THE ANATOMICAL RECORD 23252-59 (1992) Influence of Retinol on Human Chondrocytes in Agarose Culture AMY LYNN AULTHOUSE, CECILIA M. CARUBELLI, TOD M. DOW, CHRISTINE ZIEGELMAYER, AND MICHAEL BECK University of Oklahoma Health Sciences Center, College of Medicine, Anatomical Sciences, Oklahoma City, Oklahoma (A.LA.,C.M.C., T.M.D.); University of Mainz, Department of Pediatrics, Mainz, Germany (C.Z., M B J ABSTRACT Vitamin A and its congeners, collectively called retinoids, are known to have teratogenic potential and have induced craniofacial and limb malformations in numerous animal species. More importantly, retinoids are recognized as teratogenic to fetuses of pregnant women who have taken such preparations for dermatologic disorders. Information gathered from the study of animal models suggests that retinoids interfere with cartilage differentiation. If chondrogenesis in limb development is disturbed it may contribute to limb reductions and malformations. In vitro studies using various animal systems have shown that cartilage matrix macromolecules are altered to resemble those secreted by mesenchymal cells. The response of human chondrocytes to retinoids in vitro is not known. Culture of human chondrocytes in agarose maintains the cartilage phenotype and therefore serves as a model system to evaluate the influence of retinoids directly on human chondrogenesis. The studies presented in this paper were done t o determine if the expression of specific matrix macromolecules of human chondrocytes in agarose culture is altered by retinol treatment. Immunocytochemistry demonstrated enhanced labeling of type I collagen while type I1 collagen labeling was reduced in cultures treated with retinol. In addition, morphometric analyses indicated a decrease in the size and number of chondrogenic clusters and that individual cells synthesized less alcian blue matrix when compared to parallel control cultures. The size of the proteoglycan monomers, glycosaminoglycan side chains as well as the disaccharide composition were not affected. However, there was a reduction in the quantity of proteoglycan monomers produced. Vitamin A and its analogues can exert major effects on the control of pattern formation and differentiation. It has been shown that retinoic acid when applied directly to developing chick limb buds causes duplication of digits (Tickle et al., 1985) and proximo-distal serial repetition of stump tissue in regenerating limbs (Maden, 1982, 1983). Retinoic acid has been shown to mimic the activity of the zone of polarizing activity (ZPA), a posterior region of the limb bud known to cause duplications when grafted anteriorly (Tickle et al., 1982). Retinoic acid has been described to act as a morphogen. Morphogens are hypothetical signal substances emitting from an organizing center. The cells then respond to different concentrations by adopting different pathways of differentiation (Slack, 1987). Hence, certain retinoids may be responsible for proper development of chick limb buds when in specific concentrations and locations. However, recent studies by Wanek et al. (1991) have suggested that retinoic acid acts by converting anterior limb mesenchymal cells to ZPA cells which in turn would bring about digit duplication. Moreover, Noji et al. (1991) have reported that it is unlikely that exogenous retinoic acid is identical with a morphogen secreted from a graphed ZPA but that there could be an “endogenous retinoic a c i d that may serve as a morphogen. Although retinoids may be instrumental in deter0 1992 WILEY-LISS. INC mining proper skeletal development, they have also been shown to be teratogenic when given during specific times and concentrations in development. The effects of excess retinol given to pregnant mice and rats during certain periods of gestational growth resulted in pups with skeletal defects. Rat pups displayed various craniofacial anomalies while mouse pups showed numerous craniofacial and limb defects (Pillans et al., 1988; Takakubo and Eto, 1986). Retinoids are commonly used in humans in the treatment of various dermatologic disorders. Specifically, isotretinoin (Accutane) and etretinate (Tigason) have been administered orally to treat severe acne and psoriasis. Since the release of these drugs, numerous cases of spontaneous abortions and fetal abnormalities have been reported. Fetal craniofacial and limb malformations have been associated with maternal ingestion of isotretinoin during pregnancy (Barnhart, 1989; Lammer et al., 1985; McBride, 1985; Willhite et al., 1986). Multiple synostoses, facial dysmorphia, syndactylies, Received April 9, 1991; accepted June 24, 1991. Address reprint requests to Amy Lynn Aulthouse, Ph.D., University of Oklahoma Health Sciences Center, College of Medicine, Department of Anatomical Sciences, P.O. Box 26901, Oklahoma City, OK 73190. INFLUENCE OF RETINOL ON HUMAN CHONDROCYTES absence of terminal phalanges, malformations of hip, ankle, and forearm, and other skeletal defects have been reported when etretinate was taken during pregnancy (Barnhart, 1989; Grote et al., 1985; Happle et al., 1984; Willhite e t al., 1986). Animal models have shown that retinoids interfere with limb development resulting in limb reduction (Kochhar, 1973; Takakubo and Eto, 1986; Willhite et al., 1986). Both in vivo and in vitro animal studies have demonstrated that retinoids interfere with chondrogenesis and subsequently with limb development. These reports have demonstrated that chondrogenesis is interrupted concurrent with a switch in collagen phenotype, a decrease in proteoglycan synthesis, and a n increase in fibronectin synthesis. Numerous cases of limb reduction in humans have been reported (Lammer et al., 1985; Willhite et al., 1986). Currently, human studies are restricted to interpretation of clinical manifestations because no system has yet been developed to evaluate the effect(s) of retinoids directly on human chondrocytes. We utilized an agarose culture system to directly study the response of human chondrocytes to retinol treatment. METHODS AND MATERIALS Culturing Techniques Primary cell culture Human costochondral cartilage was obtained a t autopsy. Individuals with recognized skeletal abnormalities were excluded. Six cases with ages ranging from 4 days to 6 years were evaluated. Primary cultures were established as previously described (Aulthouse et al., 1989). In brief, released chondrocytes were plated in Dulbecco’s Modified Eagles Media (DMEM; Gibco, Grand Island, NY) with 10% fetal calf serum (FCS; Irvine Scientific, Santa Ana, CA), 1%L-glutamine (Irvine Scientific), and with 0.1% penicillin-streptomycin (Gibco), in 24 multiwell tissue culture dishes (Corning, Corning, NY). 53 (Corning) that were coated with 1% high temperature agarose (Biorad Cat# 162-0100). The cultures were subsequently refrigerated at 4°C for 15 minutes to allow the 0.5% low temperature agarose to gel. The cultures were fed twice weekly for 1-4 weeks. Cultures were treated daily with 10 ng all-trans retinol (Sigma Chemical Co.) dissolved in absolute ethanol. A stock solution, 1mgll ml, of retinol was kept dark and stored at -20°C. To evaluate the effects of ethanol on chondrogenesis, parallel cultures were treated with equivalent volumes of absolute ethanol. In addition, untreated normal control cultures were established. There were no detectable differences between ethanoltreated and untreated normal controls. To evaluate the potential of cells to recover after exposure to retinol, some cultures were treated for 2 weeks with retinol and then allowed to rest for 2 weeks without further treatment. Light Microscopy Detection of Cartilage Proteoglycans With Alcian Blue Treated and control cultures were rinsed in phosphate-buffered saline (PBS) and then simultaneously fixed and stained for 24 hours in a mixture of 2.5% glutaraldehyde, 0.05% alcian blue with 25 mM sodium acetate (pH 5.6), and 0.4 M magnesium chloride. Fixation was followed by brief rinses with 3% acetic acid, then 25% and 50% ethanol equivalent volumes (ETOH) with 3% acetic acid and finally 70% ETOH (Aydelotte e t al., 1986). Quantification of Alcian Blue Staining Material of Single Cells In order to quantitate the amount of alcian blue staining material produced by individual cells, single cells were randomly selected from cultures that had been treated daily with 10 nglml retinol, or equivalent volumes of ethanol, or that were untreated. The amount of alcian blue surrounding a single cell was determined by subtracting the perimeter of the cell Monolayer culture from the total perimeter of the alcian blue halo surThe biochemical assays required quantities of cells rounding the cell. A Zeiss Interactive Digital Analysis in excess of those obtained in primary cultures. There- System (ZIDAS, Zeiss, Oberkochen, West Germany) fore, cells in primary cultures were grown to conflu- was used for these measurements. Using the Crunch ence and passaged to tissue culture flasks. Cells were Statistical program (Crunch Software Corporation, rinsed twice with Hank’s balanced salt solution (HBSS) Oakland, CA), the means were calculated and a one(Irvine Scientific) and dissociated in 0.2% trypsin way analysis of variance was performed. The Newman(Gibco) with 5 mM ethylenediamine-tetraacetic acid Keuls post-hoc test was performed to compare the dif(EDTA) (Sigma Chemical Co., St. Louis, MO). The cells ferent treatment groups (4 weeks retinol, 4 weeks were then resuspended in 4 ml of media and passaged ethanol, and 4 weeks untreated). to 25 cm2 tissue culture flasks. Cells were passaged one to five times. Agarose culture Cells were dissociated from monolayer, centrifuged, resuspended, and maintained in 5 ml of complete DMEM. The cell suspension was allowed to rest for 1 hour at 4°C and then filtered through 2 ply, No. 20, nitex (Tetko Inc., Elmsford, NY). The cell suspension was centrifuged and resuspended at 5 x lo5 cells in 1 ml of 0.5% low temperature agarose (Biorad Cat#1620017, Richmond, CA) in DMEM. For alcian blue quantitation and immunocytochemistry, 10 ~l of cell suspension was plated in 35 mm tissue culture dishes Quantification of Cell Clusters as an Expression of Chondrogenesis Alcian blue stained whole cultures were photographed with a n IMT-2 Olympus inverted microscope (E A 4 objective). All single cells and cell clusters associated with alcian blue staining were counted. Using the Crunch Statistical program the means were calculated and a one-way analysis of variance was performed. The Newman-Keuls post-hoc test was performed to compare the different groups of treatment (4 weeks retinol, 4 weeks ethanol, 4 weeks untreated, and 2 weeks retinol followed by 2 weeks untreated). A.L. AULTHOUSE ET AL. 54 Fig. f , Whole cultures stained with alcian blue after culture in agarose for 3 weeks. A: Normal untreated control culture. B: Culture treated daily with absolute ethanol (equivalent volume given in (C). C: Culture treated daily with 10 ng/ml all-trans retinol. A-C, X 15. lmmunocyiochemistiy Used to identify Matrix Macromolecules Retinol-treated, ethanol-treated, and untreated control agarose cultures were rinsed three times in PBS and fixed in two, 5 minute changes of methano1:acetone (1:l) and air dried overnight. After rehydration in PBS, the cultures were incubated overnight at 4°C in monoclonal antibody to one of the following monoclonal antibodies purchased from the Developmental Studies Hydridoma Bank, University of Iowa, Iowa City, IA: human type I1 collagen (CIICI, undiluted, Holmdahl et al., 1986) or human type I procollagen (M38, undiluted, McDonald et al., 1986). Cultures incubated in conditioned medium from NS1 myeloma cell line before fusion served as negative controls. The cultures were rinsed and incubated for 1hour at room temperature in fluorescein-conjugated rabbit anti-mouse IgG (1:100, Cappel, Organon Teknika Corporation, West Chester, PA). After final rinsing, the cultures were mounted in 80% glycerine in PBS and photographed through a n Olympus Vanox microscope (DPlanApo 20 UV PL Olympus objective) equipped for fluorescence microsCOPY. Proteoglycan Biochemistry The following biochemical assays were conducted on untreated (normal controls), ethanol-treated (volume equivalents), and all-trans retinol-treated chondrocytes grown in agarose culture. CL column (1.0 x 100 cm, Pharmacia, Uppsala, Sweden). The agarose was rinsed with PBS and the proteoglycans were extracted with 4 M guanidinium chloride (GnC1) in 0.1 M sodium acetate buffer, pH 5.8, containing protease inhibitors (0.1 M 6-aminohexanoic acid, 10 mM EDTA, 5 mM benzamidine, and 10 mM N-ethylmaleimide; all from Sigma Chemical Co.) a t 4°C for 24 hours. Subsequently, the agarose was separated from the extract by centrifugation for 1hour at 30,000 rpm (Sorvall Ultracentrifuge OTD65B, TV 865 Sorvall Vertical Rotor, Dupont Sorvall Instrument, Wilmington, DE). The supernatant was then dialyzed against water and precipitated with ammonium sulfate (48 hours at 4°C). The precipitate was then centrifuged for 1 hour at 30,000 rpm. The sample was then resuspended in 4 M GnCl buffer. Fifty microliters of the crude GnCl extract was next subjected to gel chromatography on a Sepharose 2B-CL column (1.0 x 100 cm, Pharmacia). The equilibration and elution buffer was 0.1 M sodium acetate, pH 5.8, with 4 M GnC1, and protease inhibitors. The excluded and total volumes were determined with high-molecular-weight DNA from bacteriophage lambda D (Sigma Chemical Co.) and Na235S0,, respectively. One milliliter fractions were collected and assayed for radioactivity. Determination of the hydrodynamic size of glycosaminoglycan chain lengths Proteoglycan containing fractions collected from the Sepharose 2B-CL column were further analyzed. Glycosaminoglycan chains were detached from proteoglyDetermination of the hydrodynamic size of the can monomers by beta elimination (Truppe and Kresse, proteoglycan monomers 1978). The samples were dialyzed against distilled waCells were suspended in agarose as described above ter, evaporated and resuspended in 500 p1 0.15 M for 3 weeks (5 x lo5 cells in 0.5 ml agarose in medium). NaOH, and incubated for 4 hours at 37°C. An equal After incubating for 48 hours in sulfate-free medium volume of 8 M GnCl was added to each sample. The (SERVA West Germany), the cells were exposed to 100 hydrodynamic size was then determined by gel chropCi Na>S04 (Amersham-Buchler, Frankfurt/M, West matography on a 1.0 x 100 cm Sephacryl S-300 column Germany) in 1.5 ml sulfate-free medium for 48 hours. (Pharmacia) as described by Beck et al. (1988). The Proteoglycans that were secreted into the medium and equilibration and elution buffer was 0.1 M sodium acretained in the agarose were analyzed. The medium etate, pH 5.8, with 4 M GnCl and protease inhibitors. was removed, precipitated with ammonium sulfate as The excluded and total volumes were determined with described below, and then applied to a Sepharose 2B- Dextran blue and Na235S04,respectively. INFLUENCE OF RETINOL ON HUMAN CHONDROCYTES '60 O! N^ 50 50 E E v 40 40 30 30 20 20 10 10 6 0 0 NC ROL 55 Glycosaminoglycan composition Proteoglycan monomers (50 pl about 1,000-2,000 cpm) collected from Sepharose 2B-CL fractions were digested with 20 mU chondroitin AC lyase (Sigma) in 0.1 M TrisiHCL, pH 7.3, with 0.1 M sodium acetate at 37°C for 12 hours as described by Beck et al., 1988. The digest was then subjected to descending paper chromatography in butan-1-ol/l M NH,/acetic acid (2:1:3 by volume). One centimeter strips, cut into small pieces were eluted in 1 ml distilled water and measured with a Kontron MR 300 Liquid Scintillation Counter. RESULTS Light Microscopy A - N Whole cultures (untreated, ethanol-treated, and retinol-treated) stained with alcian blue are seen in Figure 1. There are fewer alcian blue staining chondrogenic clusters in retinal-treated cultures when compared to ethanol and untreated cultures. Untreated normal control and ethanol-treated cultures have numerous, large clusters of isogenic groups of chondrocytes. The retinol-treated cultures tend to have fewer clusters that often contain only one or two cells. 15 Quantification of Aician Blue Staining Matrices NC non ROL 6 200 T $ + 5 - y.- T 200 180 180 160 160 140 140 120 120 ;100 m p 100 80 80 c 60 60 6 40 40 20 20 D 0 0 NC ETOH ROL ROL2 C Fig. 2. A Morphometric values recorded after 4 weeks treatment in culture of a typical case as alcian blue staining material per single cell (ABS mm'). NC, untreated; ETOH, ethanol equivalent volumes; ROL, 10 ng retinol. Values are the mean ? SEM. NC > ROL ( P 5 0.01); NC > ETOH (P 4 0.01); ETOH > ROL (not significant). B: Morphometric values of cell size (mm') from cells recorded in (A). NC, untreated; ETOH, ethanol equivalent volumes; ROL, 10 ng retinol. Values are the mean & SEM. There were no significant differences between cell sizes. C: Representative case demonstrates the potential for recovery after retinol treatment. The chondrogenic capacity is expressed in the number of chondrogenic clusters. NC, untreated; ETOH, ethanol equivalent volumes (4weeks treatment); ROL, 10 ng retinol (4 weeks treatment); ROL2, 10 ng retinol (2 weeks treatment followed by 2 weeks no treatment). Values are the mean f SEM. NC > ROL (P 5 0.01); ETOH > ROL (P4 0.01); NC, > ROL2 (P 5 0.01), ROL2 > ROL (not significant). The amount of alcian blue staining material surrounding individual cells was quantitated to determine if differences between untreated and treated cultures existed. Morphometric analysis indicated that individual cells in cultures treated with 10 ng retinol produce significantly less alcian blue matrix (mm2)when compared to normal (untreated) controls and ethanoltreated cultures (Fig. 2A). The cell size was not affected by retinol treatment (Fig. 2B). The potential for cultures to recover from exposure to retinol was evaluated by quantitating alcian blue associated cell clusters as an expression of chondrogenesis. It was found that untreated control cultures and ethanol-treated cultures (4weeks treatment) had significantly more chondrogenic clusters than the retinoltreated cultures (4 weeks treatment). Moreover, cultures that were treated with retinol for 2 weeks followed by 2 weeks of no treatment had significantly less chondrogenic clusters than sister cultures that were untreated or were treated with ethanol for 4 weeks. However, these cultures had more chondrogenic clusters than cultures that had been treated continuously for 4 weeks, suggesting a potential for recovery in vitro (Fig. 2C). lmmunocytochemistry Four week cultures were analyzed for the presence of type I and I1 collagens. Retinol-treated cultures had a strong label for type I procollagen (Fig. 3E). In contrast, both untreated normal control cultures (Fig. 3A) and ethanol-treated cultures (Figure 3C) had no detectable immunofluorescence for type I procollagen. While type I1 collagen is readily detectable by 4 weeks in normal untreated cultures (Fig. 4A) it is also present in ethanol-treated cultures (Fig. 4C). Cultures treated with retinol, however, showed a marked decrease in fluorescent labeling (Fig. 4E). 56 A.L. AULTHOUSE ET AL. Fig. 3. Immunofluorescent detection of type I procollagen in human chondrocytes cultured for 4 weeks in agarose. A. Untreated control culture. Cell cluster (location indicated by arrowhead) is negative for type I procollagen. B Phase contrast of A. C: Culture treated with absolute ethanol (equivalent volume given in (El. Cell cluster (location indicated by arrowhead) is negative for type I procollagen. D: Phase contrast of C. E: Culture treated with 10 ng/ml all-trans retinol. Cell cluster (arrowhead) is positive for type I procollagen. F: Phase contrast of E. Cell clusters are typically surrounded by alcian blue staining matrix. The refractile ring seen in the phase contrast photographs corresponds to the outer boundaries of alcian blue matrix. A-F, x 146. Proteoglycan Analysis gel chromatography, i t was found that the size of glycosaminoglycan chains was not affected by retinol (Fig. 5B). There were no detectable differences between ethanol-treated and untreated normal controls in the size of proteoglycan monomers or glycosminoglycan side chains. Analysis of disaccharide composition by paper chromatography showed the ratio of chondroitin- It was found using Sepharose 2B-CL gel chromatography, that the size of the proteoglycan monomer was not affected by retinol treatment. However, the quantity of labeled monomers was reduced (Fig. 5A). The size of the glycosaminoglycan side chains was also determined. With beta elimination and Sephacryl S-300 INFLUENCE O F RETINOL ON HUMAN CHONDROCYTES 57 Fig. 4. Immunofluorescent detection of type I1 collagen in human chondrocytes cultured for 4 weeks in agarose. A Untreated control culture. Cell cluster (arrowhead) is positive for type I1 collagen. B: Phase contrast of A. C: Culture treated with absolute ethanol (equivalent volume given in El. Cell cluster (arrowhead) is positive for type I1 collagen. D: Phase contrast of C. E: Culture treated with 10 ngiml all-trans retinol. Cell cluster (location indicated by arrowhead) has a reduced staining intensity when compared to A and C . F: Phase contrast of E. A-F, x 146. 6-sulfate to chondroitin-4-sulfate to be approximately 1:3 in both untreated and retinol-treated cultures. The proteoglycan composition was identical in both the media and agarose. aggregate into the precartilage blastema and in the second phase, a cartilage matrix is elaborated which is replaced by a calcified bone matrix during endochrondral bone formation. Possible mechanisms for the teratogenesis of retinoids may relate to defects in the formation of the precartilage condensation and/or subsequent defects in chondrocyte growth and elaboration of matrix and/or enhanced matrix degradation. DISCUSSION Appendicular skeletal development may be divided into two phases. In the first phase, mesenchymal cells 58 A.L. AULTH()USE ET AL. Sapharose 2B-CL i VO 2oT i . O091 16 KO,= 0-0 0 10 0.4 10 ng Rol 20 30 40 50 70 60 80 90 100 Ffaction weight (9) A Sephocryl S-300 7 4 0-0 E KO,= 0 ng Rol @-0 10 ng 0.3 Rol * 2t @C) & 1 @ & 0 - 0 10 20 30 40 50 60 70 80 I 90 100 Fraction weight (4) 6 Fig. 5. Size exclusion chromatography of Na2?3O, labeled proteoglycans extracted from human chondrocyte cultures. Untreated cultures, 0 ng ROL (open circles), and cultures treated with retinol for 3 weeks, 10 ng ROL (solid circles). There were no differences between untreated and ethanol-treated cultures. In addition there were no differences between proteoglycans obtained from the media and the agarose. The graphs show the proteoglycans extracted from agarose. A: The hydrodynamic size of proteoglycan monomers is the same in both untreated and retinol-treated cultures. There are less labeled proteoglycans with retinol treatment. B: The size of the glycosaminoglycan side chains of the retinol-treated cultures is the same as in the untreated cultures. Animal studies using mice and hamsters have indicated that retinoic acid induced teratogenesis occurs during organgogenesis suggesting that blastema formation is unaffected (Pillans et al., 1988; Willhite e t al., 1986). Similarly, evaluation of clinical cases demonstrates that in humans teratogenesis occurs in organogenesis and not in initial stages of blastema formation (Willhite et al., 1986). The possible degradative effects of retinoids on cartilage matrix have also been evaluated. Initial investigations suggested that retinoids caused degradation of cartilage matrix (McElligott, 1962) probably through lysosomal proteolytic enzymes (Fell and Dingle, 1963; Lucy et al., 1961). However, i t has since been shown using membrane stabilizers and proteolytic inhibitors that retinoids acted by interfering with matrix synthesis (Shapiro and Poon, 1976; Solursh and Meier, 1973). Recently, in vitro studies have shown that retinoids may cause a decrease in andlor a modification of the synthesis of cartilage specific matrix macromolecules. In other animal studies, a decrease in proteoglycan synthesis was reported in chick sternal chondrocytes (Horton e t al., 1987; Solursh and Meier, 1973) and also in rabbit articular chondrocytes (Benya and Padilla, 1986). Similarly, in cultured rabbit chondrocytes type I1 collagen synthesis decreased while type 111collagen and fibronectin synthesis increased (Benya and Padilla, 1986). Type I1 collagen synthesis decreased while fibronectin and type I collagen increased in chick cultured chondrocytes (Horton et al., 1987). In addition, Horton et al. (1987) reported that alterations in the matrix components seen in cultured chick sternal chondrocytes were correlated with changes in specific mRNA transcripts. Nuclear runoff experiments indicated changes in the transcriptional activities of the collagen I1 and collagen I11 genes following retinoic acid treatment. From these data it was surmized that retinoic acid produces a change in chondrocyte phenotype by altering the pattern of gene expression. Our data support the findings of previous in vitro animal studies which concluded that chondrocytes exposed to retinol express a hybrid phenotype. The cartilage specific matrix macromolecules are expressed in a reduced fashion while type I collagen synthesis was enhanced. The expression of type I collagen suggests that the cells may be maintained in a modulated phenotype. However, the continued synthesis of cartilage specific proteoglycans and type I1 collagen by these chondrocytes suggests that the expression of a cartilage specific matrix is possible. It has been shown by von der Mark e t al. (1977) that both types I and I1 collagen may be expressed simultaneously. It is noteworthy in our system that a n alcian blue staining matrix was associated with cells t h a t stained positive for type I procollagen or type I1 collagen. Some in vitro studies have indicated that retinoids maintain chondrocytes in a mesenchymal-like state. Alcian blue staining nodules are not produced in micromass cultures from chick (Hassell et al., 1987; Pacifici et al., 1980), mouse (Hassell et al., 1978; Lewis et al., 1987; Pennypacker et al., 19781, or rat (Gallandre and Kistler, 1980) limb buds and the cells are maintained in a mesenchymal morphology, i.e., flat and stellate. It has been further demonstrated that retinoic acid maintains the synthesis of the mesenchymal glycoprotein, fibronectin, in both chick (Hassell et al., 1987) and mouse (Lewis et al., 1987) chondrocyte cultures. This does not seem to be the case with human chondrocytes cultured under these conditions because of the continued production of cartilage specific matrices. It is possible that this alteration of matrix (reduced cartilage specific matrix macromolecules and increased type I collagen) coupled with a decrease in the size and number of isogenous groups may modify the establishment of a normal cartilaginous template. This culture system may serve as a model system to further evaluate the importance of cartilage in appendicular skele- INFLUENCE OF RETINOL ON HUMAN CHONDROCYTES togenesis. Indeed retinol and other retinoids may elicit limb malformations by interfering with this early stage in the development of the appendicular skeleton. ACKNOWLEDGMENTS We thank Dr. Sarah Johnson and colleagues for their help in acquiring tissue, Mr. Ben Han for photographic assistance, and Dr. Daniel L. Feeback for editorial comments. This research is supported in part by research grants; OCAST grant HN9-019, OU College of Medicine Alumni Association, Presbyterian Harris Research Foundation, and Presbyterian Health Foundation #85. LITERATURE CITED Aulthouse A.L., M. Beck, E. Griffey, J . Sanford, K. Arden, M. Machado, and W.A. Horton 1989 Expression of human chondrocyte phenotypes in vitro. In vitro Cell Dev. Biol., 25:659-668. Aydelotte, M.B., R. Schleyerbach, B.J. Zeck, and K.E. Kuettner 1986 Articular chondrocytes cultured in agarose gel for study of chondrocytic chondrolysis. In: Articular Cartilage Biochemistry. K.E. Kuettner, R. Schleyerbach, and V.C. Hascall, eds. Raven Press, New York, pp. 235-256. Barnhart, E.R. 1989 Physicians’ Desk Reference. 43rd Edition. Medical Economics Co., Inc., New Jersey, pp. 1711-1716. Beck, M., K. Lingnau, and J . Spranger 1988 Newly synthesized proteoglycans in pseudoachondroplasia. Bone, 9:367-370. Benya, P.D., and S.R. Padilla 1986 Modulation of the rabbit chondrocyte phenotype by retinoic acid terminates type I1 collagen synthesis without inducing type I collagen: the modulated phenotype differs from that produced by subculture. Dev. Biol., 118:296305. Fell, H.B., and J.T. Dingle 1963 Studies on the mode of action of excess vitamin A. 6. Lysosomal protease and the degradation of cartilage matrix. Biochem. J., 87:403-408. Gallandre, F., and A. Kistler 1980 Inhibition and reversion of chondrogenesis by retinoic acid in rat limb bud cell cultures. Wilhelm Roux’s Arch., 189t25-33. Grote, W., D. Harms, U. Janig, H. Kietzman, U. Ravens, and I. Schwarze 1985 Malformations of fetus conceived 4 months after termination of maternal etretinate treatment. Lancet, It1276. Happle, R., H. Traupe, Y. Bounameaux, and T. Fisch 1984 Teratogene Wirkung von Etretinat beim Menschen. Dtsch. Med. Wochenschr., 109:1476-1480. Hassell, J.R., J.P. Pennypacker, and C.A. Lewis 1978 Chondrogenesis and cell proliferation in limb bud cell cultures treated with cytosine arabinoside and vitamin A. Exp. Cell Res., 112t409-417. Hassell, J.R., J.P. Pennypacker, K.M. Yamada, and R.M. Pratt 1987 Changes in cell surface proteins during normal and vitamin A inhibited chondrogenesis in vitro. Ann. N.Y. Acad. Sci. USA, 312: 406-409. Holmdahl, R., K. Rubin, L. Klareskog, E. Larsson, and H. Wigzell 1986 Characterization of the antibody response in mice with type I1 collagen-induced arthritis, using monoclonal anti-type I1 collagen antibodies. Arthritis Rheum., 29r400-410. Horton, W.E., Y. Yamada, and J.R. Hassell 1987 Retinoic acid rapidly reduces cartilage matrix synthesis by altering gene transcription of chondrocytes. Dev. Biol., 123t508-516. Kochhar, D.M. 1973 Limb development in mouse embryos I. Analysis of teratogenic effects of retinoic acid. Teratology, 7r289-298. Lammer, E.J., D.T. Chen, R.M. Hoar, N.D. Agnish, P.J. Benke, J.T. Braun, C.J. Curry, P.M. Fernhaff, A.W. Grix, I.T. Lott, J.M. 59 Richard, and S.C. Sun 1985 Retinoic acid embryopathy. New Engl. J. Med., 313:837-841. Lewis, C.A., R.M. Pratt, J.P. Pennypacker, and J.R. Hassell 1987 Inhibition of limb chondrogenesis in vitro by vitamin A: alterations in cell surface characteristics. Dev. Biol., 64:31-47. Lucy, J.A., J.T. Dingle, and H.B. Fell 1961 Studies on the mode of action of excess of vitamin A. 2. A possible role of intracellular proteases in the degradation of cartilage matrix. Biochem. J., 79:500-508. Maden, M. 1982 Vitamin A and pattern formation in the regenerating limb. Nature, 295t672-675. Maden, M. 1983 The effect of vitamin A on the regenerating axolotl limb. J. Embryol. Exp. Morphol., 77:273-295. McBridge, W.G. 1985 Limb reduction deformaties in child exposed to isotretinoin in utero on gestational days 26-40 only. Lancet, 1: 1276. McDonald, J.A., T.J. Broekelmann, M.L. Matheke, E. Crouch, M. Koo, and C. Kuhn 1986 A monoclonal antibody to the carboxyterminal domain of procollagen type I visualizes collagen-synthesizing fibroblasts. Detection of a n altered fibroblast phenotype in lungs of patients with pulmonary fibrosis. J. Clin. Invest., 78:1237-1244. McElligott, T.F. 1962 Decreased fixation of sulfate by chondrocytes in hypervitaminosis A. J . Pathol., 83:347-355. Noji, S., T. Nohno, E. Koyama, K. Muto, K. Ohyama, Y. Aoki, K. Tamura, K. Ohsugi, H. Ide, S. Taniguchi, and T. Saito 1991 Retinoic acid induces polarizing activity but is unlikely to be a morphogen in the chick limb bud. Nature, 350t83-86. Pacifici, M., G. Cossu, M. Molinaro, and F. Tat0 1980 Vitamin A inhibits chondrogenesis but not myogenesis. Exp. Cell Res., 129: 469-474. Pennypacker, J.P., C.A. Lewis, and J.R. Hassell 1978 Altered proteoglycan metabolism in mouse limb mesenchyme cell cultures treated with vitamin A. Arch. Biochem. Biophys., 186:351-358. Pillans, P.I., P.I. Folb, and S.F. Ponzi 1988 The effects of in vivo administration of teratogenic doses of vitamin A during the preimplantation period in the mouse. Teratology 37r7-11. Shapiro, S.S., and J.P. Poon 1976 Effect of retinoic acid on chondrocyte glycosaminoglycan biocynthesis. Arch. Biochem. Biophys., 174: 74-81. Slack, J.M.W. 1987 We have a morphogen! Nature, 327553-554. Solursh, M., and S. Meier 1973 The selective inhibition of mucopolysaccharide synthesis by vitamin A treatment of cultured chick embryo chondrocytes. Calcif. Tiss Res., 13: 131-142. Takakubo, F., and K. Eto 1986 Stage specific responses of the mesenchyme to excess vitamin A in developing rat facial processes. J. Craniofac. Genet. Dev. Biol., 2:179-185 (Suppl). Tickle, C., B. Alberts, L. Wolpert, and J. Lee 1982 Local application of retinoic acid to the limb bud mimics the action of the polarizing region. Nature, 2 9 6 5 6 4 4 6 6 . Tickle, C., J. Lee, and D. Eichele 1985 A quantitative analysis of the effect of all-trans-retinoic acid on pattern of chick wing development. Dev. Biol., 109:82-95. Truppe, W., and H. Kresse 1978 Uptake of proteoglycans and sulfated glycosaminoglycans by cultured skin fibroblasts. Eur. J. Biochem., 85:351-356. Von der Mark. K., V. Grauss. H. Von der Mark. and P. Muller 1977 Relationship between cell shape and type ofcollagen synthesised as chondrocytes lose their cartilage phenotype in culture. Nature, 267:531-532. Wanek, N., D.M. Gardiner, K. Muneoka, and S.V. Bryant 1991 Conversion by retinoic acid of anterior cells into ZPA cells in the chick wing bud. Nature, 350%-83. Willhite, C.C., R.M. Hill, and D.W. Irving 1986 Isotretinoin-induced craniofacial malformations in humans and hamsters. J . Craniofac. Genet. Dev. Biol., 2:193-209 (Suppl).