Alpha cells are endocrine cells in mammalian islets of the pancreas, which comprise 20% of the human islet (32). Alpha cells play an important role in regulating glucose homeostasis in human daily activities by synthesizing and secreting the hormone glucagon to elevate glucose levels in the blood (32). Further evidence has demonstrated that the role of glucagon appears to be vital for hyperglycemia to arise in T1D (33, 34). Interestingly, the alpha cell population is preserved in the early and late stages of autoimmune diabetes. We and other groups have reported that alpha cell mass increases through proliferation when beta cell numbers diminish in islet centers during the development of diabetes, while blood glucose levels begin to surge (6–8, 24, 31). Thus, a more efficient approach to treat T1D is reprogramming alpha cells to insulin-producing cells at the early onset of diabetes, increasing the transformed alpha cell number (following transformation) and proliferation.
T1D diabetic islets have been shown to dramatically decrease in functioning beta cells that coincide with the remaining alpha cells occupying the islet centers (3–5); however, the mass of alpha cells is also affected by invading immune cells in progressive T1D. This interesting phenomenon was observed in our experiment (Fig. 1). Based on this finding, we designed a therapeutic strategy to infect native pancreatic alpha cells using an AAV8-GCG-PM viral delivery method that requires beta cell essential genes (Padx1 and MafA) to reprogram alpha cells into insulin-producing cells. Following disease development, alpha cells proliferate, and virally infected alpha cells carrying the Pdx1 and MafA genes coincide to proliferate, which makes reprogramming more effective (Fig. 1G Schematic) and prevents the development of severe disease. In animal experiments, a common approach is to target glucagon-secreting alpha cells in vivo by crossing mice expressing Cre recombinase under the glucagon promoter (Gcg-Cre mice) with loxp reporter mice, which have a loxP site transcriptional STOP sequence upstream of the fluorescent protein gene sequence (24, 35). However, these methods are time consuming due to breeding and limited to mouse strains. We demonstrated that a combination of the sox9 promoter and AAV serotype 6 is able to deliver the GFP gene in pancreatic ductal cells (22). Based on this finding that the combination of AAV serotype tissue tropism with a specific gene promoter, together with ductal infusion of the virus, leads to successful in vivo gene delivery and labeling of pancreatic ducts, we could be able to deliver a reporter gene to target alpha cells and deliver Pdx1 and MafA genes specific to alpha cells. First, we tested beta-cell toxin alloxan (ALX)-injected C57BL/6J mice, which developed diabetes one week after degradation of the beta cells of pancreatic islets (31). We found that ALX induced diabetes with a modest blood glucose increase but a significant increase in alpha-cell mass, which is consistent with our previous observations (31). The AAV8 virus carrying Pdx1-MafA genes was delivered into mouse pancreatic ducts of ALX-induced diabetic C57BL/6J mice by retrograde infusion of 150 µl of AAV8-GCG-PM (22, 23, 25–28). From our results, we found that the short GCG promoter and AAV8 combination was the best for targeting mouse alpha cells. These results demonstrated that the human short GCG promoter is specific and sufficient to drive gene expression in vivo. Here, we demonstrated that AAV8 in combination with the human alpha cell promoter is accomplished by delivering transgenes in pancreatic alpha-cells in vivo and that the Pdx1 and MafA genes under the human alpha cell promoter are able to normalize T1D mice either in chemically induced or autoimmune-induced diabetic mice. We first showed that Pdx1 and MafA expression in vivo was able to correct hyperglycemia in both ALX-induced diabetes and autoimmune diabetic NOD/ShiLtJ mice, implying that beta cell-like reprogramming occurred in alpha cells. Among all AAV serotypes, we found that 8 and 6 were the best for infecting mouse pancreatic cells (22, 23). We chose the AAV serotype 8 vector since we found that serotype 8 had a better infection efficiency in mouse islet cells than serotype 6 (23). AAV serotype 6 infects mouse pancreatic duct cells, while serotype 8 does not, but duct cells were not the focus of the current study. Both serotypes infect mouse acinar cells well; however, combining the human GCG promoter increases tighter control of the tropism of AAV8, which could specifically infect alpha cells.
We designed and constructed various sizes of the human GCG promoter for assembly in the pAAV vector combined with the GFP reporter gene. Based on our previous work, we found that glucagon regulates its own synthesis by autocrine signaling (36) and infected the alpha TC1 clone 9 cell line with the AAV8-GCG-GFP virus and AAV8-GCG-Pdx1-MafA. We found that GFP was expressed in alpha TC1 clone 9 cells 3 days after infection. After infusion into murine pancreases, we observed specific GFP + expression in mouse alpha cells with little overlap in insulin-positive cells but almost no GFP expression in acinar cells. In the AAV GCG promoter-driven Cre virus used to infuse into the pancreas of ROSA26 tomato reporter mice, we observed 85 to 90% Cre cleaved red fluorescence in the nuclei of alpha cells. This result indicates that the GCG promoter drives the Cre gene to determine the function of cleaving the loxp sites in the alpha cells of ROSA26 tomato reporter mice, which further confirms the GCG promoter specificity. Even though this short human GCG is only 648 bp, after analyzing the results of a BLAST search, we found that the promoter sequence is 83% and 82% similar to the mouse and rat promoters, respectively, in GenBank. Our experimental results indicated that the short GCG promoter is sufficient to drive gene expression and with specificity for identifying alpha cells in the mouse pancreas.
As we observed and other groups reported that alpha cells are proliferated at early onset of diabetes and dominating in islet following the disease development, with the GCG promoter we successful, at this appropriate diabetes early onset stage, to deliver insulin producing essential transcription factors Pdx1 and MafA specific in alpha cells and reprogram the infected alpha cells into insulin producing cells. The time to trigger alpha cells to reprogram into insulin-producing cells is critical for the treatment of diabetes, especially for NOD/ShiLt mice, as they give impetus to diabetes very quickly once disease onset. The overexpression of Pdx1 and MafA in NOD/ShiLt islets completely protected against the development of spontaneous autoimmune diabetes. Additionally, in contrast to their diabetic NOD/ShiLt with AAV8-GCG-GFP control, NOD-AAV8 GCG Pdx1 and MafA animals had normal levels of serum triglycerides, free fatty acids, and beta-hydroxybutyrate, indicating that they have normal energy metabolism. AAV8 combined with GCG promoter-mediated Pdx1 and MafA gene transfer to the pancreas protected NOD/ShiLt mice against the development of autoimmune diabetes, as evidenced by the significant reduction in the incidence of spontaneous diabetes in the animals that received the therapeutic vector. Our results demonstrated that two transcription factors, Pdx1 and MafA, assembled in an AAV8 viral vector with a human GCG promoter through a retrograde infusion technique are able to specifically reprogram pancreatic alpha cells into insulin-producing beta-like cells, and blood glucose levels in autoimmune NOD/ShiLt diabetic mice can be normalized. AAV8-GCG-Pdx1-MafA-treated mice also showed preservation of beta-cell mass and normal levels of circulating insulin. Here, we showed that Pdx1 and MafA expression under the GCG promoter driven in vivo was able to correct hyperglycemia in both ALX-induced diabetes and in autoimmune diabetic NOD/ShiLt mice, signifying that beta cell-like reprogramming was occurring, and this is not suppression of the GCG promoter in alpha cells, as we saw glucagon expression is still normal, as AAV is not a genome integration virus. With this cell reprogramming, we did not observe significant immune cells in the AAV-GCG-Pdx1-MafA-infused islets in NOD/ShiLt mice. We currently do not know the reason, but the possibility is that ductal viral infusion with AAV GCG promoter-driven Pdx1 and MafA altered the autoimmunity of NOD/ShiLt mice, leading to extended survival of neogenic insulin cells. It is also possible that AAV8-GCG-Pdx1-MafA-infected alpha cells are not immune cells recognized, as they do not produce mature insulin at the preprogramming stage. Once the alpha cells reprogrammed into insulin-producing cells, they maintained glucagon production, which may confuse the immune system of NOD/ShiLt mice. In addition, the GCG promoter is significantly different from the insulin promoter; for example, the transcription factor ATF4 binds to the insulin promoter, causing beta cell dysfunction and apoptosis (37, 38), and activated ATF4 regulates cell immunity (39), which may play an attractive role in immune cell homing to islets. The mechanisms by which Pdx1 and MafA reprogram alpha cells into insulin-producing cells with protective action on transgenic islets or after AAV8-GCG-Pdx1 and MafA-mediated gene delivery have not been fully elucidated. The next research aim was to establish stable Pdx1 and MafA gene integration into the alpha cell glucagon gene locus to further study the mechanism of alpha to beta transdifferentiation.
In summary, we report here that the human short promoter-driven reporter GFP gene could target pancreatic alpha cells specifically in the mouse pancreas. The short human GCG promoter is enough to drive two transcription factors, Pdx1 and MafA, and combination with AAV serotype 8 is capable of normalizing blood glucose either in chemically induced or autoimmune diabetes in animals. We reported that normal blood glucose restoration and beta cell survival in NOD/ShiLt mice by AAV8-CMV drive Pdx1, and MafA appears to last 4 months prior to re-establishment of autoimmunity, but in our GCG promoter driven Pdx1 and MafA, euglycemia of NOD mice lasts 7 months and without immune cells found in the neogenic islets. This may be the motivation for us to better use the GCG promoter driving Pdx1 and MafA gene reprogramming alpha cells into insulin-producing beta cells for T1D.
Four weeks after infusion, we observed blood glucose normalization in ALX-induced diabetic mice, and the new insulin-positive cells were found to be derived exclusively from alpha cells. Similarly, we observed that the blood glucose of hyperglycemic NOD/ShiLt mice became normal at approximately 4 weeks and lasted more than 7 months to the end of the experimental design date without immune cell infiltration in islets. Therefore, we conclude that this alpha cell-specific promoter in combination with AAV 8 serotype viral reprogramming provides the vision to develop a novel therapy that can potentially be translated to the clinic for the treatment of T1D.