Lysophosphatidic acid signaling through its receptor initiates pro-fibrotic epithelial cell fibroblast communication mediated by epithelial cell derived connective tissue growth factor

Abstract

The expansion of the fibroblast pool is a critical step in organ fibrosis, but the mechanisms driving expansion remain to be fully clarified. We previously showed that lysophosphatidic acid (LPA) signaling through its receptor LPA₁ expressed on fibroblasts directly induces the recruitment of these cells. Here we tested whether LPA-LPA₁ signaling drives fibroblast proliferation and activation during the development of renal fibrosis. LPA₁-deficient (LPA_1−/−) or -sufficient (LPA_1+/+) mice were crossed to mice with green fluorescent protein expression (GFP) driven by the type I pro-collagen promoter (Col-GFP) to identify fibroblasts. Unilateral ureteral obstruction-induced increases in renal collagen were significantly, though not completely, attenuated in LPA_1−/−Col-GFP mice, as were the accumulations of both fibroblasts and myofibroblasts. Connective tissue growth factor was detected mainly in tubular epithelial cells, and its levels were suppressed in LPA_{1 −/−}Col-GFP mice. LPA-LPA₁ signaling directly induced connective tissue growth factor expression in primary proximal tubular epithelial cells, through a myocardin-related transcription factor-serum response factor pathway. Proximal tubular epithelial cell derived connective tissue growth factor mediated renal fibroblast proliferation and myofibroblast differentiation. Administration of an inhibitor of myocardin-related transcription factor/serum response factor suppressed obstruction-induced renal fibrosis. Thus, targeting LPA-LPA₁ signaling and/or myocardin-related transcription factor/serum response factor-induced transcription could be promising therapeutic strategies for renal fibrosis.

INTRODUCTION

Fibrosis is a common pathological feature of diseases that result in end-stage organ failure, including all forms of chronic kidney disease (CKD).¹ Renal fibrosis appears to result from abnormal wound-healing processes.^2,3 Wound-healing responses involve multiple cell types, including epithelial cells, endothelial cells, pericytes, leukocytes and fibroblasts. The interactions between these cell types have been increasingly appreciated to be central to the pathogenesis of fibrosis, ultimately resulting in the expansion of fibroblasts and their activation into myofibroblasts.^4,5 The molecular mediators of cell-cell communication in the development of fibrosis, however, remain to be fully elucidated.

We and others have implicated the bioactive lipid lysophosphatidic acid (LPA) in fibrosis of multiple organs, including the kidney.^6–10 LPA signals through specific G protein-coupled receptors (GPCRs), of which at least six have been identified and designated as LPA_1–6.¹¹ We have demonstrated that LPA signaling specifically through LPA₁ has pro-fibrotic effects on multiple cell types, promoting epithelial cell apoptosis, loss of endothelial cell barrier function, and fibroblast migration.^7,8 We have recently found that LPA contributes to fibrosis in a model of peritoneal fibrosis by inducing pro-fibrotic mesothelial cell to fibroblast communication through connective tissue growth factor (CTGF/CCN2).¹² We found that LPA induces fibroblast proliferation and activation in this model indirectly, by inducing mesothelial cell CTGF expression, which in turn drives fibroblast proliferation and myofibroblast accumulation in the peritoneum.

CTGF has been noted to stimulate fibroblast proliferation and production of extracellular matrix,^13,14 and its expression has been demonstrated to be increased in fibrotic disorders, including CKD.^15–17 Both the development of renal fibrosis and the renal CTGF expression have previously been found to require LPA-LPA₁ signaling in the UUO model.¹⁰ These investigators found that LPA induced CTGF expression in tubular epithelial cells by signaling through LPA₁.¹⁰ In contrast, other investigators have shown that LPA induces tubular CTGF expression through LPA₂, in a pathway involving activation of latent TGF-β in an αvβ6 integrin-dependent manner.¹⁸ These investigators found this integrin-TGF-β-CTGF pathway was required for renal fibrosis in an ischemia-reperfusion injury model.¹⁸ A third group of investigators had previously implicated LPA₃ in renal ischemia-reperfusion injury.¹⁹

Although several signaling pathways can produce CTGF expression, including Smad,²⁰ we previously found that LPA signaling through LPA₁ stimulates CTGF expression in mesothelial cells through a myocardin-related transcription factor (MRTF)-serum response factor (SRF) pathway.¹² SRF is a MADS box transcription factor that induces CTGF by binding to a CArG box sequence [CC(A/T)₆GG] in its promoter region.^21–23 MRTF-A and MRTF-B are transcriptional co-activators that are sequestered in the cytoplasm by binding to globular actin (G-actin) monomers. Polymerization of G-actin into filamentous actin (F-actin) liberates MRTFs from G-actin, allowing their translocation to the nucleus, where they augment SRF transcriptional activity.^24–26 We and others have demonstrated that targeting of the MRTF-SRF pathway protects mice from fibrosis in several organs, including the 12, 27–30 kidney.

In this study, we consequently investigated the hypothesis that LPA signaling specifically through LPA₁ contributes to renal fibrosis by inducing tubular CTGF through the MRTF-SRF pathway. We further hypothesized that this induction of CTGF mediates pro-fibrotic epithelial cell-to-fibroblast communication that drives fibroblast proliferation and myofibroblast differentiation. Here we report compelling evidence for these hypotheses in UUO-induced renal fibrosis model.³¹ Clarifying the mediators and signaling pathways responsible for the epithelial cell-to-fibroblast communication that results in the expansion and activation of the fibroblast pool is important not only for understanding renal fibrogenesis, but for the rational development of new therapeutic strategies to treat renal fibrosis.

RESULTS

Renal LPA₁ expression increased in the course of UUO-induced renal fibrosis

To determine the profile of renal LPA receptor subtypes in the course of renal fibrosis, we examined LPA receptors (LPA₁-₆) mRNA expression in whole kidneys (Supplementary Figure S1). In control kidneys obtained from LPA₁-sufficient (LPA₁^+/+) mice, LPA₂ and LPA₆ were most highly expressed, followed by LPA₁ and LPA₃. Control kidneys in LPA₁-deficient (LPA₁^−/−) mice did not express LPA₁ as expected, and their LPA₁ deficiency did not cause compensatory changes in their expression of other LPA receptors. In LPA₁^+/+ ligated kidneys, LPA₁ was the most highly induced LPA receptor; LPA₅ and LPA₆ expression was significantly increased as well, where LPA₃ expression was significantly decreased.

Genetic deletion of LPA₁ attenuated UUO-induced renal fibrosis

To examine the impact of LPA₁ on fibroblast accumulation and renal fibrosis, we crossed LPA^−/− or LPA₁^+/+ mice to mice with green fluorescent protein expression driven by the type I pro-collagen promoter (Col-GFP). LPA₁^−/−Col-GFP mice were significantly, though not completely protected from the development of renal fibrosis. Masson’s trichrome staining of ligated kidneys demonstrated that the renal fibrosis induced in LPA₁^+/+Col-GFP mice was reduced in LPA₁^−/−Col-GFP mice (Figure 1a). Hydroxyproline levels in non-ligated kidneys were not different between both groups; in ligated kidneys, hydroxyproline levels were significantly reduced in LPA₁-deficient mice (Figure 1b). UUO-induced increases in type I procollagen α1 chain (COLIα1) mRNA expression were similarly reduced in LPA₁^−/−Col-GFP mice (Figure 1c). These data suggest that LPA-LPA₁ signaling is involved in the development of renal fibrosis.

Figure 1. Protection of LPA₁-deficient mice from UUO-induced renal fibrosis.

(a) Masson’s trichrome-stained renal sections of LPA₁^+/+Col-GFP mice (upper lane) and LPA₁^−/−Col-GFP mice (lower lane). Representative tissue sections are shown (magnification × 200). Bars, 100 µm. (b) Biochemical analysis of UUO-induced renal fibrosis. Hydroxyproline content was measured in the kidneys of LPA₁^+/+Col-GFP mice and LPA₁^−/−Col-GFP mice ten days after UUO (n = 6 mice/group). Data are expressed as mean ± SEM. (c) Renal expression of COLIα1 mRNA in LPA₁^+/+Col-GFP mice and LPA₁^−/−Col-GFP mice ten days after UUO (n = 6 mice/group). Data are expressed as mean copies of COLIα1 mRNA relative to copies of GAPDH mRNA ± SEM.

UUO-induced fibroblast accumulation required LPA-LPA₁ signaling

Recent genetic fate mapping studies have pointed to the resident perivascular fibroblast/pericyte as the major source of myofibroblasts in renal fibrosis, indicating that expansion of the resident collagen-producing fibroblast pool is a critical step in renal fibrogenesis.^{1, 32} We consequently investigated whether expansion of the fibroblast pool requires LPA-LPA₁ signaling. As demonstrated in the representative sections, UUO induced a marked accumulation of GFP⁺ fibroblasts in LPA₁^+/+Col-GFP mice, which was significantly attenuated in LPA₁^−/−Col-GFP mice (Figure 2a and b). To determine the impact of LPA₁ signaling on fibroblast proliferation, we co-stained renal sections with anti-proliferating cell nuclear antigen (PCNA) antibody and anti-GFP antibody as previously described.³³ As demonstrated in the representative sections, the numbers of proliferating fibroblasts (GFP⁺PCNA⁺ cells) were lower in LPA₁^−/−Col-GFP mice as compared to that in LPA₁^+/+Col-GFP mice (Figure 2a and c). The percentage of proliferating fibroblasts among total fibroblasts (the percentage of GFP⁺PCNA⁺ cells among total GFP⁺ cells) was significantly reduced by the loss of LPA-LPA₁ signaling (Figure 2d). These data suggest that LPA-LPA₁ signaling is involved in UUO-induced expansion of the collagen-producing fibroblast pool through the regulation of fibroblast proliferation.

Figure 2. UUO-induced renal fibroblast accumulation requires LPA₁.

(a) Accumulation of proliferating fibroblasts (GFP⁺PCNA⁺) ten days after UUO. Representative tissue sections stained with anti-GFP antibody/anti-PCNA antibody are shown. Bars, 100 µm. (b) Numbers of GFP⁺ cells in the kidney are expressed as mean ± SEM per HPF (n = 5 mice/group). (c) Numbers of renal GFP⁺PCNA⁺ cells (proliferating fibroblasts) are expressed as mean ± SEM per HPF. (d) Percentages of renal fibroblasts that are proliferating (GFP⁺PCNA⁺ cells/total GFP⁺ cells) are expressed as mean ± SEM per HPF (n = 5 mice/group).

Genetic deletion of LPA₁ attenuated UUO-induced myofibroblast accumulation

Myofibroblasts predominate in areas of increased collagen deposition in fibrotic diseases, consistent with their identification as the effector cells that produce the pathological extracellular matrix in fibrotic tissues.⁵ We therefore assessed the impact of LPA-LPA₁ on the accumulation of myofibroblasts. As demonstrated in the representative sections in Figure 3a, a smooth muscle actin (αSMA)⁺ myofibroblasts accumulated predominantly in the interstitium by day 10 following UUO. UUO increased the area of the sections that had αSMA positive staining, and the increases in these areas of positive staining in LPA₁^+/+Col-GFP mice were attenuated in LPA₁^−/−Col-GFP mice (Figure 3a and b). UUO-induced increases in renal expression of αSMA mRNA and protein were also reduced in LPA₁^−/−Col-GFP mice (Figure 3c and d). In addition, we determined the percentages collagen-producing fibroblasts that co-expressed αSMA, to investigate the effect of LPA₁ signaling on myofibroblast differentiation. In LPA₁^+/+Col-GFP mice, 91.5 ± 3.0 % of GFP⁺ cells were positive for αSMA, while the percentage of GFP⁺αSMA⁺ cells among total GFP⁺ cells in LPA₁^−/−Col-GFP mice was significantly reduced to 70.8 ± 6.0% (Figure 3e and f). These findings suggest that LPA-LPA₁ signaling contributes to αSMA⁺ myofibroblast accumulation during renal fibrogenesis.

Figure 3. UUO-induced renal αSMA⁺ myofibroblast accumulation requires LPA₁.

(a) Accumulation of αSMA⁺ myofibroblasts ten days after UUO. αSMA-stained renal sections of LPA₁^+/+Col-GFP (upper panel) and LPA₁^−/−Col-GFP (lower panel). Representative tissue sections are shown (magnification × 200). Bars, 100 µm. (b) αSMA⁺ areas in the kidney are expressed as mean SEM per HPF (n = 6 mice/group). (c) Renal expression of αSMA mRNA in LPA₁^+/+Col-GFP mice and LPA₁^−/−Col-GFP mice (n = 6 mice/group). Data are expressed as mean copies of αSMA mRNA relative to copies of GAPDH mRNA ± SEM. (d) The expression of αSMA protein in kidney (n = 4 mice/group). Quantification was performed with Image J software and data are expressed as mean dots of αSMA bands relative to dots of GAPDH bands ± SEM. (e) Accumulation of αSMA-expressing fibroblasts (GFP⁺αSMA⁺) ten days after UUO. Representative tissue sections stained with anti-GFP antibody/anti-αSMA antibody are shown. Bars, 100 µm. (f) Percentages of renal fibroblasts that are expression αSMA (GFP⁺αSMA⁺ cells/total GFP⁺ cells) are expressed as mean ± SEM per HPF (n = 5 mice/group).

UUO-induced renal CTGF expression required LPA₁, and was predominantly attributable to tubular epithelial cells

We examined whether LPA-LPA₁ signaling regulates renal CTGF expression in vivo. CTGF was induced in ligated kidneys of LPA₁^+/+Col-GFP mice, but this increase in renal CTGF was significantly attenuated in LPA₁^−/−Col-GFP mice (Figure 4a and b). To identify which cells were responsible for the increase in CTGF, we performed CTGF immunostaining of renal tissue sections. Intense staining was induced by UUO in the kidneys of LPA₁^+/+Col-GFP mice, mainly in tubular epithelial cells, and to a lesser extent in interstitial cells (Figure 4c). The intensity of CTGF staining induced by UUO was significantly reduced in LPA₁^−/−Col-GFP mice. These findings suggest that LPA-LPA₁ signaling contributes to the induction of tubular CTGF expression in the development of renal fibrosis.

Figure 4. UUO-induced CTGF expression requires LPA₁, and is predominantly attributable to tubular epithelial cells.

(a) Renal expression of CTGF mRNA following UUO for ten days (n = 6 mice/group). Data are expressed as mean copies of CTGF mRNA relative to copies of GAPDH mRNA ± SEM. (b) The production of CTGF protein in kidney ten days after UUO (n = 6 mice/group). Quantification was performed with Image J software and data are expressed mean dots of CTGF bands relative to dots of GAPDH bands ± SEM. (c) The localization of CTGF protein in the kidney. Representative tissue sections are shown (magnification × 200). Bars, 100 µm.

LPA induction of tubular epithelial cell CTGF was partially mediated by LPA₁, and induced fibroblast proliferation and αSMA expression

Tubular epithelial cells have been thought to be a critically important source of pro-fibrotic mediators.³⁴ To investigate the ability of LPA-LPA₁ signaling to induce CTGF in tubular epithelial cells, we used mouse primary proximal tubular epithelial cells (PTECs) for in vitro studies. LPA induced CTGF mRNA expression in PTECs in a time- and dose-dependent manner (Figure 5a and b). To investigate which of LPA’s receptors mediate CTGF expression by PTECs, we determined the profile of LPA receptor expression by these cells. We found detectable levels of mRNA for each receptor investigated (LPA_1–6), with LPA₂ being the most highly expressed in these cells followed by LPA₁ (Fig. 5C). To determine the functional requirement for individual LPA receptors for the induction of CTGF, PTECs were transfected with either LPA₁ or LPA₂ siRNA (Figure 5d). We did not see any compensatory changes in the expression of other LPA receptors induced by siRNA treatment (data not shown). The induction of CTGF mRNA expression stimulated by LPA was significantly suppressed by the treatment with LPA₁ siRNA (Figure 5e), indicating that LPA signaling through LPA₁ plays an important role to induce CTGF in PTECs. Treatment with LPA₂ siRNA also significantly inhibited the expression of LPA-induced CTGF in PTECs, indicating that both LPA₁ and LPA₂ contribute to this activity of LPA (Figure 5e).

Figure 5. LPA-LPA₁-induced tubular epithelial CTGF drives fibroblast proliferation and αSMA expression.

(a, b) LPA induces CTGF mRNA expression in PTECs in a time- and dose-dependent manner (n = 3 cell preparations/group). (c) LPA receptor expression of PTECs. (d) Validation of the inhibitory effects of LPA₁ siRNA and LPA₂ siRNA on the expression of LPA₁ and LPA₂ in PTECS (n = 3 cell preparations/group). (e) Expression levels of LPA-induced CTGF were decreased by knockdown of LPA₁ and LPA₂ by siRNA in PTECs (n = 3 cell preparations/group). (f) Identification of CTGF protein in conditioned media (CM) from PTECs by Western blot. (g, h) Mouse primary renal fibroblasts were transfected with CTGF siRNA, to prevent them from making additional CTGF in response to LPA still present in the CM, and then incubated with CM obtained from PTECs for 48 hours. Fibroblast proliferation and αSMA expression levels were examined (n = 3 cell preparations/group). Data from BrdU proliferation assays are expressed as mean ± SEM of OD value (OD₃₇₀-OD₄₉₂). All data of mRNA expression are expressed as mean ± SEM.

Next, to elucidate the pro-fibrotic functions of CTGF derived from PTECs, we examined the ability of media conditioned by LPA-stimulated PTECs to induce the proliferation of fibroblasts, and their expression of αSMA. Conditioned media (CM) of LPA-stimulated PTECs contained CTGF protein that was not detectable in CM of unstimulated cells (Figure 5f). CM from LPA-stimulated PTECs also induced significantly greater fibroblast proliferation (Figure 5g) and αSMA expression (Figure 5h) than CM of unstimulated cells. CTGF protein was also not detectable in CM of LPA-stimulated PTECs transfected with siRNA targeting CTGF (Figure 5f). CM of these cells has significantly reduced effects on fibroblast proliferation and αSMA expression (Figure 5g and h) compared to the CM of LPA-stimulated PTECs transfected with control siRNA. To prevent any LPA remaining in the CM from directly stimulating CTGF expression by fibroblasts, the responding fibroblasts were also transfected with siRNA targeting CTGF. Taken together, these data support the concept that LPA-induced CTGF in PTECs promotes fibroblast proliferation and myofibroblast differentiation.

Recent studies have shown that induction of epithelial cell cycle arrest in G2/M induces the production of profibrotic cytokines, including CTGF, in these cells that can drive the progression of renal fibrosis.³⁵ We therefore investigated whether PTEC-derived CTGF, in addition to its effects on fibroblasts, could also induce PTEC G2/M arrest that would further upregulate their CTGF expression in an autocrine/paracrine manner. As shown in Supplementary Figure S2, CM of LPA-stimulated PTECs transfected with control siRNA did not induce G2/M arrest in PTECs, suggesting that PTEC-derived CTGF does not also drive progression of renal fibrosis through the induction of PTEC cell cycle arrest.

LPA-induced CTGF expression was dependent on Gα_12/13, Rho, ROCK and actin polymerization in tubular epithelial cells

We next investigated the mechanisms through which LPA induces CTGF production by PTECs. LPA receptors including LPA₁ are GPCRs, which couple to different classes of G proteins (Gα_12/13, Gα_i/o, Gα_s or Gα_q), to mediate the diverse activities of LPA.³⁶ To investigate the contribution of Gα_12/13-containing G proteins to LPA-induced CTGF expression, we transfected PTECs with siRNAs targeting Gα₁₂ and/or Gα₁₃. Following validation of siRNA-induced knockdown of Gα_12/13 expression (Figure 6a), we found that reducing expression of Gα_12/13 G proteins in PTECs significantly suppressed LPA-induced CTGF production (Figure 6b). In contrast, pre-treatment of PTECs with pertussis toxin, which is an inhibitor of Gα_i/o G proteins, did not suppress LPA-induced CTGF expression (Figure 6c). These results suggest that Gα_12/13 G proteins mediate LPA-induced CTGF expression by PTECs.

Figure 6. LPA-induced PTEC CTGF expression is dependent on Gα_12/13, Rho, ROCK and actin polymerization.

(a) Validation of the inhibitory effects of Gα₁₂ siRNA and Gα₁₃ siRNA on the expression of Gα₁₂ and Gα₁₃ in PTECS (n = 3 cell preparations/group). (b) Expression levels of LPA-induced CTGF were decreased by knockdown of Gα₁₂ and Gα₁₃ by siRNA in PTECs (n = 3 cell preparations/group). (c) Pertussis toxin (PTX)-independent induction of CTGF in response to LPA. PTECs were preincubated with vehicle or 100 ng/ml PTX for 18h followed by the stimulation with 10 µM LPA for 2h (n = 3 cell preparations/group). (d, e) PTECs were preincubated with vehicle (Co), 2.0 µg/ml C3 toxin (C3) for 10h, 5µM Y27632 (Y) for 30 min or 1µg/ml latrunculin B (LB) for 30 min. Cells were then stimulated with 10 µM LPA for an additional 2h (n = 3 cell preparations/group). All data are expressed as mean ± SEM. (f) Immunocytochemical staining for phalloidin in PTECs. PTECs were incubated in serum-free media for 24 hours followed by the stimulation with LPA (10 µM). Some cells were pre-incubated with Y27632 (Y, 5 µM) for 30 min prior to LPA stimulation. All images were captured using identical exposure settings. Scale bars, 25 µm.

Downstream of G proteins in GPCR signaling pathways are members of the Rho family of small GTPases, which can be subdivided into the Rac, Cdc42 and Rho subfamilies.³⁷ Gα_i/o G proteins typically mediate LPA-induced Rac activation, whereas Gα_12/13 G proteins are responsible for LPA-induced Rho activation, which induces activation of Rho-associated coiled-coil-forming kinases (ROCKs).³⁶ One of the major effects of ROCK activation is to shift the equilibrium between actin polymerization and depolarization in favor of filament assembly.³⁸ We recently reported that LPA-induced CTGF expression in mesothelial cells is driven by Rho-ROCK activation via Gα_12/13 leading to actin polymerization.¹² We consequently hypothesized that LPA-induced CTGF expression in PTECs would similarly require Rho-ROCK-induced actin polymerization. As shown in Figure 6d, PTEC treatment with the Rho inhibitor C3 toxin reduced LPA-mediated induction of CTGF mRNA. In addition, inhibition of ROCK or actin polymerization by Y27632 or latrunculin B respectively suppressed LPA-induced CTGF expression (Figure 6e). Inhibition of LPA-induced actin polymerization by Y27632 was confirmed by immunocytochemical staining of PTECs with phalloidin (Figure 6f). Taken together, these data indicate that LPA-induced CTGF expression in PTECs is mediated by a signaling pathway involving Gα_12/13, Rho, ROCK and actin polymerization.

LPA-induced ROCK activation promoted MRTF-A and -B nuclear translocation and SRF transcriptional activity

Next, we hypothesized that the nuclear translocation of MRTF-A and -B is driven by ROCK-induced actin polymerization and can link LPA-LPA₁ signaling to SRF transcriptional activity. Immunocytochemical staining of MRTF-A and -B in PTECs revealed that LPA significantly induced the nuclear translocation of both of these transcriptional co-activators. Pre-treatment of PTECs with Y-27632 markedly blocked LPA-induced MRTF nuclear translocation (Figure 7a and b). To investigate the impact of LPA-induced MRTF nuclear translocation on SRF transcriptional activity, we transfected PTECs with a transcriptional reporter containing luciferase under the control of a CArG box sequence. As shown in Figure 8a, LPA enhanced SRF transcriptional activity in a dose-dependent manner. Pre-treatment of PTECs with Y27632 significantly reduced the SRF transcriptional activity by LPA (Figure 8b). Taken together, these data indicate that LPA induces MRTF-A and -B nuclear translocation and SRF transcriptional activity, in a ROCK dependent manner.

Figure 7. LPA promotes the nuclear translocation of MRTF-A and MRTF-B in a ROCK-dependent manner.

(a) The subcellular distribution of MRTF-A (upper panel) and MRTF-B (lower panel) in PTECs. PTECs were incubated in serum-free media for 24 hours and then stimulated with LPA (10 µM for 30 min) or control media. PTECs were additionally pre-treated with Y27632 (Y, 5 µM) or control media for 30 min before LPA stimulation. All images were captured using identical exposure settings. (b) Quantification of the subcellular distribution of MRTF-A (upper panel) and MRTF-B (lower panel). Five random fields of view were counted per slide. Subcellular distributions were classified as nuclear (N, nuclear staining > cytoplasmic staining); equal (E, nuclear staining = cytoplasmic staining); or cytoplasmic (C, nuclear staining < cytoplasmic staining). Two independent series of PTECs were analyzed. All data are expressed as mean distribution ± SEM. Scale bars, 25 µm.

Figure 8. LPA enhances MRTF-SRF transcriptional activity through ROCK.

(a) LPA induced MRTF-SRF transcriptional activity in PTECs transfected with a transcriptional reporter containing luciferase under the control of a CArG box sequence in a dose-dependent manner (n = 3 cell preparations/group). (b) Pre-treatment with Y27632 (Y, 5 µM for 30 min) reduced the MRTF-SRF transcriptional activity in response to 10 µM LPA (n = 3 cell preparations/group). All data are expressed as mean ± SEM.

Inhibition of the MRTF-SRF pathway inhibited LPA-induced PTEC CTGF expression, and attenuated renal fibrosis

To determine whether the MRTF-SRF pathway contributed to LPA-induced CTGF expression, we transfected PTECs with siRNAs targeting MRTF-A and/or MRTF-B. Following validation of siRNA-induced knockdown of these genes (Figure 9a), we found that reducing expression of both MRTF-A and -B significantly suppressed LPA-induced CTGF production (Figure 9b). We then transfected PTECs with siRNAs targeting SRF, and we found that reducing SRF expression in these cells significantly suppressed LPA-induced CTGF production as well (Figure 9c and d). To confirm the involvement of the MRTF-SRF pathway in LPA-induced CTGF expression, we pre-treated PTECs with CCG-1423, an inhibitor of MRTF-SRF pathway.³⁹ Pre-treatment of PTECs with CCG-1423 abrogated LPA-induced CTGF expression (Figure 9e). These results suggest that MRTF-SRF pathway is significantly involved in LPA-induced CTGF expression in PTECs. We also investigated the role of the MRTF-SRF pathway in LPA-induced fibroblast CTGF expression. Renal fibroblasts were transfected with siRNAs targeting MRTF-A, MRTF-B or SRF; knockdown of either MRTF-A and -B together, or SRF alone, significantly reduced LPA-induced CTGF expression in these cells as well (Supplementary Figure S3).

Figure 9. LPA-induced PTEC CTGF expression is dependent on MRTF-A, MRTF-B and SRF.

(a) Validation of the inhibitory effects of MRTF-A siRNA and MRTF-B siRNA on the expression of MRTF-A and MRTF-B (n = 3 cell preparations/group). (b) Expression levels of LPA-induced CTGF were decreased after knockdown of MRTF-A and MRTF-B by siRNA in PTECs (n = 3 cell preparations/group). (c) Validation of the inhibitory effects of SRF siRNA on the expression of SRF (n = 3 cell preparations/group). (d) Expression levels of LPA-induced CTGF were decreased after knockdown of SRF by siRNA in PTECs (n = 3 cell preparations/group). (e) PTECs were preincubated with 10µM CCG1423 (CCG) for 16h. Cells were then stimulated with 10 µM LPA for an additional 2h (n = 3 cell preparations/group). All data are expressed as mean ± SEM.

To confirm an important role for the MRTF-SRF pathway in the development of renal fibrosis in vivo, we administered another MRTF-SRF pathway inhibitor, CCG-203971,⁴⁰ in the UUO-induced renal fibrosis model. Mice treated with CCG-203971 were protected from renal fibrosis, as indicated by Masson’s trichrome staining and by measurements of hydroxyproline content and COLIα1 mRNA levels (Figure 10a–c). Additionally, CCG-203971 treatment significantly attenuated UUO-induced increases in renal expression of CTGF and αSMA (Figure 10d and e), all suggesting that targeting the MRTF-SRF pathway has the potential to suppress renal fibrosis. To confirm targeting of this pathway by CCG-203971 in vivo, we also investigated the effects of this inhibitor on the expression of known gene targets of this pathway, including zyxin, vinculin and PDZ-LIM domain 7 (PDLIM7). mRNA expression of each of these genes was significantly reduced in CCG-203971-treated mice (Figure 10f–h), suggesting that CCG-203971 treatment inhibited the MRTF-SRF pathway in vivo.

Figure 10. Pharmacological inhibition of the MRTF-SRF pathway attenuates UUO-induced renal fibrosis.

(a) Representative Masson’s trichrome -stained sections of ligated kidneys. CCG; CCG-203971 (magnification × 200). Bars, 100 µm. (b) Hydroxyproline content in the kidney following UUO for 7 days (n = 5 mice/group). Data are expressed as mean ± SEM. (c–h) Renal expression of COLIα1, CTGF, αSMA, zyxin, vinculin and PDLIM7 following UUO for seven days (n = 5 mice/group). Data are expressed as mean copies of target gene mRNAs relative to copies of GAPDH mRNA ± SEM.

Finally, we examined if LPA-LPA₁ signaling regulates the ROCK-MRTF-SRF pathway in vivo. As shown in Supplementary Figure S4, ROCK activity was increased in the ligated kidneys of LPA₁^+/+ mice, and this UUO-induced increased in ROCK activity was significantly reduced in LPA₁^−/− ligated kidneys. In addition, UUO induced the expression of zyxin, vinculin and PDLIM7 in the ligated kidneys of LPA₁^+/+ mice, and the increased expression of each of these MTRF-SRF gene targets was significantly reduced in LPA₁^−/− ligated kidneys. Taken together, these data support our hypothesis that LPA-LPA₁ signaling contributes to the development of UUO-induced renal fibrosis through the regulation of a ROCK-MRTF-SRF signaling pathway.

DISCUSSION

In this study, we found that LPA signaling specifically through LPA₁ is required for the development of renal fibrosis. Genetic deletion of LPA₁ protected mice from renal fibrosis by UUO. Disrupting the LPA-LPA₁ pathway significantly reduced the accumulation of fibroblasts and myofibroblasts. Rather than acting directly on fibroblasts, as LPA does in mediating fibroblast migration and resistance to apoptosis, we provide evidence that LPA induced fibroblast proliferation and myofibroblast differentiation during the development of renal fibrosis indirectly, by inducing pro-fibrotic epithelial cell-to-fibroblast communication mediated by CTGF. Renal CTGF expression was dependent on LPA₁, and was predominantly attributable to tubular epithelial cells. LPA-LPA₁ signaling induced CTGF expression by PTECs, and this CTGF accounted for much of the fibroblast-activating activity. Of the several signaling pathways to induce CTGF expression, we demonstrated that LPA-induced epithelial CTGF expression was predominantly mediated by MRTF-SRF pathway. Pharmacological inhibition of MRTF-SRF-induced transcription reduced the extent of UUO-induced renal fibrosis, as well as renal expression of MRTF-SRF-dependent genes. Finally, we confirmed that LPA-LPA₁ signaling regulates MRTF-SRF pathway in renal fibrosis model. Taken together, we conclude that LPA-LPA₁ signaling contributes to the development of renal fibrosis by inducing pro-fibrotic tubular epithelial cell-to-fibroblast communication, in that LPA-LPA₁ signaling stimulates tubular epithelial cells to produce CTGF dependent on MRTF-SRF pathway, and this epithelial CTGF in turn induces fibroblast proliferation and myofibroblast differentiation (Fig. 11).

Figure 11. Proposed pro-fibrotic interaction between tubular epithelial cells and fibroblasts in the pathogenesis of renal fibrosis.

LPA-LPA₁ signaling contributes to the development of renal fibrosis through interactions between tubular epithelial cells and fibroblasts, in which LPA-LPA₁ signaling stimulates tubular epithelial cells to produce CTGF through a Gα_12/13-Rho-ROCK-MRTF-SRF pathway, and this tubular epithelial cell-derived CTGF in turn induces fibroblast proliferation and αSMA expression.

Fibroblast migration into the wound matrix and the proliferation of these cells are thought to play a critical role in fibroblast accumulation.^41,42 As noted, LPA-LPA₁ signaling directly regulates several important fibroblast activities, including their migration and persistence.^7,8 In addition to these direct effects, here we have shown that LPA-LPA₁ signaling indirectly induces the proliferation of fibroblasts, as well as the differentiation of these cells into myofibroblasts, by directly inducing epithelial CTGF expression. CTGF has been demonstrated to mediate both of these pro-fibrotic fibroblast activities, but through different domains of the CTGF protein: the C-terminal domain of CTGF mediates fibroblast proliferation, whereas the N-terminal domain of CTGF mediates myofibroblast differentiation.⁴³ CTGF’s activity on fibroblasts has been shown to be mediated by its binding to mannose 6-phosphate/insulin-like growth factor 2 receptors,⁴⁴ suggesting that this receptor may be an additional target for anti-fibrotic therapies.

The pro-fibrotic interactions between epithelial cells and fibroblasts during the development of renal fibrosis that we have demonstrated in this study may be particularly relevant to the development of fibrosis in other organs in which these two cell types normally reside in close proximity, such as the lung and liver. The development of fibrosis in these organs is thought to fundamentally involve epithelial cell-fibroblast communication. Non-resolving epithelial injury in these organs is thought to provoke fibroblast activation through epithelial cell-derived soluble molecules,^4,12,32,45 such as CTGF, as we have focused on in this study. In turn, activated fibroblasts also appear to be able to amplify epithelial injury through fibroblast-derived soluble molecules, such as a reactive oxygen species, causing a vicious cycle of epithelial cell-fibroblast interactions that result in the development of organ fibrosis.⁴ Of note, we have shown that LPA-LPA₁ signaling is involved in both fibroblast activation and epithelial apoptosis,^7,8,12 suggesting that LPA-LPA₁ signaling could be a common pathway contributing fibrosis of epithelial organs by mediating bi-directional pro-fibrotic interactions between epithelial cells and fibroblasts. Of note, since this study used mice that were deficient for LPA₁ in all cell types, effects of LPA-LPA₁ signaling of other cell types in addition to epithelial cells and fibroblasts, such as endothelial cells and innate immune cells, may also contribute to the pathogenesis of renal fibrosis.

Our elucidation of the molecular pathway downstream of LPA-LPA₁ signaling that induces CTGF expression suggests new targets for renal fibrosis therapies. CTGF expression is principally regulated at the levels of transcription, and its promoter contains a number of binding-sites for transcription factors such as Smads.¹⁵ We recently demonstrated that LPA-induced CTGF expression by mesothelial cells is dependent on LPA₁-MRTF-SRF pathway, and that pharmacological inhibition of the MRTF-SRF pathway attenuates peritoneal fibrosis.¹² We have showed here that this same molecular connection links LPA-LPA₁ signaling to CTGF expression in tubular epithelial cells, and that pharmacological inhibition of MRTF-SRF pathway ameliorates renal fibrosis as well as renal CTGF expression. MRTF-A has also recently been shown to induce epigenetic histone modifications leading to type I collagen induction in a diabetic nephropathy model.³⁰ Taken together, these data suggest that the MRTF-SRF pathway may contribute to renal fibrosis in multiple ways, in multiple contexts, and have substantial potential to be the targets of effective therapies for renal diseases.

LPA has been demonstrated to be produced in multiple tissues,^7,10,46 but the source(s) of LPA during the development of organ fibrosis remain to be established. In the UUO model, the concentrations of LPA in the effluent from the pelvis of the ligated kidney have been demonstrated to be significantly higher than those in the urinary bladder, suggesting that LPA is produced locally in the kidney during the course of renal fibrosis.⁴⁶ LPA can be produced by two major pathways. Of these, the pathway involving the enzyme autotaxin (ATX), which converts lysophospholipids such as lysophosphatidylcholine to LPA by its lysophospholipase D activity, is responsible for the most of LPA present in the circulation.⁴⁷ Whether, and how, ATX activity could be increased locally in the kidney as fibrosis develops remains to be investigated. If the pathways responsible for producing LPA during renal fibrogenesis could be identified, their inhibition could potentially complement LPA₁ antagonism in preventing or treating renal fibrosis.

In summary, we have shown that LPA-LPA₁ signaling importantly contributes to the pathogenesis of renal fibrosis, by inducing pro-fibrotic tubular epithelial cell-to-fibroblast communication mediated via MRTF-SRF-dependent CTGF expression by epithelial cells. These results suggest that LPA-LPA₁ signaling may be a common pathway in fibrotic diseases affecting most epithelial organs. In addition, the MRTF-SRF pathway may mediate an important component of these interactions, linking epithelial injury to fibroblast activation, and inhibition of this pathway may have the potential to be an effective therapeutic strategy for fibrotic diseases of the kidney and other epithelial organs.

MATERIALS AND METHODS

Materials and Methods are described in detail in Supplementary Materials and Methods.^48–50