Matrix Vol. 13/1993, pp. 235 - 241 © 1993 by Gustav Fischer Verlag, Stuttgart· Jena . New York

Influence of Retinoic Acid and TGF-~ on Dermal Fibroblast Proliferation and Collagen Production in Monolayer Cultures and Dermal Equivalents

J. K. JUTLEy1, E. J. WOOD1 and W. J. CUNLlFFE2

1 Department of Biochemistry and Molecular Biology University of Leeds, Leeds LS2 9 JT, UK and 2 Department of Dermatology Leeds GeneralInfirmary, Leeds LS13EX, UK.

Abstract

The effects of transforming growth factor-~l (TGF-~l) and all-trans retinoic acid (RA) on human dermal fibroblast proliferation and collagen production were investigated in 'attached' and 'detached' dermal equivalent (DEs) systems compared with fibroblasts grown in monolay­ers. The combined effects ofTGF-~l with epidermal growth factor (EGF) on fibroblast prolifera­tion and collagen production were also studied. Fibroblast proliferation was stimulated O.7-fold by TGF-~l in attached DEs and O.3-fold in detached DEs compared with untreated control DEs. RA stimulated fibroblast proliferation 1.1-fold in attached DEs and O.6-fold in detached DEs. Neither TGF-~l nor RA had a significant effect on fibroblast proliferation in monolayer cultures. In the presence of EGF, the action of TGF-~l on fibroblast proliferation was slightly suppressed in attached DEs. At 5 ng/ml TGF-~l, collagen production was stimulated 6.1-fold in attached DEs, by 3.5-fold in monolayer at 4 ng/ml and 2.9-fold in detached DEs at 1 ng/ml. RA at 5 X 10-10 M to 5 X 10-6 M stimulated collagen production in all three systems. The stimulation of collagen production by 5 ng/ml TGF-~l was suppressed by EGF in both attached DEs and monolayer culture. Fibroblasts in attached DEs had elongated, bipolar morphology with a tendency to line up in the same direction. Fibroblasts in detached DEs were randomly distributed and exhibited stellate morphology. These data indicate that the extracellular matrix strongly influences the actions of growth factors and of RA on dermal fibroblasts. When stimulated with TGF-~l collagen production in attached DEs was higher than that by fibroblasts in monolayer culture and detached DEs. The attached and detached DEs offer a model which resembles the in vivo situation more closely than monolayer culture with respect to collagen production and fi­broblast proliferation and morphology.

Key words: collagen, dermal equivalents, fibroblasts, retinoic acid, transforming growth factor-~.

Introduction

During the process of wound repair an extensive remod­elling of the extracellular matrix takes place during which structural materials such as collagens are both synthesised and degraded. The cells predominantly responsible for this remodelling are the fibroblasts and in normal dermis these are distributed evenly throughout the tissue and form a fairly static population. However, following wounding

they migrate from adjacent sites in response to chemical signals and undergo a localised proliferation (Irvin, 1981; Clark, 1985). They are known to synthesise both collagen and collagenase at different times in the wound healing process. Ultimately the remodelled granulation tissue fills the wound and contracts it, and the expanded fibroblast population regresses (Irvin, 1981; Clark, 1985).

Growth factors such as transforming growth factor-beta (TGF-~), epidermal growth factor (EGF), platelet-derived

236 J. K. Jutley et al.

growth factor (PDGF), fibroblast growth factor (FGF) are essential throughout the wound healing process and the actions of growth factors are reported to be modulated by the extracellular matrix (Skerrett, 1991). Many of these growth factors are recruited from the blood or are produced by platelets or by cells such as epidermal keratinocytes. One of the factors believed to be of major importance in the repair of tissue injury is transforming grown factor-beta (TGF-~). This pleiotropic factor stimulates extracellular matrix formation and inhibits its degradation. TGF-~ is also a potent chemotactic agent for fibroblasts and has been shown to cause a concentration-dependent increase in the expression of the mRNA for the a1(I} chain of type! colla­gen (Fine et ai., 1990). Any response of fibroblasts to growth factors is dependent on receptors to these growth factors being expressed on their surfaces.

Although the behaviour of fibroblasts in response to growth factors may be studied with monolayer cultures of these cells there is increasing evidence that such a system does not represent a very satisfactory model system. Fib­roblasts in monolayer culture do not behave as they do in a collagenous matrix (Bell et ai., 1979; Bellows et ai., 1981; Guidry and Grinnell, 1985; Nakagawa et ai., 1989 a). For example, fibroblasts synthesise less collagen in collagen gels than they do in monolayer cultures (Gillery et ai., 1989; Hey et ai., 1990; Lambert et ai., 1992). In addition, fibro­blasts in collagen gels under tension synthesise more colla­gen and other extracellular matrix proteins than fibroblasts in free floating gels (Lambert et ai., 1992). Collagenase activity was increased in collagen gels compared with monolayer cultures (Lambert et ai., 1992). Reduced levels of procollagen mRNA in fibroblasts in collagen gels have been attributed to the retraction of lattices and the nature of the fibrillar protein within the collagen gels (Gillery et ai., 1992). Fibroblast function is modulated by self-regulating mechanical forces in collagen gels and the regulation occurs at a pretranslationallevel (Lambert et aI., 1992). This has led several groups to develop the dermal equivalent system in which fibroblasts are incorporated into a collagen gel which, at least to a limited extent, reproduces the environ­ment in, and the connective tissue contraction that occurs, during wound repair (Bell et ai., 1979; Bellows et ai., 1981; Grinnell and Lanke, 1984; Nakagawa et aI., 1989 a; Schor, 1980). Not only are fibroblast proliferation and collagen production controlled by interactions with the extracellular matrix, but also the effects of the various growth factors are likely to be interactive.

In the present work we have investigated the influence of TGF-~ on fibroblast proliferation and collagen production in two different dermal equivalent systems, the so-called 'attached' and 'detached' systems, in comparison with its influence on fibroblast monolayers. Epidermal growth fac­tor (EGF) is also important in wound healing and EGF/ TGF-a receptors are known to be expressed by fibroblasts as well as by keratinocytes (Schultz et aI., 1991). The effect

of TGF-~ in the presence of EGF on fibroblast proliferation and collagen synthesis was therefore also investigated. In addition, we have studied the effects of all-trans-retinoic acid (RA) in these systems. Retinoids are known to regulate the proliferative and differentiative capacities of several mammalian cell types, and the possibility has been raised that some of the retinoid-mediated effects could be accounted for by the production of endogenous TGF-~ in active form (Poli et aI., 1992). TGF-~ is usually secreted in a latent form and treatment with retinoids results in the production of active TGF-~ by some cells such as keratinocytes (Glick et ai., 1989). RA also influences extracellular matrix production and has been used to treat keloids (Oikarinen et ai., 1985). It is presently being used topically as a cosmetic treatment for photoinduced ageing of the skin, where it is reported to promote the formation of new dermal collagen (Schwartz et aI., 1991).

Materials and Methods

Fibroblasts

Human skin fibroblast cultures were established from foreskin samples obtained at routine circumcision. After the separation of epidermis and dermis, fibroblasts were released from the dermis by collagenase treatment and grown in monolayer culture on Dulbecco's modified Eag­le's medium (DMEM) containing 10% newborn calf serum (NBCS) at37°Cin a 5% CO2 : 95% air atmosphere. Exper­iments were performed on (adult, 22-40yrs) cell lines A151, A146 and A164 at passage numbers 4-12.

Dermal equivalents

Dermal equivalents (DEs) were prepared using acetic acid-extracted rat tail collagen essentially according to Bell et ai. (1979). Fibroblasts were mixed with neutralized colla­gen in DMEM (Gibco-BRL, Paisley, UK) at a seeding density of 5 x 104 cells/ml. The final concentration of the collagen was 0.52 mg/ml. The collagen-fibroblast mixture was rapidly dispensed into 24-well tissue culture plates (Nunc, Paisley, UK) in aliquots of 0.2 ml per well, and allowed to polymerise to form a gel. After 60 min at 37°C, 1 ml of DMEM supplemented with 10% NBCS and 50 Ilg/ ml ascorbic acid was added to each well. After polymerisa­tion the DEs were either allowed to remain attached to the wells ("attached DEs") or were lifted off the bottoms of the wells by ringing with a sterile spatula and allowed to float ("detached DEs"). Monolayer cultures were also prepared with fibroblasts at the same passage number using 7.5 X

104 cells/mVwell in 24-well plates. Attached and detached DEs were cultured for two days, and monolayers for one day, after which the medium was changed to DMEM con­taining a reduced amount of NBCS (2%) and 50Ilg/ml ascorbic acid. After a further 24 h, this was replaced with

DMEM containing 2% NBCS, 50 !lg/ml ascorbic acid, L­[2,3 -3H] proline (NEN, Stevenage, UK; 8.3 !lCi/ml) and either 1-10 ng/ml TGF-~1 (British Biotechnology Ltd., Oxford, UK), 10-1O_10-4 M all-trans-retinoic acid (Sigma, Poole, Dorset, UK) or 1-10 ng/ml TGF-~1 plus 4 ng/ml EGF (Advanced Protein Products Ltd., UK).

Collagen biosynthesis

The amount of collagen synthesis that had taken place was determined according to the method given by Peter­kofsky and Diegelmann (1971); as modified by Postleth­waite et al. (1984) and by Nakagawa et al. (1989 a). Briefly, incorporation of tritiated proline into collagen was allowed to proceed for 24 h. At the end of the incubation 0.6 ml medium was collected in vials containing 90!l1 of 50 mM N-ethylmaleimide (NEM) and the amount of radioactivity in collagenase sensitive, TCA-precipitable material, deter­mined as follows. To 0.4 ml of each sample of medium was added Tris-calcium acetate (50 mM Tris-HCl, pH 7.2, con­taining 10 mM calcium acetate) to bring the volume to 1 ml. One half of each sample was treated with chromatograph­ically purified collagenase form III from Clostridium his­tolyticum (3 BTC units/sample, Advance Biofactures Corp., Lynbrook, USA) for 3 h at 37°C (Peterkofsky and Diegelmann, 1971). Bovine serum albumin, 0.5 mg/ml, was then added as a carrier and samples were subjected to precipitation and three washes with 10% trichloroacetic acid at 4°C. The precipitates were finally dissolved in Sol­usol (Mensura Technology Ltd., Lancs., UK) and their radioactivity measured by scintillation counting. Collagen synthesis was calculated as total less collagenase-insensi­tive, TCA precipitable radioactivity. Each incorporation experiment was performed in triplicate and the incorpora­tion data were normalized by cell number. The results are presented as means of 3 - 5 assays performed in triplicate.

Polyacrylamide gel electrophoresis

DEs were radiolabeled metabolically for 24 h with 5 !lCi/ ml of [35Sl-methionine (NEN, Stevenage, UK) added in methionine-free DMEM containing TGF-~1 or retinoic acid. The medium was removed and the gels were washed briefly in PBS and sonicated in DMEM containing unlabel­led methionine. Additional samples treated with TGF-~1 or RA were digested with collagenase to verify the identity of PAGE bands as collagen. Electrophoresis was performed using 5% polyacrylamide gels in reducing conditions (Laemmli, 1970). After electrophoresis, gels were fixed and stained with Coomassie brilliant blue R250 and auto­radiographed. Densitometric scans were used to quantify the collagen bands on the auto radiographs.

TGF-~, Retinoids and Fibroblasts 237

Cell numbers

Collagen gels were solubilized on day 5 by treatment with bacterial collagenase (0.25 units/ml) and dispase (2 mg/ml) for 2 h at 37°C. Aliquots of the samples were counted in a haemocytometer. Cells in monolayer culture were harvested on day4 using 0.05% trypsin: 10mM EDTA and aliquots counted similarly. The results pre­sented are means of 3 - 5 assays performed in triplicate. Statistical analysis were performed using Student's t-test.

Results

Influence ofTG F-~l and retinoic acid on fibroblast proliferation

When fibroblasts were grown as monolayers under the conditions described, neither TGF-~ (Fig. 1 a) nor RA (Fig. 1 b) had a very significant influence on fibroblast pro­liferation over the time-scale of these experiments (p > 0.05 compared with DEs). The range of concentration tested were 1-10ngTGF-Wmland5 X 10-1O-1O-4 MRA. The physiological concentrations are believed to be around 5 ng/ml for TGF-~ (Cromack et al., 1987) and around 10-7 M for RA (Kim et al., 1987).

In contrast, both TGF-~ and RA had significant effects on the extent of fibroblast proliferation in DEs. In general, both were stimulatory over the mid-range of concentra­tions, with RA having a slight tendency to be inhibitory at the very highest concentration tested. The influence of TGF-~ was somewhat different in attached compared with detached DEs (Fig. 1 a). In attached DEs, TGF-~ at 2 ng/ml, significantly stimulated proliferation (72 ± 14% above basal levels, i.e. with no added TGF-~) compared with monolayers (p = 0.05). The stimulation was reduced at higher TGF-~ concentrations. However, in detached DEs, proliferation was stimulated at 5 ng/ml (34 ± 19% above basal levels) but the results were not statistically significant (p > 0.05) compared with monolayers or attached DEs. Stimulation was inhibited at higher concentrations tested (above 9 ng/ml). Similarly, RA stimulated proliferation at 5 X 10-10 to 10-6 M in both DE systems, but was inhibited at higher concentrations. In attached DEs, RA stimulated cell proliferation by up to 111 ± 23% above basal levels. RA stimulated proliferation in detached DEs at 5 X 10-10

to 10-6 M (up to 62% above basal levels).

Effect ofTG F-f3 on fibroblast collagen production

There was a striking difference in the response of fibro­blasts to the addition of TGF-~ depending upon their envi­ronment. Thus TGF-~ strongly stimulated collagen produc­tion by fibroblasts in attached DEs (Fig. 2). In attached DEs the maximal effect (610 ± 19% above basal level, i.e. with no added TGF-~) occurred at 5 ng TGF-Wml (p = 0.02,

238 J. K. Jutley et al.

A

'2 (a) 1: zoo 8 '5 160 'G-c

.g IZO IL ..... ::;.;..--r--..R::...-~ ~ '2 80 Q.

~ c 40 :0 o .. £ O~--.----r---r---.---'

Z 4 6 8 m [TGF -IJl ( ngl mil

Fig. 1 a. Effect of TGF-~ on fibroblast proliferation with cell line A151. TGF-~ 1-10ng/ml was added 24h prior to assay. Control samples contained no TGF-~. 0, attached DE; e, detached DE; 0, monolayer culture. Each point represents the mean of five experi­ments performed in triplicate (± S. E.).

'0 ( b) .. Z40 ..

c 0 u ... ZOO 0

'!:. c 160 .2 .. c .. .. IZO .'!: '0 .. Q.

80 .. til C :0 0 40

.Q u:

0 lerO 10-9 10'8 10-7 10.6 10-5 10.4

B [All-trans rctinoic acid] 1M)

Fib. 1 b. Effect of RA on fibroblast proliferation with cell line A151.RA5 x to-IOt05 x to-4 Mwas added 24 hprior to assay. Control samples contained no RA. 0, attached DE; e, detached DE; 0, monolayer culture. Each point represents the mean of three experiments performed in triplicate (± S.E.).

compared with monolayers). In monolayers a similar, if less marked, stimulation of collagen production was observed (346 ± 16% increase on basal level) at 4 ng/ml, but in both cases the stimulatory effect fell off quite sharply at both high and lower TGF-~ concentrations. In contrast, how­ever, detached DEs showed very little response to TGF-~ between 4-10 ng/ml (Fig. 2). There was a significant stimu­lation of collagen production at 1 ng/ml TGF-~ (p < 0.05) compared with monolayers. Thus fibroblasts modulate their response to growth factors such as TGF-j3 depending upon whether they are in contact with collagen fibres that are in a state of tension (attached DEs) or not (detached

800

'0 700 .. ...

c 600 0

u

'5 "6- 500

c .2 400 .. u :;, '0 0

300 .. Q.

C .. go ZOO c '0 u

100

0 Z 4 6 8 10

[TGF-/l]( ng/mlJ

Fig. 2. Effect of TGF-~ on collagen production by fibroblast of cell line 151. TGF-~ 1-l0ng/ml was added 24h prior to assay. 0, attached DE; e, detached DE; 0, monolayer culture. Each point represents the mean of five experiments performed in triplicate. Results are expressed as % of control; and corrected for cell number (± S.E.).

c: (bl .~ 300 -u ..... ::1-'tIO e.!:: 200 Il~ c: u ~ '0 100 ~~ 0 ...... u

,§ t; (al

~~200E=: :::e i1100 .................................... . -0

e~ .Q ......

i.<: 2.6810 [TGF'B).4ng/ml EGF

2 4 6 8 10 (TGF-B)+4ng/ml EGF

Fig. 3. Effect of TGF-~ + EGF on (a) fibroblast proliferation and (b) collagen production by cell line Al51. TGF-~ 1-l0ng/ml was added in the presence of 4ng/ml EGF 24h prior to assay. 0, attached DE; 0, monolayer culture.

DEs). In fact, fibroblasts in attached DEs tend to present a different morphology from those in detached DEs. In attached DEs the cells always showed a bipolar morphol­ogy with a tendency to line up in the same direction. In detached DEs some cells had a stellate morphology and were randomly distributed in every case (data not shown).

Z40

'0 ... .. zoo c 8 ... 0 .,. 160

c .2 IZO -U :::J "0 0 80 ... Q.

C

" C/I 40 ~ "0 u

0 10.10 10.9 10·S 10-7 10-6 10-5 10-4

l AII- trans rltinoic acid]( M I

Fig. 4. Effect of RA on collagen production by fibroblast of cell line AlS1. RAS x 1O-10 toS x 1O-4 Mwas added 24h prior to assay. 0, attached DE; ., detached DE; 0, monolayer culture. Each point represents the mean of three experiments performed in tripli­cate. Results are expressed as % of control; and corrected for cell number (± S.E.).

Effect ofTG F-~ in the presence of EG F

The interaction between TGF-~ and EGF in the attached dermal equivalent system, compared with what happens in mono layers was studied. Collagen production and fibro­blast proliferation were investigated using 1-10 ng/ml TGF-~ in the presence of 4ng/ml EGF (Fig. 3). There were only small effects on fibroblast proliferation (Fig. 3 a), but the presence of EGF suppressed collagen production in monolayers (compare Fig. 3 b and Fig. 2). In attached DEs, maximal collagen production was observed at somewhat higher TGF-~ concentrations (7 ng/ml compared with 5 ng/ ml) and overall was suppressed (max. lA-fold stimulation compared with 6. I-fold in the absence ofEGF).

Effect of retinoic acid on fibroblast collagen production

RA stimulated collagen biosynthesis in all three systems at physiological concentrations (Fig. 4). The greatest stimu­lation was observed in monolayers over the concentration range 5 X 10-10-5 X 10-7 M, with maximal increase at 5 X 10-9 M (108 ± 20% above basal level). Fibroblasts in attached and detached DEs showed very little response to stimulation of collagen production by RA compared with monolayer cultures. The highest stimulation of collagen production in attached DEs was 44 ± 21 % (above basal levels) and 35% (above basal levels) in detached DEs.

Fibroblasts from cell line A151 not treated with growth factors (other than those contained in the added newborn calf serum) synthesised less collagen in attached (2696 ± 289 D.P.M.) and detached (2068 ± 165 D.P.M.) DEs than fibroblasts in monolayer cultures (4657 ± 127D.P.M.).

Experiments were repeated with cell lines A146 and

TGF-~, Retinoids and Fibroblasts 239

A164, essentially the same data were obtained for both TGF-~ and RA.

Polyacrylamide gel electrophoresis

Densitometric scanning of SDS-PAGE autoradiographs revealed no preferential stimulation of synthesis of ul(I) or u2(I) chains.

Discussion

From the data presented above, it is clear that the physi­cal and chemical nature of the extracellular matrix strongly influences the behaviour of fibroblasts. One of the reasons may be that because TGF-~ (and other growth factors) adhere to collagen (Nakagawa et al., 1989 a), their action is in some way controlled or potentiated, or that growth factor receptors are expressed on fibroblasts in vitro but not in vivo (Terracio et al., 1988). However, these are not the only factors, because the response of the fibroblasts also depends on whether the collagen gel in which they find themselves is under tension (attached DEs) or whether they are allowed to contract the gel, so removing the tension (detached or floating DEs). Using attached and detached DEs, Nakagawa et al. (1989a) reported that fibroblasts in attached gels had an elongated, bipolar morphology and that the collagen fibrils were aligned in the plane of cell spreading. They also reported that collagen fibrils in detached gels were randomly arranged and that the fibro­blasts had stellate morphology: our observations are con­sistent with this.

The action of fibroblasts on contracting the collagen gel may have a number of consequences. The fact that the gel shrinks to a small fraction (i.e. 10-30%) of its initial volume means that not only are the cells brought closer together and so may start to experience contact inhibition, but also they may be brought closer to TGF-~ molecules adhering to the collagen, so increasing the effective concen­tration of this growth factor. The difference in collagen production between detached and attached gels may be partially explained by increased collagenase activity observed in detached DEs (Hey et al., 1990).

Fibroblast proliferation was stimulated in both attached and detached DEs compared with monolayers, but at dif­ferent concentrations of TGF-~. Similarly TGF-~ at 5 ng/ml greatly stimulated collagen production in attached DEs and to a slightly less extent, in monolayer cultures at 4 ng/ml. However, at these concentrations, TGF-~ had very little effect on collagen production in detached DEs. In detached DEs, maximal stimulation was observed at 1 ng/ml. After 5 days in culture, fibroblasts in attached DEs secreted colla­gen into medium and incorporated collagen into the matrix (data not shown). The response to TGF-~ by dermal fibro­blasts is dose dependent. Our data are in agreement with

240 J. K. Jutley et al.

previous in vitro and in vivo studies on dermal fibroblasts (Nakagawa et aI., 1989 a; Raghow et aI., 1987) and human embryonic lung fibroblasts (Fine et aI., 1990). Mauch et aI. (1988) reported a reduction in collagen synthesis by fibro­blasts in 'floating' (detached) DEs to 5% of the value found in monolayer cultures. Our data on levels of collagen syn­thesis in detached DEs are consistent with those of Nakagawa et aI. (1989a). This reduction was shown to be due to reduction in biosynthesis of pro-al (I), pro-a2(1) and pro-a 1 (III) collagen mRNAs (Mauch et aI., 1988).

Collagen production was stimulated to a small extent in both attached and detached DEs. SDS-PAGE autoradio­graphs of medium showed that fibroblasts in attached DEs secreted relatively more collagen into the medium than did fibroblasts in detached DEs. Collagen was also incorpo­rated into the matrix (data not shown). In monolayer cul­tures, stimulation was observed between 5 X 10-10 M and 5 X 10-7 M RA. Varani et aI. (1990) reported RA stimula­tion of type I collagen production in monolayer cultures at 1. 7 X 10-6 M, but at higher concentrations, collagen pro­duction was inhibited. However, in contrast, Oikarinen et aI. (1985) reported a dose-dependent inhibition in pro colla­gen production in human fibroblasts in culture. The reduc­tion in procollagen production in RA-treated cultures was accompanied by a marked reduction in the pro-a 2 (I) specific mRNA levels. Fibroblast proliferation in our study was stimulated in both attached and detached DEs at phys­iological concentrations of RA. However, RA at physiolog­ical concentrations had little effect on fibroblast prolifera­tion in monolayer cultures. Previous authors reported stimulation of proliferation in monolayer cultures at 1.7-6.6 X 10-6 M RA (Varani et aI., 1990). Our study suggests that fibroblast proliferation and collagen produc­tion are stimulated in dermal equivalents in vitro by RA and that these changes are similar to those observed in the dermis after the topical application of retinoids (Schwartz et aI., 1991). Thus as well as serving as a wound healing model, the DE system may be useful in evaluating these actions of RA.

EGF had little effect on the stimulation of fibroblast proliferation by TGF-~. In attached DEs, stimulation of collagen production by TGF-~ at 5 ng/ml was abolished by EGF. The effect was absent at higher TGF-~ concentra­tions. However, in monolayer cultures, the 3.5-fold stimu­lation was inhibited by 4 ng/ml EGF and the effect was not reversed at higher TGF-~ concentrations. Similar findings were reported by Roberts et aI. (1986) who showed that collagen production stimulated by 50 pM TGF-~ was inhi­bited by the addition of EGF at 40 (by 20%) and 200 pM (by 110%). They also reported that these antagonistic effects could not be ascribed to protease or collagenase activation by EGF and that the inhibition was not due to effects on amino acid uptake.

Wound healing has evolved to be a complex process involving interactions between growth factors, cells and the

extracellular matrix. The present data with attached and detached DEs offer a highly simplified model system in which some of the interactions may be studied and such models are worthy of study and of further development. It is possible to place keratinocytes on the DEs to produce a so-called 'living skin equivalent' (Bell et aI., 1979) and this may provide an attractive model for studying keratinocyte­fibroblast interactions.

Acknowledgements

We wish to thank Mrs Mollie Depledge and Mr A. Russell for their skilled technical assistance. This work was supported by the Leeds Foundation for Dermatological Research.

References

Bell, E., Ivarsson, B. and Merrill, c.: Production of a tissue-like structure by contraction of collagen lattices by human fibro­blasts of different proliferative potential in vitro. Proc. Natl. Acad. Sci. USA 76: 1274-1278, 1979.

Bellows, C. G., Mekher, A.H. and Aubin,J.E.: Contraction and organization of collagen gels by cells cultured from periodontal ligament, gingiva and bone suggest functional differences between cell types.]. Cell Sci. 50: 299-314,1981.

Clark, R.A.F.: Cutaneous tissue repair: Basic biologic considera­tions. 1.]. Am. Acad. Dermatol. 13: 701-725,1985.

Cromack,D.T., Sporn,M.B., Roberts,A.B., Merino,M]., Dart, L. L. and Norton, J. A.: Transforming growth factor beta levels in rat wound chambers.]. Surg. Res. 42: 622-628,1987.

Fine,A., Poliks,C.F., Smith,B.D. and Goldstein,R.H.: The accumulation of type I collagen mRNAs in human embryonic lung fibroblasts stimulated by transforming growth factor-~. Connective Tissue Res. 24: 237-247,1990.

Gillery, P., Bellon, G., Coustry, F. and Borel,J.-P.: Cultures of fi­broblasts in fibrin lattices: models for the study of metabolic activities of the cells in physiological conditions.]. Cell Physiol. 140: 483-490,1989.

Gillery,P., Leperre,A., Coustry,F., Maquart, F.-X. and Borel,J.­P.: Different regulation of collagen I gene transcription in three­dimensional lattice cultures. FEBS 296: 297-299, 1992.

Glick,A.B., Flanders,K.C., Danielpour,D., Yuspa,S.H. and Sporn, M. B.: Retinoic acid induces transforming growth factor -~2 in cultured keratinocytes and mouse epidermis. Cell Regul. 1: 87-97,1989.

Grinnell, F. and Lanke, F. R.: Reorganization of hydrated collagen lattices by human skin fibroblasts.]. Cell Sci. 66: 51-63,1984.

Guidry, C. and Grinnell, F.: Studies on the mechanism of hydrated collagen gel reorganisation by human skin fibroblasts.]. Cell Sci. 79: 67-81,1985.

Hey, K. B., Jutley,]. K., Cunliffe, W.J. and Wood, E.].: Growth factors modulate collagen production and collagenase action by skin fibroblasts in a dermal equivalent model system. Biochem. Soc. Trans. 18: 899,1990.

Irvin, T. T.: Wound Healing. In: Principles and Practice, ed. by Irvin, T. T., Chapman and Hall, London and New York, 1981.

Kim,K.H., Stellmach, V., Javors,J. and Fuchs, E.: Regulation of human mesothelial cell differentiation: Opposing roles of retinoids and epidermal growth factor in the expression of intermediate filament proteins.]. Cell Bioi. 105: 3039-3051, 1987.

Laemmli, U. K.: Cleavage of structural proteins during the assem­bly of the head of bacteriophage T4. Nature 227: 680-685, 1970.

Lambert, C. A., Soudant, E. P., Nusgens, B. V. and Lapiere, C. M.: Pretranslational regulation of extracellular matrix mac­romolecules and collagenase expression in fibroblasts by mechanical forces. Lab. Invest. 66: 444-451,1992.

Mauch, c., Hatamochi, A., Scharffetter, K. and Krieg, T.: Regula­tion of collagen synthesis in fibroblasts within a three-dimen­sional collagen gel. Exp. Cell Res. 178: 493-503,1988.

Nakagawa, S., Pawelek, P. and Grinnell, F.: Extracellular matrix organization modulates fibroblast growth and growth factor responsiveness. Exp. Cell Res. 182:572-582, 1989a.

Nakagawa, S., Pawelek, P. and Grinnell, F.: Long-term culture of fibroblasts in contracted collagen gels: Effects on cell growth and biosynthetic activity. ]. Invest. Dermatol. 93: 792-798, 1989b.

Oikarinen, H., Oikarinen, A. I., Tan, E. M. L., Abergel, R. P., Meeker, C.A., Chu,M.-L., Prockop,D.j. and Uitto,J.: Modu­lation of procollagen gene expression by retinoids. Inhibition of collagen production by retinoic acid accompanied by reduced type I pro collagen messenger ribonucleic acid levels in human skin fibroblast cultures.]. Clin. Invest. 75: 1545 -1553,1985.

Peterkofsky, B. and Diegelmann, R.: Use of a mixture of protein­ase-free collagenases for the specific assay of radioactive colla­gen in the presence of other proteins. Biochemistry 10: 988-994,1971.

Poli, G., Kinter, A. L., Justement,j. S., Bressler, P., Kehrl,j. H. and Fauci, A. S.: Retinoic acid mimics transforming growth factor ~ in the regulation of human immunodeficiency virus expression in monocytic cells. Proc. Nat. Acad. Sci. USA 89: 2689-2693, 1992.

Postlethwaite, A. E., Smith, G. N., Mainardi, C. L., Seyer,j. M. and Kang, A. H.: Lymphocyte modulation of fibroblast function in vitro: Stimulation and inhibition of collagen production by different effector molecules. ]. Immunol. 132: 2470-2477, 1984.

TGF-~, Retinoids and Fibroblasts 241

Raghow, R., Postlethwaite, A. G., Keski-Oja,J., Moses, H. L. and Kang, A. H. : Transforming growth factor-~ increases steady state levels of type I procollagen and fibronectin messenger RNAs posttranscriptionally in cultured human dermal fibro­blasts. J. Clin. Invest. 79: 1285 -1288,1987.

Roberts,A.B., Sporn,M.B., Assoian,R.K., Smith,j.M., Roche, N.S., Wakefield,L.M., Heine, U.l., Liotta,L.A., Falanga, V., Kehrl,J. H. and Fauci, A. S.: Transforming growth factor type­~: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc. Natl. Acad. Sci. USA 83: 4167 -4171, 1986.

Schor, S. L.: Cell proliferation and migration on collagen substrata in vitro.]. Cell Sci. 41: 159-175, 1980.

Schultz, G., Rotatoni, D.S. and Clark, W.: EGF and TGF-a in wound healing and repair.]. Cellular Biochem. 45: 346-356, 1991.

Schwartz, E., Cruickshank,F.A., Mezick,j.A. and Kligman, L. H.: Topical all-trans-retinoic acid stimulates collagen synthe­sis in vivo.]. Invest. Dermatol. 96: 975 -978,1991.

Skerrett, P.j.: 'Matrix Algebra' Heals Life's Wounds. Science 252: 1064-1066,1991.

Terracio, L., Ronnstrand, L., Tingstrom, A., Rubin, K., Claesson­Welsh, L., Funa, K. and Heldin, C.-H.: Induction of platelet­derived growth factor receptor expression in smooth muscle cells and fibroblasts upon tissue culturing. ]. Cell Bioi. 107: 1947-1957,1988.

Varani,J., Mitra, R. S., Gibbs, D., Phan, S. H., Dixit, V.M., Mit­ra, R., Wang, T., Siebert, K.J., Nickoloff, B.j. and Voorhees, J.j.: All-trans retinoic acid stimulates growth and extracellular matrix production in growth-inhibited cultured human skin fi­broblasts.]. Invest. Dermatol. 97: 717 -723, 1990.

Dr. J. K. Jutley, Department of Biochemistry and Molecular Biol­ogy, University of Leeds, Leeds LS2 9JT, UK.