Epac ameliorates tubulointerstitial inflammation in diabetic nephropathy via C/EBPβ/SOCS3/STAT3 pathway CURRENT

Background: Tubulointerstitial inflammation plays a pivotal role in the progression of diabetic nephrology (DN), and tubular cells act as the driving force in the inflammation cascade. The aim of the study is to investigate whether activation of Epac alleviates tubulointerstitial inflammation through SOCS3/JAK/STAT3 pathway under DN. Methods: The kidney morphological changes and the level of oxidative stress were observed in db/db mice treated with Epac agonist (8-pCPT-2’-O-Me-cAMP). The expression of inflammatory cytokines, fibrosis markers, C/EBPβ, SOCS3 and p-STAT3 in renal tissues were assessed. In in vitro study, the changes of above genes and proteins were also detected in human proximal tubular cell line (HK-2) incubated with high glucose (HG) with and without Epac agonist. Overexpression of SOCS3 plasmid, SOCS3 siRNA and C/EBP-β siRNA were employed and the translocation of C/EBP-β and p-STAT3 were observed by immunofluorescence. Macrophage migration assay were detected in Transwell system. Results: In db/db mice, we observed that the expression of MCP-1 was up-regulated in renal tubular cells, which concurrent with increased influx of macrophages, cytokines production (MCP-1, TNF-α, and IL-6), and tubulointerstitial fibrosis. Moreover, we also found that these changes were accompanied by reduced C/EBP-β expression and increased SOCS3 and phosphorylated STAT3 (p-STAT3) expression. However, after treated db/db mice with 8-O-cAMP, an Epac activator, the expression of SOCS3 further up-regulated while other protein expression profile was partially reversed. In vitro , high glucose (30mM, HG) treatment down-regulated C/EBP-β expression while increased SOCS3, p-STAT3 and MCP-1 expression in HK-2 cells; Activation of Epac by 8-O-cAMP further increased the gene and protein

DN has not been traditionally considered an inflammatory disease until recently. Emerging data have demonstrated that tubulointerstitial inflammation is strongly correlated with the development and progression of DN (2,3). Macrophage infiltration in the kidney tissues has been postulated to be a prominent pathological feature and key inflammatory mediator during the progression of DN. Studies from human biopsies and animal models have identified the presence of macrophages in diabetic kidneys, even in the early stage of the disease (4). Moreover, the degree of tubulointerstitial macrophage infiltration is closely associated with the renal function decline and interstitial fibrosis (5,6). Depletion of macrophage by silencing macrophage scavenger receptor-A expression in animal models alleviated STZ-induced diabetic renal damage (7). Therefore, diminishing macrophage recruitment and activation may be a novel therapeutic approach for the treatment of DN. 4 Notably, it was found that tubular damage may be the primary event in DN and may be strongly associated with macrophage infiltration (8,9). High glucose induced renal tubular epithelial cells secreting many inflammatory cytokines, including MCP-1, IL-6 and TNFα (10)(11)(12). Among them, MCP-1 is one of the well-documented and important chemotactic factors (13,14). Recently, Kuehn and his colleagues have found that the integrity of cilia of tubular cells modulated macrophages recruitment via an MCP-1 dependent way (15).
Downregulated MCP-1 expression alleviated STZ-induced diabetic renal damage and macrophage infiltration (16). Therefore, MCP-1 signaling cascade plays a key role in the cross-talk between injured tubular cells and infiltrating macrophages in the progression of DN.
Exchange protein activated by cAMP (Epac) is a cyclic AMP-regulated guanine nucleotide exchange factor (GEFs) that activates the small GTPase Rap(17, 18). Epac1 and Epac2 are two isoforms of Epac and they have different tissue distribution and expression levels (19). Generally, Epac1 is higher expression in the kidney than Epac2, especially in renal tubular epithelial cells (20,21). Studies have shown that activation of Epac might exert a renoprotective effect in various CKD model. For instance, activation of Epac decreased the Ang II-induced ROS generation and inflammatory cytokines (IL-1β, IL-6, IL-8, and TNF-α) production in renal tubular cells (22). Subsequent study further indicated that Ang IIinduced expression of inflammatory cytokine were regulated by Epac1-Rap1a-NHE3 pathway in tubular cells (23), indicating that Epac may play an inhibitory effector in response to renal inflammatory. However, whether Epac is involved in the antiinflammatory signaling by targeting MCP-1 in the proximal renal cells expose to hyperglycemia remains unclear.
Previous studies have suggested that hyperglycemic stimulus enhanced the activation of the JAK/STAT signaling cascade in renal tubular cells, which contributed to renal tubular 5 damage in DKD (24,25). Overexpression of SOCS3 by renal delivery adenovirus, an important endogenous suppressor of JAK/STAT3 pathway (26), significantly reversed STAT3 activation-induced MCP-1 upregulation and macrophage infiltration in DN (10), suggesting a renoprotective role of SOCS3. Therefore, the manipulation of SOCS3 expression would be imperative and is a therapeutic target for DN. Additionally, it has been reported that Epac1 induces SOCS3 gene expression in various types of cells, including cardiocyte (27), vascular endothelial cells (28) and retinal ganglion cells (29). Based on these perspectives, we hypothesized that activation of Epac may protect against the inflammation injury by targeting SOCS3 gene expression in the context of hyperglycemia.
In the present study, we investigated the role of pharmacological activation of Epac by 8-pCPT-2'-O-Me-cAMP (8-O-cAMP) in both in vitro and in vivo models of DN. We found that activation of Epac decreased macrophage infiltration and subsequent inflammatory cytokines production in the kidney of db/db mice, and significantly ameliorated renal tubulointerstitial fibrosis. Mechanistically, selective activation of Epac reduced HG-induced STAT3 activation and MCP-1 expression via increased SOCS3 gene expression in proximal tubular cells. These results revealed that the beneficial effect of activation of Epac by attenuating macrophage-mediated renal injury in DN.

Animal experimental design
The Animal Care and Use Committee of Second Xiangya Hospital of Central South University approved all animal procedures. C57BL/KS db/db and C57BL/KS db/m mice were used for the animal experiments. These mice were purchased from the Aier Matt Experimental Animal Company (Suzhou, China) and then housed in Second Xiangya Hospital Animal Center, Central South University. All mice access to standard chow and filtered tap water ad libitum. At 8 weeks of age, they were organized into the following 6 three groups for animal experiments: a db/m group (control, n=5), a db/db group (n=5) and a db/db group receiving an intraperitoneal injection (i.p.) of 8-pCPT-2'-O-Me-cAMP (Sigma, 1.47mg/kg, once every other day for 8 weeks, n=5). Every two weekly body weight and fasting blood glucose were monitored throughout the study and at 16 weeks of age, all the mice were euthanized under intraperitoneal pentobarbital sodium anesthesia (50 mg/kg body weight).

Assessment of physiological features and renal function
Body weights and blood glucose levels were measured every two weeks. Blood glucose levels were detected by tail vein blood samples using a blood glucose monitor (Boehringer Mannheim, Mannheim, Germany). Urine albumin concentrations were measured using a mouse urine albumin ELISA kit (Bethyl Laboratories, USA), and serum creatinine levels were tested by HPLC, liquid chromatography/tandem mass spectrometry.

Immunohistochemistry
Four-micron thick paraffin-embedded kidney sections were utilized for immunohistochemistry (IHC) studies as previously described (31). In brief, the sections were dewaxed, rehydrated, and incubated with primary antibodies overnight at 4 °C and later incubated with horseradish peroxidase-conjugated secondary antibody and diaminobenzidine (DAB) substrate. After counterstaining with hematoxylin and dehydration, the sections were mounted and analyzed with a Nikon microscope.

Measurement of superoxide generation and immunofluorescence assay
Dihydroethidium (1μM, DHE, Sigma-Aldrich) were used to assess intracellular ROS production in 4-mm-thick kidney cryostat sections as described previously (32). The mean fluorescence intensity (MFI) of DHE was calculated by ImageJ. For immunofluorescence assay, the cells or frozen sections were fixed with 4% paraformaldehyde for 10 min at room temperature. Then, the sections were blocked with 0.1%TritonX-100 and 5%BSA mixture for 60 min at room temperature and then incubated with primary antibody overnight at 4°C. Later incubated secondary antibodies conjugated with Alexa Fluor 488 (green) or 594 (red) for 1 hour. These cells or slides were counterstained with 4′,6diamidino-2 phenylindole (DAPI) and their fluorescent signals were visualized using an LSM 780 META laser scanning microscope (Zeiss, Thornwood, NY). The MFI of p-STAT3 and C/EBP-β was calculated by ImageJ.

Cell culture studies
The human proximal tubular cell line HK-2 (ATCC, Rockville, USA) was used for our in vitro studies. Time-dependent experiments were performed using 30 mM D-glucose for 0-48 h.
The HK-2 cells were pretreated with 8-pCPT-2'-O-Me-cAMP for 5 hours before exposure to 30mM glucose. C/EBP-β or SOCS3 siRNA were transfected into the HK-2 cells using Lipofectamine 2000 reagent (Life Technologies, USA) for in vitro studies.

Real -Time PCR analysis
Total RNA from renal sections and HK-2 cells was prepared using TRIzol (Takara). Firststrand cDNA was prepared by two-step RT-qPCR (Takara). Relative gene expression was determined using SYBR Green quantitative real-time PCR assays on Light Cycler 96 System (Roche). The cycling conditions were as follows: denaturing at 50 °C for 2 min, 95 °C for 10 min. Preincubation at 95 °C for 30 sec, then 40 cycles of 95 °C for 5 sec, 60 °C for 30 8 sec and 97 °C for 1 sec. All reactions were carried out in triplicate with a control.
Expression was calculated using the 2 − ΔΔCt method and normalized to ACTB.

Western blot assay
The protein extraction of kidney cortex from db/db mice, HK-2 cells and controls were lysed in radio immunoprecipitation assay (RIPA) buffer containing protease inhibitors and phosphatase inhibitors. BCA Protein Assay kit (CoWin Biosciences, China) was used to determine protein concentration. Proteins were separated using 8% and 12% SDS-PAGE and transferred to PVDF membranes, which were probed with primary antibodies against C/EBP-β, MCP-1, TNF-α (Abcam, USA), p-STAT3 (Cell Signaling Technology, USA), STAT3, and SOCS3 (Proteintech, CHINA), respectively. β-Actin (Proteintech, CHINA) was used as a loading control. Following incubation with an appropriate secondary antibody, the protein expression was detected by western blot analysis, as described previously (31).

Macrophage migration assay
HK-2 cells pretreated with or without 8-O-cAMP were incubated with 30 mM D-glucose, and HK-2 cells under basal conditions served as a control. The cell supernatants were collected after 24 hours. THP-1 cells cultured in the upper chambers of Transwell plates treated with 100ng/ml PMA for 48h, and then cultured in fresh complete medium that in absence of PMA for 24h. While the cell supernatants derived from HK-2 cells exposed to HG were added to the lower chambers. After 24h incubation, a cotton swab was used to remove the cells on the upper surface of the membrane; Cells that migrated to the bottom surface of the membrane were fixed and stained with DAPI, and then observed under a Nikon microscope.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 7.0 software (GraphPad, San Diego, CA, USA). Statistical different between two groups were analyzed by unpaired t test. One-way analysis of variance (ANOVA) test was used for multiple independent sample analysis. Comparisons between groups were performed with Tukey's test post hoc analysis. Values are shown as the Mean ± SD; p < 0.05 was considered to be statistically significant. HE and PAS staining revealed mesangial expansion and tubular atrophy in the kidneys of db/db mice compared to db/m mice (Fig. 1e); Masson's trichrome staining showed increased interstitial fibrosis in the kidney of db/db mice (Fig. 1e). However, Activation of Epac with 8-O-cAMP dramatically ameliorated these abnormal changes (Fig. 1e).

Activation of Epac
Quantitative analysis of tubular and glomerular damage score confirmed this tendency ( Fig. 1f, g). These data indicated that Epac activator 8-O-cAMP may exert a therapeutic role in the diabetic kidney injury.
Activation of Epac alleviates the inflammation response, oxidative stress, and interstitial fibrosis in the kidney of db/db mice.
Inflammation and oxidative stress contribute critically to the initiation and extension of 10 DN. We therefore evaluated renal inflammatory cytokines and reactive oxidative stress (ROS) production the kidneys of db/db mice receiving 8-O-cAMP treatment. By IHC and Western blot assay, TNF-α, IL-6 expression was found to be up-regulated in the tubules of db/db mice compared to those of controls ( Fig. 2a (e, f vs b, a)); DHE staining demonstrated increased ROS production in the tubular cells of the kidneys of db/db mice ( Fig. 2b (b vs a)). Finally, the expression of FN and Col-1 were also detected, as one can see, a notable increase in FN and Col-1 expression in the tubulointerstitium of db/db mice ( Fig. 2a (g, h vs c, d)) and similar result was demonstrated by Western blot assays (Fig.   2c). Of note, these above abnormal changes were reversed after treated with 8-O-cAMP

Activation of Epac attenuates renal macrophage infiltration in db/db mice.
To further understand the role of Epac agonist 8-O-cAMP in the inflammation response in the kidney of db/db mice, we detect the expression of macrophage chemotactic factor MCP-1 and macrophage accumulation in the kidney. IHC assay showed that the expression of MCP-1 significantly increased in the tubules of db/db mice compared to the db/m mice ( Fig. 3a (b vs a)). Similar results of MCP-1 protein and gene expression were demonstrated by western blot assays (Fig. 3b) and qPCR (Fig. 3c), respectively. Moreover, immunofluorescence staining results showed that the expression of F4/80 was significantly up-regulated in in renal interstitium of db/db mice, indicating the increased macrophage infiltration, which were significantly inhibited by 8-O-cAMP treatment ( Fig. 3a and a2).  Fig.3a and a3) in db/db mice. These above results indicated that activation of Epac may exert anti-inflammatory effects by reducing macrophage accumulation and activation.
The transcription factor, C/EBP-β, has been reported to bind directly to the SOCS3 promoter region (33). Here, we found that the expression of C/EBP-β in db/db mice was significantly reduced compared with db/m mice (Fig. 4a (b vs a)). Interestingly, the expression C/EBP-β was significantly increased by the treatment of 8-O-cAMP (Fig. 4a (c) ).
Considering that the MCP-1 is a classical STAT-responsive inflammatory gene and SOCS3 is known as a negative regulator of STAT3 signaling pathway. Next, we examined the effect of 8-O-cAMP on the expression of SOCS3 and p-STAT (Tyr705). Immunohistochemistry revealed a lower level of SOCS3 and p-STAT3 expression in db/m mice (Fig. 4a (d, g)), and broad distribution of them in db/db mice (Fig. 4a (e, h)). Moreover, activation of Epac with 8-O-cAMP further up-regulated SOCS3 expression (Fig. 4a (f)), while inhibited p-STAT3 expression in db/db mice (Fig. 4a (i)). Similar results were observed with respect to expression of C/EBP-β, SOCS3 and p-STAT3, as detected by western blot assays (Fig. 4c).
These data suggested that activation of Epac might increase SOCS3 expression and inhibit STAT3/MCP-1 pathway by recovering C/EBP-β expression.
Sequential changes in expression of SOCS3/STAT3/MCP-1 in HK-2 cells exposed to high glucose.
Western blot assays and real-time PCR demonstrated that HG (30 mM, D-glucose)-induced the mRNA and protein expression levels of SOCS3 was highest at 8h and then gradually decreased to normal level ( Fig. 5a and b). However, the protein expression of p-STAT3 was increased in a time-dependent manner, peaking at 24-48h following treatment with HG ( Fig. 5a1).In line with this, a time-dependent increase in the mRNA and protein expression of MCP-1 was observed and was maximal at 48h (Fig. 5a and c).

Activation of Epac reduces MCP-1 expression and decreases macrophage recruitment via SOCS3.
In order to confirm the effect of Epac on p-STAT expression in HK-2 cells expose to HG.
Immunofluorescence assay revealed that the p-STAT3 expression was quite low in HK-2 cells under basal conditions (5.6 mM D-glucose) and HG treatment resulted in the activation of STAT3 (Fig. 6a (a, b) and b), while it was largely reduced by the treatment with 8-O-cAMP (Fig. 6a (c) and b).
To further determine whether the inhibitory effect of Epac on the STAT3 activation was relied on SOCS3, we transfected HK-2 cells with SOCS3 siRNA contaminant with 8-O-cAMP treatment. As shown in Figure 6B, western blot results showed the inhibitory effect of Epac activation on p-STAT3 and MCP-1 expression was abolished by knockdown SOCS3 expression in HK-2 cells exposed to HG ambience. Likewise, we also observed that HGinduced increased p-STAT3 and MCP-1 expression in HK-2 cells were significantly reduced by SOCS3 overexpression, suggesting modulation of SOCS3 expression is sufficient to mimic the activation of Epac. Real-time PCR demonstrated the mRNA expression of SOCS3 and MCP-1 was in accordance with the protein expression. (Fig. 6c and d). These results suggested that Epac-induced SOCS3 expression reduced MCP-1 expression by preventing the STAT phosphorylation in HK-2 cells under HG condition. Furthermore, macrophage migration assay was used to demonstrate activation of Epacreduced MCP-1 expression can effectively decrease the recruitment of macrophages. The number of THP-1 cells migrated from the upper side to the bottom side of the membrane were increased by adding the hyperglycemia-stimulated HK-2 cells supernatant (Fig. 6f (b)).While supernatant from HK-2 cells treated with 8-O-cAMP under HG condition, the 13 number of migrated-macrophages were significantly decreased (Fig. 6f (c)).

Activation of Epac increases SOCS3 expression via C/EBP-β.
By IF staining, C/EBP-β was found to be highly expressed in the nuclear in HK-2 cells under basal conditions (Fig. 7a (a)). The HG exposure reduced nuclear fluorescence of C/EBP-β ( Fig. 7a (b)), which was restored by 8-O-cAMP treatment (Fig. 7a (c)). The transcription factor, C/EBP-β, has been reported to bind directly to the SOCS3 promoter region(33), we next determined whether Epac modulates the expression of SOCS3 by C/EBP-β. HK-2 cells were transfected with C/EBP-β siRNA before HG treatment. Western blot analysis showed that the expression of C/EBP-β significantly decreased in HK-2 cells exposure to HG, and activation of Epac by 8-O-cAMP restored C/EBP-β and SOCS3 expression, while this effect was neutralized by C/EBP-β siRNA transfection (Fig. 7b). Similar results of SOCS3 mRNA expression were demonstrated using real-time PCR (Fig. 7c). These data confirmed the role of C/EBP-β in controlling the Epac-dependent induction of the SOCS3 gene in HK-2 cells.

Discussion
In the present study, we demonstrated that Epac activation by 8-O-cAMP could alleviate inflammatory damage by down-regulated MCP-1 expression in kidney tissues of diabetic mice. We also observed that SOCS3 contributed to this process by suppression the phosphorylation of STAT3 in vivo and vitro studies. At the molecular level, activation of Emerging evidence has identified that inflammation plays a central role in the pathogenesis of DN, and macrophage infiltration in the interstitial compartment is closely related to the progression of DN (34,35). The monocyte chemokine, MCP-1, was increased in the glomerular and tubular epithelium, and as a major contributor to macrophage infiltration in the kidney tissues of DN is well established (16,34,36,37). MCP-1, a member of the CC chemokine family acting through its receptor CCR2 (38), is a critical factor in recruiting monocytes/macrophages into inflammation site (39). The expression of MCP-1 in tubular cells and urinary excretion of the protein were both upregulated in human and experimental diabetes, which coincided with macrophage infiltration in both glomeruli and interstitial and correlated with levels of albuminuria and renal function decline (34,37,39,40). Pharmacologic inhibition of MCP-1 or CCR2 effectively reduced the recruitment of macrophage into the interstitial compartment and improved tubulointerstitial damage (41,42).In diabetic animal models treated with STZ, deficiency of MCP-1 decreased the number of macrophage accumulation and prevented early diabetic renal injury (34,40). In agreement with these previous studies, we observed that upregulated renal MCP-1 expression was accompanied by increased macrophage infiltration in tubular interstitial in diabetic mice. However, administration of Epac activator 8-O-cAMP dramatically reversed these changes ( Fig. 1 and Fig. 3). These data indicated that Epac reduced macrophage-mediated renal injury by inhibiting the expression of MCP-1 in DN. But the specific mechanism by which Epac reduced MCP-1 expression is still not known. JAK/STAT pathway has been shown to participate in various inflammation response, including in DN (10). Phosphorylated STAT3 translocates to the nucleus and initiates inflammatory-related gene transcription, such as CCL2 (the gene name of MCP-1) (43).
Hyperglycemia induces the activation of JAK/STAT3 signaling in glomerular mesangial cells, resulting in cell proliferation and synthesis of extracellular matrix molecules (25).
Giving STAT3 inhibitor dramatically decreased renal macrophage infiltration and attenuated inflammation and fibrosis in diabetic rats (44). Moreover, Lu's group constructed streptozocin-induced STAT3 knockdown diabetic mice and found that STAT3 deficiency dramatically reduced the expression of inflammation markers, like IL-6, MCP-1 (45). Indicating that activation of STAT3 may play a crucial role in the pathogenesis of DN. However, studies of the relationship of STAT3 and DN are mostly focus on mesangial cells, the role of phosphorylated STAT3 in the proximal tubular cells exposed to high glucose is under elucidated yet. In agreement with previous studies in renal disease, we observed that the expression of p-STAT3 in the proximal tubule was up-regulated in diabetic mice (Fig. 4). Activation of Epac with 8-O-cAMP inhibited p-STAT3 expression which dramatically reduced the expression of MCP-1( Fig. 3 and Fig. 4). Thus, suggesting that Epac activation may inhibit MCP-1 expression by preventing STAT3 phosphorylation.
We also investigated the molecular mechanisms by which Epac regulated the STAT3 inactivation under hyperglycemic conditions. SOCS proteins are natural inhibitor family for JAK/STAT signaling pathway and have been demonstrated participating in the regulation of inflammatory signaling (26,46). Overexpression of SOCS1 or SOCS3 has been shown effectively prevented JAK/STAT activation in cultured renal cells and subsequent inflammatory response (10) (47). Besides, it has been reported that activation of Epac1 induces SOCS3 gene expression in various types of cells, i.e. cardiocyte (27), vascular endothelial cell (28) and retinal ganglion cell (29). In this study, we observed that Epac activation significantly up-regulated the expression of SOCS3 while inhibited the expression of p-STAT3 and MCP-1 in kidney tissues (Fig. 4) and in HK-2 cells (Fig. 6); These effects were imitated by transfection with SOCS3 overexpression plasmid in HK-2 cells exposed to HG; More importantly, knockdown of SOCS3 expression could reverse the inhibitory effect of Epac on STAT3 activation and MCP-1 expression (Fig. 6). These data indicated that Epac attenuated renal inflammation may be associated with the suppression of STAT3 activation at least by SOCS3. Nonetheless, we cannot exclude the other mediators for Epac targeting STAT3 pathway. For example, in our study, we also found that activation of Epac ameliorated renal tubular oxidative stress in db/db mice (Fig. 2).
Previous studies have demonstrated that hyperglycemia-induced the production of ROS could also activate JAK/STAT pathway (48). Therefore, the suppression of JAK/STAT activation by Epac might by inhibiting ROS production in DN, but it needs further investigation.
Next, we attested the role of transcription factor C/EBP-β in Epac-induced SOCS3 expression in HK-2 cells. Since our previous study found that Rap1b, a downstream molecule of Epac, could modulate C/EBP-β transcription activity in DN, and studies in other cell types also shown that Epac regulated SOCS3 gene expression in a C/EBP-β dependent manner (33,43). As expected, in this study, activation of Epac by 8-O-cAMP could restore C/EBP-β expression and promote its translocation to the nuclear in HK-2 cells (Fig. 7). A loss-of function assay on C/EBP-β shown that knockdown C/EBP-β expression would inhibit the Epac activation-induced SOCS3 gene expression. Taken together, these data suggested C/EBP-β as Epac-activated transcription factors that mediate the induction of the SOCS-3 gene in tubular cells. However, future study like luciferase reporter assay would be taken to the evaluate whether stimulation of Epac in tubular cells leads to enhanced binding of C/EBP-β to the SOCS-3 promoter.

Conclusion
In summary, we use in vivo and in vitro diabetic models to demonstrate the beneficial effects of activation of Epac on macrophage-mediated inflammatory injury and the mechanisms underlying its effects. Epac regulates C/EBP-β-mediated SOCS3 transcription and inhibits the expression of MCP-1, thereby improving renal macrophage infiltration and inflammatory response, ultimately alleviating interstitial fibrosis (Fig. 8). This finding suggests that activation of Epac may provide a new therapeutic target for preventing renal inflammation injury in DN.