Colchicine Reverts Cell Shape But Not Collagen Phenotypes in Corneal Endothelial Cells Modulated by Polymorphonuclear Leukocytes

Transcription

1 Colchicine Reverts Cell Shape But Not Collagen Phenotypes in Corneal Endothelial Cells Modulated by Polymorphonuclear Leukocytes EunDuck P. Kay and Suk Oh The authors have shown previously that polymorphonuclear leukocytes (PMN) modulate rabbit corneal endothelial cells (CEC) into cells that irreversibly acquire the characteristics of fibroblasts, including multilayering of spindle-shaped cells and deposition of interstitial extracellular matrix composed predominantly of type I collagen. In an attempt to determine if the changes in cell shape caused by the disruption of cytoskeleton are correlated with the alteration of collagen phenotypes in fibroblastic corneal endothelial cells (FCEC), colchicine and cytochalasin B (CB) were used. A series of dose-response studies were performed, and correlated with exposure time. When cells were exposed to the drugs (ranging from /ig/ml) 24 hr after plating, the majority of cells treated with colchicine dramatically changed from fibroblastic to polygonal shape: cells became flattened and cytoplasmic processes disappeared. Conversely, no apparent changes were observed in the CB-treated cells. On removal of colchicine, the cells resumed fibroblastic morphology within 24 hr, most of the cells again developed cytoplasmic processes. When collagen phenotypes were analyzed by electrophoresis, types I, III, and V collagen were present in either the colchicine or CB-treated cells, regardless of the concentration of drug used. However, synthesis of type I trimer and type HI collagen was significantly increased in the cells treated with colchicine at concentrations S: 1.0 Mg/ml; the al:a2 ratio was approximately 4.5, and type III accounted for 35-40% of the total collagen. CB did not induce a similar alteration. These observations indicate that changes in cell shape are not related to the switch of collagen phenotypes in FCEC. Furthermore, transiently induced polygonal-shaped FCEC maintain the ability to produce interstitial collagens in a manner identical to this observed in early stages of growth of FCEC. Invest Opthalmol Vis Sci 28: ,1987 Polymorphonuclear leukocytes (PMN) have been reported to modulate polygonal corneal endothelial cells into fibroblastic cells that synthesize predominantly type I procollagen rather than type IV collagen, which are seen in Descemet's membrane-corneal endothelium complex. 1 " 4 The modulation process appears to be composed of several stages: selection of sensitive cells, induction to type I collagen-synthesizing cells, and enhancement of their growth until they reach irreversible metaplasia. There are striking coincident changes in cell shape and the expression of collagen phenotypes: nonresponding polygonal endothelial cells maintain synthesis of type IV collagen, and the responding spindle-shaped endothelial cells synthesize type I collagen in the presence of PMN-conditioned medium. From the Department of Ophthalmology, Southwestern Medical School, University of Texas, Dallas, Texas. Supported by grant EY06431 from the National Eye Institute. Submitted for publication: July 24, Reprint requests: EunDuck Kay, DDS, Department of Ophthalmology, University of Texas HSCD, 5323 Harry Hines Boulevard, Dallas, TX There have been experimental studies that support the hypothesis that cell shape and the cytoskeleton may affect the expression of certain genes. For example, cell shape changes owing to the loss of actin filaments are prerequisites to initiation of biosynthesis of lipogenic enzymes in certain subpopulation of 3T3 cells 5 ' 6 ; with synovialfibroblastcultures, changes in cell shape correlate with induction of collagenase. 7 * 8 Several studies also indicate that cell shape influences chondrogenic gene expression. 9 " 11 The differentiated phenotypes of chondrocytes in vitro are enhanced when cells are maintained in a round configuration by culturing them in agarose, whereas chondrocytes dedifferentiate when exposed to retinioic acid, with resultant loss of polygonal morphology, or when grown on lens capsule. 12 ' 13 During initial modulation of corneal endothelial cells by PMN-conditioned medium, there were striking coincident changes in the cell shape and in the expression of collagen. Therefore, the hypothesis that reversion of cell morphology back to polygonal shape can influence the expression of collagen was tested in the study described herein by using cytoskeleton-disrupting agents to alter cell morphology. 826

2 No. 5 COLCHICINE REVERTS CELL SHAPE OUT NOT COLLAGEN IN FCEC / Koy and Oh 827 Cell Cultures Materials and Methods Isolation and establishment of primary rabbit corneal endothelial cells in culture was as previously described. 1 " 4 Briefly, the authors used sequential treatments of Descemet's membrane-corneal endothelium with 0.2% tryspin (Cooper Biomedical; Malvera, PA), 0.2% collagenase (Cooper Biomedical) and 0.5% hyaluronidase (Cooper Biomedical), and 0.2% trypsin containing 1 mm dithiothreitol. Cultures were maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco; Grand Island, NY) supplemented with 10% fetal calf serum (Hazleton; Denver, PA) and 50 Mg/ml gentamicin in a humidified atmosphere of 7.5% CO 2 in air. Modulated fibroblastic corneal endothelial cells (FCEC) were established by using PMN-conditioned medium. 1 PMNs were obtained from the peritoneal cavity of New Zealand rabbits after glycogen stimulation. PMNs were purified as previously described. 14 To obtain PMN-conditioned medium, cells were plated in tissue culture dishes (3.2 X 10 6 cells/ml) and incubated at 37 C for 24 hr in a humidified atmosphere of 7.5% CO 2 in air. Then medium was collected and sterilized. This medium was used for initiation of modulation and maintenance of all modulated endothelial cells. For these studies, FCEC from the 50th passage were plated at 1.27 X 10 4 cells/cm 2. Drug Treatments Colchicine and cytochalasin B (CB) were obtained from Sigma Chemical Co. (St. Louis, MO). CB was kept as stock solution of 1 mg/ml in dimethyl sulfoxide (DMSO) at -20 C. The drugs were applied to cells at final concentrations of Mg/ml, either 1 or 2 days after plating. For reversibility experiments, drug treatment was initiated on day 4 of culture followed by subculture on day 7. The subcultures were maintained in the drug-containing medium for 2 days, followed by removal of drugs 24 hr, 48 hr, or 6 days later. Protein Synthesis The cells were labeled for 20 hr with DMEM containing 200 MCi[5-3 H]proline (27 Ci/mM) (Dupont; Boston, MA), 2% fetal calf serum, 50 Mg/ml 2-aminopropionitrile, and 50 Mg/ml ascorbate. Medium was collected and the cell debris was removed by centrifugation. Protease inhibitors (1 mm phenylmethylsulfonyl fluoride, 10 mm N-ethylmaleimide, 4 mm ethylenediamine tetraacetic acid) were added. The medium fraction was then concentrated with ammonium sulfate to 45% saturation. The precipitate was collected by centrifugation at 10,000 X g for 15 min at 4 C and dissolved in Buffer A (5 M NaCl, 0.05 M Tris-HCl, ph 7.4,% Triton-X-100) followed by dialysis in the same buffer. An aliquot of the precipitate was dialyzed against 0.5 M acetic acid and % Triton-X-100, followed by pepsin treatment (100 Mg/ml) at 4 C for 24 hr. The enzyme reaction was stopped by raising ph to 8.0, and the reaction mixture was dialyzed against Buffer A. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) Polypeptides were electrophoresed under the conditions described by Laemmli. 15 Gels were fixed in 20% trichloroacetic acid for 30 min and processed for fluorography as described. 16 For the quantitation of individual collagen types, each monomeric band separated by electrophoresis and localized by fluorography was hydrolyzed with 0.4 N NaOH for 16 hr at 55 C, neutralized with 1 M acetic acid, and counted. Enzyme Digestion Bacterial collagenase digestion was performed as described by Peterkofsky and Diegelman. 17 Results The fibroblastic corneal endothelial cells (FCEC) in tissue culture have been well characterized in the authors' previous works. In order to document the modulation process, every fifth passage of FCEC cultures has been examined for its collagen phenotypes; each subculture, regardless of passage number, expresses collagen phenotypes identical to those observed in the earlier passage cultures obtained immediately after irreversible modulation. Because cell attachment required at least 1 day, during which few cells divided, treatment with each drug was initiated either 24 or 48 hr after plating. Dose Dependence of Cell-Shape Changes: Comparison Following Colchicine and CB Treatment The effect of colchicine and CB on the cell shape was determined by culture for 4 days in the presence of drugs at concentrations ranging from 10 ng/ml to 4 Mg/ml; phase contrast micrographs were taken daily at each concentration. Colchicine at a concentration of Mg/ml was effective in altering the fibroblastic morphology to polygonal shape (Fig. 1), which was essentially complete within 24 hr of exposure, at which time most of the cells acquired normal polygonal shape. Morphologic changes resulting from one day of exposure were similar to those produced by 48-hr exposure (data not shown). In contrast, CB was not effective in altering the cell shape of FCEC at the con-

3 828 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / May 1987 Vol. 28 Fig. 1. Phase-contrast micrographs of fibroblastic cornea] endothelial cells after colchicine or CB treatment for 4 days. (A) untreated cultures; (B) cultures treated with fig/ml colchicine; (C) cultures treated with 1.0 n%/m\ colchicine; (D) cultures treated with 2.0 jtg/ml colchicine; (E) cultures treated with 2.0 ng/ml CB; and (F) cultures treated with 4.0 Mg/ml CB. centrations nor times used. Four days of treatment with colchicine at higher concentrations (2.0 and 4.0 ng/ ml) enhanced cytotoxixity, and CB-treated cells (2.0 Mg/ml) essentially reached fibroblastic confluency. CB at 4.0 ^g/ml resulted in cytotoxicity of FCEC, albeit at a low level. Of interest is that disruption of microtubules by colchicine causes polygonal morphology of FCEC; this is in contrast to CB, which does not alter the morphology of the same cell line. Control experiments with keratocytes and normal corneal endothelial cells (data not shown) provide further evidence that alteration of cell shape in FCEC is a reaction specific to these cells, and is not a general property of colchicine on cell lines in general. Collagen Synthesis: Comparison of Colchicine With CB-Treated Cells The fibroblastic corneal endothelial cells is characterized by synthesis of interstitial collagens (types I, III, and V). 1 " 3 FCEC, when multilayered, maintain a stationary expression of different collagen types: type I, 75-80%, type III, 10-15%; and type V, the remainder. 3 In contrast, the polygonal, normal corneal endothelial cells synthesize type IV collagen 4 and type VIII collagen. l8 ' 19 It is of importance, however, to note that FCEC maintain a rather active state of collagen metabolism, which appears to correlate with cell numbers and stage of cell growth; on day 2 of culture, the a\:al ratio of type I collagen is larger than 6, which then decreases to 2 as the cells multilayer. Concomitantly, the proportion of type III collagen increased to one third of total collagen and such transient alteration reached a stationary level during later days in culture. 3 In order to determine the differences induced by the drugs alone, cells were labeled with radioactive precursor on day 7 of culture. The same cultures used to study cell shape changes induced by the drugs were analyzed using SDS-PAGE of collagen chains in order to determine whether there were dose-dependent qualitative changes in collagen phenotypes (Fig. 2). The collagen phenotypes in both colchicine-treated cells and CB-treated cultures are types I, III, and V collagen, indicating that there is no switch in collagen phenotypes. The striking finding is that the polygonal FCEC induced by colchicine ( or 1.0 A*g/ml) maintain synthesis of interstitial collagens. The most apparent alteration is the significant decrease of a2(i) in colchicine-treated cultures (Fig. 2), the degree of which is correlated with the concentration of colchicine. CB, which was not effective in alterating cell shapes, did not affect the stoichiometry of type I

4 No. 5 COLCHICINE REVERTS CELL SHAPE BUT NOT COLLAGEN IN FCEC / Kay and Oh 829 type ill- _ 02(0- ~ Fig. 2. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of collagen phenotypes synthesized after 4 days of treatment with various concentrations of colchicine or CB. Cells were treated with drugs and labeled with radioactive precursors as described in the text. Medium fractions were treated with pepsin before electrophoresis on a 4.5% gel. Colchicine (lanes 2-6) or CB (lanes 8-12) were present during incorporation. Untreated cultures (lanes 1, 7); 0.01 jig/ml (lanes 2, 8); /*g/ml (lanes 3, 9); 1.0 Mg/ml (lanes 4, 10); 2.0 /ig/ml (lanes 5, 11); and 4.0 #ig/ml (lanes 6, 12). collagen. When Figure 2 is translated into mathematical quantitation (Table 1), the degree of deviation in the a\:a2 ratio of type I collagen was correlated with concentration of colchicine; there is no apparent correlation between cell shape and collagen phenotypes: polygonal FCEC induced by Mg/ml of colchicine maintain the correct stoichiometry of type I collagen, and polygonal FCEC induced by 1.0 Mg/ml of colchicine have an increased ratio. There are striking coincident decreases in a2(i) synthesis and in the proportion of type I collagen: the proportion of type I collagen concomitantly decreased with 1.0 Mg/ml of colchicine, which appeared to have the maximum effect. However, the possibility that a relative increase in a 1(1) synthesis rather than a decrease in a2(l) synthesis cannot be excluded. Another possibility of specific loss of a2(i) owing to its more susceptible property during proteolysis was partly eliminated by differential salt precipitation of some procollagen fractions before pepsin treatment (data not shown). In contrast, neither the stoichiometry nor the proportion of type I collagen decreased below the value of stationary expression in the CB-treated cells. Time-Course: Colchicine Effect on Type I Collagen Colchicine at Mg/ml was able to induce morphologic changes in FCEC but did not induce the altered expression of type I collagen observed with higher concentrations of colchicine. The authors wanted to de- Table 1. Effect of colchicine and cytochalasin B on collagen* Drug Colchicine Cytochalasin Concentration (tig/ml) aid) a 2(1) 2.46 ± ± ± ± ± ± ± ± ± ± ± ± 0.05 % of type I 84.3 ± ± ± ± ± ± ± ± ± ± ±1.4 * The samples analyzed in Figure 2 were quantitated for the individual chains as described in the text.

5 830 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / Moy 1987 Vol. 28 Table 2. Time course of colchicine effect on type I collagen* Days in culture (passage no.) 3(50) 7(50) 10 (50) 2(51) Concentration (^g/ml) a 1(1) a2(i) 5.0 ± ± ± 2.5 ± ± ± 4.9 ± ±0.4 * Quantitation was performed as described in the text. % of type I 55 ± ± ± ± ± ± ± ±1.2 termine whether sustained morphologic alteration triggered an altered expression of type I collagen. Cells were treated continuously with /ig/ml colchicine (except for 4 hr immediately after subculture, during which time cells were maintained in DMEM containing 10% fetal calf serum for plating), a concentration at which no cytotoxicity was observed during the period of examinination (initial exposure and one subculture). Cell morphology remained altered throughout the period (data not shown). During 10 days of culture, FCEC showed characteristic features on stoichiometry and the proportion of type I collagen (Table 2) as a function of culture age and cell numbers, which is designated as cell-density-dependent phenotypic modulation: during early days in culture, the a\\a2 ratios elevated because of the low cell density and exponential growth potential. As cells reached confluency and became multilayered, the stoichiometry and proportion of the collagen resumed normal values. Continuous treatment with colchicine neither enhanced the values at early days (day 3), nor induced higher values (days 7, 10). Furthermore, no quantitative changes in the a\:a2 ratio and the amount of type I collagen were observed with subcultures originated from the drug-treated cultures and sustained under similar conditions. These findings indicate that sustained changes in the morphology by colchicine did not induce altered expression of type I collagen. Reversibility of Effects of Colchicine: Cell Shape and Collagen Synthesis The cultures passaged from colchicine-treated cells ( Mg/ m 0 w ere further used to monitor the reversibility of the cell shape. Colchicine was removed on day 2 and cultures further maintained for 1, 2, or 6 days in the absence of colchicine. Control cultures, at each of the different time intervals, showed fibroblastic morphology (Figs. 3A, C, E, G), and cells treated with Mg/ml colchicine maintained the acquired polygonal shape (Fig. 3B). On removal of the drug, the disappearance of polygonal shape was complete within 1 day, at which time most of the cells reverted to spindle shape with prominent cytoplasmic processes (Fig. 3D). During the following 24 hr, cells appeared to resume their capacity for proliferation (Fig. 3F). On day 6 after removal of the drug, the cells reverted back to normal fibroblastic confluency (Fig. 3H). During the period of reversal, the cultures previously treated with colchicine appeared to have fewer cells, most likely because of the antimitotic effect of this drug. The same cultures were studied for collagen synthesis (Table 3). Control cultures in this experiment were composed of two sets: either for drug or age dependency. The stoichiometry of type I collagen was compared. FCEC, either in the presence or absence of colchicine, maintained elevated values at day 3; indicating that colchicine does not enhance the already elevated value of a\:a2 ratio. At this stage, low cell numbers and subsequent high degree of cell division appear to regulate the stoichiometry. On removal of colchicine, the spindle-shaped cells maintained the elevated level during thefirst2 days because of their relative low cell density, and their control cultures demonstrated age and cell number dependency of the stoichiometry (FCEC at day 3 have 4.3 in contrast to 2.6 at day 4). The confluent cells on day 6 after removal of colchicine resumed normal values. Thesefindingsindicate an apparent correlation between collagen metabolism and cell numbers, and suggest that colchicine does not further enhance the cell density-dependent phenotypic modulation in FCEC. Discussion These results demonstrate that colchicine causes changes in cell shape in the modulated endothelial cells (FCEC); thefibroblasticcells assume polygonal shape. The change is both dose and time dependent before reaching cytotoxic levels. The onset concentration tested is /ig/ml of colchicine and major changes are complete 24 hr after initiation of treatment. Such acquired polygonal shapes are maintained in the presence of colchicine, but when colchicine is removed, the polygonal cells revert back to fibroblastic morphology within 24 hr. On the other hand, cytochalasin B in subcytoxic doses does not cause any alteration in cell shape. These observations suggest that microtubules rather than actin filaments control the shape of the fibroblastic corneal endothelial cells. Further evidence with other cell lines (stromal fibroblasts and normal corneal endothelial cells) suggests that alteration of cell shape by colchicine is a specific function of FCEC. The collagen phenotypes in both colchicine-treated cells and

6 No. 5 COLCHICINE REVERTS CELL SHAPE DUT NOT COLLAGEN IN FCEC / Kay and Oh 801 Fig. 3. Phase-contrast micrographs of fibroblastic corneal endothelial cells following removal of colchicine. Cells were treated with jtg/ml colchicine for 4 days followed by subculture. The subcultured cells were allowed to plate for 4 hr in the absence of colchicine and maintained in drug containing media for 2 days. Colchicine was then removed and cultures maintained for 1 day (D), 2 days (F), or 6 days (H) in the absence of colchicine. (A) Control cultures on day 3; (B) cultures on day 3 treated with ^g/ml colchicine; and control cultures on day 4 (C), day 5 (E), and day 8 (G). CB-treated cultures are types I, III, and V collagen, indicating that there is no switch in collagen phenotypes. Polygonal FCEC induced and maintained by colchicine for more than one passage maintain synthesis of interstitial collagens, suggesting that alteration in cell shape does not cause phenotypic switches of collagen in FCEC, unlike the phenomena reported in other systems in vitro (ie, polygonal shape versus fibroblastic chondrocytes and their distinctive collagen phenotypes). It appears that colchicine influences collagen metabolism by controlling cell numbers: the antimitotic effect of the drug results in fewer cells, which in turn regulates collagen metabolism. The authors' previous studies demonstrated an apparent relationship between cell numbers and collagen metabolism in Table 3. Stoichiometry of type I collagen on reversibility (R) after colchicine treatment* Samples Control R-1 R-2 R-3 Colchicine (ng/mi) Days after removal ~2 5.0 ± ± ± ± ± ± ± ±0.3 * Drug treatment was initiated on day 4 of culture followed by subculture on day 7. The subcultures were maintained in the drug-containing medium for 2 days followed by removal of drugs 24 hr (R-1), 48 hr (R-2), or 6 days (R-3) later.

7 832 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1987 Vol. 28 FCEC: during the early stage of growth, when cells exponentially grow, «2(I) synthesis and total type I collagen synthesis markedly decrease, in contrast to the transient increase in type III collagen. The expression of collagen phenotypes are essentially stabilized with increased cell numbers. 3 Such a phenomenon (cell density-dependent phenotypic modulation) also is observed in FCEC treated with colchicine. For example, at 1.0 Mg/ml concentration, cell number is one fifth that of control cultures (Unpublished data: E. Kay) and the spectrum of changes in collagen phenotypes resembles that observed in early stage of growth in FCEC (ie, marked increase in type I trimer and decrease of overall type I collagen synthesis). The authors also have observed that colchicine does not further influence the phenotypic expression of collagen in FCEC. Furthermore, there is no apparent correlation between cell shape and collagen phenotypes; polygonal FCEC induced by Mg/ml colchicine retain the correct stoichiometry and the proportion of type I collagen, in contrast to the polygonal FCEC induced by 1.0 Mg/ml colchicine, which have altered levels. There have been speculations and evidence acquired that cell shape plays an important role in growth control and in gene expression. 6 " 11 Alteration in cell morphology caused by a variety of agents has shown to be correlated with induction of collagenase in rabbit synovial fibroblasts. 7 An in-vitro study of chondrogenic differentiation of early limb bud mesenchyme demonstrates that it is influenced by cell shape; the shape of limb bud cells is controlled by actin filaments rather cytoplasmic microtubules. 9 In addition to the cytoskeleton-disrupting agents, there are other processes that control cell shapes. 20 ' 21 However, although the underlying mechanism, in general, seems to be associated with the organization of the cytoskeleton of the cell, the biochemical mechanism has yet to be elucidated. Despite experimental evidence of a correlation between cell shape (and the cytoskeleton) and gene expression in several experimental systems, there is no apparent correlation between cell shape and collagen phenotypes in the colchicine treated FCEC. This finding is rather important, for there were striking coincident changes in the cell shape and in the expression of collagen during the initial modulation of corneal endothelial cells by PMN-conditioned medium: polygonal nonresponding cells continue to synthesize type IV collagen, and elongated responding cells completely switch collagen phenotypes to those of interstitial collagens. However, the authors do not know whether the initial modulation process represents corelationship between cell shape and gene expression or if it is a rather striking coincidence of two distinct processes. Subsequent but independent modulation of the initially modulated corneal endothelial cells demonstrates that there is neither correlation nor coincidence between cell shape and collagen phenotypic changes. The present findings suggest another important aspect of corneal endothelial modulation: transiently induced and sustained polygonal cell morphology does not induce switches in collagen phenotypes. Regardless of the cell shape, collagen phenotypes are expressed by the characteristic cell density-dependency process of FCEC: polygonal FCEC or spindle-shaped FCEC produce type I trimer at low cell density. In an attempt to find avenues of therapy and prevention of posterior collagenous layer, which appears between Descemet's membrane and corneal endothelium, the hypothesis that reversion of cell shape of fibroblastic endothelial cells to polygonal configuration may induce changes in collagen gene expression was tested in the present study. Unlike other normal cells (chondrocytes, synovial fibroblasts, and so forth), FCEC that have acquired irreversible metaplasia demonstrate resistance to cytoskeleton-disrupting agents, which readily cause alteration in both cell shape and gene expression in some other cells lines. Further studies of the correlation between cell shape and collagen gene expression should be performed in fibroblastic corneal endothelial cells because they may yield useful information concerning the control of gene expression of collagen, and may indicate avenues of therapy of the posterior collagenous layer in cornea. Key words: colchicine, cytochalasin B, collagen, modulation, corneal endothelial cells References 1. Kay EP, Smith RE, and Nimni ME: Type I collagen synthesis by corneal endothelial cells modulated by polymorphonuclear leukocytes. J Biol Chem 260:5139, Kay EP, Nimni ME, and Smith RE: Modulation of endothelial cell morphology and collagen synthesis by polymorphonuclear leukocytes. Invest Ophthalmol Vis Sci 25:502, Kay EP: Rabbit corneal endothelial cells modulated by polymorphonuclear leukocytes are fibroblasts: comparison with keratocytes. Invest Ophthalmol Vis Sci 27:891, Kay EP, Smith RE, and Nimni ME: Basement membrane collagen synthesis by rabbit corneal endothelial cells in culture. Evidence for an alpha chain derived from a larger biosynthetic precursor. J Biol Chem 257:7116, Spiegelman BM and Farmer SR: Decreases in tubulin and actin gene expression prior to morphological expression of 3T3 adiopocytes. Cell 29:53, Spiegelman BM and Ginty DA: Fibronectin modulation of cell shape and lipogenic gene expression in 3T3-adiopocytes. Cell 35:657, Aggeler J, Frisch SM, and Werb Z: Changes in cell shape correlate with collagenase gene expression in rabbit synovial fibroblasts. J Cell Biol 98:1662, Harris ED Jr, Reynolds JJ, and Werb Z: Cytochalasin B increases collagenase production by cells in vitro. Nature (Lond) 257:243, 1975.

8 No. 5 COLCHICINE REVERTS CELL SHAPE BUT NOT COLLAGEN IN FCEC / Kay and Oh Zanetti NC and Solursh M: Induction of chondrogenesis in limb mesenchymal cultures by disruption of the actin cytoskeleton. J Cell Biol 99:115, Benya PD and Shaffer JD: Dedifferentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels. Cell 30:215, Glowacki J, Trepman E, and Folkman J: Cell shape and phenotypic expression in chonochocytes. Proc Soc Exp Biol Med 172:93, Benya PD and Padilla S: Modulation of rabbit chondrocyte phenotype by retinoic acid terminates type II collagen synthesis without introducing type I collagen. The modulated phenotype differs from that produced by subculture. Dev Biol 118:296, Tsunematsu Y: Effects of the lens capsule on cellular flattening, cell growth, and expression of the differentiated traits of chondrocytes cultured in vitro. Dev Growth Differ 21:437, Meyers RL and Pettit TH: Chemotaxis of polymorphonuclear leukocytes in corneal inflammation: tissue injury in herpes simplex virus infection. Invest Ophthalmol 13:187, Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680, Laskey RA and Mills AD: Quantitative film detection of 3 H and I4 C in polyacrylamide gels by fluorography. Eur J Biochem 56: 335, Peterkofsky B and Diegelman R: Use of a mixture of proteinasefree collagenases for the specific assay of radioactive collagen in the presence of other proteins. Biochemistry 10:988, Benya PD and Padilla SR: Isolation and characterization of type VIII collagen synthesized by cultured rabbit corneal endothelial cells. J Biol Chem 261:4160, Benya PD: EC collagen: Biosynthesis by corneal endothelial cells. and separation from type IV without pepsin treatment or denaturation. Renal Physiol 3:30, Folkman J and Moscona A: Role of cell shape in growth control. Nature (Lond) 273:345, Tomasek JJ, Hay ED, and Fujiwara K: Collagen modulates cell shape and cytoskeleton of embryonic corneal and fibroma fibroblasts: Distribution of actin, a-actin and myosin. Dev Biol 92: 107, 1982.