FoxO1 gain of function in the pancreas causes glucose intolerance, polycystic pancreas, and islet hypervascularization

Abstract

Genetic studies revealed that the ablation of insulin/IGF-1 signaling in the pancreas causes diabetes. FoxO1 is a downstream transcription factor of insulin/IGF-1 signaling. We previously reported that FoxO1 haploinsufficiency restored β cell mass and rescued diabetes in IRS2 knockout mice. However, it is still unclear whether FoxO1 dysregulation in the pancreas could be the cause of diabetes. To test this hypothesis, we generated transgenic mice overexpressing constitutively active FoxO1 specifically in the pancreas (TG). TG mice had impaired glucose tolerance and some of them indeed developed diabetes due to the reduction of β cell mass, which is associated with decreased Pdx1 and MafA in β cells. We also observed increased proliferation of pancreatic duct epithelial cells in TG mice and some mice developed a polycystic pancreas as they aged. Furthermore, TG mice exhibited islet hypervascularities due to increased VEGF-A expression in β cells. We found FoxO1 binds to the VEGF-A promoter and regulates VEGF-A transcription in β cells. We propose that dysregulation of FoxO1 activity in the pancreas could account for the development of diabetes and pancreatic cysts.

Figure 1. Hyperglycemia and impaired glucose tolerance in TG mice.

(A) Blood glucose levels (mg/dl) in fasted and ad-lib fed TG mice and their control littermates at 8 weeks of age. (B) Plasma insulin levels (ng/ml) before and 15 min after glucose (1.2 g/kg) intraperitoneal injection. (C) Intraperitoneal glucose tolerance tests (IPGTT) in male (left) and female (right) TG mice that had normal blood glucose levels in fasted and fed states at 8 weeks of age. (D) Insulin tolerance test (ITT) in male (left) and female (right) TG and control mice at 12 weeks of age. (E) Plasma insulin (left) and glucagon (right) levels in fasted and fed conditions in TG and control mice at 12 weeks of age. (F) The size-matched islets were isolated from male TG and control mice at 12 weeks of age, and glucose-stimulated insulin secretion levels were measured in those islets. The data were normalized by DNA content in those islets (G and H) TG mice that had normal blood glucose levels at 4 weeks of age and their control littermates were fed a HFHSD for 8 weeks further, then, fasted blood glucose levels were measured (G) and IPGTT performed (H) in both male (left) and female (right) mice. In each experiment (A–H), at least six male or female mice for each genotype were analyzed. In the line graphs, closed triangle indicates TG mice and open circle indicates control mice. Data represent mean ± SEM. * or ** indicates P<0.05 or P<0.01 by ANOVA.

Figure 2. Reduced β cell mass and decreased expression of Pdx1 and MafA in TG mice.

(A) Double immunohistochemistry with anti-insulin and anti-glucagon antibodies were conducted in pancreatic sections from TG and control mice. Representative images of low (upper panels) and high (lower panels) magnifications are shown. (B) The % area of β cells (left panel) or α cells (right panel) vs. whole pancreatic area was scored as described in Methods. (C and D) Double immunostaining of insulin with Pdx1 (C) or MafA (D) were conducted in pancreatic sections from TG and control mice. Arrows indicate individual cysts in TG pancreas. Representative images are shown. (E) Islets were isolated from TG and control mice and used for analyses by real time RT-PCR. In each experiment (A–E), at least six male and six female mice of each genotype and ten sections per mouse were analyzed. All the results were normalized using GAPDH. There was no significant difference between male and female mice. Data represent mean ± SEM. An asterisk indicates P<0.05 by ANOVA.

Figure 3. Enhanced proliferation of pancreatic duct epithelial cells and development of polycystic pancreas in TG mice.

(A) Immunohistochemistry with anti-Ki67 antibody (red) along with Lectin-DBA (green) was conducted in pancreatic sections from 2 month old TG and control mice. Representative results are shown. (B) Photographs of the pancreas in 12 month old TG and control mice. Representative pictures are shown. (C) Hematoxylin and eosin (HE) staining (left panel indicates ×40 magnification and right ×400) in pancreatic sections from 12 month old TG mice. Representative images are shown. (D) Immunohistochemistry with anti-Ki67 antibody (red) along with Lectin-DBA (green) was conducted in pancreatic sections from 12 month old TG mice. Representative images of pancreatic cysts are shown. Arrow indicates Ki67 positive cells.

Figure 4. Islet vasculature and VEGF-A expression in β cells are increased in TG mice via FoxO1 regulation of VEGF-A transcription.

(A) Hematoxylin and eosin (HE) staining and immunohistochemistry with anti-PECAM1 or anti-VEGF-A antibodies in pancreatic sections from 2 month old TG and control mice. Representative images are shown. (B) Islets were isolated from TG and control mice and used for analyses by real-time RT-PCR to quantify PECAM1 and VEGF-A mRNA levels. In each experiment (A and B), six male and six female mice were analyzed for each genotype. There was no difference between male and female mice. Data represent the mean ± SEM fold-increase relative to control mice. An asterisk indicates P<0.05 by ANOVA. (C) β TC3 or MIN6 cells were infected with adenoviruses expressing FoxO1 ADA or GFP at indicated MOI, 24 hrs later mRNA was isolated from the cells and used for real-time RT-PCR to quantify VEGF-A mRNA levels. The results of (B) and (C) were normalized using GAPDH. (D) VEGF-A promoter driven luciferase activities were measured in MIN6 cells infected with adenoviruses expressing FoxO1 ADA or GFP at indicated MOI. For (C) and (D), data represent the mean ± SEM fold-increase relative to GFP infection. An asterisk indicates P<0.05 by ANOVA. (E) The promoter region of the mouse VEGF-A gene. Square indicates the location of the forkhead responsive element (FHRE). For ChIP assays, a primer set spanning three FHREs and control primers were used. An arrow indicates the initiation site of RNA synthesis and is designated +1. (F) ChIP assays of the VEGF-A promoter are shown. We performed ChIP assays in MIN6 cells with indicated antisera. NRS indicates normal rabbit serum used as a control antisera. The crosslinked DNA was amplified by PCR using FHRE and Cont primers. Input represents DNA extracted from chromatin prior to immunoprecipitation. We used input DNA with 1∶100 dilution.

Figure 5. Enhanced VEGF-A expression in β cells but not α cells in TG mice.

(A) Double immunostaining of VEGF-A with insulin or glucagon in pancreatic sections from TG mice and control mice. Enhanced VEGF-A staining (red) in TG mice is completely merged with insulin staining (green to yellow) but not with glucagon staining (green). (B) Quantitative analysis of VEGF-A staining levels in the islets of TG and control mice. We measured the staining levels of VEGF-A per unit area by using NIS-Elements (Nikon) and Image-J (NIH). Six mice were analyzed for each genotype and representative images are shown.

Figure 6. Islet vascularity and VEGF-A expression in HFHSD fed mice and db/db mice.

(A) Immunohistochemistry with anti-PECAM1 or anti-VEGF-A antibodies in pancreatic sections from HFHSD fed mice, db/db, and control mice. Six male mice from each genotype and six sections per mouse were analyzed. Representative images are shown. (B) Islets were isolated from HFHSD fed mice, db/db, and control mice and used for analyses by real-time RT-PCR to quantify PECAM1 and VEGF-A mRNA levels. Six male mice were analyzed for each genotype. The results were normalized using GAPDH. Data represent the mean ± SEM fold-increase relative to control mice. An asterisk indicates P<0.05 by ANOVA.