Interferon-γ enhances the therapeutic effect of mesenchymal stem cells on experimental renal fibrosis

Mesenchymal stem cells (MSCs) administered for therapeutic purposes can be activated by interferon-γ (IFN-γ) secreted from natural killer cells in injured tissues and exert anti-inflammatory effects. These processes require a substantial period of time, leading to a delayed onset of MSCs’ therapeutic effects. In this study, we investigated whether pretreatment with IFN-γ could potentiate the anti-fibrotic ability of MSCs in rats with ischemia–reperfusion injury (IRI) and unilateral ureter obstruction. Administration of MSCs treated with IFN-γ strongly reduced infiltration of inflammatory cells and ameliorated interstitial fibrosis compared with control MSCs without IFN-γ treatment. In addition, conditioned medium obtained from IFN-γ-treated MSCs decreased fibrotic changes in cultured cells induced by transforming growth factor-β1 more efficiently than that from control MSCs. Most notably, secretion of prostaglandin E2 from MSCs was significantly increased by treatment with IFN-γ. Increased prostaglandin E2 in conditioned medium obtained from IFN-γ-treated MSCs induced polarization of immunosuppressive CD163 and CD206-positive macrophages. In addition, knockdown of prostaglandin E synthase weakened the anti-fibrotic effects of MSCs treated with IFN-γ in IRI rats, suggesting the involvement of prostaglandin E2 in the beneficial effects of IFN-γ. Administration of MSCs treated with IFN-γ might represent a promising therapy to prevent the progression of renal fibrosis.

IL-18, which are known as DAMPs, are released from damaged tissues. Released IL-18, which is also induced by HMGB1 8 , contributes to secretion of IFN-γ 19 . To confirm expression of HMGB1, IL-18, and IFN-γ induced by IRI, rats were sacrificed to evaluate their expression in the kidney at 1 day (Post IRI Day 1) and 7 days (Post IRI Day 7) after the IRI procedure. As shown in a previous study 31 , the protein level of HMGB1 was increased by the IRI procedure (Fig. 1a). Immunostaining revealed that IL-18 and IFN-γ-positive areas were increased strongly in the kidney at 1 day after IRI. However, their increase was attenuated at 7 days after IRI (Fig. 1b,c). These changes are similar to those seen in a previous study 32 . Furthermore, because IL-18 promotes IFN-γ secretion from natural killer cells, IFN-γ expression might be associated with IL-18 secretion. MSCs are activated by released IFN-γ and exert immunosuppressive effects [22][23][24][25][26] . Therefore, administration of MSCs cultured in IFN-γcontaining medium might have beneficial effects on preventing the progression of renal fibrosis.

IFN-γ enhances the ability of MSCs to attenuate fibrosis induced by IRI.
To evaluate anti-fibrotic effects of MSCs, we injected PBS, rat MSCs (rMSCs) treated with IFN-γ (IFN-γ rMSCs) or untreated rMSCs (control rMSCs) into the abdominal aorta of rats after ischemic reperfusion. Twenty-one days later, the rats were sacrificed, and injured kidneys were collected to evaluate the degree of fibrosis by western blotting. The protein levels of α-smooth muscle actin (α-SMA) and transforming growth factor-β1 (TGF-β1), markers for drivers of fibrosis, were increased in the kidney of PBS-injected rats and their levels were suppressed by injection of control rMSCs (Fig. 2a). Furthermore, injection of IFN-γ rMSCs decreased IRI-induced fibrotic changes more significantly than that of control rMSCs (Fig. 2a). Immunostaining of α-SMA, collagen type I (Col-I), and collagen type III (Col-III) (extracellular matrix proteins) was also performed to assess renal fibrosis. α-SMA, Col-I, and Col-III-positive areas were increased in the PBS group. Similar to the results from western blotting, administration of IFN-γ rMSCs reduced α-SMA, Col-I, and Col-III-positive areas more strongly compared with that of control rMSCs (Fig. 2b,c). These results suggest that IFN-γ-preconditioned rMSCs have a strong anti-fibrotic effect.

IFN-γ enhances the ability of MSCs to inhibit inflammation induced by IRI. As shown above,
IFN-γ rMSCs strongly suppressed renal fibrosis in IRI rats. We next examined the anti-inflammatory effects of IFN-γ rMSCs in the acute phase of renal injury in IRI rats. rMSCs with or without IFN-γ treatment were injected into rats after IRI induction, and 7 days later, the rats were sacrificed to investigate infiltration of inflammation cells into the kidney. Immunostaining showed that infiltration of CD3 (T cell marker)-and CD68 (macrophage marker)-positive cells was markedly increased in the PBS group (Fig. 3a,b). Injection of control rMSCs suppressed the infiltration of these cells, and the suppression was more significant in the IFN-γ rMSCs group (Fig. 3a,b). However, the infiltration of CD163 and CD206 (immunosuppressive macrophage markers)-positive cells was increased by injection of control rMSCs, and injection of IFN-γ rMSCs resulted in a more significant increase in CD163 and CD206-positive cells (Fig. 3a,b), indicating that IFN-γ-preconditioned rMSCs have a stronger anti-inflammatory effect.
Administration of IFN-γ-preconditioned MSCs strongly attenuates UUO-induced fibrosis. A UUO model was used to confirm the anti-fibrotic effect of IFN-γ rMSCs. Four days after UUO, rats were injected with PBS, control rMSCs, or IFN-γ rMSCs through the tail vein. Seven days after injection, the rats were sacrificed and evaluated for renal fibrosis. Western blotting showed that expression of α-SMA was induced by UUO in the PBS group. The protein levels of α-SMA were suppressed by injection of control rMSCs, and the suppression of α-SMA levels by IFN-γ rMSCs was more significant than that by control rMSCs (Fig. 4a). Immunostaining also revealed increases in α-SMA, Col-I, and Col-III-positive areas in the PBS group (Fig. 4b,c). The α-SMA and Col-III-positive areas were suppressed by injection of control rMSCs. Administration of control rMSCs also reduced Col-I-positive area, but it did not reach significance (p = 0.08). Stronger suppressive effects on fibrosis were observed in rats injected with IFN-γ rMSCs compared with control rMSCs (Fig. 4b,c).

IFN-γ enhances secretion of PGE2 from MSCs. It has been reported that secretion of PGE2 from
MSCs exerts anti-inflammatory effects 24 . Thus, the amount of PGE2 contained in CM from IFN-γ hMSCs and control hMSCs was measured by an enzyme-linked immunosorbent assay (ELISA). The amount of PGE2 was www.nature.com/scientificreports/ significantly elevated in CM from control hMSCs. Furthermore, the levels of secreted PGE2 were more markedly increased in CM from IFN-γ hMSCs (Fig. 5c).  www.nature.com/scientificreports/ The samples derive from the same experiment and that gels were processed in parallel. Data are presented as the mean ± SD. *p < 0.05, **p < 0.01. www.nature.com/scientificreports/ with phorbol 12-myristate 13-acetate (PMA) to induce differentiation into macrophages. Then, the medium was replaced with 0.1% fetal bovine serum (FBS)-containing medium or CM from either control hMSCs or IFN-γ hMSCs. Forty-eight hours after medium replacement, the cells were collected and expression of CD163, CD206 and CD68 was evaluated. Although there was no significant difference in the protein levels of CD68, those of CD163 and CD206 were upregulated by CM from control hMSCs (Fig. 5d). Furthermore, CM from IFN-γ hMSCs resulted in stronger induction of CD163 and CD206 compared with that from hMSCs, indicating increased infiltration of CD163 and CD206-positive immunosuppressive macrophages.

Scientific Reports
Knockdown of prostaglandin E synthase weakens the ability of IFN-γ-preconditioned MSCs to change the phenotype of macrophages to immunosuppressive M2. We hypothesized that the potent anti-fibrotic and anti-inflammatory effects of IFN-γ MSCs were due to increased secretion of PGE2. First, we investigated whether prostaglandin E synthase (PTGES) siRNA affected the anti-fibrotic ability of IFN-γ hMSCs using HK2 cells and CM. hMSCs were transfected with PTGES siRNA or negative control siRNA (NC siRNA) for 12 h. After transfection, hMSCs were stimulated by IFN-γ for 24 h. IFN-γ hMSCs transfected with PTGES siRNA showed significant decreases in PTGES mRNA ( Supplementary Fig. S1a). Although the protein levels of α-SMA in HK-2 cells were decreased by incubation in CM from IFN-γ hMSCs transfected with NC siRNA, there was no significant difference between NC siRNA and PTGES siRNA groups ( Supplementary  Fig. S1b). Next, to identify the effects of PTGES siRNA on the phenotypic change of macrophages, THP-1 monocytes were stimulated with PMA. After 48 h, the medium was replaced with 0.1% FBS-containing medium or CM from IFN-γ hMSCs transfected with NC siRNA or PTGES siRNA. After the medium replacement, the cells were collected to evaluate expression of CD163 and CD68. The protein levels of CD163 were increased by CM from IFN-γ hMSCs transfected with NC siRNA (Fig. 6a). However, the upregulation of CD163 was suppressed by CM from IFN-γ hMSCs transfected with PTGES siRNA.

Discussion
Here, we provide the first evidence that administration of MSCs treated with IFN-γ strongly ameliorates renal fibrosis and inflammation in rat IRI and UUO models compared with that of untreated MSCs. Furthermore, CM from MSCs stimulated by IFN-γ inhibits TGF-β/Smad signaling in vitro. IFN-γ stimulation promotes the secretion of PGE2 from MSCs, and increased PGE2 induces polarization of immunosuppressive CD163 and CD206positive macrophages. Knockdown of PTGES attenuates the anti-fibrotic effect of IFN-γ-treated MSCs in IRI rats. These findings suggest that MSCs treated with IFN-γ have a great potential to suppress renal fibrosis (Fig. 8a). DAMPs, such as HMGB1 and IL-18, are involved in the induction and repression of inflammation [6][7][8] . HMGB1 released from damaged tissues cooperates with cytokines, such as those secreted from macrophages, to induce inflammation [33][34][35] . HMGB1 is also involved in inducing IL-18 production 8 . Induced IL-18 promotes secretion of IFN-γ from natural killer cells 21 . When MSCs cultured under normal conditions are administered to renal injured rats, MSCs can be activated by IFN-γ secreted from natural killer cells in injured tissues and suppress renal inflammation. These processes might require several days, leading to a delay of the onset of MSCs' immunosuppressive effects. Because anti-inflammation and anti-fibrotic effects of MSCs treated with IFN-γ were enhanced in advance, their strong therapeutic effects appear to be exerted immediately after administration (Fig. 8b).
PGE2 promotes polarization of immunosuppressive M2 macrophages through induction of the cyclic adenosine monophosphate pathway 27 . In this study, MSCs treated with IFN-γ produced increased amounts of PGE2, leading to more pronounced polarization of M2 macrophages. Therefore, we speculated that PGE2 might contribute to the immunosuppressive action of MSCs in vivo. In fact, we found that knockdown of PTGES weakened the anti-fibrotic effect of MSCs treated with IFN-γ in IRI rats. Thus, the upregulation of PGE2 in MSCs plays an important role in anti-inflammation and anti-fibrotic effects of MSCs treated with IFN-γ.
The TGF-β/Smad signaling pathway is considered as one of the processes that cause renal fibrosis. TGF-β is a soluble factor that binds to a cell surface receptor, and phosphorylates and activates Smad2. Phosphorylated Smad2 induces expression of the αSMA gene and promotes fibrosis 36 . The damaged tissues produce TGF-β that induces renal fibrosis via the TGF-β/Smad signaling pathway. Whereas, TGF-β exerts anti-inflammatory effects, www.nature.com/scientificreports/ and TGF-β blockade might lead to autoimmune disease 37 . Thus, inhibition of the TGF-β/Smad signaling pathway, but not TGF-β itself, might be a key mechanism by which MSCs treated with IFN-γ suppress renal fibrosis. Therefore, in this study, we investigated whether CM from IFN-γ-preconditioned MSCs strongly inhibits the induction of p-Smad2 and α-SMA by TGF-β. Western blotting showed that CM from IFN-γ-preconditioned MSCs significantly decreased TGF-β-induced fibrotic changes compared with that from untreated MSCs. These results suggest that IFN-γ-preconditioned MSCs exert strong anti-fibrotic effects through direct inhibition of the TGF-β/Smad signaling pathway. We show that renal fibrosis and inflammation are strongly attenuated by injection of IFN-γ-treated MSCs in IRI and UUO rats. The anti-inflammation and anti-fibrotic effects of MSCs treated with IFN-γ are due to direct inhibition of TGF-β/Smad signaling and enhancement of PGE2 secretion, which polarization of immunosuppressive M2 macrophages. As a result, administration of MSCs treated with in IFN-γ might provide a promising therapeutic approach to prevent the progression of renal fibrosis. Other materials. Rat and human IFN-γ were obtained from PEPROTECH (Rocky Hill, NJ). A PGE2 high sensitivity ELISA kit was purchased from Enzo Life Science (ADI-930-001; Villeurbanne, France). Dulbecco's modified Eagle's medium (DMEM) was obtained from Sigma-Aldrich.

Methods
Cell culture. rMSCs were isolated from bone marrow of the SD rat femur and tibia. rMSCs were cultured in DMEM containing 10% FBS (Sigma-Aldrich) for the primary culture. MSCs were passaged four times before used for administration or in vitro experiments. These cells were confirmed as MSCs by promoting their differentiation into osteocytes and adipocytes with specific differentiation media 38 . Furthermore, we confirmed that standard MSC surface markers CD44 and CD90 were positive in these cells by flow cytometry ( Supplementary  Fig. S3). hMSCs derived from human bone marrow and human THP-1 monocytes were purchased from Riken BRC (Ibaraki, Japan). HK-2 cells, a human proximal tubular cell line, were obtained from the American Type Culture Collection (Manassas, VA). These cells were cultured as described previously 29 . Cell preparation for in vivo experiments. rMSCs (3 × 10 5 cells/100-mm dish) were seeded and cultured in DMEM supplemented with 10% FBS. At 80% confluence, the medium was replaced by fresh medium with or without 200 ng/ml IFN-γ, and the cells were cultured for 24 h. The cells were resuspended in PBS and subjected to in vivo analyses as IFN-γ rMSCs or control rMSCs.

Preparation of CM.
After hMSCs or rMSCs were grown to 80% confluence in DMEM supplemented with 10% FBS, the medium was replaced by fresh medium with or without 200 ng/ml IFN-γ ( Supplementary Fig. S4). Twenty-four hours later, the medium was replaced by DMEM supplemented with 0.1% FBS, which was collected after 24 or 48 h.
Cell treatment with TGF-β1. After incubation in DMEM supplemented with 0.1% FBS, CM from control hMSCs, or CM from hMSCs treated with IFN-γ for 24 h, human HK-2 cells were treated with 10 ng/ml recombinant human TGF-β1 (R&D Systems, Minneapolis, MN). Thirty minutes or 24 h later, the cells were collected and subjected to analysis of TGF-β1-induced fibrotic changes.
Polarization of M2 macrophages. To  www.nature.com/scientificreports/ tems) using Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific). Cells were cultured in DMEM supplemented with 10% FBS for 5 days and then subjected to in vitro and in vivo analyses.

Preparation of an IRI model and MSC administration. A rat IRI model, an acute kidney injury (AKI)
model, was used to analyze the progression from AKI to CKD. Rats were anesthetized by intraperitoneal injection of three types of mixed anesthetic agents (butorphanol, medetomidine, and midazolam). The left renal artery was clamped to induce ischemia. Sixty minutes later, the clamp was opened and reperfusion of blood was confirmed. After confirmation of reperfusion, PBS (vehicle), control MSCs, or IFN-γ rMSCs (5 × 10 5 cells/rat) were injected into the abdominal aorta. After 7 or 21 days, the rats were sacrificed and their kidneys were collected to evaluate inflammation and chronic fibrosis.
Preparation of a UUO model and MSC administration. A UUO model, which is an experimental renal fibrosis model, was prepared to investigate the anti-fibrotic effect of MSCs. Rats were anesthetized by the same procedure described for the IRI model. The left ureter was exposed after an abdominal midline incision and ligated to induce UUO. At 4 days after the UUO operation, PBS, control rMSCs, or IFN-γ rMSCs (2.5 × 10 6 cells/rat) were injected into the tail vein. After 7 days, the rats were sacrificed and examined for renal fibrosis.
Western blotting. Sample preparation and western blotting were performed as described previously 29 . The protein bands were quantified using ImageJ software (version 1.47v; National Institutes of Health) and normalized to GAPDH levels.
Immunohistochemical analysis. Left kidneys of IRI and UUO rats were fixed in 10% formaldehyde for 18 h. Fixed samples were embedded in paraffin and cut into 4 µm-thick sections. Immunohistochemical staining of kidneys was performed for light microscopic observation. Five fields (× 100) of the renal cortex were selected randomly. CD3-, CD68-, and CD163-positive cells and areas positive for α-SMA, Col-I, and Col-III were assessed using ImageJ software.

ELISA.
To evaluate the amount of PGE2 in CM, the ELISA was performed according to the manufacturer's instructions. PGE2 concentrations were normalized to the total protein content of MSCs.
Quantitative real-time reverse transcription-PCR. RNA extraction and quantitative reverse transcription-PCR were performed as described previously 39 . mRNA levels were normalized to the level of 18 s rRNA. Primers and TaqMan probes (TaqMan Gene Expression Assay) were obtained from Applied Biosystems (Foster City, CA). The probe set ID for rat PTGES is Rn00572047_m1 and for human PTGES is Hs00610420_m1.

Statistical analysis.
All results are expressed as the mean ± standard deviation (SD). Statistical analysis for multiple comparisons was performed using one-way ANOVA followed by Bonferroni's post-hoc test. The Student's t-test was performed to compare the difference between two groups. p < 0.05 was defined as statistically significant.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request. www.nature.com/scientificreports/