MicroRNA-205 promotes keratinocyte migration via the lipid phosphatase SHIP2

Abstract

microRNA-205 (miR-205) and miR-184 coordinately regulate the lipid phosphatase SHIP2 for Akt survival signaling in keratinocytes. As the PI3K-Akt pathway has also been implicated in regulating the actin cytoskeleton and cell motility, we investigated the role that these 2 miRNAs play in keratinocyte migration. We used antagomirs (antago) to reduce the levels of miR-205 and miR-184 in primary human epidermal keratinocytes (HEKs) and corneal epithelial keratinocytes (HCEKs) as well as direct SHIP2 silencing using siRNA oligos. Treatment of HEKs and HCEKs with antago-205 increased SHIP2 levels and impaired the ability of these cells to seal linear scratch wounds compared with untreated or irrelevant-antago treatments. In contrast, AKT signaling was enhanced and wounds sealed faster in HCEKs where miR-184 was suppressed, enabling miR-205 to inhibit SHIP2. Similar increases in migration were observed following direct SHIP2 silencing in HEKs. Furthermore, down-regulation of miR-205 resulted in an increase in Rho-ROCKI activity, phosphorylation of the actin severing protein cofilin, and a corresponding diminution of filamentous actin. The connection among miR-205, RhoA-ROCKI-cofilin inactivation, and the actin cytoskeleton represents a novel post-translational mechanism for the regulation of normal human keratinocyte migration.—Yu, J., Peng, H., Ruan, Q., Fatima, A., Getsios, S., Lavker, R. M. MicroRNA-205 promotes keratinocyte migration via the lipid phosphatase SHIP2.

Keratinocyte migration is essential for maintaining epidermal homeostasis and re-epithelialization following wounding (refs. 1, 2 and references therein). Much has been learned about the regulation of keratinocyte migration, including the roles that cell-matrix interactions and the actin cytoskeleton play in this process. This keratinocyte motility machinery is tightly regulated by both extrinsic and intrinsic factors. For example, soluble growth factors and ECM act via cell surface receptors and integrins, respectively, to trigger a number of intracellular signaling pathways that commonly converge on the regulation of Rho family GTPases, reorganization of the actin cytoskeleton, and cell migration (ref. 3 and references therein). At the same time, the baseline activation state of these various signaling pathways likely contributes to the ability of keratinocytes to respond to environmental cures and direct migration. To date, most of our knowledge has focused on the genes and proteins involved in the extrinsic regulation of keratinocyte motility machinery; however, it is now appreciated that microRNAs (miRNAs) play a significant role in determining the intrinsic potential of cells to migrate.

miRNAs exhibit as wide a variety of developmental and tissue distribution patterns as we are accustomed to seeing with the more conventional mRNA-coding proteins (4). Although the exact function of the majority of miRNAs remains unknown, these endogenous silencing RNAs have been shown to play important roles in development and differentiation, cell migration, cellular stress responses, stem cell regulation, and cancer (ref. 5 and references therein). The cancer studies have revealed important roles for miRNAs in invasion and metastasis (refs. 6–8 and references therein). For example, the microRNA-200 (miR-200) family and miR-205 are part of a regulatory loop with the transcription factor ZEB1 that regulates the epithelial to mesenchymal transition (9), which is likely to prove important for tumor metastasis (10). Less is known about how miRNAs regulate migration in normal cells.

miR-205 is expressed in a wide variety of epithelial tissues. In stratified squamous epithelia such as the epidermis and the anterior ocular surface, miR-205 is detected in the basal, suprabasal, and superficial layers (11, 12). This broad epithelial expression has led to the idea that miR-205 may be important for establishing epithelial cell lineages during early development and maintaining epithelial homeostasis in the adult (9, 13). Despite such a potentially fundamental role in epithelial biology, most information on miR-205 comes from investigations on cancer where, depending on the neoplasia, this miRNA has been proposed to function either as an oncomir (14–17) or a tumor suppressor (18–22). Using transformed cell lines (e.g., MDA-MB-231, MCF7, SK-LU-1, U87), low-density lipoprotein receptor-related protein 1 (17), ErbB3 (14), VEGF-A (14), and HER3 (15) have been shown to be targets of miR-205.

With respect to normal epithelial cells, we reported that miR-205 represses SH2-containing phosphoinositide 5′-phosphatase 2 (SHIP2) in primary cultures of human epidermal keratinocytes (HEKs) and human corneal epithelial keratinocytes (HCEKs) (13). SHIP2 is a ubiquitous lipid phosphatase that dephosphorylates phosphatidyinositol 3,4,5-triphosphate (PIP3), a critical second messenger in several cell signaling pathways including Akt, phosphoinositide-dependent kinase-1, guanosine diphosphate-GTP exchange factors, and adaptor proteins (GAB-1) (23–25). Furthermore, we demonstrated that the corneal epithelial-preferred miR-184 (11) can interfere with the ability of miR-205 to suppress SHIP2 levels (13). Our studies also indicated that miR-205 enhanced the Akt-signaling pathway via SHIP2 suppression, leading to improved cell survival (13). Interestingly SHIP2 and the PI3K-AKT pathway have also been implicated in regulating the actin cytoskeleton and cell migration (26–29), although more studies have focused on the upstream regulation of Akt signaling by the related lipid phosphatase and tumor suppressor PTEN (30–32). Moreover, the coordinated regulation of PTEN and PI3K activity results in increased PIP3 at the leading edge of fibroblasts and promotes directional migration of fibroblasts (33). Thus intrinsic factors that are capable of regulating PIP3 levels (i.e., SHIP2, PTEN) and downstream Akt signaling are likely to be key determinants of keratinocyte migration potential.

As miR-205 and miR-184 coordinately regulate SHIP2 to modulate Akt signaling in keratinocytes (13), we investigated whether these 2 miRNAs play a role in keratinocyte migration. We report that suppression of miR-205 in HEKs and HCEKs results in a decrease in keratinocyte migration. Antagonizing miR-205 also causes major cytoskeletal reorganizational events including a marked diminution of filamentous actin and an increase in focal contacts (FCs). Suppression of miR-184 in HCEKs, which enables miR-205 to inhibit SHIP2, or direct SHIP2 silencing using siRNA oligos, both of which lead to an enhancement in Akt signaling, results in a more rapid sealing of wounds. Moreover, the ability of miR-205 and miR-184 to regulate SHIP2 levels and Akt activity converged on RhoA signaling in keratinocytes, leading to an increase in the activation state of the downstream effector, cofilin (34, 35), and reorganization of the actin cytoskeleton. Collectively, these observations define a novel intrinsic regulatory network for the keratinocyte motility machinery by defining the set-point for Akt signaling potential through the modulation of miR-205 and miR-184 expression.

MATERIALS AND METHODS

Cell culture

Primary cultures of HEKs were maintained in keratinocyte serum-free medium (medium 154; Cascade Biologics, Portland, OR, USA) containing HKGS growth supplements and 70 μM CaCl₂. Primary HCEKs were cultured in CnT-50 with supplements (CellnTech, Bern, Switzerland) on collagen IV-coated plates (BD Biosciences, San Jose, CA, USA). HeLa cells were obtained from American Type Culture Collection (Manassas, VA, USA) and grown in Ham's F12 medium with 10% FBS.

Constructs and oligonucleotides

The 3′UTR of the human SHIP2 mRNA was cloned between the SpeI and HindIII sites of pMIR-Report (Ambion, Austin, TX, USA). The mutant of the SHIP2 sequence was created by using a QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). miRNA mimics specifying miR-205 and nontargeting controls were obtained from Dharmacon (Lafayette, CO, USA). Antagomirs were synthesized by Dharmacon. Sequences were 5′-mC_smA_smG_smAmCmUmCmCmGmGmUmGmGmAmAmUmGmAmA_smG_smG_smA-Chol 3′ (antago-205), 5′-mG_smG_smC_smAmUmUmCmAmCmCmGmCmGmUmGmCmC_smU_smU_smA-Chol 3′ (irrelevant-antago), 5′-mA_smC_smC_smCmUmUmAmUmCmAmGmUmUmCmUmCmCmGmU_smC_smC_smA-Chol 3′ (antago-184). “mN” represents 2′-O-methyl-modified oligonucleotide, the subscript “s” represents a phosphorothioate linkage, and “Chol” represents linked cholesterol.

Transfections and assays

Transfection and a luciferase report assay were conducted as described previously (13). For Western blot and immunocytochemical analyses, HeLa cells were serum-starved for 48 h after transfection.

Antagomirs directed against miR-184, miR-205, and miR-124 were synthesized by Dharmacon and were added to culture medium to a final concentration of 1000 nM. HEKs and HCEKs were grown in normal culture medium to a 70% confluent state and then treated with antagomir-containing culture medium for 48–72 h.

SHIP2 siRNA and control siRNA pools were synthesized by Dharmacon and transfected into HEKs (100 nM) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Medium was changed every 24 h, cells were cultured for 72 h and harvested for Western blot analyses as described below.

Western blots

Western blots were performed as described previously (13). The following antibodies were used: SHIP2, p-Akt(Ser-473), Akt, p-Paxillin(Tyr118), Paxillin, p-cofilin(Ser-3), cofilin, p-ezrin(Thr567)/radixin(Thr564)/moesin(Thr558), ezrin/radixin/moesin, desmin, vimentin (Cell Signaling Technologies, Danvers, MA, USA), p-FAK(Tyr576/577), FAK (Santa Cruz Biotechnology, Santa Cruz, CA, USA), GAPDH (Abcam, Cambridge, MA, USA), and α-tubulin (Invitrogen).

The Rho-associated kinase I (ROCKI) inhibitor Y-27632 (Sigma-Aldrich, St. Louis, MO, USA) was added to the culture medium at a concentration of 10 μM. Following treatment of NHEKs with antagomirs for 48 h, cells were incubated with Y-27632 for 1 h.

Affinity-precipitation of cellular GTP-Rho and GTP-Rac

HEK cells were washed with ice-cold TBS-M (150 mM NaCl; 50 mM Tris, pH 7.6; and 5 mM MgCl₂) and lysed on ice with either buffer A (50 mM Tris, pH 7.6; 1% Triton X-100; 0.5% sodium deoxycholate; 0.1% SDS; 500 mM NaCl; 5 mM MgCl₂; 10 μg/ml each of leupeptin and aprotinin; 200 μM orthovanadate; and 1 mM PMSF) or buffer B (1% Triton X-100; 150 mM NaCl; 50 mM Tris, pH 7.6; 5 mM MgCl₂; 10 μg/ml each of leupeptin and aprotinin; 200 μM orthovanadate; and 0.1 mM PMSF). Cell lysates were centrifuged at 14,000 g at 4°C for 4 min. For Rho-GTP pulldown, equal volumes of cell lysates prepared with buffer A were incubated with GST-RBD (30 μg) beads at 4°C for 30 min. For GTP-Rac puldown, equal volumes of cell lysates prepared with buffer B were incubated with GST-PBD (30 μg) beads at 4°C for 30 min. The beads were washed 3 times with buffer B. Bound Rho and Rac were detected by Western blotting using anti-RhoA and anti-Rac (Cell Signaling Technologies), respectively. Equal volumes of cell lysates were blotted with anti-Rho A, anti-Rac, and anti-cdc42 as loading controls.

Scratch wound assay

HEKs and HCEKs were grown to confluence on 12-well plastic dishes and treated with antagomirs or siRNAs. Then 48 h posttreatment, linear scratch wounds (in triplicate) were created on the confluent monolayers using a 200 μl pipette tip. Immediately after wounding (time 0) and at 6 h intervals for 30 h, images of the untreated, antago-treated, siRNA-treated HEKS, and/or HCEKs were taken using a Zeiss AxioCam MR digital camera mounted on a Zeiss Axiovert 40 CFL inverted light microscope (Carl Zeiss, Oberkochen, Germany). The percentage decrease in the wound gaps were calculated using Axiovision computer-assisted image analysis and normalized to the time 0 wounds.

To remove cells from the cell cycle prior to wounding, mitomycin C (5 μg/ml) was added into the medium of antago-treated HEKs and HCEKs after 2 h incubation. Linear scratch wounds were created and evaluated as described above.

Cell adhesion assay

HEKs were seeded onto 6-well plates (5×10⁵ cells/well) the day before transfections were performed. Cells (≈70% confluent) were transfected with SHIP2 siRNA or control siRNA pools (100 nM/well) using Lipofectamine 2000 (Invitrogen), or were treated with antago-205 or an irrelevant antagomir (1000 nM). After 48 h, HEKs were harvested in PBS with EDTA, resuspended in culture medium with the respective siRNAs or antagomirs, and seeded onto collagen-coated 6-well plates (BD Biosciences). After incubation for 1 h at 37°C, cells were washed (3×) in PBS, and adherent cells were fixed with 10% formalin and stained with methylene blue. The number of adherent HEKs was visualized and quantified using a Zeiss Axiovert light microscope and Axiovision image analysis software.

Immunohistochemistry and light microscopy

HeLa, HEK, and HCEK cultures grown on glass coverslips were fixed in 4% paraformaldehyde at room temperature for 20 min. After washing in PBS, cells were blocked and permeabilized in PBS containing 2.5% goat serum and 0.1% Triton X-100 at room temperature for 90 min. Filamentous actin was detected by incubating coverslips for 2 h at room temperature with rhodamine-conjugated phalloidin (1:50, Sigma). For SHIP2 staining, human SHIP2 antibody (Cell Signaling Technologies) was incubated at 1:25 overnight at 4°C, and then Alexa Fluor 488 goat anti-rabbit IgG (1:500; Invitrogen) was incubated at room temperature for 1 h. As a negative control, antibodies against rabbit IgG were used. Cells were viewed and photographed with a Zeiss UV LSM 510 confocal microscope.

RESULTS

Down-regulation of miR-205 decreases keratinocyte migration

In primary keratinocytes, miR-205 and miR-184 interact to regulate SHIP2 levels, and such interaction regulates Akt signaling (13). Since the PI3K-Akt signaling pathway regulates many functions in cells, one of which is cell migration (26, 36), we asked whether modulation of miR-205 levels would affect the ability of keratinocytes to migrate. Wound healing assays, where a scratch is made on a confluent monolayer of cells, are commonly used for assessing epithelial cell migration (37). Linear scratch wounds were made on confluent sheets of irrelevant antago- (antago-124: a neuronal specific miRNA; ref. 38), antago-205-treated, or untreated HEKs, and the extent and time required to seal the wounds was determined. Untreated and irrelevant antago-treated HEKs completely sealed linear scratch wounds between 24 and 30 h, whereas antago-205-treated HEKs sealed only ∼60% of the wounded area after 30 h (Fig. 1A, B). Furthermore, HEKs treated with antago-205 showed a marked increase in SHIP2 levels by Western blot analysis when compared with the irrelevant antagomir-treated or untreated NHEKs (Fig. 1C). To eliminate the contribution of proliferation to wound closure, a separate series of experiments were conducted where, prior to wounding, antago-treated HEKs and HCEKs were exposed to mitomycin C to remove the cells from the cell cycle. No difference in wound closure was observed in the antago-205/mitomycin C-treated cells compared with the antago-205-treated cells (Supplemental Table 1), indicative that the changes seen reflected migration without any major contributions due to proliferation.

Figure 1.

miR-205 and SHIP2 levels regulate keratinocyte migration. A) HEKs were treated with an irrelevant antagomir (ir-antagomir) or an antagomir to miR-205 (antago-205) for 48 h. At this time, confluent HEKs were scratch wounded and then allowed to migrate for 30 h in the presence of the ir-antagomir or antago-205. Phase-contrast micrographs were taken immediately after wounding and at 6 h intervals for up to 30 h. B) Antago-205-treated HEKs sealed wounds slower when compared with ir-antagomir-treated or untreated HEKs. Wounds (in triplicate) were photographed, and the percentage of wound closure from a representative experiment (n=3) was measured temporally using AxioVision software. Data are means ± se. C) Immunoblotting of SHIP2 and GAPDH in HEKs shows an increase in SHIP2 expression 48 h after treatment with antago-205. Numbers between panels represent the densitometry values of the protein normalized to total levels.

Silencing SHIP2 enhances keratinocyte migration

SHIP2 has been shown to regulate cytoskeletal remodeling, cell adhesion, and spreading (28, 39). Since we established that SHIP2 is a target of miR-205 in HEKs and HCEKs (13), we reasoned that silencing SHIP2 should enhance the ability of HEKs to seal scratch wounds. Following 48 h of treatment with either a siSHIP2 or a control siRNA, HEKs reached confluence, linear scratch wounds were made, and wounds were allowed to seal in the presence of siSHIP2 or the control siRNA. As predicted, siSHIP2-treated HEKs sealed the wounds at a faster rate than siRNA control-treated or untreated HEKs (Fig. 2A, B). Western blot analysis of HEKs treated with siSHIP2 showed a reduction in SHIP2 protein (Fig. 2C).

Figure 2.

SHIP2 restricts keratinocyte migration. A) HEKs were treated with a siSHIP2 or a control siRNA for 48 h. Scratch wounds were made in confluent cells, and cells were allowed to migrate for 30 h in the presence of the siSHIP2 or control siRNA. B) siSHIP2-treated HEKs sealed wounds faster when compared with control siRNA-treated and untreated HEKs. Data are means ± se. C) Immunoblotting of SHIP2 and GAPDH in HEKs shows a decrease in SHIP2 expression 48 h after treatment with siSHIP2. Numbers between panels represent the densitometry values of the protein normalized to total levels.

We used an alternative approach to down-regulate SHIP2 and assess its effects on cell migration by taking advantage of the unique relationship between SHIP2 and miR-205 in HCEKs. As mentioned previously, miR-184 antagonizes miR-205 to maintain SHIP2 levels in these keratinocytes (13). Since we showed that antago-184 “released” miR-205 to negatively regulate SHIP2 (13), we reasoned that treatment of HCEKs with antago-184 prior to wounding would result in faster healing. Conversely, we expected that treatment of HCEKs with antago-205 (a positive control) should mimic the results obtained in HEKs and retard keratinocyte migration. Indeed, antago-184-treated HCEKs sealed wounds 33% faster than irrelevant antago-treated HCEKs 24 h postwounding (Fig. 3A). As predicted, antago-205-treated HCEKs sealed wounds 50% slower than irrelevant-treated HCEKs (Fig. 3B). SHIP2 levels in HCEKs decreased after treatment with antago-184 and increased after treatment with antago-205 compared with irrelevant antagomir-treated cells (Fig. 3C). Taken together these findings support the idea that miR-205 functions to accelerate cell migration in keratinocytes via a down-regulation of SHIP2.

Figure 3.

Down-regulation of miR-184 and miR-205 have reciprocal effects on HCEK migration. A) HCEKs were treated with either antago-184, antago-205, or an irrelevant antagomir for 48 h. At this time HCEKs were confluent, scratch wounded, and then allowed to migrate for 24 h in the presence of antago-184, antago-205, or the irrelevant antagomir. Phase-contrast micrographs were taken immediately after wounding and 24 h postwounding. B) Antago-184-treated HCEKs sealed wounds faster at 24 h compared with irrelevant antagomir-treated HCEKs. Conversely, antago-205-treated HCEKs sealed wounds slower at 24 h compared with irrelevant antagomir-treated HCEKs. Wounds (in triplicate) from a representative experiment (n=3) were photographed, and the percentage of wound closure at 24 h was determined using AxioVision software. Data are means ± se. C) Immunoblotting of SHIP2 and GAPDH in HCEKs that were treated with an irrelevant antagomir, antago-184, or antago-205 for 48 h. Reduction in SHIP2 is seen following antago-184 treatment, whereas SHIP2 is increased after miR-205 treatment. Numbers between panels represent the densitometry values of the protein normalized to total levels.

Down-regulation of miR-205 enhances cell adhesion and focal contacts

It is well known that adhesion plays a central role in cell migration. Therefore we investigated the effects that modulating miR-205 would have on keratinocyte adhesion. HEKs treated with antago-205 for 48 h were harvested and allowed to adhere to collagen-coated wells. After 1 h incubation, HEKs were washed in PBS, and the number of adherent cells quantified. Antago-205-treated cells were ∼40% more adherent when compared to irrelevant antago-treated cells (Fig. 4A). We also transfected siRNA oligonucleotides specific for SHIP2 into HEKs and then performed adhesion assays on collagen coated wells. siSHIP2-treated HEKs were ∼45% less adherent when compared with control siRNA-treated HEKs (Fig. 4B). Collectively, these findings suggest that miR-205-induced acceleration in keratinocyte migration may be due in part to SHIP2-mediated decreases in HEK adhesion to the ECM.

Figure 4.

Down-regulation of miR-205 increases cell adhesion and focal contacts. A, B)HEKs were treated with antago-205 or an irrelevant antago (A) or were transfected with SHIP2 siRNA or control siRNA pools for 48 h and then tested for their ability to adhere to collagen-coated plates (B). After a 1 h incubation, antago-205 treated HEKs were significantly more adherent than irrelevant antago-treated cells. Conversely, si-SHIP2-treatment diminished adhesion to collagen. Values are means ± sd of 3 experiments. **P < 0.01; *P < 0.05. C) Immunoblotting of pFAK, FAK, p-paxillin, and paxillin in HEKs that were treated with an irrelevant antagomir or antago-205 for 48 h, showing that down-regulation of miR-205 increases p-FAK and p-paxillin. α-Tubulin serves as a loading control. Numbers between panels represent the densitometry values of phosphorylated protein normalized to total levels.

FCs help anchor keratinocytes to the ECM (40). Therefore, we determined the effect of antago-205 on the expression of FAK and paxillin, which are biomarkers of FCs (ref. 41, 42). We used phospho-FAK (tyrosine 576/577) and phospho-paxillin (tyrosine 118) antibodies to indicate kinase domain activity of FAK and paxillin, respectively. Consistent with the increases seen in cell adhesion, HEKs treated with an antago-205 showed an increase in p-FAK and p-paxillin levels by Western analyses when compared with the irrelevant antago, whereas no change was noted in the total levels of FAK or paxillin (Fig. 4C). Furthermore, increased attachment to the ECM most commonly corresponds to a reduction in the migratory ability of keratinocytes (43).

miR-205 increases filamentous actin

Actin filaments are involved in cell migration, cell adhesion to the ECM, and FCs (ref. 44 and references therein). To study the role of miR-205/miR-184 and SHIP2 in actin organization we used HeLa cells, which lack miR-205 and miR-184 but contain high levels of SHIP2 (13, 28). Consistent with our previous studies, transfecting HeLa cells with a miR-205 mimic and luciferase reporter constructs carrying the entire 3′UTR of SHIP2 mRNA for 48 h caused a dramatic reduction in luciferase activity when compared with the irrelevant mimic, confirming that SHIP2 is a target of miR-205 (Supplemental Fig. 1A). Such transfections markedly reduced SHIP2 protein on immunoblots (Supplemental Fig. 1B) and immunostaining (Fig. 5A), whereas a nontargeting (irrelevant) mimic caused little reduction in SHIP2 protein. Interestingly, HeLa cells treated with the miR-205 mimic had increased levels of filamentous actin (F-actin) as shown by phallodin staining compared with untreated and irrelevant mimic-treated cells (Fig. 5A).

Figure 5.

miR-205 is a positive regulator of actin organization. A) Immunofluorescence microscopy of HeLa cells stained with anti-SHIP2, phalloidin (F-actin), and DAPI (nuclei), showing a marked decrease in SHIP2 staining and an increase in phalloidin staining 48 h after treatment with miR-205 mimic. B, C) Immunofluorescence microscopy of HEKs (B)and HCEKs (C) stained for phalloidin, showing a marked decrease in filamentous actin 48 and 72 h following treatment with antago-205. DAPI staining highlights the nuclei.

Since ectopic expression of miR-205 in HeLa cells resulted in an increase in filamentous actin, we reasoned that down-regulation of miR-205 should result in disorganization of this cytoskeletal component in HEKs and HCEKs, which express miR-205 and SHIP2 (13). To test this, we treated HEKs and HCEKs for 48 and 72 h with antago-205 and an irrelevant antago and compared the distribution of F-actin. As expected, a reduction of miR-205 in HEKs and HCEKs resulted in a marked diminution in F-actin (Fig. 5B, C). Thus, in keratinocytes, miR-205 appears to maintain actin filament organization.

Down-regulation of miR-205 increases p-cofilin via Rho activation

Our observations that down-regulation of miR-205 decreased F-actin and retarded the ability of keratinocytes to seal scratch wounds prompted us to investigate the possible mechanisms that controlled these processes. RhoA GTPase is a member of the Rho family of small GTPases that regulate the actin cytoskeleton with downstream targets such as myosin light chain kinase, diaphanous-related formins, and cofilin, having the potential to affect actin polymerization (34, 45, 46). Therefore, we determined the status of activated Rho in keratinocytes treated with antago-205, using a GST-fusion protein containing the Rho-binding domain of mouse Rhotekin. Immunoblotting revealed that the levels of activated RhoA were increased in antago-205-treated HEKs when compared with the irrelevant-antago treatment (Fig. 6A); RhoB was not detected. We did similar experiments to assess the levels of activated Rac and observed no changes between irrelevant-antago and antago-205 treatment (Supplemental Fig. 2). Cofilin is a major downstream target of RhoA and an important regulator of actin dynamics and cell motility (ref. 47 and references therein). We therefore investigated the effects of miR-205 on this actin binding and severing protein (34). Using Western blot analysis, we noted an increase in p-cofilin following a 48 h treatment of HEKs with antago-205 when compared with the irrelevant antagomir (Fig. 6B, C). No change was seen in total cofilin levels (Fig. 6C).

Figure 6.

Down-regulation of miR-205 inactivates cofilin via Rho activation. A) Rho-pulldown assays with GST-Rhotekin to detect relative amounts of RhoA-GTP and total RhoA by immunoblotting following treatment of HEKs with antago-205 or an irrelevant antagomir for 48 h. Antago-205 increased the amount of active RhoA. B) HEKs were treated with antago-205 or an irrelevant antagomir for 48 h, and cells were then incubated with the ROCKI inhibitor Y-27632 for 1 h. Cell lysates were immunoblotted with p-cofilin and total cofilin antibodies. Antago-205-treated-HEKs exposed to the Y compound had decreased levels of p-cofilin, when compared to antago-205-treated HEKs, indicating involvement of the Rho/ROCK pathway in miR-205-related cofilin inactivation. C) Immunoblotting of SHIP2, p-cofilin, and cofilin in HEKs treated as described in panel A, showing marked increases in SHIP2 and p-cofilin following exposure to antago-205. D) Immunoblotting of p-ERM, total ERM, and p-Akt in HEKs treated as described in panel A, showing marked decreases in p-ERM and p-Akt following exposure to antago-205. Top upper band in the total ERM blot represents ezrin and radixin; bottom band represents moesin. GAPHD serves as a loading control. Numbers between panels represent the densitometry values of phosphorylated protein normalized to total levels.

In a separate series of experiments, antago-205 treated- and irrelevant-antago treated-HEKs were exposed to Y-27632, an inhibitor of ROCKI (48), which is a downstream effector of Rho (35). Western blot analysis with an antibody to p-cofilin showed that antago-205 affected phosphorylation of cofilin in a Y compound-sensitive manner since antago-205-treated-HEKs exposed to the Y compound had decreased levels of p-cofilin when compared to antago-205-treated HEKs (Fig. 6B). Collectively these results suggest that the changes in F-actin following down-regulation of miR-205 occur, in part, through Rho-ROCK activation, which ultimately inactivates cofilin via phosphorylation (45). Such changes in cofilin might partially account for the miR-205 induced increase in keratinocyte migration. To our knowledge this is the first demonstration of a link between a miRNA, cofilin, and the actin cytoskeleton in normal human keratinocytes.

Finally, we investigated SHIP2, phosphorylated Akt, and the ERM (ezrin/radixin/moesin) family because these actin-associated proteins are capable of regulating the PI3K/Akt pathway and possibly contribute to regulation of actin during cell migration (26–29, 49). We observed an increase in SHIP2 (Fig. 6C) and a decrease in p-Akt and p-ERM (Fig. 6D) following treatment with antago-205 when compared with irrelevant antago-treated HEKs (Fig. 6C, D). However, no major change in total ERM or Akt levels was observed. The effects that miR-205 down-regulation have on the phosphorylation status of cofilin (Fig. 6C) and the ERM proteins (Fig. 6D) could contribute to the decrease observed in keratinocyte migration. Since miR-205 has also been implicated in epithelial to mesenchymal transitions (EMT; ref. 9), we determined whether miR-205 affected the mesenchymal marker desmin and the EMT marker vimentin. No changes were seen in either of these 2 proteins in HEKs following treatment with antago-205 (Supplemental Fig. 3).

DISCUSSION

miR-205 negatively regulates SHIP2: implications for cell migration

Cell migration is vital for the homeostasis of self-renewing tissues such as the epidermis and the epithelia of the ocular anterior surface. Migration is also a key component of re-epithelialization, as this process is initiated by keratinocyte movement at the edges of a wound. Cell migration is a dynamic process (ref. 44 and references therein), and major advances have been made in our understanding of cell polarization, rearrangements in the cytoskeleton, assembly and disassembly of cell-cell contacts, and interactions between the cell and the ECM, coordinated events necessary for successful cell movement. We now identify miR-205 as a positive regulator of keratinocyte migration by altering F-actin organization and decreasing cell-substrate adhesion. miR-205 also suppresses the lipid phosphatase SHIP2 (13), and we show that such suppression decreases cell-substrate adhesion and enhances migration. Collectively these findings indicate that in normal human keratinocytes, SHIP2 has a negative role in migration, and this is consistent with a tumor suppressor function suggested for SHIP2 (13, 50).

Our findings imply that SHIP2 may be functioning in a manner similar to the lipid phosphatase, PTEN, which also dephosphorylates PIP₃ (30, 32). Importantly, PTEN levels are not affected in keratinocytes treated with antago-205 (13). Similar to what we observed with SHIP2, decreases in PTEN have been correlated with increased cell migration and increased cortical actin (31). It should be noted that our observations regarding SHIP2's role in keratinocyte migration are in opposition to a recent report describing a positive role for SHIP2 in the migration of MDA-MB-231 breast cancer cells (51). This discrepancy in SHIP2-related migration could reflect inherent differences between primary cultures of keratinocytes vs. transformed cells. Furthermore, SHIP2 has the potential to exert either a positive or a negative influence on migration (51) depending on the levels of the local phosphinositide pools, which can affect cell adhesion and migration (52, 53). Nonetheless, it is clear that in normal keratinocytes, one of miR-205's functions is to reduce SHIP2, which enhances survival and cell migration. It is also recognized that miR-205 has many other potential targets (54) and that some of our findings on keratinocyte migration may be the consequences of a combination of effects on multiple targets.

miR-205 enhancement of cell migration and actin-based motility pathways

It is well established that actin dynamics play a major role in migration and that the changes in actin during this process are highly regulated by a series of actin-associated proteins (ref. 55 and references therein). Therefore, it is not surprising that the suppression of F-actin in HEKs and HCEKs following down-regulation of miR-205 would affect the ability of these keratinocytes to seal linear scratch wounds as well as alter levels of p-Akt, activated Rho, p-cofilin, and ERM, proteins associated with actin remodeling and migration (29, 34, 36, 47, 49, 56). Although regulation of growth and apoptosis are commonly associated with the Akt signaling pathway (57, 58), actin remodeling and cell migration can also be regulated, in part, via Akt signaling (36, 56).

We show that antago-205-treated HEKs have decreased levels of p-Akt and increased p-cofilin expression. Cofilin is regulated by phosphorylation and is rendered inactive in actin binding when phosphorylated (59). Cofilin severs actin filaments and regulates actin polymerization and depolymerization during migration (47). Furthermore, a reduction in cofilin inhibits motility, whereas increases in this protein enhance motility (ref. 60 and references therein). We propose that the alterations in F-actin observed after modulating miR-205 levels are due, in part, to increased p-cofilin (inactive), which diminishes keratinocyte motility. Support for this idea comes from our findings that antago-205-treated HEKs have increased levels of activated RhoA and that antago-205 increases in p-cofilin could be reversed by inhibiting ROCKI.

The phosphoinositide PIP₂ has also been proposed as a means of cofilin regulation via its ability to bind to cofilin and inhibit cofilin-actin interactions in vitro (61, 62). This is reasonable when considered in the context of miR-205/SHIP2 interactions. SHIP2 dephosphorylates PIP₃ to yield PIP₂ (24, 25). Treatment of keratinocytes with antago-205 results in an increase in SHIP2 (13); this could amplify the cellular pool of PIP₂, making more of this phosphoinositideavailable to potentially act to further inhibit cofilin. Regardless of the mechanism, our finding that a connection exists between miR-205, cofilin inactivation, and the actin cytoskeleton is a novel mechanism for the regulation of normal human keratinocyte migration.

The ERM proteins link the cortical cytoskeleton and are involved in establishing a leading edge and polarized migration (63, 64). ERM-directed migration occurs via C-terminal phosphorylation, which results in the activation of this family of proteins (64–67). These proteins are also involved in the Rho and PI3K-Akt signaling pathways (ref. 68 and references therein). Our observation that antago 205-treated HEKs have lowered levels of activated (phosphorylated) ERM proteins is compatible with a decrease in keratinocyte migration. Alternatively, the lowered levels of p-ERM in HEKs following treatment with antago-205 might be related to keratinocyte survival. Recently, it has been shown that p-Akt levels parallel those of p-ERM (69), which is consistent with our observations that antago-205 treatment results in a reduction in p-Akt and p-ERM. Activation of the PI3K/Akt pathway, necessary for cell survival, requires phosphorylation of ezrin (68). This raises the possibility that the pro-cell survival function ascribed to miR-205 in keratinocytes (13) is working through 2 signaling pathways: 1) inactivation of SHIP2, which increases p-Akt (13) and 2) activation of p-ERM, which regulates Akt-dependent cell survival.

miR-205 contributes to corneal epithelial wound healing

The corneal epithelium is unique in that it exhibits distinct as well as overlapping expression of miR-205 and miR-184 (11). Specifically, miR-205, the second most abundant corneal epithelial miRNA, is expressed in all viable layers of the corneal, limbal, and conjunctival epithelia (11). In contrast, mir-184, the most abundant corneal epithelial miRNA, is present only in the basal and immediately suprabasal layers of the corneal epithelium; miR-184 is noticeably absent in the limbal and conjunctival epithelia as well as the epidermis (11). The corneal epithelium is also special in that miR-184 negatively regulates miR-205 to maintain SHIP2 levels (13). Twenty-four hours following corneal wounding in mice, miR-184 was not expressed in the re-epithelialized epithelium but was strongly expressed in the corneal epithelium adjacent to the wound, indicating that wounding down-regulates miR-184 (11). In contrast, miR-205 expression was maintained in the re-epithelialized epithelium and in the corneal epithelium adjacent to the wound (Supplemental Fig. 4). This makes excellent biological sense from the perspective of miR-205's ability to positively influence cell migration. We propose that after a corneal epithelial wound, the antagonistic effect of miR-184 on miR-205 is removed, which ultimately results in the down-regulation of SHIP2; cell migration is enhanced, and this contributes to the rapid healing observed postcorneal wounding. Evidence in support of this hypothesis comes from our in vitro scratch assays where HCEKs treated with antago-184, which is analogous to the 24 h in vivo wound condition, have decreased levels of SHIP2 (13). Such a treatment enhanced the ability of HCEKs to seal wounds (Fig. 3). These findings suggest that miRNAs could be used therapeutically to promote corneal epithelial wounding. Specifically, topical application of antago-184 should “release” miR-205 to down-regulate SHIP2, thereby increasing cell migration and speeding the reepithelialization process.