Wip1 regulates the immunomodulatory effects of murine mesenchymal stem cells in type 1 diabetes mellitus via targeting IFN-α/BST2

Mesenchymal stem cells (MSCs) show significant therapeutic effects in type 1 diabetes mellitus (T1DM) as regulating the inflammatory processes. However, little is known about the detailed process of MSCs immunosuppression in T1DM. In this study, we investigated the effects of wild-type p53-induce phosphatase 1 (Wip1) on regulating MSCs immunosuppressive capacities in T1DM mice. We found that Wip1 knockout (Wip1−/−) MSCs had lower therapeutic effects in T1DM mice, and displayed weaker immunosuppressive capability. In vivo distribution analysis results indicated that Wip1−/−MSCs could home to the damaged pancreas and increase the expression of tumor necrosis factor-α (TNF-α), interleukin-17a (IL-17a), interferon-α(IFN-α), IFN-β, and IFN-γ, while decrease the expression of IL-4 and IL-10. Moreover, we confirmed Wip1−/−MSCs exhibited weaker immunosuppressive capacity, as evidenced by enhanced expression of bone marrow stromal cell antigen 2(BST2) and IFN-α. In conclusion, these results revealed Wip1 affects MSCs immunomodulation by regulating the expression of IFN-α/BST2. Our study uncovered that Wip1 is required to regulate the therapeutic effects of MSCs on T1DM treatment, indicating a novel role of Wip1 in MSCs immunoregulation properties.


INTRODUCTION
Type 1 diabetes mellitus (T1DM), an autoimmune disease induced by multiple factors, causes pancreatic infiltration of T lymphocyte and destruction of β islet cells, which ultimately leads to a significant decline in insulin release [1][2][3]. Currently, mesenchymal stem cells (MSCs) have been extensively utilized in treating T1DM and its complications as they contribute to alleviate inflammation and reduce the apoptosis of β islet cells [4,5]. However, the molecular cues of MSCs in T1DM are still unclear.
Wild-type p53-induce phosphatase 1 (Wip1), encoded by the PPM1D gene, is a serine/threonine phosphatase newly identified in wild-type p53-induced phosphatase of the PP2C family [6]. It has been reported to be closely associated with tumorigenesis, cell proliferation, as well as development and aging processes [7]. Meanwhile, it played important role in regulating the function of immune cells [8][9][10]. Recently, several studies demonstrated that Wip1 could modulate MSCs migration, proliferation, and senescent growth arrest [11,12]. Wip1 depletion in mouse embryonic fibroblasts could lead to the reduction of insulin in diabetes [10,13]. Also, Wip1 could regulate islet cell proliferation and regeneration through modulating the p38 MAPK pathway [14]. These indicated Wip1 was essential for the biological characteristics of MSCs and islet cells. Nevertheless, the effects of Wip1 on the immunomodulatory function of MSCs need to elucidation.
To investigate the effects of Wip1 on immunosuppressive properties of MSCs and the therapeutic effects of T1DM, we isolated MSCs from Wip1 −/− mice and injected into T1DM mice. Wip1 − / − MSCs showed weak anti-inflammatory activity and impaired therapeutic efforts in T1DM mice. We also found that BST2 contributed to the high expression of interferon-α (IFN-α) in Wip1 −/− MSCs, which was an important factor in the pathogenesis of T1DM. Importantly, administration of Wip1 −/− MSCs to T1DM mice potently promoted IFN-α expression and aggravation of inflammatory responses in the pancreatic microenvironment. In a word, we uncovered that Wip1 played vital roles in the therapeutic effects of MSCs in T1DM mice, which provided a novel mechanism for understanding the immunosuppressive capacity of MSCs.

RESULTS
Isolation and identification of Wip1 −/− MSCs from murine compact bones MSCs were isolated from Wip1 +/+ and Wip1 −/− mice by cultivating the digested compact bone. Fibroblast-like cells sprouted from the bone fragments and adhered to the flask 48 h (Fig. 1A). These adherent cells could be readily expanded in vitro by successive cycles of trypsinization, seeding, and culture every 3 days without visible morphologic alteration (Fig. 1B, C).
We then determined the adipogenic and osteogenic differentiation of Wip1 −/− cells in vitro. As shown in Fig. 1D, E, many Oil Red-O-positive lipid droplets were available in murine Wip1 −/− MSCs. Consistent with histochemical staining results, the mRNA expression of C/EBPα and PPAR-γ in Wip1 −/− MSCs showed elevation (Fig.1F). Under osteogenic conditions, the cells from Wip1 −/− mice displayed alkaline phosphatase activity (Fig. 1G, H). In line with the results of differentiation assays, qPCR demonstrated transcriptional expression of Osteocalcin and Runx2 under specific osteogenic cultures in adherent cells (Fig. 1I). The immunophenotype of the Wip1 −/− adherent cells was assessed by flow cytometry. The culture-expanded adherent cells were positive for CD90, CD29, and Sca-1 but were negative for CD31, MHCII, CD11b, CD34, and CD45 (Fig. 1J). The morphologic, immunophenotypic, and differentiation assays strongly indicated that the adherent cells isolated from Wip1 −/− mice were MSCs. In addition, Wip1 +/+ MSCs showed similar biological characteristics (data not shown).
Therapeutic effect of Wip1 −/− MSCs in T1DM mice To assess the therapeutic effect of Wip1 −/− MSCs in T1DM mice, Wip1 +/+ MSCs and Wip1 −/− MSCs were intravenously administered to STZ-induced T1DM mice. As shown in Fig. 2A, the body size in Wip1 −/− MSCs mice was smaller compared with Wip1 +/+ MSCs mice. The blood glucose levels were higher in Wip1 −/− MSCs compared with the Wip1 +/+ MSCs group (Fig. 2B). Additionally, the slow gain in weight caused by hyperglycemia was controlled by MSCs infusion. The average body weight of the Wip1 +/+ MSCs group was larger than the Wip1 −/− MSCs group (Fig. 2C).
To further evaluate whether Wip1 −/− MSCs infusion could reduce the pathological damages in T1DM mice, the histological examination of islet tissues was performed by H&E staining and immunohistochemical staining on day 28. As shown in Fig. 2D, E, islets of the Wip1 −/− MSCs group exhibited mild inflammation. Additionally, there were a few islet areas with preserved islet morphology, indicating that Wip1 −/− MSCs infusion could prevent the destruction of islets in T1DM. Moreover, the islet area and the mean density of insulin-positive cells in pancreas treated with Wip1 +/+ MSCs were larger than that of the Wip1 −/− MSCs group, despite that the effects of Wip1 −/− MSCs were better than T1DM mice (Fig. 2F, G). These findings indicated the therapeutic effects of Wip1 −/− MSCs were weaker than Wip1 +/+ MSCs in T1DM mice.
Wip1 −/− MSCs failed to reduce the inflammatory response in T1DM mice Previous studies indicated that the immunoregulation of MSCs was closely involved in the Th1 inflammatory processes in T1DM [15,16]. To investigate whether Wip1 regulated the immunomodulatory function of MSCs in T1DM. Flow cytometry was used to detect IFN-γ in Th1 cells (CD4 + IFN-γ + T cells) of spleen lymphocytes in each group. As shown in Fig. 3A, B, the percentage of Th1 cells (CD4 + IFN-γ + T cells) in the T1DM group was significantly higher than that in other groups (P < 0.01). Both Wip1 −/− MSCs and Wip1 +/+ MSCs infusion significantly decreased the percentage of Th1 cells (P < 0.05). However, the percentage of Th1 cells in the Wip1 −/− MSCs group was lower than that of the T1DM group and was higher than that of the Wip1 +/+ MSCs group (P < 0.05).
Subsequently, we investigated whether Wip1 could affect proinflammatory factors in T1DM mediated by MSCs infusion. Compared with the Wip1 +/+ MSCs group, serum IFN-γ and IL-17a increased and IL-4 decreased in the Wip1 −/− MSCs infusion group (Fig. 3C). These data indicated that Wip1 is critical for the immunomodulatory activity of MSCs in T1DM mice.

Wip1 interacted with BST2
To investigate the involvement of Wip1 in MSCs immunomodulatory activity, we then determined the gene expression profiles in Wip1 +/+ MSCs and Wip1 −/− MSCs by microarray analysis. Bioinformatics methods were used to analyze the microarray results, which predicted that 23 genes were associated with immunomodulatory responses. Among the upregulated genes in Wip1 −/− MSCs, BST2 was identified to be related to inflammatory processes, and highly expressed in Wip1 −/− MSCs (Fig. 4A, B and Table S1). To further explore whether BST2 was upregulated in Wip1 −/− MSCs, we examined the expression of BST2 in Wip1 +/+ MSCs and Wip1 −/− MSCs. As shown in Fig.4C-F, compare with Wip1 +/+ MSCs, the protein level of BST2 increased in Wip1 −/− MSCs.
Subsequently, we further investigated whether Wip1 interacted with BST2. The 293 T cells were transfected with Wip1-EGFP, followed by determining the expression of Wip1 and BST2. EGFP expression indicated high transduction efficiency, which showed a transfection efficiency of 61 ± 10.81% and 69 ± 7.93% for the EGFP group and Wip1-EGFP group, respectively (Fig. 4G, H). As shown in Fig. 4I, after overexpression of Wip1 in 293 T cells, thereby the BST2 expression decreased. Furthermore, endogenous CO-IP demonstrated that Wip1 and BST2 interact in MSCs (Fig. 4J). Collectively, these results suggest that Wip1 is associated with BST2 in MSCs.

IFN-α secreted by Wip1 −/− MSCs aggravated inflammatory response in the pancreatic microenvironment
Several studies had shown that MSCs preferentially homed to the sites of tissue damages, where they enhanced wound healing and were involved in modulating the balance of inflammatory response [21]. To determine whether Wip1 −/− MSCs could migrate to the damaged sides in the pancreas, red fluorescence stained Wip1 −/− MSCs and Wip1 +/+ MSCs were administrated into T1DM mice. After staining for 72 h, red fluorescence occurred around blood vessels in pancreatic tissues (Fig. 6A), which indicated that Wip1 −/− MSCs could home to the pancreatic tissues in T1DM mice.
IFN-α was indicated to play important roles in T1DM immune disorders. To explore whether IFN-α increased due to Wip1 −/− MSCs infusion, the expression of IFN-α in the pancreatic supernatant of T1DM mice was detected by ELISA. The expression of IFN-α in the pancreatic supernatant of the Wip1 −/− MSCs group showed a significant increase (2103.14 ± 797.1 pg/mL) compared with the control group and Wip1 +/+ MSCs group (P < 0.05). Compared with the T1DM model group, there was no significant difference. These data demonstrated that Wip1 −/− MSCs could promote IFN-α secretion after homing to pancreatic tissue (Fig. 6B). Besides, we further examined the expression of IFN-α, IFN-β, and IFN-γ in splenic lymphocytes. The expression of IFN-α, IFN-β, and IFN-γ in the Wip1 −/− MSCs group was higher than that of the Wip1 +/ + MSCs group, which indicated that Wip1 −/− MSCs also could promote the expression of IFN-α, IFN-β, and IFN-γ in spleen lymphocytes, and further aggravate the inflammatory response (Fig. S1).

DISCUSSION
T1DM is a public health challenge worldwide with a hallmark of autoimmune attack of β-cells leading to specific loss of the insulinsecreting cells [22]. Previous clinical studies demonstrated that moderate immunosuppression in T1DM could prevent further loss of insulin production, which then attenuated the clinical symptoms [23][24][25]. With the capacity to modulate immune responses, MSCs have been widely employed to treat various inflammatory diseases including T1DM [25][26][27], however, the potential mechanisms remain largely elusive.
The study indicated wip1 effected the immunomodulatory functions of MSCs in T1DM via targeting IFN-α/BST2. Also, we provided a new mechanistic insight in the Wip1 regulated the immunomodulatory function of MSCs based on in vitro and in vivo evidence. Wip1 −/− MSCs had impaired therapeutic effects in the T1DM model, including failed to decrease inflammatory response and increase IFN-α in the pancreas microenvironment.
Wip1 is expressed in hematopoietic progenitors, stem cells, neutrophils, macrophages, B and T lymphocytes [28,29]. Recent studies demonstrated that Wip1 could regulate immune cells' proliferation, activation, and differentiation activities [8,28,30,31]. Moreover, Wip1 was crucial for the modulation of MSCs migration, but little is known about the effects of Wip1 on the immunosuppression capacity of MSCs. In this study, we showed that the therapeutic effect of Wip1 −/− MSCs on T1DM in mice was significantly lower than that of the Wip1 +/+ MSCs group. Wip1 −/ − MSCs triggered no decline of the Th1 frequency, and the level of serum IFN-γ and IL-17a was nearly the same in T1DM mice.
Previously, Th1 pro-inflammatory autoimmune response led to a failure of immune tolerance to β-cells [25], while MSCs decreased the frequency of Th1 cells and reduced IFN-γ and IL-17a in T1DM. Therefore, the immunosuppressive effects of MSCs on T1DM were more likely to depend on Wip1. However, little is known about the molecular mechanism involved in the MSCs immunomodulation by Wip1. The Wip1 −/− MSCs microarray assay showed that the BST2 expression increased. Moreover, Western blot and CO-IP assay indicated that Wip1 not only inhibits the expression of BST2 but also Wip1 and BST2 interact in MSCs. Previous studies reported that the expression of BST2, serving as a surface protein involved in viral vesicle budding in response to interferons [32,33]. Our data also indicated that IFN-α in Wip1 −/− MSCs was significantly upregulated in protein and mRNA levels. Moreover, the BST2 deficiency significantly inhibited the expression of IFN-α. These results confirmed that high expression of IFN-α in Wip1 −/ − MSCs was related to endogenous BST2 expression.
IFN-α, encoded by the type 1 interferon gene, is known to be crucial for the pathogenesis of autoimmunity [34,35], and contributed to the evolution of autoimmunity causing damages to β cells. Moreover, previous studies indicated MSCs could preferentially home to damage sites in certain tissues, which then enhanced wound healing and tissue regeneration, and inhibited the inflammatory reprocess [36,37]. We revealed that IFN-α was highly expressed in Wip1 −/− MSCs. Wip1 −/− MSCs also could home to the pancreas in T1DM mice, leading to an increase of IFN-α, IL-17a, and TNF-α, and decrease of IL-4 and IL-10 in the pancreas, which exacerbated the inflammatory response and deterioration of clinical symptoms in T1DM mice. Therefore, our studies, together with previous investigations, demonstrated that there was high expression of IFN-α in Wip1 −/− MSCs, which was mediated by BST2. However, in our study, the exact mechanisms of the Wip1-BST2-IFN-α axis in MSCs are still not well defined, which requires further investigation.
In summary, Wip1 deletion contributed to IFN-α expression in MSCs via increasing the expression of BST2, which thereby impaired the therapeutic effects of MSCs in T1DM mice. Wip1 −/−  MSCs migrated to the pancreas in T1DM mice, which led to secretion of IFN-α and augmentation of inflammatory processes. Additionally, we indicated the importance of the Wip1-BST2-IFN-α axis in the therapeutic effects of MSCs in T1DM mice.

CONCLUSIONS
This study identified the role of the Wip1 in regulating the immunomodulatory capabilities of MSCs. Furthermore, Wip1 was shown to directly impact the immunosuppression effect of MSCs in T1DM mice. The mechanism of Wip1 in MSCs immunomodulatory properties is pivotal for claiming the Wip1-BST2-IFN-α axis in the therapeutic effects of MSCs in T1DM mice.

T1DM induction and experimental therapies
To induce the T1DM model as previously described [38,39], male C57BL/6j mice were fed ad libitum for 48 h followed by food and water deprivation for 10-16 h before T1DM induction. Subsequently, streptozotocin (STZ, 50 mg/kg, Sigma-Aldrich, v900890) dissolved in sodium citrate buffer was administrated by intraperitoneal injection in the lower left abdomen once per day for 5 days. The ad libitum diet was resumed immediately after the first injection. T1DM was defined as a blood glucose of ≥16.7 mmol/L for three consecutive random samplings.

Cell preparation
Murine primary MSCs were isolated from mouse compact bone and identified according to the previous description [40]. One-week-old mouse pups were collected and tailed for genotyping. The femur and tibia of Wip1 +/+ and Wip1 −/− mice were isolated under aseptic conditions. The bone fragments were digested using type II collagenase at 37°C for 45 min. Afterward, the mixture was transferred to α-MEM complete medium containing 10% fetal bovine serum. Adherent cells were maintained with medium replenishment every 3 days. The third passage cells were used for measuring differentiation characterization.

Histology analysis
Pancreatic sections were prepared from at least six mice in each group for hematoxylin-eosin (H&E) staining. Mouse insulin antibody (1:100; Cell Signaling Technology, 3014) was used for immunohistochemical and immunofluorescent staining of sections.
Gene knockdown using siRNA

Western blotting
Total protein was extracted from the cell pellet with RIPA lysis buffer (eBioscience, 89901). The protein concentration was determined by the BCA method using a commercial kit (Bio-Rad, M60-009RDPD). Protein lysis was separated by SDS-PAGE gel and then was transferred to a PVDF membrane. The PVDF membrane was blocked for 2 h in TBST buffer containing 5% skim milk powder and then tested with WiP1 Rabbit mAb (CST, 11901), BST2 Rabbit mAb (ABclonal, A12315), and β-Actin Rabbit mAb (CST,4970). Proteins were visualized using HRP-conjugated secondary antibodies. The bands were exposed by ECL software.

Co-immunoprecipitation (Co-IP) assay
MSCs were rinsed twice with PBS, and lysed with RIPA lysis buffer (eBioscience, 89901) on 4°C for 30 min, followed by centrifugation at 12,000×g for 15 min. Lysis supernatants were incubated with Wip1 Rabbit mAb or control IgG (Abcam, ab190475) overnight on a rotator, followed by the addition of 30 μl prewashed protein A/G agarose beads (CWBIO, CW0349S) for another 2 h. After extensive washing with a diluted lysis buffer, the lysate was used for western blot analysis.

Quantitative real-time PCR (qPCR)
Total RNA was extracted from cultured cells using the TRIzol Reagent (Invitrogen, 15596018) and reverse-transcribed using SuperRT cDNA Synthesis Kit (CWBIO, CW0741). qPCR amplification was performed on a 7500 Real-Time system. The qPCR primers (1.0 OD per vial) purchased from Sangon (Sangon Biotech) were listed in Table S2. Relative quantification was performed using mouse-specific GAPDH primers.

Statistical analysis
Data were presented as mean ± standard error of mean (SEM). The statistical significance of the differences between treatment groups was analyzed by Student's t-test or one-way analysis of variance. Data analysis was performed using GraphPad Prism Version 6.0 software. P < 0.05 was considered to be statistically significant.

DATA AVAILABILITY
All data generated or analyzed during this study are included in this published article and its supplementary information files.