Mesenchymal stem cell-conditioned medium accelerates skin wound healing: An in vitro study of fibroblast and keratinocyte scratch assays

We have used in vitro scratch assays to examine the relative contribution of dermal fibroblasts and keratinocytes in the wound repair process and to test the influence of mesenchymal stem cell (MSC) secreted factors on both skin cell types. Scratch assays were established using single cell and co-cultures of L929 fibroblasts and HaCaT keratinocytes, with wound closure monitored via time-lapse microscopy. Both in serum supplemented and serum free conditions, wound closure was faster in L929 fibroblast than HaCaT keratinocyte scratch assays, and in co-culture the L929 fibroblasts lead the way in closing the scratches. MSC-CM generated under serum free conditions significantly enhanced the wound closure rate of both skin cell types separately and in co-culture, whereas conditioned medium from L929 or HaCaT cultures had no significant effect. This enhancement of wound closure in the presence of MSC-CM was due to accelerated cell migration rather than increased cell proliferation. A number of wound healing mediators were identified in MSC-CM, including TGF-β1, the chemokines IL-6, IL-8, MCP-1 and RANTES, and collagen type I, fibronectin, SPARC and IGFBP-7. This study suggests that the trophic activity of MSC may play a role in skin wound closure by affecting both dermal fibroblast and keratinocyte migration, along with a contribution to the formation of extracellular matrix.     

Fetal Fibroblasts and Keratinocytes with Immunosuppressive Properties for Allogeneic Cell-Based Wound Therapy

Fetal skin heals rapidly without scar formation early in gestation, conferring to fetal skin cells a high and unique potential for tissue regeneration and scar management. In this study, we investigated the possibility of using fetal fibroblasts and keratinocytes to stimulate wound repair and regeneration for further allogeneic cell-based therapy development. From a single fetal skin sample, two clinical batches of keratinocytes and fibroblasts were manufactured and characterized. Tolerogenic properties of the fetal cells were investigated by allogeneic PBMC proliferation tests. In addition, the potential advantage of fibroblasts/keratinocytes co-application for wound healing stimulation has been examined in co-culture experiments with in vitro scratch assays and a multiplex cytokines array system. Based on keratin 14 and prolyl-4-hydroxylase expression analyses, purity of both clinical batches was found to be above 98% and neither melanocytes nor Langerhans cells could be detected. Both cell types demonstrated strong immunosuppressive properties as shown by the dramatic decrease in allogeneic PBMC proliferation when co-cultured with fibroblasts and/or keratinocytes. We further showed that the indoleamine 2,3 dioxygenase (IDO) activity is required for the immunoregulatory activity of fetal skin cells. Co-cultures experiments have also revealed that fibroblasts-keratinocytes interactions strongly enhanced fetal cells secretion of HGF, GM-CSF, IL-8 and to a lesser extent VEGF-A. Accordingly, in the in vitro scratch assays the fetal fibroblasts and keratinocytes co-culture accelerated the scratch closure compared to fibroblast or keratinocyte mono-cultures. In conclusion, our data suggest that the combination of fetal keratinocytes and fibroblasts could be of particular interest for the development of a new allogeneic skin substitute with immunomodulatory activity, acting as a reservoir for wound healing growth factors.     

The connexin mimetic peptide Gap27 increases human dermal fibroblast migration in hyperglycemic and hyperinsulinemic conditions in vitro

Significant increases in skin wound healing rates occur by reducing connexin-mediated communication (CMC). Gap27, a connexin (Cx) mimetic peptide targeted to the second extracellular loop of Cx43, which inhibits CMC, increases migration of human keratinocytes and dermal fibroblasts. To examine the efficacy of Gap27 in a hyperglycemic and hyperinsulinemic in vitro environment, cell migration, gap junction, and Cx hemichannel functionality and cell-substrate adhesion assays were performed on human dermal fibroblasts and diabetic fibroblast and keratinocytes. To investigate fibroblast genes involved in these processes, extra-cellular matrix (ECM) and adhesion gene expression was determined with a PCR array. Gap27 increased fibroblast migration in both euglycemia/euinsulinemia and hyperglycemia/hyperinsulinemia, and influenced migration in diabetic keratinocytes. Hyperglycemia/hyperinsulinemia reduced gap junction coupling in fibroblasts and Gap27 reduced CMC and cell adhesion to substrata in fibroblasts cultured in high glucose. Migrating dermal fibroblast ECM and cell adhesion genes were found to be differentially regulated by Gap27 in euglycemia and hyperglycemia. The PCR array showed that Gap27 upregulated 34 genes and downregulated 1 gene in euglycemic migrating fibroblasts. By contrast in hyperglycemia, Gap27 upregulated 1 gene and downregulated 9 genes. In euglycemic conditions, Gap27 induced upregulation of genes associated with ECM remodeling, whereas in hyperglycemia, ECM component genes were downregulated by Gap27. Thus, Gap27 improves cell migration during scrape-wound repair in hyperglycemia/hyperinsulinemia conditions in vitro, although migration of diabetic cells is less influenced. Our results suggest that this increase in motility may occur by decreasing gap junction and hemichannel activity and altering gene expression in the adhesion and ECM pathway.     

Fibroblasts change spreading capability and mechanical properties in a direct interaction with keratinocytes in conditions mimicking wound healing

Keratinocytes are predominant in the uppermost layer of the skin, while fibroblasts dominate in the dermal layer. These cells interact with each other directly when fibroblasts migrate to a region of the wound where they induce keratinocytes proliferation through double paracrine signalling. Since a response from both keratinocytes and fibroblasts dominates during the inflammatory and proliferative phases, the exact knowledge how these two types of cells interact with each other is crucial for deeper understanding of mechanisms involved in the wound healing process. The aim of this study was to quantify alterations in mechanical properties of cells, i.e. fibroblasts and keratinocytes, in conditions mimicking direct cellular interactions observed in wound healing. Single cell elasticity was measured using atomic force microscope. To verify the influence of keratinocyte neighbors on fibroblasts elasticity (and vice versa), the effect of cellular confluency was studied in parallel. Our results enabled us to distinguish cellular density-related effects from intercellular interactions occurring between fibroblasts and keratinocytes. While the presence of keratinocytes affects fibroblasts spreading capability and mechanical properties, the keratinocytes remain unaffected by the fibroblasts. These results highlight the importance of the cellular deformability in understanding of the role of biomechanics in double paracrine signalling as fibroblast-keratinocyte interaction can change the potential of the wound healing.     

Induction of proteins and mRNAs after uv irradiation of human epidermal keratinocytes

uv sensitivity of cultured human epidermal keratinocytes was analyzed at different growth conditions and compared with the sensitivity of dermal fibroblasts derived from the same skin specimen. No significant differences in survival curves were found between these two cell types, although keratinocytes grown under standard conditions were slightly more resistant to uv irradiation than fibroblasts. The extracellular concentration of calcium appeared to be critical not only in the regulation of keratinocyte proliferation and differentiation, but also in the uv sensitivity of these cells: keratinocytes grown under conditions which favor cell proliferation (low calcium concentration) are more resistant to uv irradiation than those grown under conditions favoring differentiation (high calcium concentration). Two-dimensional protein gel electrophoresis was used to detect a possible effect of uv irradiation on the accumulation of specific mRNAs in the cytoplasm and/or on the synthesis of specific proteins. Proteins were pulse labeled in vivo with [35S]methionine or synthesized in vitro in rabbit reticulocyte lysates on mRNA isolated from keratinocytes that were irradiated with different uv doses at different periods of time prior to isolation. Alterations in expression were demonstrated for several proteins in both in vivo and in vitro experiments.     

Induction of Invasion and Tubulogenesis in Primary Human Keratinocyte Culture

In this study we used invasion and tubulogenesis assays to measure the morphogenetic potential of primary human keratinocytes in vitro. We also used human embryonic and dermal fibroblasts in coculture with keratinocytes and compared their influence on keratinocyte behavior in these different cell culture systems. The keratinocytes retained morphogenetic potential in vitro. In the presence of embryonic fibroblasts, the keratinocytes formed tubes and invaded collagen gel. A three-dimensional architecture of invasion clusters was reconstructed. We conclude that the cellular mechanisms of invasion and tubulogenesis are similar.     

Interaction of keratinocytes and fibroblasts modulates the expression of matrix metalloproteinases-2 and -9 and their inhibitors

Disruption of epidermal-mesenchymal communication due to a delay in epithelialization, increases the frequency of developing fibrotic conditions in skin. As matrix metalloproteinases-2 (MMP-2) and -9 (MMP-9) are two key enzymes involved in wound healing and tissue remodeling, here we examined the efficacy of keratinocyte-fibroblast interaction on modulation of these enzymes and their inhibitors. The conditioned media derived from keratinocytes and fibroblasts grown in upper and lower chambers of a co-culture system, respectively, were analyzed for MMP-2 and -9. Keratinocyte or fibroblast conditioned medium (FCM) was used as a control. Gelatinolytic activity analyzed by zymography showed that keratinocytes mainly express MMP-9 and to a lesser extent MMP-2; while fibroblasts express only MMP-2. In a co-culture system, the activities of both MMP-2 and MMP-9 markedly increased in conditioned media collected from bottom chambers. These findings were consistent with the level of MMP-2 and MMP-9 measured by Western blot. Using the same experimental setting, the levels of tissue inhibitors of MMPs (TIMPs) secreted by keratinocytes and fibroblasts grown in the same co-culture system were also evaluated. Western blot showed that fibroblasts secrete only TIMP-1 and TIMP-2 whose levels were increased by co-culturing fibroblasts with keratinocytes. In contrary the level of TIMP-3, which was mainly expressed by keratinocytes, increased by co-culturing these cells with fibroblasts. In conclusion, interaction of fibroblast-keratinocyte modulates the levels of MMP-2 and -9 and their inhibitors produced by these cells and this interaction may be critical for a better healing quality at a late stage of the wound healing process.     

In vitro paracrine regulation of human keratinocyte growth by fibroblast-derived insulin-like growth factors.

Human keratinocytes isolated from a skin biopsy and cultured in vitro on a feeder-layer of irradiated fibroblasts reconstitute a stratified squamous epithelium suitable for grafting onto patients suffering from large burn wounds. Since conditioned medium from 3T3-J2 cells can partially substitute for the intact feeder-layer, we studied the possible involvement of insulin-like growth factors acting in a paracrine fashion. IGFs were measured (after Sephadex G-50 gel-chromatography in acid conditions) in media conditioned by a feeder-layer of lethally irradiated 3T3-J2 fibroblasts on which keratinocytes were grown. Immunoreactive (IR) IGF-I, IGF-II, and IGF binding activity were present in the medium conditioned by the feeder-layer. The medium conditioned by keratinocytes showed nearly undetectable amounts of IR IGF-I and IGF-II, suggesting that keratinocytes are unable to synthesize IGFs peptides. Recombinant IGF-I and IGF-II, and conditioned medium from 3T3-J2 cells, caused a dose-dependent increase of 3H-thymydine incorporation in cultured keratinocytes. The stimulatory effect of IGF and of 3T3-J2 conditioned medium was inhibited by the MoAb Sm 1.2, which recognizes both IGF-I and IGF-II but not insulin, and by the MoAb alpha IR-3, which is a specific antagonist of type-I IGF receptor. Fetal mouse-derived 3T3-J2 cells and adult human skin fibroblasts were equally able to sustain keratinocyte growth and in both cases addition of Sm 1.2 MoAb causes a 50% decrease in the keratinocyte number. When the non-IGF-producing BALB/c 3T3 cells were used as a feeder-layer, the keratinocytes number was similar to that observed with 3T3-J2 and with human fibroblasts plus the Sm 1.2 MoAb. IGF-I and IGF-II restored the BALB/c 3T3 growth promoting activity to the level of 3T3-J2 and of normal human fibroblasts. Our results suggest that fetal mouse 3T3-J2 and human fibroblasts synthesize IGF peptides, while keratinocytes do not. Fibroblast-derived IGFs stimulate keratinocyte growth in a paracrine fashion, suggesting their role in the regulation of keratinocyte proliferation in skin growth and in wound healing.     

Interaction of keratinocytes and fibroblasts modulates the expression of matrix metalloproteinases-2 and -9 and their inhibitors

Disruption of epidermal-mesenchymal communication due to a delay in epithelialization, increases the frequency of developing fibrotic conditions in skin. As matrix metalloproteinases-2 (MMP-2) and -9 (MMP-9) are two key enzymes involved in wound healing and tissue remodeling, here we examined the efficacy of keratinocyte-fibroblast interaction on modulation of these enzymes and their inhibitors. The conditioned media derived from keratinocytes and fibroblasts grown in upper and lower chambers of a co-culture system, respectively, were analyzed for MMP-2 and -9. Keratinocyte or fibroblast conditioned medium (FCM) was used as a control. Gelatinolytic activity analyzed by zymography showed that keratinocytes mainly express MMP-9 and to a lesser extent MMP-2; while fibroblasts express only MMP-2. In a co-culture system, the activities of both MMP-2 and MMP-9 markedly increased in conditioned media collected from bottom chambers. These findings were consistent with the level of MMP-2 and MMP-9 measured by Western blot. Using the same experimental setting, the levels of tissue inhibitors of MMPs (TIMPs) secreted by keratinocytes and fibroblasts grown in the same co-culture system were also evaluated. Western blot showed that fibroblasts secrete only TIMP-1 and TIMP-2 whose levels were increased by co-culturing fibroblasts with keratinocytes. In contrary the level of TIMP-3, which was mainly expressed by keratinocytes, increased by co-culturing these cells with fibroblasts. In conclusion, interaction of fibroblast-keratinocyte modulates the levels of MMP-2 and -9 and their inhibitors produced by these cells and this interaction may be critical for a better healing quality at a late stage of the wound healing process. (Mol Cell Biochem 269: 209–216, 2005)     

Agent based modelling helps in understanding the rules by which fibroblasts support keratinocyte colony formation

Background

Autologous keratincoytes are routinely expanded using irradiated mouse fibroblasts and bovine serum for clinical use. With growing concerns about the safety of these xenobiotic materials, it is desirable to culture keratinocytes in media without animal derived products. An improved understanding of epithelial/mesenchymal interactions could assist in this.

Methodology/Principal Findings

A keratincyte/fibroblast o-culture model was developed by extending an agent-based keratinocyte colony formation model to include the response of keratinocytes to both fibroblasts and serum. The model was validated by comparison of the in virtuo and in vitro multicellular behaviour of keratinocytes and fibroblasts in single and co-culture in Greens medium. To test the robustness of the model, several properties of the fibroblasts were changed to investigate their influence on the multicellular morphogenesis of keratinocyes and fibroblasts. The model was then used to generate hypotheses to explore the interactions of both proliferative and growth arrested fibroblasts with keratinocytes. The key predictions arising from the model which were confirmed by in vitro experiments were that 1) the ratio of fibroblasts to keratinocytes would critically influence keratinocyte colony expansion, 2) this ratio needed to be optimum at the beginning of the co-culture, 3) proliferative fibroblasts would be more effective than irradiated cells in expanding keratinocytes and 4) in the presence of an adequate number of fibroblasts, keratinocyte expansion would be independent of serum.

Conclusions

A closely associated computational and biological approach is a powerful tool for understanding complex biological systems such as the interactions between keratinocytes and fibroblasts. The key outcome of this study is the finding that the early addition of a critical ratio of proliferative fibroblasts can give rapid keratinocyte expansion without the use of irradiated mouse fibroblasts and bovine serum.     

Neutrality of the canonical NF-kappaB-dependent pathway for human and murine cytomegalovirus transcription and replication in vitro

Cytomegalovirus (CMV) is known to rapidly induce activation of nuclear factor kappaB (NF-kappaB) after infection of fibroblast and macrophage cells. NF-kappaB response elements are present in the enhancer region of the CMV major immediate-early promoter (MIEP), and activity of the MIEP is strongly upregulated by NF-kappaB in transient-transfection assays. Here we investigate whether the NF-kappaB-dependent pathway is required for initiating or potentiating human and murine CMV replication in vitro. We show that expression of a dominant negative mutant of the inhibitor of NF-kappaB-alpha (IkappaBalphaM) does not alter the replication kinetics of human or mouse CMV in cultured cells. In addition, mouse embryo fibroblasts genetically deficient for p65/RelA actually showed elevated levels of MCMV replication. Mutation of all NF-kappaB response elements within the enhancer of the MIEP in a recombinant mouse CMV containing the human MIEP (hMCMV-ES), which we have previously shown to replicate in murine fibroblasts with kinetics equivalent to that of wild-type mouse CMV, did not negatively affect replication in fibroblasts. Taken together, these data show that, for CMV replication in cultured fibroblasts activation of the canonical NF-kappaB pathway and binding of NF-kappaB to the MIEP are dispensable, and in the case of p65 may even interfere, thus uncovering a previously unrecognized level of complexity in the host regulatory network governing MIE gene expression in the context of a viral infection.     

Development of phyllanthin containing microcapsules and their improved biological activity towards skin cells and Staphylococcus aureus

Chitosan based microcapsule which encapsulated with phyllanthin was developed by simple coacervation. The composition and surface morphology of phyllanthin containing microcapsules were analyzed by Fourier Transform Infrared spectroscopy and Scanning Electron Microscopy, respectively. The release of phyllanthin from the microcapsules was found to be more than 60% after 120 h. In vitro biological assays demonstrated that these phyllanthin containing microcapsules showed a stronger anti-oxidation potential on both human fibroblasts and keratinocytes as well as a better growth inhibitory activity towards Staphylococcus aureus.     

Multicomponent origin of cytomegalovirus lytic-phase DNA replication.

Cytomegalovirus (CMV) lytic-phase DNA replication requires both trans-acting factors, such as the virus-coded DNA polymerase, and a previously undefined cis-acting element, the origin, within which initiation occurs. We have located a candidate origin of CMV lytic-phase DNA replication, oriLyt, in both simian and human strains by assessing the ability of cloned restriction fragments to mediate phosphonoformic acid-sensitive DNA replication after transfection into human fibroblasts when required trans-acting factors were supplied by infection. In initial experiments the simian CMV-like strain Colburn EcoRI D fragment directed DNA replication; this fragment contains all of the single-stranded DNA-binding protein gene (dbp) and about 7 kbp of upstream sequence. A larger region upstream of human CMV dbp also mediated replication in transient assays. Subsequent subcloning and deletion analyses defined a CMV strain Colburn region sufficient for origin function, spanning about 1,300 bp in the apparently noncoding region upstream of dbp. The nucleotide sequence of this region revealed four distinct domains, containing (i) a 9-bp repeated sequence, (ii) an A+T-rich segment, (iii) an 11-bp direct repeat, and (iv) a 47-bp direct repeat. At least some part of each of these domains was required for origin function. Therefore, like the Epstein-Barr virus lytic-phase origin of DNA replication, CMV oriLyt appears to be structurally complex.     

CD98, a novel marker of transient amplifying human keratinocytes

Identification of plasma membrane markers of basal keratinocytes is essential for sorting basal cells and, subsequently, adult epidermal stem cells. In this study, we isolated caveolin-1-enriched microdomains from human HaCaT keratinocytes and identified proteins representing potential cell surface markers of the epidermis by a proteomic approach. The purification of this caveolae domain allowed us to characterize 53 proteins of which 26% were transmembrane and 32% associated-membrane proteins. One of them, CD98, was found to be co-localized with beta1 integrin at the plasma membrane of the basal keratinocytes of healthy human epidermis. We then isolated CD98-positive keratinocytes from fresh skin biopsies. Using clonogenic assays, we demonstrate that CD98 may be considered as a marker of transient amplifying human keratinocytes.     

Growth of normal human keratinocytes and fibroblasts in serum-free medium is stimulated by acidic and basic fibroblast growth factor

Keratinocytes and fibroblasts isolated from human neonatal foreskin can be plated and grown through multiple rounds of division in vitro under defined serum-free conditions. We utilized these growth conditions to examine the mitogenic potential of acidic and basic fibroblast growth factor (aFGF and bFGF) on these cells. Our results demonstrate that both aFGF and bFGF can stimulate the proliferation of keratinocytes and fibroblasts. aFGF is a more potent mitogen than bFGF for keratinocytes. In contrast, bFGF appears to be more potent than aFGF in stimulating the growth of fibroblast cultures. Heparin sulfate (10 micrograms/ml) dramatically inhibited the ability of bFGF to stimulate the proliferation of keratinocytes. In comparison, heparin slightly inhibited the stimulatory effect of aFGF and had no effect on epidermal growth factor (EGF) stimulation in keratinocyte cultures. In fibroblast cultures the addition of heparin enhanced the mitogenic effect of aFGF, had a minimal stimulatory effect on the mitogenic activity of bFGF, and had no effect on EGF-stimulated growth. Our results demonstrate that the proliferation in vitro of two normal cell types found in the skin can be influenced by aFGF and bFGF and demonstrate cell-type specific differences in the responsiveness of fibroblasts and keratinocytes to these growth factors and heparin.     

Receptor phenotype underlies differential response of hepatocytes and nonparenchymal cells to heparin-binding fibroblast growth factor type 1 (aFGF) and type 2 (bFGF)

Summary Heparin-binding fibroblast growth factors (HBGF) have been implicated in the regeneration of both parenchymal and nonparenchymal cells of the liver. The response to and phenotype of hepatocyte receptors for HBGF-1 (acidic fibroblast growth factor) and HBGF-2 (basic fibroblast growth factor) were compared to keratinocytes, fibroblasts, and endothelial cells. HBGF-1 stimulated DNA synthesis in hepatocytes, keratinocytes, fibroblasts, and endothelial cells whereas activity of HBGF-2 was limited to fibroblasts and endothelial cells. HBGF-2 antagonized the mitogenic activity of HBGF-1 for hepatocytes and keratinocytes. Hepatocytes and keratinocytes exhibited both high- and low-affinity, nonmatrix receptor sites for HBGF-1, but only low-affinity sites for HBGF-2. The mesenchymal cells displayed only high-affinity sites for both HBGF-1 and HBGF-2. Northern blot and immunochemical analysis revealed that the expression of HBGF receptor genesbek andflg are partitioned between normal hepatocytes and nonparenchymal cells, respectively. Expression of epithelial cell-specific, mesenchymal cell-derived HBGF-7 (keratinocyte growth factor) mRNA in regenerating liver tissue was undetectable relative to HBGF-1. The results support a multifunctional role of HBGF-1 acting through different receptor phenotypes in hepatocyte and nonparenchymal cells during liver regeneration.     

PTH-1R responses to PTHrP and regulation by vitamin D in keratinocytes and adjacent fibroblasts

Vitamin D and PTHrP are essential for the differentiation of keratinocytes and epidermal development. The action of PTHrP on skin is mediated via its PTH-1R receptors present in both epidermal and dermal cells. This suggests that PTHrP may have a paracrine/autocrine role, and its receptors may act in association or in negative cooperativity. We compared the intracellular signaling pathways in response to PTHrP (1-34) and to various PTHrP peptides, the N-terminal (1-34), Mid region (67-89), and C-terminal (107-139) fragments, and the possible modulation of PTHrP and its receptor mRNA expressions by vitamin D. Adjacent dermal fibroblasts as freshly isolated keratinocytes expressed both PTHrP and PTH-1R mRNAs, and responded to the various PTHrP fragments. bPTH and PTHrP(1-34) increased both cellular cAMP and [Ca(2+)]i in keratinocytes and fibroblasts. In contrast, PTHrP (107-139) increased [Ca(2+)]i but not cAMP in the two cell types. PTHrP (67-89) had no effect in keratinocytes, and only increased [Ca(2+)]i in fibroblasts. Vitamin D deficiency in weaned rats increased the expression of PTHrP mRNA in keratinocytes, and decreased it in fibroblasts and kidneys. Vitamin D deficiency increased PTH-1R mRNA expression in keratinocytes and kidneys, but not in fibroblasts. Although keratinocytes and skin fibroblasts are target cells for PTHrP and express PTH-1R, the two adjacent cell types differ as regards their intracellular signaling in response to PTHrP peptides. Moreover vitamin D regulates PTHrP and PTH-1R in a cell-specific manner.     

The interaction of epidermal keratinocytes and fibroblasts during the formation of a live skin equivalent]

The effect of human fetal fibroblasts and adult keratinocytes on collagen contraction was studied. Keratinocytes embedded in collagen lattices did not spread and produced only a slight contraction. When keratinocytes were seeded on the surface of tht gel, the contraction began within 24 h and correlated with the formation of epithelial colonies. Transplantation of multilayered epithelial sheets on the gel significantly accelerated the onset of contraction. Keratinocytes seeded on and fibroblasts grown in collagen lattices cooperatively contracted the gel, and keratinocytes were able to stimulate gel contraction even when they had no contact with the collagen roughly populated with fibroblasts. Swiss 3T3 cells remained spherical in collagen lattices and did not contract the gel but when cultivated with keratinocytes they stimulated gel contraction. In their turn, keratinocytes influenced the behaviour of Swiss 3T3 cells which elongated and produced processes. We suggest that both keratinocytes and mesenchymal cells can affect gel contraction 1) by a direct contact with collagen lattices, and 2) through potentiation of the ability of another cell type to contract the gel.