Reg3α Overexpression Protects Pancreatic β Cells from Cytokine-Induced Damage and Improves Islet Transplant Outcome

Ying Ding,1,2 Yuemei Xu,2 Xuanyu Shuai,2 Xuhui Shi,2 Xiang Chen,3 Wenbin Huang,4 Yun Liu,5 Xiubin Liang,5 Zhihong Zhang,1 and Dongming Su2,3,5

1Department of Pathology, the First Affiliated Hospital of Nanjing Medical University, Nanjing, China; 2Department of Pathology, Nanjing Medical University, Nanjing, China; 3Center of Cellular Therapy, the Second Affiliated Hospital of Nanjing Medical University, Nanjing, China; 4Department of Pathology, Nanjing First Hospital, Nanjing, China; and 5Center of Metabolic Research, Nanjing Medical University, Nanjing, China

The process of islet transplantation for treating type 1 diabetes has been limited by the high level of graft failure. This may be overcome by locally delivering trophic factors to enhance engraftment. Regenerating islet-derived 3α (Reg3α) is a pan- creatic secretory protein which functions as an antimicrobial peptide in control of and cell proliferation. In this study, to investigate whether Reg3α could improve islet engraftment, a marginal mass of syngeneic islets pretransduced with aden- oviruses expressing Reg3α or control EGFP were transplanted under the renal capsule of streptozotocin-induced diabetic mice. Mice receiving islets with elevated Reg3α production exhibited significantly lower blood glucose levels (9.057 ± 0.59 mmol/L ver- sus 13.48 ± 0.35 mmol/L, P < 0.05) and improved glucose-stimulated insulin secretion (1.80 ± 0.17 ng/mL versus 1.16 ± 0.16 ng/mL, P < 0.05) compared with the control group. The decline of apoptotic events (0.57% ± 0.15% versus 1.06% ± 0.07%, P < 0.05) and in- creased β-cell proliferation (0.70% ± 0.10% versus 0.36% ± 0.14%, P < 0.05) were confirmed in islet grafts overexpressing Reg3α by morphometric analysis. Further experiments showed that Reg3α production dramatically protected cultured islets and pancre- atic β cells from cytokine-induced apoptosis and the impairment of glucose-stimulated insulin secretion. Moreover, exposure to cytokines led to the activation of MAPKs in pancreatic β cells, which was reversed by Reg3α overexpression in contrast to control group. These results strongly suggest that Reg3α could enhance islet engraftments through its cytoprotective effect and advance the therapeutic efficacy of islet transplantation. Online address: http://www.molmed.org doi: 10.2119/molmed.2014.00104

INTRODUCTION unsolved hurdle for clinical islet trans- IFN-γ released in the aforementioned Type 1 diabetes is characterized by a plantation (2,3). Vulnerable islets are sus- pathological processes may account for gradual loss of β-cell mass and function. ceptible to the injuries caused by the hy- the graft loss and dysfunction in islet In recent years, islet transplantation has poxia, various immunological damages, transplantation (5,6). Therefore, more ef- become a promising therapeutic ap- the extent of islet revascularization and fective control of cytokine-induced dam- proach for type 1 diabetic patients (1). other factors, which lead to poor engraft- age to islet engraftment for achieving Long-term insulin independence requires ment after the procedure (3,4). Several prolonged maintenance of islet grafts has large quantities of islet infusions. How- studies reported that inflammatory cy- been suggested. ever, chronic loss of islet grafts is still an tokines, including IL-1β, TNF-α and Regenerating (Reg) family pro- teins (REGs) were first identified as se- cretory in the pancreatic juice of rats with induced acute pancreatitis (7), Address correspondence to Dongming Su, Department of Pathology, Nanjing Medical and have been proven to be a multifunc- University, 140 Hanzhong Road, Nanjing 210029, China. Phone: +86-25-86862085; Fax: +86- tional factor involved in cell proliferation 25-86862085; Email: [email protected]. or differentiation (8). Although the mem- Submitted May 15, 2014; Accepted for publication December 16, 2014; Epub bers of this family contain a similar (www.molmed.org) ahead of print December 17, 2014. C-type -like domain in the se- quence, the suggested function of the various REGs are startlingly diverse (8). As one Reg family protein, regenerating

548 | DING ET AL. | MOL MED 20:548-558, 2014 RESEARCH ARTICLE

islet-derived protein 3α (Reg3α), also (Changsha, Hunan, China). The shuttle chow diet. All animal experiments were named hepatocarcinoma-intestine- plasmid pYr-adshuttle-4 containing two carried out in accordance with the Guide /pancreatitis-associated protein expression cassettes for mouse Reg3α and for the Care and Use of Laboratory Animals (HIP/PAP), has been found to affect bio- EGFP fluorescence gene with a titer of 1 × (19) and permitted by the Institutional logical activities in multiple organs 1010 plaque-forming units (pfu)/mL was Animal Care and Usage Committee of (9–11). In the pancreas, Reg3α has been constructed, each gene independently ex- Nanjing Medical University. observed to express in the acinar epithe- pressed. The rAd-EGFP vector expressing lium and in non β cells in islets (12,13), EGFP fluorescence gene with a titer of 1 × Intraperitoneal Glucose Tolerance which has been shown to protect pancre- 1010 pfu/mL was used as control. Test (IPGTT) and Glucose-Stimulated atic acinar cells both in vitro (14) and in Insulin Release vivo (15). Meanwhile, growing evidence Animal Studies Mice were fasted for 6 h and intraperi- links Reg3α and other Reg proteins to Adult male Sprague Dawley rats (Slac, toneally injected with 3 g/kg body weight the regeneration of not only exocrine but Shanghai, China) were induced with dia- of glucose. Blood glucose levels were also of the endocrine tissue of the pan- betes using bolus streptozotocin (Sigma- measured at 0, 15, 30, 60, 120 min after creas (8). Increased Reg3α expression Aldrich, St Louis, MO, USA) treatment glucose loaded. Area under curve (AUC) also was detected in the cadaver islets with a single dose of 70 mg/kg as de- of IPGTT was calculated to compare the with type 2 diabetic individuals (16) and scribed previously (3,18). With nonfasting differences among each group. During islets from pregnant rats, both processes blood glucose concentrations measuring IPGTT, aliquots (20 μL) of blood samples associated with elevated islet mass (17), greater than 16.7 mmol/L for three con- before and 30 min after glucose infusion indicating that Reg3α may influence islet secutive days, these rats were killed for were collected from tail vein, and serum regeneration and β-cell proliferation. pancreas morphometric analysis. insulin levels were measured by mouse Therefore, with respect to the potential Male eight-week-old inbred BALB/c insulin ELISA kit (Millipore, Darmstadt, impact of Reg3α on pancreatic islets and mice (Slac) were used in this experiment Germany) following its protocol. the fact that Reg3α plays an important as donors and recipients. Diabetes was role in tissue regeneration and inflamma- induced by intraperitoneal injections of Morphometric Analysis tion, we hypothesized that elevated streptozotocin (Sigma-Aldrich) at one Harvested rat pancreata were fixed Reg3α production locally in islet grafts dose of 200 mg/kg as described previ- and embedded for immunostaining. would protect islets from inflammatory ously (3,18). With nonfasting blood glu- Paraffin sections were immunostained cytokine-induced apoptosis, contribute to cose concentrations measuring greater with rabbit anti-glucagon (1:100, improved preservation of islet mass and than 16.7 mmol/L for three consecutive SAB4501137; Sigma-Aldrich) or goat anti- better glycemic control in transplants. To days, these mice were randomly assigned Reg3α (1:100, sc-50969; Santa Cruz elucidate the effect of Reg3α on islet sur- to three groups for transplantation. Islets Biotechnology, Santa Cruz, CA, USA) vival and function during transplanta- were isolated and cultured using a previ- primary antibodies. tion, a mouse model of syngeneic mar- ously described procedure (18). Aliquots After nephrectomy under nonfasting ginal islet mass transplantation was used of islets (n = 150) were mock treated with conditions, pancreatic islet graft–bearing in this study. Further investigation of the PBS buffer or incubated with Reg3α or kidneys were fixed and embedded for im- possible underlying mechanisms also EGFP vector at a defined multiplicity of munohistochemistry and immunofluores- was carried out in cultured islets and a infection (MOI, 100 pfu/cell, 1000 cence as described previously (20,21). For β murine pancreatic -cell line INS-1. cell/islet) at 37° C, 5% CO2 for 16 h, fol- morphology analysis, hematoxylin and lowed by transplantation under the renal eosin (H&E) staining was used. Paraffin MATERIALS AND METHODS capsule (Reg3α, n = 11; EGFP, n = 9; sections were also immunostained with mock, n = 7) using the established proce- rabbit anti-insulin (1:100, 4590S; Cell Sig- Vector dure (3,18). After transplantation, the naling Technology, Danvers, MA, USA) or The plasmid vector pcDNA3.1-Reg3α food intake and body weight were re- goat anti-Ki-67 (1:100, sc-7846; Santa derived by inserting mouse Reg3α cDNA corded. Blood glucose was monitored Cruz) primary antibodies. β-cell apoptosis (GenBank NM_011259) into the Eco RI randomly every five days using Accu- in islet grafts was confirmed by the termi- and Kpn I restriction sites of pcDNA3.1 Chek (Roche, Mannheim, Germany). At nal deoxynucleotidyl transferase dUTP (Invitrogen [Thermo Fisher Scientific, 30 d after transplantation, nephrectomy nick end labeling (TUNEL) assay with an Waltham, MA, USA]) was a kind gift was performed. The blood glucose levels in situ death detection kit (Roche). The from Jun-Li Liu of Montreal Diabetes Re- were recorded for another 3 d. pancreatic β-cell proliferative and apop- search Centre at Canada. Adenovirus Animals were maintained in a 12-h totic ratio in grafts were calculated as the rAd-mReg3α was obtained commercially light–dark cycle at a temperature of 23°C number of Ki-67/insulin or TUNEL/ from Yingrun Biotechnologies Inc. with free access to water and regular insulin double-staining cells divided by

MOL MED 20:548-558, 2014 | DING ET AL. | 549 Reg3α IMPROVES ISLET TRANSPLANT OUTCOME

the number of insulin-positive cells in the were transfected using X-tremeGENE assessed by Student t test, or one-way insulin-positive region of interest, respec- transfection reagents (Roche Diagnostics) ANOVA, then Tukey multiple-comparison tively. For vessel analysis, insulin/lectin according to the manufacturer’s instruc- posttest was performed for statistical anal- (1:50, FL-1201, Vector Laboratories Inc., tions. The expression of Reg3α was de- ysis of multiple groups. P values <0.05 Burlington, Ontario, Canada) double termined in both culture media and cell were considered statistically significant. staining was used to visualize blood ves- lysate to confirm the efficacy of transfec- sels in interested areas. The vessel density tion. After 24 h, INS-1 cells were incu- All supplementary materials are available was calculated as described in legends. bated in the presence or absence of cy- online at www.molmed.org. For TUNEL staining of cultured islets, tokine cocktail as mentioned above for aliquots of isolated mouse islets (n = 50) another 24 h. Then the cell viability was RESULTS either mock treated with PBS or trans- determined by the MTT reduction con- duced with Reg3α or EGFP vector for version assay, Annexin V/PI staining Reg3α Production in Islet Grafts 16 h were incubated in the presence or and flow cytometry analysis as described Improved Syngeneic Islet Engraftment absence of the inflammatory cytokine previously (23,24). The ROS generation in Streptozotocin-Induced Diabetic cocktail (5 ng/mL human IL-β, 30 ng/mL was detected using commercial kits Mice TNF-α, and 30 ng/mL IFN-γ; R&D Sys- (Beyotime, Shanghai, China) following Before transplantation, initial experi- tems, Abingdon, UK) for 24 h. Cytokine- the protocol as described previously (25). ments prepared and characterized the induced islet cell apoptosis was confirmed rAd-mReg3α vector (Supplemental Fig- by the TUNEL assay as mentioned Western Blotting ures 1A–D). There was no significant above. The apoptotic ratio was calculated After different treatments, the proteins change in food intake and body weight as the number of TUNEL-positive stain- in the conditioned media and lysate of within each group (Supplemental Fig- ing nuclei adjusted for areas in each islet. isolated islets and INS-1 cells were sub- ure 2A) 30 d after transplantation. After In each morphological analysis, three jected to SDS-PAGE and blotted onto 30 d, although a rapid decrease of blood nonconsecutive sections were viewed PVDF membranes. The proteins were glucose was observed in all recipients under microscope (DP70, Olympus, probed with primary antibodies against after islet transplantation, diabetic mice Tokyo, Japan) or confocal microscope Reg3α (1:500, AF3907; R&D), caspase-3 transplanted with Reg3α-transduced (FV1000, Olympus). (1:1000, 9665S; Cell Signaling Technology), islets had lower random blood glucose Bcl-2 (1:1000, 2876S; Cell Signaling Tech- (RBG) levels (9.057 ± 0.59 mmol/L) in Glucose-Stimulated Insulin Secretion nology), PCNA (1:1000, BS1289; Bioworld comparison with that in diabetic mice re- (GSIS) Technology Inc. St. Louis Park, MN, ceiving the same number of islets pre- Isolated mouse islets either mock treated USA), p27KIP1 (1:500, SAB4500068; transduced with EGFP vector (13.48 ± with PBS or transduced with Reg3α or Sigma-Aldrich), Pdx-1 (1:2000, ab47267; 0.35 mmol/L) or mock-treated islets EGFP vector were plated in 24-well plates Abcam, Cambridge, UK), MafA (1:800, sc- (13.43 ± 2.01 mmol/L, Figure 1A). No (10 islets per well). Apoptosis was induced 27140; Santa Cruz), phosphorylated detectable differences were found in RBG by cytokine cocktail (5 ng/mL human IL-β, Thr202/Tyr204-p44/42 MAPK (Erk1/2) levels between EGFP and the mock 30 ng/mL TNF-α and 30 ng/mL IFN-γ) (1:1000, BS5016; Bioworld) and total group (see Figure 1A). To investigate the treatment for 24 h. After being preincu- p44/42 MAPK (Erk1/2) (1:1000, BS3627; glucose disposal rates in different groups, bated in glucose-free Krebs–Ringer bicar- Bioworld), phosphorylated Thr180/ a glucose tolerance test was performed at bonate (KRB), the islets were treated with Tyr182-p38 MAPK (1:1000, 9211S; Cell d 15 and 30 after transplantation. At 15 d KRB containing 5 mmol/L glucose for Signaling Technology) and total p38 after transplantation, a similar pattern 1 h, and followed for another 1 h in the MAPK (1:1000, 9212S; Cell Signaling Tech- was observed in diabetic mice receiving presence of 25 mmol/L glucose KRB. The nology), phosphorylated Thr183/ Tyr185- Reg3α-transduced islets, which showed insulin concentrations in collected super- JNK (1:1000, BS4322; Bioworld) and total significantly improved blood glucose natants were measured by ELISA kit as JNK (1:1000, AP0370; Bioworld), GAPDH profiles compared with EGFP or mock mentioned above. (1:5000, AP0063; Bioworld) and tubulin group at 15 min (21.29 ± 2.93 mmol/L (1:10000, PA1-41331; Thermo Fisher Scien- versus 23.43 ± 1.25 mmol/L or 23.25 ± Cell Culture, Cell Viability and tific), followed by incubation with HRP- 1.28 mmol/L, respectively; P < 0.05) and Reactive Oxygen Species (ROS) conjugated secondary antibodies for ana- 30 min (19.80 ± 3.10 mmol/L versus Detection lyze as described previously (26). 23.18 ± 1.88 mmol/L or 23.83 ± Rat insulin-secreting cell line INS-1 0.93 mmol/L, respectively; P < 0.05) after (passage 15-30) was cultured as de- Statistics glucose challenge. There was a 30.8% re- scribed previously (22). The pcDNA3.1- Data are expressed as mean ± SEM. duction of AUC in Reg3α group versus Reg3α and control pcDNA3.1 vectors Comparisons between two groups were EGFP group and a 30.1% reduction ver-

550 | DING ET AL. | MOL MED 20:548-558, 2014 RESEARCH ARTICLE

could be detected in all diabetic mice after removing the islet grafts by unilateral nephrectomy. No significant differences in blood glucose levels were detected among Reg3α, EGFP and mock group (see Fig- ure 1A), suggesting that islet grafts were the major responsible factors for the gly- cemic control in all recipients.

Reg3α Production in Islet Grafts Preserved Functional Islet Mass by Influencing β-Cell Apoptosis and Replication TUNEL assay and Ki-67 immunofluo- rescence staining were performed to de- tect β-cell apoptosis and replication rates in grafts, respectively. The apoptotic rate was reduced by about 50% in Reg3α group versus EGFP or mock group (0.57% ± 0.15% in Reg3α, 1.06% ± 0.07% in EGFP, P < 0.05; 1.16% ± 0.11% mock group, P < 0.05, Figures 2B, E), while β-cell replication in Reg3α group was al- most two-fold higher than EGFP group Figure 1. Reg3α overexpression in islet grafts improved glycemic control in streptozotocin- (0.70% ± 0.10% in Reg3α, 0.36% ± 0.14% induced diabetic mice. (A) Random blood glucose levels in diabetic mice receiving in EGFP group, P < 0.05; 0.45% ± 0.06 in Reg3α -transduced islets (Reg3α, n = 11) or EGFP vector-transduced islets (EGFP, n = 9), and mock-treated mice (Mock, n = 7) groups were measured after transplantation. (B) IPGTT mock group, P = 0.07, Figures 2C, F). was performed at 15 d after transplantation. Blood glucose levels were measured after α glucose loading at indicated time. (C) Area under curve (AUC) of IPGTT was calculated to Reg3 Production Protected Primary compare the differences among each group. (D) During IPGTT, serum insulin levels were Cultured Islets from Cytokine-Induced determined before and 30 min after glucose infusion. Values represent the means ± SEM, Damage *P < 0.05 denoted significance between Reg3α group and EGFP or Mock group. Islets pretransduced by Reg3α or con- trol EGFP vector or receiving mock treat- ment were incubated in the presence/ sus mock group (Figures 1B, C). Similar 0.05; versus 1.26 ± 0.10 ng/mL in mock absence of the inflammatory cytokine results of IPGTT were shown at d 30 group, P < 0.05, see Figure 1D). Similar re- cocktail (IL-β, TNF-α and IFN-γ) for 24 h, after transplantation (Supplemental Fig- sults also were observed at d 30 after five- to six-fold of apoptotic cells could be ures 2B, C). transplantation (Supplemental Figure 2D). observed in cytokine-treated islets com- Accordingly, we found that diabetic The effect of Reg3α production on islet pared with normal cultured islets (5.36 ± mice receiving Reg3α-transduced islets grafts also was illustrated with morpho- 0.84–fold in cytokine-treated mock group, had relatively higher plasma insulin lev- metric analysis. The retrieved graft- Figures 3A, B). However, TUNEL staining els (0.92 ± 0.12 ng/mL versus 0.67 ± bearing kidneys were embedded and sec- showed that there were lower numbers of 0.06 ng/mL in the EGFP group, P < 0.05; tioned followed by H&E staining and apoptotic cells in Reg3α-transduced islets versus 0.72 ± 0.08 ng/mL in the mock immunohistochemical staining using an than in control EGFP-transduced islets group, P = 0.13, Figure 1D) under fasting insulin antibody. There were no obvious and normal cultured islets under cytokine conditions. Furthermore, 30 min after glu- differences in morphology of grafts stimulation (3.42 ± 0.58–fold in Reg3α- cose (3 g/kg) was administrated to the among each group 30 d after transplanta- transduced, 6.84 ± 1.10–fold in EGFP mice, glucose-stimulated plasma insulin tion, but optical density of insulin staining group, P < 0.05; 5.36 ± 0.84–fold in mock levels were increased significantly in in Reg3α group was increased signifi- group, P < 0.05, see Figures 3A, B). Reg3α-transduced mice in comparison cantly compared with EGFP or mock Afterward, measurement of GSIS in pri- with control EGFP vector and mock- group (Figures 2A, D). mary cultured mouse islets after 24 h cy- treated group (1.80 ± 0.17 ng/mL versus At 30 d after transplantation, an imme- tokine (IL-1β) treatment revealed that al- 1.16 ± 0.16 ng/mL in EGFP group, P < diate recurrence of severe hyperglycemia though cytokine-injured insulin secretion

MOL MED 20:548-558, 2014 | DING ET AL. | 551 Reg3α IMPROVES ISLET TRANSPLANT OUTCOME

Figure 2. Reg3α overexpression in islet grafts preserved functional islet mass by influencing β-cell apoptosis and replication. (A) Functioning islet grafts under the renal capsule of diabetic BALB/c mice were retrieved on d 30 after transplantation. H&E staining and immunohisto- chemical staining of insulin using 3,3′-diaminobenzidine tetrahydrochloride (DAB) were analyzed. Scale bar = 100 μm and referred to all panels. (B) Apoptotic and (C) replicating β cells were stained with TUNEL and Ki-67 conjugated with red fluorescence in islet grafts, respec- tively. Insulin showed green and nuclei were stained with blue DAPI dye. Scale bar = 50 μm and referred to all panels. (D) Insulin staining was analyzed semiquantitatively by the average optical density of insulin+ cells adjusted for islet areas. (E-F) The apoptotic (E) and prolifer- ative (F) ratio of pancreatic β cells in grafts was calculated as averaging the number of TUNEL+/insulin+ or Ki-67+/insulin+ cells divided by the number of insulin+ cells in islet grafts, respectively. At least three sections from animals were analyzed in each treatment group. Values represent the means ± SEM, *P < 0.05 denoted significance between Reg3α (n = 11) and EGFP (n = 9) or Mock (n = 7) group. was decreased compared with normal cul- tion also helped to increase GSIS compared in this experiment to overexpress Reg3α tured islets (2.78 ± 0.16 ng and 5.23 ± with control and EGFP vector–treated in the INS-1 cell line (Supplemental Fig- 0.33 ng, respectively, P < 0.05), Reg3α pro- islets, although the differences did not ure 1E). To investigate whether Reg3α duction significantly restored cytokine- quite achieve significance (see Figure 3C). production had a similar protective ef- induced impairment of GSIS in response fect on insulin-secreting cell lines such to 25 mmol/L glucose stimulation (4.64 ± Reg3α Production Attenuated as primary cultured islets, cell viability 1.05 ng in Reg3α, 2.78 ± 0.16 ng in EGFP Pancreatic β-Cell Injury Induced by of INS-1 cells with different treatments group, P < 0.05; 2.78 ± 0.2 ng in mock Cytokines was measured by MTT assay. Reg3α group, P < 0.05, Figure 3C). Interestingly, Plasmid vector pcDNA3.1-Reg3α and production could restore cytokine- without cytokine treatment, Reg3α produc- control vector were prepared and used induced impairment significantly in cell

552 | DING ET AL. | MOL MED 20:548-558, 2014 RESEARCH ARTICLE

labeled Annexin V staining and the red fluorescence-labeled PI staining was per- formed simultaneously to distinguish between the late apoptotic and necrotic cells. Results showed that Reg3α pro- duction in INS-1 cells decreased signifi- cantly in both early and late apoptotic cell numbers compared with the pcDNA3.1-transfected group (Figure 4B). Similar results were found by the flow cytometry analysis (10.51% ± 2.73% versus 29.01% ± 3.73%, P < 0.01, Figures 4C, D). And there were no obvious dif- ferences in cytokine-induced apoptosis between pcDNA3.1-Reg3α and the nega- tive vector-transfected group (see Figure 4D). Elevated levels of cleaved caspase-3 were obviously found in cytokine-stimu- lated INS-1 cells (Figures 5A, D). How- ever, Reg3α production suppressed the deleterious effect of cytokines obviously (see Figures 5A, D). Bcl-2 protein levels also were determined. Higher expres- sion levels of Bcl-2 were found in the pcDNA3.1-Reg3α-transfected group in comparison with pcDNA3.1-transfected group under cytokine stimulation (Fig- ures 5A, C). In addition to the amelioration of β-cell apoptosis, the impact of Reg3α on β-cell proliferation also was studied. It was found that after 24 h treatment with cytokines, the expression levels of PCNA (proliferation cell nuclear anti- gen) and a key transcriptional activator related to β-cell proliferation called Figure 3. Reg3α overexpression protected primary cultured islets from cytokine-induced Pdx-1 (pancreatic and duodenal home- damage. (A) Cytokine-induced islet cell apoptosis was determined by TUNEL staining. obox 1) were reduced, while the TUNEL-positive cells showed red fluorescence in nuclei, and all nuclei in each islet were p27Kip1 (cyclin- dependent kinase in- marked with DAPI (blue). The transduction efficiency also was observed by green fluores- hibitor 1B) protein levels were obvi- cent protein (GFP) gene under the fluorescence microscope. Scale bar, 100 μm (refers to all panel-A subpanels). (B) The apoptotic ratio of islet cells was calculated as the number ously increased (Figures 5B, F–H). α of TUNEL-positive nuclei adjusted for islet area. At least 10 islets were analyzed in each However, Reg3 production in INS-1 group. (C) Insulin secretion under stimulation of glucose (5 mmol/L and 25 mmol/L, re- cells could observably restore these cy- spectively) was measured in each group. Values represent the means ± SEM, n = 3. *P < tokine-induced detrimental effects (see 0.05 denoted significance; n.s. was not significant. Figures 5B, F–H). Moreover, we detected the expression levels of MafA (v-maf avian muscu- viability compared with pcDNA3.1- with pcDNA3.1- transfected and normal loaponeurotic fibrosarcoma oncogene ho- transfected INS-1 cells (0.5 ± 0.03 and INS-1 cells (see Figure 4A). Furthermore, molog A), a key transcriptional activator 0.31 ± 0.03, respectively, P < 0.05, Figure Annexin V/PI staining and flow cytom- related to insulin biosynthesis. Reg3α 4A). Interestingly, in the absence of cy- etry analysis were performed. The early production restored cytokine-induced tokines, Reg3α production also in- apoptotic cells were identified by direct decreases in MafA expression levels in creased cell viability slightly compared visualization of the green fluorescence- INS-1 cells (Figures 5B, E).

MOL MED 20:548-558, 2014 | DING ET AL. | 553 Reg3α IMPROVES ISLET TRANSPLANT OUTCOME

epithelium (27–29). We further showed that Reg3α expression was restricted mainly to pancreatic α cells of streptozo- tocin-induced diabetic models, the colo- calization of Reg3α and glucagon im- munoreactivity was predominant. This report proves Reg3α has a well conserved expression in diabetic islets among differ- ent species which could imply a potential role for this protein in impaired islet turnover. Whether Reg3α immunoreactiv- ity represents a precursor endocrine cell remains a focus of our current investiga- tion. Additionally, compared with the low expression level in normal islets, islet Reg3α-immunoreactive cells were upreg- ulated significantly after streptozotocin treatment, indicating its strong induction of islet protection/regeneration in re- sponse to stress/damage. Increased ex- pression of Reg3α in diabetic islets may have been a consequence of direct toxicity of hyperglycemia, since Reg3α was identi- fied as a glucose metabolism–related gene in the pancreas for the glucose-dependent preservation of β-cell function (30). α β Figure 4. Reg3 overexpression attenuated pancreatic -cell injury induced by cytokines. During the early posttransplant period, α (A) Vector or Reg3 -transfected cells and normal cultured cells were incubated in the islets are susceptible to damage exerted presence or absence of the inflammatory cytokine cocktail for 24 h, cell viability was by inflammatory cytokines IL-1β, TNF-α measured by MTT assay. (B) Direct visualization of Annexin V/PI staining was imaged by a and IFN-γ, which results in the survival microscope. Annexin V showed green fluorescence while PI showed red. Scale bar, 150 μm (refers to all panel-B subpanels). (C) Cytokine-induced apoptosis were analyzed and subsequent engraftment of only a by flow cytometry. (D) FCM results were summarized as apoptosis rates. Values represent fraction of transplanted islets (<30% the means ± SEM, n = 3. *P < 0.05, **P < 0.01 denoted significance. transplanted cell mass) (1,31). Syngeneic islet transplantation in streptozotocin- induced diabetic recipient mice, a model Reg3α Production Suppressed on ERK phosphorylation (Figures 6D–F). representing early host nonspecific im- Cytokine-Induced Activation of p38 However, Reg3α production seems to mune responses including exposure to in- MAPK and ERK Phosphorylation have no consistent impact on cytokine- flammatory cytokines (18,32,33), allowed α To define the probable mechanisms con- activated c-Jun NH2-terminal kinase (JNK) us to focus the effects of Reg3 produc- tributing to the protective effect of Reg3α phosphorylation (data not shown). In ad- tion on transplant outcome. Local pro- production on cytokine-induced apoptosis dition, Reg3α production had no signifi- duction or pretreatment with trophic fac- and dysfunction, cytokine-activated β-cell- cant effect on MAPKs pathway in the ab- tors in islet graft may provide an ideal specific mitogen-activated protein kinase sence of cytokines (see Figure 6). window to protect against cytokine-me- (MAPKs) was examined. The p38 mitogen- diated dysfunction and subsequent im- activated protein kinase (MAPK) and an DISCUSSION provement in islet function (18). In this extracellular signal-regulated kinase As increasingly diverse functions of research, we attempted to abrogate the (ERK) phosphorylation were increased ob- Reg3α were reported (9–11), we refined detrimental effect of inflammatory cy- viously after cytokine treatment in INS-1 the investigation to focus on the expres- tokines by elevating Reg3α production cells (Figure 6). However, p38 MAPK sion and localization of Reg3α immunore- locally in syngeneic islet grafts. It was phosphorylation was reversed in Reg3α- activity in the murine diabetic pancreas shown that diabetic mice receiving a mar- overexpressed cells under cytokine treat- (Supplemental Figure 3). Reg3α previ- ginal islet mass pretransduced by Reg3α ment (Figures 6A–C). Meanwhile, Reg3α ously was considered as expressing in vector displayed better glycemic control production also showed inhibitory effects non-β-cells of NOD islets and the acinar and glucose-stimulated insulin secretion

554 | DING ET AL. | MOL MED 20:548-558, 2014 RESEARCH ARTICLE

Figure 5. Western blotting analysis of Reg3α effect on INS-1 cells viability and function. (A) After different treatment, Western blotting of Bcl-2, cleaved caspase-3 and total caspase-3 expression was showed and adjusted for tubulin expression. (B) Western blotting of Pdx-1, MafA, PCNA and p27Kip1 expression was performed. Tubulin was used as a loading control. A representative blot was presented from three independent experiments. (C-H) The ratio of interest protein to loading control was shown. Values represent the means ± SEM, n = 3. *P < 0.05 denoted significance. compared with the control group. The We performed nephrectomy of the graft- tion in the pancreas of streptozotocin-in- mechanism by which Reg3α may pro- bearing kidney at 30 d after transplanta- duced diabetic mice. As opposed to the duce this beneficial effect appeared to be tion, which showed in a recurrence of hy- Reg3α group, the control group main- related to a decrease in β-cell apoptosis, perglycemia in different groups. Thus, tained hyperglycemia along with lower coupled with an increase in β-cell replica- although Reg3α gene was reported to in- amplitudes of glucose- stimulated insulin tion from the morphological analyses of duce human pancreatic ductal tissue into release, correlating with reduced intra- the islet grafts. Furthermore, these find- insulin-producing cells and to improve graft insulin-positive cells. Indeed, Reg3α ings were in keeping with in vitro studies glycemic control in streptozotocin- has been shown to promote the prolifera- that showed the deleterious effect of cy- induced diabetic mice (34), the better gly- tion of different cell types (11,34,35). It tokines on islet viability and function was cemic control in Reg3α overexpressed would be of interest to determine the po- significantly mitigated in cytokine-treated grafts could be due to the functional islet tential capability of Reg3α on islet graft islets with elevated Reg3α production. transplants rather than to β-cell regenera- revascularization, as the rate and extent

MOL MED 20:548-558, 2014 | DING ET AL. | 555 Reg3α IMPROVES ISLET TRANSPLANT OUTCOME

Figure 6. Reg3α overexpression suppressed cytokine-induced activation of p38 MAPK and ERK but not JNK phosphorylation. Vector or Reg3α-transfected cells and normal cultured cells were incubated in the presence/absence of the inflammatory cytokine cocktail from 0-120 min. The phosphorylation of p38 and ERK were detected by Western blotting. (A–C) A time course of p38 phosphorylation was in- dicated by Western blotting and adjusted for total p38 MAPK. The phosphorylated ratio of p38 MAPK adjusted for normal cultured group also was shown in different treated groups. (D-F) A time course of ERK phosphorylation was indicated by Western blotting and adjusted for total ERK. The phosphorylated ratio of ERK adjusted for normal cultured group was also shown. GAPDH was used as a load- ing control. A representative blot was presented from at least three independent experiments. Values represent the means ± SEM, n = 3. *P < 0.05 indicated significance. of islet revascularization has been consid- antiapoptosis and pro-proliferation, but scription factors responsible for β-cell dif- ered to determine the fraction of func- not pro-angiogenic activities. ferentiation and insulin biosynthesis that tional implanted islets (18). However, Molecular mechanisms mediating the fundamentally contribute to β-cell death islet grafts with elevated Reg3α produc- biological effect of Reg3α seemed varying (37,38). Larsen, et al. found that ERK acti- tion did not display a significantly higher in different cell types and physiological vation was required for cytokine-induced content of intragraft endothelium than circumstances (9,11,35,36). Activation of expression of iNOS (38) which plays an did the control group (Supplemental Fig- cytokine-induced MAPKs plays a crucial important role in oxidative stress-induced ure 4), which means the beneficial effect role in controlling multiple gene regula- β-cell death. Meanwhile, the execution of of Reg3α on improving islet transplanta- tory networks, including the expression of cell death was highly related to the activa- tion is likely due to a combined action of pro- and antiapoptotic and tran- tion of MAPKs via triggering of ER stress

556 | DING ET AL. | MOL MED 20:548-558, 2014 RESEARCH ARTICLE

and by the release of mitochondrial death passing MAPKs, to mediate β-cell prolif- protein exhibits cytoprotective properties signals (38). In this research, we found eration and function. Recent studies have on islet survival and function, contribut- that Reg3α overexpression abrogated cy- shown that Reg3α activates PI3K-Akt sig- ing to better preservation of functional tokine-induced activation of ERK and p38 naling pathway to regulate cell prolifera- islet mass and euglycemia in marginal in INS-1 cells, thereby protecting β-cell vi- tion and/or differentiation (11,34,35). Fur- mass islet infusions. Our findings impli- ability and function. These data are con- ther study is needed to prove whether cate the potential of Reg3α to improve sistent with earlier findings that pharma- other signaling pathways are correlated transplant outcomes, thereby reducing cological inhibition of MAPKs could with Reg3α response to β-cell survival. the number of donor islets and sustain- diminish cytokine-induced β-cell apopto- As early graft damage was induced by ing long-term function. The identification sis (37,39–41). Furthermore, Pdx-1 and multiple processes (46), we admit that of Reg3α function provides insights into MafA were already known as conserved Reg3α could have other effects that po- mechanisms associated with early graft transcriptional activators that regulate tentially contribute to graft survival. In- loss, which may ultimately lead to the β-cell survival and insulin gene expres- flammatory cytokines have been verified identification of opportunities to control sion (42,43), we proved that Reg3α over- to upregulate several chemokines in the local environment of transplanted expression upregulated the protein level islets, directing lymphocytes and islets as well as to the achievement of of both. We also proved that Reg3α pro- macrophage infiltration toward the site prolonged insulin independence. duction upregulated the expression of an- of transplantation. This effect can trigger tiapoptotic protein Bcl-2 and downregu- local inflammation in islet grafts, which ACKNOWLEDGMENTS lated the expression of cell death executor exacerbates β-cell damage and leads to a We thank JL Liu from McGill University caspase-3 in the presence of cytokines. negative clinical outcome (47). Since pre- Health Centre for the kind gift of plasmid These data are in line with the upregula- vious studies represented an antiinflam- pcDNA3.1-Reg3α. Also we would like to tion of PCNA and p27Kip1, two indica- matory role of Reg3α in inflammatory thank YB Ge from Nanjing Medical Uni- tors of cell proliferation and DNA replica- bowel disease (IBD) and skin inflamma- versity and P Sun from Shanghai Institute tion (44,45), and suggest that Reg3α tion (11,48,49), studies are needed to ad- of Materia Medica for excellent technical overexpression influences the fate of β dress whether Reg3α production led to assistance with histology and islet isola- cells undergoing cytokine-induced apo- inhibition of chemokine expression in tion. We are grateful to X Han, XR Wang ptosis. We also investigated the possible cultured islets and suppression of local and H Zhou from Nanjing Medical Uni- effect of Reg3α overexpression on ROS inflammation in islet grafts of diabetic versity for helpful discussion and valuable generation. It was found that Reg3α over- mice. Although the exostosin-like glyco- advice during the development of the expression could reduce cytokine-induced syltransferase 3 (EXTL3) has been study. This research was supported by the ROS generation in INS-1 pancreatic β cells strongly suggested to be a receptor of grants from National Basic Research Pro- (Supplemental Figure 5), suggesting that Reg family protein to mediate Reg sig- gram of China (2011CB504000, 973 Pro- not only an antiinflammatory but also an naling for islet regeneration and it has gram), the National Natural Science Foun- antioxidative mechanism could be in- been demonstrated that human Reg3α dation of China (81070656, 81170252, volved in the cytoprotective effect of depends on EXTL3 to mitigate skin in- 81170715), the Open Research Project of Reg3α on pancreatic β cells. Hence we jury (11,33,50), further investigation is Shanghai Key Laboratory of Diabetes Mel- used mechanistic researches to document under way to confirm specific receptors litus (SHKLD-KF-1102) and the Distin- the inhibition of inflammatory cytokines’ on pancreatic β cells, including down- guished Medical Experts Program of damage and ROS generation via MAPKs stream mediators and other interacting Jiangsu Province, China. pathway in vitro by Reg3α. One intriguing proteins that were poorly explored. Pre- speculation is that Reg3α production vious reports displayed that pancreatic DISCLOSURE might be an intermediator involved in islet-specific overexpression of Reg3β The authors declare that they have no regulating the cytokine-activated signal protein protected mice against competing interests as defined by Molecu- molecules, partially through inhibit of the streptozotocin-induced diabetes (51). lar Medicine, or other interests that might downstream kinases that have been impli- To refine our investigation, we would be perceived to influence the results and cated in both the p38 MAPK and MAPK- like to ask whether transgenic overex- discussion reported in this paper. ERK 1/2 signaling pathways, therefore af- pression of Reg3α in islets can stimulate fecting the common final pathway for islet cell regeneration and/or shield the REFERENCES β-cell failure. However, without cytokine- islets from experimental damage. 1. Shapiro AM, et al. (2000) Islet transplantation in induced activation of MAPKs, Reg3α pro- seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive duction also exerts its effect on β cells CONCLUSION regimen. N. Engl. J. Med. 343:230–8. under normal conditions. This suggests In conclusion, our present study is of 2. Bennet W, et al. (1999) Incompatibility between that Reg3α activates other molecules, by- significance, as we elucidate that Reg3α human blood and isolated islets of Langerhans: a

MOL MED 20:548-558, 2014 | DING ET AL. | 557 Reg3α IMPROVES ISLET TRANSPLANT OUTCOME

finding with implications for clinical intraportal 19. Committee for the Update of the Guide for the and CDK4 in insulinoma cells. Growth Factors. islet transplantation? Diabetes. 48:1907–14. Care and Use of Laboratory Animals, Institute 27:195–202. 3. Zhang N, et al. (2004) Elevated vascular endothe- for Laboratory Animal Research, Division on 36. Koblas T, et al. (2012) Activation of the Jak/Stat lial growth factor production in islets improves Earth and Life Studies, National Research Coun- signalling pathway by leukaemia inhibitory fac- islet graft vascularization. Diabetes. 53:963–70. cil of the National Academies. (2011) Guide for the tor stimulates trans-differentiation of human 4. Bennet W, Groth CG, Larsson R, Nilsson B, Care and Use of Laboratory Animals. 8th edition. non-endocrine pancreatic cells into insulin-pro- Korsgren O. (2000) Isolated human islets trigger Washington (DC): National Academies Press. ducing cells. Folia Biol. (Praha). 58:98–105. an instant blood mediated inflammatory reac- 20. Zhang N, et al. (2006) Sirolimus is associated 37. Eizirik DL, Mandrup-Poulsen T. (2001) A choice of tion: implications for intraportal islet transplan- with reduced islet engraftment and impaired death—the signal-transduction of immune-medi- tation as a treatment for patients with type 1 dia- beta-cell function. Diabetes. 55:2429–36. ated beta-cell apoptosis. Diabetologia. 44:2115–33. betes. Ups. J. Med. Sci. 105:125–33. 21. Chen L, et al. (2012) Diverse mechanisms of an- 38. Cnop M, et al. (2005) Mechanisms of pancreatic 5. Contreras JL, et al. (2004) Activated protein C tidiabetic effects of the different procyanidin beta-cell death in type 1 and type 2 diabetes: preserves functional islet mass after intraportal oligomer types of two different cinnamon species many differences, few similarities. Diabetes. 54 transplantation: a novel link between endothelial on db/db mice. J. Agric. Food Chem. 60:9144–50. Suppl 2:S97–107. cell activation, thrombosis, inflammation, and 22. Guo J, et al. (2010) Blockage of ceramide metabo- 39. Larsen CM, et al. (1998) Interleukin-1beta-in- islet cell death. Diabetes. 53:2804–14. lism exacerbates palmitate inhibition of pro-insulin duced rat pancreatic islet nitric oxide synthesis 6. Amrani A, Verdaguer J, Thiessen S, Bou S, Santa- in pancreatic beta-cells. Mol. Cell requires both the p38 and extracellular signal- maria P. (2000) IL-1alpha, IL-1beta, and IFN- Biochem. 338:283–90. regulated kinase 1/2 mitogen-activated protein gamma mark beta cells for Fas-dependent de- 23. Meng ZX, et al. (2009) Activation of liver X recep- kinases. J. Biol. Chem. 273:15294–300. struction by diabetogenic CD4(+) T lymphocytes. tors inhibits pancreatic islet beta cell proliferation 40. Maedler K, et al. (2004) Glucose- and interleukin- J. Clin. Invest. 105:459–68. through cell cycle arrest. Diabetologia. 52:125–35. 1beta-induced beta-cell apoptosis requires Ca2+ 7. Keim V, Rohr G, Stockert HG, Haberich FJ. (1984) 24. Yin X, et al. (2013) Hydroxycamptothecin induces influx and extracellular signal-regulated kinase An additional secretory protein in the rat pan- apoptosis of human tenon’s capsule fibroblasts (ERK) 1/2 activation and is prevented by a sul- creas. Digestion. 29:242–9. by activating the PERK signaling pathway. Invest. fonylurea receptor 1/inwardly rectifying K+ chan- 8. Fleming A, Rosenberg L. (2007) Prospects and Ophthalmol. Vis. Sci. 54:4749–58. nel 6.2 (SUR/Kir6.2) selective potassium channel challenges for islet regeneration as a treatment 25. Wang T, et al. (2014) Cinnamtannin D-1 Protects opener in human islets. Diabetes. 53:1706–13. for diabetes: a review of islet neogenesis associ- Pancreatic beta-cells from palmitic acid-induced 41. Zhuang S, Schnellmann RG. (2006) A death- ated protein. J. Diabetes Sci. Technol. 1:231–44. apoptosis by attenuating oxidative stress. J. Agric. promoting role for extracellular signal-regulated 9. Cavard C, et al. (2006) Overexpression of regener- Food Chem. 62:5038–45. kinase. J. Pharmacol. Exp. Ther. 319:991–7. ating islet-derived 1 alpha and 3 alpha genes in 26. Chen F, et al. (2011) Dynamic regulation of PDX-1 42. Aramata S, Han SI, Kataoka K. (2007) Roles and human primary liver tumors with beta-catenin and FoxO1 expression by FoxA2 in dexametha- regulation of transcription factor MafA in islet mutations. Oncogene. 25:599–608. sone-induced pancreatic beta-cells dysfunction. beta-cells. Endocr. J. 54:659–66. 10. Viterbo D, et al. (2008) Pancreatitis-associated Endocrinology. 152:1779–88. 43. Miyatsuka T, Matsuoka TA, Kaneto H. (2008) protein 2 modulates inflammatory responses in 27. Gurr W, et al. (2002) A Reg family protein is over- Transcription factors as therapeutic targets for di- macrophages. J. Immunol. 181:1948–58. expressed in islets from a patient with new-onset abetes. Expert Opin. Ther. Targets. 12:1431–42. 11. Lai Y, et al. (2012) The antimicrobial protein type 1 diabetes and acts as T-cell autoantigen in 44. Georgia S, Bhushan A. (2006) p27 regulates the REG3A regulates keratinocyte proliferation and NOD mice. Diabetes. 51:339–46. transition of beta-cells from quiescence to prolif- differentiation after skin injury. Immunity. 37:74–84. 28. Gurr W, Shaw M, Li Y, Sherwin R. (2007) RegII is eration. Diabetes. 55:2950–2956. 12. Wang Y, et al. (2011) Coordinated age-dependent a beta-cell protein and autoantigen in diabetes of 45. Rachdi L, et al. (2006) Differential effects of p27 in and pancreatic-specific expression of mouse NOD mice. Diabetes. 56:34–40. regulation of beta-cell mass during development, Reg2Reg3alpha, and Reg3beta genes. Growth Fac- 29. Fetaud-Lapierre V, et al. (2013) Time-course pro- neonatal period, and adult life. Diabetes. 55:3520–8. tors. 29:72–81. teomic analysis of taurocholate-induced necrotiz- 46. Gibly RF, et al. (2011) Advancing islet transplan- 13. Taylor-Fishwick DA, Bowman A, Korngiebel- ing acute pancreatitis. J. Proteomics. 85:12–27. tation: from engraftment to the immune re- Rosique M, Vinik AI. (2008) Pancreatic islet im- 30. Coffey R, Nam H, Knutson MD. (2014) Microar- sponse. Diabetologia. 54:2494–505. munoreactivity to the Reg protein INGAP. J. His- ray analysis of rat pancreas reveals altered ex- 47. SoRelle JA, et al. (2013) Withaferin A inhibits pro- tochem. Cytochem. 56:183–91. pression of alox15 and regenerating islet-derived inflammatory cytokine-induced damage to islets 14. Malka D, et al. (2000) Tumor necrosis factor alpha genes in response to iron deficiency and over- in culture and following transplantation. Dia- triggers antiapoptotic mechanisms in rat pancre- load. PLoS One. 9:e86019. betologia. 56:814–24. atic cells through pancreatitis-associated protein I 31. Boker A, et al. (2001) Human islet transplanta- 48. Gironella M, et al. (2005) Anti-inflammatory ef- activation. Gastroenterology. 119:816–28. tion: update. World J. Surg. 25:481–6. fects of pancreatitis associated protein in inflam- 15. Norkina O, Graf R, Appenzeller P, De Lisle RC. 32. Udayasankar J, et al. (2009) Amyloid formation matory bowel disease. Gut. 54:1244–53. (2006) Caerulein-induced acute pancreatitis in results in recurrence of hyperglycaemia follow- 49. van Beelen Granlund A, et al. (2013) REG gene mice that constitutively overexpress Reg/PAP ing transplantation of human IAPP transgenic expression in inflamed and healthy colon mu- genes. BMC Gastroenterol. 6:16. mouse islets. Diabetologia. 52:145–53. cosa explored by in situ hybridisation. Cell Tissue 16. Marselli L, et al. (2010) Gene expression profiles 33. Rink JS, Chen X, Zhang X, Kaufman DB. (2012) Res. 352:639–46. of Beta-cell enriched tissue obtained by laser cap- Conditional and specific inhibition of NF-kappaB 50. Kobayashi S, et al. (2000) Identification of a re- ture microdissection from subjects with type 2 in mouse pancreatic beta cells prevents cytokine- ceptor for reg (regenerating gene) protein, a pan- diabetes. PLoS One. 5:e11499. induced deleterious effects and improves islet creatic beta-cell regeneration factor. J. Biol. Chem. 17. Xue Y, et al. (2010) Study on pancreatic islet survival posttransplant. Surgery. 151:330–9. 275:10723–6. adaptation and gene expression during preg- 34. Levetan CS, et al. (2008) Discovery of a human 51. Xiong X, et al. (2011) Pancreatic islet-specific over- nancy in rats. Endocrine. 37:83–97. peptide sequence signaling islet neogenesis. En- expression of Reg3beta protein induced the ex- 18. Su D, et al. (2007) Angiopoietin-1 production in islets docr. Pract. 14:1075–83. pression of pro-islet genes and protected the mice improves islet engraftment and protects islets from 35. Cui W, et al. (2009) Overexpression of Reg3alpha against streptozotocin-induced diabetes mellitus. cytokine-induced apoptosis. Diabetes. 56:2274–83. increases cell growth and the levels of cyclin D1 Am. J. Physiol. Endocrinol. Metab. 300:E669–80.

558 | DING ET AL. | MOL MED 20:548-558, 2014