contraction and organization of collagen gels by … · 2005. 8. 21. · collagen gels plated in 60...

J. Cell Sci. 50, 299-314 (1981) 299Printed in Great Britain © Company of Biologists Limited 1 g8i

CONTRACTION AND ORGANIZATION OF

COLLAGEN GELS BY CELLS CULTURED FROM

PERIODONTAL LIGAMENT, GINGIVA AND

BONE SUGGEST FUNCTIONAL DIFFERENCES

BETWEEN CELL TYPES

C. G. BELLOWS, A. H. MELCHER AND J. E. AUBINMedical Research Council Group in Periodontal Physiology, Faculty of Dentistry,University of Toronto, Toronto, Ontario, Canada

SUMMARY

Monkey periodontal ligament fibroblasts (MPLF cells), human gingival fibroblasts (HGFcells), rat embryonic calvaria cells (REC cells), porcine periodontal ligament epithelial cells(PPLE cells) and rat osteosarcoma 17/2 cells (ROS cells) were incorporated into 3-dimensionalcollagen gels plated in 60 mm Petri dishes in order: first, to measure the capacity of these celltypes to contract; second, to investigate cell-collagen and intercellular relationships duringcontraction; and third, to define the cellular contribution to tissue contraction in an in vitrosystem. Measurements at times up to 72 h on 3 ml gels containing 5x10 ' cells and with acollagen concentration of r jomg/ml showed that MPLF cells contracted the gels at a sig-nificantly greater rate (P < o-ooi) than did the other cell types. In addition, contractionstarted sooner and was of greater extent than with the other cells. HGF cells contracted thegels more rapidly than REC and PPLE cells, while ROS cells caused no contraction. Severalstages of gel compaction could be defined: (1) the attachment of cells to collagen; (2) cellularspreading within the collagen fibre matrix; (3) organization and alignment of collagen fibresby cell processes; (4) cell migration; (5) establishment of intercellular contacts; and (6) thedevelopment of a cellular reticular arrangement within the gel and the extension of this arrange-ment into a 3-dimensional, tissue-like, honeycomb network. Electron microscopic observationson o-i ml gels containing MPLF cells showed that, in the early contractile phase, numerouscell processes attached to or enclosed collagen fibrils. These processes contained micro-filamentous material and few organelles. In compacted gels, the cells contained an increasedamount of distended rough endoplasmic reticulum and Golgi membranes. Since MPLF cellshave the capacity for vigorous contraction of the collagen gels and since they develop a reticular,3-dimensional structure in compacted gels that is reminiscent of the relationship of periodontalligament fibroblasts to collagen fibres in vivo, it is suggested that they could provide the majorforce necessary for tooth eruption in vivo. This system also provides a well-defined in vitromodel to study the sequential stages that occur during contraction processes.

INTRODUCTION

Contraction of cells is believed to play a role in a number of biological processes;for example, contraction of wounds, retraction of blood clots and eruption of teeth.Contraction during repair of skin wounds results in a scar that occupies a considerablysmaller area than the original wound (James, 1964). This contraction is believed to be

Address for correspondence: Dr C. G. Bellows, MRC Group in Periodontal Physiology,4384 Medical Sciences Building, University of Toronto Toronto, Ontario M5S 1A8, Canada.

300 C. G. Bellows, A. H. Melcher andj. E. Aubin

mediated by fibroblasts within the newly formed granulation tissue (Abercrombie,Flint & James, 1956) or perhaps by cells designated as myofibroblasts (Gabbiani etal.1972; Gabbiani, Majno & Ryan, 1973). Contraction also occurs when platelets areactivated in the presence of fibrin, leading to retraction of the blood clot (Cohen,1979). Microfilamentous material and microfilament bundles containing actin appearto be ubiquitous in fibroblasts and many other non-muscle cells (Abercrombie,Heaysman & Pegrum, 1971; Lazarides & Weber, 1974; Pollard & Weihing, 1974).Microfilaments have been associated with many aspects of cell behaviour, includingcell contraction, migration and phagocytosis (Abercrombie etal. 1971; for reviews seeHuxley, 1973; Pollard, 1975; and Korn, 1978). Observations on the ultrastructuralcharacteristics of fibroblasts in the periodontal ligament have led to the postulate thattheir putative motility and contractile properties could provide the major force fortooth eruption (Beertsen, Everts & van den Hooff, 1974; Beertsen, 1975; Melcher &Beertsen, 1977). However, Shore & Berkovitz (1979) have challenged this view.

Collagen in the form of films, rafts or gels has been used as a natural biologicalsubstrate for cell growth and differentiation (Ehrmann & Gey, 1956; Schor & Court,1979). Human foreskin fibroblasts have been shown recently to contract 3-dimensionalcollagen gels established in Petri dishes to which cells do not attach (Bell, Ivarsson &Merrill, 1979). It occurred to us that this model could provide a useful in vitro systemin which to measure the capacity of different cell types to contract, to investigatecell-collagen and intercellular relationships during contraction and to clarify thecellular contribution to tissue contraction. We have found that fibroblast-like cellsfrom the periodontal ligament contract collagen gels at a significantly greater rate thando 4 other cell types, suggesting that they exhibit a strong capacity to contract in vitro.Further, these cells establish numerous intercellular contacts and form a characteristic3-dimensional network within the compacted gel, which is reminiscent of theirarrangement in vivo. These results suggest that such fibroblast-like cells in theperiodontal ligament could play a role in tooth eruption.

MATERIALS AND METHODS

Cells and cell culture

Cells from 5 different sources were compared in this study. (1) Monkey periodontal ligamentfibroblasts (MPLF cells), isolated by the explant procedure of Marmary, Brunette & Heersche(1976), were used from subcultures 3 to 15. (2) Human gingivalfibroblasts(HGFcells),obtainedfrom the American Type Culture Collection (Rockville, Md, U.S.A.), were used at sub-cultures s and 6. (3) Rat embryonic calvaria cells (REC cells) were isolated from term embryonicrat calvaria according to the procedure of Rao, Ng, Brunette & Heersche (1977). Briefly, cellsuspensions from the dissected calvaria were collected by 5 sequential incubations withcollagenase, each lasting 10 or 20 min. Cells from the final 2 collagenase digests were pooled toyield a population of osteoblast-like cells, and these were used from subcultures 1 to 7. (4) Epi-thelial cells from the rests of Malassez of porcine periodontal ligament (PPLE cells) wereobtained by either the explant or the single cell suspension procedures described by Brunette,Melcher & Moe (1976). Cells between subcultures 1 and 4 were used. (5) Rat osteosarcoma17/2 cells (ROS cells) were cloned from a continuous cell line and were a gift from Dr G. Rodan,School of Dental Medicine, University of Connecticut, Farmington, Conn.

All cells were grown and maintained in a. Minimal Essential Medium (aMEM) containing

Differences in collagen gel contraction 301

100 fig/ml penicillin G (Sigma Chemical Co., St Louis, Mo.), 50 /Jg/ml Gentamicin sulphate(Sigma) and 0-3/Jg/ml Fungizone (Flow Laboratories; McLean, Va) plus 15% foetal calfserum (FCS). The cultures were incubated at 37 °C in a humidified atmosphere of 95 % airand s % CO,. Subculturing, and collection of cells for incorporation into gels, was performedwith o-oi % trypsin in citrate/saline.

Preparation of collagen gels

Stock collagen solutions comprising the following were prepared on ice in prechilled sterilebottles: (1) 0-3 ml of either 10 x cone. McCoy's 5a medium or 10 x cone. aMEM; (2) 0-3 ml0-26 M-NaHCO, buffer; (3) 0-3 ml FCS; (4) 0-3 ml of 10 x cone, antibiotic mixture in aMEM;(5) o-i2 ml o-i M-NaOH; (6) 1-2 ml Vitrogen 100 (Collagen Corp., Palo Alto, Calif.); and (7)0-48 ml of a cell suspension (1-05 x io'cells/ml in aMEM + 15% FCS). The first 5 con-stituents were mixed well by shaking before the addition of the Vitrogen 100. Thereafter, thesolution was mixed with a pipette before and after the addition of the cell suspension to givea final collagen protein concentration of i'2o mg/ml and a final cell concentration of approx.1-66 x io6 cells/ml (5 x io6 cells/60 mm dish). Three-ml aliquots of the cold stock solutionwere pipetted into 60 mm Petri dishes kept at room temperature (Falcon 1007; Falcon Plastics;Oxnard, Calif.). Cells do not attach to the floor of these dishes. The plated collagen solutionswere immediately incubated at 37 °C in a humidified atmosphere of 95 % air and 5 % CO,and were found to gel after 5-10 min. In experiments to determine the effects of gel volumeupon the rate of contraction, 1, 2, 3, 4 and 5 ml of the collagen solution were pipetted into andspread over the floor of 60 mm Petri dishes. In some experiments, 1 ml gels were not spreadand therefore covered only a portion of the floor of the dish. Two forms of control cultureswere used. (1) Collagen solutions without cells but containing an equivalent amount of aMEMplus 15% FCS and antibiotics, were pipetted into 60 mm Petri dishes. (2) Cell-containingcollagen solutions were pipetted into 60 mm tissue culture dishes (Falcon 3002) to which boththe gel and the cells adhere.

Measurement of contraction

The diameter of the collagen gels was measured after 6, 12 and 24 h, 2 days and 3 days, bythe method of Bell et al. (1979). This was done by placing the Petri dishes containing collagengels upon a transparent metric ruler against a dark background. The longest and the shortestaxes of the gel were measured and the mean of the 2 measurements was recorded.

Observations in vitro

Events during contraction of the gels were observed using a Leitz Diavert phase-contrastmicroscope. Selected contracting gels were filmed by time-lapse photomicrography, at a speedof 1 frame per min for periods up to 24 h after plating. To do this an electronically controlledBolex H-16 movie camera (Hommel Electronics, Toronto, Canada) mounted on a ReichertBiovert microscope was located in a warm room maintained at 37 °C, and the culture disheswere contained in a chamber through which was passed a constant flow of air/5 % CO,. Thecamera was loaded with Kodak Plus X reversal film no. 7276.

Histology

Some of the collagen gels were fixed for 24 h in 10 % neutral buffered formalin after 1, 6,12 and 24 h, and 2, 4 and 9 days of contraction. Portions of these gels were dehydrated, clearedand embedded in paraffin and sectioned either parallel to or perpendicular to the floor of theculture dish at a setting of either 5, 10 or 20 fim. Sections were stained with either Haematoxylinand Eosin or van Gieson's picro-fuchsin counterstained with Celestine blue-Haematoxylin(Drury & Wallington, 1967).

302 C. G. Bellows, A. H. Melcher and J. E. Aubin

Electron microscopy

For electron microscopic observation, o-i-ml aliquots of the collagen solution containingeither MPLF cells or REC cells were pipetted into Petri dishes and incubated as describedpreviously. These small contracting gels were measured and collected after i, 6, 12 and 24 h,fixed for 2 h at 4 °C in 2 % glutaraldehyde in o-i M-cacodylate buffer, pH 7-3, and post-fixedin 1 % osmium tetroxide in o-i M-cacodylate buffer for the same time. The collagen gels werethen trimmed,' dehydrated through an ascending series of ethanols, embedded in Epon-Araldite (Voelz & Dworkin, 1962), and sectioned on a Reichert-Jung Ultracut microtomeparallel to the floor of the culture dish at either 1-2 fim or, to exhibit a silver interferencecolour. The thick sections for viewing by light microscopy were mounted on glass slides andstained with Toluidine blue. The thin sections were stained with 2 % uranyl acetate in 100 %ethanol and 1-5 % lead citrate (Venable & Coggeshall, 1965) and viewed in a Philips EM 300electron microscope.

RESULTS AND OBSERVATIONS

During contraction in Petri dishes, the translucent collagen gel changed to anopaque white tissue-like structure, which floated beneath the surface of exudedmedium. This tissue-like structure was sufficiently firm to be handled easily. Gelsprepared from the same stock solution and containing the same cell populationexhibited very similar rates of contraction. Slight differences were noted, however,between different experiments using the same cell type and collagen concentration.Cell-containing gels established in tissue culture dishes and maintained up to 7 daysshowed no change in diameter; however, their thickness appeared to be reduced tothat of a thick film. No measurements of changes in gel thickness were made.

Rates of contraction

The 2 different culture media used in gel preparation, McCoy's 5 a medium andaMEM, had no effect upon the rates of contraction. However, the rate at which thegels contracted was influenced by the cells that they contained (Table 1). MPLF cellscontracted the collagen gel at a significantly more rapid rate than any of the other cellpopulations (P < o-ooi at all times). HGF cells contracted the gel significantly morerapidly than REC cells up to 48 h (P < o-ooi between 6 and 24 h, and


Table 1. The capacity of different cell populations to contract collagen gels*

Time(h)

0

61 2

244872

Controlgels

53-o53'0S3'o53-o53-o53-o

MPLF

53O457 ±2-238-7 ±4-9I3-6±O7u-5±o-8ICVO± !•!

Cell population

HGF

5 3 052-3 ± 1 - 248-2! 1122-8 ±1-321-810-720-710-7

REC

S34©

53-oS2-6± 1-242-3 ±4-629-4 ±7-1177 ± 2 4

PPLE

53O53-oS3-o5 3 047-2 ± 1 - 447-2 ± 1 - 4

ROS

5 3 053'05 3 °53-o53-o5 3 0

• Each value represents the mean in mm ± s.D. of measurements from 4 to 7 gels in each of2 separate experiments. Approximately 5x10* cells in 3-0 ml collagen gels (protein concn =1 -2 mg/ml) were plated in 60 mm plastic Petri dishes. The inner diameter of the dish was53-0 mm. See the text for statistical significance of the data.

Table 2. Effect of volume upon the contraction rate of collagen gels by MPLF cells*

Volume of gel in Petri dish (ml)

Time (h)

o6

122448

• Each value represents the mean diameter!s.D. of 3 gels at each volume. The cell densitywas approx. 1 -66 x 10s cells/ml and the protein concn was 1-20 mg/ml. See the text forstatistical significance of the data.

Table 3. Contraction rates of o-i ml collagen gels by MPLF and REC cells

Cell population

I

53-o53O45-8 ±o-840-0 ± 1-4357 ±o-i

2

53'050-7 ± 1 740-5 ± I - I2 i - s ±4-315-5 ± I - I

3

53-o477 ±o-233-2! i-615-810-613-8 ± 0 6

4

53-o45-5 ± 027-010-I5'5±o-12-7 ±o-

885

5

53-o46-210-626-310-5i53±o-212-510-4

Time (h) MPLF REC

1 7'9lo-7* 7-210-56 4'3 ±O'5 7-2 ±0-5

12 2-810-3 6-210-524 2-110-3 2-510-4

• Mean diameter of 18 gels 1 s.D.

C. G. Bellows, A. H. Melcher andj. E. Aubin

§

t


a more rapid rate than their counterparts in 53 mm gels. Gels with the smaller andthe larger volume had the same ratio between cells and collagen protein concentration(Table 3). After 6 h, MPLF cells in o-i ml gels had contracted the gel by 45-5%while REC cells had caused no contraction. During this time period, MPLF cells hadcontracted the 3 ml, 53 mm gels by only 13-8%. After 12 h, MPLF cells had con-tracted 01 ml gels by 64-5 % and REC cells had contracted these gels by 8-6%. Theequivalent values for 3 ml, 53 mm gels were 45-9 % and o-8 %, respectively. By 24 h,MPLF cells had contracted o-i ml gels by 73*4% and REC cells had contracted themby 65*3% while MPLF cells had contracted the 53 mm gels by 8I - I % and RECcells by 22-1%.

Cellular events in gel contraction

ROS cells appeared to attach very poorly to the collagen fibres as judged by thefact that they did not spread within the collagen lattice. They did not extend cyto-plasmic processes, align collagen fibres, or contract the collagen gel.

In contrast to ROS cells, PPLE cells appeared to attach well to collagen fibres,since collagen fibres were seen to extend radially from the circumference of the cellsafter gelation (Figs. 1, 2). These cells adopted an almost spherical morphology,existed mostly as single cells throughout the gel and did not produce long cytoplasmicextensions. The attachment of cells to the collagen fibres was probably facilitatedthrough extensive surface blebbing (Fig. 3). When 2 adjacent cells attached to a

Fig. 1. PPLE cells; 6 h in gel. Collagen fibres surround the circumference of the cellsin a radial fashion (arrows) while other fibres are aligned between cells and clumps ofcells and appear to be under tension (arrowheads). The presence of the relatively thick,translucent gel and the resulting depth of focus precluded obtaining crisp phase-contrast photomicrographs of these cultures. Phase-contrast photomicrograph of a2-3 mm thick gel in situ, x 300.Fig. 2. PPLE cells; 12 h. Collagen fibres are attached to, and oriented between PPLEcells (arrow) but no intercellular contact between cell processes or gel compactionwas observed. Van Gieson-stained section. X455.Fig. 3. PPLE cells; 6 h. Surface blebs (arrows) may be involved in the attachmentof cells to collagen fibres. Van Gieson-stained section, x 455.Fig. 4. REC cells; 1 h. Cells are evenly distributed in single-cell suspension throughoutthe gel and have begun to spread within the gel. Several cells have short processes(arrowheads). A small number of collagen fibres are aligned (arrows). Phase-contrastphotomicrograph of a 2-3 mm thick gel in situ, x 300.Fig. 5. REC cells; 5 h. Many collagen fibres have been aligned between cells (arrows).Cell processes in different focal planes within the 3-dimensional gel are extendedalong the oriented collagen fibres (arrowheads). The cell processes appear darker andare more refractile than aligned collagen fibres. Phase-contrast photomicrograph ofa 2-3 mm thick gel in situ, x 300.Fig. 6. REC cells; 4 days; i-o ml gel. Cell-collagen relationships can be seen moreclearly in these thin gels than in 3 ml gels. Although collagen alignment is extensive(arrows), few intercellular contacts can be identified and contraction proceeded at aslow rate. Spread cells are indicated by arrowheads. Phase-contrast photomicrographof a thin collagen gel in situ, x 300.

C. G. Bellows, A. H. Mekher andj. E. Aubin


collagen fibre bundle, the cells appeared to exert tension upon it as judged by the lossof the normal wavy appearance of the collagen, and the observed orientation of thefibres between the cells (Figs. 1, 2). As many fibres became aligned between adjacentcells, a network composed of cells and collagen fibres was established (Fig. 2). Theestablishment of these compact collagen bundles probably resulted in a small amountof the liquid phase of the gel being squeezed out.

REC cells existed in single-cell suspension throughout the collagen gel. Many cellspossessed short cytoplasmic extensions that were attached to collagen fibres, and thisresulted in some alignment of these fibres between cells after only 1 h (Fig. 4). Thecells then oriented themselves along the axis represented by the oriented collagenfibres. Slender cell processes, which increased in length with time, became orientedalong the aligned collagen fibres (Fig. 5). This established a lattice network consistingof oriented cell processes and bundles of collagen fibres apparently under tensionthroughout the gel (Fig. 5). This network, which was easily seen after 5 h (Fig. 5),was established prior to measurable contraction (Table 1). The cell processes eventu-ally extended along these oriented collagen fibres to establish intercellular contactsbetween adjacent cells. Periodic observations on the contracting gels suggested thatthe formation of intercellular contacts between REC cells resulted from extension andelongation of cytoplasmic processes along collagen fibre bundles under tension ratherthan by active migration by the cells. Intercellular contacts developed only in regionswhere cells were close enough for the cytoplasmic extensions to bridge the spacebetween them. Few intercellular contacts were seen in 10 ml gels that were spreadto cover the bottom of the Petri dish, probably because of the relatively large distancesbetween the cells (Fig. 6).

Fig. 7. MPLF cells; 3 h. Cells (c) are either spheroid or exhibit slight orientation andsome collagen fibres are aligned (arrows). Intercellular contact via cell processes isapparent (arrowheads). Note that the cell processes appear to be darker and morerefractile than the collagen. Phase-contrast photomicrograph of a 2-3 mm thick gelin situ, x 300.

Fig. 8. MPLF cells; 6 h. The developing reticular or honeycomb arrangement, formedby oriented cells and intercellular contacts, is established during the first 6 h aftergelation. Phase-contrast photomicrograph of a 2-3 mm thick gel in situ, x 75.Fig. 9. MPLF cells; 12 h. Many of the rectangular, honeycomb-like structures havebecome compacted centripetally to form a dense cellular peripheral ring of cells(between arrows). Several cells are indicated by arrowheads. The collagen outsidethis ring of cells is oriented perpendicular to the cells (crossed arrow) and appearsto be attached to cells within the ring. Phase-contrast photomicrograph of a 2-3 mmthick gel in situ, x 300.Fig. 10. MPLF cells; 10 h. The honeycomb units are wholly surrounded by orientedcells or their processes resulting in a reticular-like structure. Spherical-appearingcells (arrows) are oriented perpendicular to the base of the honeycombs. Collagenfibres within these honeycombs have been condensed. Phase-contrast photomicro-graph of a 2-3 mm thick gel in situ, x 300.Fig. 11. MPLF cells; 9 days. Two chains of cells (arrows) in a section of a compactedgel that has been cut at right angles to the floor of the dish. These chains form a partof a 3-dimensional organized network. Haematoxylin and eosin. x 187.


Following gelation, MPLF cells were suspended singly throughout the gel andwere more spread than REC cells (Fig. 7). Although the cells appeared to be attachedto the collagen fibres, the alignment of the fibres into well-defined bundles betweencell processes was not as evident as in the case of REC cells (Fig. 7). The cellspossessed numerous short processes, many of which were connected to similar pro-cesses from adjacent cells (Fig. 7). Many cells also possessed long cytoplasmic exten-sions similar to REC cells. Observations on time-lapse sequences showed that when2 cells in close proximity were connected to each other by collagen fibres, each cellextended and withdrew short cell processes thereby pulling on the collagen fibres.This suggests that the collagen fibres were under tension and that the tension wascreated by the action of the cells.

The elongation of MPLF cell processes along oriented collagen fibres was muchless extensive than was observed to occur with REC cells; however, there appearedto be much greater migration of cells through the gel. Time-lapse films showed cellmigration to occur by means of extension and withdrawal of many cell processes,which frequently established contacts of either temporary or permanent nature withprocesses of other cells or cell bodies. During the pre-contractile and early contractilestage (Fig. 8), cell migration, reorientation of cell bodies and elongation of cell pro-cesses led to the formation of a reticular or honeycomb network of cells and alignedcollagen within the gel. Two methods by which the honeycomb arrangement wascompleted were observed in time-lapse films. In some cases, cells formed a smallaggregation at one pole of a bundle of collagen fibres. One of the cells in the aggregatethen established contact with a cell at the opposite pole of the collagen bundle, afterwhich it migrated to a position between the 2 poles and assumed an elongated, spindle-shaped morphology. In other instances, a cell made contact with an oriented collagenfibre bundle or a long cytoplasmic extension and, having oriented itself parallel to thissubstrate, established contacts with cells located at each of the poles.

All populations examined except ROS cells developed a compact ring of cells at theperiphery of the gel during the very early contractile phase. This peripheral ring wasmost prominent in gels of 1 or 2 ml. In 1 ml unspread gels where only a portion ofthe gel was in contact with the wall of the Petri dish, the ring developed only in thisregion of contact but gel contraction was not affected. Because of the greater numberof cells present in the ring area, the ring may have represented a more advanced stageof the contractile process that was occurring centrally within the gel. The peripheralring resulted from the honeycomb pattern becoming modified to form a series ofrectangular chains with most of the cells oriented parallel to the periphery of thePetri dish. A9 contraction proceeded, these rectangles became more condensed untilthe cells eventually lay parallel to one another (Fig. 9). Collagen fibres peripheral tothis ring were oriented perpendicularly, in a radial fashion and many were attachedto cells within the ring (Fig. 9). A small number of elongated cells were situatedperipheral to the ring. Migration of these cells is suggested by the observation thatcollagen fibres were attached and compacted perpendicular to the orientation of thecell body.

The borders of individual honeycombs were better delineated, and the collagen

Differences in collagen gel contraction

Fig. 12. MPLF cells; 12 h. Collagen fibrils are surrounded by processes of one ormore cells (crossed arrows) or are embedded in plasma membrane-bound recesses inthe cell surface (arrow). The cell processes contain few organelles but much micro-filamentous material (arrowheads), x 24700.Fig. 13. MPLF cells; 12 h. Portions of 2 adjacent cells. Cell a contains rough ER andmany mitochondria while cell b contains more polysomes and microfilamentousmaterial. Collagen fibrils in cross-section are either compacted between their plasmamembranes (arrows) or surrounded by cell processes (arrowheads), x 11 000.

310 C.G. Bellows, A. H. Melcher andj. E. Aubin

fibres within these honeycombs were progressively more aligned and compacted at10 h (Fig. 10). Many of the honeycombs decreased in size and cell aggregates becamemore numerous during the contractile phase. All cells attached to collagen appearedto become involved in the formation of the honeycombs. The borders of the honey-combs were formed by oriented cells or cell aggregates, which made contact withadjacent cells or with spherical-appearing cells. When the gel was sectioned at rightangles to the floor of the dish, the spherical-appearing cells (Fig. 10) were seen to beelongated and oriented perpendicular to the floor of the dish. These cells formed longoriented chains running throughout the thickness of the gel (Fig. n ) . Thus, the cellsformed a 3-dimensional network in the collagen.

Electron microscopic observations of gels containing MPLF cells

In the pre-contractile phase, 1 h after the gel was poured, the cells remained largelyspherical in shape with few intercellular contacts evident. After 6 h gelation and inless-advanced regions of 12 h gels, collagen fibrils lay either adjacent to, or wereattached to, the plasma membranes of the cells. Some fibrils were surrounded by thecell processes of one or more cells or were embedded in plasma membrane-boundrecesses in the cell surface (Fig. 12). The most characteristic feature of cells duringthis actively contracting phase of the gels was the large number of cell processes orcytoplasmic extensions that were nearly devoid of organelles and were packed withmicrofilamentous material (Fig. 12). Narrow contacts were present between theseprocesses. After most of the gel contraction had occurred (in 24 h, and portions of12 h gels), the cells were packed together and many of them had assumed a morerounded shape with areas of intercellular contact over large regions of the cell mem-brane. Rough endoplasmic reticulum and Golgi membranes were often prominent incells in these fully compacted gels in contrast to cells in actively contracting gels(Fig. 13). Some cells in the compacted gels also contained many polysomes, whichwere concentrated in localized areas of cytoplasm within the cell. Many of these cellsalso exhibited aligned microfilament bundles parallel to the cell periphery. Collagenfibrils were frequently compacted either singly or in small bundles between theplasma membranes of adjacent cell membranes or were surrounded by small cellprocesses (Fig. 13). Several cells in 12 h gels contained collagen profiles which, inmany cases, appeared to be intracellular.

DISCUSSION

The collagen gels provided a 3-dimensional lattice to which the cells embeddedwithin them could attach and spread, and within which they themselves could form a3-dimensional meshwork. Contraction of the gels resulted in compaction of the solidphase, that is the collagen fibres, and this was accompanied by separation of the liquidphase and a decrease in the diameter and thickness of the gel. As gels lacking cellsdid not get smaller, it is clear that the cells were responsible for the changes in dimen-sion. Cell-containing gels poured into tissue culture dishes to which both cells andcollagen adhere showed no change in diameter but did appear to decrease in thickness.


A substantial tensile force can be exerted by fibroblasts cultured in collagen gels. Thisis exemplified by the fact that when gels are attached at 4 equidistant points to theperiphery of a Petri dish, they do not compact radially. Instead, they form a squarebetween the attachment points, and the cells become oriented between pairs of thesepoints (Bell et al. 1979). The alignment of collagen fibres oriented between cells asseen in the present study is reminiscent of the sheets of cells that become alignedunder tension between pairs of dentine particles when the 2 are cultured togetherin vitro (Bellows, Melcher & Brunette, 1980).

The process of gel contraction seen here can be divided into several consecutivesteps: (1) the attachment of cells to collagen; (2) the spreading of cells within thecollagen fibre matrix; (3) the organization and alignment of collagen fibres by cellprocesses; (4) cell migration; (5) the establishment of intercellular contacts either bycell migration or by extension of cell processes along aligned collagen fibres; (6) theorganization of the cells into a network exhibiting a 3-dimensional tissue-like honey-comb arrangement. As judged by cellular morphology and the ability of the cells tospread, all cell types, with the exception of ROS cells, attached to the collagen fibres.The reason for the lack of attachment of ROS cells is not known, although it mayreflect cell-surface changes concomitant with transformation (Hynes, 1976; Yamada,Olden & Pastan, 1978).

Expression of the other activities believed to contribute to compaction of the geldiffered between the different cell types. The slight contraction of gels by PPLE cellscan be attributed to the alignment and compaction of collagen fibres by attachmentof these fibres to surface blebs, since protrusion of cell processes essential for theestablishment of intercellular contacts was not apparent. The epithelial cells did notform tissue-like structures within the gel. REC cells efficiently aligned collagen fibresand established a reticular-like cell network chiefly through the elongation of cellprocesses along collagen fibres during the pre-contractile phase. However, after con-tration the MPLF cells exhibited a more organized arrangement than did the RECcells. Furthermore, while MPLF cells aligned fewer collagen fibres into bundles thandid REC cells, they were quicker to establish intercellular contacts and a reticular-likeappearance. HGF cells behaved similarly to MPLF cells, but the honeycomb networkwas less well developed and intercellular contacts formed more slowly.

The extent of cellular organization and the number and rate at which intercellularcontacts are formed may be directly related to the efficiency of contraction in acollagen gel. This interpretation is consistent with a number of observations. Theonset and the rate of contraction of collagen gels containing MPLF cells was signi-ficantly greater than that of gels containing any of the other cell populations tested.Since MPLF cells also contracted the gel to a smaller diameter than did any of theother cell populations, these cells appeared to have generated a greater contractileforce. HGF cells began contraction ahead of REC cells and contracted gels at a morerapid rate during the earlier time periods. However, REC cells had contracted thegels to a smaller diameter than HGF cells after 72 h. This is consistent with theobservation that REC cells migrated and formed intercellular contacts at a muchslower rate than did either MPLF or HGF cc-lls.

312 C.G. Bellows, A. H. Mekher andj. E. Aubin

The model system used here may help to clarify the mechanisms and the specificrole of certain cells in various contractile phenomena. The contraction of collagen gelsis similar in many ways to the process that occurs during contraction of connectivetissues during healing of wounds. Wound contraction is mediated by the fibroblastswithin the wound (Abercrombie et al. 1956) and is associated with an absolute reduc-tion in the wet weight of the wound content (James, 1964). The peripheral ringsurrounding contracting gels may be analogous to the 'picture frame' or specializedzone of cells at the wound margin that has been shown not to play an active role inwound contraction (Abercrombie, James & Newcombe, i960). This peripheral ringis not essential for gel contraction in this model system either, since it did not developin contracting gels whose periphery did not make contact with the sides of the Petridish. The finding that cell-containing gels plated in dishes to which the cells andcollagen fibres can attach are not reduced in diameter could explain the virtual absenceof contraction and scaring from some wounds; for example, those in attached gingivawhere the lamina propria is firmly attached to the periosteum of the underlying bone(Melcher, 1969).

The cells that are responsible for the contractile phenomenon in wounds arebelieved by some investigators to be a specialized type of fibroblast, to which the name'myofibroblast' has been applied (Majno et al. 1971; Gabbiani et al. 1972). Thesemyofibroblasts have: (1) nuclei possessing many folds and indentations indicative ofa contracted state and similar to those seen in contracted smooth muscle cells; (2)cytoplasm containing large numbers of microfilament bundles; (3) cell surfacematerial similar to basal lamina; and (4) intercellular contacts described as desmo-somes (Gabbiani et al. 1973). The presence of gap junctions, which may function tosynchronize contraction between myofibroblasts, has also been described (Gabbiani,Chaponnier & Huttner, 1978). We have found that the ultrastructural morphology ofMPLF cells in contracted collagen gels is different from that exhibited in the earlycontractile phase, and the details of these observations will be published elsewhere(C. G. Bellows, A. H. Melcher, U. Bhargava & J. E. Aubin, unpublished). In theearly contractile stage, the cells possessed large processes packed with microfilamen-tous material but containing few organelles, while after gel contraction the cells weremore rounded and contained extensive rough endoplasmic reticulum and Golgimembranes. While many cell nuclei exhibited indentations, structures resemblingbasal lamina and desmosomes were not observed. MPLF cells of the type used in thisexperiment have been shown previously to exhibit normal characteristics of fibro-blasts, for example, ultrastructure (Brunette et al. 1977) and collagen production(Limeback & Sodek, 1979). The observations made in our study, therefore, suggestthat MPLF have the capacity to contract vigorously and that their cytological appear-ance may be different during different stages of the process. This casts doubt on theneed to ascribe to a specific cell type (the myofibroblast) the role of contraction in thegranulation tissue of a wound, and suggests that a phenotypic modulation of fibro-blasts to cells possessing the required contractile machinery can occur when needed.

It has been suggested that the motility of periodontal ligament fibroblasts providesthe major contractile force for tooth eruption (Beertsen et al. 1974; Beertsen, 1975).


Melcher & Beertsen (1977) have suggested that the presence of microtubules andmicrofilaments in cells indicates a structural basis for a motile system whereby exten-sion of cell processes and attachment to collagen fibrils or other cells, and the sub-sequent withdrawal of these processes, could draw the collagen fibrils or cells closertogether. This could result in shortening of the connective tissue joining bone andtooth and, because of the alignment of the fibres, movement of the tooth in an occlusaldirection. The present experiment has demonstrated that MPLF cells are capable ofvigorous contraction, the onset, rate and extent of which exceeds that of HGF cellsand REC cells, and that they can attach to and pull on collagen fibres. Although thisdoes not provide direct evidence that periodontal ligament fibroblasts are responsiblefor tooth eruption, it shows for the first time that they have the capacity to do in vitrowhat it has been postulated that they do during tooth eruption in vivo. Furthermore,it is striking that the 3-dimensional arrangement adopted by the MPLF cells incollagen gels in the present experiment is reminiscent of their arrangement relative tothe collagen fibres of the periodontal ligament in vivo (Melcher, 1980).

REFERENCES

ABERCROMBIE, M., FLINT, M. H. & JAMES, D. W. (1956). Wound contraction in relation tocollagen formation in scorbutic guinea-pigs. J. Embryol. exp. Morph. 4, 167-175.

ABERCROMBIE, M., HEAYSMAN, J. E. M. & PEGRUM, S. M. (1971). The locomotion of fibro-blasts in culture. IV. Electron microscopy of the leading lamella. Expl Cell Res. 67, 359-367.

ABERCROMBIE, M., JAMES, D. W. & NEWCOMBE, J. F. (i960). Wound contraction in rabbit skin,studied by splinting the wound margins. ,7. Anat. Land. 94, 170-182.

BEERTSEN, W. (1975). Migration of fibroblasts in the periodontal ligament of the mouse incisoras revealed by autoradiography. Archs oral Biol. 20, 659-666.

BEERTSEN, W., EVERTS, V. & VAN DEN HOOFF, A. (1974). Fine structure of fibroblasts in theperiodontal ligament of the rat incisor and their possible role in tooth eruption. Archs oralBiol. 19, 1087-1098.

BELL, E., IVARSSON, B. & MERRILL, C. (1979). Production of a tissue-like structure by contrac-tion of collagen lattices by human fibroblasts of different proliferative potential in vitro.Proc. natn. Acad. Sci. U.S.A. 76, 1274-1278.

BELLOWS, C. G., MELCHER, A. H. & BRUNETTE, D. M. (1980). Orientation of calvaria andperiodontal ligament cells in vitro by pairs of demineralized particles. J. Cell Set. 44, 59-73.

BRUNETTE, D. M., KANOZA, R. J., MARMARY, Y., CHAN, J. & MELCHER, A. H. (1977). Inter-actions between epithelial and fibroblast-like cells in cultures derived from monkey perio-dontal ligament. J. Cell Sci. 27, 127—140.

BRUNETTE, D. M., MELCHER, A. H. & MOE, H. K. (1976). Culture and origin of epithelial-likeand fibroblast-like cells from porcine periodontal ligament explants and cell suspensions.Archs oral Biol. 21, 393-400.

COHEN, I. (1979). The contractile system of blood platelets and its function. Meth. Achiev. exp.Pathol. 9, 40-86.

DRURY, R. A. B. & WALLINGTON, E. A. (1967). Carleton's Histological Technique, 4th edn.London: Oxford University Press.

EHRMANN, R. L. & GEY, G. O. (1956). The growth of cells on a transparent gel of reconstitutedrat-tail collagen. J. natn. Cancer Inst. 16, 1375-1390. \

GABBIANI, G., CHAPONNIER, C. & HUTTNER, I. (1978). Cytoplasmic filaments and gap junc-tions in epithelial cells and myofibroblasts during wound healing. ,7. Cell Biol. 76, 561-568.

GABBIANI, G., HIRSCHEL, B. J., RYAN, G. B., STATKOV, P. R. & MAJNO, G. (1972). Granulationtissue as a contractile organ: a study of structure and function. J. exp. Med. 135, 719-734.

GABBIANI, G., MAJNO, G. & RYAN, G. B. (1973). The fibroblast as a contractile cell: the myo-fibroblast. In Biology of fibroblast (ed. E. Kulonen & J. Pikkarainen), pp. 139-154. London,New York: Academic Press.

11 CEL 50


HUXLEY, H. E. (1973). Muscular contraction and cell motility. Nature, Lond. 343, 445-449.HYNES, R. D. (1976). Cell surface proteins and malignant transformation. Biochim. biophys.

Acta 458, 73-107-JAMES, D. W. (1964). Wound contraction - a synthesis. In Advances in Biology of Skin, vol. 5

(ed. W. Montagna & R. E. Billingham), pp. 216-230.KORN, E. (1978). Biochemistry of actomyosin-dependent cell motility (a review). Proc. natn.

Acad. Sci. U.S.A. 75, 588-599.LAZARIDES, E. & WEBER, K. (1974). Actin antibody: the specific visualization of actin filaments

in non-muscle cells. Proc. natn. Acad. Sci. U.S.A. 71, 2268-2272.LIMEBACK, H. F. & SODEK, J. (1979). Procollagen synthesis and processing in periodontal

ligament in vivo and in vitro. Eur. J. Biochem. 100, 541-550.MAJNO, G., GABBIANI, G., HIRSCHEL, B. J., RYAN, G. B. & STATKOV, P. R. (1971). Contraction

of granulation tissue in vitro: similarly to smooth muscle. Science, N.Y. 173, 548-550.MARMARY, Y., BRUNETTE, D. M. & HEERSCHE, J. N. M. (1976). Differences in vitro between

cells derived from periodontal ligament and skin of Macacairus. Archs oral Biol. 21, 709-716.MELCHER, A. H. (1969). Healing of wounds in the periodontium. In Biology of the Periodontium

(ed. A. H. Melcher & W. H. Bowen), pp. 449-529. London, New York: Academic Press.MELCHER, A. H. (1980). Periodontal ligament. In Orban's Oral Histology and Embryology,

9th edn (ed. S. N. Bhasbar), pp. 204-239. St Louis, Mo.: C. V. Mosby Co.MELCHER, A. H. & BEERTSEN, W. (1977). The physiology of tooth eruption. In The Biology of

Occlusal Development (ed. J. A. McNamara, Jr), monograph no. 7, Craniofacial GrowthSeries, pp. 1-23. Ann Arbor, Mich. Centre for Human Growth and Development: Universityof Michigan.

POLLARD, T. D. (1975). Functional implications of the biochemical and structural propertiesof cytoplasmic contractile proteins. In Molecules and Cell Movement (ed. S. Inoue &R. E. Stevens), pp. 259-286. New York: Raven Press.

POLLARD, T. D. & WEIHING, R. R. (1974). Actin and myosin and cell movement. C.R.C. Crit.Rev. Biochem. a, 1-65.

RAO, L. G., NG, B., BRUNETTE, D. M. & HEERSCHE, J. N. M. (1977). Parathyroid hormone-and prostaglandin E^response in a selected population of bone cells after repeated sub-culture and storage at —80 °C. Endocrinology 100, 1233-1241.

ScHOR, S. L. & COURT, J. (1979). Different mechanisms in the attachment of cells to nativeand denatured collagen. J. Cell Sci. 38, 267-281.

SHORE, R. C. & BERKOVITZ, B. K. B. (1979). An ultrastructural study of periodontal ligamentfibroblasts in relation to their possible role in tooth eruption and intracellular collagendegradation in the rat. Archs oral Biol. 24, 155-164.

VENABLE, J. H. & COGGESHALL, R. (1965). A simplified lead citrate stain for use in electronmicroscopy. J. Cell Biol. 25, 407-408.

VOELZ, H. & DWORKIN, M. (1962). Fine structure of Myxococcus xantaus during morpho-genesis. J. Bact. 84, 943-952.

YAMADA, K. M., OLDEN, K. & PASTAN, I. (1978). Transformation-sensitive cell surface protein:isolation, characterization, and role in cellular morphology and adhesion. Ann. N. Y. Acad.Sci. 312, 256-277.

(Received 1 December 1980)

Note added in proofSince acceptance of this manuscript, Harris, Stopak & Wild (1981, Nature 290, 249) have

demonstrated that cellular traction differs greatly between cell types, that fibroblasts exertforces greater than those needed for locomotion and that fibroblasts can orient fibres in acollagen gel. These observations are consistent with our previous results indicating thatorientation of cells can depend upon the establishment of tensional forces (Bellows, Melcher& Brunette (1980) J. Cell Sci. 44, 59) and our observations here on the orientation of fibresin collagen gels.

contraction and organization of collagen gels by … · 2005. 8. 21. · collagen gels plated in 60...

Documents