Inhibition of Renal Stellate Cell Activation Reduces Renal Fibrosis

Interstitial fibrosis is a common feature of chronic kidney disease, and platelet-derived growth factor receptor-β (PDGFR-β)-positive mesenchymal cells are reportedly the major source of scar-producing myofibroblasts. We had previously demonstrated that albumin and its derivative R-III (a retinol-binding protein-albumin domain III fusion protein) inhibited the transdifferentiation/activation of hepatic stellate cells (HSCs) to myofibroblasts and that R-III administration reduced liver fibrosis. In this study, we isolated cells (referred to as renal stellate cells, RSCs) from rat kidney tissues using the HSC isolation protocol and compared their morphological and biochemical characteristics with those of HSCs. RSCs shared many characteristics with HSCs, such as storage of vitamin A-containing lipid droplets and expression of HSC markers as well as pericyte markers. RSCs underwent spontaneous transdifferentiation into myofibroblasts in in vitro culture, which was inhibited by albumin expression or R-III treatment. We also evaluated the therapeutic effects of R-III in unilateral ureteral obstruction (UUO)-induced renal fibrosis in mice. Injected R-III localized predominantly in cytoglobin/stellate cell activation-associated protein (Cygb/STAP)-positive cells in the kidney and reduced renal fibrosis. These findings suggest that RSCs can be recognized as the renal counterparts of HSCs and that RSCs represent an attractive therapeutic target for anti-fibrotic therapy.


Introduction
Chronic kidney disease (CKD) represents a global health burden that affects >10% of adults worldwide, and renal fibrosis, particularly tubulointerstitial fibrosis, is the common final outcome of almost all progressive CKD [1]. Renal tubulointerstitial fibrosis is characterized by loss of tubular architecture with extracellular matrix (ECM) accumulation and tubular atrophy, resulting in organ dysfunction and kidney failure [2]. The traditional view is that myofibroblasts, responsible for excessive

Isolation of Rat HSCs and RSCs and Cell Culture
HSCs were isolated from male Sprague-Dawley rats (nine weeks old) as described previously [19]. For the isolation of RSCs, the kidneys were perfused in situ with phosphate-buffered saline (PBS) and then with Hank's buffered salt solution (HBSS) supplemented with collagenase, pronase (Sigma-Aldrich, St, Louis, MO, USA), and DNase (MP Biomedicals, Santa Ana, CA, USA) via the inferior vena cava. The perfused kidneys were dissected, and the outer membrane was removed. The kidney cell suspensions were further digested with HBSS supplemented with collagenase, pronase, and DNase for 10 min in a 37 • C water bath. The cells were then washed and centrifuged in a 13.4% Nycodenz gradient at 1400× g for 20 min without brake. The interface containing the enriched RSCs was collected and washed with HBSS. Then, the isolated RSCs were cultured in DMEM supplemented with 10% FBS. The purity of RSCs was assessed by microscopic observation. After reaching confluence in primary culture, the RSCs were passaged and used as activated RSCs.

Reverse Phase HPLC (RP-HPLC) Quantitation of Retinoids
Cells were quantified and extracted as described previously [23]. RP-HPLC was carried out on an AKTA Explorer HPLC system (GE Healthcare Life Sciences, Piscataway, NJ, USA). The chromatographic conditions were the same as previously described [24].

Quantitative Real-Time PCR
Total RNA was isolated using TRIzol (Ambion, Austin, TX, USA), and was used to synthesize the cDNA. Real-time PCR was performed on an ABI QuantStudio TM 3 Real-Time PCR system. To control for variations in the reaction, the PCR products were normalized against the mRNA levels of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The primers used are listed in Supplemental Table S1.

Western Blotting
Cell lysates were prepared for analyses by electrophoresis and immunoblotting as described previously [22]. The primary antibodies used are listed in Supplemental Table S2.

Animal Experiments
Eight-week-old male C57BL/6 mice were used for unilateral ureteral obstruction (UUO). Left kidney UUO was performed in mice (n = 18) with an established protocol, and the sham-operated group (n = 6) was subjected to the same surgical procedure without ureter ligation. For determining the therapeutic effects of R-III, UUO-treated mice were randomly divided into two groups and injected via the tail vein with saline or R-III (30 µg) every day for seven days, starting on day eight after UUO. Mice were euthanized 15 days post-surgery, and the kidney tissues were weighed and collected for analyses. All experiments were performed in duplicate.

Immunohistochemistry
Formalin-fixed, paraffin-embedded kidney tissues were sectioned (5-µm thick) and stained with PAS for histological analysis or MT for analysis of collagen deposition. Tissue sections were also subjected to immunohistochemistry with the antibodies listed in Supplemental Table S2, and 10 random fields were graded semiquantitatively, and the average score was calculated as described previously [26]. All histological examinations were performed by a trained pathologist in a blinded manner. The degree of tubulointerstitial fibrosis and injury was determined by assessing the MT-stained sections. Briefly, images were captured by digital imaging sequentially over the entire sagittal section incorporating the cortex and outer medulla (10-20 images), and each image was divided into a grid of 252 squares and scored for the presence of tubular injury. The final score was the percentage of squares in each image with a positive score, averaged for all images from a single kidney.

Double Immunostaining
Double immunostaining was performed on Ventana Benchmark XT immunostainer (Ventana Medical System, Tucson, AZ, USA) following the manufacturer's instructions.
Briefly, for His-tag:α-SMA and His-tag:Cygb/STAP, samples immobilized on slides were incubated with His-tag antibody (Bio-Rad, Hercules, CA, USA). The bound primary antibody was detected with the OptiView DAB IHC Detection Kit (Ventana Medical System). Subsequently, the samples were incubated with α-SMA or Cygb/STAP antibody, and the bound primary antibody was detected using the ultraView Universal AP Red Detection Kit (Ventana Medical System). For His-tag:CD31, His-tag:desmin, and His-tag:F4/80, the samples were incubated with an antibody against CD31, desmin, or F4/80, respectively, and the bound primary antibody was detected with the OptiView DAB IHC Detection Kit. Subsequently, the samples were incubated with His-tag antibody and the bound primary antibody was detected by the ultraView Universal AP Red Detection Kit. The samples were then counterstained with hematoxylin (Ventana Medical Systems).

Statistical Analysis
Results are expressed as the means ± standard deviation (SD). A paired t-test was used for analysis of in vitro data. Data from in vivo studies were analyzed by either the Wilcoxon rank-sum test or the Kruskal-Wallis test, followed by Dwass-Steel-Critchlow-Fligner (DSCF) multiple comparison test. Statistical analyses were performed using SPSS for Windows, version 20.0 (SPSS, Inc., Chicago, IL, USA). A p-value < 0.05 was considered significant.

Cells Resembling HSCs Are Present in the Kidney Tissues
Previous studies have suggested that vitamin A-storing cells are also present in the kidneys [15] and demonstrated that both Cygb/STAP and cellular retinol-binding protein 1 (CRBP-1), which are markers of HSCs, are expressed in the renal interstitium [27,28]. These findings prompted us investigate the possibility that renal counterparts of HSCs may exist. We isolated cells from the rat kidney tissues using the HSC isolation protocol with minor modifications and compared their morphological and biochemical characteristics with those of HSCs. 1~3 × 10 5 cells (referred to as renal stellate cells, RSCs) were isolated per rat and seeded into culture dishes. Results showed that, like HSCs on day three after seeding (HSCs d3; early-activated cells), RSCs on day three after seeding (RSCs d3) exhibited a flattened polygonal shape and contained lipid droplets in the cytoplasm, as assessed by transmission electron microscopy and oil red O staining ( Figure 1A,B, Supplemental Figure S2). Their cell size and lipid droplet size were, however, smaller compared with those of HSCs d3. In addition, similar to HSCs, which when cultured on plastic undergo spontaneous transdifferentiation/activation in vitro [10], RSCs on day seven after seeding (RSCs d7) and RSCs after passage two (RSCs P2) transformed into myofibroblast-like cells and displayed loss of lipid droplets and increased expression of collagen type I and α-SMA, both well-known markers for activated HSCs/myofibroblasts ( Figure 1B-D, and Figure 2A). Next, reverse-phase HPLC analysis revealed that RSCs contained retinol (vitamin A) whose level was decreased in RSCs d7 compared to that in RSCs d3, whereas the level of all-trans retinoic acid (ATRA), a bioactive vitamin A metabolite, increased with time in culture ( Figure 1E,F). The observed changes in the retinoid levels in RSCs are in agreement with previous findings for HSCs [24,29], although the retinoid levels were lower in RSCs than in HSCs. Furthermore, real-time PCR analysis showed that, as observed in HSCs, the mRNA expression levels of markers of quiescent HSCs, such as PPARγ, C/EBPα, angiogenic factor with G patch and FHA domains 1 (Aggf1), and albumin [19,[30][31][32], were decreased in RSCs d7 compared with those in RSCs d3 ( Figure 1G,H). These finding suggest that the cells, which we isolated from the kidney tissues, are renal counterparts of HSCs. RSCs contained retinol (vitamin A) whose level was decreased in RSCs d7 compared to that in RSCs d3, whereas the level of all-trans retinoic acid (ATRA), a bioactive vitamin A metabolite, increased with time in culture ( Figure 1E,F). The observed changes in the retinoid levels in RSCs are in agreement with previous findings for HSCs [24,29], although the retinoid levels were lower in RSCs than in HSCs. Furthermore, real-time PCR analysis showed that, as observed in HSCs, the mRNA expression levels of markers of quiescent HSCs, such as PPARγ, C/EBPα, angiogenic factor with G patch and FHA domains 1 (Aggf1), and albumin [19,[30][31][32], were decreased in RSCs d7 compared with those in RSCs d3 ( Figure 1G,H). These finding suggest that the cells, which we isolated from the kidney tissues, are renal counterparts of HSCs.

Albumin Expression and R-III Treatment Inhibited the Transdifferentiation/Activation of RSCs In Vitro
Confocal microscopy showed that RSCs d3 (early-activated cells), but not RSCs P2 (activated/myofibroblastic cells), expressed albumin and emanated autofluorescence of vitamin A. On the other hand, α-SMA expression was detected in RSCs P2 ( Figure 2A). Intriguingly, when RSCs P2 were transfected with a plasmid encoding albumin, they transformed into distinct cell types with autofluorescent lipid droplets reappearing and reduced α-SMA expression, as observed in HSCs [19]. Treatment of RSCs P2 with R-III also induced a phenotypic reversion into fat-storing cells and decreased the mRNA expression levels of α-SMA and collagen type I ( Figure 2B and 2C). However, treatment with bovine serum albumin, which scarcely enters RSCs, exerted no significant effect on α-SMA expression (Supplemental Figure S3). Next, we analysed the expression of other SC markers in RSCs. Western blotting revealed that lecithin retinol acyltransferase (LRAT), an enzyme responsible for retinyl ester formation and a marker for quiescent HSCs [33], was expressed in freshly isolated were analyzed by real-time PCR. The data represent the means ± SD for three independent experiments. p-value was calculated using paired t-test (compared with cells on day 3 after seeding). * p < 0.05, ** p < 0.01, *** p < 0.001. (E) Retinol levels in whole-cell extracts of RSCs d3 and HSCs d3 were measured by reverse-phase HPLC. (F) Graphical representation of intracellular retinol and ATRA levels in RSCs and HSCs on days 3 and 7 after seeding.

Albumin Expression and R-III Treatment Inhibited the Transdifferentiation/Activation of RSCs In Vitro
Confocal microscopy showed that RSCs d3 (early-activated cells), but not RSCs P2 (activated/myofibroblastic cells), expressed albumin and emanated autofluorescence of vitamin A. On the other hand, α-SMA expression was detected in RSCs P2 ( Figure 2A). Intriguingly, when RSCs P2 were transfected with a plasmid encoding albumin, they transformed into distinct cell types with autofluorescent lipid droplets reappearing and reduced α-SMA expression, as observed in HSCs [19]. Treatment of RSCs P2 with R-III also induced a phenotypic reversion into fat-storing cells and decreased the mRNA expression levels of α-SMA and collagen type I ( Figure 2B,C). However, treatment with bovine serum albumin, which scarcely enters RSCs, exerted no significant effect on α-SMA expression (Supplemental Figure S3). Next, we analysed the expression of other SC markers in RSCs. Western blotting revealed that lecithin retinol acyltransferase (LRAT), an enzyme responsible for retinyl ester formation and a marker for quiescent HSCs [33], was expressed in freshly isolated RSCs, but not in RSCs P2 ( Figure 2D). Immunofluorescence and real-time PCR analyses also confirmed the expression of the HSC markers Cygb/STAP and CRBP-1, respectively ( Figure 2E,F) [25,34]. CRBP1 expression was not detected in L6 rat myoblasts, which was used as a negative control. Thus, our findings suggest that albumin and its derivative R-III inactivates RSCs. RSCs, but not in RSCs P2 ( Figure 2D). Immunofluorescence and real-time PCR analyses also confirmed the expression of the HSC markers Cygb/STAP and CRBP-1, respectively ( Figure 2E,F) [25,34]. CRBP1 expression was not detected in L6 rat myoblasts, which was used as a negative control. Thus, our findings suggest that albumin and its derivative R-III inactivates RSCs.   Figure S4. (E) RSCs d3 and RSCs P2 were analyzed by immunofluorescence using anti-Cygb/STAP antibody. (F) Total RNA was extracted from RSCs d3, RSCs P2, and L6 rat myoblasts, and the CRBP-1 expression was analyzed by real-time PCR. The data represent the means ± SD for three independent experiments. p-value was calculated using paired t-test (n = 3). * p < 0.05, ** p < 0.01.

R-III Administration Reduced UUO-Induced Renal Fibrosis
Based on the in vitro effects of R-III on RSCs, we evaluated its therapeutic effects on renal fibrosis. UUO was performed in C57BL/6J mice and R-III was injected via the tail vein daily for seven days, starting on day eight after UUO (Supplemental Figure S5). Histological analysis by Masson's trichrome (MT) and periodic acid-Schiff (PAS) staining revealed that UUO induced interstitial fibrosis and tubular atrophy, which was significantly reduced by R-III treatment ( Figure 3A). The histopathologic degree of tubulointerstitial fibrosis was scored in MT-stained sections, and we found that R-III administration reduced fibrosis by~55% ( Figure 3A, right panel). Next, we performed immunohistochemical staining to visualize the pro-fibrotic protein expression in the renal tissue. The expression of TGF-β1, plasminogen activator inhibitor 1 (PAI-1), and collagen type 1 and type IV was increased in UUO kidneys compared with that in sham kidneys; however, this effect significantly inhibited by R-III treatment ( Figure 3B).
(B) and real-time PCR for the expression of α-SMA and collagen type I (C). (D) Cell lysates from freshly isolated RSCs and RSCs P2 were analyzed by western blotting. α-tubulin was used as a loading control. Full-length blots are presented in Supplementary Figure S4. (E) RSCs d3 and RSCs P2 were analyzed by immunofluorescence using anti-Cygb/STAP antibody. (F) Total RNA was extracted from RSCs d3, RSCs P2, and L6 rat myoblasts, and the CRBP-1 expression was analyzed by real-time PCR. The data represent the means ± SD for three independent experiments. p-value was calculated using paired t-test (n = 3). * p < 0.05, ** p < 0.01.

R-III Administration Reduced UUO-Induced Renal Fibrosis
Based on the in vitro effects of R-III on RSCs, we evaluated its therapeutic effects on renal fibrosis. UUO was performed in C57BL/6J mice and R-III was injected via the tail vein daily for seven days, starting on day eight after UUO (Supplemental Figure S5). Histological analysis by Masson's trichrome (MT) and periodic acid-Schiff (PAS) staining revealed that UUO induced interstitial fibrosis and tubular atrophy, which was significantly reduced by R-III treatment ( Figure 3A). The histopathologic degree of tubulointerstitial fibrosis was scored in MT-stained sections, and we found that R-III administration reduced fibrosis by ~55% ( Figure 3A, right panel). Next, we performed immunohistochemical staining to visualize the pro-fibrotic protein expression in the renal tissue. The expression of TGF-β1, plasminogen activator inhibitor 1 (PAI-1), and collagen type 1 and type IV was increased in UUO kidneys compared with that in sham kidneys; however, this effect significantly inhibited by R-III treatment ( Figure 3B). Kidney sections from sham-and UUO-and UUO+R-III-treated mice were subjected to immunohistochemistry against TGF-β1, PAI-1, and collagen type I and type IV. Representative images from each study group are shown. Semiquantitative analysis of the staining intensity for each group is shown (right). Data are expressed as the means ± SD (p-value; Kruskal-Wallis test, followed by DSCF multiple comparison test). * p < 0.05, ** p < 0.01, *** p < 0.001.

R-III Administration Reduced Pro-Fibrotic Marker Expression and Macrophage Infiltration in UUO Kidneys
Consistent with the immunohistochemical staining results ( Figure 3B), real-time PCR revealed that the mRNA expression of TGF-β1, PAI-1, and collagen type I was increased in UUO kidneys compared with that in sham kidneys; however, this effect was suppressed by R-III administration ( Figure 4A). Immunohistochemical analysis also showed that the protein expression of myofibroblast markers, such as α-SMA, desmin, and fibronectin was enhanced in UUO kidneys, but this effect was inhibited by R-III treatment ( Figure 4B). Next, we assessed macrophage infiltration, as increased infiltration of macrophages has been reported in UUO kidneys [35]. Infiltration of F4/80-positive macrophages was remarkably increased in UUO kidneys compared with that in sham kidneys; however, this effect was suppressed by R-III treatment ( Figure 4C). In addition, the mRNA expression level of monocyte chemoattractant protein 1 (MCP-1), a pro-inflammatory gene, also roughly paralleled the intensity of F4/80 immunostaining ( Figure 4A).

R-III Was Delivered to RSCs In Vivo
We then investigated the cellular distribution of injected His-tagged R-III by immunohistochemistry. Unlike the sham kidneys, the UUO+R-III-treated kidneys showed distinct His-tag staining. The His-tag-positive cells were predominantly located in the peritubular and perivascular region, although a weak signal was also detected in epithelial cells ( Figure 5A). Importantly, double immunostaining revealed that ~70% of Cygb/STAP + cells and ~40% of α-SMA + Data are expressed as the means ± SD (p-value; Kruskal-Wallis test, followed by DSCF multiple comparison test). * p < 0.05, ** p < 0.01, *** p < 0.001 (B) Kidney sections from sham-, UUO-, and UUO+R-III-treated mice were subjected to immunohistochemistry against α-SMA, desmin, and fibronectin. Representative images from each study group are shown. Semiquantitative analysis of the staining intensity in each group is shown (right). (C) The kidney sections from treated mice were subjected to immunohistochemistry against F4/80. Arrows indicate F4/80-positive cells. The number of F4/80-positive cells was quantified (right).

R-III Was Delivered to RSCs In Vivo
We then investigated the cellular distribution of injected His-tagged R-III by immunohistochemistry. Unlike the sham kidneys, the UUO+R-III-treated kidneys showed distinct His-tag staining. The His-tag-positive cells were predominantly located in the peritubular and perivascular region, although a weak signal was also detected in epithelial cells ( Figure 5A). Importantly, double immunostaining revealed that~70% of Cygb/STAP + cells and~40% of α-SMA + cells showed His-tag signals ( Figure 5B). However, no significant overlap was found with other cellular markers, such as CD31 and F4/80, indicating that His-tagged R-III was not randomly distributed in different types of cells but successfully delivered to Cygb/STAP-positive RSCs. PDGFR-β has been shown to be expressed in myofibroblasts and renal mesenchymal cells, including mesangial cells, interstitial fibroblasts and pericytes, and vascular smooth musle cells [36]. Intriguingly,~55% of interstitial PDGFR-β + cells expressed Cygb/STAP ( Figure 6C). cells showed His-tag signals ( Figure 5B). However, no significant overlap was found with other cellular markers, such as CD31 and F4/80, indicating that His-tagged R-III was not randomly distributed in different types of cells but successfully delivered to Cygb/STAP-positive RSCs. PDGFR-β has been shown to be expressed in myofibroblasts and renal mesenchymal cells, including mesangial cells, interstitial fibroblasts and pericytes, and vascular smooth musle cells [36]. Intriguingly, ~55% of interstitial PDGFR-β + cells expressed Cygb/STAP ( Figure 6C).

EMT Is Not Involved in the Anti-Fibrotic Effect of R-III
As EMT was previously reported to contribute to renal fibrosis [4], we examined the effects of R-III on TGF-β-induced EMT. TGF-β1 treatment caused a decrease in E-cadherin and increase in α-SMA levels in NRK-52E rat renal tubular epithelial cells, and R-III treatment partially inhibited TGF- For His-tag:CD31, His-tag:desmin, and His-tag:F4/80, His-tag antibody stains red, and antibodies against CD31, desmin, and F4/80 stain brown. For His-tag:Cygb/STAP and His-tag:α-SMA, His-tag antibody stains brown, and antibodies against Cygb/STAP and α-SMA stain red. His-tag staining largely overlaps with Cygb/STAP staining. Arrows indicate co-localization of His-tag and selected markers. Semiquantitative scoring of His-tag + selected marker + double-positive cells as a percentage of selected marker-stained cells is shown on the right. Data are expressed as the means ± SD (C) Double immunohistochemistry was performed for kidney sections from UUO-treated mice. For PDGFR-β:Cygb/STAP, PDGFR-β antibody stains brown, and Cygb/STAP antibody stains red. Figure 6. RSCs express pericyte markers. (A) NRK-49F rat kidney fibroblast cells were treated with TGF-β1 (50 ng/mL) in the presence or absence of R-III (0.5 μM), and α-SMA expression was analyzed by real-time PCR. The data represent the means ± SD for three independent experiments. p-value was calculated using paired t-test (n = 3). * p < 0.05. (B) Kidney sections from sham-, UUO-, and UUO+R-III-treated mice were subjected to immunohistochemistry against STRA6. Arrows indicate STRA6positive cells. Semiquantitative analysis of STRA6 staining is shown (right). Data are expressed as the means ± SD (p-value; Kruskal-Wallis test, followed by DSCF multiple comparison test). ** p < 0.01, *** p < 0.001. (C) Total RNA was extracted from the kidneys of sham-, UUO-, and UUO+R-III-treated mice, and STRA6 expression was analyzed by real-time PCR. (D) Transmission electron microscopy showing the lipid droplet-containing cells in the peritubular region of UUO+R-III-treated kidneys. SC, stellate cells; TE, tubular epithelial cells; L, lipid; C, collagen fibers. (E,F) Total RNA was isolated from RSCs (E) and HSCs (F) on days 1 and 7 after seeding, and the expression of collagen type I, α-SMA, PDGFR-β, and NG2 was analyzed by real-time PCR. p-value was calculated using paired t-test (n = 3, compared with cells on day 1 after seeding). * p < 0.05, ** p < 0.01.  5 µM), and α-SMA expression was analyzed by real-time PCR. The data represent the means ± SD for three independent experiments. p-value was calculated using paired t-test (n = 3). * p < 0.05. (B) Kidney sections from sham-, UUO-, and UUO+R-III-treated mice were subjected to immunohistochemistry against STRA6. Arrows indicate STRA6-positive cells. Semiquantitative analysis of STRA6 staining is shown (right). Data are expressed as the means ± SD (p-value; Kruskal-Wallis test, followed by DSCF multiple comparison test). ** p < 0.01, *** p < 0.001. (C) Total RNA was extracted from the kidneys of sham-, UUO-, and UUO+R-III-treated mice, and STRA6 expression was analyzed by real-time PCR. (D) Transmission electron microscopy showing the lipid droplet-containing cells in the peritubular region of UUO+R-III-treated kidneys. SC, stellate cells; TE, tubular epithelial cells; L, lipid; C, collagen fibers. (E,F) Total RNA was isolated from RSCs (E) and HSCs (F) on days 1 and 7 after seeding, and the expression of collagen type I, α-SMA, PDGFR-β, and NG2 was analyzed by real-time PCR. p-value was calculated using paired t-test (n = 3, compared with cells on day 1 after seeding). * p < 0.05, ** p < 0.01.

EMT Is Not Involved in the Anti-Fibrotic Effect of R-III
As EMT was previously reported to contribute to renal fibrosis [4], we examined the effects of R-III on TGF-β-induced EMT. TGF-β1 treatment caused a decrease in E-cadherin and increase in α-SMA levels in NRK-52E rat renal tubular epithelial cells, and R-III treatment partially inhibited TGF-β1-driven EMT (Supplemental Figure S6A,B). Next, we prepared cell lysates from the kidney tissues and analyzed EMT marker expression by western blotting. In line with our immunohistochemical data ( Figure 4B), α-SMA protein levels were increased in UUO kidneys, but this effect was inhibited by R-III administration (Supplemental Figure S6C,D). However, we did not find any statistically significant changes in the levels of other EMT markers, such as E-cadherin and fibroblast-specific protein 1 (FSP-1), suggesting that the in vivo anti-fibrotic effect of R-III was not due to inhibition of EMT.

R-III Does not Affect NRK-49F Renal Fibroblasts
It has been previously suggested that myofibroblasts in the kidney represent an activated population of residual fibroblasts [37]. Hence, we examined the effect of R-III on NRK-49F rat renal fibroblasts. TGF-β1 treatment increased α-SMA mRNA levels in NRK-49 cells, which was consistent with previous findings [38], but R-III treatment had no significant effect ( Figure 6A). Thus, it is unlikely that the in vivo anti-fibrotic effect of R-III is derived from its action on resident renal fibroblasts.

STRA6 Is Expressed in the Kidneys
As the entry of R-III into HSCs is dependent on STAR6, a specific membrane receptor for RBP [21,22], the expression of STRA6 was examined in the kidney. Immunohistochemistry showed that STRA6 staining was mainly detected in the peritubular and perivascular cells, resembling His-tag staining ( Figure 5A), and significantly increased in UUO kidneys compared with that in sham kidneys ( Figure 6B). Real-time PCR analysis revealed that STRA6 mRNA expression levels roughly paralleled the intensity of STRA6 immunostaining ( Figure 6C).

Lipid Droplet-Containing Cells Appeared in R-III-Treated UUO Kidney
As R-III treatment induced phenotypic reversion of RSCs P2 (activated/myofibroblastic cells) into lipid droplet-containing cells in vitro ( Figure 2B), we examined whether the reappearance of fat-storing cells also occurs in vivo. Consistent with previous findings [15], electron microscopy data revealed that lipid droplet-containing cells were not present in UUO kidneys ( Figure 6D). Lipid droplet-containing cells were, however, detected in the peritubular region of the UUO+R-III-treated kidneys.

RSCs Express Pericyte Markers
HSCs are pericytes found in the perisinusoidal space of the liver [9]. To verify the identity of our isolated RSCs, we examined the expression of the commonly used pericyte markers, including α-SMA, PDGFR-β, and chondroitin sulfate proteoglycan neural/glial antigen 2 (NG2) [39], by real-time PCR. Results showed that the mRNA expression levels of α-SMA, PDGFR-β, and NG2 in RSCs were increased during cell culture ( Figure 6E). However, the NG2 mRNA level was decreased in HSCs d7 ( Figure 6F). Notably, the mRNA levels of collagen type I and α-SMA in RSCs d7 (activated cells) were >1000-fold higher when compared to those in RSCs d1 (pre-activated cells). These results suggest that our isolated cells from the kidney tissues may be, at least in part, renal pericytes and that the in vivo anti-fibrotic effect of R-III is probably due to its action on RSCs.

Discussion
Hepatic and pancreatic stellate cells were first isolated and characterized from the rat liver and pancreas in mid 1980s and late 1990s, respectively [40,41]. In this study, we isolated stellate cells from rat kidneys using the standard protocol for HSC isolation and characterized them. We also tried to isolate RSCs from mouse kidneys but failed to obtain sufficient cells for analysis, probably due to either technical reasons or a low quantify of cell population or both. (Similarly, it is much more difficult to isolate mouse HSCs than rat HSCs.) RSCs displayed morphological and biochemical characteristics similar to those of HSCs. RSCs d1 (pre-activated) or d3 (early-activated) contained autofluorescent, cytoplasmic lipid droplets and showed little or no α-SMA expression, as has been observed in HSCs and PSCs [10,13]. By contrast, culture-activated RSCs d7, P1, or P2 lost most of their lipid droplets and displayed a myofibroblastic phenotype with increased expression of α-SMA and collagen type I. Such a transition into myofibroblasts during cell culture resembles the activation process of HSCs and PSCs. Furthermore, RSCs were found to contain Vitamin A and express the stellate cell markers, such as LRAT, Cygb/STAP, and CRBP-1. These findings indicate that our isolated cells are renal counterparts of HSCs/PSCs and differ from fibroblasts.
Previous studies have suggested that scar-producing myofibroblasts following kidney injury are mainly derived from pericytes and perivascular fibroblasts, based on their perivascular localization and expression of PDGFR-β [6,7]. A question then arises; are the isolated RSCs are renal pericytes? Our study showed that RSCs express common pericyte markers, PDGFR-β and NG2, and that Cygb/STAP (or His-tag)-positive cells are located in perivascular region. This suggests that RSCs may be, at least in part, renal pericytes as HSCs are liver pericytes, although clear evidence is lacking. Further study is required to address this issue, but there is a serious technical problem that there are no specific markers for stellate cells and pericytes. Commonly used markers, including PDGFR-β, NG2, and Cygb/STAP, are expressed in different types of interstitial, mesenchymal stromal cells [36,42,43].
Consistent with a previous finding that the expression of NG2 was increased in pericytes/coll1a1 + cells after UUO [6], we found that the mRNA expression of NG2 in RSCs was increased during culture activation. By contrast, NG2 expression was decreased in HSCs d7 (activated HSCs) compared to that in HSCs d1. The mechanism underlying the differential regulation of NG2 expression and its physiological significance are not yet clear and require further investigation.
Despite intensive research efforts over the last 30 years, the molecular mechanism of HSC/PSC activation remains elusive. Quiescent HSCs store retinoids in the form of retinyl esters in the cytoplasmic lipid droplets. When the lipid droplets are rapidly lost during SC activation, a portion of their retinoid content is likely released and metabolized into retinaldehyde, which is further irreversibly oxidized to retinoic acid (RA) by retinaldehyde dehydrogenase. A role of retinoids in HSC activation has been proposed, but previous reports regarding the effects of exogenous retinoids on HSC activation and liver fibrosis are controversial [44]. Recently, we demonstrated that endogenous RA plays a role in HSC activation [22], which was in agreement with a previous study that showed that the expression of aldehyde dehydrogenase 1 family member A2, also known as retinaldehyde dehydrogenase 2, was strikingly upregulated in myofibroblasts in UUO kidney compared to that in their precursors [45]. Biophysical studies have shown that albumin readily binds to RA [18,46]. This explains why albumin expression or R-III treatment markedly reduces intracellular RA levels and RA signaling in activated HSCs [22]. The mechanism of action of R-III downstream of RA sequestration is now under investigation.
In this study, we showed that R-III treatment induces phenotypic reversion of activated/myofibroblastic RSCs into fat-storing phenotype but exerts no significant effects on renal fibroblasts. In addition, R-III administration led to the reappearance of lipid droplet-containing cells in R-III-treated kidneys. Thus, it is reasonable to assume that the in vivo anti-fibrotic effect of R-III is due to its action on RSCs. The UUO mouse model was used because (1) this significantly reduced the amount of R-III needed for these critical in vivo experiments and (2) a recent Meta-analysis study of the UUO model suggests that both mouse and rat models equally contribute to the study of renal fibrosis [47].
The stellate cells of the liver and extrahepatic organs, such as the pancreas and kidneys, show striking similarities with respect to morphology and perivascular location, which suggests that their activation contributes to the myofibroblast population present in fibrotic tissues [13]. There is an imperative demand for the development of tissue fibrosis therapy and several studies have been previously conducted with an aim to inactivate SCs and reduce tissue fibrosis, but unfortunately an effective therapy for fibrosis is not yet available.

Conclusions
We demonstrated that our isolated RSCs from the kidney tissues are the renal counterparts of HSCs, and that R-III exerts anti-fibrotic effects on RSC transdifferentiation/activation and renal fibrosis, similar to its effects on HSC activation and liver fibrosis. Hence, our findings suggest that R-III, designed for stellate cell targeting, may serve as a novel anti-fibrotic drug candidate.
Supplementary Materials: The following are available online at http://www.mdpi.com/2227-9059/8/10/431/s1, Figure S1: Schematic diagram of R-III, a retinol-binding protein (RBP) -albumin domain III fusion protein, in comparison with full-length albumin and RBP. The numbers indicate amino acids, Figure S2: Differential interference contrast images of oil red O staining for renal stellate cells on day 3 after seeding, Figure S3: Renal stellate cells after passage 2 were treated with bovine serum albumin (BSA; 0.5 µM) or R-III (0.5 µM) for 20 h, and α-smooth muscle actin (α-SMA) expression was analyzed by real-time PCR, Figure S4: Uncropped full-length blot images, Figure S5: Schedules for developing the murine kidney injury model using unilateral ureteral obstruction (UUO), Figure S6: Effects of R-III on epithelial-mesenchymal transition (EMT) in vitro and in vivo, Table S1: Primers used in this study, Table S2: Primary antibodies used in this study, Supplemental Method: Epithelial-mesenchymal transition (EMT).