Beta cell regeneration after single-round immunological destruction in a mouse model

Abstract

Aims/hypothesis: Achieving a better understanding of beta cell regeneration after immunological destruction is crucial for the development of immunotherapy approaches for type 1 diabetes. In previous type 1 diabetes models, sustained immune activation eliminates regenerating beta cells, thus limiting the study of the regenerative capacity of beta cells upon immunological destruction. Here, we employed an adeno-associated virus 8 (AAV8) vector for beta cell-targeted overexpression of a foreign antigen to induce single-round immunological destruction of existing beta cells.

Methods: Young and aged C57BL/6J mice were treated with AAV8 vectors expressing the foreign antigen luciferase. Islet inflammation and regeneration was observed at 3, 6, 10 and 22 weeks post-AAV delivery.

Results: In young C57BL/6J mice, robust humoral and cellular immune responses were developed towards antigen-expressing beta cells, leading to decreased beta cell mass. This was followed by beta cell mass replenishment, along with enhanced proliferation of insulin-positive cells, recruitment of nestin/CD34-positive endothelial cells, displacement of alpha cells and mobilisation of cytoplasmic neurogenin 3-positive cells. Mice with recovering beta cells showed normal or reduced fasting blood glucose levels and faster glucose clearance than controls. Although aged mice demonstrated similar responses to the treatment, they initially exhibited notable islet scarring and fluctuations in blood glucose levels, indicating that beta cell regeneration is slower in aged mice.

Conclusions/interpretation: Our hit-and-run, beta cell-targeted antigen expression system provides an opportunity to monitor the impact of single-round immunological beta cell destruction in animals with diverse genetic backgrounds or ageing status.

Fig. 1

Systemic administration of AAV8 vectors containing mIP2 facilitates beta cell-targeted transgene expression. (a) Schematic representation of AAV vectors. L-ITR, left inverted terminal repeat; R-ITR, right inverted terminal repeat; prom, promoter. (b) AAV8-mediated luciferase expression was assessed 14 days after i.p. injection (control n = 3, AAV8-Luc n = 3). Luminescence scale bar, ×10⁴. (c) EmGFP expression was assessed in pancreatic sections EmGFP (n = 2). Selective EmGFP expression (green) was observed in insulin-positive beta cells (red), but not in acinar cells. Nuclei were counterstained with DAPI (blue). Scale bars, 100 μm

Fig. 2

Beta cell-targeted luciferase expression induces specific humoral and cellular immune responses. (a) Luciferase expression was monitored 3, 6 and 10 weeks p.i. (b) RT-PCR analysis of total pancreatic lysates for firefly luciferase transcripts in saline controls and AAV8-Luc (white and black bars, respectively); data represents fold change vs age-matched controls set as 1; logarithmic scale, base 10; ***p < 0.001 vs controls. (c) Western blot detection of anti-luciferase antibodies in mouse plasma samples. (d) Lysates from 293T cells overexpressing luciferase and AAV8 capsid proteins and from uninfected controls were used as an antigen source to detect antibodies against AAV8 capsid. Plasma samples were from mice 6 weeks p.i. (n = 4). Anti-VP3 antibody was used as a control. (e) An ELISpot assay was performed to detect cytotoxic T cells from splenocytes of control (white bars) and AAV8-Luc (black bars) infected mice that react to the luciferase epitope peptide LMYRFEEEL. Ovalbumin (OVA) peptide was used as a negative control (grey bars). Splenocytes were analysed after 3, 6, 10 and 22 weeks. Representative wells and the means ± SEM of spot-forming units (SFU) from four mice per group per time point are shown. Logarithmic scale, log base 10. (f) Representative islets used for insulitis scoring with an anti-CD45 antibody at 4 weeks p.i. The table shows the number of random islets counted and the average (Ave.) intra- and peri-islet CD45 cells (n = 3 per group). (g) Serial pancreatic sections were stained with antibodies against the immune cell markers CD11b, F4/80, CD4 and CD8 (green). Nuclei were counterstained with DAPI (blue). Scale bars, 50 μm. INS, insulin; Luc, luciferase

Fig. 3

Beta cell-targeted immune responses result in a transient reduction in beta cell mass followed by robust beta cell regeneration. (a) Pancreatic islets were visualised by staining with anti-glucagon (GCG) and anti-insulin (INS) antibodies. Images of the islets at 6 weeks p.i. are shown. Scale bar, 100 μm. (b) Insulin-positive area as a percentage of the total pancreatic area in treated and control mice (n = 4 for each time point; *p < 0.05, **p < 0.01). (c) Representative pancreatic islets from AAV8-Luc treated or control mice stained with anti-insulin and anti-Ki67 antibodies. Insulin-positive and Ki67-positive islet cells from AAV8-Luc treated or control mice were counted, and the percentage of insulin/Ki67 double-positive cells was determined. Data are presented as means ± SEM. Scale bar, 100 μm. *p < 0.05 vs control. (d) Groups of single insulin-positive and Ki67-positive cells detected at 10 weeks p.i. are shown. Scale bars, 100 μm. (e) Fasting blood glucose levels were monitored (n = 4 per group). *p < 0.05, **p < 0.01. (f) Glucose tolerance test on control and treated mice (n = 4 per group) at 6 weeks p.i. *p < 0.05. (g) Fasting plasma insulin levels were determined by ELISA at 3 and 10 weeks p.i. (n = 3 per group per time point). (b), (c), (e–g) white bars/squares, saline controls; black bars/squares, AAV8-Luc treatment

Fig. 4

Displacement of glucagon-positive cells and activation of NGN3-positive cells following immunological beta cell damage. (a) The location of insulin (INS)-positive and glucagon (GCG)-positive cells was determined by staining control and immunologically damaged islets with anti-insulin (green) and anti-glucagon (red) antibodies at 3, 6 and 10 weeks p.i. (b) NGN3 expression was detected using an anti-NGN3 antibody in age-matched control islets at 3 and 6 weeks p.i.. Note cytoplasmic (arrows) and nuclear localisation (arrowheads) in normal islets. (c) Immunostaining demonstrated both nuclear (arrowheads) and cytoplasmic (arrow) NGN3 expression in age-matched control islets near ducts. (d) NGN3 expression was analysed in islets from AAV8-Luc-transduced mice at 3, 6 and 10 weeks p.i. Insulin expression was detected using an anti-insulin antibody. (e) Islets from AAV8-Luc-transduced mice at 6 weeks p.i. were analysed for NGN3 and PDX1 (green) expression. Note induction of PDX1 expression in ductal lining cells (arrows). (f) Representative islets showing NGN3, PDX1 and glucagon localisation after immune response. Nuclear PDX1-positive cells (arrows) were identified in ductal lining cells. (g) Immunostaining of the region between recovering islets and ducts with anti-NGN3 and glucagon antibodies. (h) A representative image of single insulin-positive cells (green) identified at 10 weeks p.i. is shown with anti-NGN3 co-staining (red). (a–h) Dashed lines correspond to ductal epithelium. Scale bars, 50 μm; w, weeks. (i–p) Quantitative RT-PCR analysis of pancreatic transcripts for common islet- and beta cell-associated factors. Results show AAV8-Luc transcripts (black bars) relative to age-matched control (white bars) pancreas, set as 1; logarithmic scale, base 10. Data are presented as means ± SEM, representing four mice per group per time point. *p < 0.05, **p < 0.01, ***p < 0.001; INS, insulin

Fig. 5

Induced immunological beta cell destruction leads to beta cell regeneration in aged mice. (a) Mice received AAV8-Luc vector through i.p. injection. Immunohistochemistry was performed at 6 and 10 weeks p.i. Scale bars, 50 μm. (b) Percentage insulin-positive area was determined as described in Fig. 3b. (c) Beta cell proliferation was determined as described in Fig. 3c. Images of islets at 6 weeks are shown. Beta cell proliferation was determined by quantification of total insulin- and Ki67-positive cells in five random islets from age-matched controls (n = 4 per time point) and AAV8-Luc-transduced mice (n = 4 per time point). *p < 0.05 vs control. (d) A representative islet in AAV8-Luc-treated mice at 10 weeks p.i., showing a dense population of Ki67-positive cells with a few isolated insulin-positive and Ki67-positive cells. (e) Fasting blood glucose levels were determined over the 10 week period for treated (n = 4) and control mice (n = 4). **p < 0.01 vs control. (f) Glucose tolerance tests were conducted at 5 weeks p.i. when vector-treated mice showed higher fasting blood glucose levels (see Fig. 5e). GCG, glucagon; INS, insulin; (b, c, e, f) white bars/squares, saline controls; black bars/squares, AAV8-Luc treatment