International Review of Cytology

VOLUME 142 SERIES EDITORS Geoffrey H. Bourne James F. Danielli Kwang W. Jeon Martin Friedlander 1949-1988 1949-1 984...

Author: Author Unknown

10 downloads 972 Views 20MB Size Report

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

DOWNLOAD PDF

VOLUME 142

SERIES EDITORS Geoffrey H. Bourne James F. Danielli Kwang W. Jeon Martin Friedlander

1949-1988 1949-1 984 19671984-

ADVISORY EDITORS Airnee Bakken Eve Ida Barak Howard A. Bern Robert A. Bloodgood Dean Bok Stanley Cohen Rene Couteaux Marie A. DiBerardino Donald K. Dougall Charles J. Flickinger Nicholas Gillham Elizabeth D. Hay Mark Hogarth Keith E. Mostov

Audrey Muggleton-Harris Andreas Oksche Muriel J. Ord Vladimir R. Pantic' M. V. Parthasarathy Lionel I.Rebhun Jean-Paul Revel L. Evans Roth Jozef St. Schell Hiroh Shibaoka Wilfred Stein Ralph M. Steinrnan M. Tazawa Alexander L. Yudin

Edited by

Kwang W. Jeon Department of Zoology The University of Tennessee Knoxville. Tennessee

Martin Friedlander Jules Stein Eye Institute and Department of Physiology UCLA School of Medicine Los Angeles, California

VOLUME 142

Academic Press, Inc. Harcourt Brace Jovanovich, Publishers San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @ Copyright 0 1992 by ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Academic Press, Inc. 1250 Sixth Avenue, San Diego, California 92101-431 1 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NW1 7DX Library of Congress Catalog Number: 52-5203 international Standard Book Number: 0-12-364545-X

PRINTED IN THE UNITED STATES OF AMERICA

92 93 94 95 96 97 98

EB

9 8 7 6 5 4 3 2 1

Contributors .........................................................................

ix

Biological Problems of Regenerative Cementogenesis: Synthesis and Attachment of Collagenous Matrices on Growing and Established Root Surfaces Hubert E. Schroeder I. Introduction ......................................... II. Origin, Types, and Function of Root Cementum on Hum 111. Spontaneous Cementogenesis and Matrix Formation on Growing ...... Root Surfaces ....................................... IV. Matrix Formation on Established Root Surfaces in Vitro ...... V. Regenerative Cementogenesis on Established Root Surfaces in Vivo . . . . . . . . . . . . . . VI. Concluding Remarks and Perspectives . . . . . . . . . . .......................... References ............................. .....................

1 2 5 28 43 51 52

lmmunocytochemical Localization of Proteins in Striated Muscle Marvin H. Stromer I. II. 111. IV. V. VI.

Introduction ................................ Localization of Proteins in Skeletal Muscle Cells . . . . . . . . . . . . . . . . . . . Localization of Proteins in Cardiac Muscle Cells ................................. Sarcoplasmic Reticulum, Transverse Tubules, and the Sarcolemma . . . . . . . . . . . . . . . Other Proteins .................................... ..................... Conclusions and Outlook ............................ References . . . . . . . . . . . . . . . . .................................. V

61 62 102 119 127 128 129

vi

CONTENTS

Recent Developments in Vertebrate Cell Culture Technology Satish J. Parulekar. Thomas Hassell. and Satish C . Tripathi I. II. 111. IV. V. VI .

....................... Traditional Cultures . . . . . ........................... Three-Dimensional Cultures . . . . . . . . . . ............................ Commercial Scale Bioreactors . . . . . . . . . . . . . . . . .

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

Design and Optimization Considerations ..................... Concluding Remarks . . . . . . . . . . . . ...........................

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

145 147 153 162 192 201 204

Transdifferentiation in Medusae Volker Schmid I II 111 IV

Introduction ....................... ............................ .............. The Concept of Transdifferentiation . . . . . . . . . . . . . . Transdifferentiation in Hydromedusae ...................... Concluding Remarks . . . . . . . ............................ References ............................. ......................

213 214 218 256 258

Symplast as a Functional Unit in Plant Growth Kiyoshi Katou and Hisashi Okamoto ............ I. Introduction ..................................... II. Electrophysiological Structure of the Plant Germ Axis . . . . . . . . . . . 111 . Role of Spatially Separated Proton Pumps in Stem Elongation .................... ..................... IV. Lockhard Equations and Action of Auxin . . . . V . Integration of the Activity of the Symplast in P .................................. VI . ........................................

263 265 273 277 285 299 300

lntracellular Ca2+Messenger System in Plants Shoshi Muto I. Introduction ................................................................... I1. Receptors ..................................................................... Ill. G Proteins ....................................................................

305 306 308

vii

CONTENTS

IV. Regulation of lntracellular Ca2+ Concentration .................................. V . Phosphatidylinositol Turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

................ .......................................................... VIII . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References

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

Index ...............................................................................

311 321 328 332 338 339 359

This Page Intentionally Left Blank

Numbers in parentheses indicate the pages on which the authors' contributions begin.

Thomas Hassell' (145), Celltech Limited, Slough SL1 4EN, Berkshire, England Kiyoshi Katou (263), Biological Laboratory, College of General Education, Nagoya University, Nagoya 464-01, Japan Shoshi Muto (305),Institute of Applied Microbiology, University of Tokyo, Tokyo 113, Japan Hisashi Okamoto (263), Biology Department, Graduate School of Integrated Science, Yokohama City University, Kanazawa 236, Japan Satish J. Parulekar (145), Department of Chemical Engineering, lllinois Institute of Technology, Chicago, Illinois 60616 Volker Schmid (213), Institute of Zoology, Pheinsprung 9, CH-4051 Basel, Switzerland, and Friday Harbor Laboratories, Friday Harbor, Washington 98250 Hubert E. Schroeder ( l ) ,Department of Oral Structural Biology, Dental lnstitute, University of Zurich, CH-8028 Zurich, Switzerland Marvin H. Stromer (61), Department of Animal Science, Muscle Biology Group, Iowa State University, Ames, Iowa 5001 1 Satish C. Tripathi (1 45), Department of Life Sciences, IlT Research Institute, Chicago, lllinois 60616 'Present address: iAF BioVac Inc., Ville de Laval, Quebec, Canada H7N 422.

ix


Biological Problems of Regenerative Cementogenesis: Synthesis and Attachment of Collagenous Matrices on Growing and Established Root Surfaces Hubert E. Schroeder Department of Oral Structural Biology, Dental Institute, University of Zurich, 8028 Zurich. Switzerland

1. Introduction

During the past decade, regeneration of periodontal tissues has received increasing and worldwide attention, from both oral biologists and dental clinicians. “Periodontal regeneration” is defined as the restoration of the various components of the periodontium, i.e., alveolar bone, periodontal ligament, root cementum, and gingiva lost due to disease, “in their appropriate locations, amounts, and relationships to each other” (Aukhil, 1991). In contrast to more simple goals such as the reattachment of the periodontal ligament and supraalveolar connective tissue to the torn cell/fiber tissue at the dental root surface, following their short-term separation, or as the spontaneous repair resulting from unguided wound healing, “periodontal regeneration” requires an enormously well-conducted action of various cell populations to appear and function in space and time in order to reconstitude both structural normality and functional integrity. As the development, structure, and function of the human periodontium, being dissimilar in details to that of laboratory animals such as rodents, dogs, and non-human primates, are immensely complex biologically and not entirely understood (Schroeder, 1986), clinical and laboratory experiments in humans and other animals designed to study the regenerative potential of the various tissue components under conditions of different defects and treatment modalities are exceedingly difficult to analyze. In fact such experiments necessarily address the periodontium as a whole rather than being able to examine separately the response of component tissues. In addition, the situation is rendered even more complex by the InlPrnati~lnu/Reuirn, of

Cyro/oas. V o / . 142

1

Copyright Q 1992 by Academic Press. Inc. All rights of reproduction in any form reserved.

2

HUBERT E. SCHROEDER

fact that the creation of artificial defects or the pretreatment of spontaneous lesions caused by bacterial infection and inflammation introduce artificial tissue alterations that interfere with repair and regeneration. For these reasons, clinical and laboratory experiments on periodontal regeneration, reviewed by Egelberg (1987), Nyman et al. (1989), and Minabe (1991), have yielded widely different results that remain insufficient from a clinical, and invalid from a biological, point of view, results that to a varying extent appear as biologically undecodable messages from a blackbox world. Not surprisingly, periodontal regeneration is loaded with problems, clinical, such as reinfection and mechanical disturbance of the wound healing processes, as well as biological. The latter include the physical and chemical denaturation of treated root surfaces; epithelial migration over the exposed root surface sites that should be the substrate for regenerative tissue formation; the different turnover and growth potentials of the four periodontal tissue components; and innumerable problems regarding cells, mediators, growth factors, etc. (Abdallah et al., 1988; Terranova er al., 1989a; Messadi and Bertolami, 1991; Aukhil, 1991). One reason for the unsatisfying situation was and still is the fact that human cementogenesis remained undiscovered for too long. Indeed, at the beginning of the research focusing on periodontal regeneration, our knowledge of dental root cementum was incredibly meager, resting on antique rather than medieval information. Root cementum represents, however, the most cardinal periodontal tissue component that is primary and indispensible for regenerating the tooth-bone connection, i.e., for reconstituting tooth anchorage. This review provides some of the missing data needed to discuss spontaneous and regenerative cementogenesis and unveils some of the biological problems involved.

II. Origin, Types, and Function of Root Cementum on Human Teeth

A new and increasingly accepted classification of root cementum on human teeth was proposed by Jones (1981) and with modifications adopted by Schroeder (1986). It differentiates among four varieties according to the absence or presence of cells and to the source of collagen fibers, i.e., the major matrix component contained within. Consequently, this classification distinguishes between acellular afibrillar cementum (AAC); acellular extrinsic fiber cementum (AEFC); cellular intrinsic fiber cementum (CIFC) that may also occur as an acellular variety (AIFC; Bosshardt and Schroeder, 1990);and cellular, mixed stratified cementum (CMSC). These varieties are summarized in Table I.

3

REGENERATIVE CEMENTOGENESIS TABLE I Types of Human Root Cementum

Terms

Abbreviation

Acellular, afibrillar cementum

AAC

Acellular, extrinsic fiber cementum

AEFC

Cellular, intrinsic fiber cementum

CIFC

Acellular, intrinsic fiber cementum

AIFC

Cellular, mixed, stratified cementum (AEFC + CIFCI AIFC)

CMSC

Organic components Homogeneous matrix, no cells, no collagen fibrils Collagen fibrils as Sharpey’s fibers, no cells Intrinsic collagen fibrils and fibers, cementocytes

Intrinsic collagen fibrils and fibers, no cells Intrinsic collagen fibrils and fibers, collagen fibrils as Sharpey’s fibers, cementocytes

Location At dentinoenamel junction, on enamel Cervical to middle root

Function Unknown

Tooth anchorage

Apical and interradicular root surfaces, resorption lacunae, fractures Apical and interradicular root surfaces

Adaptation, repair

Apical and interradicular root surfaces

Adaptation, root anchorage

Adaptation

Apart from a homogeneous and mineralized ground substance of unknown composition, AAC contains neither cells nor collagen fibrils. In humans, it is found as coronal cementum covering patchwise the cervical enamel, and as an occasional part of cervical AEFC (Schroeder 1986, 1988). Its origin and function are unknown, but it may represent serumderived organic material coprecipitated with mineral (Beertsen and Van Den Bos, 1991). The AEFC lacks cells and is composed entirely of densely packed, well-oriented bundles of collagen fibrils, i.e., the so-called fibers of Sharpey. These fibers continue into the periodontal ligament and connect the root to the alveolar bone. Thus, all AEFC fibers are extrinsic. About 30,000 fibers insert into 1 mm2 of AEFC surface, each fiber being about 4 pm in diameter (Schroeder, 1986). In humans, the rather thin (20 to 250 pm), densely mineralized AEFC shows parallel incremental lines and is found primarily on the cervical and middle root regions, but it may extend further apically. It is formed by fibroblasts of the dental follicle

4

HUBERT E. SCHROEDER

proper, i.e., ectomesenchymal derivative, and later of the periodontal ligament, and serves exclusively for tooth anchorage (Schroeder 1986, 1988). The CIFC contains cells, the cementocytes, but its collagen fibers, being again the major matrix component, are all intrinsic and run a circular or spiral course around the root, i.e., more or less parallel to the root surface. Thus, CIFC lacks Sharpey's fibers. In humans, the less well mineralized CIFC is found mainly in situations of repair, filling resorption lacunae or connecting root fracture fragments. It is also part of CMSC, forming initially on apical root portions and patchwise between layers of AEFC. The CIFC is formed by cementoblasts of the dental follicle proper and later of the periodontal ligament; its function is associated with repair and adaptation. On particular portions of the root, AIFC may form without leaving cementocytes behind. CMSC is a mixture of pure AEFC and CIFC/AIFC; the latter part may contain cementocytes with uneven distribution and density. The CMSC is usually a stratified tissue, with consecutive or alternating layers of AEFC and CIFC/AIFC being unpredictably superimposed on one another. In humans, the inhomogeneously mineralized and, in part, porous CMSC is variably thick, ranging from 100 to 600 p m or more, and occurs primarily in the apical third of the roots and in the furcations. It serves the functions of adaptation, i.e., adynamic reshapeningofthe root surfaceas the tooth shifts and drifts in its socket, and, if superficially covered by AEFC, of root anchorage (Table I; Schroeder, 1986, 1988). Measurements of sequential fluorochrome labeling lines in all deciduous teeth and the permanent molars of one -13-month-old M . fascicularis monkey (Bosshardt er al., 1989) and in the alveolar bone surrounding the first molars (Schroeder er al., 1992) provided data for the formation rate of two cementum varieties, in comparison to that of dentine, and bone TABLE II Rates of Formation of Cementum, Dentine, and Bone

Tissue

Abbreviation

Acellular, extrinsic fiber cementum Cellular, intrinsic fiber cementum Initial layer Appositional layers Crown dentine (first molars) Root dentine (deciduous teeth) Root dentine elongation (deciduous teeth) Alveolar bone crest (first molars) Alveolar bone septum (first molars)

AEFC CIFC

CD RD RDE ABC ABS

Formation-rate (pmlday)" x + s

< 0.10 + 0.02 0.4-3. I 0.1-0.5 3.1 k0.2 2.7-4.6 12.0-36.0 5.0-14.0 I3 .O-22.0

" Measured in one sequentially fluorochrome-labeled M . fuscicularis monkey; from Bosshardt et a / . (1989) and Schroeder et a / . (1992).

REGENERATIVE CEMENTOGENESIS

5

and to root elongation (Table 11). These data demonstrated that AEFC is an extremely slowly forming tissue, initially as well as later in life (Sequeira et al., 1992). In humans, its daily rate of formation, i.e., increase in thickness, is smaller that 0.1 p m , possibly as low as 0.005 to 0.01 p m (Dastmalchi et al., 1990; Sequeira et al., 1992). In contrast, initial CIFC is formed at a fast rate, ranging from 3.1 to 0.4 pm/day. Subsequently, appositional cementum layers, possibly of the AIFC variety, may still form at a faster rate than AEFC, i.e., 0. I to 0.5 pm/day. Comparatively, CIFC may form as rapidly as crown and root dentine and not much slower than alveolar bone (Table 11). 111. Spontaneous Cementogenesis and Matrix Formation on Growing Root Surfaces

In contrast to amelogensis and dentinogenesis, i.e., to rather well-defined developmental systems characterized by particular classes of cells and their morphologically and biochemically defined matrix products, cementogenesis on human teeth was essentially unknown until 1985, although some fragmentary information was available for other mammalian species such as rodents. This information had been derived from studies on mouse incisors and molars (Selvig 1963, 1964, 1967) and on rat molars (Paynter and Pudy, 1958; Diab and Stallard, 1965; Lester, 1969; Formicola et al., 1971 ;Owens, 1980). Although such molars are also covered by both acellular and cellular cementum, on their roots, and albeit most recent investigations of Cho and Garant (1988, 1989) and Yamamoto and Wakita (1990, 1991, 1992) demonstrated some similarity in matrix production and attachment to dentine, there are a number of reasons for the argument that root development and cementogenesis in rodent molars might be unlike that in humans (see below). Therefore, this review focuses primarily on cementogenesis in human teeth and recent observations in rodents will be used only comparatively. A. Acellular Extrinsic Fiber Cementum

In human teeth, AEFC covers the cervical root surfaces and extends from the cemento-enamel junction apically. In single-rooted teeth (i.e., incisors, canines, and most premolars), AEFC coats 60 to 90% of the total root length that varies between 13 (central incisors) and 15.5 mm (canines; Schumacher and Schmidt, 1983; Schroeder, 1988). AEFC is first formed while the roots develop. AEFC formation begins at and along the growing root edge and the AEFC slowly increases in thickness in the coronal direction. As shown in human premolars with incomplete roots developed

6

HUBERT E. SCHROEDER


7

50 to 60% of their final length (Bosshardt and Schroeder, 1991a), initiation of AEFC formation, early matrix production, and its attachment to root dentine all take place within a zone of about 300 pm, extending coronally from the advancing root edge (Fig. 1). This zone probably first develops at the outer surface of the initially formed root portion and later shifts in the apical direction as the root grows in length. As a consequence, thin layers of established AEFC are encountered in coronal root regions, while initial AEFC genesis continues apically, i.e., near the advancing root edge. From data defining the time period necessary for root development, it is known that human premolars attain their complete root length within about 6 years following the completion of their crown (Schroeder, 1987). Based on a rough calculation, the time period during which AEFC is formed initially on the growing root spans 43 to 65 months, i.e., 60 to 90% of 6 years. Because AEFC coats 8 to 10 mm (i.e., 60 to 75%) of the final root length in human premolars, AEFC development at and along the growing root may proceed at a rate of 4.6 to 6.9 pm/day (Schroeder, 1987, 1988). These rates, calculated from clinical and morphometric measurements rather than based on direct evidence which is unavailable, are very much lower than the rates of root elongation measured in all deciduous and permanent molar teeth of the M.fuscicularis monkey, ranging between 12 and 36 pmlday (Table 11). As a preliminary estimate, it can be inferred from these data that initiation of AEFC formation within the 300-pm-wide zone proceeds apically with an average speed of 5 to 6 pmlday. In all probability, AEFC is a product of a particular class of fibroblasts (Beertsen and Everts, 1990; Bosshardt and Schroeder, 1991a,b). The advancing root edge includes the leading edge of newly produced predentine and the inorganic edge of mineralized dentine. The latter follows the former within a short distance of up to 50 pm. From this edge, predentine continues both over the pulpal surface along the dentine-odontoblast interface and over the external surface of the newly formed root dentine. At the latter site, it can be followed coronally over a distance of about 250

FIG. 1 Schematic drawing illustrating topographically the initial stages of AEFC genesis on human premolars developed to 5040% of their final root length: 1, fibroblasts contact root/ predentine and become committed: 2, fibroblasts start to form and attach collagen fibrils: 3, inital fiber fringe with maximum fiber density is established: 4, cell fiber fringe meshwork is established and the mineralization front approaches the base of the fringe: 5, mineralization front progresses into initial fiber fringe. AEFC. acellular extrinsic fiber cementum: MD, mineralized dentine: ERM, epithelial rests of Malassez: FPF, fringe-producing fibroblasts: FF, collagenous fiber fringe; PD, predentine;MF, mineralizationfront; NMD, nonmineralized dentine or predentine;CF, committed fibroblasts;ARE, advancing root edge; HRS, Hertwig’s epithelial root sheath;FFB, fiber fringe base. Modified from Bosshardtand Schroeder (1991 a).

HUBERT E. SCHROEDER


9

to 300 pm from the advancing edge (Fig. 1). At this edge, the diaphragm, i.e., the most apical part of Hertwig’s epithelial root sheath, touches the predentine but, lateral or external to this edge, the root sheath deviates from the surface of newly formed dentine, continues coronally as a short strand, and eventually breaks up into the discontinuous epithelial rests of Malassez (Fig. 1). In humans, Hertwig’s root sheath, including its diaphragm, consists of the former inner and outer layers of the enamel epithelium, extends by continuous proliferation of the diaphragm (Diab and Stallard, 1965; Kenney and Ramfjord, 1969; Formicola et al., 1971), disintegrates coronally in accordance with its rate of proliferation, and is surrounded by a basal lamina. The latter actually contacts the leading root edge of predentine (Schroeder, 1986). In contrast to previous statements in most current textbooks, Hertwig’s root sheath does not cover much of the external surface of newly formed predentine, at least in human premolars. Rather, that surface at the advancing root edge is almost from its beginning accessible to connective tissue cells of the dental follicle proper. In the triangular region between the laterally deviating root sheath and the surface of newly formed predentine, connective tissue cells with the morphological appearance of fibroblasts can always be encountered. These cells are slender or bulky, are basophilic, and reveal an activated euchromatin-rich nucleus, displaying an -50-nm-thick nuclear fibrous lamina and a cytoplasm with numerous strands of rough endoplasmic reticulum cisternae and a prominent Golgi field. These cells are connected to one another by desmosome-like junctions and project numerous, slender cytoplasmic processes that contact and insert between collagen fibrils of the not yet mineralized outer predentine matrix (Fig.2a). These features are typical for the most apical 30 to 50 pm along the surface of the newly formed root. In about that distance to the advancing root edge, cells of similar appearance begin to produce the first AEFC matrix portions in the form of tiny but discrete bundles of collagen fibrils, slowly increasing in length and density in the apico-coronal direction. Thus, a short fiber fringe is produced, with most of the collagen fibrils arranged in parallel and oriented more or less perpendicular to the root surface. These fibrils

FIG. 2 (a) Electron micrograph depicting the most apical zone of not yet mineralized predentine matrix (PD) contacted and penetrated by cytoplasmic processes (CP) of fibroblast-like cells ( F )that begin to produce cementa1 collagen fibrils (see Fig. I , parts 1 and 2). (b) Electron micrograph of the initial fringe of collagen fibrils and fibril bundles (FF) that insert into the not yet mineralized predentine (PD);the mineralization front (MF) has not reached the future dentino-cementa1 junction (see Fig. 1, parts 3 and 4). Magnification: a,b, x6700. (a) From Bosshardt and Schroeder (1991a). (b) Courtesy Dr. D. D. Bosshardt.

10

HUBERT E. SCHROEDER


11

terminate within and intermingle with the collagen fibrils of the dentinal matrix (Fig. 2b). Because of their orientation and parallelity, the fringe fibers are well demarcated from the randomly arranged fibrils of predentine. Within about 200 pm from the advancing root edge, this fringe attains maximum density (i.e., 130-150 fibers per I mm root surface length), which remains constant further coronally (Bosshardt and Schroeder, 1991b; Sequeira et al., 1992). In other words, the base of the fringe at its interface with predentine is fully established within a distance of 200 pm, or, based on the above calculation, within about 30 days of matrix production (Figs. 2b and 3a). Matrix production is not fully understood at present, but it is apparently carried out by the one to three cell layers that form a three-dimensional interwoven cell-fiber fringe meshwork. It is possible that the sequential, intracellular events in the synthesis and secretion of collagen are analogous to those described for human gingival fibroblasts (Yajima et al., 1980), for periodontal ligament fibroblasts in young Balb-C mice (Cho and Garant, 1981)and in 20-day-old rats (Marchi and Leblond, 1983, 1984), and for osteoblasts and odontoblasts in 20-day-old rats (Weinstock and Lebland, 1974;Weinstock, 1975; Leblond, 1989). In these cells, the precursors of Type-I collagen are synthesized in the rough endoplasmic reticulum and processed along the Golgi-secretory granule pathway, the resulting procollagens being released by exocytosis at the cell surface. Extracellularly , these procollagens are then transformed into fibrillar collagen. However, the particular class of fibroblasts forming the initial fiber fringe that represents the first AEFC matrix, in addition to synthesis and release of procollagen, is also engaged in fibril assembly, bundle formation, and fibril bundle orientation. Whereas the assembly and orientation of collagen fibrils is a primary task, bundle condensation and elongation is a subsequent step, increasing the bundle density and allowing AEFC matrix to increase in thickness. Selvig (1964), studying cementogenesis in albino mice, Schroeder (1986), examining early AEFC on human teeth, and Yamamoto and Wakita( 1992),investigating bundle formation during the genesis of acellular cementum in 20-dayold Wistar rats, observed that in tangential and cross sections through the developing and the established fiber fringe, as well as the fiber matrix covering the first, already mineralized layer of AEFC, the fibroblasts

FIG. 3 (a) Electron micrograph depicting the established cell fiber fringe meshwork with the established fringe (FF) of collagen fibril bundles attached to mineralized dentin (MD); the mineralization front (MF) has surpassed the dentino-cementa1 junction (see Fig. I , part 5 ) . (b) Electron micrograph of a tangential section through the cell-fiber fringe meshwork, with cytoplasmic extensions of the fibroblast-like cells (F)encircling compartments for individual collagen fibril bundles (FB). Magnification: a,b x4470. (a,b) Courtesy Dr. D. D. Bosshardt.

12

HUBERT E SCHROEDER

associated with AEFC genesis form circular cytoplasmic recesses and sheet- or plate-like extensions that partially of fully surround the assembly of cross-cut collagen fibrils (Fig. 3b). In fact, cytoplasmic processes of adjacent cells overlap, forming closed compartments (Yamamoto and Wakita, 1992). The formation of collagen fibril bundles has also been studied in both the corneal stroma and the tendon of chick embryos (Birk andTrelstad, 1984, 1985,1986; Fleischmajer er a / . , 1988; Birk et a/., 1990). In both cases, collagen fibrils are produced and bundled into oriented fibers. Bundling occurs because the extracellular space adjacent to fibroblasts is partitioned by the fibroblast surface and its extensions. Individual compartments, probably assuming the form of a cylinder, that house developing fibril bundles are first formed by surface recesses and are later separated from one another by cytoplasmic extensions. Within such compartments, collagen fibril segments of varying lengths can be added to the already existing group of fibrils (Birk et a/., 1990). By means of linear and lateral fusion of fibril segments, it is assumed that collagen fibrils grow in length and fibers in thickness. Birk e t a / . (1990) stated, “it is our hypothesis that synthesis, posttranslational processing, packaging, discharge, and assembly into fibril segments, fibrils, bundles, and tissue-specific macroaggregates are closely regulated by the cell at each stage,” including “the formation of distinct compartments in which the structural elements of the developing matrix would serve to physically position these elements within the developing matrix.” The generation of dense bundles of collagen fibrils that are packed nearly parallel and are oriented nearly perpendicular to the root surface and their elongation are the major events in AEFC matrix production, both initially and subsequently, as AEFC increases in thickness. With increasing density and diameter of the fibril bundles, cytoplasmic extensions may even retract, allowing lateral bundle coalescence that results in particularly thick collagen fibers (Yamamoto and Wakita, 1992). The macroaggregate of oriented fringe fibers, which is very particular for the AEFC matrix, consisting exclusively of Type I ([al(I)]* a2) collagen and noncollagenous components that amount to about 19% of the total matrix, certainly derives from a guiding activity of that particular class of fibroblasts (Christner er a / . , 1977; Smith er al., 1983). In the rat molar, early stages of AEFC matrix formation have been described differently. Hertwig’s epithelial root sheath covers the newly formed predentine externally until the latter starts to mineralize. Instead of deviating from the growing root edge and the newly formed root surface, as occurs in humans, the epithelial sheath has to break up or be penetrated in order to allow connective tissue cells to arrive at the newly formed root surface (Formicola et al., 1971; Yamamoto, 1986; Cho and Garant, 1988, 1989; Yamamoto and Wakita, 1990). As in humans, the latter retains an I-pm-thick, not yet mineralized, layer of predentine, the matrix of which

-


13

is loosely knit and reveals only a few collagen fibrils (Owens, 1980; Yamamot0 and Wakita, 1990). The first primitive and still loose collagen fibril bundles that penetrate into this superficial predentine layer appear rapidly. Thereafter, a ruthenium red-positive, phosphotungstic acid-positive, and silver-stain-negative material is deposited into and onto the predentine, giving rise to a layer 1-2 pm thick (Yamamoto, 1986; Yamamoto and Wakita, 1990). Under an electron microscope, this layer shows a reticular structure that includes fine granular and filamentous material. In addition, radiolabeled mannose is deposited on the predentine surface, but not on older portions of the root surface already covered with acellular cementum. As this material first appears within cementoblast-like cells, i.e., cytoplasmic granules, and later in association with collagen fibrils of the cementa1 matrix, both newly formed collagen fibrils and mannose-containing material are believed to be the product of the cementoblast-like cells (Cho and Garant, 1989). Among the known mannose-containing glycoproteins of the connective tissue are fibronectin, structural glycoprotein, and the carboxyl terminal propeptides of procollagen (Cho and Garant, 1989). The origin of the ruthenium red-positive layer is less clear. However, the cementoblast-like cells have been traced by Cho and Garant (1989). They reported that “disruption of the epithelial root sheath appears to be a consequence of directed cell migration by cells of the dental follicle proper which undergo differentiation into precementoblasts.” The latter eventually contact the newly formed outer predentine by means of cytoplasmic processes, become cementoblasts, and temporarily produce part of the acellular cementum matrix. Thereafter, they withdraw from the root surface and assume a fibroblast-like morphotype, referred to as postcementogenic fibroblasts (Cho and Garant, 1989). Later in the process, prior to calcification, further collagen fibril bundles are formed that penetrate only a short distance into the ruthenium red layer and become denser and more bulky; these bundles serve as the major component of the acellular cementum matrix in rat molars. Although there are some similarities of AEFC genesis in rats and humans, it is not known whether a ruthenium red-positive layer and mannosecontaining material also appear in human cementogenesis. Because in the latter all cells, i.e., contacting the predentine and associated with the fiber-fringe meshwork, are apparently engaged in the production of collagen fibrils and bundles thereof, it is likely that all of them are fibroblasts (young and mature), possibly belonging to a particular but still unspecified class. It is possible that the cementoblast-like cells reported on by Cho and Garant (1988, 1989) are also typical fibroblasts, as they were similar ultrastructurally and their profiles were only slightly larger than that of the fibroblasts. Moreover, in humans, none of the connective tissue cells associated with AEFC genesis is comparable in size and structure to

14

HUBERT E. SCHROEDER


15

cementoblasts associated with matrix formation of either AIFC or CIFC produced more apically (See 111,B;Bosshardt and Schroeder, 1990, 1991a, 1992). As soon as the first fiber fringe has attained maximum density, the external mineralization front in human premolar roots approaches the zone of interdigitation of the fringe fibers and the dentinal matrix (Bosshardt and Schroeder, 1991a). This occurs at about 0.3 to 1 mm coronal to the advancing root edge (Figs. 1 and 3a). Subsequently, the base of the fiber fringe begins to mineralize (Bosshardt and Schroeder, 1991b). Thus, the fiber fringe becomes identical to the AEFC matrix, although the fringe continues to exist by fiber elongation, whereas the mineralization front involves more and more of the fringe matrix. In humans, early mineralization of the fiber fringe is associated with the appearance of spherical mineralization centers that initially occur as isolated patches, bound to particular bundles of collagen fibrils ahead of the mineralization front, and later fuse with the advancing front, eventually becoming incorporated in mineralized AEFC (Fig. 4a, Schroeder, 1986; Bosshardt and Schroeder, 1991b). In rat molars, early calcification at the dentino-cementa1 junction was shown to proceed in the presence of spherical bodies resembling matrix vesicles (Yamamoto, 1986). Apparently, early and active AEFC mineralization follows a globular pattern, but it is unclear whether in humans this is guided by matrix vesicles. However, it is known that periodontal ligament cells are rich in alkaline phosphatase activity (Chomette et al., 1987; Oshima et al., 1988; Piche et al., 1989) and this enzyme is believed to act in the process of cementum mineralization (Beertsen and Van Den Bos, 1991). Further development of AEFC involves fiber fringe elongation in proportion to the advancement of the mineralization front, which eventually becomes smooth and stabilized (Fig. 4b). The remaining, elongated and not or not yet mineralized fringe of fibril bundles shows still the same, constant density (Bosshardt and Schroeder, 1991b; Sequeira er al., 1992), implying that at the base of fiber implantation, i.e., along the dentinocementa1junction, no additional fibers are added to this base and that the

FIG. 4 (a) Electron micrograph depicting the actively progressing mineralization of the initial fiber fringe (FF), with the first -5-pm-thick portion of acellular extrinsic fiber cementum (AEFC) attached to dentine (D) and the dentino-cementa1 junction, and with mineralization centers ahead of the mineralization front (MF) and within AEFC. (b) Electron micrograph with the of an established first layer, 10 to I5 pm thick, of AEFC attached to dentine (D), former fringe fibers extending into the periodontol ligament (PL); the mineralization front (MF) is smooth and has reached stability. DCJ, dentino-cementa1 junction. Magnification: a,b, x3020 (a) From Bosshardt and Schroeder (1991b). (b) Courtesy Dr. D. D. Bosshardt.

16

HUBERT E. SCHROEDER Apical Root Portion

0 Apposition

of AlFC

Apposition

of ClFC

of ClFC Matrix Attachment


17

initially formed fibers become elongated and possibly thicken in diameter, as outlined previously. Once the first, 15- to 20-pm-thick layer of AEFC has been formed, its fiber fringe becomes connected to the principal fibers of the periodontal ligament that run from the tooth surface to bone and the gingiva. From that particular point in development, the AEFC matrix fiber fringe appears as the terminal ends of periodontal ligament fibers, which insert into AEFC as fibers of Sharpey. The fiber region closely abutting the AEFC surface may serve as cementoid giving rise, through mineralization, to the next and overnext layers of AFEC. The mechanisms of fiber connection and the alteration of fiber portions into cementoid and its mineralization are still unknown, both in humans and in rodents. However, it is known that AEFC, after its initial establishment, thickens with age. In human premolars, this amounts to an increase by about 20 pm between ages 10 and 17 years (Sequeira et a/., 1992). Later in life, AEFC continues to grow in thickness at an approximate rate of 1.5 to 3.0 pm/year, and this is true for erupted and functioning as well as for impacted, nonfunctioning teeth (Zander and Hurzeler, 1958; Azaz et a / . , 1974, 1977; Dastmalchi et al., 1990). B. Cellular/Acellular Intrinsic Fiber Cementum

In human teeth, CIFC/AIFC are a normal part of CMSC that covers the apical and furcational root surfaces and more coronally merges with pure AEFC. In single-rooted teeth (i.e., incisors, canines, and most premolars), CMSC coats the apical 10 to 40% of the total root length, i.e., about 1.0 (mandibular second incisors) to 5.7 mm (mandibular first premolars) of the apical root (Schroeder, 1988). Frequently, CIFC appears as the first layer of CMSC; thus it forms on newly produced dentine when the advancing root edge has reached the respective apical region. In this situation, the

FIG. 5 Schematic drawing illustrating topographically the initial stages of CIFC genesis on human premolars, developed to about 75% of their final root length: I , committed clone of precementoblasts contacts root predentine and produces first matrix fibrils; 2, cementoblasts form initial collagenous matrix attached to predentine; 3, first formed and mineralized CIFC, including cementocytes, grows by apposition; 4. CIFC is covered with a layer of AIFC. AIFC, acellular intrinsic fiber cementum; CB/u, cementoblasts with unipolar matrix production; CIFC, cellular intrinsic fiber cementum; ERM, epithelial rests of Malassez; MD, mineralized dentine; CBlc, cementoblasts with potential to become cementocytes; CC, cementocytes; PD, predentine; MF, mineralization front; CM, initial cementum matrix; CB/m, cementoblasts with multipolar matrix production; NMD, nonmineralized dentine or predentine; CPCB, committed precementoblasts; ARE, advancing root edge; HES, Hertwig’s epithelial root sheath.

18

HUBERT E. SCHROEDER


19

initial layer of CIFC is a very fast forming tissue that later increases in thickness and forms still faster than AEFC (Table 11). As shown in human premolars with incomplete roots developed to about 75% of their final length (Bosshardt and Schroeder, 1990, 1992), initiation of CIFC formation, early matrix production, and its attachment to root dentine take place along a very short, i.e., about 100- to 200-pm-long, zone extending coronally from the advancing root edge (Fig. 5). The CIFC is a product of cementoblasts; it resembles bone as a product of osteoblasts (Selvig, 1965; Furseth, 1967, 1969; Schroeder, 1986; Bosshardt and Schroeder, 1990, 1992). The formation of early CIFC follows an apico-coronal gradient that is established over a distance very much shorter than that of AEFC (Fig. 5 ) . In the triangular space between the deviating epithelial root sheath and the newly formed predentine, which may extend coronally up to 100 pm from the advancing root edge, a cluster of large, basophilic, closely abutting cells is situated (Figs. 6a and 6b). These cells, considered to be precementoblasts, have an activated, euchromatin-rich nucleus and abundant cytoplasm rich in rough endoplasmic reticulum cisternae and Golgi fields (Fig. 6b). At the precementoblast-predentine interphase, they form numerous slender, fingerlike, partially branching and randomly oriented cytoplasmic processes extending into the first, still very narrow, newly formed CIFC matrix. The latter consists of loosely and randomly arranged collagen fibrils of varying diameter, which interdigitate with the predentine fibrils (Fig. 9b). Along that interphase, this matrix is not well demarcated from the dentinal matrix, and it gradually becomes thicker in the coronal direction, while precementoblasts withdraw. Immediately coronal to the precementoblast cluster, there is a variably wide and short region of rapid matrix production (Figs. 5 , 6a). Mature cementoblasts, separated from one another, become embedded in CIFC matrix that accumulates between and around them (Fig. 6a, inset). This matrix appears denser than that at the predentine surface and consists mainly of randomly oriented collagen fibrils. However, the cementoblasts

FIG. 6 (a) Light micrograph depicting the root surface zone of rapid CIFC genesis and appositional growth, coronal to the advancing root edge (ARE, see Fig. 5, parts 1-3). (b) Electron micrograph of the cluster of precementoblasts (PC) along the most recently formed root predentin (PD; see Fig. 5, part 1). Coronal to this cluster, cementoblasts (CB) with shallow surface recessions and cytoplasmic processes are surrounded by collagen fibrils of newly produced CIFC matrix (a, inset). Further coronally, cementoblasts (CB) reside along the CIFC surface (a). CC, cementocyte; CIFC, cellular intrinsic fiber cementum; D, dentine; HRS, Hertwig’s epithelial root sheath; MF. mineralization front. Magnifications: a, x570; b, ~ 3 0 2 0 ;inset, X6700 (a,b). From Bosshardt and Schroeder (1992).

20

HUBERT E. SCHROEDER


21

enclosed in this matrix have shallow surface recesses bordered by cytoplasmic processes, which begin to compartmentalize the newly formed collagen fibrils (Fig. 6, inset). Thus, in sections cut vertical to the root surface in a corono-apical direction, most of the collagen fibrils, with varying diameters, appear cross-cut in these recessions. As the latter are distributed irregularly in all areas of the cementoblast periphery, these cells are considered to produce CIFC matrix in a multipolar fashion (Bosshardt and Schroeder, 1990). This is supported by the observation of Formicola et al. (1971), that tritiated proline accumulates around such cells in the rat molar. As a consequence, the CIFC matrix rapidly increases in thickness, and some cementoblasts, surrounded by their products, become cementocytes (Figs. 6a, 7b). This area of rapid matrix production continues for a variably short distance along the root surface. It ends approximately level with the mineralized edge of the growing root; i.e., the mineralization front follows an incline up to 45" to the root axis (Figs. 5 and 6a). Cementoblasts that remain at the CIFC surface at this point continue to produce matrix, although more slowly (Bosshardt et al., 1989). This matrix forms a variably narrow layer of cementoid at the mineralized CIFC surface and consists mainly of very densely packed collagen fibrils of rather similar diameter, which to a variable extent are bundled into fibers of varying size. Most of the fibril bundles at the surface of and within mineralized CIFC/AIFC appear cross-cut; i.e., they run parallel to and probably around the root surface rather than perpendicular to it (Figs. 7a and 8b; Schroeder, 1986; Bosshardt and Schroeder, 1990, 1992). Cementoblasts that cover the cementoid of this most recently formed CIFC show typical, deep-seated, partially circular surface recessions and long, slender cytoplasmic processes forming numerous compartments densely filled with cross-cut collagen fibrils (Fig. 7a). These cells also show an activated, euchromatin-rich nucleus with a dense fibrous lamina and a large cytoplasm rich in rough endoplasmic reticulum cisternae and Golgi fields. They also extend finger-like, thin cytoplasmic processes into the bundles of collagen fibrils, which run parallel to and even include such fibrils. Young cementocytes located within already mineralized CIFC appear morphologically identical, and still show slight surface recesses and cytoplasmic processes of varying lengths (Fig. 7b). They are sur-

FIG. 7 (a) Electron micrographs depicting a cementoblast (CB). with deep surface recessions and cytoplasmic processes that compartmentalize the extracellular collagen fibril matrix at the surface of CIFC cementoid (CM: see Fig. 5 . part 3). (b) Electron micrograph of a cernentocyte (CC) within newly formed and mineralized CIFC. surrounded by a matrix halo (MH; see Fig. 5 , part 3). CIFC, cellular intrinsic fiber cementum. Magnifications: a , x 10.000; b, ~ 9 0 0 0 (a) . Courtesy Dr. D. D. Bosshardt. (b) From Bosshardt and Schroeder (1992).

22

HUBERT E. SCHROEDER

rounded by a narrow, halo-like zone of nonmineralized matrix including loosely and randomly arranged collagen fibrils that fill the space between mineralized CIFC and the cementocyte. The mineralization front is indistinct both around cementocytes and along the CIFC surface; focal mineralization centers are absent. The mineralized matrix is very dense and shows a heterogeneous pattern composed of randomly oriented collagen fibrils and numerous bundles of such fibrils cut cross-sectional or in tangential planes. Cementoblasts in the zone of rapidly forming CIFC matrix and that covering the newly formed matrix, i.e., cementoid, at the surface of established and already mineralized CIFC consistently reveal single intracytoplasmic collagen fibrils, either within slender cytoplasmic processes or at the cytoplasm periphery. These fibrils appear to be enclosed in membranebound, electron-lucent compartments. In central portions of the cementoblast-cytoplasm, groups of fibrils in parallel orientation appear in large membrane-bound intracytoplasmic channels (Bosshardt and Schroeder, 1992). It is unclear whether such fibrils are in fact intracytoplasmic or are located in narrow and deep recesses continuous with the extracellular space, as described for tendon and corneal fibroblasts (Birk and Trelstad, 1984, 1986; McBride et al., 1985). Intracytoplasmic, membrane-bound collagen fibrils have been the subject of numerous studies, predominantly concerning gingival and periodontal ligament fibroblasts. Whereas most of these investigations have claimed that such collagen fibrils are undergoing enzymatic degradation (Ten Cate, 1972; Listgarten, 1973; Ten Cate and Deporter, 1974; Eley and Harrison, 1975; Frank et a/., 1976; Garant, 1976; Ten Cate et al., 1976; Rose et al., 1980; Melcher and Chan, 1981; Schellens et al., 1982), a few authors have associated intracytoplasmic collagen fibrils with polymerization and secretion of collagen (Garant and Cho, 1979; Cho and Garant, 1981, 1985). The suggestion that collagen is degraded intracytoplasmatically has been supported by showing either respective enzymatic activity or suggestive ultrastructural details, such as lysosomal bodies. In the case of the periodontal ligament, synthesis and degradation of collagen are part of its specific turnover characteristics (for review, see Schroeder, 1986). Another line of reasoning is that intracytoplasmic collagen fibrils occur “when the rate of procollagen synthesis exceeds the rate of transport to the extracellular compartment,” i.e., in situations of accelerated collagen biosynthesis (Enwemeka, 1991). As no ultrastructural details typical for collagen degradation appear in cementoblasts and because the phenomenon is associated with a process of rapid matrix formation, it has been suggested that intracytoplasmic collagen fibrils of cementoblasts are features of fibril formation and fibril assembly rather than of fibril degradation (Bosshardt and Schroeder, 1992). However, it cannot be ruled out that


23

cementoblasts also screen their products and actively participate in removing unsuitable components from the newly formed matrix. In summary, human CIFC genesis involves cementoblasts that may function either in a multipolar, rapid or in a unipolar, slow mode of matrix production (Bosshardt and Schroeder, 1990,1992). In the former case, the rapid matrix development may be one of the reasons for the incorporation of cementocytes, as was suggested by Paynter and Pudy (1958) and Formicola et d. (1971). A dense cementocyte distribution in CIFC may, therefore, signal rapid CIFC genesis. On the other hand, the same cementoblast may be a slow matrix producer, resulting in either CIFC with low cementocyte density or AIFC (Bosshardt and Schroeder, 1990). However, there seems to be a significant difference between the slow-rate matrix production of CIFC and that leading to AIFC. In the former case, cementoblasts form a discontinuous and not strictly unicellular layer at the CIFC surface and the resulting matrix is less well organized. The AIFC is formed by a continuous, unicellular layer of cementoblasts that show typical surface recessions and matrix compartmentalization at the cell-rnatrix interphase, and this matrix shows the highest degree of fibril-bundle organization (Figs. 8a and 8b; Bosshardt and Schroeder, 1990, 1992). Thus, the different modes of matrix production that a cementoblast may be associated with result in dramatic differences in terms of the speed of genesis and the structural pattern of the product. Much less is known about the genesis of cellular cementum in the rat. Apparently, the cellular variety of rat molar cementum is different from human CIFUAIFC in part because cells of the disintegrating epithelial root sheath of Hertwig become trapped between dentine and the newly forming cementum and are thus incorporated (Paynter and Pudy, 1958; Diab and Stallard, 1965; Lester, 1969; Lester and Boyde, 1970; Schroeder, 1986). In addition, Yamamoto and Wakita (1991) have shown that the genesis of cellular cementum in rat molars is associated with the simultaneous formation of Sharpey’s fibers, i.e., compact bundles of collagen fibrils that run perpendicular to and insert into the newly forming cementum matrix of cross-cut or randomly arranged collagen fibrils. In human cementum, AEFC may be formed later on top of already established CIFC/ AIFC, but in the early stages of CIFC genesis, the incorporation of Sharpey’s fibers is a rare event. Cementoblasts are very similar, morphologically, to osteoblasts and may even function analogously, as CIFC resembles bone. In the latter, “protein secretion is generally polarized toward the bone surface, but at regular intervals along the surface of newly forming bone an osteoblast will secrete matrix away from the surface, eventually surrounding itself to become an osteocyte” (Marks and Popoff, 1988). Another way for an osteoblast to become an osteocyte has been proposed by Nefussi er al.

24

HUBERT E. SCHROEDER


25

(1991). Osteocyte inclusion may “occur without a matrix synthesis inversion by the future osteocyte and with maintenance of close cell contacts with the replacing cell.” The future osteocyte may cease its matrix-producing activity while an adjoining and replacing preosteoblastic cell turns to actively produce new matrix that covers and embeds the osteocyte. However, Frank and Frank (1969) demonstrated that the osteoblast may also behave as a polarized cell with an alternating secretory activity, i.e., directed either toward the osteoid layer or the opposite side. Similarly, cementoblasts seem to have the option of functioning in either a multipolar or an unipolar fashion. It should be added that the mineralization front may also assume a different progression speed, either rapidly following rapid matrix production or slowly advancing behind slow appositional matrix production (Bosshardt et al., 1989). C. Attachment of Matrices t o Growing Dentine

Attachment of the early, first-formed cementum matrix to root dentine is one of the cardinal aspects of root development. In both humans and rats, cementum matrix attachment occurs on a previously formed dentine surface while the respective tooth erupts. In other words, the dentine surface serving as an attachment site moves in the coronal direction while attachment is secured. The speed of this movement may vary between 10 and 40 pm/day and may be faster during AEFC than CIFC initiation (Schroeder, 1991b; Schroeder et al., 1992). Obviously, this attachment concerns the newly formed cementum matrix rather than the cells producing that matrix. In both AEFC and CIFC the newly formed cementum matrix becomes attached to an outer layer of still nonmineralized dentine. In the case of AEFC, an essential part of this attachment is the implantation and interdigitation of collagen fibril bundles perpendicular to and with the randomly oriented collagen fibrils of the dentinal matrix (Fig. 9a; Bosshardt and Schroeder, 1991a). Whether noncollagenous material, such as fibronectin, other mannose-containing matrix components, or unknown glycoproteins, is necessary to facilitate or enhance matrix attachment is unclear at present, at least for human cementogenesis. However, the cells

FIG. 8 (a) Light and (b) electron micrographs depicting a continuous, unicellular layer of cementoblasts (CB), with surface recessions and fibril compartmentalization against the cementoid layer (CM) covering acellular intrinsic fiber cementum (AIFC), including numerous cross-cut bundles of collagen fibrils (b). Magnifications: a. x 1000; b, ~ 3 0 2 0(a,b) . From Bosshardt and Schroeder (1990).

26

HUBERT E. SCHROEDER


27

producing the collagen fibrils and organizing the fiber fringe do not attach to predentine. Rather, by means of cytoplasmic processes, they initially contact the dentinal matrix and this contact may be mandatory in triggering their commitment to function as a particular class of AEFC-forming fibroblasts. In the case of CIFC, precementoblasts appear to attach to the most recently formed part of root dentine at the advancing root edge, but even in this situation contact for cell commitment may be all that is required, as these precementoblasts immediately start to produce the first cemental matrix against that same surface. This early matrix of CIFC is loose and irregular, but its randomly oriented collagen fibrils also intermingle with those of the predentine (Fig. 9b; Bosshardt and Schroeder, 1992). In both cases, AEFC and CIFC, the attachment of first-formed matrix fibrils has to be met and supplemented by two requirements. First, as the attachment site, i.e., the surface of predentine, is moving and newly produced portions of root dentine are continually offered for further attachment, there is a demand for new cementogenic cells to be generated. Whether these new cells move in from the surrounding dental follicle proper, as suggested for rat cementogenesis (Cho and Garant, 1989), or originate by cell division from the cluster of precementoblasts seen in human CIFC genesis is unknown. It is tempting, however, to consider the cluster of human CIFC precementoblasts a pool of dividing cells that do not move but remain located with the deviating epithelial root sheath, furnishing the moving root surface with the population of cementoblasts necessary to continue CIFC genesis. Second, the initial interdigitation and attachment of cementum matrix to predentine must be tightly secured by mineralization. In the case of AEFC, the mineralization front that temporarily remained behind the dentinal root surface starts to advance and gradually approaches and overruns the interphase between the interdigitating cemental matrices and the dentinal matrices, i.e., the future cemento-dentinal junction, as soon as the maximum fiber fringe density and, thus, the maximum collagen fibril interdigitation are established (Bosshardt and Schroeder, 1991a). In other words, the primary interdigitation of the two matrix fibril populations before mineralization and the secondary reenforcement by the mineralizing process allow an intimate fibrous linkage and guarantee a force-resistant cemento-dentinal junction.

FIG. 9 Electron micrographs depicting the interdigitation of cemental with predentinal collagen fibrils at the cernento-dentinal junction of acellular extrinsic fiber cementum (a, AEFC) and cellular intrinsic fiber cementum (b, CIFC). PD, predentine. Magnifications: a, x 14,400; b, x 10,000. (a,b) Courtesy Dr. D. D. Bosshardt.

20

HUBERT E. SCHROEDER

The same is true for CIFC, but in this case, the process of early matrix production and the reenforcement of the rather loosely arranged interdigitation of the cementa1 and dentinal fibril populations by mineralization are much faster. In fact, the mineralization front, being inclined to the root axis, moves in an oblique corono-apical direction rather than parallel to the root surface. Based on such differences in development, structure, and reenforcement of the attachment zone, i.e., the cemento-dentinal junction, it could be assumed that the force resistance between AEFC and dentine is greater than that between dentine and CIFC.

IV. Matrix Formation on Established Root Surfaces in Vitro

Based on a series of experiments (Bellows et al., 1980, 1981, 1982a,b) indicating that “periodontal ligament fibroblasts contract vigorously in vitro, that they attach to exogenous collagen fibers with which they are cultured, and that they orient themselves and the fibers between pairs of mineralized tissue particles,” i.e., dentine, “to which they are attached” (Pitaru and Melcher, 1983),a new interest developed to study the behavior of cells and their products in an in vim-simulated interdental space. In particular, the orientation and attachment of cells and collagen matrices to root surfaces and the formation of a matrix resembling that of natural cementum along such surfaces were the subjects of a series of investigations. A. Culture Material and Experimental Design

The in uirro model used for studying orientation and attachment of cells and endogenously produced collagen fibers to root surfaces in a simulated interdental or periodontal space was first introduced by Pitaru and Melcher (1983) and later modified first by Aukhill and Fernyhough (1986) and then by Bernstein et af. (1988). In principle, it consisted of a cultural system in which cells were grown in the presence of denuded root discs, about 150-300 pm thick, that were cut transversely from middle and apical portions of extracted or otherwise removed, healthy teeth. Following different kinds of pretreatment, the denuded root discs were placed in pairs and in part mechanically secured on the bottoms of plastic culture dishes, with a space of 0.1 to 0.5 mm left between the paired discs. In a modified system (Aukhil and Fernyhough, 1986), 600-pm-thick root discs were placed into slightly wider rings of cortical bone, with a space of 0.1-1 .O mm between tooth and bone. Cells were either added to pretreated


29

discs or precultured prior to disc placement. The complete system was cultured under standard conditions supplemented with ascorbic acid for periods up to 90 days. In the original and the modified model, cells and discs were in contact with the bottom of the culture dish, whereas in the model of Bernstein et al. (19881, discs and cells rested on a filter that was suspended in the culture medium (Fig. 10a). The materials that were used to compose the model systems varied widely, at least originally. Cells used were human gingival fibroblasts obtained from the American Type Culture Collection (Rockville, MD; Pitaru and Melcher, 1983,1987; Pitaru et al., 1984a,b; Melcher et al., 1986) or grown from explants of gingival biopsies (Pitaru et al., 1984b; Aukhil and Fernyhough, 1986; Quarnstrom and Page, 1986; Fernyhough et al., 1987),and the periodontal ligament cells grown from the roots of rat molars (Melcher et al., 1986). Root discs originated from porcine mandibular premolars (Pitaru and Melcher, 1983, 1987; Pitaru er al., 1984a,b) or primary molars (Melcher er al., 1986) removed from animal cadavers, and from extracted human teeth of unspecified type (Aukhil and Fernyhough, 1986; Quarnstrom and Page, 1986; Fernyhough er al., 1987). Bone rings were prepared from cortical bovine bone of unspecified origin (Aukhil and Fernyhough, 1986; Fernyhough er al., 1987). It is clear, retrospectively, that a model composed of material, living or dead, derived from such a variety of species (rat, pig, human) and tissues (gingiva, periodontal ligament, teeth, bone) is extremely heterogeneous, inaccurate, and not necessarily valid, biologically. These original model systems were used to answer a series of questions, such as whether mineralized or demineralized root discs are a favored substrate for cell and matrix attachment, whether cell attachment is a prerequisite for fiber attachment, whether a root surface still covered with natural cementum or one with exposed root dentine enhances cell attachment and fiber orientation, whether a matrix analogous or similar to the connective tissues in the interdental and periodontal spaces develops in uitro, and whether some forms of cementum matrix could be generated in uitro (Pitaru and Melcher, 1983, 1987; Pitaru et al., 1984a,b; Aukhil and Fernyhough, 1986; Melcher et al., 1986; Quarnstrom and Page, 1986; Fernyhough et al., 1987). The original model was modified twofold, first by using autologous human material and standardized methods of microscopic evaluation (Preisig and Schroeder, 1988) and second by employing a new culture system in which nourishment is provided from both above and below the cellular matrix (Fig. IOa; Bernstein et al., 1988). In contrast to the original model used primarily to examine the fibro-cellular “tissue” that develops between two opposing root discs, or the root disc and bone, the new culture system was employed to focus primarily on the collagenous matrix that develops at, on, and along the established root

30

HUBERT E. SCHROEDER


31

surfaces. This system, using human periodontal ligament cells and alveolar bone cells, both grown from explants of the same subject who also provided the extracted teeth for root disc preparation (Bernstein et d., 1988), had the disadvantage that a particular experiment was based on autologous material from an individual patient and, therefore, restricted in scope.

B. Formation and Attachment of Collagenous Matrices Any in uitro model devised to study aspects of periodontal regeneration obviously must attempt to develop structures similar to those of normal tissues in uiuo, in particular to the tissue most essential for that regeneration, i.e., AEFC. Neither the original nor the lately modified model systems (see Section IV,A) have achieved that goal, although using the system of Bernstein et al. (1988, 1989a,b, 1990) two findings were helpful. First, a collagenous fiber fringe attached perpendicular to and intermingled with the matrix fibrils of established but demineralized root cementum developed occasionally, a fringe similar to that in early AEFC genesis. Second, this autologous and rather standardized system furnished some indications why such model systems are necessarily inefficient, at least for the time being. In the original model of Pitaru and Melcher (1983), employed for 20-30 days and later for short-term culturing, multiple layers of gingival fibroblasts and collagen fibers developed over 10 days in about 60% of the culture sets. These layers formed sheets that radiated perpendicularly from the root disc surfaces and bridged and eventually filled the interdisc space. The fiber component, being phase-contrast refractile and histochemically positive for collagen, assumed parallel orientation and extended across the interdisc space as well as from the root disc periphery not opposed by another disc. Because the refractile material was also seen after 1 to 6 days of culturing, with the discs placed on top of a confluent layer of gingival fibroblasts precultured for 4 to 6 days (Pitaru et al., 1984a,b), the interpretation of this refractile material bridging the interdisc space as though resembling the arrangement and distribution of transeptal and dentogingival fibers in uiuo was slightly premature (Pitaru and

FIG. 10 (a) Schematic drawing of experimental culture model. (b,c) Light and (d) electron micrographs depicting the interdisc space (b) with a matrix forming a dense fringe (FF) of collagen fibrils (c,d) perpendicular to and intermingled with cementa1 fibrils (d) along cementum-lined (C) root surfaces in a 42-day-old culture of human periodontal ligament cells cocultured with autologous root discs (RD). Magnifications: b, X 10; c , X 150; d, x7000. (a-d) From Bernstein et al. (1988).

32

HUBERT E. SCHROEDER

Melcher, 1983). In addition, the idea that precultured fibroblasts migrate toward and attach to the root discs within 24 hr (as seen by time-lapse photography of the disc surfaces in whole-mount culture preparations) because they are attracted by them (Boyko et al., 1980; Pitaru et al., 1984a) gave rise to a series of attachment assays introduced some years later (see Section IV,C). Furthermore, it was suggested that partially demineralized substrates favor cell matrix orientation and attachment, whereas demineralized cementum is more efficient in maintaining that attachment (Pitaru and Melcher, 1983; Pitaru et a / . , 1984a,b). The first light- and electron-microscopic studies on the development and composition of the interdisc matrix synthesized in uitro were reported by Quarnstrom and Page (1986), Fernyhough et al. (1987), and Pitaru and Melcher (1987). Using totally or very briefly decalcified root discs of unspecified human teeth, planed to create a smooth dentine surface, and a preculture of confluent human gingival fibroblasts of the 5th to the 10th passage, Quarnstrom and Page (1986) described the development and composition of the interdisc matrix histologically and biochemically in cultures 5 days to 13 weeks old. Initially, cells attached to and migrated up the vertical root surfaces. Eventually, they filled the entire interdisc space with extracellular matrix. Attachment and mobility of cells to and along the demineralized root surface was explained by the presence of fibronectin and the exposure of collagen fibrils. The first interdisc matrix was shown to contain hyaluronic acid, chondroitin sulfate, dermatan sulfate, and large amounts of fibronectin. Collagen fibrils appeared only after 6 weeks in culture and fibers became denser with time. In late cultures the matrix contained collagen of Types I, 111, and V (Quarnstrom and Page, 1986). Although some fibroblasts of the interdisc matrix assumed a shape indicating a tendency to compartmentalize extracellular collagen fibrils, no fiber fringe structures were observed along the vertical disc-dentine surfaces. Similar results, albeit restricted to microscopic evaluation, were obtained in 20- to 30-day cultures of bovine disc pairs and human gingival fibroblasts (Pitaru and Melcher, 1987). These authors, however, reported that along decalcified cementum surfaces, cells in high density were oriented mainly perpendicular to that surface, whereas along nondecalcified surfaces, less numerous cells lay with their long axis parallel to the disc surface. And the same was true for matrix collagen fibrils. Initial cell attachment at the demineralized cementum surface was established either by cell processes penetrating between cementum collagen fibrils or by the plasma membrane being closely apposed to such fibrils, possibly suggesting a fibronectin-independent or fibronectin-mediated process. Pitaru and Melcher (1987) felt that “the continuous layer of gingival fibroblasts oriented parallel to the non-decalcified cementum and the similarly oriented extracellular fibrillar matrix is reminiscent of a capsular structure,”


33

whereas the cell matrix orientation at demineralized surfaces would resemble the in uiuo arrangement. However, the findings of Pitaru and Melcher (1987) as illustrated in their paper were very much unlike the in uiuo situation and the same was true for the observations of Fernyhough et al. (1987), who used the root disdbone model. The modified culture system of Bernstein et al. (1988) yielded more consistent results, in particular with respect to the formation of a dense fiber matrix along established cementum surfaces. In a series of experiments (Preisig and Schroeder, 1988; Bernstein et al., 1988, 1989a,b, 1990) using autologous human periodontal ligament cells and demineralized root discs precoated with fibronectin in 6- to 13-week-old cultures that rested on a Puropor-200 filter supported by a wire-mesh grid (Figs. 1Oa and lob), it was confirmed that dentine-lined root surfaces were often encapsulated by cells and collagen fibrils running parallel to that surface. Along cementum-lined root surfaces, a variably dense fringe of collagen fibrils that interdigitated with those of the cementum matrix developed within 3 to 6 weeks (Figs. 1Oc and 1Od).This fringe with its fibrils oriented perpendicular to the cementum surface extended for a short distance into the interdisc space or the disc surroundings. However, the fringe was not continuous and covered patchwise variably large areas of the root surface. Fibroblasts were positioned irregularly. This fringe being rather short and occasional was not to be equated with the refractile material of cell fiber sheets, radiating from disc surfaces (Pitaru and Melcher, 1983; Pitaru et al., 1984a,b). In older cultures, the interdisc matrix and its cells appeared to deteriorate, whereas new cell matrix layers had formed on top of that space. In other words, the fiber fringe fully developed after 6 weeks did not increase in density with time. When autologous cells derived from human alveolar bone were added to that model system, a more continuous and very much denser fiber fringe developed along most of the cementumlined disc surfaces and this fringe remained relatively unchanged over 13 weeks in culture (Figs. 1Oc and 1Od; 1 la and I lc). Because of a high degree of fibril interdigitation, the interface between the newly formed matrix and the established but demineralized cementum attained structural continuity (Fig. 1 Ic). Thus, the combination of autologous human periodontal ligament cells and alveolar bone cells produced an improved fiber fringe intimately attached to established cementum and indeed somewhat reminiscent of early AEFC matrix and its attachment to dentine (see Section 111,A). The majority of collagen fibrils produced by that combination had a diameter within the 50- to 80-nm range, whereas the fibrils produced by periodontal ligament cells alone were mainly in the 20- to SO-nm range (Table 111). The former distribution was indeed similar to that in cementum (Table 111). In addition, as a less frequent finding, single periodontal ligament cells or single cells in ligament/bone cell mixtures were found to be

34

HUBERT E. SCHROEDER


35

surrounded by their collagenous fibril product in a halo fashion (Fig. 1 Ib). This was possibly an indication for some cells assuming a cementoblast phenotype, producing an early form of CIFC matrix (see Section 111,B). Intentionally, the autologous model system of Bernstein et al. (1988, 1989a) was not supplemented with Na-P-glycerophosphate, which had been shown to allow the production and calcification of bone-like matrix nodules (Williams et al., 1980; Ecarot-Charrier et al., 1983, 1988; Sudo et al., 1983; Bhargava et a / . , 1988) and even the assembly, by osteoblasts, of a bone-specific macrostructure in uitro (Gerstenfeld et al., 1988). Bernstein er al. (1989a,b) merely examined the contribution bone cells would make to the development of a “cementoid” fiber fringe matrix attached to the established cementum surface. When Melcher et al. (1986, 1987) cultured fetal rat calvaria cells in the presence of porcine root discs in medium supplemented with ascorbic acid and Na-P-glycerophosphate, they claimed to have found varieties of “cementum” newly formed on the root disc surfaces. However, in no instance did they report a true fibrous attachment with a dense collagen fibril arrangement, and the nodules forming on cementum-lined discs and elsewhere in their cultures were similar to bone-like centers produced in systems without root discs (Bhargava et al., 1988; Maniatopoulos el a / . , 1988). The finding of an improved fiber fringe matrix attached to scaled, cementum-lined but demineralized root surfaces by the addition of alveolar bone cells to autologous periodontal ligament cells (Bernstein et al., 1989a) supported the hypothesis that, in uiuo, cells originating in alveolar bone may migrate to the periodontal ligament and function as cementum-producing cells (McCulloch et al., 1987; Melcher et al., 1987). However, in a subsequent experiment using the same model system with demineralized and nondemineralized discs precoated with normal culture medium, fibronectin, or autologous serum, and with the culture medium in part being supplemented with 10% autologous serum for the first 17 days, Bernstein et a/. (1989b) were unable to further improve the fringe matrix. The collagen fibrils were again in a diameter range of 40-60 nm or beyond (Table 111). The fiber fringe did not become denser or more consistent, and the newly formed matrix again displayed a patchwise distribution. A dense collagenous fiber matrix was also produced along nondemineralized, ce-

FIG. l l (a) Light and (b,c) electron micrographs depicting (a) the interdisc matrix (IDM) with cells surrounded by discrete halos of collagen fibrils (b, arrowheads) or a dense fiber fringe (FF) continuous and intermingled with cementa1 fibrils (c, inset) along cementurn-lined ( C ) root surfaces in a 56-day-old culture of human periodontal ligament cells and alveolar bone cells cocultured with autologous root discs (RD). Magnifications: a, x580; b,c, x2700; inset, ~ 5 8 0 (a,c) . From Bernstein er a / . (1989a).

TABLE Ill Diameter (nm) of Collagen Fibrils of the Fiber Fringe Matrix Grown in Cultures, Compared to Cementum Fibrilsa % Diameter distribution (nm)

Root discs (decalcified) Days in culture 56 56 124 124 Cementum Cementum

F-coated F-coated F-coated F-coated -

-

Cells

N

Diameter (x ? s)

HPLC HPLC + HBC HPLC HPLC + HBC -

80 80 160 160 320 320

42+ 65 ? 472 63 ? 642 71 +

Diameter (x + s)

Range

20-40

40-60

552 9 51 + 10 58 & 10

32- 83 26- 86 26- 86

7 15 6

69 67 59

9 14 9 I1 11

13

Range

20-50

50-80

80- 100

25- 64 38-101 32- 79 39- I05 32- 97 37-109

79 10 61 8 9 4

21 78 39 86 83 72

0 12 0 6 8 24

% Diameter distribution (nm)

Days in culture 35-42 35-42 35-42

Root discs (decalcified)

Cells

N

M-coated F-coated AS-coated

HPLC + HBC HPLC + HBC HPLC + HBC

180 180 180 ~~

" Data from Bernstein

~~

60-80 24 18 35

~

er al., 1988, 1989a,b; N = number of measurements; F, fibronectin; M. medium; AS, autologous serum; HPLC, human

periodontal ligament cells; HBC, human alveolar bone cells.


37

mentum-lined, sterile root discs derived from previously diseased and experimentally planed teeth with advanced periodontitis (Bernstein et al., 1990). However, the smooth, hard, and chemically as well as mechanically denatured root surface did not appear to be a suitable substrate for attaching a newly produced matrix. Such surfaces do not provide exposed collagen fibrils with which a new fibril population could intermingle, and this is exactly the situation for clinically treated root surfaces that are needed to serve as attachment sites for regenerating cementum.

C. Factors and Failures Retrospectively, attempts to study regenerative phenomena such as interdental connective tissue matrix development and cementogenesis along established root surfaces in uitro have failed so far. All in uirro models used have focused on an experimental situation in which two factors, i.e., cells and established but denatured and modified root surfaces, were elected to interact. Aspects of normal wound healing processes, including blood clot formation and clearance of traumatized wound edges by phagocytosis and resorption and the several humoral factors that participate in would healing, were intentionally excluded. The most simplified model situation indeed did not permit the observation of anything but the behavior of cells and the arrangement of matrix products in the presence and along denatured root surfaces. Also, both of these model components were heterogeneous, nonspecified biologically, and most often of unknown origin and ontogeny. This is particularly true of the cells used in such model systems, i.e., gingival and periodontal ligament fibroblasts and (alveolar) bone cells freed by enzyme treatment (Melcher et al., 1986). All of these cell populations were heterogeneous mixtures of functionally unknown properties, both when harvested from biopsies and after propagation in culture. Gingival fibroblasts grown from human gingival biopsies represent a mixture of ill-defined groups of cells that include subpopulations of different functional capacity and subpopulations that may be phenotypically stable both in uiuo and in uitro or regulated by extracellular factors (Narayanan and Page, 1983; McCulloch and Bordin, 1991). For example, human gingival fibroblasts grown from normal gingiva include subpopulations with a heterogeneously expressed Clq-receptor (Bordin et al., 1983,1984). Cells with a surface marker for the collagen-like domain of Clq are activated by that molecule and increase their DNA synthesis, rate of proliferation, and total protein production, the latter about threefold, with 40% of this production directed to collagen synthesis, in particular of Types I11 and V. These properties remain stable in culture, at least over 12 doublings

38

HUBERT E. SCHROEDER

(Korotzer et al., 1980; Bordin et a / . , 1983, 1984; Bordin and Page, 1988, 1989). Mass cultures of human gingival fibroblasts derived from one individual biopsy were shown to differ widely in proliferation rate and production of protein, in particular the synthesis of collagens and glycosaminoglycans (Hassel and Stanek, 1983). The types of collagens produced by such mixed populations are I, 111, IV, and V, the latter two as minor species (Hurum et al., 1982; Narayanan and Page et al., 1983), and most cells can produce Types I and I11 collagens simultaneously, whereas some produce only Type I or none (Engel et al., 1980).The two main glycosaminoglycans synthesized by human gingival fibroblasts are hyaluronic acid and heparan sulfate, but chondroitin sulfate and dermatan sulfate (Bartold and Page, 1985) are also synthesized. Clones of human gingival fibroblasts, varying in size, morphology, and growth rates, are rather similar in their ability to synthesize collagen, collagenase, and collagenase inhibitors and to degrade native collagen mats, indicating that such clones can both synthesize and degrade collagen (Hurum et al., 1982). However, any particular fibroblast can either secrete or degrade collagen but cannot do both simultaneously (Yajima, 1988). Furthermore, human gingival fibroblasts can be functionally modified by extracellular factors: prostaglandin E2 inhibits their growth and collagen synthesis (KOet al., 1977); concanavalin A stimulates their synthesis of collagenase about 10-fold (Hurum et al., 1982); guanidine/EDTA extracts of proteins from bovine cementum, dentine, and alveolar bone stimulate their total protein production and collagen synthesis (Somerman et al., 1987a,b). However, it is not known whether these modifications in cellular activity occur or can be induced in all or only in particular subsets of human gingival fibroblasts that, like human dermal fibroblasts, can be made to increase their collagen production, independent of cell growth, by a factor of 2 to 3 in the presence of ascorbic acid (Russell et al., 1981). It is also unclear whether there are stem cells for fibroblast propagation. McCulloch and Knowles (1991) demonstrated that in hamster gingiva and cultures thereof, 40% of the fibroblasts divide in uiuo and in v i m , a further 40% of these cells divide only in uitro, and about 10% do not divide in either situation. That may suggest that in the hamster gingiva, there is a population of actively cycling progenitor fibroblasts, another population that is growth-inhibited in uiuo but capable of growing upon explantation, and a third population of terminally differentiated cells. Truly colony-forming progenitor fibroblasts, i.e., stem cells, may represent a very small population in uivo (i.e., about 0.5% of all isolated cells: McCulloch and Knowles, 1991). For all these reasons, subcultures of human gingival fibroblasts (and that of animal models) in fact are heterogenous mixtures of cell subsets of unknown properties and of unknown proportions that may vary from culture to culture and donor-to-donor tissue. Nevertheless, in continuation


39

of the in uitro model work discussed here (Pitaru et al., 1984a,b), such illdefined mixtures of human gingival fibroblasts have been used to study their migration and attachment to untreated and modified porcine root slices and various alloys in short-term cultures (Lowenberg et al., 1985, 1986, 1987) as well as to diseased and nondiseased human roots treated by citric acid- or EDTA-demineralization and collagenase (Fardal and Lowenberg, 1990). Similar attachment studies were performed with mixtures of human gingival fibroblasts added to diseased and nondiseased roots (or slices thereof) freed from cementum and coated with human plasma (Abbas et al., 1987) or fibronectin (Fernyhough and Page, 1983), and to Petri dishes or wells coated with guanidine/EDTA protein extracts of unspecified bovine and human cementum (Somerman et al., 1991). Also, human gingival fibroblasts were cultured for 2 hr on fibronectincoated and/or tetracycline-treated blocks of bovine dentine (Terranova el al., 1986). All authors reporting such experiments usually claimed that cell-attachment assays used to screen conditions of favored or enhanced attachment to established but modified root tissue surfaces such as blank cementum or dentine would contribute directly to the clarification of a principal phenomenon required for regenerative cementogenesis. Similar experiments and arguments were applied to study the behavior of so-called periodontal ligament cells. When cultured, such cells are derived from explants of the residues of natural periodontal ligament tissue that remains at the healthy root surface after tooth extraction. Developmentally, periodontal ligament cells are of ectomesenchymal origins and residents of the dental follicle proper that invests tooth germs. In the mouse, such ectomesenchymal cells of the dental follicle have been shown to give rise to and differentiate into cementogenic, osteogenic, and fibroblastic lineages forming the functionally united root cementum, alveolar bone proper, and periodontal ligament (Ten Cate et al., 1971; Barrett and Reade, 1981; Yoshikawa and Kollar, 1981; Palmer and Lumsden, 1987). However, this totipotency of periodontal ligament cells is restricted to the time period of root development and has so far not been detected in any of the cells residing in the periodontal ligament of fully developed and functioning teeth (McCulloch and Bordin, 1991). For culturing, periodontal ligament cells have been harvested from a variety of different species such as pigs, cattle, dogs, non-human primates, and humans. Following tooth extraction, at least some of such cells can be kept viable at their original site at the root surface for up to 1 year (Litwin et al., 1971) and harvested by trypsinization or scraping (Blomlof and Otteskog, 1981; Ragnarsson et al., 1985; Oikarinen and Seppa, 1987). Morphologically, cultured human periodontal ligament cells are fibroblastlike and spindle-shaped or stellate, with numerous, long cytoplasmic extensions protruding in the direction of their long axis (Ragnarsson et al.,

40

HUBERT E. SCHROEDER

1985; Preisig and Schroeder, 1988).Ultrastructurally, these cells reveal an euchromatin-rich nucleus, and a cytoplasm containing the usual set of organelles and single and compact bands of microfilaments, i.e., vimentin and possibly actin, numerous endocytotic vesicles, and variable amounts of glycogen storage granules (Chomette et al., 1987;Preisig and Schroeder, 1988; Rose et al., 1987). Surprisingly, probably because of the presence of a preosteoblast subset, periodontal ligament cells reveal a high alkaline phosphatase activity similar to that of osteoblasts (Kawase et al., 1986; Piche et al., 1989) and this is also found in periodontal ligament cells of bovine incisors, which also produce increasing amounts of 3’, 5’-cyclic adenosine monophosphate in response to human PTH, and a marker for osteoblasts (i.e., a protein cross-reactive with bovine bone Gla protein; Nojima et al., 1990). Clones of porcine or monkey periodontal ligament cells are capable of synthesizing collagen Types I, 111, and V in varying proportions, i.e., 10 to 30% of the total collagen for Type 111and extremely high or low amounts of Type I, from clone to clone (Limeback et al., 1978, 1983). This is a clear indication that such periodontal ligament cells are a heterogenous mixture of cells producing distinctly different ratios of collagen types. Some clones of human periodontal ligament cells produce a heat-resistant factor that inhibits the PTH-induced resorption process in fetal long bone cultures (Giniger et al., 1991). Further evidence for the heterogeneity of periodontal ligament cells is derived from in uiuo studies. In both wounded and normally functioning periodontal ligament of the mouse, paravascular cells with small nuclei, rather undifferentiated ultrastructurally, appear to have progenitor or stem cell characteristics (Gould et al., 1977, 1980; McCulloch and Melcher, 1983b; Gould, 1983). These paravascular cells constitute 15 to 20% of all ligament cells (McCulloch and Melcher, 1983b),have the capacity for a relatively high rate of proliferation, migrate to the bone and cementum surfaces (McCulloch and Melcher, 1983a), cycle more slowly than cells outside the paravascular region (McCulloch, 1985), and are found in the endosteal spaces of the alveolar bone, from which they migrate via Volkmann’s channels into the periodontal ligament (McCulloch et al., 1987). These cells, being considered a stem cell population of clonal distribution (McCulloch, 1985) may be able to augment fibroblasts, cementoblasts, or osteoblasts because of their totipotency (McCulloch and Bordin, 1991).This is supplemented by the finding that periodontal ligament fibroblasts and preosteoblasts express numerous specific binding sites for epidermal growth factor in uiuo, whereas fibroblasts of other connective tissues have only a few binding sites (Cho et al., 1988). These ligament cells may be of common origin and related to paravascular stem cells. Whether paravascular cells of similar characteristics are present in the human periodontium is, however, unknown.


41

Comparative studies with human gingival fibroblasts and human periodontal ligament cells have revealed that both mixed populations are not alike. In vivo studies in the mouse demonstrated that the collagen turnover rate is five times higher in the periodontal ligament than in gingiva (Sodek, 1976, 1977), and in the rat, periodontal ligament cells including preosteoblasts consist of subpopulations that respond with proliferation patterns different from those to traumatic wounding (Gould et al., 1977, 1980), endocrine (Roberts, 1975a,b)and electrical (Davidovitch et al., 1980)stimulation, and orthodontic forces (Roberts et al., 1982; for review see Schroeder, 1986). In vitro, mixed human periodontal ligament cells are significantly more active in their protein and collagen synthesis and have a significantly greater alkaline phosphatase activity than gingival fibroblasts (Somerman er al., 1988; Arceo et al., 1991). Furthermore, mixed human periodontal ligament cells, but not gingival fibroblasts, are “capable of producing mineral-like nodules in vitro” (Arceo et al., 1991). Porcine periodontal ligament cells, either in clones or in mass cultures, are relatively homogeneous in the expression of collagen Type I and fibronectin; i.e., 99% of all cells express both. In contrast, porcine gingival fibroblasts are considerably heterogeneous; only 57 to 78% of the cells express both collagen and fibronectin, 21 to 42% express only collagen, and 0.4 to 0.8% express only fibronectin (Connor et a/., 1983). In attachment assays, human periodontal ligament cells respond preferentially to fibronectin-coated blocks of dentine or wells, whereas gingival fibroblasts are promoted to attach to surfaces coated with bone phosphoprotein (i.e., osteopontin) and with guanidine/EDTA extracts of bone and cementum (Somerman er al., 1987c,d, 1989;Terranova et al., 1987,1989b). However, as cultures of both periodontal ligament cells and gingival fibroblasts are heterogeneous mixtures of different cell subsets, it is unclear whether such differences are typical for all or any particular set of cells. Finally, alveolar bone cells that can be freed from the marrow spaces and bone surfaces are again mixtures of heterogeneous cell populations that differ due to their source and origin. None of the isolation procedures used provides a pure homogeneous population of structurally and functionally identified cells (Nijweide et al., 1986; Marks and Popoff, 1988). Cementoblasts or precementoblastic precursor cells have not been isolated or detected yet at all. In fact, it is entirely unknown where they come from, whether they proliferate, how long and under what conditions they function, and where they go to or reside in reserve. And this is true for the particular class of fibroblasts forming AEFC as well as for the osteoblastlike cementoblasts producing CIFC/AIFC (see Sections III,A and 111,B). The second component of the in vitro model for studying regenerative matrix production, i.e., denuded, planed, and artificially modified root

42

HUBERT E. SCHROEDER

surfaces of human and mostly porcine teeth, is likewise of heterogeneous, often unidentified nature. Not only were cementum- and/or dentine-lined surfaces, denuded or denatured, still mineralized or demineralized, employed side by side, but there was also no differentiation between AEFC and CIFC or CMSC (see Section 11). As these varieties of root cementum differ from one another in terms of their genesis, structure, and composition (Schroeder, 1986), it is possible that their ability to serve as a favored substrate for regenerative cementogenesis is distinctly different as well. For example, if interdigitation of collagen fibril populations of dentinal and cementa1 matrices is required, decalcified AEFC, with its fibrils running perpendicular to the root surface, is a much better candidate than CMSC or CIFC/AIFC, in which most fibrils run parallel to the root surface. The latter is also true for dentine. Apart from such structural differences, the various cementum varieties may, in addition, possess different biochemical characteristics. For example, studies have suggested that extracts of healthy human (and bovine) cementum modify the migration, attachment, and growth of gingival or periodontal ligament fibroblasts (Miki et al., 1987; Somerman et al., 1987a-d, 1989; Nakae et al., 1989; Nishimura et al., 1989; McAllister et al., 1990) and that human and bovine cementum contains two sialoproteins, i.e., bone sialoprotein I1 and osteopontin (Somerman et al., 1990; Olson et al., 1991). However, in none of these studies was root cementum defined as one or the other variety. In conclusion, in vitro models simulating interdental or periodontal relationships have been used under conditions lacking any possibility to define the model component parameters. Under such conditions, it is not possible to study the sequential evolvement of the cells and matrices or the origin and character of a particular matrix or its attachment to established root surfaces. This has already been pointed out by Bernstein er al. (1989b). Furthermore, it is questionable whether the fact that mixed gingival fibroblasts attach to denuded but variably coated dentine has anything in common with biological problems of periodontal regeneration, specifically regenerative cementogenesis. As shown in Section I1 of this review, it is not clear whether cells, acting in wound healing, in the presence of a dense blood clot and during the subsequent development of a granulation tissue, indeed do need to attach to a substrate surface prior to forming and attaching their matrix product. The requirements for future in uitro investigations focusing on periodontal regeneration appear to be of a different order: 1. to characterize the various phenotypically different or modulationdependent cells in the human periodontal ligament; 2. to identify, biologically and functionally, the fibroblast-like class of cells forming AEFC and the cementoblasts forming CIFC/AIFC;


43

3. to devise methods for harvesting and culturing these cells and to study their behavior under simplified conditions of an in uitro model; and 4. to focus on matrix attachment rather than cell attachment.

V. Regenerative Cementogenesis on Established Root Surfaces in Vivo Clinical studies focusing on the healing and regenerative response of the periodontal tissues to various treatment modalities, although attractive for many decades, have been increasing in number since about 1975 and respective reports are abundant. Reviews attempting to summarize the essential results and failures of such studies have been provided by Selvig (1983), Polson (1986, 1987), Egelberg (1987), Nyman et al. (1989), and Minabe (1991). In the context of this review, selected and restricted groups of original contributions will be used to discuss the major biological problems involved in spontaneous and guided regenerative cementogenesis. A. Clinical and Experimental Conditions

In addition to numerous clinical studies in humans, various animal models have been developed, employing mostly non-human primates and dogs, to explore the periodontal regeneration potential. Such models included (a) autotransplantation of single-rooted teeth (Butcher and Vidair, 1955; Andreasen and Kristerson, 1981; Proye and Polson, 1982), ( 6 )submucosal implantation of tooth roots at the bone-connective tissue interface (Nyman et al., 1980; Karring et al., 1980), ( c ) the window technique creating artificial submucosal defects in the mid-root periodontium ( Jansen et al., 1955; Nyman et uf., 1982a), ( d )experimental creation of chronic periodontitis lesions around single- or multirooted teeth by means of attaching rubber bands or silk ligatures subgingivally (Caton and Zander, 1975; Caton and Kowalski, 1976; Nyman et af., 1980), and ( e ) the experimental production of furcation defects in multirooted teeth (Ellegaard el al., 1973; Johansson et al., 1978; Crigger et al., 1978). In the diseased human periodontium as well as in all such model systems, the principal obstacle during wound healing and tissue regeneration is the fact that one side of the wound defect is the blank root surface lined by varieties of cementum or exposing dentine, i.e., a hard, smooth tissue wall denuded from organic material and compressed by planing instruments. Such surfaces, which

44

HUBERT E. SCHROEDER

primarily had been exposed to the inflammatory environment of a periodontal pocket and covered by bacteria, have undergone pathological changes of both their organic and their inorganic components, such as hypermineralization, accumulation of rare inorganic elements (e.g., fluoride), and contamination with endotoxins (Selvig, 1983). Periodontal wound healing along such surfaces can possibly and eventually provide only four different types of restoration: ( a ) the exposed and treated root surface is lined by epithelium, i.e., in the form of a long junctional epithelium that grows out from the gingival wound edge and covers and protects the denatured root surface; ( b ) the exposed hard tissue being recognized as an inert foreign body is encapsulated by connective tissue with collagen fibers running parallel to its surface; (c) the exposed root surface undergoes initial resorption followed by either bone contact and ankylosis or the formation of new cementum and a periodontal ligament; and ( d ) the exposed and treated root surface is covered by new cementum of one or the other variety, without undergoing resorption (Selvig, 1983; Nyman et al., 1989;Schroeder, 1991a).These options depend on the proliferational speed with which the tissues bordering the defect can react, i.e., the epithelium and connective tissue of the gingiva, the alveolar bone, and the periodontal ligament. That speed is high in the gingival epithelium and low in the periodontal ligament (see Schroeder, 1986, 1991a).

B. Spontaneous Regeneration Using autotrans- or replantation of healthy teeth as a model, the quality and extent of periodontal regeneration depend on the root surface conditions. In freshly extracted teeth, with vital remnants of the periodontal ligament, i.e., cells and torn collagen fibers, remaining attached to the cementum of their root surface, a scar-like healing results in a reconnection of these fibers with that of the surrounding connective tissue, with normal structure and function of the bone-ligament-cementum complex being restored eventually (Butcher and Vidair, 1955; Polson and Caton, 1982; Houston et al., 1985; Blomlof et al., 1988). The healing result is the same when a healthy root is replanted into an alveolar bone socket of artificially reduced height (Polson and Caton, 1982; Houston et al., 1985). Healing and fiber restoration may be completed in about 3 weeks (Hurst, 1972) and circumscribed, artificial notches placed prior to replantation in the midroot may be lined with new AEFC and fiber attachment (Houston et al., 1985). A periodontitis-affected or a healthy root partially denuded from organic material including some or all old cementum, however, is rapidly covered by a longjunction epithelium (i.e., within 2 weeks; Karring et al., 1984),


45

even when that root is autotransplanted into a socket of normal height (Stones, 1934; Polson and Proye, 1982, 1983; Lindhe et a/., 1984; Houston et al., 1985, Bowers et al., 1989~;Blomlof et a / . , 1988). If epithelium is excluded from wound healing, i.e., by crown amputation and mucosal closure, a great portion of the denuded root surface will be lined by connective tissue with its collagen fibers running parallel to that surface, or the latter may undergo resorption with or without ankylosis (Karring et af., 1980, 1984; Nyman et al., 1980; Houston et a/., 1985). In such experiments, an observation made in clinically treated periodontitis patients has been repeated (Skillen and Lundquist, 1937; Listgarten and Rosenberg, 1979; Cole et al., 1980; Bowers et a/., 1985, 1989a,b,c), i.e., that at the apical termination of the denuded root surface, usually marked by an artificial notch, and close to the most apically located healthy or restored part of the periodontal ligament and its fiber attachment to old cementum, a variably short strand of new cementum develops, with or without fiber attachment. This occurs more frequently and to a greater extent if epithelium is excluded (Bowers et a / . , 1989a,b,c). As seen in classical histological sections, this new cementum is variably thin, slightly mineralized (Ogura et a/., 1991), of a cellular variety, and present over nonresorbed old cementum and dentine, but does not, due to tissue shrinkage, stick to the old root tissues (Listgarten and Rosenberg, 1979; Bowers et a/., 1989a,b,c). These experiments suggested that: ( a ) gingival and periodontal connective tissue reattaches to the root surface as long as the latter carries vital remnants of periodontal ligament, ( 6 ) bone regrowth and periodontal ligament regeneration are unrelated phenomena, and (c) cementum regeneration and new fiber attachment are possible provided the denuded root surface is first repopulated by cells derived from periodontal ligament and, possibly, the alveolar bone. In fact, using precultured fibroblasts from that ligament and other tissues such as gingiva, dermis, periosteum, and fascia to envelop replanted roots denuded by mechanical scaling, Andreasen and Kristerson (1981) and Boyko et al. (1981) demonstrated that periodontal ligament cells are unique in regenerating cementum and fiber attachment and cannot be replaced by other cells, except perhaps those of a dental follicle. Such cells, therefore, appear as an indispensible prerequisite for regenerative cementogenesis in uiuo. This conclusion was corroborated by experiments using the window technique, i.e., a surgically created submucosal fenestration in the alveolar wall, combined with a local removal of periodontal ligament and the cementum, covered by a Millipore filter (Nyman et al., 1982a; Pettersson and Aukhil 1986; Aukhil et a/., 1986b; Iglhaut et a/.,1988; Knox and Aukhil, 1988; Selvig er a / . , 1988). In this situation, at least part of the exposed root dentine will be covered within about 3 weeks to 6 months,

46

HUBERT E. SCHROEDER

with thin, patchy new cementum, possibly AEFC (Pettersson and Aukhil, 1986), albeit often clearly not AEFC (Nyman et al., 1982a),and the defect will be filled by bone ingrowth and a new periodontal ligament with partial fiber attachment, but root resorption and ankylosis occur frequently. In non-human primates (M. fuscicularis), cells from both the surrounding, intact periodontal ligament and the cut bone surface, labeled by [3H]thymidine, arrive in the window defect 3 days after wounding, and at 21 days, a thin “cementoid” matrix is deposited along the exposed dentine (Iglhaut et al., 1988). In rats, a 2-week window is filled with a fibroblast-rich connective tissue and after 4 weeks, an electron-dense material lines the denuded dentine, from which dense aggregates of inserted collagen fibrils radiate perpendicularly (Knox and Aukhil, 1988). This initial stage of regenerative cementogenesis somewhat resembles the beginning of normal genesis of acellular cementum in the rat (see Section 111,A).Similar results, although with direct attachment of collagen fibrils to decalcified dentine, were obtained in dogs (Selvig et al., 1988). In these cases, regenerative matrix attachment seemed to follow the principle of interdigitation of collagen fibril populations as seen in human cementogenesis (see Section 111,C).

C. Guided Regeneration

The principle of guided periodontal tissue regeneration was born as an extension of the window technique and first introduced by Nyman et al. (1982b). It is based on the argument that cells proliferating from the periodontal ligament bordering the periodontal defect should preferentially and selectively repopulate the denuded root surface. To achieve that purpose, a topographically dressed piece of a physical barrier, i.e., a nonresorbable membrane such as Millipore or Teflon (Gore-Tex) or a biodegradable membrane such as collagen or Vicryl (Polyglactin 910; Zappa, 1991a,b) is placed over the denuded root surface and covered by mucosal tissue. In that position, the membrane extends from the outer bone surface apical and lateral to the root defect coronally underneath the gingival margin, and allows cells of the adjacent periodontal ligament and of alveolar bone to repopulate the blood clot that fills the space between root and barrier, and later the denuded root surface. At the same time, the barrier prevents gingival epithelium and connective tissue from contacting that surface. Such barrier membranes have been used clinically in humans (Nyman et al., 1982b, 1987; Gottlow et al., 1986; Pontoriero et al., 1987, 1988, 1989; Stahl et al., 1990; Stahl and Froum, 1991a,b) as well as in experimental animals such as dogs (Aukhil et al., 1983, 1986a; Pitaru et al., 1987, 1988, 1989; Caffesse el al., 1988, 1990, 1991; Magnusson et al.,


47

1988, Claffey et al., 1989) and non-human primates (Gottlow er al., 1984; Magnusson et al., 1985; Aukhil and Iglhaut, 1988; Ettel et al., 1989; Gottlow et al., 1990) presenting with artificial periodontitis produced by the ligature technique (see Section V,A). In part, such experiments were combined with crown amputation and mucosal closure in order to exclude the epithelium from wound healing (Gottlow et al., 1984; Claffey et al., 1989). Unfortunately, the guided tissue regeneration technique is hampered by clinical problems such as incorrect barrier placement due to topographical complexity, postoperative gingival recession, and shrinkage resulting in the exposure of the coronal edge of the barrier membrane, reinfection along the exposed membrane, and premature loss of the latter. On the other hand, the barrier membrane may help to protect the blood clot from mechanical disturbances (Claffey et al., 1989; Selvig et al., 1990). In the most favorable circumstances, a new cementum matrix is formed along up to 70% of the denuded root surface (Minabe, 1991), extending from the apical or lateral border of the still intact periodontium coronally for distances varying between 1 and 5 mm. This regenerated cementum matrix, presumably calcified in uiuo, is variably thin; tapers coronally ; overlaps the old cementum apically; includes cells (i.e., cementocytes or osteocytes ?); and, unless the denuded root surface has preceedingly undergone superficial resorption by dentoclasts, does not adhere to the established root tissues (Fig. 12), neither old cementum nor dentine (Gottlow er al., 1984, 1990; Ettel er al., 1989; Zappa, 1991b). The outer surface of this new cementum matrix is covered by connective tissue that, in the human, rarely shows dense aggregates of collagen fibers oriented perpendicularly to and inserting into that new cementum, although, on the basis of classical histologic sections, some authors claimed to have demonstrated new fiber attachment after about 3 months of healing (Gottlow et al., 1986). This is different in dogs and non-human primates in which a dense and oriented fiber arrangement in the periodontal ligament and fiber insertion into a new cementum matrix is observed more frequently (Aukhil et al., 1986a; Ettel et al., 1989; Caffesse et al., 1988, 1990, 1991; Pitaru et al., 1988; Gottlow et al., 1990). In contrast, new alveolar bone regenerates more easily and may restore interradicular septi entirely (Caffesse et al., 1990). However, that portion of the denuded root surface not covered by either epithelium or new cementum frequently and unpredictably undergoes resorption and ankylosis. On the other hand, if root resorption is followed by the production of new cementum matrix, the latter is intimately bound and irreversibly attached to that root surface (Aukhil et al., 1983, 1986a; Caffesse et al., 1990, 1991). In fact, Egelberg (1987) stated very clearly that “if initial surface resorption occurs, newly formed collagen fibrils could interdigitate . . . with collagen fibrils from the cementum or the

40

HUBERT E. SCHROEDER


49

dentrine matrix, being exposed by the resorptive process. . . . In its absence, the newly formed matrix splits away from the root surface.” Furthermore, along the intact periodontium bordering the denuded root defect, cells of the periodontal ligament react with a proliferative burst only within a very limited zone of 200 to 300 p m during the first 7 days after surgery and there is some evidence, only from experiments with the window technique, that such cells actually migrate toward the denuded root surface (Aukhil and Iglhaut, 1988; Iglhaut et a / . , 1988). There is no evidence, however, to indicate that cells derived from the alveolar bone participate in the regenerative process leading to new cementum matrix formation. But, when the influx of cells from intact portions of bone and the periodontal ligament is blocked by a barrier membrane placed apically around the root, regenerative processes fail to occur (Aukhil et a/., 1987).

D. Biological Problems After several decades of research, it seems clear at present that under both spontaneous and guided conditions, periodontal tissue regeneration, in particular regenerative cementogenesis, along established but formerly diseased and denatured root surfaces is not achieved in the true sense of the term. Even under the best possible, so-called guided conditions, unpredictable repair rather than regeneration (i.e., the true morphological and functional restoration of the periodontium) results from wound healing in that area, at least after short-term observations in the human. The newly formed tissue that in part develops along such root surfaces is a cellular, rapidly generated variety of either cementum or bone, and this tissue, with or without a dense collagen fiber insertion, is not functionally attached to the old, denatured root tissues, be it cementum or dentine. There is no indication that a fiber fringe matrix and AEFC are deposited initially or

FIG. 12 (a-c) Light micrographs depicting newly formed cellular cementum (NC) atop and along a thick layer of old acellular extrinsic fiber cementum (AEFC), with and without root resorption (a). Where resorption (R)took place. NC is firmly attached to the old cementum (a,b); without resorption, NC is simply superimposed and separates (arrows) from the old cementum due to preparatory shrinkage (a.c). The same is true at the apical termination, marked by an artificial notch (N), of the mechanically treated root surface (inset). There is no attachment or insertion of new connective tissue (CT) fibers in NC. [The material originated from human patients undergoing guided regeneration treatment (with a resorbable Vicryl membrane) of lower incisors, after 3 months of regenerative wound healing.] CC, old cellular cementum. Courtesy Dr. Urs Zappa, Bern. Magnifications: a, x 100; b,c, ~ 3 2 0 ; inset. x100.

50

HUBERT E. SCHROEDER

later in time. The reasons for this may be found in basic biological issues regarding cementogenesis. Apart from clinical problems mentioned above, the basic biological issues center around two questions: ( a )the repopulation of the blank root surface and its covering blood clot with cells appropriate for regenerating cementum and the periodontal ligament, and ( b )the functional attachment of the regenerated tissue to the denatured root surface. In his review, Minabe (1991) concluded that “at present, the attainment of a new functional periodontal ligament over extensive portions of treated root surfaces may not be realistic,” because “without a method for selective repopulation of the root surface, the support system of the cementum, periodontal ligament, and bone cannot be reliably renewed.” When teeth develop preeruptively, cells forming root cementum derive directly or indirectly from the dental follicle proper that envelops the growing root. Apparently the appropriate cells receive their differentiation signal in a time-dependent fashion, isochronous with root elongation in the apical direction. Various varieties of cementum are formed along a corono-apical gradient. On the other hand when periodontal defects with denatured root surfaces enter protected wound healing, the respective cells, supposedly deriving from still intact portions of the periodontal ligament and alveolar bone, have to follow an apico-coronal pathway, with the gradient of cementogenesis reversed. In this situation, the questions arise: how fast do these cell populations proliferate, are the appropriate cementogenic cells present or available through precursors, how fast and, how far are they able to migrate, and what are the signals to turn on their matrix-producing activity at a particular wound or defect locality? At present, these questions have no answers. Periodontal ligament cells are developmentally, topographically, and functionally different from gingival fibroblasts, and they represent heterogeneous mixtures of phenotypically or through environmental modulation different subsets of largely unknown ability and proportion. Although osteoblasts precursors might occur, it is unkown whether cementogenic cells or their precursors are part of that mixture or not. But even when present, such cells may not divide frequently, may be unable to respond to differentiation signals or to migrate extensively, or may decrease in number with age. This is particularly valid topographically, i.e., for the remaining, intact ligament tissue around the root apex or in the furcational region. In fact, it is not known whether the cells producing the new, cellular “cementum” derive from the ligament or the alveolar bone. However, the fact that root resorption and ankylosis are frequent and unpredictable side effects, particularly when epithelium is artificially excluded from the wound healing process, indicates that connective tissue and cells other than those with the appropriate cementogenic capacity have arrived first and proliferate along the root surface.


51

Whatever origin and composition the new cementum-like tissue may have, its matrix does not attach to the hard and denatured root surface. Clear evidence to the mere deposition of that matrix is the fact that during histologic processing, it splits away from the established root tissues. The reason is lack of matrix fibril interdigitation secured by mineralization as it occurs during natural cementogenesis. Believing that demineralized root surfaces held an induction potential for bone and cementum regeneration, Register (1973) and Register and Burdick (1975, 1976), employing hydrochloric, lactic, citric, phosphoric and other acids, initiated a series of studies showing that superficial demineralization of the denuded root surfaces facilitated and accelerated fiber reattachment (Holden and Smith, 1983). Later on, it was shown in the dog that citric acid demineralization to a depth of 3 to 5 pm indeed favors the interdigitation of the newly formed and acid-exposed collagen fibril populations (Ririe et al., 1980; Selvig, 1983),mimicking the first step of a spontaneous resorption process, i.e., demineralization of the crystals along the resorptive surface (Frank et al., 1974). A similar result, i.e., dense collagen fibril reattachment in superficially demineralized (about 1 to 2 pm deep) dentin, was observed along the apical portion of a natural defect in a 72-year-old woman (Heritier, 1983). Unfortunately, under routine clinical conditions and in further animal experiments root demineralization has met with controversial success.

VI. Concluding Remarks and Perspectives Although some first steps have been made in supplying urgently needed information, the genesis of the various types of cementum on growing human teeth, the conditions under which they increase in thickness, and the functional roles they might fulfill are still unknown. The task of deciphering all this is very complex because ( a ) the various types of cementum developing in selected areas, as a pure or mixed variety, simultaneously or in alternating layers, often behave unpredictably in time and space, ( 6 ) the full spectrum of all possible combinations is not yet known, and ( c )human teeth are not freely available and rarely subject to experimentation. Nevertheless, the need for further information is imperative. One aspect, however, appears secured, i.e., that a force resistant union between cementum and dentine is generated only via matrix fibril interdigitation and subsequent mineralization. The failure in establishing that kind of union is one of the basic problems in regenerative cementogenesis. Another is that the cells producing the various types of cementum are virtually unknown. For that reason, neither in vitro nor in vivo models for

52

HUBERT E. SCHROEDER

studying regenerative cementogenesis are particularly promising, unless much more is known about where these cells come from, when they do what and why, and where they eventually go. Animal models, such as dogs and non-human primates, if still available in the future, may project a response picture too optimistic to be used in analogy to aging humans. Very little is known about the animal's spontaneous cementogenesis, their particular types of root cementum, and their regenerative response potential in general. Future perspectives probably lie along the route of dynamic labeling studies, and the imperative task is to establish knowledge, for the various types of root cementum, about their developmental system characterized by particular classes of cells and their morphologically, biochemically, and functionally defined matrix products.

Acknowledgments The author is indebted to Dr. D. D. Bosshardt, Zurich, for his generosity in permitting the use of part of his unpublished material on developmental cementogenesis; to Dr. U. Zappa, Dental Clinics, University of Bern, for allowing photography and reproduction of part of his experimental material on guided tissue regeneration: to Mr. H. Maag, Department of Orthodontics, University of Zurich, for translating the author's raw sketches into representative drawings; to Mrs. Susi Munzel-Pedrazzoli for competent help in designing and producing the illustrative plates: and to Mrs. Rosmarie Kroni for handling the reference list and for typing and editing the manuscript.

References Abbas, D. K., Albandar, J. M., Helgeland, K., and Johansen, J. R. (1987). Acta Odontol. Scand. 45, 353-360. Abdallah. F., Kon, S., and Ruben, M. P. (1988). J . West. SOC. Periodontol. (Periodont. Abstr.) 36, 53-61. Andreasen, J. O., and Kristerson, L. (1981). fnt. J . Oral. Surg. 10, 189-201. Arceo, N., Sauk, J. J., Moehring, J., Foster, R. A,, and Somerman, M. J. (1991). J . Periodonrol. 62, 499-503. Aukhil. 1. (1991). Dent. Clin. North. Am. 35, 459-467. Aukhil, I . , and Fernyhough, W. S. (1986). J . Periodontol. 57, 405-412. Aukhil, I . . and Iglhaut, J. (1988). 1. Clin. Periodonrol. 15, 374-382. Aukhil, I . , Pettersson, E., and Suggs, C. (1986a). J. Periodontol. 57, 727-734. Aukhil, I . Simpson. D. M., Suggs. C., and Pettersson, E. (l986b). J . Clin. Periodontol. 13, 862-868. Aukhil, I., Pettersson, E., and Suggs, C. (1987). J . Periodonrol. 58, 71-77. Aukhil, I., Simpson, D. M., and Schaberg, T. V. (1983). J . Periodont. Res. 18, 643-654. Azaz, B., Michaeli, Y.,and Nitzan, D. (1977). Oral Surg. Oral Med. Oral Parhol. 43, 572-578. Azaz, B., Ulmansky, M., Moshev, R., and Sela, J. (1974). Oral Surg. 38, 691-694. Barrett, A. P., and Reade, P. C. (1981). J . Periodont. Res. 16, 456-465.


53

Bartold. P. M.. and Page. R. C. (1985). J . Periodont. Res. 20, 284-292. Beertsen, W., and Everts, V. (1990). J . Dent. Res. 69, 1669-1673. Beertsen, W., and Van Den BOS,T . (1991). J . Dent. Res. 70, 176-181. Bellows, C. G.. Melcher, A. H., and Aubin. J. E. (1981). J . Cell Sci. 50, 299-314. Bellows, C. G., Melcher, A. H., Bhargava, U., and Aubin. J. E. (l982a). J . Ultrastrucr. Res. 78, 178-192. Bellows, C. G., Melcher, A. H., and Aubin, J . E. (1982b). J . Cell Sci. 58, 125-138. Bellows, C. G., Melcher, A. H., and Brunette, D. M. (1980). J . Cell Sci. 44, 59-73. Bernstein, A. B., Preisig, E., and Schroeder, H. E. (1988). Cell Tissue Res. 254, 659-670. Bernstein. A. B., Preisig, E., and Schroeder, H . E. (1989a). Cell Tissue Res. 255,631-639. Bernstein, A. B., Preisig, E., Pajarola, G., and Schroeder, H. E. (1989b). Cell Tissue Res. 258, 125-135. Bernstein, A. B., Preisig. E., and Schroeder, H. E. (1990). Cell Tissue Res. 259, 603-606. Bhargava, U., Bar-Lev, M . , Bellows, C. G., and Aubin, J. E. (1988). Bone 9, 155-163. Birk, D. E., and Trelstad, R. L. (1984). J . Cell Biol. 99, 2024-2033. Birk, D. E., and Trelstad, R. L. (1985). Ann. N . Y . Acad. Sci. 460, 258-266. Birk, D. E., and Trelstad, R. L . (1986). J . CellBiol. 103, 231-240. Birk, D. E., Zycband, E. I., Winkelmann, D. A,. and Trelstad, R. L. (1990). Ann. N . Y . Acad. Sci. 580, 176-194. Blomlof. L., Lindskog. S. , and Hammarstrom, L. (1988). Swed. Dent. J . 12, 101-112. Blomlof, L., and Otteskog, P. (1981). Scand. J . Dent. Res. 89, 43-47. Bordin, S ., Kolb, W. P., and Page, R. C. (1983). J . fmmunol. 130, 1871-1875. Bordin, S . , and Page, R. C . (1988). In Vitro 24, 719-726. Bordin, S . , and Page, R. C. (1989). Mol. Immunol. 26, 677-685. Bordin. S . , Page, R. C.. and Narayanan, A. S. (1984). Science 223, 171-173. Bosshardt, D. D . . Luder, H. U., and Schroeder, H. E. (1989). J . Biol. Buccale 17, 3-13. Bosshardt, D. D., and Schroeder, H. E. (1990). J . Clin. Periodontol. 17, 663-668. Bosshardt, D. D., and Schroeder, H . E. (1991a). Cell Tissue Res. 263, 311-324. Bosshardt. D. D.. and Schroeder, H. E. (1991b). Cell Tissue Res. 263, 325-336. Bosshardt, D. D.. and Schroeder, H. E. (1992). Cell Tissue Res. 267, 321-335. Bowers, G. M., Chadroff, B., Carnevale, M., Mellonig, J . , Corio, R . , Emerson, J . , Stevens, M., and Romberg, E. (1989a). J . Periodontol. 60,664-674. Bowers, G. M., Chadroff, B., Carnevale, M., Mellonig, J., Corio. R.. Emerson, J., Stevens, M., and Romberg, E. (1989b). J . Periodontol. 60, 675-682. Bowers, G. M., Chadroff, B., Carnevale, M., Mellonig, J . , Corio, R., Emerson, J., Stevens, M., and Romberg, E. (1989~).J . Periodontol. 60, 683-693. Bowers, G. M., Granet, M., Stevens, M.. Emerson, J.. Corio, R., Mellonig, J . , Lewis, S. B.. Peltzman, B., Romberg, E., and Risom, L. (1985). J . Periodontol. 56, 381-395. Boyko, G. A., Brunette. D. M., and Melcher, A. H. (1980).J . Periodont. Res. 15,297-303. Boyko, G. A., Melcher, A. H., and Brunette, D. M. (1981).J . Periodont. Res. 16, 73-88. Butcher. E. O., and Vidair, R. V. (1955). J . Dent. Res. 34, 569-576. Caffesse, R. G., Dominguez, L. E., Nasjleti, C. E., Castelli, W. A., Morrison, E. C., and Smith, B. A. (1990). J . Periodontol. 61, 45-50. Caffesse. R. G . , Nasjleti, C. E., Anderson, G. B., Lopatin, D. E., Smith, B. A,, and Morrison, E. C. (1991). J . Periodontol. 62, 21-29. Caffesse, R. G . , Smith, B. A., Castelli, W. A,. and Nasjleti, C. E. (1988). J . Periodontol. 59,589-594. Caton, J. G.. and Kowalski, G. J . (1976). J . Periodontol. 47, 506-510. Caton, J . G., and Zander, H . A. (1975). J . Periodontol. 46, 71-77. Cho, M.-I., and Garant, P. R. (1981). Anat. Rec. 199, 309-320. Cho, M.-I., and Garant, P. R. (1985). J . Periodont. Res. 20, 185-200.

54

HUBERT E. SCHROEDER

Cho, M.-I., and Garant, P. R. (1988). J . Periodont. Res. 23, 268-276. Cho, M.-I., and Garant, P. R. (1989). Anat. Rec. 223, 209-222. Cho, M.-I., Garant, P. R., and Lee, Y. L. (1988). J . Periodont. Res. 23, 287-294. Chomette, G., Auriol, M., Armbruster, D., Szpirglas, H., and Vaillant, J. M. (1987). J. Biol. Buccale 15, 217-224. Christner, P., Robinson, P., and Clark, C. C. (1977). Calcif. Tissue Res. 23, 147-150. Claffey, N., Hahn, R., and Egelberg, J. (1989). J. Clin.Periodontol. 16, 371-379. Cole, R. T., Crigger, M., Bogle, G., Egelberg, J., and Selvig, K. A. (1980). J. Periodont. Res. 15, 1-9. Connor, N. S., Aubin, J. E., and Sodek, J. (1983). J . Cell Sci. 63,233-244. Crigger, M., Bogle, G., NilvCus, R., Egelberg, J., and Selvig, K. A. (1978). J. Periodont. Res. 13, 538-549. Dastmalchi, R., Polson, A., Bouwsma, O., and Proskin, H. (1990). J . Clin.Periodontol. 17, 709-7 13. Davidovitch, Z . , Finkelson, M. D., Steigman, S., Shanfeld, J. L., Montgomery, P. C., and Korostoff, E. (1980). Am. J. Orthod. 77, 33-47. Diab, M. A., and Stallard, R. E. (1965). Periodontics 3, 10-14. Ecarot-Charrier, B., Glorieux, F. H., van der Rest, M., and Pereira, G. (1983). J. Cell Biol. 96,639-643. Ecarot-Chanier, B., Shepard, M., Charette, G., Grynpas, M., and Glorieux, F. H. (1988). Bone 9, 147-154. Egelberg, J. (1987). J. Periodont. Res. 22, 233-242. Eley, B. M., and Harrison. J. D. (1975). J. Periodont. Res. 10, 168-170. Ellegaard, B., Karring, T., Listgarten. M., and Loe, H. (1973). J. Periodontol. 44,209-217. Engel, D., Schroeder, H. E., Gay, R., and Clagett, J. (1980). Arch. OralBiol. 25, 283-296. Enwemeka, C. S. (1991). Tissue Cell 23, 173-190. Ettel, R. G., Schaffer, E. M., Holpuch, R. C., and Bandt, C. L. (1989). J . Periodontol. 60, 342-351. Fardal, O., and Lowenberg, B. F. (1990). J . Periodontol. 61, 529-535. Fernyhough, W., Aukhil, J., and Link, T. (1987). J. Periodontol. 58, 762-769. Fernyhough, W., and Page, R. C. (1983). J . Periodontol. 54, 133-140. Fleischmajer, R., Perlich, J. S., Timpl, R., and Olsen, B. R. (1988). J. Histochem. Cytochem. 36, 1425-1432. Formicola, A. J., Krampf, J. I., and Witte, E. T. (1971). J . Periodontol. 42, 766-773. Frank, R. M., Fellinger, E., and Steuer, P. (1976). J . Biol. Buccale 4, 295-313. Frank, R. M., Fiore-Donno, G., Cimasoni, G., and Matter, J. (1974). J. Periodontol. 45, 626-635. Frank, R. M., and Frank, P. (1969). Z . Zellforsch 99, 121-133. Furseth, R. (1967). Acta Odontol. Scand. 25, 613-645. Furseth, R. (1969). Archs Oral. Biol. 14, 1147-1 158. Garant, P. R. (1976). J . Periodontol. 47, 380-390. Garant, P. R., and Cho, M.4. (1979). J. Periodont. Res. 14, 95-106. Gerstenfeld, L. C., Chiprnan, S. D., Kelly, C. M., Hodgens, K. J., Lee, D. D., and Landis, W. J. (1988). J. Cell. Biol. 106, 979-989. Giniger, M. S., Norton, L., Sousa, S . , Lorenzo, J. A., and Bronner, F. (1991). J. Dent. Res. 70, 99-101. Gottlow, J., Karring, T., and Nyman, S. (1990). J. Periodontol. 61, 680-685. Gottlow, J., Nyman, S., Karring, T., and Lindhe, J. (1984). J . Clin. Periodontol. 11,494-503. Gottlow, J., Nyman, S., Lindher, J., Karring, T., and Wennstrom, J. (1986). J. Clin.Periodontol. 13, 604-616. Could, T. R. L. (1983). J. Dent. Res. 62, 873-876.


55

Gould, T. R. L., Melcher. A. H., and Brunette, D. M. (1977). Anat. Rec. 188, 133-142. Gould. T. R. L., Melcher, A. H., and Brunette, D. M. (1980).J. Periodont. Res. 15, 20-42. Hassel, T. M., and Stanek. E. J. (1983). Arch. Oral Biol. 28, 617-625. Held, A.-J. (1951). Oral Surg. Oral Med. Oral Pathol. 4, 53-67. Heritier, M. (1983). J. Periodontol. 54, 515-521. Holden, M. J., and Smith, B. A. (1983). J . West. Soc. Periodontol. (Periodont. Abstr.) 31, 45-56. Houston, F., Sarhed, G . , Nyman, S., Lindhe, J., and Karring, T. (1985).J . Clin. Periodontol. U ,716-727. Hurst, R. V. V. (1972). J . Dent. Res. (Suppl.) 51, 1183-1192. Hurum, S., Sodek, J., and Aubin, J. E. (1982). Biochem. Eiophys. Res. Commun. 107, 357-366. Iglhaut, J., Aukhil, I., Simpson, D. M., Johnston, M. C., and Koch, G. (1988). J . Periodont. Res. 23, 107-1 17. Jansen, M. T., Coppes, L., and Verdenius, H. H. W. (1955). J . Periodontol. 26, 292-300. Johansson, 0.. NilvCus, R., and Egelberg, J. (1978). J . Periodont. Res. 13, 525-531. Jones, S.J. (1981). In “Dental Anatomy and Embryology” ( J . W. Osborn, ed.), pp. 193-205, 286, 294. Blackwell Scientific. Oxford. Karring, T., Nyrnan, S., and Lindhe, J. (1980). J. Clin. Periodontol. 7, 96-105. Karring, T., Nyrnan, S.. Lindhe, J., and Siriat, M. (1984). J. Clin. Periodontol. 11, 41-52. Kawase, T., Sato, M., Yarnada, M., Hirayama, A., Miake, K., and Saito, S. (1986). Bone Mineral Res. Suppl. 1,63. Kenney, E. B., and Ramfiord, S. P. (1969). J . Dent. Res. 48, 114-119. Knox, B., and Aukhil, I. (1988). J. Periodont. Res. 23, 60-67. KO, D. S . , Page, R. C., and Narayanan, A. S. (1977). Proc. Natl. Acad. Sci. U . S . A . 8, 3429-3432. Korotzer, T. I., Clagett, J . A., Kolb, W. P., and Page, R. C. (1980). J . Cell Physiol. 105, 503-512. Leblond, C. P. (1989). Anat. Rec. 224, 123-138. Lester, K. S. (1969). J. Ultrastruct. Res. 27, 63-87. Lester, K. S., and Boyde, A. (1970). J . Ultrastruct. Res. 33, 80-94. Limeback, H. F., Sodek, J., and Aubin, J. E. (1983). J. Periodont. Res. 18, 242-248. Limeback, H. F., Sodek, J., and Brunette, D. M. (1978). Eiochem. J . 170, 63-71. Lindhe, J.. Nyman, S., and Karring, T. (1984). J. Clin. Periodontol. 11, 33-40. Listgarten, M. A. (1973). J . Periodonf. Res. 8, 335-345. Listgarten, M. A., and Rosenberg, M. M. (1979). J. Periodontol. 50, 333-344. Litwin, J., Lundquist, G., and Soder, P.O. (1971). Scand. J. Dent. Res. 79, 536-539. Lowenberg, B. F., Aubin, J. E., Deporter, D. A,, Sodek, J., and Melcher, A. H. (1985). J . Dent. Res. 64, 1106-1 110. Lowenberg, B. F., Pilliar, R. M., Aubin, J. E., Fernie, G. R., and Melcher, A. H. (1987). J . Dent. Res. 66, 1000-1005. Lowenberg, 9. F., Thibault, J., Lawrence, C., and Sodek. J. (1986). J . Dent. Res. 65, 1010-10 15.

Magnusson, I., Batich, C.. and Collins, B. R. (1988). J. Periodontol. 59, 1-6. Magnusson, I., Nyman, S., Karring, T., and Egelberg, J. (1985). J . Periodont. Res. 20, 20 I -208. Maniatopoulos, C., Sodek, J., and Melcher, A. H. (1988). Cell Tissue Res. 254, 317-330. Marchi, F., and Leblond, C. P. (1983). Am. J. Anat. 168, 167-197. Marchi, F., and Leblond, C. P. (1984). J. Cell Eiol. 98, 1705-1709. Marks, S. C., and Popoff, S. N. (1988). A m . J. Anat. 183, 1-44. McAllister, B., Narayanan, A. S., Miki, Y., and Page, R. C. (1990). J . Periodont. Res. 25, 99- 105.

56

HUBERT E. SCHROEDER

McBride, D. J., Hahn. R. A,, and Silver, F. H. (1985). I n t . J. Biol. Mncromol. 7, 71-76. McCulloch, C. A. G. (1985). Anat. Rec. 211, 258-262. McCulloch, C. A. G., and Bordin. S. (1991). J . Periodont. Res. 26, 144-154. McCulloch, C. A. G., and Knowles, G. (1991). Cell Tissue Res. 264, 87-94. McCulloch, C. A. G., and Melcher, A. H. (1983a). J . Periodont. Res. 18, 339-352. McCulloch, C. A. G., and Melcher. A. H. (1983b). Am. J . Anat. 167, 43-58. McCulloch, C. A. G . , Nemeth, E., Lowenberg, G., and Melcher, A. H . (1987). Anat. Rec. 219, 233-242. Melcher, A. H., and Chan, J. (1981). J . Ultrastruct. Res. 77, 1-36. Melcher, A. H., Cheong, T., Cox, J., Nemeth, E., and Shiga. A. (1986). J . Periodont. Res. 21, 592-612. Melcher, A. H., McCulloch, C. A. G . , Cheong, T.. Nemeth, E., and Shiga, A. (1987). J . Periodont. Res. 22, 246-247. Messadi, D. V . , and Bertolami, C. N. (1991). Dent. Clin. North. A m . 35, 443-457. Miki, Y., Narayanan, A. S., and Page, R. C. (1987). J. Dent. Res. 66, 1399-1403. Minabe, M. (1991). J . Periodonr. Res. 62, 171-179. Nakae, H. Ikrayanan, A. S . , and Page, R. C. (1989).J . Dent. Res. 68,365, Abstr. No. 1439. Narayanan, A. S . , and Page, R. C. (1983). Collagen Re/. Res. 3, 33-64. Nefussi. J. R., Sautier, J . M.. Nicolas. V., and Forest, N. (1991). J . Biol. Buccale 19, 75-82. Nijweide. P. J., Burger, E. H., and Feyen. J . H. M. (1986). Physiol. Rev. 66, 855-886. Nishimura, K., Hayashi, M., Matsuda, K., Shigeyama, Y.. Yamasaki. A., and Yamaoka, A. (1989). J . Periodonr. Res. 24, 146-148. Nojima, N., Kobayashi, M., Shiomone, M., Takahashi. N.. Suda. T., and Hasegawa, K . (1990). J . Periodont. Res. 25, 179-185. Nyman, S ., Gottlow, J., Lindhe. J., Karring. T., and Wennstrom. J . (1987). J . Periodont. Res. 22, 252-254. Nyman, S . , Karring, T., Lindhe, J., and Planten, S . (1980). J. Clin. Periodontol. 7,394-401. Nyman, S., Lindhe, J., and Karring, T. (1989). I n “Textbook of Clinical Periodontology” ( J . Lindhe, ed.), 2nd edition, pp. 450-476. Munksgaard, Copenhagen. Nyman, S., Gottlow, J., Karring, T., and Lindhe, J. (1982a).J. Clin. Periodontol. 9,257-265. Nyman, S ., Lindhe, J . , Karring, T., and Rylander, H. (1982b). J. Clin. Periodontol. 9, 290-296. Ogura, N. Mera, T., Sato, F., and Ishikawa, I. (1991). J . Periodontol. 62, 284-291. Oikarinen, K. S ., and Seppa, S . T. (1987). Endod. Dent. Traumatol. 3, 95-99. Olson, S., Arzate, H., Narayanan, A. S. , and Page. R. C. (1991). J . Denf. Res. 70, 12721277. Oshima, M., Kuwata, F.. Otsuka, K., and Suzuki, K. (1988). I n “Recent Advances in Clinical Periodontology” ( J . Ishikawa, H. Kawasaki, K. Ikeda. and K. Hasegawa, eds.), pp. 573-577. Excerpta Medica, Amsterdam. Owens, P. D. A. (1980). Arch. Oral B i d . 24, 901-907. Palmer, R. M., and Lumsden, A. G. S . (1987). Arch. Oral B i d . 32, 281-289. Paynter, K. J.. and Pudy, G. (1958). Anat. Rec. 131, 233-251. Pettersson, E. C., and Aukhil, I . (1986). J. Periodont. Res. 21, 543-552. Piche, J. E., Carnes, D. E., and Graves, D. T. (1989). J . Dent. Res. 68, 761-767. Pitaru, S . , and Melcher, A. H. (1983). J . Periodont. Res. 18, 483-500. Pitaru, S . . and Melcher, A. H. (1987). J . Periodont. Res. 22, 6-13. Pitaru. S ., Gray. A., Aubin, J. E., and Melcher, A. H . (1984a). J . Periodont. Res. 19, 408-41 8. Pitaru, S ., Aubin. J. E.. Gray, A., Metzger, Z . , and Melcher, A. (1984b).J . Periodont. Res. 19. 66 1-665.


57

Pitaru, S.. Tal, H., Soldinger. M., Azar-Avidan, 0..and Noff, M. (1987). J . Periodont. Res. 22, 331-333. Pitaru, S . , Tal. H., Soldinger, M.. Grosskopf, A.. and Noff, M. (1988). J . Periodontol. 59, 380-386. Pitaru, S.. Tal, H., Soldinger, M., and Noff, M. (1989). J . Periodont. Res. 24, 247-253. Polson. A. M. (1986). J . Clin. Periodonrol. 13, 995-999. Polson, A. M. (1987). Endod. Dent. Traumatol. 3, 45-57. Polson, A. M.. and Caton, J. (1982). J . Periodontol. 53, 617-626. Polson. A. M., and Proye. M. P. (1982). J . Clin. Periodontol. 9, 441-454. Polson, A. M.. and Proye, M. P. (1983). J . Periodontol. 54, 141-147. Pontoriero, R.. Lindhe, J.. Nyman, S. , Karring, T., Rosenberg, E., and Sanavi, F. (1988). J . Clin. Periodonrol. 15, 247-254. Pontoriero, R., Lindhe, J., Nyman. S . , Karring. T., Rosenberg, E.. and Sanavi, F. (1989). J . Clin. Periodontol. 16, 170-174. Pontoriero, R., Nyman, S. , Lindhe, J., Rosenberg, E., and Sanavi. F. (1987). J . Clin. Periodontol. 14, 618-620. Preisig, E., and Schroeder, H. E. (1988). J . Periodont. Res. 23, 211-221. Proye, M., and Polson, A. M., (1982). J . Periodontol. 53, 379-389. Quamstrom, E. E., and Page, R. C. (1986). J . Cell Sci. 84, 183-200. Ragnarsson, B.. Carr, G.. and Daniel, J. C. (1985). J . Dent. Res. 64, 1026-1030. Register. A. A. (1973). J . Periodontol. 44, 49-54. Register, A. A.. and Burdick, F. A. (1975). J . Periodontol. 46, 646-655. Register, A. A., and Burdick, F. A. (1976). J . Periodontol. 47, 497-505. Ririe, C. M., Crigger, M., and Selvig. K. A. (1980). J . Periodont. Res. 15, 314-327. Roberts, W. E. (1975a). Arch. Oral Biol. 20, 465-471. Roberts, W. E. (l975b). Am. J . Anal. 143, 363-370. Roberts, W. E., Mozsary. P. G., and Klingler, E. (1982). A m . J . Anat. 165, 373-384. Rose, G. G., Yajima, T., and Mahan. C. J. (1980). J . Periodont. Res. 15, 53-70. Rose, G. G.. Yamasaki, A,, Pinero, G. J., and Mahan, C. J. (1987). J . Periodont. Res. 22, 20-28. Russel, S. B., Russel, J. D., and Trupin. K. M. (1981). J . Cell. Physiol. 109, 121-131. Schellens, J. P. M.. Everts, F., and Beertsen, W. (1982). J . Periodont. Res. 17, 407-422. Schroeder. H. E. (1986). “The Periodontiurn,” Vol. V/5. “Handbook of Microscopic Anatomy.” Springer-Verlag. Berlin. Schroeder, H. E. (1987). “Orale Strukturbiologie,” pp. 28-33. Georg Thieme Verlag, Stuttgart. Schroeder, H. E . (1988). In “Periodontology Today” (B. Guggenheim, ed.), pp. 32-40. Karger, Basel. Schroeder. H. E. (1991a). “Pathobiologie oraler Strukturen,” 2, “uberarbeitete und erweiterte Aufl,” pp. 205-214. Karger, Basel. Schroeder. H. E. (1991b). Schweiz. Monatsschr. Zahnmed. 101, 279-284. Schroeder, H. E., Luder, H . U., and Bosshardt, D. D. (1992). Schweiz. Monatsschr. Zahnmed. 102, 20-31. Schuhmacher, G. H . , and Schmidt, H. (1983). “Anatomie und Biochemie der Zahne.” 3rd edition, pp. 438. Fischer. Stuttgart, New York. Selvig, K. A. (1963). Acta Odontol. Scand. 21, 175-186. Selvig, K. A. (1964). Acta Odontol Scand. 22, 105-120. Selvig, K . A. (1965). Acta Odontol Scand. 23, 423-441. Selvig, K. A. (1967). “Studies on the Genesis, Composition and Fine Structure of Cementum.” Theses. University of Bergen, Norway. Selvig, K. A., (1983). J . B i d . Buccule 11, 79-94.

58

HUBERT E. SCHROEDER

Selvig, K. A., Bogle, G., and Claffey, N. M. (1988). J . Periodontol. 59, 758-768. Selvig, K. A., N i l v h s , R. E., Fitzmorris, L., Kersten, B., and Khorsandi, S. S. (1990). J . Periodontol. 61, 515-520. Sequeiria, P., Bosshardt, D. D., and Schroeder, H. E. (1992). J . Periodontol. Res. 27, 134- 142. Skillen, W. G., and Lundquist, G. R. (1937). J. Am. Dent. Assoc. 24, 175-185. Smith, A. J., Leaver, A. G., and Smith, G. (1983). Arch. Oral Biol. 28, 1047-1054. Sodek, J. (1976). Biochem. J . 160, 243-246. Sodek, J. (1977). Arch. Oral Biol. 22, 655-665. Somerman, M. J., Archer, S. Y., Hassel, T. M., Shteyer, A., and Foster, R . A. (1987a). Arch. Oral Biol. 32, 879-883. Somerman, M. J., Perez-Mera, M., Merkhofer, R. M., and Foster, R. A. (1987b).J . Periodontol. 58, 349-351. Somerman, M. J., Archer, S. Y., Shteyer, A., and Foster, R. A. (1987~). J . Periodont. Res. 22, 75-77. Somerman, M. J., Prince, C. W., Sauk, J. J., Foster, R. A., and Butler, W. T. (1987d). J . Bone Mineral Res. 2, 259-265. Somerman, M. J., Archer, S. Y., Imm,G. R., and Foster, R . A. (1988). J . Dent. Res. 67, 66-70. Somerman, M. J., Foster, R. A., Imm, G. M., Sauk, J. J., and Archer, S. Y. (1989). J . Periodontol. 60,73-77. Somerman, M. J., Sauk, 1. J., Foster, R. A.. Dickerson, K., Norris, K., and Argraves, W. S. (1991). J . Periodont. Res. 26, 10-16. Somerman, M. J., Shroff, B., Argraves, W. S.. Morrison, G., Craig, A. M., Denhardt, D. T., Foster, R. A., and Sauk, J. J. (1990). J. Biol. Buccale 18, 207-214. Stahl, S. S., and Froum, S. J. (1991a). J . Clin. Periodontol. 18, 69-74. Stahl, S . S., and Froum, S. J. (1991b). J . Clin. Periodontol. 18, 685-689. Stahl, S. S., Froum, S., and Tarnow, D. (1990). J . Clin. Periodontol. 17, 191-198. Stones, H. H. (1934). Proc. R . Soc. Med. 27, 728-744. Sudo, H., Kodama, H.-A., Amagai, Y.. Yamamoto, S . , and Kasai, S. (1983). J . Cell Biol. 96, 191-198. Ten Cate, A. R. (1972). J. Anat. ll2,401-414. Ten Cate, A. R., and Deporter, D. A. (1974). Arch. Oral Biol. 19, 339-340. Ten Cate, A. R., Deporter, D. A., and Freemann, E. (1976). Am. J . Orthod. 69, 155- 168. Ten Cate, A. R., Mills, C., and Solomon, G. (1971). Anat. Rec. 170, 365-380. Terranova, V. P., Franzetti, L. C., Hic. S., DiFlorio, R. M., Lyall, R. M., Wikesjo, U. M. E., Baker, P. J., Christersson, L. A., and Genco, R. J. (1986). J . Periodontol. Res. 21,330-337. Terranova, V. P., Hic, S., Franzetti, L., Lyall, R. M., and Wikesjo, U. M. E. (1987). J . Periodontol. 58, 247-257. Terranova, V. P., Jendresen, M., and Young, F. (1989a). Adu. Dent. Res. 3, 69-79. Terranova, V. P., Odziemiec, C., Tweden, K. S., and Spadone, D. P. (1989b).J . Periodontol. 60,293-301. Weinstock, M . (1975). I n “Extracellular Matrix Influences on Gene Expression” (H. C. Slavkin and R. C. Greulich, eds.) pp. 119-128. Academic Press, New York. Weinstock, M., and Leblond, C. P. (1974). J . Cell Biol. 60,92-127. Williams, D. C., Boder, G. B., Toomey, R. E., Paul, D. C., Hillman, C. C., King, K. L., Van Frank, R. M., and Johnston, C. C.. Jr. (1980). Calcif. Tissue Int. 30, 233-246. Yajima, T. (1988). Histochemistry 90, 245-253. Yajima, T., Rose, G. G., and Mahan, C. J. (1980). J . Periodont. Res. 15, 267-287.


Yamamoto, T. (1986). Arch. Hisrol. Jpn. 49, 459-481. Yamamoto, T., and Wakita, M. (1990). J . Periodonr. Res. 25, 113-119. Yamamoto, T., and Wakita, M. (1991). J . Periodonr. Res. 26, 129-137. Yamamoto, T., and Wakita, M. (1992). J . Periodonr. Res. 27, 20-27. Yoshikawa, D. K . , and Kollar, E. J. (1981). Arch. Oral Biol. 26, 303-307. Zahnder, H. A . , and Hurzeler, B. (1958). J . Dent. Res. 37, 1035-1044. Zappa, U . (1991a). Schweiz. Monatsschr. Zahnmed. 101, 1147-1 151. Zappa, U. (1991b). Schweiz. Monatsschr. Zahnmed. 101, 1321-1324.

59


lmmunocytochemical Localization of Proteins in Striated Muscle Marvin H. Stromer Muscle Biology Group, Department of Animal Science, Iowa State University, Ames, Iowa 5001 1

1. Introduction

Localization of specific muscle proteins via immunofluorescence and with immunoelectron microscopy are both well-established approaches. The availability of monoclonal antibodies (mAb) now has made it feasible to locate specific epitopes, to probe functional sites on proteins, and to locate structural domains and their arrangement in s h . Although many valuable contributions have been made to our understanding of striated muscle by laboratories around the world, the papers cited in Table I are especially noteworthy for their enhancement of our knowledge about striated muscle. The proteins are listed in the same order in which they appear in this review. This review provides an overview of the results obtained by using antibodies to understand striated muscle better. Emphasis has been on the most recent papers, but when needed for purposes of perspective, a limited number of earlier papers have been included. Results from some studies where antibodies have been used to probe functional sites or domains on proteins have been cited, as have some recent reviews that provide important background information. Reviews by Groschel-Stewart (1980), Pepe (1983), and Pette and Staron (1990) provide valuable information on topics related to this chapter. Because of the breadth of the topic, the very large number of papers that deal with various aspects of this topic, and length limitation, it was impossible to include all papers on this subject. I apologize in advance to anyone who feels that a favorite paper of theirs should have been included. Such omissions were unintentional. /~ll~~r~lulK ~ PoUm i ClM/ 1Jf C?lo/il~?,VlJI. 142

61

Copyright 0 1992 by Academic Press. Inc. All right5 of reproduction in any form reserved.

62

MARVIN H. STROMER

TABLE I Examples of lmmunocytochemical Studies That Have Made Major Contributions to Our Understanding of Striated Muscle

Protein Myosin

Reference

Contribution

Pepe (1967b)

Provided the basis for thick-filament model and understanding how myosin molecules are arranged in thick filaments. Located C-protein isoforms in the A band of C-protein Dennis et a/. (1984) different muscle types. Troponin Ohtsuki (1975) Located the three components of the troponin complex along the thin filament. Described the behavior of nebulin filaments Nebulin Wang and Wright (1988) anchored at the Z line during muscle shortening. Demonstrated that titin filaments extend from Titin Fiirst et a/. (1988) the Z line to near the M line. Described the behavior of titin epitopes in the Maruyama et a/. (1989) A and I bands during muscle shortening and stretch. Desmin Granger and Lazarides (1978) Demonstrated that desmin IFs exist at the periphery of Z lines. Demonstrated that desmin IFs span Z line to Z Richardson et a / . (1981) line between adjacent myofibrils.

11. Localization of Proteins in Skeletal Muscle Cells A. Thick-Filament Proteins and A-Band-Associated Proteins

The early recognition that muscle contraction involved interactions between thick and thin filaments focused attention on the need to know which proteins were present in the A band and how they were arranged. The complexity of this task increased as additional proteins associated with thick filaments were identified. Current emphasis is on the relationship between the specific location of a protein and/or its isoforms and the functional properties of thick filaments. The protein that has received the most attention is myosin.

1. Myosin a . Probing the Structure of the A Band Pepe (1966; 1967a,b; 1968) was among the early workers to use fluorescein-conjugated myosin antibodies for localization with the fluorescence microscope and unlabeled myosin antibodies for electron microscope (EM) localization. Pepe’s success in

PROTEINS IN STRIATED MUSCLE

63

isolating antibodies to specific regions of the myosin molecule by absorbing myosin antibodies with tryptic fragments of myosin permitted him to demonstrate that antigenic sites on thick filaments were accessible to these antibodies only in certain portions of the A band. For example, myosin antibodies absorbed with the heavy meromyosin (HMM) tryptic fragment of the myosin molecule would provide antibodies that would only recognize the light meromyosin (LMM) antigenic sites. Absorption with LMM from a 60-min trypsin digest of myosin would produce an antibody fraction that would recognize both the HMM and the trypsin-sensitive sites. Absorption with LMM from a 15-min trypsin digest would yield an antibody fraction that recognizes only HMM because short-digest LMM contains both the LMM and the trypsin-sensitive sites. Antibodies specific for LMM bound in the outer thirds of each half of the A band (Pepe, 1971). Antibodies specific for the trypsin-sensitive region were localized in seven stripes in the middle third of each half of the A band. HMM-specific antibodies bound to sites in the inner third of each half of the A band only if sarcomere lengths were 2.5 pm or greater so that no actin filaments overlapped with this region of the A band. Reduction or abolition of antiHMM staining at shorter sarcomere lengths indicated that the actin-HMM interaction blocked the epitope required for anti-HMM binding. These myosin antibody staining patterns were an important part of the information used by Pepe (1967a, 1971) to propose a model for how myosin molecules were packed into thick filaments. In a subsequent study, Wachsberger et al. (1983) used a mAb specific for an epitope on the S1 fragment of skeletal myosin that was not related to the ATPase site. When myofibrils in relaxing medium were exposed to this antibody, there was binding along the entire length of the cross-bridge region in the A band. In the absence of relaxing medium, the antibody bound to narrow regions on each side of the bare zone and at the outer edges of the A band. Because this antibody binding pattern was not affected by linkage of myosin crossbridges to actin filaments, it was suggested that differences in molecular packing along the thick filament could cause differences in availability of the antigenic site. A rnAb that recognized an epitope 92 ? 5 nm from the C-terminal end of the LMM or tail fragment of myosin was used by Shimizu et al. (1985a) to map this epitope in the A band and in isolated thick filaments. In contradiction to the earlier results of Pepe (1967b) and of Lowey and Steiner (1972), who observed a nonuniform fluorescent staining of the A band with antibody to LMM, this antibody stained along almost the entire length of the A band. Thin sections of labeled muscle bundles contained 50 stripes spaced 14.3 nm apart in each half of the A band. Isolated thick filaments decorated with this LMM antibody had stripes of variable density that repeated every 14.6 nm along the filament. This pattern was interpre-

64

MARVIN H. STROMER

ted to mean there was an uninterrupted axial spacing of myosin rods in thick filaments and as support for Huxley’s (1963) proposal for how myosin molecules are packed in thick filaments.

b. Mapping Functional Sites on Myosin Molecules The vertebrate skeletal myosin molecule consists of two 223-kDa heavy chains (Strehler et al., 1986) and four light chains (LCs). Each molecule contains two 19-kDa LC2 or 5,5-dithiobis-(2-nitrobenzoate)(DTNB) light chains and combinations of two different alkali light chains, a 20-kDa LCl or A1 chain and/ or a 16-kDa LC3 or A2 chain. Each of the two heads of the myosin molecule contains 1 of LC2 and 1 of either LCl or LC3. The original nomenclature of the alkali light chains (A1 and A2) and the DTNB light chain was based on the method used to isolate them. Now the three types are usually referred to as LC 1, LC2, and LC3 in decreasing order of chain weight. Because of the complexity of both the molecule itself and the functions of myosin, including, among others, ATP hydrolysis, conformation changes, and calcium binding, there has been interest in locating sites on myosin that would affect either the function and/or the structure of the molecule. Shimizu et al. (1985b) found that a mAb that recognized LC2 in fast skeletal muscle would only stain myofibrils after incubation in EDTA or if the myofibrils were denatured by acetone or ethanol. When calcium or magnesium occupies the metal binding site on LC2, there may be a conformational change that inhibits antibody binding. Removal of the metal with EDTA or denaturing LC2 with organic solvents reverses the conformational change and the antibody can bind. Electron microscopy of rotary-shadowed myosin-antibody complexes shows that the antibody binding site and the calcium binding site are located near the head-rod junction of myosin. The question of the spatial location of the N-terminus of myosin heavy chains (MHC) in the head of myosin molecules was investigated by Sutoh et al. (1987) by using an antibody to the N-terminal eight residues. Rotaryshadowed HMM-antibody complexes showed that the antibody was located 12 nm from the head-rod junction and indicated that the N-terminus of the MHC is located in the middle region of the head. Miyanishi et al. (1988) produced antibodies to two of the three tryptic fragments (23, 50, and 20 kDa) of the S1 portion of HMM. Rotary shadowing of the myosin-antibody complexes showed that one-third of the myosin molecules had antibodies to 50-kDa fragments bound at the tip of the head and two-thirds bound 16 nm from the head-rod junction on the side of the head. Antibodies and Fab fragments of antibodies to the 23-kDa fragments both bound 14-18 nm from the head-rod junction at the middle of the head. Dan-Goor et al. (1990) also used trypsin to cleave S1 into 27-, 50-,


65

and 20-kDa fragments aligned in order from the N-terminus. One mAb to the 27-kDa fragments reacted with an epitope in the N-terminal23 residues of the 27-kDa N-terminal fragment. A second mAb to the 50-kDafragments reacted with a 3-kDa N-terminal region of the 50-kDa fragment. The 27and the SO-kDa epitope were 14 and 17 nm, respectively, from the head-rod junction of myosin. Binding SI to actin decreased the affinity of antibodies to 50-kDa fragments for S1 but did not affect the affinity of antibodies to 27-kDa fragments for S 1. Antibodies to 50-kDa fragments inhibited the K+ (EDTA) and the actin-activated ATPase activity of SI, but antibodies to 27-kDa fragments had no effect. This suggests that either the 50-kDa epitope is near the actin or ATP binding sites or there is some form of “communication” between these sites and the antibody binding site. The location of the alkali light chains (LCl and LC3) was determined by using antibodies specific for the N-terminal region (Waller and Lowey, 1985). These light chains are located in the “neck” region of the myosin head, very close to the LC2 chain. Tokunaga et al. (1987) used mAbs to determine that the N-terminus of LC2 is located at the head-rod junction and that the N-terminus of LCI is 11 nm from the head-rod junction. This proximity between LC2 and the alkali light chains may explain why LC2 affects interactions between alkali light chains and MHC. Additional detailed information on the location of both the alkali light chains and LC2 was provided by labeling cysteine residues with the thiol-specific reagent 5-(iodoacetamido) fluorescein (Katoh and Lowey, 1989). Each of the two alkali LCs has a single cysteine near the C-terminus, and LC2 was labeled at either cysteine 125 or cysteine 154. Fluorescyl antibodies were then added to the labeled myosin, and the complexes were rotary-shadowed. The cysteine of the alkali light chains was 9 nm closer to the tip of the myosin head than the cysteine of LC2, which was close to the head-rod junction. This indicates that the two types of light chains are not arranged colinearly in the myosin head. The effect of antibodies to the S2 subfragment of myosin on the contractile behavior of myofibrils was examined by Harrington et al. (1990). Antibody-treated myofibrils contracted at an initial rate similar to that of untreated myofibrils. This was followed, in antibody-treated myofibrils, by a much slower contraction that ended with I-pm sarcomere spacing still visible. Untreated myofibrils contracted into nonstriated globular structures. Skinned muscle fibers were able to develop much less isometric force in the presence of antibodies to S2. These results may indicate that the S2 region of the myosin molecule also may be involved in force production in working muscle. Edman er al. (1988) used two MHC mAbs in conjunction with studies on unloaded shortening velocity and myofibrillar

66

MARVIN H. STROMER

ATPase activity and concluded that variability in shortening velocity and myofibrillar ATPase activity that exists among frog twitch fibers is based on differences in MHC composition. c. Location of MHC

mRNA The approach most recently used to locate MHC mRNA is in situ hybridization with a biotinated cDNA probe. An antibody to biotin is then added, followed by a second antibody that is specific for the species in which the biotin antibody was produced and that is complexed with a label (e.g., colloidal gold). Pomeroy et al. (1991)found that MHC mRNA was localized mainly in the peripheral regions of 14-day chick pectoral muscle cells where developing myofibrils are abundant. Some MHC mRNA was also observed near the nucleus. Most MHC mRNA was associated with 4- or 10-nm-diameter filaments believed to be actin and intermediate filaments, respectively. Although MHC mRNA is concentrated near assembling myofibrils, most mRNA is 0.1-0.5 pm from the nearest myofilaments. This suggests that an unknown intervening process such as diffusion or transport associated with cytoskeletal filaments may be involved in transporting protein chains to the assembly sites. Wenderoth and Eisenberg (1991) used a similar approach with cryosections and with sections of LR white-embedded papillary muscle. In both types of preparations, MHC mRNA was in the intermyofibrillar space adjacent to both A and I bands but was not identified as being associated with any structures. Enzymatic detection of MHC mRNA with a light microscope had previously shown that this mRNA is concentrated at the myotendinous junction of stretched fibers, where there is active myofibril assembly (Dix and Eisenberg, 1990), under the sarcolemrna and, possibly, between myofibrils in rabbit leg muscle (Dix and Eisenberg, 1988). Hesketh et al. (1991) used a sulfur-35-labeled riboprobe prepared from slow MHC cDNA to show that, in the rat soleus, about 70% of the myosin mRNA is present in the intermyofibrillar cytoplasm and the remainder is concentrated in a thin subsarcolemmal rim. A second approach to locating MHC mRNA involves a combination of cDNA probes and antibody localization. The rat extraocular (EO) musculature simultaneously expresses at least six different MHC genes at the mRNA level because of a sequential activation of MHC genes without complete repression of any of the genes (Wieczorek et al., 1985).Antibody localization showed that four different isozymes are present and tend to segregate into different cells, but some contain more than one MHC type. Isoform transitions in rat EO muscles seem to be arrested at different stages in different groups of cells; this suggests that tissue-specific or environmental factors may modify the expression of MHC genes. Miller et al. (1989) prepared four mAbs to epitopes on the head fragment of


67

fast MHC and used a two-part cDNA expression approach to map their respective epitopes. Each of the epitopes was at or near the ATPase site. One of them (F59) was conserved to varying degrees through vertebrate evolution, as shown by antibody binding to a spectrum of vertebrate muscles. A fifth mAb (F27) recognized an epitope located on the red portion of fast MHC and was also similarly conserved through vertebrate evolution.

d. Use of AntibodiesforFiber Typing Because of the existence of several distinct classification systems for vertebrate twitch-type muscle fibers and because of the evolution of some of these systems, the papers chosen for inclusion in this section will be described by using the fiber-type terminology used by the respective authors. Fiber typing has traditionally been done by using histochemical staining for enzymatic activity, normally myofibrillar ATPase, after preincubation at pH values 4.3, 4.6 or 9-10. The three fiber types identified by this approach are often named slow-twitch-oxidative (type I, or red), fasttwitch-oxidative-glycolytic (type IIA, or intermediate), and fast-twitchglycolytic (type IIB, or white). Since the advent of reliable methods to prepare striated muscle for study in the transmission electron microscope, ultrastructural characteristics such as relative numbers of mitochondria per cell and Z-line width also have been used as indicators of fiber type. Because the speed of contraction of muscle fibers is related to the expression of different MHC isoforms, the immunocytochemical detection of MHC isoforms has been and continues to be emphasized. Production of polyclonal or monoclonal antibodies to specific isoforms of myosin increased not only the sensitivity of detection of fiber types but also the apparent complexity of fiber types. Several early studies compared fiber typing based on histochemical staining or ultrastructural criteria with results from immunocytochemical staining with specific antibodies. Gauthier (1979) found that fast-twitch rat muscle contained red, intermediate, and white fibers on the basis of mitochondria1 content. Some of the red fibers bound antibodies specific for white or fast myosin, whereas others did not. All red fibers had wide Z lines, regardless of their myosin type. This led to the suggestion that rat diaphragm muscle contains four fiber types: red fast, red slow, intermediate, and white. Fiber typing of adult mouse muscle on the basis of Z-line width differed from the type I and I1 classification obtained after using antibodies to fast and slow myosin isozymes (te Kronnie et al., 1980). Human type I fibers have been unambiguously identified by using colloidal gold labeling and a mAb specific for slow MHC (Semper et al., 1988). A combination of enzyme histochemistry and immunohistochemistry for slow and fast myosin demonstrated that human type I fibers had a slow

68

MARVIN H. STROMER

myosin, type IIA and IIB had fast myosin only, and type IIC had various proportions of slow and fast myosin (Billeter et al., 1980). Bormioli et al. (1980) raised antibodies against slow myosin from the chicken anterior latissimus dorsi (ALD) and from the guinea pig soleus (SOL). Fibers from various vertebrate species that stained with anti-ALD myosin corresponded to slow-tonic fibers. Mammalian fibers stained with anti-SOL myosin, however, corresponded to slow-twitch fibers. In mammals, only a minority of EO fibers and nuclear bag fibers were labeled with anti-ALD myosin, Only afew avian slow fibers reacted with anti-SOL myosin. Thus, vertebrate skeletal muscle may have two antigenically distinct, though partly cross-reacting, types of slow myosin. The larger avian posterior adductor profundus (AP) muscle was analyzed with mAbs to myosin and C-protein isoforms and found to contain essentially pure slow-tonic fibers that were similar to those of the ALD (Zhang et al. 1985). This could make the posterior AP muscle a valuable source of slow-tonic fibers. The fast avian adductor superficialis (AS) muscle stained as a type IIA fiber by the histochemical ATPase test, but reacted with both antibodies to slow and fast MHC. The heterogeneity of this muscle is due to the presence of two MHC isoforms and not to the presence of variable amounts of slow myosin in its fibers (Zhang et al., 1989). By a combination of enzyme histochemistry and immunohistochemistry, Schiaffino et al. (1989) and Gorza (1990) have identified a type IIX MHC in rat skeletal muscles. The IIX MHC is localized in a large subset of type I1 fibers, is coexpressed with type IIA or IIB MHC in a small number of fibers, and may be a novel MHC. Muscles with predominantly type IIX MHC have a shortening velocity intermediate between muscles containing type I and type IIB MHC. Antibodies to myosin isoforms also have been used to relate isoform expression to fiber types and function in certain specialized vertebrate striated muscles. A polyclonal antibody to fast MHC from bovine EO muscle did not react with MHC of other myosins (Sartore et al., 1987). Most fibers with a type I1 ATPase stain pattern were labeled with EO antibodies, but some reacted with both anti-EO and either anti-IIA or antifetal myosin isoforms. Sartore et al. (1987) suggested that the EO isoform may exist because these muscles are exceptionally fast contracting. The EO myosin isoform is seemingly unrelated to the superfast myosin in the muscles closing a cat’s jaw. Extraocular antibodies do not stain the muscles closing a cat’s jaw (e.g., posterior temporalis). Similarly, antibodies to rat posterior temporalis myosin do not stain EO muscles. In EO muscle fibers in the rat, both single- and multiple-innervated fibers in the orbital layer labeled along their length with a mAb specific for neonatal/embryonic mammalian myosin, except in the endplate region. All fibers labeled in the endplate region with a mAb specific for fast myosin (Jacoby et al., 1990). These authors hypothesized that twitch or tonic innervation may act locally to suppress either neonatal-like myosin or fast myosin, respectively.


69

Myosin isoforms also have been studied in human masticatory muscles and in ear muscles from various mammals. The myosin composition of human masticatory muscles was more heterogeneous than limb muscles and included various proportions of slow and fast myosins, heavy as well as light chains (Thornell et al., 1984a). Many fibers of the human masseter muscle contain more than one myosin isoform (Butler-Browne et al., 1988). This muscle contains, in addition to adult fast and slow myosin, neonatal MHC and embryonic myosin light chains, two components of developing muscle. Myosin isoform-specific antibodies showed that there is no typical fiber-type composition in middle-ear muscles of several species of carnivores and some primates (Mascarello et al., 1983). This suggests that different functional requirements may dictate the myosin isoform composition of middle-ear muscles in different mammals. Staining with mAbs to MHC isoforms demonstrated that different isoforms existed in a single fiber and that there were discrepancies when fiber typing by mAbs was compared with fiber typing by histochemical staining in tensor tympani and stapedius muscles of the rat (van den Berge and Wirtz, 1989a,b). Collectively, the properties of these muscles indicate that they can contract rapidly, are fatigue-resistant, and, thus, are able to protect the inner ear from noise damage.

e. Myosin Isofomis in Muscle Spindles Muscle spindles are specialized receptors in skeletal muscle that respond to stretch of either the entire muscle or of the equatorial region of the spindle as a result of contraction of the intrafusal fibers. Sensory nerve endings exist at the equatorial region, and motor nerve endings are present at the ends of the spindle. Three fiber types are present in spindles: nuclear bag 1 , nuclear bag 2, and nuclear chain fibers. Because the intrafusal fibers express myosin isoforms different from those of the surrounding musculature, there has been interest in the patterns of isoform expression by intrafusal fibers. Species differences exist in the binding of mAbs against embryonic MHC to intrafusal fibers (Maier et al., 1988a). In cat, rat, and rabbit muscle spindles, nuclear bag 1 fibers stained strongly at both polar and juxtaequatorial regions. Nuclear bag 2 fibers stained little or not at all at the poles and moderately at juxtanuclear regions. In rat and rabbit muscle spindles, nuclear chain fibers did not stain but, in the cat, nuclear chain fibers stained as strongly as bag 1 fibers. Intrafusal fibers in EO muscles of the sheep, cow, and pig were studied with antibodies specific for slow-tonic or slowtwitch myosin (Scapolo et al., 1990). In the sheep and cow, bag fibers reacted with both antibodies, but one bag fiber in the pig was labeled with antibodies to slow-tonic myosin but not with antibodies to slow-twitch myosin. In rat and rabbit tibialis anterior muscle spindles, however, nuclear bag 2 fibers and nuclear chain fibers contain two or more myosin isoforms, but nuclear bag 1 fibers contain only one (Maier and Zak, 1989).

70

MARVIN H. STROMER

In adult rat SOL, slow-tonic and neonatal MHC are expressed only in spindle fibers (Pedrosa et al., 1989). Each of the three intrafusal fiber types has a unique and variable composition of MHC isoforms and M-band proteins. Maier and Zak (1990) also found that intrafusal fibers in chicken tibialis anterior muscle spindles and extensor digitorum longus (EDL) muscle spindles contained a nonhomogenous population of MHC. Carpenter et al. (1988) found that all intrafusal fibers in adult rat skeletal muscle reacted with antibodies to chick fast-twitch MHC and with antibodies to chicken slow-tonic MHC. Extrafusal fibers in adult rat skeletal muscle reacted with antibodies to chicken fast-twitch MHC, but not with antibodies to chicken slow-tonic MHC. Carpenter et al. (1988) suggested that extrafusal fibers with similar innervation are immunologically related and that the type of innervation may be an additional way to classify avian and rat fibers. Kucera and Walro (1988) severed either the sensory or the motor nerve supply to neonatal rat SOL muscles. In the neonatal rat, spindles are present but intrafusal fibers are structurally and immunocytochemically immature. Deafferented fibers did not react with antibodies to slow-tonic MHC or with antibodies to fast-twitch MHC. This indicates that intact sensory nerves are required for MHC expression during the postnatal development of rat spindles. Kucera and Walro (1991) used an antibody to a slow-tonic MHC isoform to show that the spread of expression of this isoform from the equator of prenatal rat intrafusal fibers toward the poles of bag 1 and bag 2 fibers may be modulated by sensory and motor neurons. An a-cardiac MHC isoform has been detected with monoclonal antibodies to human a-cardiac MHC in rat nuclear bag spindle fibers (Pedrosa er al., 1990). Neonatal deefferentation showed that motor innervation influences the expression of a-cardiac MHC along the spindle fibers. PedrosaDemellof et al. (1991) have pointed out that some of the discrepancies in the results from different labs with respect to the appearance of MHC isoforms in developing muscle spindles are due to differences in antibody sensitivity and specificity. f . Myosin Isoform Changes during Muscle Development Because of species and muscle differences, comments in this section are grouped by species in the following order: avian, human, rat, mouse, and cat. The avian embryo, especially chicken, has been a popular source of developing skeletal muscle cells. The sequential appearance of three broad categories of fast MHC isoforms at certain developmental stages has led to their designation as embryonic, neonatal, and adult isoforms (Bader et al., 1982; Bandman et al., 1982; Lowey et al., 1983). It is now clear that different isozymes may simultaneously appear in the same fiber. Zhang and Shafiq (1987) found that at Embryonic Day 16 (E16) all ALD fibers reacted strongly with a mAb that recognized slow MHCs and reacted weakly with a second mAb to slow MHC and with a mAb to fast MHC.


71

In the early stages after hatching, many fibers reacted with none of these antibodies. The adult pattern of strong reaction with both monoclonals to slow MHC and weak reaction with the antibody to fast MHC was not established until at least 9 weeks after hatching. Until E8, limb bud fibers produce both a ventricular MHC and a fast skeletal MHC (Sweeney et al., 1989). Young myotubes in adult muscle had a similar MHC pattern. At E6, some limb bud fibers that later became slow fibers began to express slow MHC. Sweeney er al. (1989) suggested that muscle progenitor cells commit to a single skeletal muscle lineage and that appearance of a fibertype-specific lineage occurs after localization of cells in premuscle masses. Muscle colonies formed by cloning myoblasts from E4-6 embryo hind limbs were of three types. Most numerous were colonies producing only fast MHC. Less numerous were colonies in which all myotubes produced fast and slow classes of MHC and most rare were colonies producing only slow MHC (Miller and Stockdale, 1986a). Colonies from E10-12 myoblasts contained myotubes that produced only fast MHC. Monoclonal antibodies were used to identify the myotube types. These results, together with data obtained by using monoclonal antibodies to study both chicken and quail embryos (Miller and Stockdale, 1986b), indicated that early myoblasts in the avian embryo are committed to three lineages, independent of innervation, that form primary myotubes differing in their MHC content. In a later phase, secondary myotubes form from myotubes in a single lineage, with maturation and fiber diversity dependent on innervation. Crow and Stockdale (1986) used mAbs to light- and to heavy-chain subunits of myosin to demonstrate that the distribution of embryonic fiber types in chicken thigh muscles was in a spatial pattern that predicted future fiber composition of the muscle. Innervation during the fetal phase is required for maintaining the pattern and controlling the myosin content of the cells. An earlier study, in which postmitotic myoblasts in clones derived from El0 chicken breast muscle were stained with antibodies to M-type creatine kinase and to skeletal MHC, also concluded that myogenic precursor cells were a heterogeneous population (Quinn and Nameroff, 1983). Cerny and Bandman (1986) inhibited spontaneous contraction of cultured chick muscle cells with either potassium (12 mM) or tetrodotoxin and found that neonatal MHC disappeared from virtually all myotubes. Immunocytochemical analysis of the fast fibers of the red strip in the adult chicken pectoralis showed that they contain embryonic fast MHC (Shear et al., 1988). During normal development, these fibers only transiently express neonatal and adult MHC. After denervation, however, the adult isoform is reexpressed, which may indicate that innervation represses certain MHC isoforms in the chicken. A special type of normally quiescent myoblast , the satellite cell, exists between the basal lamina and the sarcolemma of adult muscle cells (Campion, 1984). During muscle repair or regeneration, satellite cells prolifer-

72

MARVIN H. STROMER


73

ate, fuse and differentiate into specific muscle fiber types. Feldman and Stockdale (1991) used mAbs specific for avian fast and slow MHC to show that satellite cells from adult chicken and quail muscles are committed to form at least two fiber types. One type formed only fast fibers and the second type formed mixed (fast/slow) fibers. Different proportions of these two types of satellite cells are found in fast and slow muscles. Repeated subculturing of the progeny of these satellite cells resulted in no change in MHC isoform in cells from the chicken pectoralis major, an increase in fast versus mixed fibers in cells from the chicken ALD, and a rapid change from fast to mixed fibers in cells from quail ALD. The intracellular location of myosin isoforms in developing chicken muscle has been studied with both immunofluorescence and immunoelectron microscopy. Gauthier (1990) found that there were populations of myofibrils in pectoralis muscle cells at 7 days after hatching that contained more of a specific isoform than did other myofibrils in the same cell (Fig. 1). Monoclonal antibodies and immunoelectron microscopy have been used to track epitope changes on avian MHC during development (Winkelmann et al., 1983). An epitope located 14 nm from the head-rod junction in the N-terminal fragment of MHC is present in both adult and embryonic pectoralis myosin. Another epitope near the C-terminus of the myosin rod is present in adult fast myosin but not in early developing pectoralis myosin. A third epitope that is present on a MLC near the head-rod junction exists throughout development. Taylor and Bandman (1989) isolated myosin filaments from chicken pectoral muscle at different stages of development. Filaments from muscle at E12, 10 days after hatching, and at 1 year reacted only with mAbs to embryonic, neonatal, and adult myosins, respectively. Three classes of filaments were isolated from El9 muscle. One class reacted only with antibodies to embryonic myosin, a second reacted only with antibodies to neonatal myosin, and a third was recognized by antibodies to both embryonic and neonatal myosins. Three classes of thick filaments also existed in muscle 44 days after hatching, as indicated by reaction with antibodies to neonatal or adult fast-MHC antibodies or to both. Because the neonatal MHC antibody preferentially bound to the center of heterogeneous filaments, Taylor and Bandman

FIG. 1 Chicken pectoralis muscle, 7 days after hatching. Lowicryl section showing the response to gold-labeled anti-embryonic myosin. Specificity is demonstrated by the absence of labeling in the I bands. This mAb discriminates between two types of myofibrils. In the lower right, gold particles are localized in the A bands, except for the central bare zone. In the upper left, only a few particles are present in the A band (X25.000). (Reproduced from the Journal of CcIlBiology, 1990, Vol. 110, p. 697, by copyright permission of the Rockefeller University Press and by courtesy of G . F. Gauthier.)

74

MARVIN H. STROMER

(1989) speculated that neonatal MHC may have a unique role in thickfilament assembly. Monoclonal antibodies specific for adult and for neonatal myosin isoforms have been used to show that avian MHCs form predominantly homodimeric molecules with either two neonatal or two adult MHCs (Lowey et al., 1991). The use of antibodies also has enhanced our understanding of muscle fiber development in human embryonic muscle. Moore et al. (1984) used mAbs to adult fast and slow MHC and could detect adult slow MHC in 14-week-old human embryos, much earlier than could be done with histochemical stains. Zhang et al. (1987) also found that, in human fetal muscle at 19-20 weeks gestation, mAbs could identify three fiber types, those reacting with antibodies to slow MHC, with antibodies to fast MHC, or with neither, but histochemical staining could not discriminate these fibers. Slow MHC was detected with mAbs in all primary myotubes at 9 weeks of gestation but was not present in secondary or tertiary myotubes until 29 weeks of gestation (Draeger et al., 1987). These three types of myotubes appear sequentially and could be detected immunocytochemically but not histochemically . Monoclonal antibodies were also used to show that, in human muscle at 17-20 weeks of gestation, slow MHC continued to be present only in primary myotubes (Ecob-Prince et al., 1989b). Secondary and tertiary myotubes contained neonatal MHC and different levels of fast and embryonic MHC. A fetal MHC that is expressed at a decreasing rate during fetal life was detected by a mAb (Pons et al., 1986). Adult slow MHC was detected by immunostaining in a few fibers in 14- to 16-week fetal muscle and preceded the appearance of adult fast MHC. Antibodies to slow MHC have been used to show that embryonic rat muscle primary fibers initially produce slow myosin (Kelley and Rubinstein, 1980; Harris et al., 1989a). Narusawa et al. (1987) found, however, that in the 16-day rat fetus, all primary myotubes in the anterior tibial, EDL, and SOL muscles expressed both slow and embryonic myosin. Condon et al. (1990a) agreed that primary rat myotubes contain both slow and embryonic myosin and found that the loss of slow myosin is accompanied by the expression of neonatal myosin. As muscle development continues, the expression of slow myosin is inhibited in some fibers, which then express neonatal (Condon et al., 1990a) or fast isoforms (Narusawa et al., 1987). Narusawa et al. (1987) denervated the neonatal SOL and found that the amount of slow MHC declined, but when the neonatal EDL was denervated, the content of slow MHC increased. Narusawa et al. (1987) suggested that the nerve protects and amplifies production of the predominant MHC isoform and inhibits other isoforms. Lyons et al. (1983) stained fibers from the slow SOL and fast EDL at 20 days gestation and found that all fibers reacted strongly with antibodies to adult fast

75 myosin, and some also reacted with antibodies to adult slow myosin. The axial distribution pattern of fibers that recognized the slow myosin antibody resembled the distribution of slow fibers in the adult SOL. Sartore et al. (1982) used an antibody to fetal MHC to show that fetal myosin, which has a heterogeneous fiber distribution in fetal and neonatal rat muscle, is transiently expressed in regenerating muscle. In rat SOL, half the fibers contained embryonic and slow myosin 1 week after birth, and later contained only slow myosin (Butler-Browne and Whalen, 1984). A second group contained embryonic and neonatal myosin at 1 week, and later most contained only fast myosin. A part of this second group began at 4 weeks t o acquire slow myosin. These data were interpreted to suggest that the myosin isozyme sequence is embryonic to neonatal to adult fast myosin. An individual developing fiber may be induced to produce slow myosin as synthesis of other isoforms is repressed. Frozen sections of somites and early limb buds from 10- to 12-day mouse embryos and cultured cells from the same age limb buds stain with antibodies to both slow and embryonic fast myosin (Vivarelli et al., 1988). Thus, it seems that mouse and rat embryonic myoblasts are similar with respect to MHC synthesis, but both differ from avian embryonic myoblasts, which are divided into fast, slow, and mixed. Cells from limbs of 13-day mouse embryos have reduced affinity for the antibodies to slow myosins. By E15, cultured cells from the limb express only embryonic fast myosin, but cryostat sections show that primary fibers contain both slow and embryonic fast myosin and that secondary fibers contain only embryonic fast myosin. Whalen et al. (1984) found that 10% of the fibers in mouse EDL and semimembranosus muscles stained with antibody to adult slow myosin during the first month after birth and that, in the adult, only 0 to 0.8% of the fibers recognized the slow myosin antibody. In the newborn cat, 4.8% of the fast EDL and 26% of the slow SOL primary fibers stained strongly for slow myosin and stained weakly for fetal/embryonic myosin (Hoh et al., 1988a). The secondary fibers all reacted strongly with an antibody specific for fetal/embryonic myosin. With increased maturity, secondary fibers in the fast muscle became fast fibers and those in the slow muscle became slow fibers. In the newborn kitten posterior temporalis muscle, anti-fetal myosin stained almost all fibers uniformly but some fibers stained with antibodies to superfast myosin and lightly with antibodies to slow myosin (Hoh et al., 1988b). By 50 days, slow myosin staining disappeared and superfast myosin replaced fetal myosin, so that the muscle was nearly all superfast myosin, as in the adult. The masseter muscle in the adult cat has a unique myosin isoform composition, with a majority of the fibers expressing superfast myosin and the remainder expressing slow myosin (Hoh and Hughes, 1989). Antibody staining of fetal masseter muscle suggests there are two types of primary PROTEINS IN STRIATED MUSCLE

76

MARVIN H. STROMER

and two types of secondary fibers. Slow primary fibers stain with antibodies to slow myosin and stain transiently with anti-embryonic myosin. Superfast primary fibers stain with embryonidfetal and slow myosin antibodies during the perinatal period but later stain with antibodies to superfast myosin. Superfast secondary fibers express embryonic/fetal myosins in the fetus and replace these with superfast myosins after birth. Slow secondary fibers first express embryonidfetal myosins; shortly after birth, they express slow myosin or slow and superfast myosin, and later they express only slow myosin.

g . Effects of Innervation on Myosin Isoforms The early experiments by Buller et al. (1960), which demonstrated that cross-innervation caused a transformation in contraction speed (i.e., fast muscles became slow), created interest in the effects of the nerve supply on myosin isoforms. Gauthier et al. (1983) cross-innervated the normally fast cat flexor digitorum longus (FDL) with the SOL nerve and similarly cross-innervated the normally slow SOL muscle with the FDL nerve. The FDL muscle, after about 48 weeks, was nearly completely converted to slow fibers. The cross-innervated SOL, however, continued to recognize antibodies against slow myosin but also stained with antibodies to fast myosin light chains. Strips of the superfast cat posterior temporalis muscle were transplanted in the fiber-free bed of either the fast EDL or the slow SOL muscles (Hoh and Hughes, 1988). After 210 days, transplanted muscle strips in the EDL beds reacted only with anti-superfast myosin, but those in the SOL beds reacted mostly with anti-slow myosin, and a few fibers reacted with antisuperfast myosin. The EDL and SOL muscles regenerating in their own beds expressed fetal, slow, and fast myosins only. Because the transplanted superfast strips early in regeneration recognized antibodies to fetal, slow, and superfast MHC, Hoh and Hughes (1988) concluded that the expression of adult myosin isoforms can be modulated by the nerve. Denervation by surgical or chemical methods has also been used to study isoform transformations. Reinnervation of mouse SOL muscles 1 month after sectioning the SOL nerve resulted in the presence of hybrid fibers that stained with mAbs to both type I and type IIA myosin (Desypris and Parry, 1990). Denervation or chronic paralysis of rat calf muscles at El5 or earlier caused most primary myotubes to express only embryonic and neonatal MHC by El9 but not to express slow MHC (Harris et al., 1989b). Denervation or paralysis after El5 did not cause this effect. In these rat muscles, the sequence of myosin isoforms is dependent on active innervation during a particular embryonic period. Normal 1-month-old and adult rat gastrocnemius muscle type IIB (fast glycolytic) fibers can be divided into three subgroups by their graded staining with a mAb to MHC (Leung et al., 1987). Neonatal denervation, however, caused gastrocne-

77 mius fibers to stain uniformly with the antibody. This suggests that the nerve supply is responsible for maintaining the heterogeneity of myosin isoforms in the fibers. Brachial levels of the neural tube were surgically removed at E2 in avian embryos (Phillips et a / . , 1986). At E l I , muscles that had developed aneurally were smaller, but the distribution of fast and slow myotubes was unchanged compared with normal embryos. This finding in the avian embryo seems to differ from the situation in the rat, which has an apparent requirement for innervation to produce different myosin isoforms. Curare treatment of El8 chick ALD, however, causes the expression of a second fiber type that stains with an antibody to fastmyosin LCl (Gauthier et a/., 1984). The authors speculated that the change in the motoneuron pool may have induced the expression of this second myosin isoform. Denervation or tetrodotoxin block of the sciatic nerve in adult rats cause reexpression of embryonic and neonatal MHCs in type IIA fibers (Schiaffino et al., 1988). Immunoreactivity differed along the same fiber, implying that myosin isoform expression within a fiber is coordinated by the nerve. Injection of P-bungarotoxin into rat fetuses before muscle innervation caused no changes in either the fiber types or their intramuscular distribution as detected by antibodies to myosin when compared with controls (Condon et a / . , 1990b). The ultimate fiber type was achieved in some fibers without the normal sequence of isoform changes, and some slow fibers lost their ability to express slow myosin. Six months after adult cats were spinalized at T12-Tl3, the normally slow SOL muscle showed an increase in the percentage of fibers that stained with antibodies to fast MHC or to both fast and slow MHC (Jiang et a/., 1990). Direct electrical stimulation of muscles to affect myosin isozyme transformation also has been a popular approach. Among the early studies, Rubinstein et a/. (1978) continuously stimulated fast rabbit muscles at 10 Hz and found that this caused preexisting fibers to switch from synthesis of fast myosin to synthesis of slow myosin. Stimulating the fast EDL muscle of the rabbit at 10 Hz for 12 hr/day resulted in transformation to a slow-type muscle (Maier et a/., 1986a). Examination of this muscle after 6 to 21 days stimulation showed that fast-twitch fibers were degenerating and that small fibers, often with central nuclei, that reacted with antibodies to embryonic MHC were present. Maier et al. (1986a) concluded that, in addition to the isoform switches in adult fibers, newly formed slow fibers also could contribute to the muscle transformation. Maier et al. (1988b) subsequently found that chronically stimulated fast muscle from adult rabbits contains fibers that react with antibodies to neonatal/embryonic and embryonic MHCs, two isoforms normally not present in adult muscle. Low-frequency stimulation seemingly influences some of the newly formed myotubes to become slow fibers because, as duration of stimulaPROTEINS IN STRIATED MUSCLE

70

MARVIN H. STROMER

tion increased, fewer fibers stained with the neonatallembryonic and embryonic MHC antibodies. Immunogold labeling and examination in the electron microscope were used by Franchi et al. (1990) to study stimulation-induced fiber-type transformation in rabbit muscle. After 4 weeks of stimulation, fibers of the fast tibialis anterior muscle reacted with mAbs to both slow and fast myosins. After 6 weeks, only labeling with antibodies to slow myosin was seen. After 7 weeks of stimulation and 3 weeks of recovery, fibers again reacted with both antibodies. The distribution of gold particles indicated that newly synthesized myosin is incorporated throughout the length and cross section of the A-band. The opposite fibertype transformation was done by Gorza et al. (1988a), who stimulated rat adult slow SOL muscles with fast (100 Hz) stimuli. All stimulated fibers labeled strongly with antibodies to fast myosin, and at least 90% were also weakly labeled with antibodies to slow myosin. The labeling with antibodies to fast myosin could first be detected 7 days after stimulation was begun. No fibers reacted with antibodies to fetal myosin, indicating that the slow-to-fast transformation occurred in preexisting adult fibers in the rat SOL. Coculturing muscle cells and nerves is another approach to studying the effect of nerves on myosin isoform expression. Culturing section of embryonic mouse spinal cord with explanted adult mouse muscle fibers results in the degeneration of the adult fibers and the proliferation of satellite cells to form new myotubes (Ecob et al. 1983). Some of these myotubes become innervated by the embryonic mouse spinal cord neurons and contain, after 3 to 5 weeks, adult fast MHC. Ecob-Prince et al. (1986) subsequently found that fast MHC was absent both in new rat myotubes in cocultures and in older myotubes cultured without spinal cord tissue. The blockage of nerve-induced contractions with a-bungarotoxin did not decrease the number of fibers in cocultures that expressed fast MHC. Thus, it seems that the synthesis of adult fast MHC is dependent on the presence of nerves but not on the nerve-induced contractile activity.

h. OtherExternalInJluences on Myosin Isoforms The effects of a variety of external factors, principally on myosin but on other myofibrillar proteins as well, have been investigated with immunocytochemistry. Stretch hypertrophy of the chicken ALD muscle caused 28-52% of the fibers to stain with antibodies to fast MHC by Day 12 to 19 of stretch compared with

International Review of Cytology

International Review of Cytology

International Review of Cytology