Lysophosphatidic Acid Promotes Cell Migration through STIM1- and Orai1-Mediated Ca²⁺_i Mobilization and NFAT2 Activation

Abstract

Lysophosphatidic acid (LPA) enhances cell migration and promotes wound healing in vivo, but the intracellular signaling pathways regulating these processes remain incompletely understood. Here we investigated the involvement of agonist-induced Ca²⁺ entry and STIM1 and Orai1 proteins in regulating nuclear factor of activated T cell (NFAT) signaling and LPA-induced keratinocyte cell motility. As monitored by Fluo-4 imaging, stimulation with 10 μℳ LPA in 60 μℳ Ca²⁺_o evoked Ca²⁺_i transients owing to store release, whereas addition of LPA in physiological 1.2 mℳ Ca²⁺_o triggered store release coupled to extracellular Ca²⁺ entry. Store-operated Ca²⁺ entry (SOCE) was blocked by the SOCE inhibitor diethylstilbestrol (DES), STIM1 silencing using RNA interference (RNAi), and expression of dominant/negative Orai1^R91W. LPA induced significant NFAT activation as monitored by nuclear translocation of green fluorescent protein-tagged NFAT2 and a luciferase reporter assay, which was impaired by DES, expression of Orai1^R91W, and inhibition of calcineurin using cyclosporin A (CsA). By using chemotactic migration assays, LPA-induced cell motility was significantly impaired by STIM1, CsA, and NFAT2 knockdown using RNAi. These data indicate that in conditions relevant to epidermal wound healing, LPA induces SOCE and NFAT activation through Orai1 channels and promotes cell migration through a calcineurin/NFAT2-dependent pathway.

Introduction

Lysophosphatidic acid (LPA) is a potent bioactive phospholipid that regulates diverse cellular processes including proliferation, differentiation, adhesion, and migration among several cell types, principally through interaction with a range of specific cell-surface receptors (Noguchi et al., 2009). In the context of wound repair, LPA is released into acute wounds from serum and platelets (Eichholtz et al., 1993; Inoue et al., 1995; Goetzl and An, 1998; Baker et al., 2002; Mazereeuw-Hautier et al., 2005) and is present in physiologically relevant concentrations in blister fluid (Mazereeuw-Hautier et al., 2005). LPA has been shown to promote migration of cultured epidermal keratinocytes (Sauer et al., 2004) and to enhance wound closure following topical application to experimental wounds in rats and mice (Demoyer et al., 2000; Balazs et al., 2001). LPA-induced signaling is complex and involves the activation of specific G protein–coupled receptors (LPA_1–6), as well as transactivation of other plasma membrane receptors (Sekharam et al., 2000; Sauer et al., 2004; Noguchi et al., 2009). Binding of LPA to its specific receptors also elicits mobilization of intracellular calcium (Ca²⁺_i) from IP₃-dependent stores and influx of extracellular Ca²⁺ (Ca²⁺_o) mediated by incompletely defined mechanisms (Noguchi et al., 2009). Keratinocytes are known to express LPA_1–3 and respond to LPA with Ca²⁺ transients (Anliker and Chun, 2004; Roedding et al., 2006; Ross et al., 2007; Lichte et al., 2008), but the physiological relevance of LPA-induced Ca²⁺ fluxes and the downstream signaling pathways in mediating specific physiological effects remain to be defined.

Recent studies have identified the endoplasmic reticulum Ca²⁺ sensor STIM1 and the plasma membrane channel protein Orai1 as critical components of agonist-induced Ca²⁺ entry (Frischauf et al., 2008; Martin et al., 2009), as well as being crucial for tumor cell migration (Vassilieva et al., 2008; Yang et al., 2009; Zuo and Chen, 2009). Our previous data suggested a role for STIM1 in mediating Ca²⁺ entry in keratinocytes (Ross et al., 2007, 2008), but there is currently no information regarding the role of STIM1 or Orai1 in regulating keratinocyte migration.

The calcineurin/nuclear factor of activated T cell (NFAT) signaling cascade is characteristically recruited by the sustained Ca²⁺_i mobilization observed during Ca²⁺ entry (Macian, 2005; Gwack et al., 2007). NFAT factors have been determined to be crucial for regulating cell motility in several cell types (Jauliac et al., 2002; Corral et al., 2007). We and others have shown that NFAT factors are functionally active in keratinocytes and epidermis, and calcineurin/NFAT activity has also been directly implicated in the regulation of keratinocyte growth and differentiation (Santini et al., 2001; Al-Daraji et al., 2002; Mammucari et al., 2005). The relationship between Ca²⁺ fluxes and NFAT activation in keratinocytes remains to be determined, and furthermore, the involvement of calcineurin/NFAT signaling in LPA-induced cell motility remains unknown.

In this study, we therefore investigated the hypothesis that LPA promotes keratinocyte migration through Ca²⁺_i mobilization and examined the role of STIM1, Orai1 and calcineurin/NFAT recruitment within these physiological phenomena.

Results

Stimulation of keratinocytes with LPA in physiological Ca²⁺_o triggers Ca²⁺ entry

As previous studies in lymphocytes have indicated that sustained increases in Ca²⁺_i encountered during Ca²⁺ entry (Gwack et al., 2007) are required for NFAT activation (Dolmetsch et al., 1997), we first determined the experimental conditions in which LPA could evoke Ca²⁺ entry in primary keratinocytes. LPA is known to induce short transient elevations of Ca²⁺_i due to the release of Ca²⁺ from intracellular stores, but Ca²⁺ entry has not been reported (Ross et al., 2007, 2008; Lichte et al., 2008). As the relevance of Ca²⁺_o in the regulation of LPA-induced Ca²⁺ mobilization in keratinocytes has largely been ignored, we investigated LPA-induced Ca²⁺ mobilization in the context of varying Ca²⁺_o.

Keratinocytes were routinely cultured in medium containing 60 μℳ Ca²⁺_o. The addition of 10 μℳ LPA in these conditions induces a transient (±200 seconds) peak elevation of Ca²⁺_i (Figure 1a), indicative of store release as expected (Ross et al., 2007, 2008; Lichte et al., 2008). Adjustment of the Ca²⁺_o to 1.2 mℳ in the absence of LPA resulted in a weak increase in Ca²⁺_i (Figure 1a) over an extended period of time (>10 minutes), similar to previous reports (Bikle et al., 1996; Tu et al., 2001). To mimic the contact of keratinocytes with LPA in a serum-relevant Ca²⁺_o, we analyzed the effect of adding LPA in 1.2 mℳ Ca²⁺_o. Under these conditions, LPA induced a transient peak in Ca²⁺_i, followed by a sustained (>10 minutes) plateau elevation of Ca²⁺, which suggests store release coupled to entry of extracellular Ca²⁺ (Figure 1a).

Figure 1.

Lysophosphatidic acid (LPA) stimulation in 1.2 mℳ Ca²⁺_o evokes Ca²⁺ entry. Keratinocytes were subjected to Ca²⁺_i imaging. (a) Stimulation with 10 μℳ LPA in 60 μℳ or 1.2 mℳ Ca²⁺_o or with 1.2 mℳ Ca²⁺_o. (b) Uncoupling of store release and entry using LPA in Ca²⁺-free medium and adding back Ca²⁺. Ca²⁺ entry was blocked by the addition of 500 μℳ MnCl₂ or 10 μℳ DES. (c) Thapsigargin (Tg)-induced store-operated Ca²⁺ entry was inhibited by adding diethylstilbestrol (DES). (d) LPA added after exposure to Ca²⁺_o during de-esterification (45 minutes, n=49) caused only store release. (e) Cells were exposed to 1.2 mℳ Ca²⁺_o for 60 seconds before LPA (gray, n=134) or LPA was added in 1.2 mℳ Ca²⁺_o (n=83). No significant difference in Ca²⁺ entry was observed (two-way analysis of variance at t=600 seconds).

We then used an addback protocol to verify the occurrence of Ca²⁺ entry by uncoupling store release from influx (Figure 1b). Manganese quenching (Sage et al., 1989) by the addition of 500 μℳ MnCl₂ during the plateau phase quickly reduced Ca²⁺_i to the baseline level (Figure 1b). In addition, we used Ca²⁺ entry–blocking pharmacological agents to further define the Ca²⁺-mobilizing effect of LPA. We initially validated the pharmacological agents through studies on thapsigargin (Tg)-induced Ca²⁺ entry. Although 2-aminoethoxydiphenyl borate has been classically used to block store-operated Ca²⁺ entry (SOCE), this molecule is increasingly recognized to be rather nonspecific (Schindl et al., 2008). We found that 2-aminoethoxydiphenyl borate did not block Tg-induced Ca²⁺ entry at a variety of concentrations (data not shown). Similarly, we found that the addition of gadolinium ions (Gd³⁺) did not reliably block Tg-induced SOCE (data not shown). We therefore used the lesser known SOCE inhibitor diethylstilbestrol (DES), which blocks SOCE channels and inwardly rectifying current but exerts no effect on TRPM channels (Zakharov et al., 2004; Dobrydneva et al., 2010). DES was indeed found to inhibit Tg-induced SOCE in keratinocytes (Figure 1c), and the addition of DES during the plateau phase of LPA-induced Ca²⁺ entry blocked further entry (Figure 1b), providing further evidence that LPA induces Ca²⁺ entry in keratinocytes through SOCE channels.

To further characterize the impact of Ca²⁺_o on LPA-induced Ca²⁺ mobilization, keratinocytes were incubated in medium containing 1.2 mℳ Ca²⁺ for an extended time (45 minutes) before LPA stimulation. Under these conditions, LPA induced store release (Figure 1d) in a manner similar to previous studies (Ross et al., 2007; Lichte et al., 2008), but no entry was apparent during the observed time span (>10 minutes). Next, we uncoupled the addition of agonist and Ca²⁺ by adjusting the Ca²⁺_o to 1.2 mℳ for 60 seconds before adding LPA (Roedding et al., 2006) (Figure 1e). There was no significant difference detected between Ca²⁺_i fluxes observed in these conditions and when LPA and millimolar Ca²⁺ were added simultaneously, suggesting that simultaneous stimulation of LPA receptors and adjustment to millimolar Ca²⁺ does not just overwhelm the plasma membrane channels and thereby cause nonspecific entry.

LPA induces nuclear translocation of NFAT2 and upregulates NFAT transcriptional activity

We then investigated the effect of LPA on NFAT activation. First, we used a real-time live-cell imaging method to simultaneously analyze Ca²⁺_i variations (using FuraRed) and NFAT activation (by visualizing nuclear translocation of green fluorescent protein (GFP)-tagged NFAT2). The proof-of-principle of this method was achieved by triggering Tg-mediated SOCE, which resulted in a sustained Ca²⁺_i increase as expected. This correlated with progressive and strong nuclear accumulation of NFAT2-GFP within 10 minutes of exposure (Figure 2a).

Figure 2.

LPA-induced Ca²⁺ mobilization correlates with nuclear translocation of GFP-tagged nuclear factor of activated T cell 2 (NFAT2). Keratinocytes expressing green fluorescent protein (GFP)-tagged NFAT2 were subjected to Ca²⁺_i imaging using FuraRed. Ca²⁺_i variations are indicated as 1/(F_t/F₀) (red data points), whereas the ratio of nuclear versus cytoplasmic localization of NFAT2-GFP is represented by green data points±SEM in histograms. Representative FuraRed and NFAT2-GFP snapshots are shown for illustration purposes. (a) For proof-of-principle, store-operated Ca²⁺ entry was induced using 1 μℳ thapsigargin in 1.2 mℳ Ca²⁺_o. (b) Stimulating the cells with 10 μℳ lysophosphatidic acid (LPA) in 1.2 mℳ Ca²⁺_o resulted in store release and influx correlating with NFAT2 translocation. (c) Adjusting Ca²⁺_o to 1.2 mℳ did not trigger noticeable NFAT2 translocation. Bar=10 μm.

Simultaneous stimulation with LPA and 1.2 mℳ Ca²⁺_o resulted, as expected, in Ca²⁺ store release coupled to entry, which correlated with progressively increasing nuclear accumulation of NFAT2 (Figure 2b). Adjustment of Ca²⁺_o to 1.2 mℳ did not induce significant NFAT2 translocation within 10 minutes of treatment (Figure 2c). In all experimental conditions, nuclear translocation of NFAT2 was prevented by preincubating the cells for 1 hour with the calcineurin inhibitor cyclosporin A (CsA, 1 μℳ) before agonist challenge (data not shown).

Second, we investigated the effect of LPA on NFAT activity using a luciferase reporter under the transcriptional control of NFAT (Wilkins et al., 2004). LPA stimulation in 1.2 mℳ Ca²⁺_o prompted a significant and sustained upregulation of NFAT activity to a level ∼10-fold over baseline within 4 hours after stimulation (Figure 3a). In comparison, LPA stimulation in 60 μℳ Ca²⁺_o increased NFAT activity within 4 hours before reaching a plateau level ∼2.5-fold over the control level (Figure 3a). Adjustment of Ca²⁺_o to 1.2 mℳ in the absence of LPA induced a weak increase of NFAT activity (∼2.5-fold) over 24 hours (Figure 3a). As shown in Figure 3b, preincubation of keratinocytes for 1 hour with 1 μℳ CsA before LPA and millimolar Ca²⁺_o challenge prevented the induction of NFAT transcriptional activity. To further investigate the role of Ca²⁺ entry in LPA-induced NFAT activation, DES was added during agonist challenge, which resulted in a significantly impaired induction of NFAT activity (Figure 3c).

Figure 3.

Lysophosphatidic acid (LPA)-induced Ca²⁺ entry triggers sustained calcineurin-dependent upregulation of nuclear factor of activated T cell (NFAT) transcriptional activity. (a) Keratinocytes were stimulated as indicated and assessed for NFAT-dependent luciferase activity expressed as mean±SEM. (b, c) Calcineurin activity (b) and Ca²⁺ entry (c) were required for LPA-induced NFAT activation (*P<0.05, **P<0.01, one-way analysis of variance).

Characterization of the mechanism of LPA-induced Ca²⁺ entry

The implication of STIM1 in LPA-induced Ca²⁺ entry was studied by analyzing the effect of RNA interference (RNAi-mediated) knockdown of STIM1 expression using nucleofection, which achieves efficient transfection of primary keratinocytes (Supplementary Figure S1 online). Figure 4a–c illustrates that keratinocytes transfected with small interfering RNA (siRNA) directed against STIM1 exhibit a significant decrease in STIM1 mRNA and protein content 24 hours post transfection. Figure 4d shows that LPA-induced Ca²⁺ entry was impaired in STIM1 siRNA-transfected keratinocytes compared with scrambled siRNA-transfected cells.

Figure 4.

Lysophosphatidic acid (LPA)-induced Ca²⁺ entry requires sufficient levels of STIM1 and functional Orai1. (a–c) Relative efficiency of small interfering RNA (siRNA)-mediated STIM1 silencing compared to scrambled (scr) control by quantitative PCR and western blotting. (e) Fura-PE3-based Ca²⁺ imaging showed impaired LPA-induced Ca²⁺ entry in STIM1-knockdown cells. (e, f) Efficient transfection of Orai1-encoding plasmids demonstrated by western blotting for myc-tag. (g) LPA-induced Ca²⁺ entry was found to be blocked in Orai1^R91W-expressing cells. (h) Keratinocytes were cotransfected with Orai1, nuclear factor of activated T cell (NFAT)-directed, and Renilla luciferase plasmids. Cells were then treated with LPA and 1.2 mℳ Ca²⁺ and assessed for NFAT transcriptional activity. Overexpression of Orai1^R91W significantly impaired NFAT activation. Data represent mean±SEM; n⩾3 unless otherwise stated, *P<0.05, **P<0.01, ***P<0.001, analysis of variance.

The involvement of Orai1 in LPA-induced Ca²⁺ influx was investigated by evaluating the impact of overexpressing Orai1^R91W, an Orai1 mutant that substantially attenuates Ca²⁺ entry through Orai1 channels through a dominant-negative effect (Liao et al., 2007). Figure 4e and f illustrates efficient transfection of Orai1-encoding plasmids to keratinocytes by nucleofection. As shown in Figure 4g, LPA-induced Ca²⁺ influx was inhibited in Orai1^R91W-expressing keratinocytes, whereas entry levels were comparable in wild-type Orai1-expressing and empty vector–transfected cells. Notably, subsequent LPA-induced upregulation of NFAT transcriptional activity was found to be impaired in Orai1^R91W-expressing cells, but not in wild-type Orai1- or empty vector–transfected cells (Figure 4h). A similar effect was observed following siRNA-mediated knockdown of Orai1 (Supplementary Figure S2 online).

LPA-induced keratinocyte migration requires STIM1, calcineurin, and NFAT2

To test the involvement of STIM1 in mediating LPA-induced keratinocyte migration, keratinocytes were subjected to STIM1 knockdown before performing three-dimensional (3D) chemotactic migration assays. Cells were allowed to migrate over 14 hours in medium containing 60 μℳ Ca²⁺ or 1.2 mℳ Ca²⁺ to mimic the contact of keratinocytes with LPA in high Ca²⁺_o, as would occur following wounding of skin. As described previously, exposure to millimolar Ca²⁺_o dampened overall migration rates (Figure 5a and b) (Magee et al., 1987). In both Ca²⁺_o conditions, LPA treatment induced a significant increase in keratinocyte motility, as expected (Sauer et al., 2004). However, LPA-induced migration was significantly impaired in the 3D assays (Figure 5b) following STIM1 knockdown, suggesting that normal levels of STIM1 are required for LPA-induced keratinocyte motility.

Figure 5.

Lysophosphatidic acid (LPA)-induced keratinocyte migration requires STIM1, calcineurin activity, and sufficient levels of nuclear factor of activated T cell 2 (NFAT2). (a–d, h, i) Keratinocyte cultures were subjected to three-dimensional (3D) chemotactic migration assays, which were performed in medium containing 60 μℳ or 1.2 mℳ Ca²⁺. The average numbers of migrated cells are represented as bar graphs±SEM. Keratinocytes were treated with scrambled (scr) or STIM1-targeted small interfering RNA (siRNA) (a, b) or NFAT2-targeted siRNA (h, i) for 24 hours prior to 3D chemotactic migration assays. In both Ca²⁺ conditions, 10 μℳ LPA significantly increased migration rates. RNA interference (RNAi)-mediated knockdown of STIM1 and NFAT2 resulted in a significant impairment of LPA-induced keratinocyte motility (a, b, h, i, ***P<0.001, **P<0.01, two-way analysis of variance (ANOVA)). (c, d) Preincubation of cells with 1 μℳ cyclosporin A significantly impaired LPA-induced migration (n⩾3, **P<0.01, *P<0.05, two-way ANOVA). (e–g) Keratinocytes were transfected using scr or NFAT2-targeted siRNA. After 24 hours, cells were subjected to (e) real-time PCR analysis of NFAT2 expression (n=6), (f) western blot analysis, and (g) densitometry (n=3). RNAi-mediated knockdown of NFAT2 is significant (*P<0.05, t-test). Bar=50 μm.

To investigate the involvement of the calcineurin/NFAT pathway in mediating LPA-induced migration, cells were pretreated using the calcineurin inhibitor CsA before being subjected to two-dimensional scratch wounding (Supplementary Figure S3 online) and 3D chemotactic migration assays (Figure 5c and d). LPA-induced migration was significantly impaired in CsA-treated keratinocytes. Moreover, RNAi-mediated knockdown of NFAT2, the efficiency of which was verified by real-time PCR (Figure 5e) and western blotting (Figure 5f and g), significantly prevented the induction of 3D migration of keratinocytes by LPA in 60 μℳ and 1.2 mℳ Ca²⁺_o (Figure 5h and i).

Discussion

The data presented in this paper (schematically summarized in Figure 6) demonstrate that stimulation of keratinocytes with LPA induces Ca²⁺_o-regulated mobilization of Ca²⁺_i, leading to the activation of calcineurin/NFAT2 signaling and keratinocyte migration. LPA stimulation in physiological millimolar Ca²⁺_o evoked agonist-induced Ca²⁺ entry, which required sufficient levels of STIM1 and functional Orai1. Downstream of LPA-induced Ca²⁺ mobilization, we observed nuclear translocation of NFAT2 and activation of NFAT-dependent transcription. Finally, LPA-induced keratinocyte motility required sufficient levels of STIM1 and NFAT2 and calcineurin activity. Our study thus provides important insight into the signaling mechanism triggered by LPA to promote keratinocyte migration.

Figure 6.

Proposed mechanism of lysophosphatidic acid (LPA)-induced Ca²⁺ signaling involved in promoting keratinocyte migration. On the basis of our results, this model proposes that in physiological conditions, LPA promotes migration of keratinocytes through Ca²⁺ mobilization mediated by STIM1 and Orai1 and subsequent activation of calcineurin and nuclear factor of activated T cell 2 (NFAT2). ER, endoplasmic reticulum.

The physiology of keratinocytes is well known to be tightly regulated by Ca²⁺_o (Tu et al., 2004). Consistent with this, our data show differential Ca²⁺ mobilization and NFAT activation upon stimulation with LPA in physiological millimolar Ca²⁺_o compared with micromolar levels. Indeed, LPA induced Ca²⁺ entry when the Ca²⁺_o was adjusted to 1.2 mℳ simultaneously with or briefly before stimulation. However, keratinocytes left in 1.2 mℳ Ca²⁺_o for a prolonged time before LPA stimulation only exhibited store release, but no Ca²⁺ entry was detectable within the observed time span, which is in accordance with previous studies (Ross et al., 2007, 2008; Lichte et al., 2008) and suggests that prolonged exposure to millimolar Ca²⁺ results in the inhibition of acute LPA-induced Ca²⁺ entry. Our data indicating NFAT activation caused by Ca²⁺ entry are consistent with previous studies performed in other epithelial cell types, including prostate (Thebault et al., 2006) and kidney (Schlondorff et al., 2009) cells.

The role of store- and receptor-operated Ca²⁺ entry (SOCE/ROCE) during physiological activation of primary cells has not been extensively investigated, and thus there is little information on the roles of STIM and Orai proteins in primary cells except in hepatocytes, in which STIM1 and Orai1 are involved in mediating Ca²⁺ entry leading to agonist-induced oscillations (Jones et al., 2008). To our knowledge, the establishment of functional roles for STIM1 and Orai1 in primary keratinocytes as mediators of agonist-induced Ca²⁺ entry has previously been unreported. Our data showing that LPA-induced Ca²⁺ entry depends on normal levels of STIM1 and functional Orai1 are consistent with studies in which STIM1 knockdown (Hirano et al., 2009; Potier et al., 2009) and expression of dominant/negative Orai1^R91W (Liao et al., 2007, 2008) lead to the inhibition of SOCE or ROCE in a variety of cell lines.

Our data concur with an increasing body of evidence indicating an emerging role for Ca²⁺ signaling as a critical regulator of cell motility in a variety of physiological and pathological contexts. For instance, STIM1 and Orai1 have been found to mediate agonist-induced Ca²⁺ entry and regulate Ca²⁺ flux–dependent cell migration in epithelial tumor cells (Yang et al., 2009) and vascular smooth muscle cells (Potier et al., 2009). In addition, functional inhibition of calcineurin and NFAT2 impaired the migration rate of epithelial tumor cells (Jauliac et al., 2002), and wound closure in scratched myocyte cultures was reduced by silencing NFAT2 expression (Chow et al., 2008).

NFAT transcription factors have been previously attributed to important roles in the control of keratinocyte growth and differentiation (Santini et al., 2001; Mammucari et al., 2005; Horsley et al., 2008) and in the response of keratinocytes to UV-mediated stress (Canning et al., 2006; Flockhart et al., 2008). In conclusion, our work underscores the importance of Ca²⁺ signaling and NFAT factors in the control of epidermal physiology and potentially in cutaneous wound healing.

Materials and Methods

Reagents

All chemicals were purchased from Sigma-Aldrich (Gillingham, UK) unless otherwise specified.

Cell culture and treatments

Normal human epidermal keratinocytes were isolated from redundant foreskins (Todd and Reynolds, 1998) or breast or abdominal skin (Jans et al., 2004), and primary cultures were expanded in human keratinocyte growth supplement EpiLife medium (Invitrogen, Paisley, UK). The study adhered to the Declaration of Helsinki protocols for use of human materials, was approved by the Newcastle and North Tyneside local ethics committee, and written informed patient consent was obtained. Keratinocytes were used between passage 2 and 5. Treatments with agonists were performed by diluting the agonists directly in medium. For inhibition of calcineurin activity, cells were pretreated with 1 μℳ CsA (Calbiochem, San Diego, CA) for 1 hour before agonist challenge. The concentration of LPA used throughout the study was 10 μℳ, which is a generally accepted concentration (Ross et al., 2007, 2008; Lichte et al., 2008).

Ca²⁺ imaging

Keratinocytes seeded in Willco glass-bottomed dishes (Intracel, Royston, UK) were subjected to Ca²⁺_i imaging using Fluo-4-AM (Invitrogen) or Fura-PE3-AM (Calbiochem) as described (Ross et al., 2007). Changes in Ca²⁺_i were monitored at 10-second intervals with a Pathway HTS imaging system (Becton-Dickinson, Rockville, MD). Images were captured using an Olympus (Melville, NY) × 20 objective and a Hamamatsu 1394 (Hamamatsu, Japan) ORCA-ERA CCD camera. Fluo-4 was excited at 488 nm, collecting emitted fluorescence through a 515LP filter. Fura-PE3 was excited at 340 and 380 nm, and emitted fluorescence was collected through the 515LP filter. Quantification was performed using Volocity (Improvision, Coventry, UK). When using Fluo-4, changes in Ca²⁺_i are expressed as the ratio of the initial fluorescence to the temporal fluorescence (F_t/F₀). When using Fura-PE3, changes in Ca²⁺_i are expressed as ratiometric measurements (F_340 nm/F_380 nm).

Dual imaging of Ca²⁺_i and NFAT2-GFP

Keratinocytes transduced with NFAT2-GFP-delivering retroviruses (Flockhart et al., 2008) were seeded on Willco dishes. After 24 hours, the cells were loaded using 5 μℳ FuraRed-AM (Invitrogen). Dual imaging was then performed using a Leica TCS SP2 confocal laser scanning microscope equipped with an argon laser (Leica, Milton Keynes, UK). Fluorescence excitation of GFP and FuraRed was performed with the 488-nm line of the laser. Fluorescence emission of GFP was collected through a 500–550-nm window of the detector, whereas FuraRed emission was collected through a 650–700-nm window. Images were captured with a × 63 Plan Apo objective (NA1.32) (Leica, Milton Keynes, UK) at 10-second intervals. Changes in Ca²⁺_i are expressed as the inverted ratio of the initial fluorescence to the temporal fluorescence (1/(F_t/F₀)). The nuclear-to-cytoplasmic ratio of NFAT2-GFP was assessed by outlining the nucleus and cytoplasm of each cell as separate region of interest, calculating the mean green fluorescence in each regions of interest and frame in Volocity. Nuclear-to-cytoplasmic ratios were then calculated in Microsoft Excel.

NFAT transcriptional activity assay

NFAT transcriptional activity was measured using an NFAT firefly luciferase reporter plasmid (pGL3; Promega, Southampton, UK), containing nine copies of an NFAT-binding site from the IL-4 promoter (5′-TGGAAAATT-3′) positioned 5′ to a minimal promoter from the α-myosin heavy-chain gene, gratefully obtained from J Molkentin, Cincinnati, OH (Wilkins et al., 2004). Transfection efficiency and cell viability were controlled by cotransfecting a Renilla luciferase control vector (pRLTK; Promega). Keratinocytes were seeded in 12-well plates and transfected using 0.5 μg of firefly reporter DNA plus 0.05 μg of Renilla luciferase DNA using Fugene 6 (Roche Applied Sciences, Burgess Hill, UK) at a 6:1 Fugene:DNA ratio. After 24 hours, the cells were stimulated as indicated, lysed, and assayed for dual luciferase activity. Luciferase assays were performed using the dual luciferase assay system (Promega). Reporter firefly luciferase values were normalized to the Renilla values.

RNA interference to knock down STIM1 and NFAT2 expression

Chemically modified (to prevent off-target effects) scrambled siRNA and NFAT2-specific siRNA were obtained from Qiagen (Crawley, UK) (Flockhart et al., 2009). STIM1-targeted siRNA (5′-GAAGCCUCUUAAGAAGUAG-3′ and 5′-AGACCUCAAUUACCAUGAC-3′) was obtained from Eurofins/MWG (Ebersberg, Germany). siRNA was transfected into keratinocytes using the Human Keratinocyte Nucleofector kit and device from Amaxa/Lonza (Cologne, Germany) according to the manufacturer's protocol (2 μg siRNA per 10⁶ cells). For STIM1, 1 μg of each siRNA was used to transfect 10⁶ cells. After 24 hours, the cells were trypsinized and subjected to the 3D chemotactic migration assay. Knockdown of STIM1 and NFAT2 was verified by real-time PCR and western blotting using standard techniques (Hampton et al., 2012) and anti-STIM1 antibody (clone 5A2; Abnova, Taipei, Taiwan) or anti-NFAT2 antibody (a kind gift from Dr Nancy Rice) (Hampton et al., 2012), respectively. For NFAT2 real time, Poly-A RNA was isolated using a Direct mRNA isolation kit (Sigma-Aldrich, Gillingham, UK), retrotranscribed using a High-Capacity cDNA RT kit (Applied Biosystems, Paisley, UK) and subjected to real-time PCR using iQ SYBR Green Supermix and a Chromo4 cycler (Bio-Rad, Hemel Hempstead, UK). Primers used to detect NFAT2 were 5′-CTTCTCCAACACCAAAGTCC-3′ (forward) and 5′-CGTACCCGTGTGTTCTTCCT-3′ (reverse). NFAT2 expression was normalized to glyceraldehyde-3-phosphate dehydrogenase expression (forward primer: 5′-GTCAGTGGTGGACCTGACCT-3′ and reverse primer: 5′-AGGGGTCTACATGGCAACTG-3′). For STIM1 real time, total RNA was isolated using the Arcturus PicoPure RNA Isolation kit (Applied Biosystems) and reverse transcribed using the SuperScript III First-Strand Synthesis System (Invitrogen); TaqMan gene-expression assays Hs00162394_m1 (STIM1) and Hs99999905_m1 (glyceraldehyde-3-phosphate dehydrogenase) (Applied Biosystems) were performed on a 7900HT system (Applied Biosystems).

Effect of dominant/negative Orai1^R91W on LPA-induced Ca²⁺ entry

pCMV-myc plasmids encoding wild-type Orai1 or mutated dominant/negative Orai1^R91W were kindly donated by Professor L Birnbaumer (Liao et al., 2007) (NIEHS, Research Triangle Park, NC). Keratinocytes (10⁶ cells) were transfected by nucleofection with 2 μg of Orai1-encoding plasmid and 0.5 μg of pEGFP-N1 (Clontech, Mountain View, CA) to mark the transfected cells (Zarayskiy et al., 2007). The expression of myc-tagged Orai1-encoding plasmids was verified by standard western blotting using anti-myc antibody (clone 9E10; Santa Cruz Biotechnology, San Diego, CA). Cells were seeded into Willco dishes at 0.5 × 10⁶ cells per dish. After 24 hours, the cells were processed for Ca²⁺_i imaging using Fura-PE3. Before performing Ca²⁺ imaging, the sample field was analyzed for GFP expression using the appropriate filter sets.

Migration assays

Keratinocyte motility was assessed using standard two-dimensional scratch wounding and 3D chemotactic migration assays. Chemotactic migration of keratinocytes was assayed in a manner similar to that described in Sauer et al., 2004. Briefly, 20,000 keratinocytes were seeded on top of uncoated Transwell filters (12 mm diameter, 8 μm pore size; Corning Life Sciences, Schiphol-Rijk, The Netherlands) in a 10-μl drop. Care was taken to deposit the drop in the middle of the filter without it touching the plastic support walls, in order to prevent a meniscus effect causing unequal distribution. The cells were left to attach for 30 minutes at room temperature before adding 300 μl of EpiLife medium containing agonists into the lower compartment. The cells were left to migrate overnight at 37 °C before fixation using 4% (w/v) formaldehyde. Unmigrated cells were scraped off the top portion of the filters using cotton buds. The filters were stained using hematoxylin and eosin, cut out and mounted using DPX. The migrated cells were counted on a Leica photomicroscope using a × 10 objective (3–5 fields per sample).

Statistical analysis

Statistical analysis was performed using Prism 5 (GraphPad Software, San Diego, CA). Data represent mean±SEM. Experiments were repeated at least three times using different keratinocyte strains unless otherwise specified.

Glossary

CsA

cyclosporin A

DES

diethylstilbestrol

LPA

lysophosphatidic acid

NFAT

nuclear factor of activated T cell

RNAi

RNA interference

SOCE

store-operated Ca²⁺ entry

STIM1

stromal interaction molecule 1

Tg

thapsigargin

The authors state no conflict of interest.