Cdc42 and p190RhoGAP activation by CCN2 regulates cell spreading and polarity and induces actin disassembly in migrating keratinocytes

Abstract

Cell migration requires spatiotemporal integration of signals that regulate cytoskeletal dynamics. In response to a migration‐promoting agent, cells begin to polarise and extend protrusions in the direction of migration. These cytoskeletal rearrangements are orchestrated by a variety of proteins, including focal adhesion kinase (FAK) and the Rho family of GTPases. CCN2, also known as connective tissue growth factor, has emerged as a regulator of cell migration but the mechanism by which CCN2 regulates keratinocyte function is not well understood. In this article, we sought to elucidate the basic mechanism of CCN2‐induced cell migration in human keratinocytes. Immunohistochemical staining was used to demonstrate that treatment with CCN2 induces a migratory phenotype through actin disassembly, spreading of lamellipodia and re‐orientation of the Golgi. In vitro assays were used to show that CCN2‐induced cell migration is dependent on FAK, RhoA and Cdc42, but independent of Rac1. CCN2‐treated keratinocytes displayed increased Cdc42 activity and decreased RhoA activity up to 12 hours post‐treatment, with upregulation of p190RhoGAP. An improved understanding of how CCN2 regulates cell migration may establish the foundation for future therapeutics in fibrotic and neoplastic diseases.

Introduction

Cell migration is a fundamental process, taking place throughout various biological and pathological events such as embryogenesis, inflammatory response, tissue repair and regeneration 1, 2. During wound healing, the directional migration of keratinocytes is necessary for successful re‐establishment of an intact epidermis, and cell migration can be viewed as a cycle of cytoskeletal rearrangements that includes cell polarisation, protrusion and adhesion, followed by translocation of the cell body and retraction of the rear of the cell 3, 4. These events are dependent on the remodelling of the actin cytoskeleton and coordinated by transient signalling networks 5, 6, 7.

Directed cell migration can be initiated by selected growth factors and cytokines, and CCN2 has recently emerged as a promoter of keratinocyte migration 8, 9, 10. CCN2 [also known as connective tissue growth factor (CTGF)] is the second member of the CCN family of matricellular proteins, known to interact contextually with cell surface receptors, cytokines, growth factors and proteases 11, 12. CCN2 has several roles during tissue repair, with the overall goal to promote angiogenesis and tissue integrity 13. During wound healing, CCN2 is best known as a profibrotic factor that stimulates mesenchymal cells to produce extracellular matrix (ECM) 14, 15. However, in addition to promoting cell proliferation and adhesion, CCN2 can regulate cell migration. The first three members of the CCN family have been shown to stimulate cell migration in several mesenchymal cell types, whereas in contrast, overexpression of CCN4 and CCN5 has been shown to inhibit cell migration 16, 17, 18, 19, 20, 21, 22. CCN2 has been shown to enhance the motility of mesenchymal cells, smooth muscle cells, hepatic stellate cells and corneal epithelial cells 23, 24, 25. As a regulator of keratinocyte motility, CCN2 has a potential role in promoting re‐epithelialisation and wound healing. However, the mechanism by which CCN2 promotes keratinocyte migration is poorly understood.

At the time of this study, no unique receptor for CCN2 has been identified, but instead, CCN2 has been shown to directly interact with several cell surface receptors such as integrins 26, 27. We have recently shown that CCN2 promotes keratinocyte adhesion and migration through integrin α5β1 and focal adhesion kinase (FAK) 28. FAK is a cytoplasmic protein tyrosine kinase that functions as a scaffolding platform for the binding of signal and adaptor proteins 29, 30. Upon activation, FAK is autophosphorylated at tyrosine 397, creating a binding site for the Src‐homologue 2 (SH2). The activated FAK‐Src complex interacts with several molecular switches as a key regulator of the actin cytoskeleton during cell spreading, protrusion and migration 31, 32. Although FAK expression has been linked to increased directional cell movement, the molecular mechanisms of FAK in CCN2 promoted keratinocyte migration are not known.

In keratinocytes, activation of FAK has been shown to be associated with altered activity of Rho GTPase signalling 33. The best studied members of the Rho family of GTPases are RhoA, Rac1 and Cdc42 with evidence that Rho and Rac directly regulate the assembly and organisation of F‐actin 34, 35. Later work revealed that Cdc42 induced filopodia formation and also demonstrated crosstalk within the family of Rho GTPases 36, 37. Since then, the Rho GTPases have been shown to be involved in the regulation of actin skeleton dynamics as well as the turnover of stress fibres, lamellipodia or filopodia, respectively 38. CCN2 has been shown to activate Rho GTPases in a variety of cell types, but the role of Rho proteins in CCN2‐induced keratinocyte motility is not known.

The Rho proteins are involved in dynamic cellular processes and transduction of signals occurs through the exchange of GDP for GTP. The cycling between the nucleotide‐dependent activation states of Rho proteins is regulated by guanine nucleotide exchange factors (GEFs) and GTPase‐activating proteins (GAPs) 39. The GEFs facilitate GTP loading by catalysing the release of GDP while the GAPs promote GTP hydrolysis and subsequent Rho inactivation 40. The interplay between GEFs and the GAPS is critical for a better understanding of the regulation of Rho GTPases. We hypothesise that CCN2 modulates the actin cytoskeleton via Rho signalling. Therefore, we set out to examine the role of Rho GTPases in CCN2‐promoted keratinocyte spreading and polarity.

Materials and methods

Cell culture

Cells were isolated from human skin obtained from patients undergoing routine plastic surgery procedures. Primary human keratinocytes were isolated from epidermis as previously described 41. Briefly, epidermis was separated from the dermis by overnight incubation at 4°C with 10 U/ml dispase (Invitrogen, Carlsbad, CA) in Dulbecco's Modified Eagle's Medium (DMEM) (Sigma Aldrich, St. Louis, MO). A single‐cell suspension of keratinocytes was obtained by treatment with 0·05% trypsin and 0·01% ethylenediaminetetraacetic acid (EDTA; Invitrogen) for 10 minutes and subsequently plated on collagen‐I‐coated culture dishes (Becton Dickinson, Bedford, MA). Keratinocytes were cultured in EpiLife keratinocyte medium supplemented with bovine insulin (5 µg/ml), hydrocortisone (0·5 µg/ml), human recombinant epidermal growth factor (0·1 µg/ml), 0·4% bovine pituitary extract and 60 μM calcium chloride (GIBCO, Grand Island, NY). Keratinocytes from passages 1 to 3 were used in the experiments.

Scratch wound model

A scratch wound model was used as described previously 41. Cultured cells at the edge of a scrape‐wounded monolayer show characteristic phenotypes of activated cells at the wound edge 42. Human keratinocytes were plated on fibronectin‐coated 24‐well plates, at a density of 5 × 10⁴ cells per well, and cultured to subconfluence in serum‐free keratinocyte medium. Cell monolayers were scraped with a 200‐µl plastic pipette tip, creating a cell‐free 2 mm wide zone in the centre of each well and displaced cells were removed by washing the plate with Phosphate buffered saline (PBS). The proliferative component was excluded by repeating the scratch assay in the presence of 0·2 M hydroxyurea (Sigma‐Aldrich) for 60 minutes. To evaluate the potentially toxic effect of hydroxyurea, cell viability was assessed by the lack of trypan blue uptake after 0, 12 and 24 hours incubation (data not shown). Recombinant human CCN2 (Invitrogen) was added at concentration 50 ng/ml, and serum‐free keratinocyte medium served as a negative control. Scratch wound closure was documented after 0, 12 and 24 hours. Non‐adherent cells were removed and adherent cells were fixed in 4% formaldehyde solution in PBS for 5 minutes at room temperature. Slides were mounted with aqueous mounting medium containing 4′,6‐diamidino‐2‐phenylindole (DAPI) (ProLong Gold, Invitrogen). Scratch wound closure was expressed as a percentage of scratch surface area covered by keratinocytes. The amount of resurfacing was quantified using NIS‐Elements D3·0 digital image analysis software (Nikon Corporation, Tokyo, Japan). All experiments were performed thrice independently with four cultures in each experiment.

Spreading assay

Keratinocytes were plated onto coverslips coated with fibronectin (10 µg/ml from bovine plasma; Chemicon International, Temecula, CA) and incubated with or without 50 ng/ml recombinant human CCN2. As indicated in the figure legend (see Figure 2), some groups were pretreated with 20 mM FAK inhibitor PF573228 (Sigma‐Aldrich) for 30 minutes prior to plating. After 1, 6 or 12 hours incubation, non‐adherent cells were removed and adherent cells were fixed in 4% formaldehyde solution in PBS for 5 minutes at room temperature. Spread cells were identified under phase contrast microscopy as cells with polarity or radially distributed lamellipodia. Non‐spread cells were categorised as highly retractile and round 43. The percentage of spread cells was determined by counting at least 100 cells from randomly chosen high power fields. The cells were scored as either spread or non‐spread. Scoring was performed by three independent observers in a blinded manner. All experiments were performed thrice independently with four cultures in each experiment.

Figure 2.

Inhibition of FAK prevents CCN2‐induced keratinocyte spreading. Keratinocytes were cultured on fibronectin‐coated coverslips and treated with control serum‐free keratinocyte media (A), 20 mM FAK inhibitor PF573228 (B), 50 ng/ml CCN2 (C) or 20 mM FAK inhibitor PF573228 and 50 ng/ml CCN2 as described in Materials and methods. After 1, 6 and 12 hours, cells were fixed in formalin and assessed for spreading under phase contrast microscopy. Scale bar equals 20 µm in A–D. Representative images from three different experiments are shown.

Immunocytochemistry

For immunocytochemistry, cells were plated on coverslips coated with fibronectin as described previously and cultured with 50 ng/ml CCN2 for 12 hours. Cells cultured in serum‐free keratinocyte medium served as negative control. Following treatment, cells were fixed in 4% formaldehyde solution in PBS for 20 minutes at room temperature. After 5 minutes of permeabilisation with 0·1% Triton X‐100, non‐specific binding sites were blocked using 1% bovine serum albumin (BSA) in PBS for 1 hour in room temperature. F‐actin was visualised with Alexa Fluor 488 phalloidin (Invitrogen) according to the manufacturer's protocol. Slides were mounted with aqueous mounting medium containing DAPI (ProLong Gold, Invitrogen) and assessed by fluorescence microscopy.

Rho GTPase pull‐down assay

Rho, Rac1 and Cdc42 activity was determined using the Pierce EZ‐detect activation kits (Pierce Biotechnology, Rockford, IL). Keratinocytes were grown to subconfluence on fibronectin‐coated dishes and treated with 50 ng/ml CCN2 for the times indicated in the figure legends. After treatment, the cells were washed with ice‐cold PBS and scraped into 0·5 ml of lysis buffer. Samples were vortexed and centrifuged at 13 000 g for 15 minutes at 4°C, and the supernatants were transferred to new tubes. Equal volumes of lysates were then incubated at 4°C for 1 hour with glutathione‐Sepharose 4B‐bound GST‐Rhotekin‐PBD (for RhoA‐GTP) or GST‐PAK1‐PBD (for Rac1‐GTP and Cdc42‐GTP) fusion protein. The beads were washed four times with lysis buffer and boiled in sodium dodecyl sulphate (SDS) sample buffer. The total amounts of Rho proteins from cell lysates were separated by SDS–polyacrylamide gel electrophoresis (PAGE) and measured by western blotting as described previously using specific antibodies for RhoA, Rac1 and Cdc42 41. All experiments were performed thrice independently.

Immunoprecipitation

Protein expression was assessed through immunoblotting. Briefly, keratinocytes were treated with 50 ng/ml CCN2 for 6, 12 or 24 hours with cells treated with serum‐free keratinocyte medium serving as negative control. In some treatment groups, keratinocytes were pretreated with 20 mM FAK inhibitor PF573228 (Sigma‐Aldrich) for 1 hour prior to the addition of CCN2. Cells were lysed in modified radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, pH 7·2, 150 mM NaCl, 0·25% deoxycholate, 1% NP‐40, protease and phosphatase inhibitors). Cell extracts were pre‐cleared with 50 µl of protein G Plus/A agarose beads (Calbiochem, San Diego, CA) for 30 minutes at 4°C. Protein concentration was determined using the BCA Protein Assay Reagent Kit (Bio‐Rad, Hercules, CA) according to the manufacturer's instructions. Two micrograms of anti‐p190 monoclonal antibody (Chemicon International) was incubated with 500 µg of extract for 2 hours at 4°C with gentle mixing. Fifty microlitres of protein G Plus/A agarose beads was added, and the immune complexes were incubated for 1 hour at 4°C. Immunoprecipitates were washed thrice with 0·5 ml of modified RIPA buffer, mixed with sample buffer and separated by SDS–PAGE for immunoblot analysis. Even loading was confirmed by stripping and reprobing the blot with GAPDH. The bands were visualised with chemiluminescence detection (ECL, GE Health Care, Westborough, MA). Quantitative analysis was performed using ImageJ Software (IH‐Image – http://rsbweb.nih.gov/ij). All experiments were performed thrice independently.

Golgi orientation

Keratinocytes were plated on fibronectin‐coated glass slides and cultured to subconfluence in serum‐free keratinocyte medium for 24 hours, prior to incubation with or without 50 ng/ml recombinant human CCN2. After 2 hours of treatment, cells were fixed in 4% formaldehyde in PBS for 20 minutes at room temperature. After 5 minutes of permeabilisation with 0·1% Triton X‐100, non‐specific binding sites were blocked using 1% BSA in PBS and Golgi was detected with the monoclonal anti β‐COP antibody (clone M3 A5, Sigma‐Aldrich) diluted 1:20. Subsequently, slides were incubated with Alexa Fluor 546 goat anti‐mouse IgG, diluted 1:500 and F‐actin was visualised with Alexa Fluor 488 phalloidin (Invitrogen) following the manufacturer's instructions. Slides were mounted with aqueous mounting medium DAPI (ProLong Gold, Invitrogen) and assessed by fluorescence microscopy. To quantify Golgi reorientation, a square was drawn over the nucleus and divided into quadrants. Quadrant A was assigned as the area of the cell between the nucleus and the leading edge. A reoriented Golgi was indicated by β‐COP staining entirely within, and only quadrant A. Analyses were performed on 100 cells per treatment group and the percentage of cells with reoriented Golgi was calculated 44. All experiments were performed thrice independently with four cultures in each experiment.

Statistical analysis

Values are given as mean ± standard deviation (SD). Student's t‐test was used when comparing two groups. Statistical comparison of more than two groups was performed using one‐way analysis of variance (ANOVA) with Dunn's post hoc test. Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad, LaJolla, CA), and P values of <0·05 were considered significant.

Results

CCN2‐induced cell spreading requires FAK

To confirm the ability of CCN2 to promote keratinocyte migration, primary keratinocytes were cultured to subconfluence on fibronectin‐coated 24‐cell plates and a straight line was made through the monolayer using a sterile pipette tip. The addition of 50 ng/ml CCN2 led to an increase in the number of keratinocytes covering the scratch wound area (Figure 1A–I). After 12 and 24 hours of incubation with CCN2, the scratch wound closure was 55 ± 9% and 94 ± 5% as compared with 26 ± 5% and 46 ± 11% scratch wound closure in keratinocytes treated with media control, respectively (P < 0·05; n = 4). (Figure 1J). To exclude the proliferative component of scratch wound closure, the scratch assay was repeated in keratinocytes pretreated with hydroxyurea for 60 minutes to induce growth arrest. Treatment with CCN2 promoted scratch wound closure in growth‐inhibited keratinocytes. After 24 hours treatment with CCN2, scratch wound closure was 91 ± 8% as compared with 46 ± 11% in keratinocytes treated with media control (P < 0·05; n = 4). These results indicate that CCN2 induces scratch wound closure primarily by promoting cell migration.

Figure 1.

CCN2 promotes keratinocyte migration. Photographs of one representative scratch wound assay show that treatment with 50 ng/ml CCN2 promotes keratinocyte scratch wound closure. Keratinocytes were treated with CCN2 or control growth media for 0, 12 or 24 hours and nuclei were stained with 4′,6‐diamidino‐2‐phenylindole (DAPI, blue). The scratch assays were photographed at 0 hours (A–C), 12 hours (D–F) and 24 hours (G–I) after wounding. In repeated experiments, keratinocytes were growth‐arrested by incubation with 0·2 M hydroxyurea (B, E, H). CCN2 stimulated scratch wound closure in growth‐arrested keratinocytes (J). Scale bar equals 100 µm in A–I. Graphs show mean ± standard deviation (SD) from three independent experiments, n = 4. ^* P < 0·05 versus 0 ng/ml CCN2 (control). HU, hydroxyurea.

In unwounded skin, the basal keratinocytes are cuboidal. Upon migration, keratinocytes acquire a migratory phenotype and the cells become flattened and elongated. To investigate the effect of CCN2 on keratinocyte morphology, keratinocytes were plated onto fibronectin‐coated coverslips and incubated with 50 ng/ml CCN2 for 0 to 12 hours. Spread cells were identified under phase contrast microscopy as cells with polarity and cells with radially distributed lamellipodia. Non‐spread cells were categorised as highly retractile and round (Figure 2A). As shown in Figure 2B, keratinocyte spreading was enhanced by the addition of CCN2. The addition of CCN2 increased the percentage of spreading keratinocytes from 38 ± 5% to 53 ± 9% (P < 0·05; n = 4), and the cells that spread under the presence of CCN2 covered a greater surface area (Figure 2C).

Keratinocytes bind to fibronectin using αβ1 integrins and we have previously shown that CCN2 promotes keratinocyte adhesion to fibronectin via integrin α5β1 28. FAK is recruited to sites where integrin binds to fibronectin, and to examine the role of FAK in CCN2‐induced cell spreading, keratinocytes were pretreated with PF573228 for 60 minutes before plating onto fibronectin‐coated coverslips. PF573228 is a selective FAK inhibitor that prevents phosphorylation on Tyr 397, thereby interfering with FAK function 45, 46. As expected, inhibition of FAK blocked CCN2‐induced cell spreading on fibronectin and the cells displayed an impaired spreading phenotype, with round cell morphology and less filopodia‐like processes (Figure 2D). These results suggest that CCN2‐induced keratinocyte polarity requires FAK.

CCN2 induces actin cytoskeleton disassembly

Cell spreading and migration depend on the rearrangement of the actin cytoskeleton. To study the effect of CCN2 on the actin cytoskeleton, keratinocytes were cultured with or without CCN2 for 24 hours and F‐actin was visualised with fluorescein‐conjugated phalloidin. Keratinocytes cultured without CCN2 extended lamellipodia uniformly around their circumference, and had intact and visible actin stress fibres (Figure 3A). In contrast, keratinocytes cultured in the presence of CCN2 assumed a well‐polarised phenotype, with a large leading edge lamellipodium that is typical of migrating keratinocytes (Figure 3B).

Figure 3.

CCN2 promotes actin cytoskeleton disassembly. The effect of CCN2 on the integrity of the actin cytoskeleton was investigated by immunocytochemistry. Primary human keratinocytes were cultured with or without 50 ng/ml CCN2 for 12 hours, after which the cells were fixed and stained for F‐actin. Actin was labelled with green and cell nuclei with blue. Intact actin stress fibres were seen in cells cultured in control growth media (A). Treatment with CCN2 resulted in a disassembly of actin stress fibres (B). Scale bar equals 20 µm in A and B and representative images of three different experiments are shown.

CCN2 regulates the activity of GTPases in human keratinocytes

Rho GTPases are important regulators of cytoskeletal dynamics, and each of the GTPases contributes to cell motility by regulating actin cytoskeletal rearrangements 47. To investigate the role of Rho GTPase family in CCN2‐induced actin disassembly, keratinocytes were treated with 50 ng/ml CCN2 and cell lysate was assayed for active RhoA, Rac1 and Cdc42 over the period of 0–12 hours. Treatment with CCN2 decreased the activity of RhoA over the course of 12 hours (Figure 4A). In addition, CCN2 increased the activity of Cdc42 during the same period (Figure 4B). However, CCN2 had no significant effect on Rac1 (Figure 4C).

Figure 4.

CCN2 modulates the activity of members of the Rho GTPase family. Primary human keratinocytes were treated with 50 ng/ml CCN2 for 0, 1, 6 or 12 hours and active RhoA, Cdc42 and Rac1 were determined by pull‐down assays. Treatment with CCN2 induced a decrease in activity of RhoA (A) and an increase in the activity of Cd42 (B). In contrast, Rac1 activity was unaffected by CCN2 (C). Results shown are representative of three separate experiments.

CCN2 activates p190RhoGAP in keratinocytes via FAK

RhoA is a well‐known regulator of the actin rearrangements associated with formation of focal adhesions, actin stress fibres and membrane protrusions 40. Previous studies have shown that RhoA is antagonised by p190RhoGAP during early stages of cell adhesion and spreading 48. To determine whether the decreased activity of RhoA was correlated with p190RhoGAP activation, 190RhoGAP was analysed after treatment with 50 ng/ml CCN2. Increased levels of p190RhoGAP were detected after 1 hour of incubation with CCN2 (Figure 5). Notably, the levels of phosphorylated p190RhoAGAP diminished after 6 hours and by 12 hours approached baseline levels. Pretreatment with FAK‐specific inhibitor PF573228 decreased the levels of CCN2‐induced p190RhoGAP, indicating that CCN2 activates p190RhoGAP in keratinocytes via FAK.

Figure 5.

CCN2 activates p190RhoGAP in keratinocytes. Human keratinocytes were cultured to subconfluence in serum‐free media prior to the addition of 50 ng/ml CCN2 and the production of p190RhoGAP was measured. The production of p190RhoGAP was maximally increased after 1 hour of incubation with CCN2 (A). Pretreatment with FAK‐specific inhibitor PF573228 reduced CCN2‐induced activation of p190RhoGAP (B). Results shown are representative of three separate experiments.

CCN2 induces cell polarity and Golgi reorientation via Cdc42

In unstimulated keratinocytes, the Golgi is randomly distributed within the cell. However, upon migration the cells establish an axis of polarity, with the Golgi facing the leading edge 49, 50. Cdc42 is necessary for the establishment of cell polarity and the results show that treatment with CCN2 induces Cdc42 activation in keratinocytes 51, 52. To study the effect of CCN2 on keratinocyte polarity, cells were plated on fibronectin‐coated cover slips and cultured with or without 50 ng/ml CCN2 for 2 hours. Golgi was stained with antibodies against β‐COP, a protein that is associated with the non‐clathrin‐coated vesicles in Golgi 53. Incubation with CCN2‐induced Golgi reorientation and cells cultured with CCN2 showed immunoreactivity polarising to one side of the nucleus (Figure 6). After treatment with CCN2, 78 ± 8% of keratinocytes exhibited Golgi reorientation, compared with 40 ± 7% of cells treated with media control (P < 0·05; n = 4).

Figure 6.

Cdc42 modulates CCN2‐induced Golgi reorientation. The effect of CCN2 on the Golgi reorientation was investigated by immunocytochemistry. Keratinocytes were plated on fibronectin‐coated coverslips and cultured with 50 ng/ml CCN2 for 2 hours. Cell nuclei were labelled blue (A, E, I), actin was labelled with green (B, F, J) and Golgi was detected using anti‐β‐COP antibodies and labelled red (C, G, K). Compared with cells treated with serum free keratinocyte media (A–D), keratinocytes incubated with CCN2 exhibited reoriented Golgi (E–H). Pretreatment with Cdc42 inhibitor ML141 inhibited CCN2‐induced Golgi reorientation (I–L). Scale bar equals 20 µm in A–L. Representative images of three different experiments are shown.

To study the role of Cdc42 in CCN2‐induced keratinocyte polarisation, cells were pretreated with the specific Cdc42 inhibitor ML141 (CID‐2950007). ML141 is a selective inhibitor of Cdc42 GTPase with no effect on the other members of the Rho family of GTPases 54. Pretreatment with ML141 for 60 minutes inhibited CCN2‐induced cell polarity and the cells showed a clear reduction of immunoreactivity and a distribution to both sides of the nucleus (Figure 6 I–L). These results indicate that CCN2 induces cell polarity and Golgi reorientation via Cdc42.

Discussion

Cell motility is central to many biological and pathological processes such as embryogenesis, angiogenesis, tissue repair and regeneration, and tumour invasion. CCN2 has emerged as a regulator of cell migration and we have previously shown that CCN2 promotes keratinocyte migration and activation of FAK 28. Although previous studies have identified FAK activation and the subsequent Rho GTPase signalling as important regulators of cell motility, the mechanism behind CCN2‐induced keratinocyte migration is not known.

In this study, we propose a basic molecular mechanism for CCN2‐induced keratinocyte spreading and migration. First, we demonstrate that CCN2‐induced keratinocyte migration requires FAK to be present as a scaffold protein. Treatment with CCN2 in the presence of FAK promotes the initial stages of cell spreading through the disassembly of F‐actin filaments. Disassembly is regulated by the CCN2‐induced activity of Rho GTPase proteins. In particular, CCN2 indirectly increases the activity of Cdc42 while inhibiting RhoA. These highly regulated changes are mediated by GAPs and GEFs such as p190RhoGAP, and ultimately function to facilitate alterations in keratinocyte polarity and motility in response to external stimuli. This mechanism is similar to others regarding CCN2‐induced cell migration in cell types such as osteoblasts, chondrocytes and renal mesangial cells, but is novel in the keratinocyte 55, 56, 57.

Cell migration requires spatiotemporal integration of signals that regulate cytoskeletal changes and is largely regulated by FAK 58. FAK is a cytoplasmic protein tyrosine kinase as well as a scaffolding protein that interacts with integrins and growth factors, allowing for signalling through the formation of multi‐protein complexes 59, 60, 61. In migrating cells, upregulation of FAK activation at the leading edge promotes establishment of cell polarity and subsequent motility. Cells deficient in FAK spread more slowly and migrate poorly in response to chemotactic signals 62. Conversely, FAK is overexpressed or stabilised in certain cell lines of invasive colon, pancreatic and breast carcinomas compared with normal tissue 63, 64, 65. These results are in accordance with our finding that CCN2‐induced spreading and migration of keratinocytes requires FAK (Figure 2).

Multiple binding sites allow FAK to associate with adaptor proteins that activate Rho family GTPases such as RhoA, Cdc42 and Rac1. Cdc42 has been described as a regulator of cell spreading and migration, particularly with regard to the phenotypic changes that occur in motile cells 38. Changes in cell size, shape or morphology such as reorientation of the Golgi are critical in establishing an axis of polarity prior to migration. Similar to FAK, downregulation of Cdc42 inhibits cell migration while overexpression has been implicated in stimulating migration in cancerous cell lines 66, 67. In this study, treatment of keratinocytes with CCN2 resulted in the predicted increase in Cdc42 activity in addition to morphological changes consistent with cell migration.

It has previously been shown that the activity of RhoA is limited during the initial stages of cell spreading when the actin cytoskeleton is in a state of disassembly. Although RhoA activity is required for cell migration, precise regulation is critical as extensive RhoA activity inhibits cell movement 68, 69. Without this transient inhibition, RhoA would functionally antagonise the formation of membrane protrusions through excessive cytoskeletal contractility and thereby inhibit migration 23. The opposing actions of GEFs and GAPs orchestrate the dynamic activities of the Rho proteins during cell migration. Decreased RhoA activity has been largely attributed to the action of p190RhoGAP, a 190‐kDa multidomain protein. P190RhoGAP is the most abundant GAP for RhoA in mammalian cells and becomes activated after phosphorylation by Src kinase 70. The correlation between FAK and p190RhoGAP activation in the present study suggests that CCN2 might indirectly regulate RhoA activity through FAK‐Src‐p190RhoGAP signalling. Several studies have highlighted this pathway as a key determinant of cell polarity that occurs at the leading edge of the cell in conjunction with p120RasGAP 44, 48, 71.

The main limitation of this study is the potential for crosstalk between signalling pathways to distort the results. As previously stated, binding of CCN2 to receptors is poorly understood and complicates our understanding of CCN2 activity. In addition, the involvement of other signalling proteins remains largely unknown in CCN2‐induced signalling in the keratinocyte, but provides an avenue for additional research. In the future, it will be interesting to further explore the interplay between nucleotide exchange proteins in CCN2‐induced cell migration, particularly with regard to Cdc42 that is known to interact with an array of GEFs and GAPs, including Tuba, ECT2, Rich1 and others 72, 73.

The results of this study and those of previous studies lead us to propose that CCN2‐induced keratinocyte migration is dependent on FAK, Cdc42 and RhoA, but independent of Rac1. More specifically, the proposed mechanism supports the hypothesis that CCN2‐induced alterations in activity of Cdc42 and RhoA with involvement of p190RhoGAP results in actin cytoskeleton disassembly and subsequent migration. As a mediator of cell migration, CCN2 may play a substantial role in both wound healing and cancer. An improved understanding of CCN2‐induced keratinocyte migration may establish the foundation for future diagnostic tools and therapeutics in these fields.