

phenotype

autosomes

r-;;;;il �

Male -specific Dsxand Fruitless proteins

FIGURE 14.17 Proposed regulatory cascade for Drosophila somatic sex determination. Transcription factors from the chro­ and compete activate or the Sx/ gene, which becomes active in femal es and i nactive in males The Sex-lethal protein performs three main functions. First, it activates its own transcription, ensuring further Sxl Second, it represses the translation of msl2 mRNA, a factor that facilitates transcription from the chromosome. This equalizes the amount of transcription from the tvvo chromosomes in

mosomes (XY).

(No functional

Tra protein)

1

Female differentiation genes

( No functional Sxl prorein)

dsx

Female­ specific Dsx proteins

Female transcription rate

Sxl not activated

X

other three proteins tl1IOUgh the JAK-STAT pathway (Sefton

et al.

females with that of the single chromosome in males. Third, Sxl activates the transformer (tra) genes. The Tra proteins process dou­ bfesex pre-mRNA i n a female specific manner that provides most of the female body with its sexual fate. They process the fruit­ less pre-mRNA i n a female-specific manner, giving the fly female­ specific behavior. In the absence of Sxl (and thus the Tra proteins), dsx and fruitless pre-mRNAs are processed in the male-specific manner. (After Baker el a!. 1 987.)

X

-

also

In males" the early promoter of Sxi i s not active. How­

2000; Avila and Erickson 2007). If these factors accu­ mulate so th ey are present in amounts above a certain threshold, the Sx/ gene is activated through its early pro­ moter (Erickson and Quintero 2007; Gonzales et al. 2008). 11,e result is the transcription of Sx/ early in XX embryos. The Sxl pre-RNA transcribed from the early promoter lacks exon 3" which contains a stop codon. Thus, Sxl pro­

ever, later in development the late promoter becomes

tein that is made early is spliced in a manner such that

manner that yields eight exons, and the termination codon

exon 3 is absent and a complete protein can be made. This

is in exon

functional Sxl binds to its own late promoter to keep the

Sxl gene active.

active, and the Sxi gene is transcribed in both males and

females In XX cells, Sxl protein made from the early pro­ .

moter can bind to its own pre-mRNA and splice it in a "female" direction, producing more functional Sxl (Bell et al.

1988; Keyes et al. 1992). In )('y' cells, however, there is no

early Sxl protein, and the "male" pre-lORNA is spliced in a

3.

Protein syntb.esis thus ends at the third exon,

and the protein is nonfunctional. In femal es, RNA process­

ing yields seven exons, with the mal.e-speciiic exon 3 being

SEX DETERM I NATION

Female mRNA

Male or nonspecific mRNA

Pre-mRNA

me specific

531

e c ng sp spUci n g � I I l'I 'H'I6171 I- AAA I I -'---"---i� I.L 17L I.l: 16L I1.:: 15L II I' I 15\6171 f- AAA � ---,-, ' 1L-I-,. f-'----=-- I I I 1L---L I'I 1'1 L....- I Stop codonn Mal e -speci fi c ex l I No protei Sex-letha!

� Fe al -

'

•

.

I

'..Jl-AAA LCJ _"-------'-- I -,-

I I

2

, I ' I-AAA

Female

codonn NoStopprotei

4-1

transformer

"

' ::.J

D fa u lt''' li i

�---I-

doublesex

[iT-T I--..c ''I--' 1"-CI T " I-'. .. . .--. . ·_ I-;

I .fic exon Female-speci

f-AAA - liLllzl31

fruitless

1'11'1 1611711'1 19 1 I 10 II 1 1

FIGURE 14.18 Sex-specific RNA splicing in four major Drosophila sex-determining genes. The pre-mRNAs (shown i n the center of diagram) are identical in both male and female nuclei. In each case, the female-specific transcript is shown at the left, while the default transcript (whether male or nonspecific) is shown to the right. Exons are numbered, and the positions of ter­ mination codons are marked. Sex-lethal, transformer, and double­ sex are all part of the genetic cascade of primary sex determina­ tion. The transcription pattern of fruitless determines the secondary characteristic of courtship behavior (see pp. 533-534). (After Baker 1 989; Baker et al. 2001.)

spliced out as part of a large intron. Thus, the female-spe­ cific mRNA lacks the early-termination codon and encodes a protein of 354 amino acids, whereas the male-specific Sxl transcript contains a translation termination codon (UGA) after amino acid 48 ( Fig ure 1 4.18, top). TARGETS OF SEX-LETHAL The protein made by the female­ specific Sxl transcript contains regions that are important for binding to RNA TIlere appear to be three major RNA targets to which the female-specific Sxl transcript binds. One of these, as mentioned above, is the pre-mRNA of Sxl itself. Another target is the msl2 gene that controls dosage compensation (see below). Indeed, If the Sxl gene is non­ functional in a cell wi th two X chromosomes, the dosage compensation system will not work, and the result is cell death (hence the gene's macabre name). The third target is

B

codon L...StoNo... l pprotein

r-c-,------n ll r;, --J --c;, -,------, � -; ,--, _ I

�

Doublesex

ItI'II'I'1516171'191 1 0 1 11

Sex-lethal

I.,

I I

'

I

,

I'.

'

I- AAA

A AAA

Male Doublesex Male Fruitless

the pre-mRNA of transformer (tra)-the next gene in the cascade (BeU et at 1988; Nagoshi et al. 1988). The pre -mRNA of transformer (so named because 1055of-functi on mutations turn females into males) is spliced into a functional mRNA by Sx.l protein. The Ira pre-InRNA is made in both male and female cells; however, in the pres­ ence of Sx.l protein, the Ira transcript is alternatively spliced to create a female-specific mRNA, as well as a nonspecif­ ic InRNA that is found in both females and males. Like the male Sxf message, the nonspecific tra mRNA contains a ter­ mination codon early in the message that renders the pro­ tein nonfunctional (Boggs et al. 1987). In Ira, the second exon of the nonspecific mRNA contains the termination codon and is not utilized in the female-specific message (see Figures 14.17 and 14.18). How is it that females make a different transcript than males' The female-specific Sx.l protein activates a 3' splice site that causes Ira pre-mRNA to be processed in a way that splices out the second exon. To do this, Sx.l protein blocks the binding of splicing factor U2AF to the nonspecific splice site of the Ira message by specifically binding to the polypyrimidine tract adjacent to it (Figure 1 4.1 9; Handa et aJ. 1999). This causes U2AF to bind to the lower-affinity (female-specific) 3' splice site and generate a female-spe­ cific mRNA (Valcarcel et aJ. 1993). The female-specific Tra protein works in concert with the product of the tral1S­ fonner-2 (1m2) gene to help generate the female phenotype by splicing the doublesex gene in a female-specific manner

532

CHAPTER 1 4

teins are both present, the dsx transcr ip t is processed in a female-specific manner (Ryner and Baker 1991). The female splicing pa ttern produces a fe male spec ific protein that activates female-specific genes (such as those of the yolk proteins) and iullibits male development. If no funchonal Ira is produced, a ma le-specific transcript of dsx is made. The male transcript encodes an active transcription factor -

that inhibits female traits and

promotes male traits. In the Dsx regulates all known aspects o f sex­ u ally dimorphic gonad cell fate (Figure 1 4.20). In XX flies, the female Doublesex protei n (DsxF) com­ bines with the product of the illtersex gene (Ix) to make a transcrip tion factor complex that is responsible for promot­ ing female-specific traits. This "Doublesex complex" acti­ vates the Wingless gene (Wg), whose Wnt-farruly product embrYOnic gonad,

promotes growtll of the female portions of the genital disc. It also represses the Fgf genes responSible for making male accessory organs, activates the genes responsible for mak­ ing yolk proteins, promotes the growth of the sperm stor­ age duct, and modifies b"icabrac (bab) gene expression to give the female-specific pigmentation profile. In contrast, the male Doublesex protein (DsxM) acts directly as a tran­ scription factor and directs the expression of male-specific traits. It causes the male region of the genital disc to grow at the expense of the female disc regions. It activates the BMP homologue Decapentnplegic (Dpp), as well as stimu­ lating Fgf genes to produce the male genital disc and acces sory structures. DsxM also converts ce rtclin cuticular struc­ tures into claspers and modifies the bricnbrac gene to produce the male pigmenta hon pattern (Ahmad and Baker 2002; Christiansen et a!. 2002). According to this model, the result of the sex de term i nation cascade summarized in Figure 14.17 conles down to the type of mRNA processed from the doublesex tran-

FIGURE 14.19 Stereogram showing b inding of (ra pre-mRNA by the clef! of the Sxl protein. The bound 1 2-nucleotide RNA (GUUGUUUUUUUU) is shown in yellow. The strongly positive regions are shown in blue, while the scattered negative regions are in red. II is worth crossing your eyes to get the three�dimen­ sional effect. (From Handa et al. 1 999, courtesy of S. Yokoyama.)

Doublesex: The switch gene for sex determination

­

The dallblesex (dsx) gene is active in both males and females, but its primary transcript is processed in a sex­ specific manner (Baker et al. 1987). This alterna tiv e RNA processing is the result of the action of the Ira and trn2 gene products on the dsx gene (see Figures 14.17 and 14.18; see also Figure 2.28). I f the Tra2 and female-specific Ira pro-

(A)

+ Ix

(B)

FEMALE DsxF

Wg, Dpp

t

(NP organizer)

FGF

Femall!-specific growth genital disc

Yp genes

l

­

MALE DsxM

dac

�

FGF

t

[PJ�

(NP o rgani zer)

Female

Mal e-specificl

l

pigmentation

FIGURE 1 4.20 Roles of DsxM and Dsxf proteins in Drosophila sexual development. (A) Dsxf functions with Intersex (Ix) to pro­ mote fcma Ie-specific expression of those genes that control the growth of the genital disc, the synthesis of yolk proteins, the for­ mation of spermathecal ducts (which keep sperm stored c1fter mat4 ing), and pigment patterning. (B) Conversel y, DsxM acts as J tran4 scription factor to promote the male-specific growth of the genital

Yp genes

Wg, Dpp

bab

grovvth genital disc

l

riac

l

ICl asper I

bab

l

disc, the formation of male genitalia, the conversion of cuticle into claspers, and the male-specific pigm entation pattern. In addi­ tion, DsxF represses certain genes i nvolved in specifying male­ specific traits (such as the paragonia), and DsxM represses certain genes involved in synthesizing female-specific proteins such as yolk protein. (After Christiansen et al. 2002.)

SEX DETERM I NATION

533

(Il)

(A)

FIGURE 14.21 Subsets of neurons·expressing the mal e·spec ifi c form of iruitiess. (A) Central nervous system of a 48-hour pupa script. If there are two X chromosomes, the transcription

factors activating the early promoter of Sxl reach a critical concentratlon, and the

Sxi makes a splicing factor that causes transformer gene transcript to be spliced in a female­

specific manner. This female-specific protein interacts with the 1m2 splicing factor, causing dsx pre-rnRNA to be spliced in a female-specific manner. If the dsx transcript is not acted

(the time at which sex determination of the brain is establ ished), showing m al e-speci fi c rruifJe.�5 mRNA in particular cells (pink) of the anterior brain that are involved in cO LJ rt� h ip behaviors. (B) Some p er ip heral nervous system ce(ls express the m2l(e-sp ecific fruitless message. Fruitless protein is stained green in these taste neurons i n the labella (the spongy pari of the fly mouth). (From Stockinger et al. 2005, courtesy of B. J. D ic kson. )

on in this way, it is processed in a "default" manner to

make the male-specific message.

Brain Sex in Drosophila OUf discussion o f sexual dimorphism i n Drosophila has so far been limited to nonbehavioral aspects of development.

However, as in mammals, there appears to be a separate

ments that include following the female, tapping the

female, playing a species-specific courtship song by vibrat­

ing his wings, licking the female, and finally, curling his abdomen so that he is in a position to mate. Each of these sex-specific courtship behaviors appear to be regulated by

the products offruitless, a gene expressed in certain sets of

"brain-sex" pathway in Drosophila that provides individ­

neurons involved with male sexual behaviors (Figure

aggression behaviors. Among Drosophila, there are no par­

hearing, smell, and touel), and in total they represent about 2% of all the neurons in the adult male (Lee et aJ. 2000; Bil­

uals of each sex with the appropriate set of courtship and

ents or other conspecifics to teach "proper" mating behav­

1 4.21) . These include subsets of neurons involved in taste,

ior, and mating takes place soon after the flies emerge from

leter and Goodwin 2004; Stockinger et a\. 2005). Fruitless

into the n i sect genome!

rons in these circuits die during female development (and

their pupal cases. So the behaviors must be "hard-wired" Among male

Drosophila, there is one very simple rule

of behavior: If the other fly is male, fight it; if the other fly

is female, court it (Certel et aJ. 2007; Billeter et aJ. 2008; Wang and Anderson 2010). Although the outcomes of this

algorithm are simple, the behaviors of Drosophila courtship and mating are quite complicated. A courting male must

also retains certain male-specific neural circuits; the neu­

in frUitless mutants; see Kimura et aJ.

2005).

As with doublesex pre-rnRNA, the Tra and Tra2 proteins

splice fruitless pre-mRNA into a female-specific message; the default splicing pattern is male (see Figure 14.18). So

the female makes Tra protein and processes thefmit/ess

pre-mRNA in one way, whereas the male, lacking the Tra

first confirm that the individual he is approaching is a

protein, processes the fruitless message in another way.

toward the female and follow a specific series of move-

sequence in an early exon; therefore the female does not

female. Once this is established, he must orient his body

*This is not to say thflt flies don't leflrn; indeed, one thing they do learn is to avoid bad sexual encounters. A male who has been brushed off (quite literally) by a female because she has recently mated hesitates before starting to court another female (Siegel and Hall 1979; MacBride et al. l999).

However, female fruitless mRNA includes a termination

make functional Fruitless protein. The male, however, makes an mRNA that does not contain the stop codon (Heinrichs et aJ. 1998), and the protein it transcribes is a

zinc-finger transcription factor. Using homologous recom­

bination to force the transcription of particular splicing

forms, Demir and Dickson (2005) showed that it is Fruit-

534

CHAPTER '1 4

less, and not the flies' anatomy, that controls their sexual behavior. When female flies were induced to make the male-specific Fruitless protein, they performed the entire male courtship ritual and tried to mate with other females. In normal females, the courtship ritual is not as involved as in males. However, females have the ability to be recep­ tive to a male's entreaties or to rebuff them. The product of the retained gene (rln) is critical in this female mating behavior. Both sexes express this gene, since it is also i.nvolved in axon pathfinding. However, female flies with a loss-of-function allele of rln resist male courtship and are thus rend ered sterile by tlleir own behavior (Ditch et aJ.

histone 4 (see Figure 2.4). In this way, transcription factors gain access to the X chromosome at a much higher frequen­ cy in males than in females-hence, "hyper transcription. "

2005).

While the sex of most snakes and lizards is determined by sex chromosomes at the time of fertilization, the sex of most turtles anct all species of crocodilians is determined after fertilization, by the embrYOnic environment. In these rep­ tiles, the temperature of the eggs during a certain period of development is the deciding factor in determining sex, and small changes in temperature can cause dramatic

The splicing of the/milless transcripts not only regulates sex-specific courtship patterns, it also regulates sex-specif­ ic aggression patterns as well. Female flies having a male Fruitless protein not only tend to court females, they also will fight males and try to establish themselves at the top of a dominance hierarchy. Male flies having a mutant/ruit­ less allele will show female-specific aggression against other females (Vronto u et aJ. 2006).

Dosage Compensation

In animals whose sex is determined by sex chromosomes, there has to be some mechanism by which the amount of X chromosome gene expression is equalized for males and females. This mechanism is known as dosage compensa­ tion. In Cha pter 2 we discu sse d mammalian X-chromo­ some inactivation, whereby one of the X chro mosomes is inactivated so that the transcrip tion product level is the same in both XX cells and XY cells. In the worm Cae/lorhab­ ditis elegans, dosage compensation occurs by lowering the transcription rates of both X chromosomes so that product levels are the same as those of XO individuals. In Drosophila, the female X chromosome s are not sup­ pressed; rather, the male's single X chromosome is hyper­ activated. This "hypertranscription" is accomplished at the level of translation, and it is mediated by the Sxl protein . Sxl protein (which you will recall is made by the female cells) binds to the 5' leader sequence arid the 3' untranslat­ ed regions (UTRs) of the mls2 message. The bound Sxl inhibits the attachment of msl2 mRNA to the ribosome and prevents the ribosome from getting to the mRNNs coding region (Beckman et aJ. 2005). The result is that female cells do not produce Ms12 protein (see Figure 14.17). However, Ms12 is made in male cells, in which Sxl is not present. Mls2 is part of a protein-mRNA complex that targets the X chro­ mosome and loosens its chromatin structure by acetylating

ENVIRONMENTAL SEX DETERMINATION

Temperature-Dependent Sex Determination in Reptiles

changes in the sex ratio (Bull 1980; Crews 2003). Often, eggs incubated at low temperatures produce one sex, whereas eggs incubated at higher temperatures produce the other. There is only a small range of temperatures that p ermits both males and females to hatch from the same brood of eggs. ' Figure 14.22 sho ws the abrupt temp e rature-induced ch ange in sex ra tios for the red-eared slider turtle. If a brood of eggs is incubated at a temperature below 28"C, all the turtles hatching from the eggs will be male. Above 31"C, every egg gives rise to a female. At temperatures in between, the brood will give rise to individuals of both

sexes. Variations on this theme also exist. The eggs of the *The evolutionary advantages and disadvantages of temperature­ dependent sex determination are discussed in Chapter 19.

O? "iii

•

100

0--0

Trachemys scripta

6 75 '*' "--

.g �

Alligator mississippiensis

50

h

X

'" FIGURE 1 4. 22 Temperature-dependent sex determination in

three species of reptiles: the American all igator (Alligator missis­ sippiensis), red-eared slider turtle (Trachemys scripta elegans), and alli gator snapping turtle (Macroclemys temmincki/). (After Crain and Guillette 1 998.)

\

0

u

25

0

0 22

24

26

28

30

Temperature (OC)

32

34

\ 36

SEX DETERMI NATION

535

snapping turtle Macroclemys, for instance, become female at either cool (22°C or lower) or hot (28°C or above) tem­

of Emys is very low at the male-promoting temperature of

As we will see in Chapter 18, there can be multiple path­

period for sex determination (Desvages et al. 1993; Pieau

normal temperature conditions, the sex of the lizard

also seen in diamondback terrapins, and its inhibition mas­

peratures. Betvveen these extremes, males predominate.

ways of sex determination in the same individuaL Under

Bassiana duperreyi is determined by sex chromosomes (XY

males; XX females). However, at low temperatures, the

environmental component overrides the genetic sex-deter­

25°C. At the female-promoting temperature of 30°C, aro­

matase activity increases dramatically during the critical

et al. 1994). Temperature-dependent aromatase activity is

culinizes their gonads (jeyasuria et al. 1994). Aromatase appears to be involved in the temperature-dependent dif­

ferentiation of lizards and salamanders as well (Sakata et

mining mechanism, and all the offspring in cool nests are

a1. 2005). It is possible that aromatase expression is activat­

2008).

matase protein itself may be temperature sensitive. In other

male (even if their chromosomes are XX; Radder et al.

One of the best-studied reptiles is the European pond

turtle,

Emys

Emys obicularis.

In laboratory studies, incubating

eggs at temperatures above 30°C produces all

females, while temperatures below 25°C produce all-male

broods. The threshold temperature (at which the sex ratio

ed differently in different species. In some species, the aro­

species, the expression of the aromatase gene may be dif­ ferentially activated at high temperatures (see Murdock

and Wibbels 2006; Shoemaker et aJ. 2007).

When turtle gonads are taken out of the embryos and

placed in culture, they become testes or ovaries depend­

is even) is 28.5°C (Pieau et al. 1994). The developmental

ing on the temperature at which they are incubated

discovered by n i cubating eggs at the male-producing tem­

shows that sex determination is a local activity of the

"window" during which sex determination occurs can be

perature for a certain amount of time and then shifting them to an incubator at the female-prod ucing temperature

(and vice versa). 1n Emys, the middle third of development

appears to be the most critical for sex determination, and

it is believed that the turtles cannot reverse their sex after this period.

(Moreno-Mendoza et al. 2001; Porter et al. 2005). This gonadal primordia and not a global activity directed by

the pituitary or the brain. By analyzing the timing and reg­

ulation of sex-determining genes in such cultured gonads,

Shoemaker and colleagues (2005, 2007) proposed that

female temperatures not only activate aromatase (which would cause ovary formation), but also activate the

Wnt4

gene (which suppresses the formation of testes in mam­

The aromatase hypothesis for environmental sex determination The enzyme aromatase, which converts testosterone into

estrogen (see Sidelights & Speculations, p. 528), appears to

be a particularly important target for environmental trig­

gers. Unlike the situation in mammals, whose primary sex

mals). Conversely, male temperatures appear to activate

Sox9 to a higher level in the male than in the female, pro­

moting testis development (Moreno-Mendoza et al. 2001).

Estrogens, aromatase, sex reversal, and conservation biology

primary sex determination in reptiles and birds is inilu­

Over the last two decades, data has emerged showing that several of the polychlorinated biphenyl compounds

to develop. In reptiles, estrogen can override temperature, inducing ovarian differentiation even at masculinizing tem­

Bergeron et aJ. 1994, 1999). PCBs can reverse the sex of tur­

inhibitors of estrogen synthesis produces male offspring,

important consequences in environmental conservation

determination is a function of the X and Y chromosomes,

enced by hormones, and esh"ogen is essential if ovaries are

peratures. Similarly, experimentally e?Cposing eggs to

even if the eggs are incubated at temperatures that usual­

(PCBs), a class of widespread pollutants introduced into the environment by humans, can act as estrogens (e.g., see tles raised at "male" temperatures. This knowledge has

efforts to protect endangered species (such as turtles,

1994). The sensitive time for the effects of .estrogens and

amphibians, and crocodiles) in which hormones can effect changes in primary sex deternUnation. Indeed, some reptile

nation usually occurs (Bull et al. 1988; Gutzke and Chymiy

ments to elevate the percentage of females

ly produce females (Dorizzi et al. 1 994; Rhen and Lang their inhibitors coincides with the time when sex determi­

conservation biologists advocate using hormonal treat­ n i

endangered

1988).

species (www.reptileconservation.org).

ments mentioned above worked by blocking aromatase

from herbicides that promote or destroy estrogens. One

yield male offspring ' This correlation appears to hold

demasculinized frogs after exposure to extremely low

The estrogen-synthesis inhibitors used in the experi­

action, showing that experimentally low aromatase levels

under natural conditions as well. The aromatase activity

1994).

*One remarkable finding is that the injection of an aromatase

inhibitor into the eggs of an all-female, parthenogenelic species of lizards causes the fonnation of males (Wibbels and Crews

The survival of some amphibian species may be at risk

such case involves the development of hermaphroditic and

doses of the weed killer atrazine, the most widely used her­

bicide in the world ( Figure 14.23). Hayes and colleagues

(2002a) found that exposing tadpoles to atrazine concen­ trations as low as 0 . 1 part per billion (ppb) produced gonadal and other sexual anomalies in male frogs. At 0.1

ppb and higher, many male tadpoles developed ovaries in

536

(A)

CHAPTER 1 4

(B)

FIGURE 14.23 Demasculinization of frogs by low amounts of atrazine, (A) Testis of a frog from a natural site having 0.5 parts per billion (ppb) atrazine. The testis con­ tains three lobules that are developing both sperm and an oocyte. (8) Two testes of a frog from a natural site contain­ ing 0.8 ppb atrazine. These organs show severe testicular dysgenesis, which characterized 28% of the frogs found at that site. (C) Effect of a 46-day exposure to 25 ppb atrazine on plasma testosterone levels in sexually mature male Xenopus. Levels in control males were some tenfold higher than i n control females; atrazine-treated males had plasma testosterone levels at or below those of control females. (A,B aher Hayes et al. 2003, photographs courtesy ofT. Hayes; aher Hayes et al. 2oo2a.)

----r--,

.-:-

5

(C)

�

C

E

S 3 '"

v � 2

*

�

8

addition to testes. At 1 ppb atrazine, the vocal sacs (which a male frog must have in order to signal and

obtain a

potential mate) failed to develop properly. Atrazme induces aromatase, and aromatase can con­ 2007). In laboratory experiments, the testosterone levels

90% (to levels of con­

trol females) by 46 days of exposure to 25 ppb atrazine­

U.s. drinking water is 3 ppb, and atrazine

an ecologically relevant dose, since the allowable amount of atrazine in

levels can reach 224 ppb in streams of the midwestern United States (Battaglin et al. 2000; Barbash

et aJ.

2001).

Given the amount of atrazine in the water supply and the sensitivity of frogs to this compound, the situation

0 Control Atrazinemale ees females males

vert testosterone into estrogen (Crain et a1.1997; Fan et aJ. of adult male frogs were reduced by

2

tr

at

d

Control

location-Dependent Sex Determination Environmental factors other than temperature can also be sex determinants in some species. For example, it has been known since the nineteenth century that the sex of the echi­ uroid worm

Bonellia viridis depends on where a larva set­ Bonellia larva lands on the ocean

tles (Baltzer 1914). If a

could be devastating to wild populations. In a field study,

floor, it develops into a lO-cm-Iong female. If the larva is

Hayes and his colleagues collected leop a rd frogs and

attracted to a female's proboscis, it travels along the tube

water at eight sites across the central United States

until it enters the female's body. There it differentiates into a minute (1-3 rom long) male that is essentially a sperm­

(Hayes et al. 2002b, 2003). They sent the w a ter sa mples to two separate laboratories for the determination of atrazine, coding the frog specimens so that the techni­ cians dissecting the gonads did not know which site the animals came from. The results shoyved that all but one site contained atrazine-and this was the only site from which the frogs had no gonadal abnormalities. At con­

producing symbiont of the female (Figure 14.25).

Another species in which sex determination is affected by the location of the organism is the slipper snail Crepidu­

Ia fornicata. In this species, individuals pile up on top of

one another in a great mound. Young individuals are

always male, but this phase is followed by the degenera­

centrations as low as 0 . 1 ppb, leopard frogs displayed

tion of the male reproductive system and

testicular dysgenesis (stunted growth of the testes) or

ity. The next phase can be

conversion to ovaries. found

in the testes

In many examples, oocytes were

(see Figure 14.23A).

a period of labil­

either male or female, depend­

ing on the animal's position in the mound. If the snail is attached to a female, it will become male. If such a snail is

Concern over atrazine's apparent ability to disrupt sex

removed from its attachment, it will become female. Sim­

hormones in both wildlife and humans has resulted in bans

ilarly, the presence of large numbers of males causes some

way; Sweden, and Switzerland (Dalton 2002). Many geo­

ual becomes female, i t will not revert to being male

graphical and social concerns mediate the amount of

1936; Collin 1995; Warner et aJ. 1996).

on the use of this herbicide by France, Germany, Italy; Nor­

of the males to become females. However, once an individ­

(Cae

atrazine use ( Figure 1 4.24), and the company making

Many fish change sex based on social interactions; such

atrazine has lobbied against the work of independent

changes are mediated by the neuroendocrine system (God­

researchers whose research indicates that it might cause

win et aJ. 2003, 2009). Interestingly, altllOugh ti,e trigger of

reproductive malfunctions or cancers in wildlife and humans (see Blumenstyk 2003). Endocrine disruptors will be discussed

in more detail in Chapter 17.

the sex change may be stress hormones (e.g., cortisol) that induce sex-speCific neuropeptides, the effector of the

change may be (once again) aromatase. There are two vari-

SEX DETERMI NATION

SociaJ and geological factors

/r��T� (soil qp�'T:i,

Hydrology

Geology

Currenl

Past Zine use

Meterology

�

Atrazine exposure

Molecular biology

537

FIGURE 14.24 Poss i ble chain of ca usa­ tion leading to the feminization of male frogs and the decline of frog populations in regions where atrazine has been used to control weed populations. Socia', geo­ logical, and bio log ica l agents are shown. CYP 1 9 is the gene encoding aromatase, and it has been shown that transcription of the human CYP T 9 gene is induced by atrazine (Sanderson et al. 2(00). (From Hayes 2005.1

CYP19

gene Biochemis[ry

Biological factors

Developmental endrocrinology

�

/

Feminization

Demasculinization Physiological! behavioral organismal Evolution

/

�

Decreased reproductive success

or

in the in the gonads. Black and colleagues (2005) showed a str iking correlation between changes in brain aromatase Jevels and changes in sexual behavior during the sexual transitions of gobies, in which the school typi­

ants of aromatase in many animals, one expressed brain and one

cally contains only one male and multiple females. The

increase (>200%) in the aggressive behavior of the largest removal of the male from a stable group caused a rapid

female, wruch then became a male in about a week's time.

This transformation may have resulted from an increase in brain testosterone levels, since .within hours upon removal of the male, these dominant females developed a lower brain aromatase than the other females

(Figure

14.26). The gonadal aromatase levels, however, stayed the same, and gonadal sex change came later. In porgy fish,

I '" . � � -

10

em

I

-�+-'

1

Proboscis (may-­

extend meter) Male Bonel/in (lives symbiotically inside the reproductjve organs of the female)

aromatase inhibitors can block the natural sex change and induce male development (Lee et al.

2002).

Thus, changes

in the social group, perceived by the nervous system, became expressed by the hormonal system within hours, thereby changing the behavioral phenotype of the female fish. Interestingly, when it comes to behaviors

in fish, sex

is in the brain before it is in the gonads.

See WEBSITE 14.5 Forms of hermaphrodilism

FIGURE 14.25 Sex determ ination in Bonellia viridis. Larvae that settle on the ocean floor become female. The mature female's body is about 1 0 cm long, with a l -meter-Iong proboscis that emits chemicals which attract other B. viridis l a rvae. Larvae that land on the p(Qboscis are taken in to the female'S body, where they develop and live symbiotically as ti ny (1-3 mm longl males.

(A)

CHAPTER 1 4

538

• o •

FIGURE 14.26

Aggressive behavior and aromatase activity (AA) in the brain and gonads of the goby Lylhrypnus dalli. (A) On day 4 (prior to male removal), there was no statistical difference in aver­ age daily displacements among the largest females. On day 5, the male was removed and dominant females increased their aggres­ sive behavior. (Dominance¥phase fish have no day 5 data because they were sacrificed on day 5 just after.) (B,C) Brain (B) but not gonadal (C) AA was significantly lower in dominance-phase and sex changed individuals compared with control females. Estab­ lished males had lower brain AA than all other groups and lower gonadal AA than all groups except dominance-phase females. (After Black et al. 2005.) -

(B) Brain 12

e

e

5

20 r

Coda

Nature has many variations on her ma st rpiec In some species, including most mammals and insects, sex is deter­ mined by chromosomes; in other species, sex is a matter of environmental conditions. In yet other species, both environmental and genotypic sex determination function, often in different geographical areas. Different environ­ mental or genetic stLmuli may trigger sex determh1ation through a series of conserved pathways. As CTews and Bull (2009) have reflected, "it is possible that the develop­ mental decision of male versus female does not flow through a single gene but is instead determined by a 'par­ liamentary' system involving networks of genes that have sim ul taneous inpu ts to several components of the down­ stream cascade." We are finally beginning to understand the mechanisms by which this Umasterpiece of nature" is created.

4

Dominfemal ance­e Sex-changed in'?di--vidualO phase

Control female

or

Average for day (male present) Average for day Observation before sacrifice

8

.

�

4

-

���_� [J_ Control DomlnanceSex-changed female phase female individual e

O L-�

� � __

__

(e)

Gonads 180

Mal

Control an ee ei d- hadn female phoaseinfemal D m

c

­

d

ge S x. c n ivi ual

a

M le

Snapshot Summary; Sex Determination 1.

i

In mammals, pr m ry sex determination (the deter­ mination of gonadal sex) is a function of the sex chromosomes. XX individuals are females, XY indi­ viduals are males.

a

2. The Y chromosome plays a key role in male sex determination. XY and XX manunals both have a bipotential gonad. In XY animals, Sertoli cells differ­ entiate and enclose the germ cells within testis cords. The interstitial mesenchyme becomes the Leydig cells. 3. In XX individuals, the germ cells become surround­

ed by follicle cells in the cortex of the gonadal rudi­ ment. The epithelium of the folUdes becomes granulosa celis; the mesenchyme becomes the thecal cells.

the

4. In humans, the SRY gene is the testis-determining factor on the Y chromosome. It synthesizes a nucleic acid-binding protein that may function as either a transcription factor or as an RNA splicing factor. It activates the evolutionarily conserved SOX9 gene.

5. The SQX9 gene product also initiate testes for­ mation. Functioning as a transcription factor, it binds to the gene encoding anti-Mullerian factor is responsible for activating FGF9. Fgf9 and Sox9 pro­ teins have a positive feedback loop that activates tes­ ticular development and suppresses ovarian devel­ opment.

can

6. Wnt4 and Rspo-l are involved in ovary formation. These proteins upregulate production of �-catenin; the functions of p-ca tenin include promoting the ovarian p thw y of development while blocking testicular pathway of development

a

a

the

SEX DETERMINATION

the factors produced by the developing gonads. In

ent. The proteins translated from both dsx messages

male mammals, the Miillerian duct is destroyed by

of a set of genes involved in producing the sexually

7. Secondary sex determination in mammals involves

are active, and they activate or inhibit transcription

the AMF produced by the Sertoli cells, while testos­ terone produced by the Leydig cells enables the Wolffian duct to differentiate into the vas deferens and seminal vesicle. In female mammals, the Wolffi­ an duct degenerates with the lack of testosterone, whereas the MUllerian duct persists and is differenti­ ated by estrogen into the oviducts, uterus, cervix, and upper portion of the vagina .

S.

The conversion of testosterone to dihydrotestos­ terone in the genital rudiment and prostate gland precursor enables the differentiation of the penis, scrotum, and prostate gland.

9.

11.

dimorphic traits of the fly. B.

Sex determination of the brain may have different

Drosophila Tra proteins also activate the fruitless gene

downstream agents than in other regiOns of the body. in males (but not in females); in mammals, the Y chro­ mosome may activate brain sexual differentiation independently from the hormonal pathways.

14. Dosage compensation is critical for the regulation of gene expression in the embryo_ With the salne num­ ber of autosomes, the transcription from the X chro­

In mammals, one X chromosome of X X

mosome must be equalized for XX females and X Y

males.

Individuals with mutations of these hormones or

females is inactivated. I n Drosophiln, the single X

thejr receptors may have a discordance between

chromosome of XY males is hyperactivated.

their primary and secondary sex characteristics.

10.

539

Drosophila, sex is determined by the number of X

chromosomes in the cell; the Y chromosome does not In

15. In turtles and alligators, sex is often determined by the temperature experienced by the embryo during the time of gonad determination. Because estrogen is

play a role in sex determination. There are no sex

necessary for ovary development, it is possible that

hormones, so each cell makes

differing levels of aromatase (an enzyme that can

a

sex-determination

"decision." However, paracrme factors play impor­

convert testosterone into estrogen) distinguish male

tant roles in forming the genit'" struchlres.

from female patterns of gonadal differentiation.

The

Drosophila Sex-lethal gene is activated in females

(by the accumulation of proteins encoded on the X

an RNA splicing factor to splice an inhibitory exon

16. Aromatase may be activated by envirorunental com­ pounds, causing demasculinization of the male

chromosomes) but repressed in males. SxJ protein acts

gonads in those animals where primary sex determl·

as

na tion can be effected by hormones.

transformer (tra) transcript. Therefore, female

flies have an active Tra protein, while males d o not.

from the

12. The Tra protein also acts as

an RNA splicing factor to

splice exons from the doublesex

(dsx) transcript. The

17. In some species, such as

Bonellia and Crepidu/a, sex is In

determined by the position of the individual with regard to other individuals of the same species.

fish, numerous environmental factors-especially

dsx gene is transcribed in both XX and XY cells, but its pre-mRNA is processed to form different

ent in the population-­

age-associated degeneration are not properly part of the study of developmen­

es. Others scientists claim that the genetically determined, species-specific pat­ terns of aging are an important part of the life cycle and believe that

gerontology-the scientific study of aging-is rightly part of developmental biology. As Peter Medawar (1957) noted, "That which we call 'development' when looked at from the birth and the life end becomes 'senescence' when looked at from its close." Whatever their relationship to embryonic development, meta­ morphosis, regeneration, and aging are critical topics for the biology of the twen­ ty-first century.

METAMORPHOSIS: THE HORMONAL REACTIVATION OF DEVELOPMENT

Animals (including humans) whose young are basically smaller versions of the

adult are referred to as direct d evelope rs In most animal species, however, .

embryonic development includes a larval stage with characteristics very differ­ morphosis; these animals are indirect developers.

ent from those of the adult organism, which emerges only after a period of meta­ Very often, larval forms are specialized for some function such as growth or

dispersal, while the adult is specialized for reproduction.

Cecropia moths, for

example, hatch from eggs and develop as Wingless juveniles-caterpillars-for several months. After metamorphosis, the insects spend a day or so as fully

I'd give my right arm to know the secret of regeneration. OSCAR E. SCHOlTE

( 1 950)

542

CHAPTER 1 5

TABLE 1 5.1 Summary of some metamorphic changes in anurans System

larva

Adult

Locomotory

Aquatic; tail fins

Terrestrial; tailless tetrapod

Respiratory

Gills, skin, lungs; larval hemoglobins

Skin, lungs; adult hemoglobins

Circulatory

Aortic arches; aorta; anterior, posterior, and common jugular veins

Carotid arch; systemic arch; cardinal veins

Nutritional

Herbivorous: long spiral gut; intestinal symbionts; small mouth, horny jaws, labial teeth

Carnivorous: short gut; proteases; large mouth with long tongue

Nervous

Lack of nictitating membrane; porphyropsin, lateral line system, Mauthner's neurons

Development of ocular muscles, nictitating membrane, rhodopsin; loss of lateral line system, degeneration of Mauthner's neurons; tympanic membrane

Excretory

Largely amD1onia� some urea (ammonotelic)

Largely urea; high activity of enzymes of ornithine-urea cycle (ureotelic)

Integumental

Thin, bilayered epidermis with thin dermis; no mucous glands or granular glands

Stratilied squamous epidermis with adult keratins; well-developed dermis contains mucous glands and granular glands secreting antimlcrobial pep tides

Source: Data from Turner and Bagnara 1976 and Reilly et al. 1994.

developed winged moths and must mate quickly before they die. The adults never eat, and in fact have no mouth­ parts during this brief reproductive phase of the life cycle. As might be expected, the juvenile and adult forms often live in different environments. During metamorphosis� developmental processes are reactivated by specific hor­ mones, and the entire organism changes morphologically, phYSiologically, and behaviorally to prepare itself for its new mode of existence. Among indirect developers, there are two major types of larvae.* Secondary larvae are found among those ani­ mals whose larvae and adults possess the same basic body plan. Thus, despite the obvious differences between the caterpillar and the butterfly, these two life stages retain the same axes and develop by deleting and modifying old parts while adding new structures into a pre-existing framework. Similarly, the frog tadpole, although special­ ized for an aquatic environment, is a secondary larva, organized on the same pattern as the adult will be Uager­ sten 1972; Ralf and Ralf 2009). Larvae that represent dramatically different body plans than the adult form and that are morphologically distinct from the adult are called primary larvae. Sea urchin lar­ vae, for instance, are bilaterally symmetrical organisms that float among and collect food in the plankton of the *Although there is controversy on the subject

,

evolved after the adult form had

larvae probably

been established. In other words,

animals evolved through direct development,. and larval forms

came about as specializations for feeding or dispersal during

the

early part of the life cycle Genner 2000; Rouse 2000; Raff and Raff

2009). Even so, the biphasic life cycle may be a trait characteristic of metazoans (see Degnan and Degnan 2010).

open ocean. The sea urchin adult is pentameral (organized on fivefold symmetry) and feeds by scraping algae from rocks on the seafloor. There is no trace of the adult form in the body plan of the juvenile (see Figure 5.22).

Amphibian Metamorphosis Amphibians are named for their ability to undergo meta­ morphosis, their appellation coming from the Greek amph' ("double") and bios ("life"). Amphibian metamorphosis is associated with morpholOgical changes that prepare an aquatic organism for a primarily terrestrial existence. In urode.les (salamanders), these cllanges include the resorp­ tion of the tail fin, the destruction of the external gills, and a change in skin structure. In anUIans (frogs and toads), the metamorphic changes are more dramatic, with almost even· organ subject to modification (Table 15.1 ; see also Figure 1.1). The changes in amphibian metamorphosis are initiat­ ed by thyrOid hormones such as thyroxine (T4) and trio iodothyronine (T3) that travel through the blood to reach all the organs of the larva. When the larval organs encounter these thyroid hormones, they can respond in any of four ways: growth, death, remodeling, and respecification.

Morphological changes associated with amphibian metamorphosis GROWTH OF NEW STRUCTURES The hormone tri-iodothy­ ranine induces certain adult-specific organs to form. The limbs of the adult frog emerge from specific sites on the metamorphosing tadpole, and in the eye, both nictitating membranes and eyelids emerge. Moreover, T3 induces the proliferation and differentiation of new neurons to serve

POSTEMBRYONIC DEVELOPMENT

543

(B)

these organs. As the limbs grow out from the body axis, new neurons proliferate and differentiate in the spinal cord.

lllese neurons send axons to the newly formed limb mus­ cu l ature (Marsh-Armstrong et al. 2004). Blocking T3 activ­ ity prevents these neurons from fOrming and causes paral­ ysis of the l.imbs. One readily observed consequence of anuran metamor­ from their originally lateral position ( F igure 1 5.1 ).' The lat­ phosis is the movement of the eyes to the front of the head

eral eyes of the tadpole are typical of preyed-upon herbi­ vores, whereas the frontally located eyes of the frog befit its more predatory lifestyle. To catch its prey, the frog needs to see in three dimensions, That is, it has to acquire a binoc­

(e)

(D)

ularfield ofVisiOI1, where inputs from both eyes converge in the brain (see Chapter 1 0). In the tadpole, the right eye innervates the left side of the brain, and vice versa; there

are no ipSilateral (same-side) projections of the retinal neu­ rons. During metamorphosis, however, ipsilateral path­

ways emerge, enabling input from both eyes to reach the

same area of the brain (Currie and Cowan 1974; Hoskins

Xenopus, these new neuronal pathways result not from

and Grobstein 1985a). In

the remodelIng of existing neurons, but from the formation

of new neurons tllat differentiate in response to thyroid hor­ mones (Hoskins and Grobstein 1985a,b). The ability of tl,ese axons to project ipsilaterally results from the induction of ephrin B in the optic chiasm by the thyroid hormones (Nak­ agawa et al. 2000). Ephrin B is also found in the optic chi­

asm of mammals (which have ipsilateral p rojections throughout life) but not in the chiasm of fish and birds

(which have only contralateral projections). As shown in Chapter 10, ephrins can repel certain neurons, causing them to project in one direction rather than in another.

CEll DEATH DURING METAMORPHOSIS The hormone T3 also

induces certain larval-specific struc tures to die. Thus, T3

FIGURE 15.1 Eye migration and associated neurona l changes during metamorphosis of the Xenopus faevis tad pol e. (A) The eyes of the tadpole a re laterall y placed, so there is relatively li«le binocular field of vision. (B) The eyes migrate oo rsa lly and rostral­ Iy during metamorphosis, creating a l arge binocular field for the adu lt frog. (CD) Retinal projections of metamorphosing tadpole. The dye Dil was placed on a cut stump of the optic nerve to label the retinal projection . (C) In early and middle stages of metamor­ phosis, axons project across the midline (dashed line) from one side of the b rai n to the other. (0) In late metamorp hosis, ephri n B is produced in the o pti c chiasm as certain neurons (arrows) are formed that project i ps ilatera l ly. (A,S from Hoskins and Grobstein 1 984, cou rtesy of P. G ro ste i n ; C,D from Nakagawa et al. 2000, cou rtesy of C. E. Holt .)

b

causes the degeneration of the paddle-like tail and the oxy­

gen-procuring gills that were important for larval (but not adult) movement and respiration. While it is obvious that the tadpole'S tail muscles and skin die, is this death mur­ der or suicide? In other words, is T, telling the cells to kill themselves, or is T3 telling something else to kill the cells?

jinla et al. 2005). This was confirmed by the demonstration that the apoptosis-inducing enzyme caspase -9 is lmpor�

tant in causing cell death in the tadpole muscle cells (Rowe

et al. 2005). However, later in metamorphosis, the tail mus­ cles are destroyed by phagocytOSiS, perhaps because the

Recent evidence suggests that the first part of tail resorp­

extracellular matrix that supported the muscle cells has

tion is caused by suicide, but that the last remnants of the

been digested by proteases.

tadpole tail must be killed off by other means. When tad­

Death also comes to the tadpole'S red blood cells. Dur­

pole muscle cells were inject ed with a dominant negative

ing metamorphosis, tadpole hemoglobin is changed into

T3 re cep tor (and therefore could not respond to T,), the muscle cells survived, indicating that T3 told them to kill tllemselves by apoptosis (Nakajinla and Yaoita 2003; Naka-

adult hemoglobin, which binds oxygen more slowly and releases it more rapidly (McCutcheon 1936; Riggs 1951). The red blood cells carrying the tadpole hemoglobin have a different shape than the adult red blood cells, and these

'One of the most spectacular movements of eyes during metamor­ phosis occurs in flatfish such as flounder. Originally, a flounder's eyes, like the lateral eyes of other fish species, are on opposite sides of its face. However, during metamorphosis, one of the eyes migrates

"8 •

" �

:;:

c..

Gradient shown in �

Compartmentalization and anlerior- posterior pat­ terning i n the wing imaginal disc. (A) In the first instar l arva, the a nteri or-posterior axis has been formed and can be recognized by the expression of th e engrailed gen e in the posterior compart· ment. Engrailed, a t ranscripti on factor, activates the hedgehog gene. Hedgehog acts as a short-range paracrine factor to activate decapentaplegic (dpp) in th e anterior cells adjacent to the posteri­ or compartment, where Opp and a related protein, Glass-bottom boat (ebb), act over a l onger range. (B) Dpp and ebb proteins FIGURE 15.14

l

1990).

become the most distal structures of the leg-the claw and

(A)

i

disc epithelium (Condie et al. Using fluorescently labeled phalloidin to sta n the peripheral microfilaments of leg disc cells, they showed tha, the cells of early third instar discs are tight y arranged along the proximal-distal axis. When the hormonal Signal to dif­ ferentiate is given, the cells change their shape and Ihe leg is everted, the central cells of the disc becoming the rna distal (claw) cells of the limb. The leg structures will differ­ entiate within the pupa, so that by the time the adult fly ecloses, they are fully formed and functional. cell shape changes within the

paracrine facto r) . High concentrations of these paracrine factors cause the expressi on of the Distal-less gene. Mod­ erate concentrations cause the expression of the dachshund gene, and lower concentrations cause the expression of the

�

...... . . . . - -. . . . . .�

prOJecte gradient : ofGbb : : : + Dpp "". /'1:-- : ::

L2 L3 14

L5

,

,

'Ante·rioT ,

(B) , ' �

b,k

:

< , I

1

•

:

:

, , ,

,

,

dpp

, ,

,

,

'

: : :

'

• '

:

POfterior

sal

omb

create a concentration gradi ent of BMP-like signaling, measured by the p hosphoryl ati on of Mad (pMad). High concentrations of Dpp plus Gbb near the source activate both the spail (san and oplOmolOr blind (omb) genes. Lower concentrations (near the periphery) activate omb but ot sal. When Dpp plus ebb levels

n

drop below a certain threshold, brinker (brk) is no longer repressed. L2-L5 mark the longitudinal wing veins, with L2 being the most anterior. (After Bangi and Wharton 2006.)

POSTEMBRYONIC DEVELOPMENT (A)

(B)

t

FIGURE 15.15 surface of

(e)

555

Determining the dorsol-ventral axis. (A) The p rospective ventral the wing is stained by antibodies to Ves igial protein (green), while the prospective dorsal surface is stai ned by antibodies to Apterous protein (red). The region of yellow illustrates where the two proteins overlap in the margi n . (B) Wingless protein (purple) synthesized at the marginal juncture organizes the wing disc along the dorsal-ventral axis. The expression of Vestigial (green) is seen in cells dose to those exp ressi n g Wi ngless. (C) The dorsal and ventral portions of the wing disc te lescope out to form the tvvo-Iayered wing. Gene expression patterns are indicated on the double-layered wing. (A,B courtesy of s. Carroll a ndS. Paddock.)

Dorsal (Apt o s expression) Ventral (Vestigial expression) er u

. Margm (Wingless expression)

factor Hedgehog. Hedgehog functions only when cells have the receptor (Patched) to receive it. In a complex man­ ner, the diffusion of Hedgehog activates the gene encod­ ing Oecapentaplegic (Dpp) in a na rrow stripe of cells in the anterior reg ion of the wing disc (Ho et al. 2005). Opp and a co-expressed BMP called Glass-bottom boat (Gbb) act to establish a gradient of BMP signaling activity. BMPs activate the Mad transcription factor (a Smad protein) by phosphorylating it, so this gradient can be measured by the phosphorylation of Mad. Dpp is a short-range paracrine factor, while Gbb exhibits a much longer range of diffusion to create a gradient (Figure 15.14B; Bangi and Wharton 2006). nus signaling gradient regulates the amount of cell prolif­ eration in the wing regions and also specifies cell fates (Rogulja and Irvine 2005). Several transcription factor genes respond differently to activated Mad. At high levels, the spalt (sal) and optomotor blind (omb) genes are activated, whereas at low levels (where Gbb prOVides the primary signal), only omb is activated. Below a particular level of phosphorylat­ ed Mad activity, the brinker (brk) gene is no longer inhibit­ ed; thu s brk is expressed outside the signaling domain. Spe­ cific cell fates of the wing are specified in response to the action of these transcription factors. (For example, the fifth longitudinal vein of the wing is formed at tl1e border of opto­ motor blind and brinker; see Figure 15.14B). DORSAL-VENTRAL AND PROXIMAL-DISTAL AXES The dorsal­ ventral axis of the wing is fomled at the second instar stage by the expression of the apterolls gene in the prospective

dorsal cells of the wing disc (Blair 1993; Oiaz-Benjumea

and Cohen 1993). Here, the upper l aye r of the wing is dis­ tinguished from the lower layer of the wing blade (Bryant 1970; Garcia-Bellido et aL 1973). The vestigial gene remains "on" in the ventral portion of the wing disc (Figure 15.15A). The dorsal portion of the wing synthesizes transmembrane proteins that prevent the intermixing of the dorsal and ven­ tral cells (Milan et aL 2005). At the boundary between the dorsal and ventral comparlments, the Apterous and Ves­ tigial transcrip tion factors interact to activate the gene encoding the Wn t pa racr ine factor Wingless (Figure 1 5.15B). Neumann and Cohen (1996) showed that Wing­ less protein acts as a growth factor to promote the cell pro­ liferation that extends the wing.' Wingless also helps estab­ lish the proximal-distal axis of the wing: high levels of Wingless activate the Distal-less gene, which speCifies the most distal regions of the wing (Neumann and Cohen 1996, 1997; Zecca et aL 1996):This is in the central region of the disc and "telescopes" outward as the distal margin of the wing blade (Figure 15.15C). See WEBSITE 1 5.1 The molecular biology of wing formation See WEBSITE 1 5.2 Homologous specification *The diffusion of paracrine factors such oS Wingless and Hedgehog is facilitated when these factors cluster on lipid spheres that can travel betvoJeen cells without getting caught in the extracellular matrix (Glise et a!. 2005; Gorfinkiel et al. 2005; Panakova et al. 200S).

556

CHAPTER 1 5

Hormonal control of insect metamorphosis Although the details of insect metamorphosis differ among species, the general pattern of honnonal action is very sim­ ilar. Like amphibian metamorphosis, the me tamorphosi s of insects is regulated by systemic hormonal signals, which are controlled by neurohormones from the brain (for reviews, see Gilbert and Goodman 1981; Riddiford 1996). Insect molting and metamorphosis are controlled by two effector hormones: the steroid 20-hydroxyecdysone (20E) and the lipid juvenile hormone UH) (Figure 1 5. 16A) . 20Hydroxyecdysonet initiates and coordinates each molt and regulates the changes in gene expression that occur dur­ ing metamorphosis . Juvenile hormone prevents the ecdysone-induced changes in gene expression that are nec­ essary for metamorphosis . Thus, its presence during a molt ensures that the result of that molt is another larval instar, not a pupa or an adult.

The molting process (Figure

1 5. 1 6 B) is initiated

in the

brain, where neurosecretory cells release prothoracicotrop­

ie hormon e (PTTH)

in response to neural, hormonal, or

environmental signals. PTTH is

a peptide hormone with a molecular weight of a pp roximately 40,000, and it stimu­ lates the production of ecdysone by the prothoracic gland by activating the RTK pathway in those cells (Rewitz et a1. 2009). Ecdysone is modilied in peripheral tissues to become

cHic mRNAs

are not replaced,

and new

.rn RNA

s are syn­

thesized whose protein products inhibit the transcription 0:

the larval messages.

There are two maj or p u lses of 20E d uring Drosophi:.l first pulse occurs in the third instar larva and triggers the initiation of ("prepupal") morpho­ genesis of the leg and wing imaginal discs (as well as the death of the larval hindg ut) . The larva stops eating and migra tes to find a site to begin pupation. The second 20E pulse occurs 10-12 hours later and tells the "prepup a" te become a pupa. The head inverts and the salivary gland> degenerate (Riddiford 1982; Nijhout 1994). It appears, then that the first ecdysone pulse during the last larval ins tar triggers the processes that inactivate the l arva-specific metamorphosis. The

genes and initiates the morphogenesis of imaginal disc

structures. The second pulse transcribes pupa-specific genes and initiates the molt (Nijhou t 1994). At the imagi­ nal molt, when 20E acts in the absence of juvenile hormone, the imaginal discs fully differentiate and the molt gives rise to an

adult.

See WEBSITE 1 5.3 Insect metamorphosis The molecular biology of 20-hydroxyecdysone activity

the active molting hormone 20E. Each molt is initiated by

ECDYSON E RECEPTORS 20-Hydroxyecdysone cannot bind

one or more pulses

to DNA by itself. Like amphibian thyroid hormones, 20E first binds to nuclear receptors. These proteins, called ecdysone receptors (EcRs), are almost identical n i structure to the thyroid hormone receptors of amphibians. An EcR protein forms an. active molecule by pairing with an Ultra­ spiracle (Usp) protein, the homologu e of the amphibian RXR that helps form the active thyroid hormone receptor

of 20E. For a larval molt, the first pulse produces a small rise in the 20E concentration in the larval hemolymph (blood) and elicits a change in cellular com­

mitment in the epidermis. A second, larger pulse of 20E

associated with molting. the epidermal cells to synthesize enzymes that digest the old cuticle and s ynthesize a new one. Juvenile hormone is secreted by the c orpora allata. The secretory cells of the corp ora allata are active during lar­ initiates the differentiation events

These pulses of 20E commit and stimulate

val molts but inactive during the metamorphic molt and the imaginal molt. As long as

JH is present, the 20E-stim­

ulated molts result in a new larval instar. In the last larval instar, however, the medial nerve from the bra.in to the co[­ pora allata inhibits these glands from producing JH, and there is

a simultaneous increase in the body's ability to degrade existing JH (Sa franek and Williams 1989). Both these mechanisms cause JH levels to drop below a critical threshold value, triggering the release of PTTH from the brain (Nijhout and Williams 1974; Rountree and Bollen­ bacher 1986). PTTH, in turn, stimulates the prothoracic gland to secrete a small amount of ecdysone. The result­ ing pulse of 20E, in the absence of high levels ofjH, com­ mits the epidermal cells to pupal development. Larva-spe-

tSince its discovery in ] 954, when Butenandt and Karlson isolated 25 mg of ecdysone from 500 kg of silkworm moth pupae, 20hydroxyecdysone has gone under several names, i.ncluding ccdys­ terone, p-ecdysone, and crustecdysone.

(Koelle et al. 1991; Yao et al. 1992; Thomas et al. 1993). In

the absence of the hormone-bound EcR, the Usp protein binds to the ecdysone-responsive genes and inhibits their transcription.* This inhibition is converted into activation when the ecdysone receptor binds to the Usp (Schubiger and Truman 2000).

Although there is only one gene lor EcR, the EcR mRNA be spliced in at least three different ways to form three distinct proteins. AU three EcR proteins have the same domains for 20E and DNA binding, but they differ in their amino-terminal domains. The type of EcR present in a cell may inform that cell how to act when it receives a hor­ monal signal (Talbot et aJ. 1993; 'Ihunan et al. 1994). All cells appear to have some EcRs of each type, but the strictly lar­ val tissues and neurons that die when exposed to 20E are characterized by their abundance of the EcR-BI isoforru 01 the ecdysone receptor. Imaginal discs and differentiating transcript can

neurons, by contrast, show a preponderance of the EcR-A

*The Ultraspirade protein may be a juvenile hormone receptor, or J1iR (see Figure 15.16), suggesting mechanisms whereby JH can block 20E at the level of transcription (Jones et at 2001; Jones ct al. 2001, 2007; Sa,orith et al. 2002).

557

POSTEMBRYONIC DEVELOPMENT

�I I , "" "" �

Brain

(8 )

(A) luvenHe hormone (jH)

o

'H

cells

� ../""' Neurosecretory �-:0:::':�:�:�rOthoracic gland

OCH3 Juvenile hormone

Ecdysone

---

"

/

ProtI10raClcotroPlc " horm o e (PTTH)

Corpora

JH receptor OHR) OH

si"Mol gnalt"ing

HO

("molting"

\ '7 ��

ProthO�iC� gland

V JH-JHR

of

-�� � n allata

o

"Differentiating signal"

o

20-Hydroxyecdysone (20E) hormone) OH HO :

Ecdysone

f

e /'

20E

OH HO HO o

FIGURE 1 5,16 Regulation of insect meta­ morphosis. (A) Structures of juvenile ho r­ mone (JH), ecdysone, and the ilctive molting hormone 20-hydroxyecdysone (20E), (B) Gen eral pathway of insect metamorphosis. 20E and I H toge the r cause molts that form the next larval instar. When the concentra­ tion of I H becomes low enough, Ihe 20E­ induced molt produces a pupa i nstead of an instar. When ecdysone acts in the absence of JH, the imaginal discs differentiate and the molt gives ri se to an adult (Imago). (After Gilbert and Goodman 1 981 .)

PrQtein """h,esi,

Protein

t

t

Protein synthesis

synthesis

Adult structures

Pupal structures Cuticle

Larva

isoform. Mutations in specific codons that are found in only some of the splicing isoforms indicate that the different forms of EcR play different roles in metamorphosis and that the different receptors activate dUferent sets of genes when

they bind 20E (Davis et a!. 2005).

BINDING OF 20-HYDROXYECDYSONE TO DNA During molt­

ing and metamorphosis, certain regions of the polytene

chromosomes of Drosophila puff out in the cells of certain organs at certain times (CLever 1966; Ashbumer 1972; Ash­ burne r and Berondes 1978). These chromosome p uffs are areas where DNA is being actively transc ribe d . Moreover,

P up

a

these organ-specific patterns of chromosome puffing can be reprod uced by c ul turing larval tissue and add ing hor­ mones to the medium, or by adding 20E to an c arl y-stafie larva. When 20E is ad d ed to larval salivary glands, certain puffs are produced and others regress (Figure 1 5.1 7) . The puffing is mediated by the binding of 20E at specific places on the chromosomes; fluorescent antibodies against 20E find this hormone localized to the regions of the genome that are sensitive to it (Groneme yer and Pongs 1980). At these sites, the ecdysone bound receptor complex recruits a hislone methyltransferase that me thylatos lysine 4 of his­ tone H3, thereby loosening the nuclcosornes in that area -

-

(Sedkov et aJ.

2003).

See VADE MECUM Chromosome squash

558

(AI

CHAPTER 1 5

I I"' I!" ,� , ��\"� il .,,: I

L

FIGURE 15.17 20E-induced puffs in cultured salivary gland cells of D. mel.nogaster. (A) Uninduced control. (S-E) 20E-stimulated chromosomes at (S) 25 minutes, (C) , hour, (D) 2 hours, and (E) 4 hours. (Courtesy of M. Ashburner.)

Figure

15.18A shows a simplified schematic for the

in Drosophila. 20E binds to the EcR/USP receptor complex. It activates the "earl\' response genes," including E74 and £75 (the puffs in Figure 15. 17), as well as Broad and the EcR gene itself. The tran­ scription factors encoded by these genes activate a second series of genes, sucl1 as £75, DHR4, and DHR3. The prod­ ucts of these genes are transcription factors that work together. First, they activate {3FTZ-Fl, a gene encoding a transcription factor that enables a new set of genes to respond to the second burst of 20E. Secondly, the products of these genes shut off the early genes so that they do not interfere with the second burst of 20E. Moreover, DHR. coordinates growth and behavior in the larva. It allows the larva to stop feeding once it reaches a certain weight and to begin searching for a place to glue itself to and form a pupa (Urness and Thummel 1995; Crossgrove et al. 1996; Klng-Jones et al. 20OS). The effects of these two 20E pulses can be extremely dif­ ferent. One example of this is the ecdysone-mediated changes in the larval salivary gland. The early pulse of 20E activates the Broad gene, whicl1 encodes a family of tran­ scription factors through differential RNA splicing. The targets of the Broad complex proteins include those genes that encode the salivary gland "glue proteins." The glue proteins allow the larva to adhere to a solid surface, where it becomes a pupa (Guay and Guild 1991). So the first 20E pulse stimulates the function of the larval salivary gland. However, the second pulse of 20E calls for the destruction of this larval organ (Buszczak and Segraves 2000; Jiang et a1. 2000). Here, 20E binds to the EcR-A form of the ecdysone receptor (Figure When complexed with USP, it activates the transcription of early response genes E74, E75, and Broad. But now a different set of targets is activated. These transcription factors activate the genes encoding the apoptosis-promoting proteins Hid and i p2 gene Reaper, as well as blocking the expression of the da (whim would otherwise repress apoptosis). Thus, the first 20E pulse activates the salivary gland, and the second 20E pulse destroys it. Like the ecdysone receptor gene, the Broad gene can gen­ erate several different transcription factor proteins through differentially initiated and spliced messages. Moreover, the variants of the ecdysone receptor may induce the synthesis of particular variants of the Broad proteins. Organs such as the larval salivary gland that are destined for death during metamorphOSiS express the 21 isofonn; in1aginal eliscs des­ tined for differentiation express the 'Z2 isoform; and the cen­ tral nervous system (whim undergoes marked remodeling during metamorphOSis) expresses all isoforms, with Z3 preframework of metamorphosis

(E)

20E-regulated chromosome puffing occurs during the late stages of tl,e third instar Drosophiln larva, as it prepares to form the pupa. The puffs can be divided into three cat­ egories: "early" puffs tl,at 20E induces rapidly; "intermolt" puffs that 20E causes to regress; and "late" puffs that are first seen several hours after 20E stiml).lation. For example, in the larval salivary gland, about six puffS emerge with­ in a few minutes of hydroxyecdysone trealment. No new protein has to be made in order for these early puffs to be induced. A much larger set of puffs is. induced later in development, and these late puffs do need protein synthe­ sis to become transcribed. Ashburner (1974, 1990) hypoth­ esized tl1at the "early pulf" genes make a protein product that is essential for the activation of the "late pulf" genes and that, moreover, this early regulatory protein itself twns off the transcription of the early genes ' These insights have been confirmed by molecular analyses. "The observation that 20E (ootroHed the transcriptional units of chromosomes was an extremely important and exciting discovery. This was our first real glimpse of gene regulation in eukaryotic organisms. At the time when this discovery was made, the only examples of transcriptional gene regulation were in bacteria.

15.18B).

POSTEMBRYONIC D EVELOPMENT

(A)

Pupatio , , , 20E + EcR/USP , , , ,,

t

n

Head eversion , , 20E + EcRfUSP ,, , , , , , , , , , , , , ,

j

EeR, E74, E75, Broad

�75 {3FTZ-Fl

L-

Third instar

Ir::atel :�

_ ,, , ,, ,

_ _

Mid prepupa

late prepupa

Pupa

+

- y

20 Ecd sone

(B)

t

USP + EcRA 20E

FIGURE 15 .18 20-Hydroxyecdysone initiates deve lopmenta l cascades. (A) Schematic of the major gene expression cascade in Drosophila metamorphosis. When 20E binds to the EeR/USP receptor complex, it activates the early response genes, including £74, £75, and Broad. Their products activate the "late genes." The activated EeR/USP complex (1lso acti­ vates a series of genes whose products are transcription factors and which activate the {3FTZ-F/ gene. The PFTZ-F I protein modifies the chromatin so that the next 20E pulse acti­ vates a different set of late genes. The products of these genes also inhibit the early­ expressed genes, including the EcR receptor. (B) Postulated cascade leading from ecdysone reception to death of the larval salivary gland. Ecdysone binds to the EcR-A isoform of the ecdysone receptor. After complexing with USP, the activated transcription factor complex stimulates transcription of the early response genes E74A, E758, and the Broad complex. These make transcription factors that pmmote apoptosis in the salivary gland cells. (A after King-Jones et ,l. 2005; B after Buszczak and Segraves 2000.)

559

560

CHAPTER 1 5

dominating (Emery et a1. 1994; Crossgrove et al. 1996). Juve­ nile hormone may act to prevent ecdysone-inducible gene expression by interfering with the Broad complex of pro­ teins (Riddiford 1972; Restifo and White 1991).

See WEBSITE 1 5.4 Precocenes and synthetic JH Like those of amphibian metamorphosis, the stories of insect metamor­ phosis involve complex interactions betvveen ligands and receptors. The "target tissues" are not mere passive recip­ ients of hormonal signals. Rather, they become responsive to hormones only at particular times. For example, when there is a pulse of 20E at the middle of the fourth instar of the tobacco hornworm moth Manduca, the epidermis is able to respond because this tissue is expressing ecdysone receptors. The wing discs, however, are unaffected by ecdysone until the prepupal stage, at which time they syn­ thesize ecdysone receptors, grow, and differentiate (Nijhout 1999). Thus, the timing of metamorphic events in insects can be controlled by the synthesis of receptors in the tar­ get tissues. Metamorphosis remains one of the most striking of developmental phenomena, yet we know only an outline of the molecular bases of metamorphosis, and only for a handful of species. COORDI NATION OF RECEPTOR AND LIGAND

REGENERATION Regeneration is the reactivation of development in postem­ bryonic life to restore missing tissues. The ability to regen­ erate amputated body parts or nonfunctioning organs is so "unhuman" that it has been a source of fascination to humans since the beginnings of biolOgical science. It is dif­ ficult to behold the phenomenon of limb regeneration in newts or starfish without wondering why we cannot grow back our own arms and legs. What gives salamanders this ability we so sorely lack? In fact, experimental biology was born of the efforts of eighteenth-century naturalists to answer this question. The regeneration e�periments of Trem­ blay' (hydra), Reaumur (crustaceans), and Spallanzani (sala­ manders) set the standard for experimental research and for the intelligent discussion of one's data (see Dinsmore 1991). More than two centuries later, we are qeginning to find answers to the great questions of regeneration, and at some point we may be able to alter the human body so as to per­ mit our own limbs, nerves, and organs to regenerate. Suc-

cess would mean that severed limbs could be restored, dis­ eased organs could be removed and then regrown, and nerve cells altered by age, disease, or trauma could once again function normally. Modem medical attempts to coax human bone and neural tissue to regenerate are discussed in Chapter 17, but to bring these treatments to humanity, we must first understand how regeneration occurs in those species that already have this ability.+ Our recently acquired knowledge of the roles of paracrine factors in organ formation, and our ability to clone the genes that produce those factors, have propelled what Susan Bryant (1999) has called "a regeneration renaissance. Since ren­ aissance literally means "rebirth/' and since regeneration can be seen as a return to the embryonic state, the term is apt in many ways. Regeneration does in fact take place in all species and can occur in four major ways: II

1. Stem-cell m edia ted regeneration. Stem cells allow an

organism to regrow certain organs or tissues that have been lost; examples include the regrowth of hair shafts from follicular stem cells in the hair bulge and the con­ tinual replacement of blood cells from the hematopoi­ etic stern cells in the bone marrow. 2. Epimorphosis. In some species, adult structures can undergo dedifferentiation to form a relatively undiffer­ entiated mass of cells that then redifferentiates to form the new structure. Such epimorphosis is characteristic of planarian flatworm regeneration and also of regen­ erating amphibian limbs. 3. Morphallaxis. Here, regeneration occurs through the repatterning of existing tissues, and there is little new growth. Such regeneration is seen in Hydra (a cnidari­ an). 4. Compensatory regeneration. Here, the differentiated cells divide but maintain their differentiated functions. The new cells do not come from stem cells, nor do they corne from the dedifferentiation of the adult cells. Each cell produces cells similar to itself; no mass of undiffer­ entiated tissue forms. This type of regeneration is char­ acteristic of the mammalian liver. Numerous examples of stern-cell mediated regeneration have been discussed throughout this book. In this chapter we will concentrate on epimorphosis in the salamander limb, morphallaxis in Hydra, and compensatory regenera­ tion in the mammalian liver.

+ Mammals do have a small amount of regenerational ability. In *Tremblay's advice to researchers who would enter this new field is

addition to regenerating body parts continuously through adult

pertinent even today: he advises us to go directly to nature and to

stem cells, rodents and humans can regenerate the tips of their dig­

avoid the prejudices that our education has given us. Moreover,

its if the animal is young enough. This ability has been correlated

"one should not become disheartened by want of success, but should try anew whatever has failed. It is even good to repeat suc­ cessful experiments a number of times. All that is possible to see is not discovered, and often cannot be discovered, the first time" (quoted in Dinsmore 1991).

2003; 2(05).

2004).

with the expression of the homeodomain transcription factor MSX1 fetal digit tips express MSXl in the migrating epidermis and subja­ (Han et al.

Kumar et al.

Apparently, amputated human

cent mesenchyme, just as regenerating amphibian limbs do (Allan

et

at.

POSTEMBRYO N I C DEVELOPMENT

Distal amputation

ProximaJ amputation

561

Formation of the apical ectodermal cap and regeneration blastema When a salamander limb is amputated, a p lasma clot

Original limb

forms; within 6-12 hours, epidermal cells from the remain­ i ng StunlP migrate to cover the wOlUld s urface, forming the wound epidermis. [n contrast to wound healing in

Amputation

mammals, no scar forms, and the dermis does not move

7d

nerves innervating the limb degenerate for a short distance

with the epidermis to cover the site of amputation. The

p roxim al to the plane of amputation (see Chernoff and Stocum 1995). During the next 4 days, the extracellular matrices of the tissues beneath the wound epidermis is degraded by pro­ teases, liberating Single cells that undergo dramatic dedif­ ferentiation: bone cells, cartilage cells, fibroblas ts, and myocytes all lose their differentiated characteristics. Genes that are expressed in differentiated tissues (SUdl as the mrf4

21d 2Sd 32d

and myfS genes expressed in muscle cells) are downreguJat­

42d

ed, while there is

genes such as

72d

a dramatic increase in the expression of

msxl

that are associated with the proliferat­

ing progress zone mesenchyme

of the embryonjc limb (Simon et a1. 1995). This cell mass is the regeneration blastema, and these are the cells that will continue to pro­

FIGURE 15.19

Regeneralion of a salamander forelimb. The amputation shown on the left was made below the elbow; the amputation shown on the right cut through the humerus. In both instances, the orrec t positional information was re-specifiecl and a no rma l limb was regenerated within 72 days. (From Goss 1 969, courlesy of R. J. Goss.) c

liferate, and which will eventually redifferentiate to form the new structures of the limb

(Figure 1 5.20; Butler 1935).

Moreover, during this time, the wound epidermiS thickens to form the apical epi dermal cap (AEC), which acts simi­ larly to the apical ectodermal ridge during normal Hmb development (Han et al. 2001). Thus, the previously well-structured limb region at the cut

edge of the stump forms a proliferating mass of indis­ cap. One of the major que stions of regeneration has been: do the cells keep a "memory" of what they had been? In other tinguishable cells just beneath the apical ectodermal

Epimorphic Regeneration of Salamander Limbs When an adult salamander limb is amputated, the remain­ ing limb cells are able to reconstruct a complete new limb, with all its differentiated cells arranged in the proper order. In other words, the new cells construct only the missing structures and no more. For example, when a w rist is amputated, the salanlander forms a .t:1ew wrist and not a new elbow (Figure 1 5. 1 9). In some way, the salamander limb "knows" where the proximal-distal axis has been sev­ ered and is able to regenerate from that

point on. Salaman­ ders accomplish epimorphic regeneration by cell dediHer­ entiation to form a regeneration blastema-an aggregation

words, do new muscles arise from old muscle celis, or can

any cell

of the blastema become a muscle? Kragl and col­ (2009) found that the blastema is not a collection of homogeneous, fully dedifferentiated cells. Ra ther in the regenerating limbs of the axolotl salamander, muscle cells arise only from old muscle cells, dermal cells come only

leagues

,

from old dermal cells, and cartilage can arise only from old cartilage or old dermal cells. Thus, the blastema is not a collection of unspecified multipotential progenitor

cells. the blastema

is a heterogeneous assortment of restricted progenitor cells. Kragl and olleag ues performed an experiment in which they transplanted limb tissue from a salamander whose cells expressed green fluorescent protein (GFP) into differ­ Rather, the cells retain their specification, and

c

1 5.21). U they transplanted the GFP-expressing limb cartilage into a salamander limb that did not contain the GFP transgene, the GFP-express­ ing cartilage would integrate normally into the limb skele­

of relatively dedifferentiated cells derived from the origi­

ent regions of limbs of normal salamanders that did not

nally differentiated tissue-which then proliferates and

have the GFP transgene (Figure

redifferentiates into the new limb parts (see Brockes and Kumar 2002; Gardiner et aJ.

2002). Bone, dermis, and carti­ lage j ust beneath the site of amputation contribute 10 the regeneration blastema, as do satellite cells from nearby muscles (Morrison el at. 2006).

ton. They later amputated the limb through the region con­ taining GFP-marked cartilage cells. The blastema was

562

CHAPTER I S

(A)

... ;.--..".,...,. ".--; (B) ,...-..,.".

(C)

(D)

(E)

(F)

FIGURE 1 5.20 Regeneration in the larval forelimb of the spotted salamander Ambystoma maculatum. (A) Longitudinal section of the upper arm, 2 days after amputation. The skin and muscle (M) have retracted from the tip of the humerus. (B) At 5 days after amputation, a thin accumulation of blastema cells is seen beneath the thickened epidermis, where the apical ectodermal cap (AEC) forms. (el At 7 days, a large population of mitotically active blastema cells lies distal to the humerus. (0) At 8 days, the blastema elongates by mitotic activity; much dedifferentiation has occurred. (El At 9 days, early redif­ ferentiation can be seen. Chondrogenesis has begun in the proximal part of the regenerating humerus, H. The letter A marks the apical mesenchyme of the blastema, and U and R are the precartilaginous condensations that will form the ulna and radius, respectively. P represents the stump where the ampu­ tation was made. (F) At 1 0 days after amputation, the precartilaginous condensations for the carpal bones (ankle, C) and the first two digits (D1, 02) can also be seen. (From Stocum 1 979, courtesy of D. L Stocum.)

POSTEMBRYONIC DEVELOPMENT

(A)

563

i.\age transplantc': d Cal:t

GFP-expressing limb (E)

Graft (e)

-

Wild-type limb

Ampl1tation

Regenerate

B

�

Fa[e? Positional identity?

Blastema

e

l ast m a

Regenerate

FIGURE 15.21 Blastem'a cells retain th ei r spec i fi ca­ tion, even though they dedifferentiate. (A,B) Schemat­ ic of the procedure, wherein a pa rti c ular tissue (in this case, carti lage) is t ra nsplanted from a salamander e p ess i ng green fluorescent protein (GFP) transgene into a wild-type salamander limb. Later, the limb is x r

a

amputated through the region of the limb containing GFP exprcssion, and a blastema is formed containing GFP-expressing cel ls that had been cartilage p recur­ sors. The regenerated limb th en studied to sec if GFP is found only in the regenerated cartilage tissues or in other tissues. (C) Longitudinal section of a regen­ erated limb 30 d ay after amputation. The muscle cells

is

s

are stained red, and nuclei are stained blue. The majority of GFP cells (green) were found in the regen­ erated ca rtilage; no GFP was seen in the muscle. (After Kragl et al. 2009, c u rtesy of E. Tanaka.)

o

found to contain CFP-expressing cells, and when the blastema differentiated, the only CFP-expressing cells found were in the limb cartilage. Similarly, CFP-marked

gradient protein (nAG). This protein can cause blastema

( F igure 15.22; 11 activated nAG genes are electopo­

muscle cells gave rise only to muscle, and GFP-marked

cells to proliferate in culture, and it permits nonmal regen­

epidermal cells only produced the epidermis of the regen­

eration in limbs that have been denervated

erated limb.

Kumar et al. 2007a).

rated into the dedifferentiating tissues of limbs that have

Proliferation of the blastema cells: The requirement for nerves and the AEC

been denervated, the limbs are able to regenerate.

If nAC is

not administered, the limbs remain as stumps. Moreover, nAC is only minimally expressed in normal limbs, but it

The growth of the regeneration blastema depends on the

is induced in the Schwann cells that surround the neurons

presence of both the apical ectodennal cap and nerves. The

within

AEC stimulates the growth of the blastema by secreting

5 days of amputation.

The creation of the amphibian regeneration blastema

ridge does in normal

may also depend on the maintenance of ion currents driv­

limb development), but the effect of the AEC is only pos­

en through the stump: if this electric field is suppressed,

FgfB (just as the apical ectodermal

sible if nerves are present (Mullen et aJ. 1996). Singer (1954)

the regenera tion blastema fails to form (Altizer et al. 2002).

demonstrated that a minimum number of nerve fibers

Such fields have been shown to be necessary for the regen­

must be present for regeneration to take place. The neu­

eration of tails in the frog

rons are also believed to release factors necessary for the

an,phibian). The

Xenopus laevis

(an anuran

Xenopus tadpole regenerates its tail, and

proliferation of the blastema cells (Singer and Caston 1972;

the notochord, muscles, and spinal cord each regenerate

Mescher and Tassava 1975). There have been many candi­

from the corresponding tissue n i the stump (Deuchar 1975;

dates for such a nerve-derived blastema mitogen, but the one that is probably the best candldate is newt anterior

Slack et aI. 2004). In this frog, the V-ATPase proton pump is

activated within 6 hours after tail amputation, changing

564

CHAPTER 1 5

Right limb denervated

(A)

,

Both limbs amputated

,

Electroporation with nAG

,

--'---' -'---...; . �

-

7 days

5 days

Regeneration

FIGURE 15.22 Regeneration of newt limbs depends on nAG Inormally supplied by the limb nerves). (A) Schematic of the pro· cedure. The limb is denervated and a week later is amputated. After 5 five days, nAG is electroporated into the li mb blastema. (B) Results show that in the denervated control {not given nAG), the amputated li mb (ye ll ow star) remains a stump. The limb that is given nAG regenerates tissues and proximal-

"

�

E E

� V>

�

Second year

�

"

�

'"

c

c ·"V>

§, � �

V>

"

'"

ro �

Third year

� c

'j

CHAPTER 1 6

604

"!

Germinal vesicle

t

Progesterone

Di plotene block (no active MPF)

c-mes

Active MPF

)g::

t .

Ca2+ flux � Ferlilization CSF Metaphase block

p 34

Cydin

Inactive MPF

Calmodu l m

jI

--

--

Meiotic arrest (First meiotic prophase)

First meiotic metaphase

Second meiotic metaphase

,

FIGURE 16.22 Schematic representation of Xenopus oocyte mat­ u ration sh owing the regu l ation of meiotic cell division by proges­ terone and fertil ization. Oocyte maturation is arrested at the

d i plotene stage of first meiotic prophase by the lack of active MPF. Progesterone activates the production of the (-mDS protein . This protein initiates a cascade of phosphorylation that eventual l y p hosp hory lates the p34 subunit of MPF, allowing the MPF to become active. The MPF drives the cell cycle through the first meiotic division, but further division is blocked by eSF, a com-

n,e mediator of the progesterone signal is the c-mos

protein. Progesterone reinitiates meiosis by causing the egg to polyadenylate the maternal c-mos mRNA that has been stored

n i

,

Calpain II ---'� CSF degraded Cam- PKlI

its cytoplasm (Sagata et al. 1988; Sheets et aL 1995;

Mendez e t aL 2000). This message is translated into a 39-

Fertilization. completion of meiosis II

pou nd conta ini ng c-mos, cyclin-dependent kinase 2, and Erp l . CSF inhibits the anaphase-promoting comp lex from degrading cyel in . Upon fertilization, calcium ions released into the cy to­ plasm are bound by calmodulin and are used to activate WO enzymes, calmod u l i n-dependen t prote in kinase I I and calpain I I , which inactivate a n d degrade (SF. Second meiosis i s comp leted , and the two haploid pronuclei can fuse. At this time, cyc li n B is resynthesized, allowing the first cel l cycl e of cleavage to begin.

the first meiotic division. The proteins o f the CSF complex

interact, eventually activating Erpl by phosphorylating it. the anaphase-promoting comp l ex

Phosphorylated Erpl blocks the degradation of cyclin by

(Figure 1 6.23).

This metaphase block is broken by fertilization. The cal­

kDa phosphoprotein. This c-mos protein is detectable only

cium ion flux attending fertilization activates the ca1cium­

fertilization. Yet during its brief lifetime, it plays a major cole n i releasing the egg from its dormancy. The c-mos pro­ tein activates a phosphoryla tion cascade that phosphory­ lates and activates the p34 subunit of MPF (Ferrell and

binding protein calmodulin, and calmodulin, in tum, can

during oocyte maturation and is destroyed quickly upon

Machleder 1998; Ferrell 1999). The active MPF allows the

c-mos is inhibited by injecting

activate two enzymes that inactivate CSF. These enzymes

II, a cal ci um-depend en t protease (Watanabe et aL 1989; Lorca et aL 1993). This action promotes cell division in two ways. First,

are calmodulin-dependent protein kinase II, which inac ti­ vates cdk2, and calpain that degrades c-mos

germinal vesicle to b rea k down and the chromosomes to

withoutCSF, cyclID can be degraded, and the meiotic divi­

c�mos an tisense mRNA into the

sion can be completed . Second, calcium-dependent pro­

divide. If the translation of

breakdown and the resump tion of oocyte maturation do oocyte, germinal vesicle

not occur.

However, oocyte maturation then encounters a second block. MPF can take the chromosomes only throu gh the division. The oocyte is arres ted once again in the

first meiotic division and prophase of the second meiotic

tein kinase II also allows the centrosome to duplicate, thus forming the poles of the meiotic spindle (Matsumoto and Maller 2002). ln 1911, Frank Lillie wrote, "The nature of the inhibition that causes the need for fertilization is a most

fundamental problem." The solution to tha t p roble m appears to be oocyte-derived CSF and the sperm-induced wave of calcium ions.

metaphase of the second meiotic division. This metaphase CSF is a complex of proteins that includes c-mos, cyclin­

block is caused by cytostatic factor (CSF; Ma tsui 1 974).

Gene transcription in amphibian oocytes

dependent kinase 2 (cdk2), MAP kinase, and ErpI (Gabriel­

The amph ib ian oocyte has certain periods of very active

Erp1

RNA synthesis. During the diplotene stage, certain chro­

is the active protein, and it is synthesized immediately after

mosomes stretell out large loops of DNA, causing them to

li et aL 1993; lnoue et aL 2007; Nishiyama et aL 2007) .

CSF c

om

T H E SAGA O F T H E GERM L I N E

plex

c- m as

FIGURE 16.23 The main pathway l ead i ng to metaphase arrest in the second meiotic division. The (SF protein complex consists of c-mos, three transducer kinases, and the effector protein Erp1. Activation of c-mos activates the kinases, which eventually phos­ phorylate Erp l . Phophorylated Erp! binds to and inhibits the anaphase-pro oting com­ plex, thus blocking the deg radation of cyclin B that would allow the cell to enter anaphase. (After Inoue et al. 2007.)

MEK

I MAPK I t ce90rs0 t

Anaphase­ promoti complexng Cyetin Cde2

m

are being transcribed (Figure 1 6.24A). Electron micrographs of gene transcripts from lampbrush chromosomes also enable one to see chains of mRNA coming off each gene as it is transcribed (Figure 16.24B; also see Hill iJl1d MacGregor

� 1

Metaphase arrest

B

Metaphase II

Anaphase 11

resemble a lampbrush (which was a handy instrument for cleaning test tubes in the days before microfuges). In situ hybridization reveals these lampbrush chromosomes to be sites of RNA synthesis. Oocyte chromosomes can be incubated with radioactive RNA probe and autoradiog­ raphy used to visualize the precise locations where genes

a

(A)

605

1980).

In addition to mRNA synthesis, ribosomal RNA and transfer RNA are also transcribed during oogenesis. Fig­ ure 1 6.25A shows the pattern of rRNA and tRNA synthe­ sis during Xenopus oogenesis. Transcription appears to begin in early (stage I, 25-40 �.m) oocytes, during the diplotene stage of meiosis. At this time, all the rRNAs and tRNAs needed for protein synthesis until tile mid-blastula stage are made, and all the maternal mRNAs needed for early development are transcribed. This stage lasts for months in Xenopus. The rate of rRNA production is prodi­ gious. The Xenopus oocyte genome has over 1800 genes encoding 185 and 285 rRNA (the two large RNAs that form the ribosomes), and these genes are selectively amplified such that there are Over 500,000 genes making rRNA in the oocyte (Figure 1 6.25B; Brown and Dawid 1968). When the mature (stage Vl) oocyte reaches a certain size, its chromo-­ somes conden�e, and the rRNA genes are no longer tran-

(B)

FIGURE 16.24 In amphibian oocytes, lampbrush chromosomes are active in the diplotene germinal vesicle during first meiotic p rop hase. (A) Autoradio­ graph of ch romosome I of the newt Tritu­ rus crislalUs after in situ hybridization with radioactive histone mRNA. A his­ tone gene (or set of histone genes) is being transcribed (arrow) on one o t e loops of this lampbrush (B) Lampbrush romos m ma nder NOlOphlhalmu5 viridescens. Extended DNA (w te) transcribed i to (A

fh chromosome. ch o e of the sala­ hi loops out and is n RNA (red). from Old et al. 1 977, courtesy of H . G. Callan; B co u rtesy of M. B. Roth and J. GaiL)

606

CHAPTER 1 6

In and moths) that undergo meroistic oogenesis, in which cytoplasmic connections remain between the cells pro­ duced by the oogonium. The oocytes of meroistic insects do not pass through a transcriptionally active stage, nor do they have lampbrush chromosomes. Rather, RNA synthesis is largely confined to the nurse cells, and the RNA made by those cells is actively transported into the oocyte cytoplasm (see Figure 6.7). Oogenesis takes place in only 12 days, so the nurse cells are metabolically very active during this time. Nurse cells are aided in their transcriptional efficiency by becom­ ing polytene-instead of having two copies of each chro­ mosome, they replicate their chromosomes until they have produced 512 copies. The 15 nurse cells pass ribosomal and messenger RNAs as well as proteins into the oocyte cyto­ plasm, and entire ribosomes may be transported as well. The mRNAs do not associate with polysomes, and they are not immediately active in protein synthesis (Paglia et a1. 1976; Telfer et a1. 1981). The meroistic ovary confronts us with some interesting problems. If all 16 cystocytes derived from the PGC are connected so that proteins and RNAs can shuttle freely among them, how do 15 cystocytes become RNA-produc­ ing nurse cells while one cell is fated to become the oocyte? Why is the flow of protein and RNA in one direction only? As the cystocytes divide, a large, spectrin-rich structure called the fusome forms and spans the ring canals between the cells (see Figure 16.4A). It is constructed asynunetrical­ Iy, as it always grows from the spindle pole that remains in one of the cells after the first division (Lin and Spradling 1995; de Cuevas and Spradling 1998). The cell that retains the greater part of the fusome during the first division becomes the oocyte. It is not yet known if the fusome con­ tains oogenic determinants, or i f it directs the traffic of materials into this particular cell.

scribed. This "mature oocyte" condition can also last for months. Upon hormonal stimulation, the oocyte completes its first meiotic division and is ovulated. The mRNAs stored by the oocyte now jOin with the ribosomes to initiate protein synthesis. Within hours, the second meiotic divi­ sion has begun, and the egg is fertilized in second meiot­ ic metaphase. The embryo's genes do not begin active tran­ scription until the mid-blastula transition (Newport and Kirschner 1982). As we saw in Chapter 2, the oocytes of several species make two classes of mRNAs-those for immediate use in the oocyte, and those that are stored for use during early development. In frogs, the translation of stored oocyte mes­ sages (maternal mRNAs) is initiated by progesterone as the egg is about to be ovulated. One of the results of the MPF activity induced by progesterone may be the phosphoryla­ tion of proteins on the 3' urn of stored oocyte mRNAs. The phosphorylation of these factors is associated with the lengthening of the polyA tails of the stored messages and their subsequent translation (Paris et a1. 1991).

Meroistic oogenesis in insects There are several types of oogeneSis in insectsi but most studies have focused on those insects (including Drosophi-

(A)

�

High

tRNA

�

.� � .c

Ribosomal RNAs (IB5 and 285)

� 0

DNA

.'l • -

.�

;; 0; '"

Low

55 rRNA

�!

:��

cations

tI

Accumulation of yolk starts

�L

grown oocyte

__ __ __

Transcription

(B)

==}�!.'...;,.':" -'

"\")' (VI� -';:� � -:'�f(� :,.,.. ' ,: t '; ;< .:H

.

�'-

l

Fertilization

� �� '�

Y -y Oocyte growth 3 months

L__ __ __ __

one

Fully

�,

16

FIGURE 16. 25 po ocyte . mph bian oogenesis u ng the as 3 t (8) T a cr i ti the large RNA p ec e e er, hap o d genome. (A aher O. L. M l er,

Ribosomal and transfer RNA r duction in Xeno­ o s (A) Relative rates of DNA, tRNA, and rRNA synthesis in a i d ri l t mon hs before ovula­ tion. r ns p on of r ursor of the 285, 1 85, and 5.85 ribosomal RNAs. These units ar tand mly linked togeth­ with some 450 per l i Gurdon 1 976; 8 courtesy of i l Jr.)

pus

Maturation months

\

Transcription of ribos mai RNA

Transcription

Nontranscribed

T H E SAGA OF T H E G ERM L I N E

607

TABLE 1 6.2 Sexual dimorphism in mammalian meioses

Female oogenesis

Male spermatogenesis

One gamete produced per meiosis

FOllr gametes produced per meiosis

Meiosis initiated once in a finite population of cells Completion of meiosis delayed for months or years

Meiosis initiated continuously in a mitoticall y dividing stem cell population Meiosis completed in days or weeks

Meiosis arrested at first meiotic prophase and

Meiosis and differentiation proceed continuously without ceU cycle arrest

Differentiation of gamete occurs while diploid, in first meiotic prophase

Differentiation of gamete occurs while haploid, after meiosis ends

All chromosomes exhibit equivalent transcription and recombination during meiotic prophase

Sex chromosomes excluded from recombination and transcription during first meiotic prophase

reinitiated in a smaller pOPlllation of cells

Source: Handel nnd Eppig 19Y8.

Once the patterns of transport are established, the

Gametogenesis i n Mammals

cy toskeleton becomes actively involved in transporting

As outlined in Table 1 6.2, there are profound differences

Tbeurkauf 1994). An array of microtubules

between spermatogenesis and oogenesis in mammals. One

mRNAs from the nurse cells into the oocyte cytoplasm (Cooley and

that extends through the ring canals (see Figure 16.4C) is

of the fundamental differences concerns the timing of

meiosis onset. In females, meiosis begins in the embryon­

critical for oocyte determination. In the nurse cells, the

ic gonads; in males, meiosis is not initiated until puberty.

Ex uperantia protein binds

This critical difference in timing is due to

bicoid message to the micro­

If the microtubular array is dis·

tubules and transports it to the anterior of the oocyte (Cha et a1. 2001; see Chapter 6).

rupted (either chemically or by mutations such as bicau·

dal-D Or egalitarian), the nurse cell gene products are trans­

retinoic acid (RA)

produced by the mesonephric kidneys (Figure 1 6.26). This

RA stimulates the germ cells to undergo a new round of DNA repl ic a ti on and initiate meiosis (Baltus et al. 2006;

Bowles et .1. 2006; Lin et al. 2008). In males, however, the

mitted in all directions and a11 1 6 cells d ifferentiate into

embryonic testes secrete the RA-degrading enzyme

nurse cells (Gutzeit 1986; Theurkauf e t a l . 1992, 1993;

Cyp26bl . This prevents RA from promoting meiosis. Later,

Spra dl ing 1993).

Nanos2 will be expressed in the male germ cells, and this

The Bicaudal·D and Egalitarian proteins are probably

will also prevent meiosis and ensure that the cells follow

core components of a dynein motor system that transports

the pa thway to become sperm (Koubova et al. 2006; Suzu·

mRNAs and proteins throughout Ule oocyte (Bullock and

ki and Saga 2008).

Ish·Horowicz 2001). It is possible that some compounds transpor ted from the nurse cells into the oocyte become

VADE MECUM Gametogenesis in mammals

associated with transport prote�ns such as dynein and kinesin, which would enable them to travel along the tracks of microtubules extending through the ring canals

oskar for instance, is linked to kinesin through the Bar· entsz protein, and kinesin can transport the oskar message to the posterior of the oocyte (van Eeden et a1. 2001; see

Spermatogenesis

(Theurkauf et a1. 1992; Sun a n d Wyman 1993). The

Once mammalian PGCs arrive a t the genital ridge of

message,

male embryo, they are called gonocytes and become incor·

Figure 6.7).

Actin may become important for maintaining the polar·

a

porated into the sex cords (Culty 2009). They remain there

until maturity, at which time the sex cords hollow out to form the

seminiferous tubules. The epithelium of the

tubules differentiates into the Serto1i cells that will nOUIish

ity of transport during later stages of oogenesis. Mutations

and protect the developing spenn cells. TIle gonocytes dif·

that prevent actin microfilaments from lining the 'ring

ferentiate into a population of stem ceUs that have recent·

canals prevent the transPOlt of mRNAs from the nurse cells

ly been named the

to the oocyte, and disruption of the actin microfilaments

nia (Yoshida e t a l . 2007). These cells can reestablish

Wa tson et al. 1993). Thus, the cyt oskeleton controls the

randomizes the distribution of mRNA (Cooley et a1. 1992;

sperma togenesis when transferred into mice whose sperm

movenlent o f organelles and RNAs between nurse cells

production was eliminated by toxic chemicals. TIley appear

only in the appropriate direction.

blood vessels.

and oocyte such

that developmental cues are exchanged

und

i ffe ren ti ate d

type

A sperm atogo·

to reside in stem cell niches created by the junction of Ser­ toli cells, interstitial (testosterone·producing) cells, and

608

CHAPTER 1 6

FIGURE 16.26 Retinoic acid (RA) (A) Female germ ceUs determines the ti m ing of meiosis and sexu al di fferenti ation of m am m alian Nanos2 added Normal germ cells. (A) In fem al e mouse MesoRA RA embryos, RA secreted from the mesonephros reaches the gonad and triggers meiotic initiation via the induction of Stra8 transcription fac­ tor in female germ cells (beige). StraS .....- Nanos2 Stra8 However, if activated Nanosl genes are added to female germ ce li s they suppress StraB expression, leading the germ ce lls i n to a mal e pathway (gray) . (B) In embryonic testes Meiosis Meiosis Cyp26bl blocks RA signal ing there­ Gonad by preventing male germ cells from Male fate initiating meiosis until embryonic day 1 3 .5 (l eft pa nel). After embryon(e) RA synthesized ic day 1 3 .5, when Cyp26bl expression is decreased, Nanos2 is expressed and pre­ vents meiotic initiation by blocking Stra8 expressi on. This induces male-type differen­ tiation in the germ cells (right panel) . (C,D) Day 1 2 mouse embryos stained for mRNAs encoding the RA-synth esi zing enzyme Aldhla2 (C) and the RA-degrading enzyme Cyp26bl (D). The RA-synthesizing enzyme is seen in the mesonephros of both the male and female; the RA-degrading enzyme is seen only in the male gonad. (A,S (rom Saga 2008; C,D from Bow les et al. 2006, co urtesy of P. Koopman.) ,

, ,

=C:mJ m l mnnnm

j

t t

Male

The decision to proliferate or differentiate may involve interactions between the Wnt and BMP pathways. Wnt sig­ naling appears to promote proliferation of stem ceils, and the spermatogonia appear to have receptors for both Wnts and BMPs (Golestaneh et aJ. 2009). The initiation of sper­ matogenesis during puberty is probably regulated by the synthesis of BMPs by the spermatogenic germ ceils, the spermatogonia. When BMP8b reaches a' critical concentra­ tion, the germ cells begin to differentiate. The differentiat­ ing ceils produce high levels of BMP8b, which can then further stimulate their differentiation. Mice lacking BMPSb do not initiate spermatogenesis at puberty (Zhao et aJ. 1996). The spermatogenic germ cells are bound to the Sertoli cells by N-cadherin molecules on the surfaces of both cell types, and by galactosyltransferase molecules on the sper­ matogenic cells that bind a carbohydrate receptor on the Sertoli cells (Newton et aJ. 1993; Pratt et aJ. 1993). Sper­ matogenesis-the developmental pathway from germ cell to mature sperm--occUIs in the recesses between the Ser­ tali ceils ( F i gu re 1 6.27). FORMING THE HAPLOID SPERMAT I D The undifferentiated

type Aj spermatogonia (sometimes called the dense type

Female

(B) Male germ ceUs Before day 13.5 Meso­ RA nephros Somatic Germ line

Gonad

After day 13.5 RA

f- eyp26bl

Stra8

t

>- eyp26bl

StraB.....- Nanos2

Me i osis

i

M

(D) RA degraded

j

Male fate

Male Female

A spermatogonia) are found adjacent to the outer basal lamina of the sex cords. They are stem cells, and upon reaching maturity are thought to divide to make another type Aj spermatogonium as well as a second, paler type of cell, the type A2 spermatogonia. The A2 spermatogonia divide to produce type A3 spermatogonia, which then beget type A4 spermatogonia. The A4 spermatogonia are thought to differentiate into the first committed stem cell type, the intermediate spermatogonia. Intermediate sper­ matogonia are committed to becoming spermatozoa, and they divide mitotically once to form type B spermatogo­ nia (see Figure 16.27). These cells are the precursors of the spermatocytes and are the last cells of the line that under­ go mitosis. They divide once to generate the primary sper­ matocytes-the cells that enter meiosis_ The transition between spermatogonia and spermato­ cytes appears to be mediated by the opposing influences of glial cell line-derived neurotrophic factor (GDNF) and stem ceil factor (SCF), both of which are secreted by the Sertoli cells. GDNF levels determine whether the dividing spermatogonia remain spermatogonia or enter the path­ way to become spermatocytes. Low levels of GDNF favor the differentiation of the spermatogonia, whereas high le,-­ els favor self-renewal of the stem cells (Meng et aJ. 2000).

T H E SAGA O F T H E GERM L I N E

Type B spermatogonium

Vas deferens

609

A, um spermatogoni Type

spermatogonium

spermatocyte

spermatocyte

seminiferous tubule

FIGURE 1 6.27

Section of the seminiferous tubule, showing the relationship between $ertol i cells and the developing sperm. As germ cells mature, they progress toward the lumen of the seminiferous tubule. (Mer Dym ' 9 77.)

SCF promotes the transition to spermatogenesis

(Rossi et upregulated by

still connected to one another through their cytoplasmic

al. 2000). Since both GDNF and SCF are

bridges. The spermatids that are connected in this manner

follicle-stimulating hormone (FSH), these two factors may

have haploid nuclei but are functionally diploid, since a

serve as a link between the Sertoli cells and the endocrine

gene product made in one cell can readily diffuse into the

the testes to produce more sperm (Tadokoro et al. 2002).

During the divisions from type A, spermatogonia to

Keeping the stem cells in equilibrium-producing neither

spermatids, the cells move farther and farther away from

too many undifferentiated cells nor too many differentiat­

the basal larnina

system, and they provide a mechanism for FSH to instruct

ed cells-is not easy. Mice with the Iuxoid mutation are ster­ ile because they lack a transcription factor that regulates

cytoplasm of its neighbors (Braun et a!. 1989).

of the seminiferous tubule and closer to Figure 16.27; Siu and Cheng 2004). Thus, each type of cell can be found in a particular layer of the its lumen (see

this division. All their spermatogonia become sperm at

tubule. The spermatids are located at the border of the

once, leaving the testes

lumen, aJ1d here they lose their cytoplasruic connections

devoid of stem cells (Buaas et aJ.

2004; Costoya et al. 2004). Looking

and differentiate into spermatozoa.

at Figure 1 6.28, we find that during the sper­

matogonial divisions, cytokinesis is nC?t complete. Rather, the cells form a syncytium in which each cell communi­

In humans, the pro­

gression from spermatogonial stem cell to mature sperma­ tozoa takes 65 days (Dyrn 1994). The processes of spermatogenesis require a very

spe­

cates with the others via cytoplasmic bridges about 1 �.m

cialized network of gene expression (Sassone-Corsi 2002).

in diameter (Dym and Fawcett 1971). The successive divi­

Not only are histones substantially

sions produce clones of interconnected cells, and because

b y sperm-specific

ions and molecules readily pass through these cytoplas­

RNA polymerase n transcription factors are exchanged for

mic bridges, each cohort matures

synchronously. During

this time, the spermatocyte nucleus often transcribes genes whose

products will be used later to form the axoneme and

acrosome.

remodeled and replaced variants (see below), but even the basal

sperm-specific variants. The TFIID complex, which con­

tains the TATA-binding protein and 14 TAFs, functions in the recognition of RNA polymerase.

One of these TAFs,

TAF4b, is a sperm-specific TAF required for mouse sper­

Each primary spermatocyte undergoes the first meiot­

matogenesis (Falender et a!. 2005). Without this factor, the

ic division to yield a pair of secondary spermatocytes,

spermatogonial stem cells faU to make Ret (the receptor for

which complete the second division of meiosis. The hap­

GDNF) or the

loid cells thus formed are called spermatids, and they are

genesis fails to occur.

luxoid

transcription

factor, and spermato­

610

CHAPTER 1 6

SPERMIOGENESIS: DIFFERENTIATION OF THE SPERM The mammalian haploid spermatid is a round, unflagellated cell that looks nothing like the mature vertebrate sperm. The next step in sperm maturation, then, is spermiogenesis (or spermateliosis), the differentiation of the sperm cell. For fertilization to OCCUI, the sperm

has to meet and bind with an egg, and spermiogenesis prepares the

sperm for these functions of motility and interaction. The process of mam­

Type Aj

spermatogon ia

Type A2 spermatogonia

or �

� More type A I � spermatogonia

Type A3

spermatogonia Type A, spermatogonia Intermediate spermatogonia

Type B spermatogo nia

malian sperm differentiation was shown in Figure 4.2. The first step is the construction of the acrosomal vesicle from the Golgi apparatus. The

Secondary spermatocytes (2nd meiotic divisio n )

acrosome forms a cap that covers the sperm nucleus. As the acrosomal cap is formed, the nucleus rotates so that the cap will be facing the basal lam­ ina of the seminiferous tubule. This rotation is necessary because the fla­ gellum, which is beginning to form from the centriole on the other side of the nucleus, will extend into the

Spermatids

lumen. During the last stage of spermiogenesis, the nucleus flattens and condenses, the remaining cyto­ plasm (the residual body, or "cyto­ plasmic droplet") is jettisoned, and the mitochondria form a ring arowld the base 01 the flagellum. During spermiogenesis, the his-

-���

••••••••••o•••••

tones of the spermatogonia are often replaced by histone

variants, and widespread nucleosome dissociation takes place. This remodeling of nucleosomes might also be the

point at which the PGC pattern of methylation is removed and the maJe genome-specific pattern of methylation is

��

sperm cellS

FIGURE 16.28 Format i on of syncytial clones of human male germ cells. (After Bloom and Fawcett 1 975.)

established on the sperm DNA (see Wilkins 2005). As spermiogenesis ends, the histones of the haplOid nucleus are eventually replaced by protamines.' This replacement

In the mouse, development from stem cell to sperma­

results in the complete shutdown 01 transcription in the

tozoon takes 34.5 days: the spermatogonial stages last 8

nucleus and facilitates the nucleus a5suJ!ling an almost

days, meiosis lasts 13 days, and spermiogenesis takes

crystalline structure (Govin et aJ. 2004). The resulting

another 13.5 days. Hwnan sperm development takes near­

sperm then enter the lumen of the seminiferous tubule.

ly twice as long. Because the original type A spermatogonia

See WEBSITE 1 6.5 The Nebenkern

are stern cells, spermatogenesis can occur continuously. Each day, some 100 million sperm are made in each hwnan testicle, and each ejaculation releases 200 million spenn.

·Protamines are relatively small proteins that are over 60%, argi­ nine. Transcription of the genes for protamines is seen in the early haploid spermatids, although translation is delayed for several days (Peschon et al. 1987). The replacement, however, is not com­ plete,

Normal blood cells

�

Leukemia progenitor

Leukemia cells

FIGURE 17.18 Model of cancer stem cell production, using leukemia (a white blood cell tumor) as an exampl e. (A) A hematopoietic stem cel l (HSC) usually gives rise to normal blood progenitor cells that can become mature white blood oells. Microenvironment (8) If the HSC undergoes mutations or epigenetic changes involving gene activation it can become a cancer stem cell (esC) that can divide to produce more of itself plus other relatively differentiated (leukemic) cells. As in normal blood devel op­ ment, the CSC reta in s the abil ity for self-renewal and thereby becomes the malig­ nant portion of the cancer. IC) The CSC may also be produced by changes in the microenvironment, which allows cert.ain cells to display a stem cel l phenotype that they would not otherwise possess. (After Rosen and Jordan 2009.) ,

MEDICAL ASPECTS O F DEVELOPMENTAL BIOLOGY

dividing stem cell population that gives rise to more can­ cer stem cells and to populations of relatively slowly divid­

ing differentiated cells (Lapidot et aI.

1994; Bonnet and Dick 1997; Singh et a1. 2004; Schatton 2008). Whether the tumor

is initiated by an adult stem cell "gone bad" or by a more differentiated cell that has regained stem cell abilities is a

645

ed. Thus, one might get cancer if faulty methylation either

inappropriately methylated the tumor suppressor genes (turning them off) or inappropriately demethylated the

oncogenes (turning them on;

Figure 1 7.19).

Some genes may be oncogenes in one set of cells and

tumor suppressor genes in another set of cells. In the

2009; Schatton et aI. 2009). In some tumors, such as prostate

breast, for instance, estrogen receptors can act as oncogenes

stem cell that has escaped the control of its niche (Wang e t

receptors function as tumor-suppressor genes. Issa and col­

matter of controversy (Gupta et a1. 2009; Rosen and Jordan

cancer, the origin of the tumor is most Ukely a normal adult

for estrogen-dependent breast cancer. In the colon, how­

ever, estrogen stops the proliferation of cells, and estrogen

aI. 2oo9).

leagues

(1994) showed that in addition to the age-associ­

Cancer as a return to embryonic invasiveness: Migration reactivated

genes in colon cancers. Even the smallest colon cancers had

Another crucial point in considering cancers as diseases of

of the estrogen receptor gene.

of the malignant cell into other tissues. Like embryonic

genetic cause. Indeed, several studies indicate that these

ated methylation of estrogen receptors, there was a much

disrupted development involves metastasisi the invasion

cells, tumor cells do not usually stay put-they migrate and form colonies. Chapter 3 discussed the roles of cad­

herin proteins in the sorting-out of cells to form tissues dur­ ing development, and how cells form boundaries and seg­

regate into tissues by al tering the strengths of their

higher level of DNA methylation nearly

in the estrogen receptor

100% methylation of the cytosines in the promoter

The epigenetic causation of cancer does not exclude a

mechanisms augment one another. Numerous mutations

occur in each cancer celi, and recent evidence suggests that

as many as

14 Significant tumor-promoting mutations are 2006). Jacinco and Esteller (2007) have presented evidence that the large num­

found in each cancer cell (Sjoblom et aJ.

In cancer metastasis, this property is lost: cad­

ber of mutations that accwnuJate in cancer cells may have

ment to the extracellular matrix and other types of cells

itself from mutations. One is the editing subunits on DNA

attachments.

herin levels are downregulated, and the strength of attach­

becomes greater than the cohesive force binding the tissue together. As a result, the cells become able to spread into

other tissues (Foty and Steinberg

1997, 2004).

Another phase of metastasis involves the digestion of

extracellular matrices by metalloproteinases. These

enzymes are used by migrating embryonic cells to digest a path to their destination. They are commonly secreted by

an epigenetic cause. DNA has several means of protecting

polymerase; these "proofreaders" get rid of mismatched

bases and insert the correct ones. Another mechanism is the

set of enzymes that repair DNA when the DNA has been

damaged by light or by cellular compounds that are prod­ ucts of metabolism. In cancer cells, the genes encoding these DNA repair enzymes appear to be susceptible to inactiva­

tion by methylation. Once DNA repair enzymes have been

trophoblast cells, axon growth cones, sperm cells, and

dowmegulated, the number of mutations increases.

malignant cancer cells, allowing the cancer to invade other

linked by the common denominator of aberrant DNA

somitic cells. Metalloproteinases can be reactivated

in

tissues. The presence of these enzymes is a marker that the

tumor is particularly dangerous (see Gu et a1.

2005).

Cancer and epigenetic gene regulation In Chapter

15, we saw evidence that the methylation pat­

terns of mammalian genes change with age. We specifical­

It is therefore possible that aging and cancer may be

methylation. If metabolically or structurally important

genes (such as the estrogen receptors) become heavily

methylated, they don't produce enough receptor proteins,

and our body function suffers. If tumor suppressor genes

or the genes encoding DNA repair enzymes are heavily methylated, tumors can arise.

Tumors can be generated by a combination of genetic

ly looked at genes that might cause elements of the aging

and epigenetic means. Changes in DNA methylation can

dependent patterns of gene methylation altered the tran­

thereby initiating tumor formation. Conversely, oncogenes

Two types of genes control cell division. The first are

which also aids tumorigenesis. Moreover, the tissue envi­

sion, and prevent cell death. These are the genes that can

processes. The complexities of tumors, including their mul­

phenotype. But what would happen if the random, age­ scription of the genes regulating cell division?

oncogenes, which promote cell division, reduce cell adhe­

activate oncogenes and repress tumor-suppressor genes,

can cause the methylation of tumor suppressor genes,

ronment of the cell may be critical in regulating these

promote tumor formation and metastasis. The second set

tiple somatic mutations and their resistance to agents that

genes usually put the brakes on cell division and increase

combination of genetic and epigenetic factors rather than

of regulatory genes are the tumor suppressor genes. These

induce apoptotic cell death, may best be explained by a

the adhesion between cells; they can also induce apopto­

just by the basis of mutations. Knowledge of the epigenet­

and tumor suppressor genes has to be very finely regulat-

of cancer therapy.

sis of rapidly dividing ceUs. The interplay of oncogenes

ic causes of cancer can provide the basis for new methods

CHAPTER ·1 7

646

(A)

(B)

NORMAL CELL

NORil4AL CELL

/

Methylated

Unmethylaled CpG

\

()

Tumor-suppressor gene: Open chromatin conformation

CANCER CELL

c

J

)

pG

I It ) )

1 1 I )) )

II II I It I )

t:)) 1 )

11n

Oncogene:

Repressed chromatin conformation

CANCER CELL Hypomethylated

J

)

I

I

CpG

\ G:I)

1 1 1 Ci )

I

) 1

Repressed chromatin conformation on tumor-suppressor gene

Open chromatin conformation on oncogene

CeU cycle entry

CeU cycle entry

Blocking of apoplosis

Inappropriate gene expression

DNA repair deficits

Loss of cell adhesion

_ _ _ _ .. . L�

I

Tumor formation

I

Loss of dosage regulat ion

FIGURE 1 7.19 Cancer can arise (A) i( tumor-suppressor genes are inappropriately turned off by DNA methylation or (8) if onco­ genes are inappropriatel y demethylated (and thereby activated). (After Esteller 2007.)

SIDEL IG H TS SPECULATIONS

The Embryonic Origins of Adu lt-Onset Illnesses

T

eratogenesis is usually associated with congenital disease ( i .e., a condition appearing at birth) and is also associated with disruptions of organogenesis during the embryonic period. However, D. J. P. Ba rker and

colleagues (1 994a,b) have offered evi­ dence that certain adult-onset diseases may also result from conditions i n the uterus prior to birth. Based on epi­ demiological evidence, they hypothe­ size that there are critical periods of development during which certai n physiological insults or stimuli can cause specific changes in the body.

The "Barker hypothesis" pos tulates that .certain anatomical and physiological parameters get "programmed" during embryonic and fetal development, and that deficits in nutrition during this time can produce permanent changes in the pattern of metabolic activity­ changes that can predispose the adult to particular diseases. Speciiically, Barker and colleagues showed that i nfants whose mother experienced protein deprivation (because of wars, famines, or migra­ tions) during certa i n months of preg­ nancy were at high risk for having cer-

tain d iseases as adults. Undernutrition during a fetus's first trimester could lead to hypertension and strokes i n adult life, while those (etuses experi­ enCing undernutrition during the sec­ ond trimester had a high risk of devel­ oping heart disease and diabetes as adults. Those fetuses experiencing undernutrition rluring the third trimester were prone to blood clotting defects as adults. Recent studies have tried to deter­ mine whether there are physiological or anatomical reasons for these corre­ lations (Gluckman and Hanson 2004,

MED ICAL ASPECTS O F DEVELOPMEN TA L BIOLOGY

647

SIDELIGHTS & SPECULATIONS (Co nti n ued) Figure 17.20 Anatomical changes associated with hypertension. (A) In age-m atched individuals, the kidneys of men with hypertension had about half the number of nephrons as the kidneys of men with normal blood pressure. (B) The glomerul i of the nephrons in hypertensive kid­ neys were much larger than the glomerul i in control subjects. (After Keller et al. 2003, photographs courtesy of G. Keller.) (A) 2,250,000 2,000,000 '" � �

1,750,000

v E

1,500,000

Z

1,000,000

.M�r--:- 1,429,200

0

'0

Oh 1,250,000 " v .D

E "

750,000

'"

2> ;; � v

2005; Lau and Rogers 2005). Anatom­ ically, undernutrition can change the number of cells produced during a critical time of organ formation. When pregnant rats are fed low-protein diets at certain times during their pregnan­ cy, the resulti ng offspring are at high risk for hypertension as adult. The poor diet appears to cause low nephron numbers i n the adult kidney (see Moritz et a l . 2003). In humans, the n u mber of nephrons present i n the kidneys of men with hypertension was only about half the n u mber found i n men without hypertension (Figure 1 7.20A; Keller et a l . 2003). In addition, the glomeru l i (the blood-fi ltering unit of the nephron ) of hypertensive men were larger than those in control sub­ jects (Figure 1 7.20B). Similar trends have been reported for non-insu l i n dependent (Type II) d iabetes and glucose i n tolerance (Hales et a l . '1 991 ; Hales and Barker 1 992). Here, poor nutrition reduces the number of p cells in the pancreas and hence the ability to synthesi ze insulin. Moreover, the panc reas isn't the o n ly organ involved. Undernutri­ tion in rats changes the h i stological

Matched controls

5

2

� Patients with hypertension

6

E

�

250,000

7

4

0 "

702,379

8

E

� �

500,000

0

�E E "1

3

0

Patients with hypertension

architecture in the liver as well. A low­ protein d iet duri ng gestation appeared to increase the amount of periportal cells that produce the glucose-synthe­ sizi n g enzyme phosphoenolpyruvate carboxykinase while decreasing the number of perivenous cells that syn­ thesize the glucose-degrading enzyme glucokinase in the offpsring (Burns et a l . 1 997). These changes may be coor­ di nated by glucocorticoid hormones that are stimulated by mal nutrition and which act to conserve resources, even though such actions might make the person prone to hypertension later in life (see Fowden and Forhead 2004). (Since, historical ly, most humans died before age 50, this would not be a detrimental evolution­ ary trade-off.) Hales and Barker (200 1 ) have pro­ posed a "thrifty phenotype" hypothe­ sis wherein the malnourished fetus is "programmed" to expect an energy­ defic ien t environment. The developing

Matched cont rols

fetus sets its biochemical parameters to conserve energy and store fat. * Resulting adults who do indeed meet with the expected poor environment are ready for it and can survive better than individuals whose metabolisms were set to utilize energy and not store it as efficiently. However, if such a "deprivationally developed" person l ives in an energy- and protein-rich environment, their cells store more fats and their heart and kidneys have developed to survive more stringent conditions. Both these developments put the person at risk for several later­ onset diseases. How can conditions experienced in the uterus create anatomical and bio­ chemical conditions that will be main­ tained throughout adulthood? One place to look is DNA methylation. Lil­ Iycrop and colleagues (2005) have shown that rats born to mothers having a low-protein diet had a different pat­ tern of liver gene methylation than did

*In other words, the embryo has phenotyp i c pl as tic i ty-t he ability to modulate its phe­ notype dependi ng on the environment; this plasticity w i l l be discussed further i n Chapter 1 8.

648

CHAPTER I 7

SIDELIGHTS & SPECULATIONS (Continued)

t he offsp r i ng of mothers fed a normal diet. These differences in methylation changed the metabol ic profi Ie of the rats' livers. For instance, t he methy la­ tion of the promoter region of th e PPARa gene (which is critical in the regulation of carbohyd rate and lipid me ta bo l i s m) is 20% l ower i n the off­ spring of protei n-res tri cted rats, a n d the ge n e 's tra n sc ri pti ona l activity is tenfold greater (Figure 1 7.21). More­ over, the difference between these met hylat i on patterns can be abo l ished by includ ing folic ac i d in the protein­ restricted diet. Thus, the difference in methylati on probab l y results from changes in folate meta bo lis m caused by the l imited amount of protein avail­ able to the fetus.

I t does appear that prenatal nutri­ tion can induce long-lasting, gene­ spec i fic a l terati on s i n tra nscr i p ti ona l activity and metabolism. The preven­ tion of ad u l t di sease th rou gh p ren ata l diet could thus become a p u b l i c health issue i n the co m i n g decades.

(A)

(B) 110

." � • �

= -

8� 3 '0 '';:: 1:::: � g � v � 8 = « z Q

1200 '"

�

1000

� Q,

E

100

8-

c ??- 800

.2 :::; � o

90

8 o

0 0 v 0

0 v «

12

80

E

70

0 -

-

Control

Protein� Proteinrestricted restricted + folate

600 400 200 0

Control

Pr otei n - Proteinrestricted restricted + folate

Figure 17.21 Activity of the l iver gene for peroxisomal prol iferator-activated receptor (PPARa) is susceptible to dietary differences. (A) DNA methylation pattern of the PPARa pro­ moter region, showing highly methylated control promoters compared with poorly methylat­ ed promoters from the livers of mice whose mothers had protein-restricted diets (p < 0.00'1). Adding folate to the protein-res tricted diet abolished Ihis di fference. (8) Levels of mRNA for the PPARa gene were much higher in the mice fed the protein-restricted diel (p < 0.0001). (After lillycrop et al . 200S.)

DEVELOPMENTAL THERAPIES

can contribu te to cancer therapies is the inhibition of angio·

Know ledge gained from research in the field of deve lop­

The c ri tical p oi n t at which a node of cancerous cells becomes a rapidly growi ng tumor occurs when the node

mental biology is now

genesis (blood vessel forma tion) .

being focused on several diseases.

The abil ity to

b lock p aracri ne factors, to use stem cells to the body to become a pluripotenti.1 stem cell may enable us to block the spread of cancer, repair bodily inj uries, and

becomes vascularized. A microtumor can expand to 16,000

regenera te b ody parts, and to induce nearly any cell in

times its original volwne in the 2 weeks following vascular·

even to ameliorate genetic disease.

stances called tumor angi ogene si s factors, which often

ization (Folkman 1974; Ausprunk and Folkman 1977). To achieve vascularization, the microtumor secretes sub­ include the same factors that engender blood vessel growth in the embryo-VEGFs, Fgf2, placenta-like growth factor.

Anti-Angiogenesis

and others. Tilinor angiogenesis factors stimulate mitosis

Judah Folkman (1974) has estimated that as many as 350

in endothelial cells and direct the cell differentiation intu

billion mitoses occur i n the human

blood vessels in the direction of the tumor.

b ody every day.

With

each cell division comes the chance that the resulting cells will be

ma lignant .

Indeed , a u topsies have shown that

every person over 50 years old has

m icroscopic tumors in

their thyroid glands, although less than 1 in 1000 persons

Tu mor angiogen esis can be demonstrated by implan ti­ ng a piece of tumor tissue withi.n the layers of a rabbit or mouse cornea. The cornea itself is not vascularized, but it is surrounded by a vascu lar border, or limb us. The tumor

have thyroid cancer (Fol kman and Kalluri 2004). Folkman

tissue induces blood vessels to form and grow toward the

suggested that cells capab le of forming tumors develop at

tumor

a certain frequency, but that most never form observable

1979). Once the blood vessels enter the tumor, the tumor

rumors. The reason is that a solid tumor, like any other rap­

ceUs W'ldergo explosive growth, eventually bu rsti n g

idly div i ding tissue, needs oxygen and nutrients to sur­ vive. Withou t a b lood supply, potential tumors eW,er die or remain as dormant "microtumors," stable cell popula­ tions wherein dying cells are replaced by new cells. Thus one important area in which knowledge of developlnent

(Figure 1 7.22; Mu thu kkaruppan and A uerba ch

the eye. Other adult solid tissues do not induce blood vessels to form. It might therefore be pOSSible to block tumor devel opmen t by blocking a ngiogenesis . Numerous chem­ icals are bei ng tested as natura l and artificial angiogenesis inhibitors. These compounds act by preventing endothe-

MEDICAL ASPECTS OF DEVELOPMENTAL BIOLOGY

(A)

649

(B)

Vein

Growing tumor

Artery

/

Tumor

2 Days ---�.� 6 Days

o .. . .

I�ii>r

�

8 Days -----�.� 12 Days

FIGURE 17.22 New blood vessel growth to the site of a mam­ nlary tUnlor transplanted into the cornea of an albino mouse. (A) Sequence of events leading to vascularization of the tumor on days 2, 6, 8, and 1 2 . Both the veins and the arteries in the limbus surrounding the cornea supply blood vessels to the tumor. (B) Photograph of living cornea of an albino mouse, with new blood vessels from the limbus approaching the tumor graft. (From Muthukkaruppan and Auerbach 1 979; B courtesy of R. Auerbach.)

lial cells from responding to the angiogenetic signal of the tumor.' One of the advantages of these compounds is that the tumor cells are unlikely to evolve resistance to them, since the tumor cell itself is not the target of these agents (Boehm et al. 1997; Kerbel and Polkman 2002). In one set of clinical trials, an antibody against VEGP­ A was fOWld to be successful against colon cancer but not against breast cancer. This is probably because colon cancer is more dependent on VEGP-induced angiogenesis than are mammary tumors (Whisenant and Bergsland 2005). Antibodies against a placental form of VEGP were able to inhibit the growth of tumors without affecting healthy blood vessels (Fischer et al. 2007). Blocking the VEGP receptor VEGFR3 prevents the angiogenic sprouting need­ ed for new blood vessels (Tammela et al. 2008) and stops tumors from getting blood-borne nutrients and oxygen. Interestingly, thalidomide, the tenitogen responsible for birth defects in the 1960s, is now being used to block l1.unor-induced blood vessels. Thalidomide has been found to be a potent anti-angiogenesis factor tha t can reduce the growth of cancers in rats and mice (0' Amato et aJ. 1994; Dredge et a l. 2002). Cancer and congenital maJiorrnations are opposite sides of the san1e coin. Both involve disruptions of normal devel­ opment. Thus, as we have seen, agents that have been known to cause congenital malformations-thalidomide, retinoic acid, and cyclopamine-can be used as drugs to *This is the flip side of differentiation therapy discussed above. In differentiation therapy, paracrine factors or hormones are added to promote differentiation. Here, the substances being administered block paracrine signals in order to prevent tissue (Le., blood vessel) formation.

prevent cancers. Just as they disrupt normal development, these substances can disrupt the caricature of development that is caused by tumor cells. When angiogenesis is blocked, the tumor cells can be starved.

Stem Cells and Tissue Regeneration Embryonic stem cells As we discussed in Chapter 8, the inner cell mass of the mammalian blastocyst generates the entire embryo. 1den­ tical twins and chimeric individuals show that each of the cells of the inner cell mass are pluripotent. When cultured, the cells of the inner cell mass blastomeres can become embryonic stem (ES) cells, which remain pluripotent. Cur­ rently, pluripotent stem cells are obtained by two major techniques (Figure 17.23A). They can be derived from the inner cell mass of blastocysts, such as those left over from in vitro fertilization (Thomson et al. 1998), and they can also be generated from germ cells derived from sponta­ neously aborted fetuses. The latter are generally referred to as embryonic germ (EG) cells (Gearhart 1998). Some experimental evidence (Stre1chenko et aJ. 2004) suggests that it may also be possible to derive embryonic stem cells from late morulae, befoTe they form blastocysts. While adult stem cells are rare and do not usually remain undif­ ferentiated in culture, embryonic stem cells can be readily harvested and retain their undifferentiated state for years of culturing. The IDlportance of pluripotent stem cells in medicine is potentially enormous. The hope is that human ES cells can be used to produce new neurons for people with degener­ ative brain disorders (such as Alzheimer and Parkinson dis­ ease) or spinal cord injmies, and new pancreatic � cells for people with diabetes. People with deteriorating hearts might be able to have damaged tissu e repla ced with new heart cells, and those suffering from immune deficiencies might be able to replenish their failing immune systems. Such ther­ apies have already worked in mice. Murine ES cells have been cultured under conditions causing them to form insulin-secreting cells, muscle stern cells, glial stem cells, and

CHAPTER 1 7

650

.

(A)

Q� @ -:. . •

.

•

.

�

G:::l �

••

. . .. . . .

B lastocyst-stage em b ryo

Cells from inner cell mass

Fetus

Embryonic stem cells;

I

Primordial germ mass

cultured pluripotent stem cells

Lineage-specific stem ceUs

FIGURE 17.23 E mb ryo n i c stem ce l l ther­ a pe u ti cs IA) ES cel l s can be obtained from the inner cell mass of the blastocyst or from primordial germ cells and cul­ tured in different ways to produce lineage· spec ifi c stem cells. These cells (or the pre­ cursor cells derived from them) can then be tran sp lan ted into a host. (B) Differentia· tion of mo u se ES cells into l ineage-restrict­ ed (neuronal and glial) stem cells can be accomplished by al tering the media in which the ES cells grow. (A after Gearhart 1 998; B, photographs from Brustle et al. 1 999 and Wi cke l gren 1 999, courtesy of O. Brustle and J . W. McDonald.)

neural stem cells ( Figure McDonald et al.

1999).

1 7.236; Brustle et aJ. 1999; Dopaminergic neurons

derived from ES cells have been shown to signifi· cantly reduce the symptoms of Parkinson disease

Heart cell precursor

•

(B)

• ES cell

. .

-tI

�

:::::--...

.-

Neural precursor

I

in rodents (Bjorklund et al.

2002; Kim et al. 2002).

Wheu neural stem cells derived from germ cell· derived ES cells were transplanted into the injured

� \ "--; ----- /---' ,-------,

Transplantation therapy

brains of mice, the neural stem cells replaced neu­ rol15 and

gUal ceUs

in the forebrains of the newborn 2005). However, in most cases

mice (Mueller et al.

when human ES cells have been transplanted into the brains of immunosuppressed animals, the results have been less encouraging. Typically, trans­ plantation or thousands of ES cell-generated neu­

�:�����:;�r

Medium . . contammg basic fibroblast growth factor and platelet-derived growth factor Glial stem cells

I ansp!antation t �mto mICe

rons into animals results in very few surviving

Medium containing retinoic acid

• t _

dopaminergic neurons and a high frequency of tumors (Li et aJ. 2006). For such transp!antatiol15 to work, one must be able to find the neural stem cells made by the ES cells and to prevent undifferentiat­ ed ES cells from being transplanted. Human embryonic stem cells differ from their murine counterparts in their growth requirements.

In most ways, however, they are very similar, and have a similar, if not identical, pluripotency. Like mouse ES cells, human ES cells can be directed

Neural stem cells

Kaufman and his colleagues

down specific developmental paths. For example,

(2001) directed human

ES cells to become blood-forming stem ceUs by plac­ ing them on mouse bone marrow or endothelial cells. These ES-derived hematopoietic stem cells could further differentiate into numerous types of blood cells (Figure

1 7.24) . Human embryonic germ in

cells were able to cure virus-induced paraplegia

rats. These stern cells appear to do thls both by dif­ ferentiating into new neurons and by producing paracrine factors (BDNF and TGF-a) that prevent the death of the existing neurOI15 (Kerr et aJ.

Functional glial cells

2003).

Similarly, ES cells from monkey blastocysts have

Functional neurons

been able to CUIe a Parkinson-like condition in adult

MEDICA L ASPECTS OF DEVELOPMENTAL BIOLOGY

FIGURE 17.24 Differentiated blood cells developing from human ES cel l s cultured on mouse bone marrow. (Courtesy of The University of Wisconsin.)

monkeys whose dopaminergic neurons had been destroyed (Tagaki et al. 2005). Research is now being done to find ways of directing the differen tia tion of ES cells by using small molecules (rather than paracrine factors). Chen and colleagues (2009), for instance, have found tha t indolactam V molecules can induce Prix1 expression in ES cells, di.recting them into the pancreatic lineage. Thus, ES cells may be able to provide a reusable and readily available source of cells to heal dam­ aged tissue in adult men and worn,en. See WEBSITE 1 7.4 Therapeutic cloning

651

The newfound knowledge of the transcription factors needed to maintain pluripotency has illuminated a star­ tlingly easy way to generate embryonic stem cells that have the exact genotype of the patient.' In 2006, Kazutoshi Taka­ hashi and Shinya Yamanaka of Kyoto University demon­ strated that by inserting activated copies of four genes that encoded some of these critical transcription factors, nearly any cell in the adult mouse body could be made into a cell with the pluripotency of embryonic stem cells. Such cells are called induced pluripotent stem (iPS) cells. Using a strategy very similar to the one used to identi­ fy the active components of Spemann's organizer (see Chapter 7), Takahashi and Yamanaka obtained mRNA from mouse ES cells and made these into cDNAs, which they placed onto active viral promoters. They transfected sets of 24 of these recombinant viruses into cultured fibrob­ lasts that had a neomycin-resistance gene placed onto an Fbx15 regulatory region. The Fbx15 gene is usually turned on in ES cells, so if the transfected genes activated the Fbx15 gene, then the neomycin-resistance gene would be activat­ ed and the cells would survive in neomycin-mntaining cul­ ture medium (Figure 1 7.25A). U a cohort of 24 active genes was found to activate the Fbx15 promoter (as shown by the cells' growing in neomycin-containing medium), then the group of cloned genes was further split. Eventually, Taka­ hashi and Yamanaka discovered that only four active genes were needed to tum on the Fbx15 gene. These genes encod­ ed the transcription factors Oct4, Sox2, .Klf4, and c-Myc (Figure 1 7.256). When the researchers cultured those cells whose Fbx15 promoter was activated, they found that in many cases the cells had become pluripotent, as demon­ strated in a series of tests:

• When the cells were aggregated together, they formed a

teratoma-a tumorlike amalgam containing cell types of all three germ layers.

Induced pluripotent stem cells

The mammalian inner cell mass is known to be character­ ized by certain transcription fadors, including Nanog, Oct4, and Sox2. Knocking out these. genes in mice abolish­ es the pluripotency and self-renewal of the inner mass blas­ tomeres, eventually leading to the demise of the embryo (see Boyer et aJ. 2006; Niwa 2007). Sox2 and Oct4 can dimerize to form a transcription factor complex tha t acti­ vates thei.r own genes (Oe14 and Sox2) as well as the Nnnog gene. It appears that Oct4 not only promotes the synthesis of Sox2 and Klf4, but it also blocks the production of the miR145 microRNA that would otherwise prevent the trans­ lation of Oct4, Sox2, and KIf4 messages (Chivukula and Mendell 2009; Xu et al. 2009). These transcription factors (and microRNAs) initiate a transcription network in the inner cell mass and ES cells that is essential for pluripoten­ cy (see Welstead et al. 2008). This fi.rst set of transcription factors probably acts by activating other transcription fac­ tors, such as Nanog and FbxlS, which are critical in main­ taining the pluripotent and dividing state.

•

•

When injected onto the blastocyst of a normal mouse, the induced cells showed that they could contribute to the production of cells in each of the three germ layers (Figure 1 7.25C; Maherali et a1. 2007; Okita et a1. 2007; Wernig et a!. 2007). 'When the inner cell masses of mouse embryos were C0111pletely compo;;ed of induced pluripotential cells, normal mice were generated (Boland et al. 2009; Zhao et al. 2009).

• The transcription and DNA methylation pattern of the induced pluripotential cells was found to be almost iden­ tical to that of actual mouse embryonic stem cells. The genes for Nanos, Sox2, and other ES-cell transcription factors were hypomethylated, as they are in ES cells. Interestingly, the methylation problems that plague

:tIn order to be a successful therapeutic agent embryonic stem cells have to be the same genotype as the patient, SO that their differenti­ ated products will not be rejected. In fact, however they are not quite exact, because the viruses used to induce pluripotency are integrated into the chromosomes, as we'll see next. ,

CHAPTER 1 7

652

(A) Pbx 15

Ii

enhan�er X� F I

enhancer �

" ' &) " N � � _ Fibroblast iPS cell _

Pbx 15

II

(B)

C=I ,

•

"

N,

1Retroviral infection and neomycin selection 5' LTR II Viral II I 3'LTR promoter

N

Control

(C)

eomyc -resistant in

eDNA

FIGURE 1 7.25 Produ ction of i nduced pluripotent stem cells by expression of four transc ri pti on factors in adult mouse tail fibro­ blasts. (A) A mouse was produ ced wherein the 5' regul a to ry region of the Fbx7 5 gene was attached to a neomycin-resistance (N eo RI gene. When Fbx l5 gene is activated (as it is in ES cellsl, th e n eomycin-res stan ce gene is a ctivated and the cells are able to grow in medium containing neomycin. Various combinations of genes encoding mRNAs found in ES cells were added to the fibroblasts, and the fibroblasts grown in neomycin to see if a ny combination of genes activated the Fbx 15 promo ter. (8) Some neomyc in-resistant colonies survived . These resistant colonies were found to have genes encoding Sox2, Oct4, Klf4, and c-Myc. (C) The pluripotency of these cells was shown by their i nsertion into the inner cell mass of a mouse blastocyst. GFP-Iabeled iPS cells were subsequ e tly found in all organs of the embryo. (After Takahashi and Yamanaka 2006.1 -

i

n

normal cells. However, Nakagawa and colleagues (2007) showed that by altering the culturing techniques, one could circumvent the need for c-Myc. In other words, one should be able to obtain induced pluripotent stem cells by adding merely three active genes to a patient's cell. Another way of "getting rid of c-Myc" was to place the genes for the key ES transcription factors onto episomal viral vectors (Kaji et a1. 2009; Yu et a1. 2009). Episomal vectors are derived from viruses (such as the Epstein-Barr virus that causes mononucleosis) that do not insert themselves into host DNA. rn this manner, the genes from the vector generated the transcription factors that converted human fibroblasts into iPS cells. Once the iPS cells were produced, the cul­ ture media could be changed so that the vector would be eliminated. This technique generated human iPS cells that did not have a v:irus inserted in it. Thus, not only was the

cloned animals do not appear to be a problem for embry­ on.ic stem cells derived by somatic cell nuclear transfer. It appears that the nuclei in the small population of ES cells that survive in culture have had their methylation pat­ terns erased, thus enabling them to redifferentiate (Jaenisch 2004; Rugg-Gwm et a1. 2005).

problem of c-Myc circumvented, but so was the problem of the viral insertion capable of causing a gene mutation. Recently, adult stem cells have been found that are already "part of the way" to becoming pluripotent and do not need as many factors to make them so. Giorgetti and colleagues (2009) showed that blood stem cells from the human umbilical cord can be converted into pluripoten­ tial stem cells merely by adding activated OCT4 and SOX2

This technique does nothing less than transform any cell

genes, and Kim and colleagues (2009) found that they

in the body into an embryonic stem cell. By 2007, the

could transform human neural stem cells into iPS cells by

Yamanaka laboratory (Takahashi et a1. 2007) had used the

adding only a single activated gene, OCT4.

same set of four transcription factors to induce pluripoten­

The therapeutic potential of iPS cells was demonstrat­

tiality in adult human fibroblast cells. These induced cells

ed by the ability of iPS-derived hematopoietic stem cells

formed teratomas containing cells of all three germ layers,

to correct a sickle-cell anemia phenotype in mice

and they had the same transcription profile as the human

(Figure 17.26; Hanna et a1. 2007). Here, tail-tip fibroblasts from a

ES cells.

mouse with sickle-cell hemoglobin were made into iPS cells

Yamanaka (2007) proposed that c-Myc helped "immor­

b y being infected in culture by viruses containing Oct4,

talize" the cell, keeping it in an undifferentiated state of

Sox2, Klf4, and c-Myc. The iPS cells were selected, then

proliferation, and that Klf4 prevented apoptosis and senes­

electroporated with DNA containing wild-type globin

cence. But this situation would have created a tumor, were

genes. These genetically corrected iPS cells were cultured in

it not for Oct4 and Sox2, which redirected growth and gave

media that promoted the production of hematopoietic stem

the cell properties similar to those of germ cells. The need

cells, and the hematopoietic stem cells were injected back

for c-Myc was disconcer ting, however, since this is a well­

into the mice with sickle-cell anemia. By

known oncogene, capable of initiating tumor formation in

mia had been cured.

2 months, the ane­

MEDICAL ASPECTS OF DEVELOPMENTAL BIOLOGY

The combination of iPS cells and genetic engineering may

be able to cure certain genetic diseases. Raya and colleagues

(2009)

have cultured dermal fibroblast cells from patients

653

paracrine factors that appear to activate the heart's own stem cells to repair the damage (Cho et al.

2007; Mirotsou

2007).

et at.

with the genetic disease Fanconi illlemia and have added to them a good copy of the defective gene. These cells were

BONE REGENERATION Several stem cell therapies can be

es bearing the activated forms of OCT4, SOX2, KLF4, and c­

seen

then converted into iPS cells by the addition of retrovirus­

MYC. Moreover, these iPS cells could be directed to become

can become normal parts of

in an incredibly important area of regenerative med­

icine: forming new adult bone. VVhile fractured bones can heal, bone cells

in adults usually do not regrow to bridge

the blood stem cells needed by the patients. However, we

wide gaps. The finding that the same paracrine and

still do not know if these cells

endocrine factors involved in endochondral ossification

the patients and not produce tumors themselves.

are also involved in fracture repair (Vortkamp et al.

1998)

raises the pOSSibility that new bone could grow if the prop­ er paracrine factors and extracelJular environment were

Adult stem cells and regeneration therapy

provided. Several methods are now being tried to devel­

In the introduction to Part 1II, we mentioned the potential

op new functional bone in patients with severely fractured

these new therapies.

Bonadio and his colleagues

medical uses of adult stem cells, especially the mesenchy­

mal stem cells. We will now provide details about some of

or broken bones. One solution to the problem of delivery was devised by

(1999), who developed a

col­

lagen gel containing plasmids carrying the human parathy­

CARDIAC REGENERATION When mesenchymal stem cells

roid hormone gene. The plasmid-impregnated gel was

from the bone marrow of heart attack patients are injected

placed in the gap be tween the ends of a broken dog tibia

into their own hearts, these cells can differentiate into heart

or femur. As cells migrated into the collagen matrix, they

and vessel cells and can significantly m i prove the patients'

incorporated the plaSmid and made parathyroid hormone.

outcomes (Yousef et al.

2009).

Interestingly, in many

A dose-dependent increase in new bone formation was

1 7.27). This type of treat­

instances the stem ceUs do not create new structures them­

seen in about a month (Figure

selves to circumvent the blockage. Rather, they secrete

ment has the potential to help people with large bone frac­ tures as well as those with osteoporosis.

Humanized sickle ceU

(HbSIHbS)

anemia mouse model

Uo

,

o Transplant corrected

hematopoetic progenitors back into irradiated mice

t

� "'"

_ _ _

I

\

e Differentiate into embryoid bodjes

t;-y HbAlHbS

iPS clones

,

in

IPS cells by speCific

gene targeting

E;�y\

FIGURE 1 7.26 Protocol for curing a "human" disease in a mouse, using iPS cells fibroblas t plus recombinant genetiCS. (1 ) Tail-tip fibroblasts are taken from a mouse whose genome contains the human alleles for f) Infect with Oct4, sickle cell anemia (HbS) and no mouse Sox2, Klf4 and genes for this protein. (2) The cells are cultured and infected with viruses con­ taining the four transcription factors known to induce pluripotentiality. (3) The iPS cells are identified by their distinctive shapes and are given DNA containing the wild-type al lel e of human g l ob in (HbA). (4) The embryos are allowed to differenti­ ate in culture . They form Ilernbryoid bod­ ies" that contain blood-forming stem ce lls. (5) Hematopoietic progenitor and stem cells from these embryoid bodies are injected into ./ mouse-derived the original mouse and cure its sickl e-ce ll aneiPS dones mia. (After Hanna et aJ. 2007.)

�

o �.rrect sick1e-ceU �utation

· Harvest tail tIp fibroblasts

I

@

654

CHAPTER 1 7

b

(e) Whole on e, 53 wk

(A) Treated fracture

FIGURE 17.27 Bone formalion from collagen matrix containing plasm ids bearing the human parathyroid hormone. (A) A 1 .6-cm gap was made in a dog femur and stabilized with screws. Plasmid-containing gel was placed on the edges of the break. Radiographs of Ihe area at 2, 8, 1 2, 1 6, and 1 8 weeks afler the sur­ gery show the formalion of bone bridging the gap at 1 8 weeks. (B) Control fracture (no plasmid in Ihe gel) al 24 weeks (C) Whole bone a year after surgery, showing repaired region. (From Bonadio el aJ. 1 999, courtesy of J. Bonadio.) .

and mechanical engi neering, is called tissue engineering. Li and collea gue s (2005) made scaffolds of ma terial that resembles normal extracellular matrix and which can be molded to form the shape of bone needed. The bone

(B)

Untreat.ed fracture, 24

wk

Another approach is to find the right mixture of paracrine factors to recruit stem cells and produce normal bone. Peng and colleagues (2002) genetically modified muscle stem cells to secrete BMP4 or VEGF-A. These cells were placed in gel matrix discs, which were implanted in wowlds made in mouse skulls. The researchers found that certain ratios of BMP4 and VEGFs were able to heal the wounds by making new bone. Similarly, BMP2 has been used to heal large fractures of primate mandibles and rab­ bit femurs (Li et a1. 2002; Marukawa et a1. 2002). A third approach is to make sca ffolds that resemble those of the bone and seed them with bone nlarrow stem cells. This approach, combining developmental biology

marrow stem cells can be placed in these scaf· folds and plac ed into the existing bone. These stem cells are told how to differentiate by the local conditions, and histological and gene expression studies have shown that these cells form bone with the appropriate amounts of osteocytes and chondrocytes. Similarly, when cells are removed from the tracheae of recent cadavers, they can be reseeded with bone marrow stem cells of patients with damaged tracheae. The stem cells dif­ ferentiate into chondrocytes, and the newly reseeded tra­ cheae can be substituted for ille damaged tracheae and pro­ vide a normal airway (M acchi arini et a1. 2008). A fourth ap proach has been to let the body do the work. While bones can't regenerate if wide gaps appear, many bones can undergo normal healing of small wounds. Here, cells in the periosteum (the sheath of cells that surrounds the bone) will differentiate into new cartilage, bone, and ligaments to fill the crack. Stevens and colleagues (2005) have used this normal healing process to make new bones. They injected saline solutions between the rabbit tibial bone and its periosteum, mimicking a fracture, and kept this space open by adding a gel containing calcium (to push the per iosteal cells to differentiate into bone rather than cartilage). Within a few w eeks, these cav ities filled with new bone, which could be tra nspl a nted into sites where b one had been damaged. This technique might pro­ vide a r ela tively painless way to produce new bone for fus­ ing vertebrae and other surgical techniques. NEURONAL REGENERATION While the central nervous sys­ tem is characterized by its ability to change and make new connections, it has very little regenerative capacity. The motor neurons of the peripheral nervous system, however, have Significant regenerative powers, even in adult mam· mals. The regeneration of motor neurons involves the

MEDICAL ASPECTS OF D EVELOPMENTAL B I OLOGY

J£ the cell body of a motor neu­

655

regrowth of a severed axon, not the replacement of a miss­

myelination, and these transcription factors appear to be

ing or diseased cell body.

repressed by high levels of Wnt signaling (Arnett et a1.

ron is destroyed, it cannot be replaced.

2004; Fancy et al. 2009). Therefore, the use ofWnt inhibitors

The myelin sheath that covers the axon of a motor neu­

(such as Frzb and Dickkopf) may be a mechanism for treat­

ron is necessary for its regeneration. This sheath is made

ing this disease. Research into CNS axon regeneration may

by the Schwann cells, a type of glial cell in the peripheral

become one of the most important contributions of devel­

nervous system (see Chapter

opmental biology to medicine.

9). When an axon is severed,

the Schwann cells divide to form a pathway along which the axon can regrow from the proximal stump . This pro­ liferation of the Schwann cells is critical for directing the

Direct transdifferentiation

regenerating axon to the original Schwann cell basal lam­

One of the newest developmental therapies involves using

ina. If the regrowing axon can find that basal lamina, it can

transcription factors to convert one cell type into another

be guided to its target and restore the original connection.

without the intermediary of stem cells, a procedure known

In turn, the regenerating neuron secretes mitogens that

as transdifferentiation. In Chapter

allow the Schwann cells to divide. Some of these utitogens

transdifferentiation of exocrine pancreatic cells into insulin­

2, we

discussed the

are specific to the developing or regenerating nervous sys­

secreting pancreatic p cells by three transcription factors

tem (Livesey et a1.

(Ngn3, Mafal, and Pdxl). More recently; Kajiyama and col­

1997).

The neurons of the central nervous system cannot regenerate their axons under normal conditions. Thus,

leagues

(2010) have transfected .one of these factors, Pdxl,

into mouse adipose-derived stem cells and converted the

spinal cord injuries can cause permanent paralysis. One

stem cells into insulin-producing pancreatic cells. These

strategy to get around this block is to find ways of enlarg­

induced pancreatic cells were able to reduce the hyper­

ing the population of adult neural stem cells and to direct

glycemia in diabetic mice.

their development in ways that circumvent the lesions

Vierbuchen and colleagues (2010) found that ti,e viral

caused by disease or trauma. The neural stem cells found

insertion of three active transcription factor genes (Ascll,

.in adult mammals may be very similar to embryonic neu­

Bm2,

ral stem cells and may respond to the same growth factors

dermal fibroblasts into ftmctional neurons in vitro. These

Oohe et al.

1996; Johansson et a1. 1999;

Kerr et a1.

2003).

Another strategy for CNS neural regeneration is to cre­

and My tIl) sufficed to effiCiently convert mouse

induced neuronal cells expressed several neuron-specif­ ic proteins, were able to generate action potentials, and

ate environments that encourage axonal growth. Unlike

formed functional synapses (resembling excitatory neu­

the Schwann cells of the peripheral nervous system, the

rons of the forebrain). It remains to be seen if this trans­

myelinating glial cells of the central nervous system, the

differentiation is stable and whether it will ameliorate

oligodendrocytes, produce substances that inhibit axon

disease in vivo. However, the ability to generate pancre­

regeneration (Schwab and Caroni

atic � cells and neural cells that are genetically identical

1988).

Schwann cells

transplanted from the peripheral nervous system into a

to a patient holds the promise of significant therapies for

CNS lesion are able to encourage the growth of CNS axons

patients with diabetes and degenerative neural diseases.

to their targets (Keirstead et a1.

1999; Weidner et a1. 1999).

Three substances that inhibit axonal outgrowth have been isolated from oligodendrocyte myelin: myelin-associated glycoprotein, Nogo-l, and oligodendrocyte-myelin glyco­

1994; (hen et a1. 2000; Grand­ 2002). Each of these substances

protein (Mukhopadyay et a1. Pre et al.

2000; Wang et al.

Coda Developmental biology is gaining increasing importance in modern medicine. First, preventive medicine, pubHc health, and conservation biology demand that we learn

binds to the Nogo receptor (NgR). Thus, NgR may be the

more about the mechanisms by which industrial chemi­

critical target for therapies allowing regeneration. Chen

cals and drugs can damage embryos. The ability to effec­

and colleagues

(2009) lISed RNA interference to block

tlIe

tively and inexpensively assay compounds for potential

synthesis of NgR and found that this helped rat optic neu­

harm is critical. Second, developmental biology is pro­

rons to regenerate. The axons regenerated even better

viding us new ways of preventing and curing cancers.

when this therapy was combined with nutrients and pos­

Third, developmental biology is providing the explana­

itive growth factors.

tions for how mutated genes and aneuploidies cause their

As mentioned in the case with bone, there is a dose rela­

aberrant phenotypes. Fourth, the field of regenerative

tionship between wound healing and regeneration.

medicine is using developmental biology to provide

Patients with multiple sclerosis suffer from the demyeli­

induced pluripotent stem cells as well as the knowledge

nation of axons

of paracrine factors needed to form new cells, tissues, and

in the brain and spinal cord.

Several tran­

scription factors are needed to reinitiate the pathway to

organs in adults.

656

CHAPTER 1 7

Snapshot Summary: Medical Aspects of Developmental Biology 1. Pleiotropy occurs when several different effects are

produced by a single gene, In mosaic pleiotropy, each effect is caused independently by the expres­ sion of the same gene in different tissues . In relation­ al pleiotropy, abnormal gene expression

in one tissue

influences other tissues, even though those oilier tis­ sues do not express that gene,

2.

type,

in different individuals,

can produce different defects (or differing severities 4, Preimplantation genetics involves testing for genetic

abnormalities in early embryos in vitro, and implanting only those embryos that may develop normally. Selection for sex is also possible using preimplantation genetics,

5.

Teratogenic agents include certain chemicals such as alcohol and retinoic acid, as well as heavy metals, certain pathogens, and ionizing radiation. These agents adversely affect normal development, yield­ ing malformations and functional deficits.

6. Fetal alcohol syndrome is completely preventable. There may be multiple effects of alcohol on cells and tissues that result in this syndrome of cognitive and physical abnormalities.

7,

10, Cancer can be seen as a disease of altered develop­ ment. Some tumors revert to non-malignancy when placed in environments that fail to support rap id cell

11. Cancers can arise from errors in cell-cell communica­ tion. These errors include alterations of paracrine factor syn thesis.

Phenotypic heterogeneity arises when the same gene of the same defect)

by turning genes all,

methylation can alter metabolism and development

clivision.

Genetic heterogeneity occurs when mutations in more than one gene can produce the same pheno­

3,

inherited from one generation to the next. Such

Endocrine disruptors can bind to or block hormone receptors or block the synthesis, transport, or excre­ tion of hormones. DES is a powerful endocrine dis­ ruptor, Presently, bisphenol A and other compounds are being considered as possible agents of low sperm counts in men and a predisposition to breast cancer in women.

8. Environmental estrogens can caus� reproductive system anomalies by suppressing Hox gene expres­ sion and Wnt pathways,

9. In some instances, endocrine disruptors methylate DNA, and these patterns

01 methylation can be

12. Cancers metastasize in manners similar to embryon­ ic cell movement.

13. In many instances, tumors have a rapidly dividing as well as more quiescent and differentiated cells.

stem cell population which produces more stem cells

14, The methylation pattern 01 cancer cells is often aber­ rant� and these methylation differences can cause cancer by inappropriately inactivating tumor sup­ pressor genes or activating oncogenes.

15, Disrup ting tumor-induced angiogenesis may become an important means of stopping tumor progression.

16, Skin fibroblasts, and perhaps any normal adult cell, can be induced to form pluripotent stem cells by the incorporation of activated genes encoding certain transcription factors. These pluripotent stem cells would not be rejected by the patient from whom they were formed,

17, By altering conditions to resemble those in the embryo and by providing surfaces on which adult stem cells might grow, some adult stem cells might be directed to form numerous cell types.

19. Neurons in the central nervous system can be aided in regenerating by blocking the glial-derived

paracrine factors that stabilize neurons and prevent their growth.

20. Transdifferentiation is the viral insertion of tran­ scription factors (or transcription factor activation by small molecules), which can convert one stable cell

type into another,

For Further Reading Complete bib liographical citations for all literature cited in this chapter can be found at the free-access website www.devbio.com Anway, M, D" A. S. Cupp, M. Uzumcu and M, K. Skipper. 2005, Epigene tic transgeneration effects of endocrine dis­ ruptors and male fertility, Science 308: 146&-1469.

Baksh, D" L Song and R. S, Tuan. 2004, Adult mesenchymal stem cells: Charac­ terization� differentiation, and applica­ tion in cell and gene therapy, j, Cell Mol, Med. 8: 301-316,

Bissell, M, j" D. C. Rarlisky, A, Rizki, V, M, Weaver and 0, W. Petersen, 2002,

The organizing principle: Microenviron­ mental influences in the normal and malignant breast. Differentiation 70:

537-546.

MEDICAL ASPECTS OF DEVELOPMENTAL BIOLOGY

Foty, R. A. and M. s. Steinberg. 1997. Measurement of tumor cell cohesion and suppression of invasion by E- and P-cadherin. Cancer Res. 57: 5033-5036. Gilbert, S. F. and Epel. D. 2009. Ecologi­ cal Developmental Biology: Integratirlg Epigenetics, Medicine, and Evolution. Sin­ auer Associates. Sunderland, MA. Gluckman, P. D. and M. A. Hanson.

2004. Living with the past: Evolution, development, and patterns of disease. Science 305: 1733-1739. Gupta, P. S., C. L. Chaffer and R. A. Weinberg. 2009. Cancer stem cells: Mirage or reality?

15: 1010-1012.

Natw'e Medicine

Howdeshell, K L., A. K Hotchkiss, K. A. Thayer, ). G. Vandenbergh and F. S. vom Saal. 1999. Plastic bisphenol A speeds growth and puberty. Nature 401:

762-764.

Huang, G. T., S. Gronthos and S. Shi.

2009. Mesenchymal stem cells derived

from dental tissues vs. those from other sources: Their biology and role in regenerative medicine. J. Dental Res.

88:792-806.

The stroma as a crucial target in mam­ mary gland carcinogenesis. j. Cell Sci.

117: 1495-1502.

Nakayama, A., M. T. Nguyen C. C. Chen, K Opdecamp, C. A. Hodgkinson and H. Arnheiter. 1998. Mutations in microphthalmia, the mouse homolog of the human deafness gene MITF, affect neuroepithelial and neural crest­ derived melanocytes differently. Mech ,

Keller, G., G. Zimmer, G. Mall, E. Ritz and K Amann. 2003. Nephron number in patients with primary hypertension. New Engl j. Med. 348: 101-108. .

Lammer, E. ). and 11 others. 1985. Retinoic acid embryopa thy. N. Eng!.

Med. 313: 837-ll4 1.

I.

Lillycrop, KA., E. S. Phillips, A. A. Jack­ son, M. A. Hanson and G. C. Burdge. 2005. Dietary protein restriction of preg­ nant rats induces and folic acid supple­ mentation prevents epigenetic modifi­ cation of hepatic gene expression in the offspring. f. Nutrition 135: 1382-1386.

MacchiarinL P. and 14 others. 2008. Clinical transplantation of a tissue-engi­ neered airway. umcet 372: 2023-2030. Maffini, M. v., A. M. Soto, I. M. Calabro, A. A. Ucci and C. Sonnenschein. 2004.

Oev. 70: 155-166.

Sulik, K K 2005. Genesis of alcohol­ induced craniofacial dysmorphism. Exp. Bioi. Med. 230: 366-375.

Ta kahashi, K and S. Yamanaka. 2006. induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126:

663-676. Tammela, T. and 20 others. 2008. Block­

ing VEGFR-3 suppresses angiogenic sprouting and vascular network forma­ tion. Nat"re 454: 656-660.

Outside Sites

Go Online WEBSITE 17.1

Human embryology and genetics. This

links to other websites that will connect you to tutorials in human development, as well as to the Online Mendelian Inheritance in Man (OMlM), which details all human genetic conditions.

WEBSITE 17.2

Association of Reproductive Health Professionals:

malfonned arms and legs, and it provided the first major

resources for health care providers and their Endocrine disruptor exchange:

Reproductive Toxicology website: www.reprotox.org has summaries of over 5000

agents, exposure levels, and their effects on development and

evidence that drugs could induce congenital anomalies. The mechanism of i ts action is still hotly debated.

reproduction.

WEBSITE 17.3

The NIH website on stem cell education.

Our stolen future. nus website monitors

the environmental effects of endocrine disruptors. It is a political and consumer action site as well as a scientific clearinghouse for

endocrine disruption. Run by the authors of the book Our Stolen Future, it also provides links to the websites of people who disagree with them.

t echniqu e, thera peu tic the immune rejection

Therapeutic cloning. Prior to induced

pluripotent stem ceils, another

cloning, was seen as a way around

of stem cell-derived differentiated cells.

Vade Mecum Somites and thalidomide.

These movies are from the lab­

oratory of Jay Lash, whose insightful work on cartilage for­ mation resulted in some of our first insights into the mechansisms b y which the drug thalidomide halts limb growth.

clients.

www.arhp,orgltopics/enviro-repro-health contains

www.endocrinedisruption.com

Thalidomide as a teratogen. The drug

thalidomide caused thousands of babies to be born with

WEBSITE 17.4

657

Stern cell basics: http://stemcells.nih.gov/info/basics "All things stem cell." An informative stem cell blog: http://www. allthingsstemcell.com For information on fetal alcohol syndrome: http://www.cdc,gov/ncbddd/fasdidata.htmi . NlH, FASD: http://www.niaaa.nih.gov/AboutNIAAAllnteragency/ AboutFAS.htm

Substance Abuse and M ental Health Services Administration (SAMHSA):

http://www.fascenter.samhsa.gov

Developm ental Plasticity and Sym biosis

iT WAS LONG THOUGHT THAT THE ENVIRONMENT played only a minor role in

development. Nearly all developmental phenomena were believed to be a "read­ out" of nuclear genes, and those organisms whose development was significant­

ly controlled by the environment were considered interesting odd ities. When

environmental agents played roles in development, they appeared to be destruc­

tive, such as the roles played by teratogens and endocrine disruptors (see Chap­ ter 17). However, recent s tudies have shown that the environmental context plays significant roles in the

narmaJ d evelopment of almost aU species, and that ani­

mal genomes have evolved to respond to environmental conditions. Moreover, symbiotic associations, wherein the genes of one organism are regulated by the products 01 another organism, appear to be the rule rather than the exception. One reason developmental biologists have largely ignored the environment's eflects is that a criterion for selecting wh.ich animals to study has been their abil­

C. elegans, Drosapl.i/fl, zebra fish, Xenopus,

ity to develop in the laboratory (Bolker 1995). Given adequate nutrition and tem­

perature, all "model organisms

"-

ch.icks, and laboratory mice-develop independently 01 their particular environ­ ment, leaving us with tl,e erroneous impression that everythlng needed to form

the embryo is wW1in the fertilized egg. Today, with new concerns about the loss of organismal diversity and the effects of environmenta l pollutants, there is renewed interest in the regulation of development by the environment (see van der Weele 1999; Gilbert and EpeI 2009).

The Environment as a Normal Agent in Producing Phenotypes Although the nucleus and cytoplasm of the zygote contribute a majority of phe­ notypic instructions, everything needed for producing a pa rticula r phenotype is not pre-packaged in the fertilized egg. Rather, crucial parts of phenotypiC deter­

mination are regulated by environmental facton; outside the organism.

Phenotyp­

ic plasticity is the ability 01 an organism to react to an env ironmental input with

seen in embryonic or larval stages of animals or p la nts , this ability to change a change n i form, state, movement, or rate of activity (West-Eberhard 2003). When phenotype is often called

developmental plasticity.

We have already enco untered several exampl es of developmental plasticity in this book. When we discussed environmental sex determination in turtles,

fish, and ech.iuroid worms (see Chapter 14), we were aware that the sexual phe­ notype was being instructed not by the genome but by the environment. When we discussed in Chapter

12 the ability of shear stress to activate gene expression

in capillary, heart, and bone tissue, we Similarly were studying the effect of an

We may now tum to consider adapw­ dons towards the extenlOl environment; and fimly the direct adaptations . . . in which an animal, during irs deuelop­ mem, becomes mod ified by exremal

factors in such a way as to increase its effiCiency

in dea li ng

Wilh them.

c. H. WADDINGTON

( 1 957)

)IAN xu AND )EFFREY I. GORDON

(2003)

HOllor thy sYl1lbionts.

660

CHA PTER 1 8

(A)

(6)

IC)

(D)

Spring morph among catkins

Summer morph on twig

environmental agent on phenotype . While studies of phe­ notypic plasticity play a central role in plant developmen­

(E)

lal biology, the mechanisms of plasticity have only recent­ ly been studied in animals. These studies are showing that

the integra tion of animals into ecological communities is

accomplished largely through developmental interactions. autonomous enti ties but rather that we are co-construct­ Indeed, these studies show that we d o not develop as ,

ed by other organisms and that we respond to abiotic agents within our loea1 communities.

FIGURE 18.1 Devel opmen ta l plasti ci ty in i nsects . (A,B) Density­ induced polyphenisrn in the desert (or " plague") locu st Schistocer­ ca gregaria. (AI The low-density m orph has green pigmentation and miniature wings. (B) The high-density morph has deep pig­ mentation and wings and legs suitable for migration. (C,O) Nemo­ ria arizonaria caterpillars. (0 Caterpill ars that hatch in the spring eat young oak leaves and develop a c utic l e thai resembles the oak's flowers (catkins). (D) Caterpi llars thai hatch in Ihe summer, after the catkins are gone, eat mature oak leaves and develop a cuticle that resembles a young twig. (E) Gyne (reproductive queen) and wo rker of Ihe ant Pheidologeton. This picture shows the remarkable dimorphism between the la rge, fertile queen and the small, sterile worker (seen near the queen's antennae). The differ­ ence between these two sisters is the result of larval feed ing (A,B from Tawfik et al. 1 999, courtesy of S. Tanaka; C,D courtesy of E. Greene; E ro Mark W. Moffett/National Geographic Society.) .

recognized: reac tion norms and polyphenisms (Woitereck Two main types of phenotypiC plasticity are currently

1909; Schmalhausen 1949; Stearns et al. 1991). In a reaction norm, the genome encodes the potential for a

range of potential phenotypes, and

continuoIls

the environment the

the most adaptive one). For instance, our muscle pheno­

individual encounters determines the phenotype (usually

type is determined by the amount of exercise our body is exposed to (even though there is a genelically defined limit

to how much muscular hypertrophy is possible) Similar­ .

ly, the length of a male's horn in some dung beetle species is determined by the quantity and quality of food (Le., the

dung) the larva eats before metamorphosis (see the uext section) The upper and lower limits of a reaction norm are .

feren t phenotypes produced by environmental conditions are called morphs (or, occaSionally, ecomorphs). The second type of phenotypic plasticity, po(yphenism,

also a property of the genome that can be selected. The dif­

refers to discontinuous ("either/or") phenotypes elicited by

(Al

Hornless male

Horned male

DEVELOPMENTAL PLASTICITY AND SYMBIOSIS

661

FIGURE 18.2 Diet and Onthophagus horn p henoty pe (AI Horned and hornless males of the dung beetle Onthophagus acuminatu5 (horns have been artificially colored). Whether a male is horned or hornless is determined by the liter of juvenile hor· mone at the last molt, which in turn depends on the size of the larva. (B) There is a sharp threshold of body size under which horns fail to form and above which horn growth is linear with the size of th e beetle. This threshold effect p roduces males with no horns and males with l arge horns, but very few with horns of intermediate size. (Aher Emlen 2000; photographs courtesy of D. Emlen.) .

(6)

1.50

� 1.00

• •

1.25

� 0.75 c h

i

. _.. .. • •• • • • • •• • • • •• • • • --_ . . ••••• • •• • . .. . . ._ . . _. _.- . • •• • • . -• • •• • • • • . -_ .. . __. . • •••• • ••••••••• •••••••• • . . .

_

.c

�

•

0.50

0.25

O2.25 ��2.5 ��2.75 ��� � �4.0 3 .0 � 3� .25-3.5 �� 3.75-•

.

_.

•

.

Body size (mm)

the environment. One obvious example is sex determina­

phenotypes in genetically identical organisms. Diet is also largely responsible for the formation of fertile "queens" n i ant, wasp, and bee colonies. Among these insects, each larva has the genetic potential to become either a worker or a queen; only those larvae fed adeguately become queens ( Figure

1 8.1 E).

In honeybees, we know that extra

nutrition results in the demethylation of particular genes associated with ovary growth and general metabolic rate (Maleszka

2008; Elango et al. 2009; Foret et al . 2009) .

See WEBSITE 1 8.1 Inducible caste determination in ant colonies WHEN DUNG REAllY MAnERS

For the male dung beetle

(Ol1thaphngus), what really matters in life is the amount and

quality of the dw'g he eats as a larva. The hornless femaLe

tion in turtles, where one range of temperatures will induce

dung beetle digs tunneLs, then gathers balls of dung and

female development in the embryo, and another set of tem­

buries them in these tunnels. She then lays a Single egg on

peratures will induce male development. Between these

each dung ball; when the larvae hatch, they eat the dung.

sets of temperatures is a small band of temperatures that

Metamorphosis occurs when the dung bail is finished, and

will produce different proportions of males and females,

the anatomical and behavioral phenotypes of the

but they do not induce intersexual animals. Another

dung beetle are determined by the quality and q uantity

important example of polyphenism is the migratory locust

this maternally provided food (Emlen

Schistocerca gregaria. These grasshoppers exist either as a

Emlen

2000).

male of

1997; Moczek and

The amount of food determines the size of

short-winged, green, solitary morph or as a long-winged,

the larva at metanwrphosis; the size of the larva at meta­

Cues in the

morphosis determines the titre of juveni le hormone dur­

environment determine which morphology a larva will

ing its last molt; and the titre of juvenile hormone regulates

develop upon molting (Rogers et al.

the growth of the imaginal discs that make the horns ( Fig­

brown, gregarious morph ( Figure

Sword

1 8.1A,B).

2003; Simpson and

ure

2008).

1 8.2A; EmLen and Nijhout 1999; Moczek 2005). If juve­

nile hormone is added to tiny

Diet-induced po/yphenisms The effects of diet in development can be seen in the cater­ pillar of

Nentoria arizonaria.

When it hatches on oak trees

in the spring, it has a form that blend, remarkably with the young oak flowers

O. taunts males during the

sensitive period of their last moLt, the cuticle in their heads expands to produce horns. Thus, whether a male is horned

or hornless depends not on the male's genes but on the food his mother left for him.

Horns do not grow until the male beetle reaches a cer­

(catkil1s). But those larvae hatching from

tain size. After this threshold body size, horn growth is

their eggs in the sununer would be very obvious if they

very rapid.' Thus, although body size has a normal distri­

still looked like oak flowers. Instead, they resemble newly

bution, there is a bimodal distribunon of horn sizes: about

formed twigs. Here, it is the diet (young versus old oak

half the males have no horns, while the other half have

leaves) that determines the phenotype ( Fig ure

horns of considerable length ( Figure

Greene

1 8 . 1 C, D;

1 8.2B).

1989).

Diets having different amounts of proteins or different concentrations of methyl donors have also been fOW1d to cause different genes to be expressed in mammalian embryos. Different diets can lead to remarkably distinct

*lnterestingly, the threshold size at which the phenotype changes from hornless to homed is genetically transmitted and can change when conditions favor one morph over the other (Emlen 1996, 2000).

662

CHAPTER 1 8

that gives mice yellowish hajr color; i t also affects lipid metabolism such that the mice become fatter. The

viable-yel­ low allele of Agouti has a transposable element inserted into its cis-regulatory regions. These transposon insertion sites

are

very interesting for regulation: whereas most regions of

the adult genome have hardly any intraspecies variation in CpG methylation, there are large DNA methylation differ­ ences between individuals at the sites of transposon inser­ tion. Such CpG methylation can block gene transcription. When the promoter of the Agouti gene is methylated, the gene is not transcribed. The mouse's fur remains black, and lipid metabolism is not altered. Waterland and Jirtle fed pregnant

viable-yellow Agouti

mice methyl donor supplements, including folate, choline, and betain. They found that the more methyl supplemen­ tation, the greater the methylation of the transposon inser­ tion site in the fetus' genome, and the darker the pigmen­ tation of the offspring. Although the mice in Figure

1 8.4

are genetically identical, their mothers were fed different diets during pregnancy. The mouse whose mother did not

FIGURE 18.3 The presence or absence of horns determines the male reproducti ve strategy in some du ng beet l e speci es . Females dig tunnels in the soil beneath a pile of dung and bring dung frag­ ments i n to the lunnels. These will be the food supply of the larvae. Horned males guard the tunnel entrances and mate repeated ly with the fema les. They fight to prevent other males from entering the tunnels, and thee males with long horns usually win such contests. Sma l l er, hornless males do not guard tunnels, but dig their own to connect with those of females. They can then mate and exit, u nCha l l enged by the guard i ng male. (After Emlen 2000.)

receive methyl donor sup plementation is fat and yellow­ the Agouti gene promoter was unmethylated, and the gene was active. The mouse born to the mother that was given supplements is sleek and dark; the methylated Agouti gene was not transcribed. As we saw in Chapter 17, such dlifferential gene methyla­ tion has been linked to human health problems. Dietary restrictions during a woman's pregnancy may show up as heart or kidney problems in her adult children. Moreover,

studlies in rats showed that dlifferences in protein and methyl donor concentration in the mother 's prenatal diet affected metabolism in the pup's livers (Lillycrop et a!.

2005).

The size of the hams detennines a male's behavior and chances for reproductive success. Horned males guard the females' twmels and use their horns

to prevent other males

from mating with the female; the male with the biggest

horns wins such contests. But what about the males with no horns' Hornless males do not fight with the horned males for mates. Since they, like the fe.males, lack horns, they are able to dig their own tunnels. These "sneaker males" dig tunnels that intersect those of the females and mate with the females while the horned male stands guard at the tunnel entrance (Figure and Emlen

2000).

1 8.3; Emlen 2000; Moczek

Indeed, about half the fertilized eggs in

most populations are from hornless males. The ability to produce a hom is inherited; but whether to produce a hom and how big to make it is regulated by the environment.

DIET AND DNA METHYLATION Diet can also directly influ­ ence the DNA. As mentioned earlier, honeybee caste (queen or worker) is determined by dliet-induced changes in DNA methylation patterns. Dietary alterations can also produce changes in mammalian DNA methylation, and these methylation changes can affect the phenotype. Waterland and Jirtle

(2003) demonstrated this by

the viable-yellow aliele of Agouti.

using mice contalning

Agouti is a dominant gene

FIGURE 18.4 Maternal diet can affect p henotype. These two mice a re geneti ca l ly identical; both contain the viable-yellow all el e of the Agouti gene, whose protein product converts brown pigment to yellow and accelerates fat stora ge . The obese ye l low mouse is the offspring of a mother whose diet was not supple­

mented with methyl donors (e.g., folic acid) during her pregnancy. The embryo'S Agouti gene was not methyla ted, and Agouti protein was made. The sl eek b rown mouse was born of a mother whose pre natal diet was supplemented with methyl donors. The Agouti gene was turned off, and no Agouti protein was made. (After Waterland and J i rtl e 2003, photograph courtesy of R. L. Jirtle.)

DEVELOPMENTAL PLASTICITY A N D SYMBIOSIS

soluble filtrate from waler surrounding the predator is able to induce the changes . Chemicals that are released by a predator and can induce defenses in its prey are called kairomones Several roilier species will alter their morphology when they develop in pond water in which their predators were cultured (Dodson 1989; Adler and Harvell 1990). The predatory rotifer Asplnnchna releases a soluble compound that induces the eggs of a prey roilier species, Keralello slac­ ki, to develop into individuals with slightly larger bodies and anterior spines 130% longer than th ey otherwise would be, making the prey more difficult to eat When exposed to the effluent of the crab species that preys on it, the snail Thais lamellosa develops a thickened shell and a "tooth" in its aperture. [n a mixed snail population, crabs will not attack the thicker-shelled snails until more than half of the typicaJ-morph snails are devoured (Palmer 1985).

Predator-induced polyphenisms

Imagine an animal who is frequently confronted by a par­ tkular predator. One could then imagine an individual who could recognize soluble molecules secreted by that predator and could use those molecules to activate the development of structures that would make this individ­ ual less palatable to the predator. This ability to modulate development in the presence of predators is called preda­ tor-induced defense, or predator-induced pol ypheni sm. To demonstrate predator-induced polyphenism, one has to show that the phenotypic modification is caused by the presence of the predator, and that the modification increas­ es the fitness of its bearers when the predator is present (Adler and Harvell 1990; Tollrian and Harvell 1999). Fig­ ure 18.5A shows both the typical and predator-induced morphs for several species. In each case, the induced morph is more successful at su rviving the predator, and

(Keratella) Rotifer

pe

A

rtu re

�� t8/59

Mollusc

Barnacle (Chthama/us)

t t/43

(Thai,)

Thickened, "toothed"

r n

Preda to prese t

30/100

Predator

FIGURE 18.5 Predator-induced defenses. (A) Typical (upper row) and predator-induced (lower row) morp s of various organisms. The numbers beneath each column represent the percentages of organisms surviving predati on when both induced and uninduced individuals were presented wi th predators (in various assays). (8) Scanning electron micrograph s show predator induced (left) and typical ( right) morphs of genetically identical individuals of the water flea Daphnia. In the prese n ce of chemical signa l s from a -

/

Expanded body depth

(D)

h

(B)

(Carassius)

No predation until 50% of typical inorphs devoured

(C) Predator present

.

Carp

(A)

663

Predator absent

absent

in

predator Daphnia grows a protective "helmet." (e,D) Tadpole phenotypes. (e) Tadpoles of the tree frog Nyla chrysosce/is devel­ oping th e presence of cues from a predator's larvae develop strong trunk muscles and a red coloration (0) When predator cues are absent, the tadpoles grow sleeker, which helps them compete for food. fA aher Adler a nd Harvell 1 990 and references cited therein; B courtesy of A. A. Agrawal; C,O courtesy of T. Johnson/USGS.) ,

.

664

CHAPTER 1 8

One of the more in teres ting mechanisms of predator­

induced polyphenism is that of certain echinoderm larvae. When exposed

to the mucus of their fish pred a tor, sand budding off small groups of cells that quickly become larvae themselves. The small plutei are below the visual detection of the fish, and there­ by escape being eaten (Vaughn and Strathmann 2008; Vaughn 2009). dollar plutei clone themselves,

DAPHNIA AND THEIR KIN The predator-ind uced poly­ phenism of the parthenogenetic water flea Daphnia is ben­ eficial not only to itself but also to its offspring. When D. c"cul/atn encoun ter the predatory larvae of the fly Chaeoborus, their "helmets" grow to twice the normal size (Figure 18.5B). This increase lessens the chances th at Daph­ nia will be eaten by the fly larvae. nus same helmet induc­ tion occurs if the Daphnia are exposed to extracts of water in which the fly larvae had been swinuning. Agrawal and colleagues (1999) have shown that the offspring of such an induced Daphnia are born with this same altered head mor­ phology. It is possible that the OweabortlS kairomone reg­ ulates gene expression both in the adult and in the devel­ oping embryo. We still do not know the iden ti ty of the kairomone, the identity of its receptor, or the mechanisms by which the b indi ng of the kairomone to U,e receptor ini­

tiates the adaptive morphological changes. There are trade­ offs, however; the induced into making protective (Tollrian

Daphnia, having put resources structures, produce fewer eggs

1995).

AMPHIBIAN PHENOTYPES INDUCED BY PREDATORS Predator­

induced polyphenism is not limited to invertebrates," Among amphibians, tadpoles found in ponds or reared in the presence of other species may differ Significantl y from tadpoles reared by themselves in aquaria. For in sta nce, newly hatched wood frog tadpoles (Rona sylvetica) reared in tanks con taining the predatory la r va l dragonfly Anax (confined in mesh cages so U,ey cannot eat the tadpoles) grow smaller than those reared in similar tanks without predators. Moreover� their tail muscuJat\J-re deepens, allow­

ing faster turning and swinuning speeds (Van Buskirk and Relyea 1998). The addition of more preda tors to the tank causes a continuously deeper tail fin and tail musculature, and in fact what initially appeared to be a polyphenism may be a reaction norm that can assess the number (and type) of predators. McCollum and Van Buskirk (1996) have shown that in the presence of its predators, the tail fin of the tadpole of the tree frog Hyla chrysoscelis grows larger and becomes bright red (Figure 18.5C,D). nus phenotype allows the tad­ p ole to swim away faster and to deflect predator strikes *Indeed, when viewed biologically rather than medicilUy, the verte­ brate immune system is a wonderful example of predator-induced polyphenisru. Here, our immune cells utilize chemicilis from our

predators (viruses and bacteria) to change our phenotype so that we CCln better resist them (see Frost 1999).

toward the tail regi on . The trade-off is that noninduced tadpoles grow more slowly and survive better in preda­ tor-free environments. In some species, phenotypic plas­ ticity is reversible, and removing the predators can restore the non-induced phenotype (Relyea 2003a). The metabolism of predator-induced morphs may differ significantly from that of the uninduced morphs, and this has important consequences. Relyea (2003b, 2004) has found tha t in the presence of the chemical cues emitted by predators, the toxicity of pes ticides such as carbaryl (Sevin®j can become up to 46 times more lethal than it is without the predator cues. Bullfrog and green frog tadpoles were espeCially sensitive to carbaryl when exposed to pred­ ator chemicals. Relyea has related these findings to the global decline of amphibian populations, saying that gov­ ernments should test the toxicity of the chemicals under more natural conditions, including that of predator stress. He concludes (Relyea 2003b) that "ignoring the relevant ecology can cause incorrect estimates of a pesticide's lethal­ ity in nature, yet it is the lethality of pesticides under nat­ ural condi tions that is of utmost interest. The accumulat­ ed evidence strongly suggests that pesticides in na ture could be playing a role in the decline of amphibians ."

VIBRATIONAL CUES ALTER DEVELOPMENTAL TIMING The phe­ notypiC changes induced by environmental cues are not confined to structure. They can also include the timing of developmental processes. The embryos of the Costa Rican red-eyed treefrog (Agnlychnis callidryas) use vibrations trans­ mitted through their egg masses to escape egg-eating snakes. These egg masses (shown on this book's cover) are laid on leaves that overhang ponds. Usually, the embryos develop into tadpoles within 7 days, and these tadpoles wiggle out of the egg mass and fall into the pond wa ter. However, when snakes feed on the eggs, the vibrations they produce cue the remaining embryos inside the egg mass to begin the twitcl' ing movements that initiate their hatching (within seconds!) and dropping into the pond. The embryos are competent to begin these hatching movements a t day 5 (Figu re 1 8.6). Interes tingly, the embryos have evolved to respond this way only to vibrations given at a certain fre­ quency and interval (Warkentin et al. 2005, 2006). Up to 80% of the remaining embryos can escape snake predation in this way, and research has shown that these vibrations alone (and not smell or sight) cue these hatching movements in the embryos. There is a trade-off here, too. Although these embryos have escaped their snake predators, they are now a t greater risk from waterborne

predators than are fully of the early

developed embryos, because the musculature

hatchers has not developed fully.

Temperature as an environmental agent TEMPERATURE AND SEX There are many species in which

temperature does control whether testes or ovaries devel­ op. Indeed, among the cold-blooded vertebrates such as fish, tur tles, and all igators there are many species in which

DEVELOPME NTAL PLASTICITY A N D SYMBIOSIS

(A)

665

(B)

(e)

FIGURE 18.6 Preda tor- i nd uced polyphenism in the red-eyed tree frog Agalychnis callidryas. (A) When a snake eats a elutch of Aga/ychnis eggs, most of the remaining embryos inside the egg mass respond to the vibrations by hatching prematurely (arrow) and falling into the water. (8) I mmature tadpole, induced to hatch at day S. (C) Normal tadpol es hatch at day 7 and have better-developed musculature. (Cou rtesy of K. Warkentin.)

the environment determines whether an individual is male or female (Crews and Bull 2009). This type of environmen­ tal sex determination has advantages and disadvantages. One advantage is that it probably gives the species the ben­ efits of sexual reproduction without tying the species to a 1 : 1 sex ratio. In crococliles, in which extreme temperatures produce females while moderate temperatures produce males, the sex ratio may be as great as 10 females to each male (Woodward and Murray" 1993). 1n such instances, where the number of females limits the population size; this ratio is better for survival than the 1:1 ratio demand­ ed by genotypic sex determination. The major disadvantage of temperature-dependent sex determination may be its narrowing of the temperature lim­ its within which a species can persist. Thus thermal poUu­ tion (either locally or due to global warming) could con­ ceivably eliminate a species in a given area Uanzen and Paukstis 1991). Researchers (Ferguson and Joanen 1982; Miller et al. 2004) have speculated that dinosaurs may have had temperature�dependent sex determination and that their sudden demise may have been caused by a slight dlange in temperature creating conditions wherein only males or only females hatched. (Unlike many turtle species, whose members have long reproductive lives, can hiber­ nate for years, and whose females can store sperm, dinosaurs may have had a relatively narrow time to repro­ duce and no ability to hibernate through prolonged bad times.) See WEBSITE 1 8.2 Volvox: When heat brings out sex

Charnov and Bull (1977) argued that environmental sex determination would be adaptive in those habitats ellar­ acterized b y patchiness-that is, a habitat having some regions where it is more advantageous to be male and other regions where it is more advantageous to be female. Conover and Heins ( 1987) provided evidence for this hypothesis. In certain fish species, females benefit from being larger, since larger size translates into higher fecun­ dity. If you are a female Atlantic silverside fish (Menidia menidia), it is advantageous to be born early in the breed­ ing season, because you have a longer feeding season and thus can grow larger. (The size of males in this species doesn't influence mating success or outcomes.) In the southern range of Menidia, females are .indeed born early in the breeding season, and temperature appears to play a major role in this pattern. However, :in the northern reach­ es of its range, the species shows no environmental sex determination. Rather, a 1:1 sex ratio is generated at all tempera tures. Conover and Heins speculated tha t the more north.em populations have a very short feeding season, so there is no advantage for females in being born earlier. Thus, this fish has environmental sex determination in those regions where it is adaptive and genotypic sex deter­ mination in those regions where it is not. BUTTERFLY WINGS 1n tropical parts of the world, there is often a hot wet season and a cooler dry season. In Africa, a polyphenism of the dimorphic Malawian butterfly (Biclj­ elliS anljnafla) is adaptive to seasonal changes. The dry (cool) season morph is a mottled brown butterfly that sur­ vives by hiding in dead leaves on the forest floor. In con­ trast, the wet (hot) season morph, which routinely flies, has prominent ventral eyespots that deflect attacks from predatory birds and lizards (Figure 1 8.7).

666

CHAPTER 1 8

Dry-seas on form

Decreased amount of20-hydroxyecdysone OH HO '

� 24·C

Larva

OH

o

Increased amount of 20-hydroxyecdysone OH HO '

o

Distal-less expression

in imaginal disc OH

OH HO '

Wet-season form

OH _

o

FIGURE 18.7 Phenotypi c plasticity in Bicyclus anynana is regu­ lated by temperature during pupation. High temperature (either i n the wild o r in c ontroll ed laboratory conditions) allows t h e accu­ mulation of 20 hydrox yecdysone (20E), a hormone that is able to sustain Distal-less expression i n the pupal imaginal disc. The region of Distal-less expression becomes the focus of each eye­ spot. In cooler weather, 20E is not formed, Distal-Jess expression in the imaginal disc begins but is n ot sustained, and eyespots fa il to form. (Courtesy of S. Ca rrol l and P. Brakefield.) -

The importance of hormones such as 20E for mediating env ironmental signals controlling wing phenotypes has been docmnented in the Araschnin butterfly (Figure 1 8.8). Araschnia develops aIternative phenoty pes depending on whether the fourth and fifth instars experience a photope­ riod (hours of daylight) that is longer or shorter than a par­ ticular critical day length. Below tnis critical day length, ecdysone levels are low and the butterfly has the orange wings characteristic of spring butterflies. Above the criti­

The factor determining the seasonal pigmentation of B. nnynal1a is not diet, but the temperature d uring pupation. Low temperatures produce the dry-season morph; higher temperatures produce the wet-season morph (Brakefield and Reitsma 1991). The mechanism by which temperature regulates the Bicyclus phenotype is becoming known. In the late larval stages, transcription of the distal-less gene in the wing imagin al discs is restri cted to a set of cells that will become the signaling center of each eyespot. In the early pupa, higher temperatures elevate the formation of

20-hydroxyecdysone (20E; see Chap ter )5). This hormone sustains and expands the expression of distal-less in those regions of the wing imaginal disc, result.ing in prominent eyespots. In dry season, the cooler temperatures prevent the accumula tion of 20E in the pupa, and the foci of Dis­

tal-less signaling are not sus tained . In the absence of the

Distal-less signal, th e eyespots do not form (Brakefield et at.

1996; Koch et at. 1996). Distal-less protein is believed to

be the activating signal that determines the size of the eye­ spot (see Figure

18.7).

FIGURE 18.8 Environmentally induced morphs ofthe European map butterfly (Araschnia levana). Th e orange morph (bottom) forms i n the spri ng when levels of ecdysone in the la rva are low. The dark morph with a white stripe (top) forms i n summer, when h igher temperatures and longer photoperiods induce greater ecdysone p roduction in the larva. Linnaeus classified the two morph s as different species. (Courtesy of H. F. Nijhout.) ,

cal point, ecdysone is made and the sum mer pigmenta tion forms. The summer form can be ind uced in spring pupae by injecting 20E into the pupae. Moreover, by altering tne timing of 20E injections, one can generate a series of inter­

mediate forms not seen in the wild (Kodl and BUckmann

1987; Nijhout 2003).

DEVELOPMENTAL PLASTICITY A N D SYMBIOSIS

Environmental I nduction of Behavioral Phenotypes

667

� • Low maternal care

• Ell Plentiful maternal care

In many instances, the morphological phenotype is accom­ panied by a behavioral phenotype. This is obvious in the

l

envirorunental determination of sex, where an individual's sexual behavior generally matches the gonads and geni­

100 90

"

80

70

.g � 60 >-

talia. This is also seen in the cases of butterfly wings (fliers vs. crawlers) and dung beetle horns (fighters vs. "sneak­

-5 50 � E 40 � .S 30 � 8- 20 10 o '---u, -- .t ----.t - .-'-'--

ers"). Sometimes, however, the behavior is the major devel­ opmental phenotype induced by the environment.

Adult anxiety and environmen tally regulated DNA methylation

G chapter can be

www.devbio.com

Abzhanov, A., W. P. Kuo, C. Hartmann, P. R Grant, R Grant and C. Tabin. 2006. The calmodulin pathway and evolution of beak morphology in Darwin's finch­ es. Nature 442: 563-567.

Epigenetics, Medicine alld Evolutian. Sin­ auer Associates, Sunderland, tvtA.

Amundson, R 2005. The Changing Role of the Embryo ill Evolutionary Thought: Roots of Evo-Devo. Cambridge Universi­

Lynch, V. j., A. Tanzer, Y. Wang, F. C. Leung, B. Gelelrsen, D. Emera and G. P. Wagner. 2008. Adaptive changes in the transcription f,letar HoxA-l1 are essen­ tial for the evol ution of pregnancy in mammals. Pmc. Natl. Acad. Sci. USA. 105: 14928-14933.

ty Press, New York.

Carroll, S. B . 2006. Endless Forms Most Beautiful: The New Science of Eva-Dew. Norton, New York. Cohn, M. j. and C. Tickle. 1999. Devel­ opmental basis of limblessness and axial patterning in snakes. Nature 399: 474-479. Gilbert, S. F. and D. Epel. 2009. Ecologi­ cal Developmental Biology: lntegrating

,

Kuratani, S. 2009. Modularity, compara­ live embryology and eva-cleva: Devel­ opmental dissection of body plans. Dev. BioI. 332: 61-69.

Merino, R,j. Rodriguez-Leon, D. MaCias, Y. Ganan, A. N. &onomides and j. M. Hurle. 1999. The BMP antago­ nist Gremlin regulates outgrowth, chon­ drogenesis and programmed cell death

in the developing limb. Development 126: 5515-5522. Rockman, M. v., M. W. Hahn, N. Soran­ zo, D. B. Goldstein and G. A. Wray. 2003. Positive selection on a hurnan­ specific transcription factor binding site regulating IL4 expression. Curro BioI. 13: 2118-2123. Shapiro, M. D. and 7 others. 2004. Genetic and developmental basis of evolutionary pelvic reduction in truee­ spine sticklebacks. Natllre 428: 717-723. Shubin, N., C. Tabin and S. B. Carroll. 2009. Deep homology and the origins of evolutionary"novelty. Nature 457: 818-823. Smith, K. 2003. Time's arrow: Hetero­ chrony and the evolution of develop­ ment. Int. J. Dev. BioI. 47: 613-621.

Go Online Relating evolution to development in

have been the precursor of both protostomes and deuteros­

the nineteenth century. Immediately after Darwin's Ori­

tomes. This organism may have resembled the planula lar­

gin of Species biologists attempted to relate evolution to

vae of contemporary cnidarians.

WEBSITE 19.1

changes in development. Here the attempts of three such scientists-Frank Lillie, Edmund B. Wilson, and Ernst Haeckel-are highlighted.

WEBSITE 19.2

Correlated progression.

In many cases,

modules must co-evolve. The upper and lower jaws, for

WEBSITE 19.4

How the chordates got a head. The neu­

ral crest is responsible for forming the heads of chordates. But how did the neural crest come into existence? It is probable that ancestral deuterostomes had all the requisite genes, but only in the chordates did these genes become

instance, have to fit together properly. If one changes, so

linked together into the network that became the neural

must the other. If the sperm-binding proteins on the egg

crest cell.

change, then so must the egg-binding proteins on the sperm. This site looks at correlated changes during evolu­ tion.

WEBSITE 19.5

"Intelligent design" and evolutionary

developmental biology. Evolutionary developmental biol­ ogy explains many of the "problems" (such as the evolu­

The seach for the Urbilaterian ancestor.

tion of the vertebrate eye and the evolution of turtle shells)

Homologous genes specifying the formation of the eye,

that proponents of "intelligent design" and other creation­

heart, body axis, and nervous system enable biologists to

ists claimed were impossible to explain by evolution.

WEBSITE 19.3

intuit an "Urbilaterian ancestor"-an organism that may

G lossary

Allantois

A Achondroplasic d wa rfism

Condition

wherein chondrocytes stop proliferat­ ing earlier than usual, resulting in short limbs. Often caused by muta­ tions that activate FgJR3 prematurely.

Drosophila head; includes the brain. Acrosome (acrosomal vesicle) Cap­

Aeron

The terminal portion of the

like organeUe containing proteolytic enzymes that, together with the sperm nude us, (orms the sperm head.

The Ca2+--depend­ ent fusion of the acrOSOlfle with the sperm cell membrane, resulting i n exo­ cytosis and release of proteolytic enzymes that allow the sperm to pene­ tTate the egg extraceUular matrix and

Acrosome reaction

fertilize the egg.

Ade pith el ial cells

Cells that migrate into the insect imaginal disc early in development. During the pupal stage, these cells give rise to the muscles and nerves that serve the leg. Attachment between cells or between a cell and its extracellular substrate. The latter provides a surface

Adhesion

for migrating cells to travel along.

Adult stem cells

Stem cells found in

the tissues of organs after the organ has matured. Adult stem celJs are usually involved in replacing and repairing tis­ sues of that particular organ. and can form only a subset of ceU types.

Afferent ne u ron s

Neurons that carry inIonnation from sensory receptor cells (e.g .• sound waves from the ear, light signals from the retina. touch or pain sensations from the skin) to the central nervous system (spinal cord and brain).

Agi n g

The time-related deterioration of the physiological functions necessary for survival and fertility.

In aITUl.iote species, extraem­

bryonic membrane that stores urinary wastes and helps mediate gas exchange. It is derived from splanchnopleure at the caudal end of the primitive streak. ln mammaLs, the size of the allantois depends on how well nitrogenous wastes can be removed by the chorionic placenta. In I'ep l'i les and birds, the allantois becomes a large sac, as there is no other way to keep the toxic by-prod­ ucts of metabolism away from the developing embryo.

Allometry Developmental changes that occur when diHerent parts of an organ­ ism grow at different rates.

Alternative splicing

A means of pro­

ducing multiple different proteins

encoded by a single gene by splicing together different sets of exons to gen­ erate different types of mRNAs.

AmetaboloU5 A pattern of insect devel­ opment in which there is no larval stage and the insect undergoes direct development to a small adult form fol10wh'g a transitory pronymph stage.

A rn n i ocent.esis Medical proced ure that removes a sample of amnionic fluid around the fourth or fifth month of pregnancy. Fetal ceUs from the fluid are cultured and analyzed for the pres­ ence or absence of certain chrol1lo­ somes, genes, or enzymes. " Wa ter sac." A membrane enclosing and protecting the embryo

Amnion

and its surrounding amnionic fluid. Defining the "aITUl.iote vertebrates," this epithelium is derived from somatopleure. Edodermcil tissue sup­ plies epithelial cells, and the l1lesoclenn generates the essential blood supply. fluid A secretion that serves as a shock absorber for the developing embryo while preventing it from dry­ ing out.

Amnionic

Vertebrates whose embryos form an am.n.ion: the reptiles, birds,

Amniotes

and mammals.

Latin, "flask." The segment of the mammalian oviduct, distal to the uterus and near the ovary, where fertil­ ization takes place.

Ampulla

A n al ogo us

SLTuctures and/or their respective components whose similari­ ty arises from their performing a simi­

lar function rather than their ariSing from a common ancestor (c.g., the wing of a butterfly vs. the wing of a bird). Compare with homologous.

The cell connecting the overlying gonad to the vulval precur­ sor celJs in C. degans. If the anchor cell is destroyed.. the VPCs will not form a

Anchor cell

vulva. but instead become part of the hypodermis (skin).

Androgen insensitivity syndrome

lnterscx condition in which an XY individual has a mutation i n the gene encoding the androgen receptor pro­ tein that binds testosterone. This results in a female external phenotype, lack of a uterus and oviducts and pres­ ence of abdominal testes.

Anencephaly

A lethal congenital

defect resulting from failure to dose the anterior neuropore. The forebrain remains in contact with the am.n.ionic fluid and subsequently degenerates, so the vault of the skull fails to form.

Angioblasts From allgio. blood vessel; and blnst, a rapidly dividing cell (usu­

ally a stem cell). The progenitor ceUs of

blood vessels.

Angiogenesis

Process by which the

primary network of blood vessels cre­ ated by vasculogenesis is remodeled and pruned into a distinct capillary bcd, arteries, and veins.

G-2

GLOSSARY

Angiopoietins P�racrine factors that mediate the interaction between endothelial cells and pcricy les. Animal cap In a mph i bia ns, the roof of the blastocoel (in the animal hemi­ sphere). Animal hemisphere The non-yolk-con­ taining (upper) haJJ of the amphibian egg. During embryogenesis, cells in the animal hemisphere divide rapidly and become actively mobile ("nnimated"). Anoikis Rap id apoplosis that occurs when ep ithelia l ceUs lose their attach­ ment to the extracellular matrix. Antennapedia complex A region of Drosophila chromosome 3 containing the homeotic genes labial (lab), Anlell­

/fopet/in (Antp), sex combs reduced (sa), defonned (dfd), and probcscipedia (Pb),

specify head and thoradc seg­ ment identities. Anterior heart lield Cells 01 the heart field forming the outflow tract (conus and ITuncus arteriosus, right ventricle). Anterior intestinal portal (AlP) The posterior opening of the developing foregut region of the primitive gut tube; it opens into the future midgut region which is contiguous with the yolk sac at this stage. Anterior visceral endoderm (AVE) Mammalian equivalent to the chick hypoblast and similar to the head POI­ Han of the amphibian organizer, i t cre­ ates an anterior region by secreting antagonists of Nodal. Anterior-posterior (anteroposterior) axis The body axis extending from head to tail (or mouth 1"0 anus in those organisms that lack a head and tail). When referring to the limb, this refers . to the thumb (anterior)-pinkie (posteri­ or) axis. Anti-Mullerian factor (AMF) TGF-� family paracrine faclor secreted by the embryoniC testes thai induces apopto­ sis of the epithelium c'\nd destruction of the basal lamina of the Mullerian duct, preventing formation of the uterus and oviducts. Fonnerly known as anti­ MUllerian hormone, or AMH. Anurans Frogs dJld toads. Aorta-gonad-mesonephro5 region (AGM) A mesnchymal area in the lateral plate splanchnopleure near the aorta that produces the controversial putative definitive hematopoietic stem cells of the fetus and adult. Aortic arches These begin as symmet­ ri ca lly arranged, pclired vessels that develop within the paired pharyngeal arches and link the ascending and descending/d orsal paired aortae. Some arches degenerate. Left aortic arch IV becomes the Arch of the Aorta which

and aortic arch VI is the ductus arterio­ sus on the left. Apical ectodermal ridge (AER) A ridge along the distal margin of the limb bud that will become a major sig­ naling center for the developing limb. Its roles include (1) maintaining the mesenchyme beneath it in a plastic, proliferating state that enables the lin­ ear (proximal-distal) growth 01 the limb; (2) main taining the expression of those molecules that generate the ante­ rior-posterior axis;