Recombinant Reg3α Prevents Islet β-Cell Apoptosis and Promotes β-Cell Regeneration

Progressive loss and dysfunction of islet β-cells has not yet been solved in the treatment of diabetes. Regenerating protein (Reg) has been identified as a trophic factor which is demonstrated to be associated with pancreatic tissue regeneration. We previously produced recombinant Reg3α protein (rReg3α) and proved that it protects against acute pancreatitis in mice. Whether rReg3α protects islet β-cells in diabetes has been elusive. In the present study, rReg3α stimulated MIN6 cell proliferation and resisted STZ-caused cell death. The protective effect of rReg3α was also found in mouse primary islets. In BALB/c mice, rReg3α administration largely alleviated STZ-induced diabetes by the preservation of β-cell mass. The protective mechanism could be attributed to Akt/Bcl-2/-xL activation and GRP78 upregulation. Scattered insulin-expressing cells and clusters with small size, low insulin density, and exocrine distribution were observed and considered to be neogenic. In isolated acinar cells with wheat germ agglutinin (WGA) labeling, rReg3α treatment generated insulin-producing cells through Stat3/Ngn3 signaling, but these cells were not fully functional in response to glucose stimulation. Our results demonstrated that rReg3α resists STZ-induced β-cell death and promotes β-cell regeneration. rReg3α could serve as a potential drug for β-cell maintenance in anti-diabetic treatment.


Introduction
Destruction and dysfunction of islet β-cells exist in different types of diabetes and remain a major challenge for the cure of diabetes. Maintenance of functional β-cells was likely achieved by a short-term intensive insulin therapy, which is recently reported as not providing benefit on islet β-cells in Type 2 diabetic (T2D) patients [1]. The fasting-mimic therapy was proved, restoring pancreatic β-cell mass by the mechanism of inducing β-cell neogenesis, rather than a direct protection on the preexisting β-cells [2]. Expansion of islet β-cells through self-replication also displays an extremely low rate in adult humans and rodents [3]. At present, islet transplantation is the only approach available to supplement insulin-secreting cells in patients, while insufficient number of donors and immune rejection limit its clinical application [4]. Searching for agents to prevent β-cell loss is urgent for future anti-diabetic therapy. MIN6 cells was increased by rReg3α treatment, which was attenuated under high concentrations with 1% FBS at 48 and 72 h ( Figure 1A,B). The mitogenic effect of rReg3α was further confirmed by EdU staining (Supplementary Figure S1). In the treatment of 100 nM rReg3α, cell cycle assay showed that the proportion of the cells in S and G2 phases increased from 11% to 26% ( Figure 1C,D) and cyclin D1 and CDK4 levels were 3 and 2.5 times higher than the control ( Figure 1E,F). Phospho-ATF-2 was also detected as endogenous Reg3α is reported to increase phospho-ATF-2 which binds to cyclin D1 promoter to activate gene expression. In agreement with early reports [26][27][28][29], rReg3α induced approximately 2-, 3.5and 7-fold increases of phosphorylated Erk, Akt, and ATF-2 ( Figure 1G,H), and antagonists to Akt and Erk efficiently abolished the proliferative effect ( Figure 1I). These results indicate that rReg3α is bioactive in activating Akt and Erk phosphorylation, inducing insulinoma MIN6 cell proliferation. The relative protein levels were corrected by the corresponding total proteins. * p < 0.05, ** p < 0.01, and *** p < 0.001 using one-way ANOVA, N = 3.

rReg3α Protects MIN6 Cells from STZ-Induced Cell Apoptosis
The effect of rReg3α in β-cell protection was also investigated in MIN6 cells. In a challenge with STZ, the remaining healthy cells took less than 20% of the total cell number, which was rescued to 51% by rReg3α treatment. Reduction of late apoptotic (from 44% to 11%) and necrotic cells (from 13% to 1%), while accumulation of early apoptotic cells (from 24% to 37%), were found, possibly owing to a prevention of cells entering late apoptotic stage (Figure 2A,B). In cells challenged with STZ, caspase-3 cleavage was found twofold higher than the control, but this was significantly attenuated by rReg3α treatment. A fourfold elevation of Akt phosphorylation and 10-and 2.5-fold increases of Bcl-2 and -xL contents were detected with significant differences compared to the STZ group ( Figure 2C,D), which may contribute to the protective mechanism [21,22]. 24% to 37%), were found, possibly owing to a prevention of cells entering late apoptotic stage (Figure 2A,B). In cells challenged with STZ, caspase-3 cleavage was found twofold higher than the control, but this was significantly attenuated by rReg3α treatment. A fourfold elevation of Akt phosphorylation and 10-and 2.5-fold increases of Bcl-2 and -xL contents were detected with significant differences compared to the STZ group ( Figure 2C,D), which may contribute to the protective mechanism [21,22]. The relative cleaved caspase-3 and phosphorylated Akt contents were corrected by the corresponding total proteins and the relative Bcl-2 and -xL contents were corrected by β-actin. * p < 0.05, ** p < 0.01 and *** p < 0.001 using one-way ANOVA, N = 3.

rReg3α Alleviates STZ-Induced β-Cell Loss in BALB/c Mice
To further assess the in vivo effect in protecting islet β-cells, BALB/c mice were pretreated with rReg3α and then challenged with STZ. Blood glucose and bodyweight were recorded and the differences in variation were compared using AUC. Vehicle (PBS) treat- The relative cleaved caspase-3 and phosphorylated Akt contents were corrected by the corresponding total proteins and the relative Bcl-2 and -xL contents were corrected by β-actin. * p < 0.05, ** p < 0.01 and *** p < 0.001 using one-way ANOVA, N = 3.

rReg3α Alleviates STZ-Induced β-Cell Loss in BALB/c Mice
To further assess the in vivo effect in protecting islet β-cells, BALB/c mice were pretreated with rReg3α and then challenged with STZ. Blood glucose and bodyweight were recorded and the differences in variation were compared using AUC. Vehicle (PBS) treatment was set as the control, which exhibited as normal. STZ induced a rapid onset of hyperglycemia that peaked at 25 mmol/L, which was significantly reduced to 15 mmol/L by the treatment of rReg3α ( Figure 3A,B). The acute weight loss in the STZ group reached 16 g, whereas in the STZ+rReg3α group, it was only 7.2 g ( Figure 3C,D). An obvious serum insulin deficiency was detected in the STZ group, 8.3 µIU/mL, down to a half of the normal amount, which was partially restored to 12 µIU/mL by rReg3α treatment ( Figure 3E).  In the pancreatic sections, severe β-cell diminution was observed in the STZ group, which was, however, much alleviated by rReg3α treatment (Figure 3F,G). The β-cell mass in the STZ+rReg3α group was 66 mg/kg b.w., significantly higher than that of 45 mg/kg b.w. in the STZ group ( Figure 3H). By semiquantitative densitometric analysis, the total insulin densities per total pancreatic area, referred to as "relative pancreatic insulin density", were approximately 10, 4.2, and 6.9 (1000 × IOD/mm 2 ), respectively, in the control, STZ, and STZ+rReg3α groups, suggesting a preservation of insulin content in the pancreas by rReg3α treatment ( Figure 3I). Islet α-cell percentage (per islet cells) increased and the (I) Semi-quantification of insulin content in the whole pancreas. The value of "relative pancreatic insulin density" indicated the total insulin density per total pancreatic area. (J) α-cell percentage in the islets. * p < 0.05, ** p < 0.01, and *** p < 0.001 using one-way ANOVA, N = 5, 5, 11, 14. In the pancreatic sections, severe β-cell diminution was observed in the STZ group, which was, however, much alleviated by rReg3α treatment (Figure 3F,G). The β-cell mass in the STZ+rReg3α group was 66 mg/kg b.w., significantly higher than that of 45 mg/kg b.w. in the STZ group ( Figure 3H). By semiquantitative densitometric analysis, the total insulin densities per total pancreatic area, referred to as "relative pancreatic insulin density", were approximately 10, 4.2, and 6.9 (1000 × IOD/mm 2 ), respectively, in the control, STZ, and STZ+rReg3α groups, suggesting a preservation of insulin content in the pancreas by rReg3α treatment ( Figure 3I). Islet α-cell percentage (per islet cells) increased and the cells apparently migrated towards the central part of the islets when mice were challenged with STZ. No significant difference in α-cell percentage was found between the STZ and STZ+rReg3α groups ( Figure 3F,J).

rReg3α Induces Insulin-Producing Cell Neogenesis in the Exocrine Pancreas
After carefully examining the pancreatic sections, we found a few scattered small insulin-producing cell clusters distributed in the exocrine part from the STZ+rReg3α mice, but none in the control, and seldom in the STZ group ( Figure 4A). To distinguish these cells from regular islets, we scanned the pancreatic sections in the control mice and found that the size of the smallest islet structure was 425 µm 2 ( Figure 4A, up-left), leading to a speculation that the insulin-expressing clusters with size less than 425 µm 2 may originate from neogenesis. None of these cells was accompanied by glucagon-producing cells ( Figure 4B), suggesting a very low probability that the views with scattered insulinproducing cells were the cross sections of the islet edge. The number of these small insulin-producing clusters in the STZ+rReg3α group was 24/cm 2 (<2% of the total β-cell mass), which was significantly higher than that of 5.1/cm 2 in the STZ group ( Figure 4C). Using the 425 µm 2 threshold, the "islets" in the STZ+rReg3α mice could be classified into two subgroups, the "small insulin-producing clusters", with a size smaller than 425 µm 2 , and the "regular islets", which were larger than 425 µm 2 ( Figure 4D). By excluding the putative neogenic β-cell clusters below 425 µm 2 , the data of the regular islets perfectly complied with normal distribution ( Figure 4E), supporting that there were two distinct subgroups of the cells and one could come from β-cell neogenesis. Statistical analysis showed that the relative insulin density in the small insulin-producing clusters reached only a half of the regular islets, 314 vs. 524 (1000 × IOD/mm 2 ) ( Figure 4F), which agrees with low expression of insulin in young β-cells [30].

rReg3α Promotes Insulin-Producing Cell Neogenesis in Isolated Mouse Acinar Cells
It is reported that WGA specifically binds to N-acetyl glucosamine on the plasma membrane and can be used for acinar cell lineage tracing [31]. Immunofluorescent staining in the pancreatic section was performed to evaluate the labeling of WGA. Despite a strong background in the cytosol in all cell types, it is obvious that only the exocrine cell membrane was outlined with green fluorescence, and not the endocrine part (Supplementary Figure S2). In the ex vivo experiment, WGA could not reach the cytosol because the labeling was applied in live cells with integral membrane structure. Thus, the few contaminated β-cells were WGAwhile WGA + /Insulin + indicated the cells that originate from β-cell neogenesis. By the treatment of rReg3α, the proportion of insulin + /WGA + cells were 27% and 34%, respectively, at d 3 and 5, which were significantly increased compared to the control ( Figure 5A,B). As auto-reprogramming occurred with a higher frequency at d 5, we selected the time point d 3 for the following examination. The mRNA levels of Ins1 and Ins2 increased by 3.5-and threefold ( Figure 5C) and the content of secreted insulin rose from 0.16 to 0.66 ng/µg DNA under low glucose condition (basal). However, the insulin release was neither competent to respond to high glucose stimulation nor comparable to the level in isolated islets ( Figure 5D), implying that these insulin + cells were still immature and far from functional. Upregulation of Ngn3 expression with nuclear localization by rReg3α treatment was detected, 22% vs. that of 0.99% in the control ( Figure 5E,F), which is in accordance with the report that re-expression of Ngn3 is critical for acinar-to-beta cell transdifferentiation [19]. In addition, 3.6-and 1.7-fold elevations of Ngn3 mRNA and protein levels were further confirmed, and 1.6-fold increase of Stat3 phosphorylation was also detected (Figure 5G-I). As it is well known that Ngn3 expression is regulated by Stat3 phosphorylation and the latter is associated with Reg signaling [19,20], our results hint that rReg3α could promote β-cell neogenesis via activating Stat3/Ngn3 signaling.

rReg3α Promotes Insulin-Producing Cell Neogenesis in Isolated Mouse Acinar Cells
It is reported that WGA specifically binds to N-acetyl glucosamine on the plasma membrane and can be used for acinar cell lineage tracing [31]. Immunofluorescent staining in the pancreatic section was performed to evaluate the labeling of WGA. Despite a

rReg3α Induces GRP78 Expression and Prevents β-cell Apoptosis
As the newborn insulin-producing cells were likely still young and took only <2% of the total β-cells in the pancreas, the effectiveness of these cells to control the glycemia was questioned. In BABL/c mice, rReg3α treatment posterior to STZ exhibited elevated amount of small insulin-producing cell clusters in the pancreas but had no effect in the glycemic control ( Supplementary Figures S3 and S4). These results reveal that the neogenic β-cells were not potent enough to fight against STZ-induced diabetes, and other protective mechanism(s) may exist and contribute to the effect of rReg3α treatment.
In the mice pretreated with rReg3α, pancreatic sections early at d 2 and 7 after STZ were obtained for histological examination. At both time points, the proportion of β-cells was found in the same tendency ( Figure 6A-C) as observed at d 15 ( Figure 3). The TUNEL signal significantly increased in the STZ group, which was largely alleviated by rReg3α treatment (Figure 6D-F). Of note, by STZ challenge only 3.1% of the β-cells were TUNEL + at d 2, possibly due to the main effect of STZ inducing β-cell necrosis [11,14]. To disclose the underlying protective mechanism(s), various antibodies were used for immunofluorescent examination. Foxo1 was stained in all β-cells without any meaningful change between groups (Supplementary Figure S5A). Ki67, Aldh1a3, and Ngn3 were barely stained in the endocrine islets (Supplementary Figure S5B-D), revealing that those effects of mitogenetic, regenerative, and de-and re-differentiation were negative. Oct4 + cells were not detectable, while Sox2 + nuclei were observed in the peripheral of a few islets without significant difference (Supplementary Figure S5E,F).
GRP78 expression was found to be upregulated in the islets by rReg3α treatment. At d 2, the percentages of GRP78 high cells were 2.4%, 0.86%, and 4.2% of the total islet cells, and at d 7, those were 9.9%, 11.4%, and 27%, respectively, in the rReg3α, STZ, and STZ+rReg3α groups, all of which were significantly higher than the control ( Figure 6G-I). At d 7, the expression of GRP78 was dramatically increased in the STZ group, but which missed the best window for β-cell protection in the initial few days after STZ. By a scan of hundreds of the islets from all groups, only a minority of GRP78 high cells were found co-stained with insulin ( Figure 6J), and none was glucagon + nor somatostatin + , implying that these GRP78 high cells may originate from β-cells. GRP78 high cells appeared at d 2 after the initial rReg3α treatment ( Figure 6K), at the right time of STZ injection, which is thought to contribute to the protection on islet β-cells.

rReg3α Protects Against STZ-Induced Cell Apoptosis in Isolated Islets
The protective effect of rReg3α was further confirmed in freshly isolated mouse islets. Large numbers of early and late apoptotic cells, taking 25% and 11%, respectively, of the total cells, were observed in the STZ group. These apoptotic cells were significantly reduced to 20% and 3.6% by the treatment of rReg3α ( Figure 7A,B). The impaired insulin secretion was also largely rescued, to half of the control, under both basal and high glucose conditions ( Figure 7C). BrdU incorporation showed that rReg3α could not stimulate cell proliferation in isolated islets ( Figure 7D), which is different from the results in the insulinoma MIN6 cell line (Figure 1). The inconsistent results can be interpreted as that tumor cells have a high capacity of proliferation while in isolated primary islets the turnover rate is very low. Western blotting results showed that the contents of GRP78, Bcl-2, and -xL and phosphorylated Akt were upregulated, and cleaved caspase-3 was downregulated ( Figure 7E,F), suggesting that GRP78 and Akt/Bcl-2/-xL could be involved in the protective mechanism.

rReg3α Protects Against STZ-Induced Cell Apoptosis in Isolated Islets
The protective effect of rReg3α was further confirmed in freshly isolated mouse islets. Large numbers of early and late apoptotic cells, taking 25% and 11%, respectively, of the total cells, were observed in the STZ group. These apoptotic cells were significantly reduced to 20% and 3.6% by the treatment of rReg3α ( Figure 7A,B). The impaired insulin secretion was also largely rescued, to half of the control, under both basal and high glucose conditions ( Figure 7C). BrdU incorporation showed that rReg3α could not stimulate cell proliferation in isolated islets ( Figure 7D), which is different from the results in the insulinoma MIN6 cell line (Figure 1). The inconsistent results can be interpreted as that tumor cells have a high capacity of proliferation while in isolated primary islets the turnover rate is very low. Western blotting results showed that the contents of GRP78, Bcl-2, and -xL and phosphorylated Akt were upregulated, and cleaved caspase-3 was downregulated ( Figure 7E,F), suggesting that GRP78 and Akt/Bcl-2/-xL could be involved in the protective mechanism. The relative GRP78, Bcl-2, and -xL contents were corrected by βactin and the relative phosphorylated Akt and cleaved caspase-3 contents were corrected by the corresponding total proteins. * p < 0.05 and ** p < 0.01 using one-way ANOVA, N = 3.

Discussion
Reg protein is worthy of being named as "regenerating protein" for the evidence in support of a great potential in the treatment of many diseases [6][7][8]. We previously prepared rReg3α and proved its protective effect against acute pancreatitis in mice [22]. In the present study, not only was a protection on the islets observed, but there was also a unique bioactivity in promoting β-cell neogenesis. Although the neogenic β-cells did not have a leading effect on the ameliorated hyperglycemia, rReg3α could still serve as a potential drug for β-cell protection/regeneration in future treatment. * p < 0.05, ** p < 0.01 using one-way ANOVA, N = 3. (C) GSIS assay, * p < 0.05 and ** p < 0.01 using one-way ANOVA, N = 3. (D) BrdU incorporation, N = 5. (E) Western blotting detection of GRP78, cleaved caspase-3, phosphorylated Akt, and Bcl-2, and -xL levels. (F) Densitometric quantification of the protein levels in panel E. The relative GRP78, Bcl-2, and -xL contents were corrected by β-actin and the relative phosphorylated Akt and cleaved caspase-3 contents were corrected by the corresponding total proteins. * p < 0.05 and ** p < 0.01 using one-way ANOVA, N = 3.

Discussion
Reg protein is worthy of being named as "regenerating protein" for the evidence in support of a great potential in the treatment of many diseases [6][7][8]. We previously prepared rReg3α and proved its protective effect against acute pancreatitis in mice [22]. In the present study, not only was a protection on the islets observed, but there was also a unique bioactivity in promoting β-cell neogenesis. Although the neogenic β-cells did not have a leading effect on the ameliorated hyperglycemia, rReg3α could still serve as a potential drug for β-cell protection/regeneration in future treatment.
In mouse pancreas, Reg3α is constitutively expressed in the acini and ducts and inducible in the endocrine islets in NOD mice [32]. A single-cell transcriptomic analysis demonstrated that a subgroup of Reg3A-positive acinar cells with progenitor-like characteristics exists, neighboring the islets in human pancreas [33]. The fragment of Reg3A-derived pentapeptide can regulate the differentiation of human pancreatic progen-itor cells into functional β-cells [34]. In this study, our results again support the opinion that Reg3α (equivalent to Reg3A in human) is associated with the excellent plasticity of pancreatic cells and may play key roles in β-cell regeneration. As Liraglutide was recently reported inducing β-cell replacement in NOD mice [35], we compared the effectiveness of recombinant Reg proteins and Liraglutide in promoting β-cell neogenesis. Neither administration of rReg2 nor Liraglutide was effective in promoting β-cell regeneration using the same protocol of rReg3α treatment (Supplementary Figure S6). In another word, rReg3α is distinct from Liraglutide and other Reg isoforms and is valuable for further pharmaceutical investigation.
The main challenge for therapies of β-cell regeneration is how to generate functional β-cells. In this study, the newborn insulin-expressing cells were seemly far from functional as the insulin signal captured was extremely weak, even using a confocal microscope ( Figure 5A). The possibility that the faint insulin staining resulted from insulin internalization can be excluded because elevated endogenous Ins1 and Ins2 mRNA levels were detected. Although Ins1 mRNA level was significantly increased by 3.5 times ( Figure 5C), it should be noted that the control for the comparison was acinar cells which theoretically do not express insulin at all. Moreover, the increased amount of basal insulin secretion was neither comparable to the isolated islets nor responsive to high glucose stimulation ( Figure 5D). Thus, rReg3α seemly resembles EGF which plays a crucial role in promoting β-cell neogenesis, in which other supplemental factors are also required for β-cell maturation [19,36,37].
It is acknowledged that the terminally differentiated cells, rather than the stem cells, account for β-cell regeneration [15]. It is also reported that a group of cells in the pancreas regress to the immature state, creating a pool of progenitor cells with potential to differentiate [38]. In the pancreatic development, multipotent progenitor cells differentiate into "tip" cells, also known as unipotent acinar progenitors, and "trunk" cells, which are bipotent progenitors committed to endocrine and ductal fates [39]. In the present study, Stat3 phosphorylation and Ngn3 upregulation were detected in the cultured acinar cells ( Figure 5E-I), which is consistent with the references [19][20][21][22] and in support of rReg3α functioning through Stat3/Ngn3 pathway. Together, the islet β-cell loss in the STZ-treated mice (Figures 3 and 4) could lead acinar cells to dedifferentiate into the precursor cells [40,41] which further adapt endocrine fate by rReg3α via Stat3/Ngn3 signaling. Although it has been reported that acinar-to-beta cell transdifferentiation mainly occurs in centro-acinar cells [42], without genetic lineage tracing our circumstantial evidence does not firmly support the acinar-to-beta cell transdifferentiation. Even though most of the scattered insulin-producing cells are isolated from ductal structures, we still cannot exclude the possibility that duct-to-beta cell conversion occurs and then the cells migrate apart from the ducts [43]. The low frequency of auto-reprogramming that occurred in the STZ group ( Figure 4A,C) is likely caused by ablation of intercellular signals from islet β-cells [40]. Whether these newborn insulin-producing cells aggregate into the islets remains unknown.
As abovementioned that the neogenic β-cells were not potent enough in maintaining the blood glucose, we tested other protective mechanisms using various antibodies in immunofluorescent examination. We found that GRP78 expression was upregulated in the islets by rReg3α treatment in vivo and in vitro ( Figure 6G-K and Figure 7E,F). GRP78 is an ER molecular chaperone upregulated in the unfolded protein response (UPR), which is reported to protect against HFD-induced diabetes by the maintenance of β-cell function [44]. In islet β-cells, STZ can induce ER stress by a direct destructive effect and a secondary exhaust of β-cells owing to hyperglycemia. Elevated GRP78 production prevents protein aggregation and facilitates folding and degradation, thus relieving the ER stress. On the other hand, high insulin demand causes ER stress, subsequently activating UPR mechanism, leading to β-cell dedifferentiation [24]. The state of dedifferentiation confers the β-cells with a resistance to cell apoptosis [45,46]. Coincidently, Reg protein is also reported as a stress protein secreted and activated by trypsin cleavage. The polypeptides form insoluble complexes highly organized into fibrillar structure in ECM [47], which may be relevant to the upregulated GRP78 expression. Of note, Reg proteins are also reported to be involved in accumulation and activation of macrophage subsets [48,49]. Whether the protection on β-cells is related to immunocyte regulation deserves further study.
In summary, rReg3α administration protects islet β-cells and promotes β-cell regeneration. Upregulation of GRP78 expression and activation of Akt/Bcl-2/-xL signaling are involved in the protective mechanism. Although it is still a challenge to generate functional β-cells from terminally differentiated cells, our data suggest that rReg3α is a potential candidate for novel anti-diabetic drug development.

Cell Culture
Mouse insulinoma MIN6 cells were seeded at 3.0 × 10 3 /well in 96-well plates and cultured in DMEM (Wisent, Saint-Jean-Baptiste, QC, Canada) supplemented with 1% or 10% FBS (Wisent), 100 U/mL penicillin, and 100 µg/mL streptomycin (Wisent). The medium contained 4.5 g/L D-glucose in all experiments, except for starvation condition with 1% FBS and 1.0 g/L D-glucose. A total of 0-500 nM rReg3α was added and cells were harvested at 24, 48, and 72 h. Cells treated with PBS were set as the control. For cell cycle assay, MIN6 cells were seeded at 1.0 × 10 6 /well in 6-well plates and cultured with 1% FBS. Amounts of 10 and 100 nM rReg3α were added 24 h before PI staining using a cell cycle and apoptosis analysis kit (Beyotime, Shanghai, China). Cells were analyzed by flow cytometry and signals were captured by FACS Calibur (BD Bioscience, San Diego, CA, USA). Statistical analysis was performed using ModFit LT software (Verity Software House, Topsham, ME, USA). In the treatment of 100 nM rReg3α, cell lysates were collected at the indicated time points for Western blotting. Inhibitors against Akt (SC-66, Abcam, Cambridge, UK) and Erk (Nimbolide, Abcam) were added and cells were harvested at 24 and 48 h. Cell viability was determined by methyl thiazolyl tetrazolium (MTT) assay according to the manufacture's instruction (Sangon Biotech, Shanghai, China).
In apoptotic assay, MIN6 cells were seeded at 1.0 × 10 6 /well in 6-well plates with 10% FBS, and treated with 100 nM rReg3α for 12 h followed by a challenge with 10 mM STZ. After 12 h incubation, cells were harvested and labeled with Annexin-V-FITC and PI (Vazyme, Nanjing, China) for flow cytometry assay. Signals were captured by FACS Calibur and statistical analysis was performed using FlowJo 7.6.1 software (BD Bioscience). Cell lysates were collected at the indicated time points for Western blotting.

Western Blotting
Cells were harvested and ruptured using RIPA lysis buffer (Beyotime) containing 10 mM NaF, 1 mM Na 3 VO 4 , 1 mM PMSF, and proteinase inhibitor cocktail (Merck Millipore, Billerica, MA, USA). Cell lysates were separated by SDS-PAGE and electro-transferred to pure PVDF membranes (Merck Millipore) at 200 mA for 1 h. The blocked membranes were incubated with primary antibodies and then HRP-conjugated secondary antibodies (Table 1). Membranes were washed and incubated with hypersensitive ECL reagent (Merck Millipore), and luminescent signals were captured by a ChemiDoc XRS + System (Biorad, Hercules, CA, USA). Densitometric quantification was performed using AlphaEase software (Alpha Innotech, San Jose, CA, USA).  [13,48,49]. The α-cells percentage was quantified by counting the number of glucagon-positive cells then dividing by the total islet cell number.

Isolated Pancreatic Acinar Cells
The pancreatic acinar cells were isolated from 22-30 g male BALB/c mice using collagenase P (Roche) digestion, DNase I (Sangon Biotech), and soybean trypsin inhibitor (Sangon Biotech) [22]. Cells were resuspended in DMEM/F12 medium (Wisent) supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin, and 10% FBS. After an overnight incubation, "contaminated" cells adhered were discarded, and suspended acinar cells were collected and treated with 10 mM alloxan (Sangon Biotech) for 10 min. For lineage tracing, cells were labeled with 5 µg/mL WGA-FITC (Thermo Fisher Scientific, Waltham, MA, USA) for 24 h and then treated with 100 nM rReg3α. At d 3 and 5, cells were fixed in 4% paraformaldehyde and permeabilized in 0.2% Triton X-100 followed by an incubation with anti-insulin or anti-Ngn3 and Cy5-conjugated secondary antibodies (Table 1). Microscopic images were captured using a Zeiss LSM 5 Pascal laser scanning confocal microscope with 630× magnification.

Quantitative Real-Time PCR Assay (qRT-PCR)
Total RNA of the cells was extracted using Trizol reagent (Sangon Biotech), and cDNA was generated using an RT-PCR kit (Vazyme). qRT-PCR was performed using a ChamQ Universal SYBR qRT-PCR Master Mix kit (Vazyme) and QuantStudio 3 Real-Time PCR Systems (Applied Biosystems, Waltham, MA, USA). The certified primers synthesized (Sangon Biotech) are presented in Table 2. Table 2. Primers used in mRNA expression analysis.

Glucose-Stimulated Insulin Secretion (GSIS)
Isolated acinar cells were seeded at 1.0 × 10 6 /well in 6-well plates and treated with 100 nM rReg3α for 3 d. The cultured cells were transferred into KRBH buffer (KRB buffer supplemented with 10 mmol/L HEPES, 3 mg/mL BSA) containing 0, 2.8, or 22.2 mM glucose, in order, for 1 h each. Supernatant insulin content was determined by ELISA detection kit (Mercodia, Uppsala, Sweden) and the isolated mouse islets, 10 islets/well in 24-well plates, were set as the positive control.

Isolated Pancreatic Islets
The islets of Langerhans were isolated from 22-30 g male BALB/c mice using collagenase P digestion after common bile duct-duodenum ligation [11]. Isolated islets were handpicked individually using a stereoscopic microscope and cultured in RPMI 1640 medium supplemented with 20% FBS, 100 U/mL penicillin, and 100 µg/mL streptomycin overnight. Ten islets/well were placed in 24-well plates and treated with 100 nM rReg3α for 12 h, followed by an exposure to 4 mM STZ for another 12 h. The cultured islets were harvested and dissociated into single-cell suspension by an incubation with 5 µg/mL trypsin and 2 µg/mL DNase at 37 • C for 10 min. Dissociated cells were labeled with Annexin V-FITC and PI using apoptosis detection kit and flow cytometry was carried out. For proliferation assay, isolated islets were treated with 100 nM rReg3α for 24 h and BrdU incorporation was performed using BrdU labeling (Calbiochem, Darmstadt, Germany). GSIS and Western blotting were performed as described above.

Data Statistics
Data were presented as mean ± standard error (s.e.). Statistical analysis was performed by Student's t-test or one-way ANOVA using SigmaPlot 11.0 and GraphPad Prism 8.3.0. Significance was set at p < 0.05. Statistics of area under the curve (AUC) was performed by the trapezoidal method.