Novel Mechanism of the Pericyte-Myofibroblast Transition in Renal Interstitial Fibrosis: Core Fucosylation Regulation

Pericytes have been identified as a major source of myofibroblasts in renal interstitial fibrosis (RIF). The overactivation of several signaling pathways, mainly the TGF-β and PDGF pathways, initiates the pericyte-myofibroblast transition during RIF. Key receptors in these two pathways have been shown to be modified by fucosyltransferase 8 (FUT8), the enzyme that catalyzes core fucosylation. This study postulated that core fucosylation might play an important role in regulating the pericyte transition in RIF. The data showed that core fucosylation increased with the extent of RIF in patients with IgA nephropathy (IgAN). Similarly, core fucosylation of pericytes increased in both a unilateral ureteral occlusion (UUO) mouse model and an in vitro model of pericyte transition. Inhibition of core fucosylation by adenoviral-mediated FUT8 shRNA in vivo and FUT8 siRNA in vitro significantly reduced pericyte transition and RIF. In addition, the activation of both the TGF-β/Smad and PDGF/ERK pathways was blocked by core fucosylation inhibition. In conclusion, core fucosylation may regulate the pericyte transition in RIF by modifying both the TGF-β/Smad and PDGF/ERK pathways. Glycosylation might be a novel “hub” target to prevent RIF.

Based on the results of our previous study and other reports, core fucosylation modifies TGF-βR and PDGFRβ; furthermore, inhibition of the core fucosylation of TGF-βR alleviates RIF in rat models of unilateral ureteral obstruction (UUO) and renal tubular cell injury in vitro 20,21 . Core fucosylation of proteins is catalyzed by α1,6-fucosyltransferase (FUT8) in the Golgi apparatus, which adds fucose to the innermost GlcNAc residue of N-linked oligosaccharides on glycoproteins, and this modification is preferentially recognized by Lens culinaris lectin (LCA) [22][23][24] .
This study postulated that core fucosylation might be a potential target to simultaneously prevent these two receptors from triggering the activation of downstream intermediates. We investigated the effect of core fucosylation on the activities of both the TGF-β/Smad and PDGF/ERK pathways during the pericyte transition in RIF. The down-regulation of core fucosylation prevented the pericyte transition and RIF by inhibiting the activities of both the TGF-β/Smad and PDGF/ERK pathways. This study is the first to suggest that the core fucosylation of pericytes represents a promising "hub" target for RIF.

Core Fucosylation of Pericytes is Increased in Renal Biopsies of Patients with IgAN.
Thirtytwo patients with IgAN were divided into three groups based on T scores of renal biopsies according the Oxford IgAN classification [22][23][24] . Clinical characteristics of the patients are shown in Table 1. Serum creatinine levels were significantly increased, whereas the estimated glomerular filtration rate (eGFR) was decreased in the T2 patients compared to patients in the T0 group (P < 0.05). A greater number of pericytes (PDGFRβ+) detached from interstitial endothelial cells (CD31+) in patients with T1 and T2 scores ( Fig. 1a and b), and PDGFRβ+ cells showed enhanced expression of α-SMA (Fig. 1c). The LCA and FUT8 levels in pericytes increased as the T scores increased (Fig. 1d). According to the correlation analysis, the detachment, transition, and core fucosylation of pericytes were significantly related to T scores (Fig. 1e, r1 = 0.946, r2 = 0.948, r3 = 0.946, r4 = 0.9, P < 0.01). These findings indicated that core fucosylation of pericytes increased when pericytes were activated in the RIF process of patients with IgAN.
Core Fucosylation of Pericytes is Increased in UUO Mouse Models. We further detected the pericyte transition and core fucosylation in UUO mouse models (Fig. 2a). Pericytes began to detach from the interstitial endothelial cells as early as 1 day after UUO. A greater number pericytes subsequently detached, and the expression of α-SMA increased in renal interstitial areas over time ( Fig. 2b and c). Immunofluorescence co-staining was used to evaluate the relationship between core fucosylation, pericytes and myofibroblasts. The FUT8 and LCA levels in pericytes were substantially increased in a time-dependent manner, and similar increases were also observed in myofibroblasts ( Fig. 3a and b). According to the correlation analysis, the detachment, transition, and core fucosylation of pericytes were significantly related to the severity of RIF (Fig. 3c, r1 = 0.962, r2 = 0.813, r3 = 0.907, r4 = 0.830, P < 0.01).

Down-regulation of Core Fucosylation Alleviates the Pericyte Transition
In Vitro. We then established an in vitro pericyte transition model to further investigate the role of core fucosylation. C57BL/6 mouse kidney pericytes were isolated and cultured in vitro (Supplementary Fig. S1A-C). The profibrotic factor TGF-β1 was used to induce the pericyte transition. After 24 h of TGF-β1 induction, myofibroblast-like morphological changes were observed in pericytes, along with increased expression of α-SMA (Fig. 4a). Both immunofluorescence staining and Western blot analyses showed that FUT8 and LCA expression increased when pericytes transitioned to myofibroblasts after TGF-β1 induction in a time-dependent manner ( Fig. 4b-d). Furthermore an FUT8 siRNA was used to successfully knockdown FUT8 expression ( Supplementary Fig. S2A), and LCA was expressed at very low levels after FUT8 knockdown (Fig. 5c). Myofibroblast-like morphological changes in pericytes were substantially alleviated after FUT8 knockdown, along with a decrease in α-SMA expression ( Fig. 5a and b).

Inhibition of Core Fucosylation Prevents the Pericyte Transition and RIF in UUO Mouse Models.
To further confirm the role of core fucosylation in vivo, we knocked down FUT8 to evaluate the subsequent pericyte transition and RIF. The FUT8 shRNA recombinant adenovirus vector used for FUT8 knockdown in vivo was constructed as described in our previous research 25 (Supplementary Fig. S2B and C). Injection with the FUT8 shRNA recombinant adenovirus dramatically decreased pericyte detachment and the transition to myofibroblasts in renal interstitial areas, followed by the alleviation of RIF in UUO mice ( Fig. 6a and b). Based on these findings, core fucosylation played a key regulatory role in the pericyte transition and RIF. Core Fucosylation Regulates the Activities of Both the TGF-β/Smad and PDGF/ERK Pathways in the Pericyte Transition. We postulated that the beneficial effects of the FUT8 siRNA on RIF are due to its simultaneous inactivation of several profibrotic pathways. As TGF-β and PDGF pathways have been recognized as key and classical profibrotic signaling pathways in RIF 15-17 , we further investigated the effect of the FUT8 siRNA on these pathways. Immunoprecipitation of LCA showed that core fucosylation of both TGF-βR1 and PDGFRβ significantly increased after TGF-β1 induction over time and was alleviated by FUT8 knockdown in vitro (Fig. 7a). Moreover, the levels of phosphorylated Smad2/3 and ERK1/2 were significantly decreased in response to FUT8 knockdown (Fig. 7b). Thus, the inhibition of core fucosylation simultaneously inactivated the TGF-β/Smad and PDGF/ERK pathways and alleviated the pericyte transition.

Discussion
Recent studies have confirmed that pericytes are activated and transition to myofibroblasts, which is the direct cause of RIF, but the regulatory mechanisms are unclear. Most proteins are glycoproteins, and glycosylation has a role in protein function. Our study elucidated the regulatory mechanisms involving glycosylation. Previous reports have shown that FUT8 can globally modify surface antigens, receptors, and adhesion molecules [26][27][28][29][30][31] . TGF-βR and PDGFRβ have been reported to be modified by core fucosylation, which is directly associated with their function 20,32,33 . Our previous work showed that inhibition of core fucosylation could alleviate RIF in UUO rat models 20,21 . Thus, we hypothesized that core fucosylation modification of multiple signaling pathways might be involved in pericyte transition in RIF. IgAN is one of the most common causes of CKD, and end-stage renal disease 34 . The kidney prognosis of IgAN is more closely associated with the severity of interstitial injury and fibrosis than that of glomerular lesions [35][36][37] . In renal biopsies from patients with IgAN, we observed a significant increase in the core fucosylation of pericytes that was highly correlated with the severities of renal interstitial injuries represented by T grading. To the best of our knowledge, we are the first to report this core fucosylation disorder of pericytes in patients with IgAN. Figure 3. The pericyte-myofibroblast transition and core fucosylation were increased in UUO mouse models (n = 5). (a) Representative images of dual staining for PDGFRβ (green) and FUT8 (red) and dual staining for LCA (green) and PDGFRβ (red). (b) Representative images of dual staining for FUT8 (green) and α-SMA (red) and dual staining for LCA (green) and α-SMA (red). (c) Spearman's correlation coefficients were used to examine the differences in the fluorescence intensity of immunostaining for PDGFRβ, α-SMA, FUT8, and LCA in mice with different RIF scores. r, Spearman's correlation coefficient; r1, PDGFRβ; r2, α-SMA; r3, FUT8; r4, LCA. Scale bar, 50 μm. *P < 0.01, † P < 0.01. * and † indicate the comparison of the control group with the UUO group.
Scientific RePORtS | 7: 16914 | DOI:10.1038/s41598-017-17193-5 Although pericytes do not have specific markers, the diverse characteristics of pericytes indicate that they have large differentiation capacities. Other studies have found that pericytes lack expression of NG2 but express PDGFR-β in adult mouse kidney 7,38 . Fluorescence-activated cell sorting (FACS) is traditionally applied for PDGFRβ+ pericyte isolation from kidney 16 . However, our preliminary experiment found that the positive isolation percentage was greatly affected by the impurities during tissue preparation, such as lipids or dead cell debris. To minimize the interference of impurities, Percoll was applied to form a gradient density to remove impurities. FACS or magnetic-activated cell sorting (MACS) was then used for pericyte isolation, but MACS increased cell viability.
Consistent with previous reports 6,9,12,39 , we found that pericytes detached from endothelial cells and transitioned to myofibroblasts in primary cultured pericyte transition models and UUO mouse models. In addition, we found that core fucosylation was increased in experimental models and in patients with IgAN. However, core fucosylation did not increase in all pericytes. As shown in the study by Ching-Fang Wu et al., pericyte proliferation is substantially increased in the kidneys of subjects with UUO, and our results were consistent with their findings 17 . Thus, we postulate that some pericytes were proliferating and had not completely transitioned to myofibroblasts; thus, core fucosylation levels may start to increase when pericytes are in a pre-transition or transitioned status. As shown in Fig. 3b, almost all myofibroblasts showed high levels of fucosylation. Then, we inhibited FUT8 expression in vivo and in vitro and found that both the pericyte transition and RIF were dramatically alleviated. These findings indicated that core fucosylation was closely associated with the pericyte transition and RIF.
Pathological activation of pericytes is caused by various factors. The balance of the renal interstitial micro-environment is precisely regulated by elaborate adjacent structures between tubules, capillaries, and cross-talk among cells. Once the micro-environment balance is disrupted, pericyte activation and transition are initiated. Cross-talk, such as among TGF-β, PDGF, VEGF, CTGF, and Wnt, is commonly found among epithelial cells, endothelial cells, and pericytes, as well as within cells. Blockade of a single signaling pathway can inhibit RIF to some extent, but activation of other pathways would increase in compensation 40,41 . TGF-β1 signaling induces the phosphorylation of Smad proteins, which control the expression levels of fibrogenic genes through TGF-βRI and TGF-βRII receptor complexes 25,42 . MAPK signaling pathways in pericytes have a critical function in the progression of renal fibrosis 16,43 . In UUO mouse models, PDGF was reported to stimulate ERK1/2 activation in pericytes 44 . Both the TGF-β/Smad and PDGF/ERK pathways were activated following TGF-β1 induction in the present study, and decreased core fucosylation of receptors reduced the activities of downstream pathways. Core fucosylation has been shown to affect the binding between receptors and ligands, and further regulates the activity of downstream signaling pathways [45][46][47] . Therefore, we hypothesized that core fucosylation might be a promising "hub" target to prevent the progression of RIF and CKD.
Future research should be performed to evaluate the superiority of blocking multiple signaling pathways through core fucosylation over blockade of a single pathway in preventing RIF, to explore the underlying mechanisms regulating core fucosylation during the pericyte transition, and to clarify the precise locus of core fucosylation for the synthesis of small molecule drugs that modify glycosylation. All surgical procedures were conducted by a single surgeon under aseptic conditions in the Laboratory Animal Unit. Mice were anesthetized using an intraperitoneal injection of freshly prepared 10% chloral hydrate. A midline incision was made in the abdominal wall, and the left ureter was isolated and ligated using a 4.0 silk suture at two points along its length. The abdominal wound was closed with a silk suture, and mice were returned to their cages. Mice were humanely sacrificed on days 1, 3, and 7 (n = 6), and the kidneys were harvested for analysis.
Histological Analysis of Kidney Tissues. IgAN or mouse specimens were fixed in 10% formaldehyde for 24 h. After dehydration, they were embedded in paraffin, and 1.5 μm thick cross-sections were stained with hematoxylin and eosin (HE), Masson's trichrome, and periodic acid-Schiff (PAS). The sections were evaluated from five randomly selected fields by an independent pathologist (magnification of ×100). The extent of fibrosis was scored as 0 (negative), 1 (weak), 2 (medium), or 3 (intense). Each tissue section was observed under a light microscope (Olympus IX71, Tokyo, Japan) at magnifications of ×200 and ×400.

Immunofluorescence Analysis of IgAN Renal Biopsies. A rotary paraffin microtome (Leica RM2255,
Germany) was used to section the samples (1-2 μm). The sections were placed on polylysine-coated slides and incubated in a 60 °C oven. After dewaxing with xylene, rehydrating with 100 (5 min, twice), 95, 85, and 75% (5 min each) ethanol, and washing with PBS, the slides were placed in an EDTA antigen retrieval solution (0.01 M, pH 8.0). The slides were then placed in a microwave oven (850 W; Sharp, Osaka, Japan) until the antigen retrieval solution reached 100 °C for 10 min, cooled to room temperature for 20 min, and washed with PBS for 5 min. The slides were then placed in citrate antigen-repairing solution (0.01 M, pH 6.0) and heated in a high-pressure cooker until steam was observed. The slides were kept inside the cooker for 2 min, cooled to room temperature for 20 min, and washed with PBS for 5 min. Primary antibodies against PDGFRβ (Abcam, Cambridge, MA, USA), CD31 (Abcam, Cambridge, MA, USA), α-SMA (Abcam, Cambridge, MA, USA), and FUT8 (Santa Cruz Biotechnology, CA, USA) were incubated with the slides overnight at 4 °C, and LCA (Vector Labs) was incubated with the slides overnight at 4 °C. Secondary antibodies were incubated with the samples for 1 h at room temperature. Then, sections were mounted with Fluorescent Mounting Media containing 4′,6-diamidino-2-phenylindole (DAPI) (Abcam, Cambridge, MA, USA). Each tissue section was observed under a confocal laser scanning microscope (Leica SP8, Germany) at magnifications of ×400 and ×600, if necessary. Negative controls did not receive the first antibody.
Immunofluorescence Staining of Mouse Kidney Tissues. Freshly harvested mouse kidney tissues were fixed with 4% paraformaldehyde (PFA) for 24 h. Then, tissues were successively dehydrated in 30%, 20%, and 10% sucrose solutions for 1 h each, and embedded in optimal cutting temperature compound (OCT, Tissue Tek, Sakura, Japan). Next, 4 μm cryosections were collected on Superfrost Plus glass slides. Sections were rinsed with PBS and permeabilized with a 1% Triton solution for 5 min. Then, they were blocked with blocking buffer ( Pericyte Isolation and Pericyte-Myofibroblast Transition Model In Vitro. We used a previously described method 16 , with some modifications. Percoll was applied to remove impurities to reduce their impact on FACS analysis of pericytes. The kidney was diced and incubated with liberase (0.5 mg/ml, Roche, Mannheim, Germany) and DNase (100 U/ml, Roche) in Hank's buffered salt solution for 45 min at 37 °C. After centrifugation, cells were resuspended in Hank's buffered salt solution, and filtered (40 μm). Then, a 42% Percoll solution was used to remove impurities, and the cell layer was collected. Pericytes were purified by isolating PDGFR-β+ cells using FACS (FACSAria II, BD Biosciences) or MACS (Miltenyi, Germany), and were then cultured in pericyte medium supplemented with 2% FBS, 1% PGS, 1% P/S (ScienCell, CA, USA) at 37 °C in a 5% CO 2 atmosphere with 90% humidity. The medium was changed every three days. Cells in the first or second passage were used for experiments. Recombinant TGF-β1 was used at a concentration of 5 ng/ml to induce the pericyte-myofibroblast transition.
Immunofluorescence Staining of Pericytes. Cells were fixed with freshly prepared 4% paraformaldehyde for 10 min at room temperature. The cells were then washed three times with PBS. Each cover slip was then incubated in 1% BSA. Primary antibodies against PDGFRβ (Abcam, Cambridge, MA, USA), α-SMA (Abcam, Cambridge, MA, USA), or FUT8 (Santa Cruz Biotechnology, CA, USA) were incubated with the cells overnight at 4 °C, and LCA (Vector Labs, CA, USA) was incubated with the cells for 1 h at room temperature. After a wash with PBS, secondary antibodies conjugated with FITC or Cy3 were applied for 1 h at room temperature in a darkened humidified chamber. Finally, the preparations were washed with PBS, and mounted in fluorescent mounting medium with DAPI (Abcam, Cambridge, MA, USA). Images were acquired using a confocal laser scanning microscope (Leica SP8, Germany) at magnifications of ×200 and ×400. Negative controls did not receive the first antibody.
Inhibition of FUT8 Expression. We constructed the adenovirus carrying the FUT8 shRNA in our previous study 20 . The adenovirus carrying the FUT8 shRNA was administered by tail vein injection of 1 × 10 6 PFUs before anesthesia for UUO surgery. The FUT8 siRNA was designed by a commercial company (RiboBio, 5′-GCUACUGAUGAUCCUACUUdTdT-3′) for FUT8 knockdown in pericytes in vitro. INTERFERin TM (Polyplus-transfection SA, Bioparc, France) was used for transfection. Twenty-four hours after the start of transfection, pericytes were incubated with serum-free medium for 24 h. To confirm the knockdown of FUT8, we extracted RNA from cells and assessed FUT8 expression by reverse-transcriptase polymerase chain reaction (RT-PCR), according to the manufacturer's protocol. Beverly, MA, USA), ERK1/2 (Cell Signaling Technology, MA, USA), p-Smad2/3 (Cell Signaling Technology, MA, USA), and p-ERK1/2 (Cell Signaling Technology, MA, USA) overnight at 4 °C. Membranes were incubated with secondary antibodies for 1 h at room temperature. Blots were washed three times in a 1× PBS-Tween solution, and incubated with ECL reagents (Amersham, Pittsburgh, PA, USA) for 1 min. Proteins were visualized with super RX-N film (Fujifilm Corporation, Tokyo, Japan). A GAPDH antibody was used as a loading control to normalize the protein levels detected.
Immunoprecipitation. Cell lysates were centrifuged (12,000 rpm for 20 min at 4 °C), and the supernatant was collected and precleared by Protein G PLUS-Agarose (Santa Cruz Biotechnology, CA, USA). Cell lysates (500 μg) were then incubated with 2 μg of anti-TGFβRII or anti-ALK5 antibody for 2 h at 4 °C on a rocker platform (30 rocks/min). Protein-antibody complexes were collected with 20 μl of Protein G PLUS-Agarose on a rocker platform (30 rocks/min) overnight at 4 °C. The immunoprecipitates were washed three times with lysis buffer. Equal amounts (10 μg/lane) of proteins were subjected to 12% SDS-PAGE for lectin blotting (described below). Negative controls did not receive the first antibody.

Statistical Analysis.
Results are presented as mean values ± standard deviations. Multiple comparisons of parametric data were performed using one-way analysis of variance (ANOVA). Nonparametric data were compared with the Mann-Whitney U-test to identify differences between groups. Pearson's correlation analysis was calculated to determine the associations. A value of P < 0.05 was considered to indicate statistical significance. All statistical analyses were performed using the statistical package SPSS (version 21.0, IBM, Inc., USA).