Dihydrosphingosine 1-phosphate has a potent antifibrotic effect in scleroderma fibroblasts via normalization of phosphatase and tensin homolog levels

Authors


Abstract

Objective

Previous studies have revealed a phosphatase and tensin homolog (PTEN)–dependent interaction between the sphingolipid agonist dihydrosphingosine 1-phosphate (dhS1P) and the transforming growth factor β/Smad3 signaling pathway. This study was undertaken to examine responses of systemic sclerosis (SSc) fibroblasts to sphingosine 1-phosphate (S1P) and dhS1P and to gain further insight into the regulation of the S1P/dhS1P/PTEN pathway in SSc fibrosis.

Methods

Fibroblast cultures were established from skin biopsy samples obtained from patients with SSc and matched healthy controls. Western blotting and quantitative polymerase chain reaction were used to measure protein and messenger RNA levels, respectively. PTEN protein was examined in skin biopsy samples by immunohistochemistry.

Results

PTEN protein levels were low in SSc fibroblasts and correlated with elevated levels of collagen and phospho-Smad3 and reduced levels of matrix metalloproteinase 1 (MMP-1). Treatment with dhS1P restored PTEN levels and normalized collagen and MMP-1 expression, as well as Smad3 phosphorylation status in SSc fibroblasts. S1P was strongly profibrotic in SSc and control fibroblasts. Distribution of S1P receptor isoforms was altered in SSc fibroblasts, which had reduced levels of S1P receptor 1 and S1P receptor 2 and elevated levels of S1P receptor 3. Only depletion of S1P receptor 1 abrogated the effects of dhS1P and S1P in control dermal fibroblasts. In contrast, depletion of either S1P receptor 1 or S1P receptor 2 prevented the effects of S1P and dhS1P in SSc fibroblasts.

Conclusion

Our findings demonstrate that PTEN deficiency is a critical determinant of the profibrotic phenotype of SSc fibroblasts. The antifibrotic effect of dhS1P is mediated through normalization of PTEN expression, suggesting that dhS1P or its derivatives may be effective as therapeutic antifibrotic agents. The distribution and function of S1P receptors differ in SSc and healthy fibroblasts, suggesting that alteration in the sphingolipid signaling pathway may contribute to SSc fibrosis.

Systemic sclerosis (SSc) is a connective tissue and autoimmune disease of unknown etiology characterized by severe and often progressive cutaneous and visceral fibrosis, pronounced alterations in the microvasculature, and numerous cellular and humoral immune abnormalities (1). Excessive scarring due to overproduction of extracellular matrix (ECM) proteins is a hallmark of SSc (1). The molecular basis of fibrosis is still incompletely understood; however, there is a general consensus that transforming growth factor β (TGFβ) plays a central role in the development of SSc and other fibrotic diseases (2). TGFβ is a potent inducer of ECM and, under physiologic conditions such as wound repair, its presence is required for fibroblast induction and ECM production and contraction.

Canonical TGFβ signaling is a simple cascade that is initiated by ligand binding to TGFβ receptor type II (TGFβRII), which phosphorylates receptor type I, resulting in binding and phosphorylation of signal transducers, R-Smads, which then interact with common Smad4, translocate to the nucleus, and regulate target gene expression (3). Furthermore, more recent studies have revealed additional modes of TGFβ signaling that involve noncanonical, non-Smad pathways, including MAP kinase, Rho-like GTPase signaling, and the phosphatidylinositol 3-kinase (PI 3-kinase)/Akt pathway (4). Activation of canonical and noncanonical pathways has been demonstrated in SSc fibroblasts, including elevated levels of phosphorylated Smad3 as well as Smad1, and constitutive activation of downstream effectors of PI 3-kinase signaling such as Akt and c-Abl (2).

Sphingosine kinase is a lipid kinase that catalyzes formation of 2 bioactive lipid mediators, sphingosine 1-phosphate (S1P) and dihydrosphingosine 1-phosphate (dhS1P). Sphingosine kinase and its metabolite, S1P, have emerged as important regulators of a wide range of biologic processes, including cell growth and proliferation, cell survival and apoptosis, calcium homeostasis, angiogenesis, and vascular remodeling (5). Recent evidence suggests that S1P may also play an important role in fibrosis through crosstalk with the TGFβ pathway. It was initially reported that in keratinocytes and mesangial cells, S1P mimics the effects of TGFβ through cross-activation of Smad signaling (6, 7). In mesangial cells these effects of S1P were mediated through S1P receptor 3 and were dependent on the presence of TGFβRII (7). Mimetic of S1P, FTY720 (Fingolimod), was shown to have similar profibrotic effects, including Smad phosphorylation and up-regulation of CCN2 and collagen (8). Furthermore, similar to TGFβ, both S1P and FTY720 induced fibroblast-to-myofibroblast differentiation via activation of the S1P receptor (9). Other studies have demonstrated that sphingosine kinase was up-regulated during bleomycin-induced lung fibrosis and contributed to the TGFβ-induced myofibroblast differentiation of lung fibroblasts (10).

Our previous work has shown that, in contrast to the profibrotic function of S1P, dhS1P elicits potent antifibrotic effects through several mechanisms. TGFβ/Smad signaling and collagen synthesis in dermal fibroblasts are inhibited by dhS1P through a mechanism that involves the tumor suppressor phosphatase and tensin homolog (PTEN) (11). We have discovered a novel function of PTEN as a cofactor of the Smad3 phosphatase protein phosphatase 1A (PPM1A). Upon translocating into the nucleus, PTEN forms complexes with PPM1A and protects it against degradation in response to TGFβ signaling, thus resulting in Smad2/3 dephosphorylation (11). We have also shown that S1P and dhS1P have opposing roles in the regulation of the matrix metalloproteinase 1 (MMP-1)/tissue inhibitor of metalloproteinases 1 (TIMP-1) pathway in dermal fibroblasts (12, 13). TGFβ enhanced sphingosine kinase 1 activity and S1P production and induced prolonged up-regulation of sphingosine kinase 1 expression. In addition, we demonstrated that sphingosine kinase 1 was required for the TGFβ-induced up-regulation of TIMP-1 (13). Conversely, dhS1P up-regulated MMP-1 via activation of ERK/Ets1 signaling and was required for the tumor necrosis factor α–induced production of MMP-1 (12).

Existing evidence suggests that the sphingosine kinase metabolites, S1P and dhS1P, have distinct and often opposite effects on the TGFβ signaling pathway. Given the importance of TGFβ signaling in SSc fibrosis, the goal of this study was to evaluate the effect of dhS1P and S1P on the fibrotic features of SSc fibroblasts. Our study shows that significantly reduced protein levels of PTEN are found in SSc fibroblasts. Treatment with dhS1P normalized the profibrotic characteristics of SSc fibroblasts through the up-regulation of PTEN protein, while having only small effects in healthy dermal fibroblasts. S1P induced profibrotic changes in healthy fibroblasts as well as SSc fibroblasts. We also observed alterations in the distribution and utilization of the S1P receptor isoforms in healthy and SSc fibroblasts, suggesting that dysregulation of sphingolipid signaling may contribute to SSc fibrosis.

MATERIALS AND METHODS

Cell culture.

Human dermal fibroblast cultures were established from skin biopsy specimens obtained from the dorsal forearm of patients with diffuse cutaneous SSc (dcSSc) and from age-, race-, and sex-matched healthy controls. Informed consent was obtained from all subjects, and the study was conducted in compliance with Institutional Review Board guidelines. All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for the diagnosis of dcSSc (14). Dermal fibroblasts were cultured from the biopsy specimens as described previously (15). Normal and SSc skin fibroblasts were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic/antimycotic solution. For experiments, cells were incubated with serum-free media for 24 hours before specific treatments.

Immunohistochemistry.

The study group consisted of 7 patients with dcSSc and 7 healthy volunteers (Table 1). Skin biopsy specimens were embedded in paraffin and used for immunohistochemistry. Immunohistochemical staining of paraffin-embedded sections was performed using a Vectastain ABC kit according to the recommendations of the manufacturer (Vector). Four-micrometer–thick sections were mounted on silane-coated slides, then deparaffinized with Histoclear and rehydrated in a graded series of solutions of ethyl alcohol and phosphate buffered saline (PBS). Skin sections were treated with hydrogen peroxide for 30 minutes to block endogenous peroxidase activity and then subjected to a 45-minute antigen-retrieval treatment with antigen unmasking solution (Vector). Incubation with PTEN antibody (Cell Signaling Technology) was performed overnight in a humidified chamber at 4°C as previously described (16). After 3 rinses in PBS, binding sites of the primary antibodies were detected with biotinylated IgG, and the sites of peroxidase activity were visualized by using diaminobenzidine. The sections were then counterstained with hematoxylin. Immunostaining was detected by light microscopy. Normal rabbit IgG was used as a negative control (results not shown).

Table 1. PTEN levels in dermal fibroblasts in normal and SSc skin*
SampleAge, yearsSexRaceDisease duration, yearsTSSPTEN level
  • *

    PTEN = phosphatase and tensin homolog; SSc = systemic sclerosis; TSS = total skin score; NS = normal skin; ND = not determined.

  • − = no staining or little staining in <10% of cells; −/+ = faint, partial staining in >20% of cells; + = moderate, complete staining in >20% of cells; ++ = moderate to strong staining in >50% of cells; and +++ = strong staining in >50% of cells.

NS35237FAfrican American+++
SSc35138FAfrican American23+
NS35660FWhite++
SSc35563FWhite0.3337+/−
NS36040FAfrican American+
SSc35942FAfrican American315+/−
NS36244FAfrican American+/−
SSc36149FAfrican American424+/−
NS36454MWhite++
SSc36350MWhite1.53+
NS36648MWhite+++
SSc36542MWhite2ND
NS36845FWhite+
SSc36750FWhite1334

Reagents.

The following antibodies were used: anti-phospho Smad3, Smad2/3, and PTEN (Cell Signaling Technology), S1P receptor 1 (Santa Cruz Biotechnology), MMP-1 (Chemicon), collagen (Southern Biotechnology), PPM1A (Abcam), and β-actin (clone AC-150; Sigma). Recombinant human TGFβ1 was obtained from R&D Systems, S1P and dhS1P were from Avanti, PTEN small interfering RNA (siRNA) was from Cell Signaling Technology, and siRNA for S1P receptors 1, 2, and 3 were from Santa Cruz Biotechnology. Tissue culture reagents, DMEM, and 100× antibiotic/antimycotic solution (penicillin/streptomycin and amphotericin B) were obtained from Gibco BRL, and FBS was purchased from Hyclone. Enhanced chemiluminescent reagent and bovine serum albumin (BSA) protein assay reagent were obtained from Pierce. TriReagent was purchased from the Molecular Research Center. Primers were purchased from Operon.

Immunoblotting.

Whole cell extracts were prepared from fibroblasts using lysis buffer with the following composition: 1% Triton X-100, 50 mM Tris HCl [pH 7.4], 150 mM NaCl, 3 mM MgCl2, 1 mM CaCl2, proteinase inhibitor cocktail (Roche), and 1 mM phenylmethylsulfonyl fluoride. Protein extracts were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto nitrocellulose membranes. Membranes were incubated overnight with primary antibody, washed, and incubated for 1 hour with secondary antibody. After washing, visualization was performed by enhanced chemiluminescence (Pierce).

Small interfering RNA silencing.

For the inhibition of gene expression using specific siRNA reagents, dermal fibroblasts were grown to 80–90% confluence, serum starved for 4 hours, and transiently transfected using HiPerFect (Qiagen) with 20 μm of the gene-specific siRNA, or the corresponding concentration of scrambled nonsilencing siRNA. Twenty-four hours later, medium was changed to 10% FBS, and cells were harvested 72 hours after transfection.

Real-time polymerase chain reaction (PCR).

Total RNA was isolated from dermal fibroblasts using TriReagent according to the recommendations of the manufacturer (MRC). Two micrograms of RNA was reverse-transcribed in a 20-μl reaction mixture using random primers and a Transcriptor First Strand synthesis kit (Roche Applied Sciences). Quantitative PCR was carried out using iQ SYBR Green mixture (Bio-Rad) on an iCycler PCR machine (Bio-Rad) using 1 μl of complementary DNA in triplicate, with β-actin as the internal control. The primers used were as follows: for S1P receptor 1, forward AACTTCGCCCTGCTTGAG and reverse TCCAGGCTTTTTGTGTAGCTT; for S1P receptor 2, forward GGCCTTCATCGTCATCCTC and reverse CGCAATGAGCACCAGAAG; and for S1P receptor 3, forward TGATGAG- ATGAAACCTATTTGTAAGG and reverse CAAGAAGGCAACAGAAATGCT.

RESULTS

Treatment of SSc fibroblasts with dhS1P normalizes the fibrotic phenotype.

We have recently reported that dhS1P inhibits TGFβ-induced Smad3 signaling and collagen up-regulation in human foreskin fibroblasts through a PTEN/PPM1A-dependent pathway (11). Since SSc fibroblasts are characterized by constitutively activated TGFβ signaling, we investigated whether dhS1P would be effective in inhibiting this pathway in SSc fibroblasts. SSc and closely matched healthy control fibroblasts were treated with increasing doses (0–1 μM) of dhS1P for 12–48 hours. As previously demonstrated (17, 18), SSc fibroblasts expressed elevated levels of phospho-Smad3. Treatment with dhS1P abrogated Smad3 phosphorylation in a dose- and time-dependent manner (Figure 1A).

Figure 1.

Dihydrosphingosine1-phosphate (dhS1P) reverses constitutive phosphorylation of Smad3 in systemic sclerosis (SSc) fibroblasts through up-regulation of phosphatase and tensin homolog (PTEN)/protein phosphatase 1A (PPM1A) protein levels. A, Western blot analysis of PTEN, PPM1A, phospho-Smad3, and total Smad3 in healthy (normal skin [NS]) and SSc fibroblasts. Fibroblasts were treated with increasing doses of dhS1P for 24 hours (left) or were treated with 0.5 μM dhS1P for the indicated time periods (right). β-actin was used as a loading control. B, Western blot analysis of phospho-Smad3, total Smad3, PTEN, and β-actin in SSc fibroblasts that were treated with PTEN or nonsilencing small interfering RNA (siRNA) for 24 hours and stimulated with dhS1P for an additional 24 hours. C, Western blot analysis of PTEN and β-actin in healthy control fibroblasts that were treated with transforming growth factor β (TGFβ; 2.5 ng/ml) for the indicated time periods.

PTEN expression and PPM1A expression have not previously been evaluated in SSc fibroblasts. As shown in Figure 1A, SSc fibroblasts expressed low protein levels of PTEN and PPM1A as compared with control cells. Treatment with dhS1P normalized PTEN and PPM1A protein levels in SSc fibroblasts in a dose- and time-dependent manner, while dhS1P did not affect PTEN or PPM1A levels in control fibroblasts. Importantly, constitutive phosphorylation of Smad3 in SSc fibroblasts was inversely correlated with the increase in PTEN levels. To examine whether PTEN is responsible for the dhS1P-mediated inhibition of Smad3 phosphorylation, PTEN was depleted from SSc fibroblasts using siRNA as previously described (11). In the absence of PTEN, dephosphorylation of Smad3 by dhS1P was abrogated (Figure 1B), indicating that PTEN is required for this process. These findings represent the first demonstration that the phosphorylation status of Smad3 in SSc fibroblasts depends on endogenous PTEN levels.

We have previously shown that PPM1A protein is rapidly degraded in response to TGFβ (11). To determine whether TGFβ regulates PTEN expression, dermal fibroblasts were stimulated with TGFβ for 6–24 hours. As shown in Figure 1C, TGFβ reduced PTEN expression after 12 hours of treatment, suggesting that reduced levels of this protein in SSc fibroblasts may be due to the constitutive activation of TGFβ signaling.

We next examined 6 pairs of SSc and closely matched healthy control fibroblasts to determine the effects of dhS1P on PTEN, collagen, and MMP-1 production (Figure 2). (Results from each of the 6 pairs are available online at http://www.bumc.bu.edu/rheumatology/supplemental-data/.) Consistent with the findings of previous studies (2), all SSc cell strains produced more collagen and had reduced levels of MMP-1. In addition, all SSc cell strains demonstrated significantly reduced PTEN protein levels as compared with healthy fibroblasts. Treatment with dhS1P significantly increased the levels of PTEN in SSc fibroblasts, whereas only slight stimulatory effects were seen in control fibroblasts. Furthermore, dhS1P significantly increased MMP-1 levels and decreased collagen levels in SSc fibroblasts. In contrast, the effects of dhS1P on MMP-1 and collagen levels in control fibroblasts were not statistically significant. Taken together, these data suggest that dhS1P has a dual antifibrotic effect in SSc fibroblasts by decreasing collagen and increasing MMP-1 production.

Figure 2.

Increased sensitivity of SSc fibroblasts to the antifibrotic effects of dhS1P. Six pairs of SSc fibroblasts and closely matched control fibroblasts were stimulated with 0.5 μM dhS1P for 48 hours. Left, Western blot analysis of protein levels of PTEN, matrix metalloproteinase 1 (MMP-1), and collagen in a representative pair of normal and SSc fibroblasts. β-actin was used to normalize protein levels. Lanes 1 and 2 represent normal fibroblasts, and lanes 3 and 4 represent SSc fibroblasts. Right, Results from all pairs tested, expressed as a ratio of each protein level to β-actin level. Bars show the mean and SD. ∗ = P < 0.05; ∗∗ = P < 0.01. See Figure 1 for other definitions.

Fibroblasts from patients with SSc exhibit heightened sensitivity to the profibrotic effects of S1P.

S1P has been shown to mimic the profibrotic effects of TGFβ in several cell types, including foreskin fibroblasts (7, 10, 11). We next compared the effects of S1P on PTEN, MMP-1, and collagen production in 4 pairs of SSc and control fibroblasts. Although PTEN was already expressed at a relatively low level in SSc fibroblasts, treatment with S1P further significantly reduced PTEN protein expression (Figure 3). (Results from each pair are available online at http://www.bumc.bu.edu/rheumatology/supplemental-data/.) Likewise, S1P treatment further significantly reduced MMP-1 levels, whereas collagen levels were up-regulated in SSc fibroblasts. Similar trends were observed in control fibroblasts, but the response was more pronounced in SSc fibroblasts. Taken together, these data indicate that dhS1P and S1P have opposite effects on expression of several profibrotic genes in SSc fibroblasts.

Figure 3.

Increased sensitivity of SSc fibroblasts to the profibrotic effects of sphingosine 1-phosphate (S1P). Four pairs of SSc fibroblasts and closely matched control fibroblasts were stimulated with 0.5 μM S1P for 48 hours. Left, Western blot analysis of protein levels of PTEN, matrix metalloproteinase 1 (MMP-1), and collagen in a representative pair of normal and SSc fibroblasts. β-actin was used to normalize protein levels. Right, Results from all pairs tested, expressed as a ratio of each protein level to β-actin level. Bars show the mean and SD. ∗ = P < 0.05; # = P not significant. See Figure 1 for other definitions.

Reduction in PTEN expression in SSc skin biopsy specimens.

To further investigate the role of the PTEN pathway in dermal fibrosis in SSc, we examined the distribution of PTEN in skin specimens from 7 SSc patients and 7 healthy controls. (Representative results of staining in samples from the skin of SSc patients and healthy controls are available online at http://www.bumc.bu.edu/rheumatology/supplemental-data/.) PTEN-positive fibroblasts were counted in each specimen, and a summary of the results is included in Table 1. The analysis revealed heterogeneity of PTEN expression among SSc and control skin sections. While a majority of the SSc skin fibroblasts had either absent or low levels of PTEN expression, a similar pattern was also observed in some of the healthy skin biopsies. This finding is consistent with the results of other studies that showed low to moderate expression of PTEN in dermal fibroblasts in vivo (19). However, comparison of closely matched SSc and control skin specimens showed that, with the exception of one pair, a significantly higher proportion of PTEN-positive fibroblasts was present in healthy skin, suggesting that down-regulation of the PTEN pathway may contribute to the development of fibrosis in SSc.

Decrease in S1P receptor 1 and 2 expression in SSc dermal fibroblasts.

We have previously shown that effects of S1P and dhS1P on TGFβ-induced Smad3 phosphorylation are mediated via a single S1P receptor 1 in foreskin fibroblasts (11). We reasoned that increased sensitivity of SSc fibroblasts to dhS1P and S1P could be due to increased levels of S1P receptor 1. The distribution of S1P receptor subtypes in SSc and control fibroblasts was examined using quantitative PCR. Unexpectedly, SSc fibroblasts showed reduced expression of messenger RNA (mRNA) for S1P receptors 1 and 2; however, expression of mRNA for S1P receptor 3 was increased (Figure 4A). The expression of S1P receptor 1 protein was further investigated in SSc and control fibroblasts. Consistent with mRNA levels, S1P receptor 1 protein was expressed at the lower level in SSc fibroblasts (Figure 4B). We were unable to measure the protein levels of other S1P receptor isoforms, because of the lack of suitable antibodies. We next investigated whether TGFβ signaling regulates S1P receptor expression. As shown in Figure 4C, expression of all 3 S1P receptor isoforms was significantly down-regulated by TGFβ. Down-regulation of S1P receptor 1 was further confirmed at the protein level. These data suggest that the distribution of S1P receptor isoforms differs in SSc and healthy control fibroblasts.

Figure 4.

Distribution of sphingosine 1-phosphate (S1P) receptor isoforms differs in SSc and control fibroblasts. A, Quantitative polymerase chain reaction (PCR) analysis of expression of mRNA for S1P receptor 1 (S1P1), S1P receptor 2, and S1P receptor 3 in 5 pairs of SSc and normal fibroblasts. Bars show the mean and SD. B, Left, Western blot analysis of S1P receptor 1 protein level in a representative pair of normal and SSc fibroblasts. Note that only the lower band is specific. Right, Results from all 3 pairs tested. Bars show the mean and SD. C, Left, Quantitative PCR analysis of expression of mRNA for S1P receptors 1, 2, and 3 in 3 control fibroblast strains that were stimulated with 2.5 ng/ml of TGFβ. Bars show the mean and SD. Right, Representative Western blot of S1P receptor 1. D, Contribution of autocrine TGFβ signaling to down-regulation of S1P receptor 1 and PTEN in SSc fibroblasts. SSc fibroblasts were treated for 48 hours with 10 μg/ml of anti-TGFβ antibody (TGFβAb). Left, Representative Western blot of S1P receptor 1 and PTEN levels in SSc fibroblasts treated with the antibody as compared with levels in closely matched healthy control fibroblasts, which were used as a reference. Right, Results from 4 SSc cell strains. Bars show the mean and SD. ∗ = P < 0.05; ∗∗ = P < 0.01. See Figure 1 for other definitions.

To test the possibility that down-regulation of S1P receptor 1 and 2 isoforms in SSc fibroblasts may be mediated in part by TGFβ signaling, we blocked autocrine TGFβ signaling using TGFβ-neutralizing antibody. Addition of the TGFβ-neutralizing antibody completely abrogated TGFβ-induced phosphorylation of Smad3. (Results are available online at http://www.bumc.bu.edu/rheumatology/supplemental-data/.) Treatment of SSc fibroblasts with the neutralizing antibody resulted in up-regulation of S1P receptor 1, as well as of PTEN (Figure 4D), suggesting that the reduced levels of these genes in SSc fibroblasts may be mediated in part by autocrine TGFβ signaling. (Results from each SSc fibroblast strain are available online at http://www.bumc.bu.edu/rheumatology/supplemental-data/.)

The effects of dhS1P and S1P are mediated through S1P receptors 1 and 2 in SSc fibroblasts.

To investigate S1P/dhS1P signaling in SSc and healthy fibroblasts, we next focused on the function of the individual S1P receptor isoforms. To determine which receptor mediates the effects of S1P and dhS1P in control and SSc dermal fibroblasts, S1P receptors 1, 2, and 3 were individually depleted by >80% using specific siRNA (Figure 5A). Cells were then stimulated with 0.5 μM S1P or with a combination of TGFβ (2.5 ng/ml) and 0.5 μM dhS1P. In healthy adult dermal fibroblasts, S1P stimulated phosphorylation of Smad3, while dhS1P prevented TGFβ-induced Smad3 phosphorylation in the presence of nonsilencing siRNA (Figure 5A, left panel). Depletion of S1P receptor 1 inhibited the effects of S1P and dhS1P on Smad3 phosphorylation, while depletion of S1P receptor 2 or S1P receptor 3 did not have any appreciable effect on these responses. These results are consistent with our previous observations in foreskin fibroblasts.

Figure 5.

Depletion of distinct endogenous sphingosine 1-phosphate (S1P) receptor isoforms abrogates effects of dhS1P and S1P on Smad phosphorylation levels in normal and SSc fibroblasts. Cells were transfected with 30 nM S1P receptor 1 (S1P1), S1P receptor 2, or S1P receptor 3 siRNA or nonsilencing siRNA for 24 hours, and then serum starved overnight. Depletion of S1P receptor isoforms was assessed by quantitative polymerase chain reaction. A, Smad3 phosphorylation level in control cells that were treated with 1 μM S1P or 2.5 ng/ml of TGFβ plus 0.5 μM dhS1P for 30 minutes. B, Western blot analysis of phospho-Smad3 and total Smad3 in SSc fibroblasts that were left untreated (control [c]) or were treated with S1P (s) or dhS1P (d) for 30 minutes to assess Smad3 phosphorylation level. See Figure 1 for other definitions. Color figure can be viewed in the online issue, which is available at http://www.arthritisrheum.org.

We next examined the involvement of S1P receptors in response to S1P or dhS1P in SSc fibroblasts. S1P receptors 1, 2, and 3 were individually depleted using siRNA followed by stimulation with the agonists (Figure 5B). Interestingly, depletion of either S1P receptor 1 or S1P receptor 2 abrogated responses to dhS1P and S1P, while depletion of S1P receptor 3 had no effect. Taken together, these data suggest that normal fibroblasts mediate their responses to S1P and dhS1P through a single S1P receptor, S1P receptor 1, whereas SSc fibroblasts require 2 receptors, S1P receptor 1 and S1P receptor 2, for their responses.

DISCUSSION

Persistent TGFβ signaling is a major factor in the activation of lesional SSc fibroblasts (2). Cultured SSc fibroblasts maintain an “activated phenotype,” which is characterized by overexpression of collagen and other ECM proteins and reduced expression of the principal collagen-degrading enzyme, MMP-1. This study demonstrates that treatment of SSc fibroblasts with dhS1P effectively reverses this phenotype, including inhibition of phospho-Smad3, down-regulation of collagen, and up-regulation of MMP-1. Importantly, we show that the antifibrotic effects of dhS1P in SSc fibroblasts are mediated through the modulation of PTEN expression and that activation of the Smad3 pathway in SSc fibroblasts is directly linked to the reduced levels of PTEN. Furthermore, our data suggest that autocrine TGFβ signaling contributes to the down-regulation of PTEN in SSc fibroblasts. There was little effect of dhS1P on matrix-related genes in healthy dermal fibroblasts, consistent with its previously described role as an inhibitor of TGFβ/Smad3 signaling (11). S1P mimicked the effects of TGFβ by down-regulating PTEN and MMP-1 and up-regulating collagen protein levels. Interestingly, despite evidence of constitutive activation of the TGFβ signaling pathway, S1P effects were even more pronounced in SSc fibroblasts, suggesting that TGFβ and S1P may have an additive profibrotic effect.

There is increasing evidence that PTEN deficiency is associated with fibrosis in different organs. A previous study demonstrated that in patients with idiopathic pulmonary fibrosis, expression of PTEN was diminished in lung myofibroblasts within fibroblastic foci (20). It has also been shown that inhibition of PTEN function is necessary and sufficient for myofibroblast differentiation of lung fibroblasts. A similar role of PTEN in activation of cultured hepatic stellate cells was reported (21). The present study demonstrates a significantly lower level of PTEN in cultured SSc fibroblasts and a decreased presence of PTEN-positive fibroblasts in SSc skin in vivo. Restoration of PTEN levels in SSc fibroblasts correlated with normalization of collagen and MMP-1 expression.

The antifibrotic role of PTEN is not well understood. PTEN encodes a lipid phosphatase that dephosphorylates PtdIns(3,4,5)P3 (PIP3), leading to the inhibition of PI 3-kinase/Akt signaling. There is also evidence that PTEN, through a protein–protein interaction involving its C-terminal domain, has cellular functions that do not depend on its lipid phosphatase activity (22). Previous studies revealed a novel function of nuclear PTEN as a chaperone of the Smad3 phosphatase PPM1A. PPM1A is rapidly degraded in response to TGFβ signaling, and PTEN stabilizes PPM1A protein through formation of PTEN–PPM1A complexes (11). Accordingly, this study shows that normalization of PTEN, as well as PPM1A levels, in SSc fibroblasts leads to dephosphorylation of Smad3, suggesting that this may be one of the mechanisms whereby PTEN deficiency exerts fibrogenic effects.

It is also likely that PTEN deficiency contributes to fibrosis through activation of other fibrogenic pathways, such as Akt. In dermal fibroblasts, Akt induces collagen gene expression and inhibits MMP-1 production through a TGFβ-independent mechanism (23). Constitutive activation of the Akt pathway, which plays a central role in regulating cell growth and survival, has been demonstrated in SSc fibroblasts in vitro and in vivo; however, the pathway responsible for Akt activation in SSc was not examined (24). A recent study performed in glomerular mesangial cells has delineated the mechanism governing TGFβ activation of Akt (25). It was shown that TGFβ induces 2 microRNA, miR-216a and miR-217, which target PTEN. The decrease in PTEN increases PIP3 and leads to Akt activation. Further studies are needed to determine whether microRNA-dependent mechanisms are responsible for the down-regulation of PTEN and activation of Akt in SSc fibroblasts.

S1P and dhS1P signal through S1P receptors 1–5, which belong to the G protein–coupled receptor family (26). Different receptor subtypes couple to different G proteins, with S1P receptor 1 coupling exclusively to Gi and S1P receptors 2 and 3 coupling to Gi, Gq, and G12/13 (26). This is the first study to examine the distribution and function of S1P receptors in SSc fibroblasts. Our data show that reduced levels of S1P receptors 1 and 2 and elevated levels of S1P receptor 3 characterize SSc fibroblasts. Conversely, treatment of healthy fibroblasts with TGFβ resulted in the down-regulation of all 3 receptor isoforms; thus, altered distribution of S1P receptor isoforms in SSc fibroblasts could be only partially dependent on the activation of autocrine TGFβ signaling.

Interestingly, SSc fibroblasts differ from control cells in the utilization of S1P receptor isoforms. In SSc fibroblasts, S1P and dhS1P signal through S1P receptors 1 and 2, while in healthy fibroblasts these agonists mediate their effects via a single receptor, S1P receptor 1. The basis for this difference is not known. However, in other cell types, including cardiac fibroblasts, mesangial cells, and lung fibroblasts, profibrotic effects of S1P are mediated through S1P receptors 2 and 3 (7, 10, 27). S1P receptors 2 and 3, but not S1P receptor 1, couple to G12/13, the only G protein that activates the Rho pathway. Rho kinase has been shown to contribute to the TGFβ-induced myofibroblast differentiation in several experimental models, including SSc fibroblasts (28). Thus, it is possible that differential S1P/dhS1P signaling in SSc and healthy fibroblasts is related to the myofibroblast characteristics of SSc cells. While further studies are needed to fully understand the significance of these novel observations, this study points out a previously unappreciated role of sphingolipid signaling in SSc fibrosis.

In conclusion, our results suggest that the sphingosine kinase metabolites dhS1P and S1P may play an important role in the regulation of ECM in dermal fibroblasts through modulation of PTEN expression. The discovery that PTEN is directly involved in the regulation of Smad signaling, in addition to its well- known role as a lipid phosphatase, broadens the functional range of this tumor suppressor molecule and suggests that PTEN could also be called “fibrosis suppressor.” This study provides evidence that PTEN deficiency is present in SSc and suggests that dhS1P or its derivatives may be effective as a therapeutic antifibrotic agent. Both S1P and dhS1P are present in the circulation, with levels of S1P being an order of magnitude higher than those of dhS1P (Trojanowska M, Bielawska A: unpublished observations). Interestingly, it was recently reported that serum levels of S1P are increased in SSc, while there was no difference in dhS1P levels (29). Given the enhanced responsiveness of SSc fibroblasts to the profibrotic effects of S1P, these new data further underscore the potential contribution of S1P to SSc fibrosis and suggest that targeting the sphingolipid pathway may benefit patients with SSc.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Trojanowska had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Bu, Trojanowska.

Acquisition of data. Bu, Asano, Bujor, Highland, Hant.

Analysis and interpretation of data. Bu, Asano, Bujor, Trojanowska.

Ancillary