Insulitis and β-Cell Mass in the Natural History of Type 1 Diabetes

Abstract

Descriptions of insulitis in human islets throughout the natural history of type 1 diabetes are limited. We determined insulitis frequency (the percent of islets displaying insulitis to total islets), infiltrating leukocyte subtypes, and β-cell and α-cell mass in pancreata recovered from organ donors with type 1 diabetes (n = 80), as well as from donors without diabetes, both with islet autoantibodies (AAb(+), n = 18) and without islet autoantibodies (AAb(-), n = 61). Insulitis was observed in four of four donors (100%) with type 1 diabetes duration of ≤1 year and two AAb(+) donors (2 of 18 donors, 11%). Insulitis frequency showed a significant but limited inverse correlation with diabetes duration (r = -0.58, P = 0.01) but not with age at disease onset. Residual β-cells were observed in all type 1 diabetes donors with insulitis, while β-cell area and mass were significantly higher in type 1 diabetes donors with insulitis compared with those without insulitis. Insulitis affected 33% of insulin(+) islets compared with 2% of insulin(-) islets in donors with type 1 diabetes. A significant correlation was observed between insulitis frequency and CD45(+), CD3(+), CD4(+), CD8(+), and CD20(+) cell numbers within the insulitis (r = 0.53-0.73, P = 0.004-0.04), but not CD68(+) or CD11c(+) cells. The presence of β-cells as well as insulitis several years after diagnosis in children and young adults suggests that the chronicity of islet autoimmunity extends well into the postdiagnosis period. This information should aid considerations of therapeutic strategies seeking type 1 diabetes prevention and reversal.

Figure 1

Lobular variability in insulin⁺ islets and insulitis. Islets were imaged from a 13-year-old patient with T1D for 5 years (nPOD 6243). Serial paraffin sections were stained for Ki67 plus insulin (INS) (A, C, and E) and CD3 plus glucagon (GCG) (B, D, and F), and islets were subtyped as described in research design and methods. Five insulin⁺ islets (A) are seen in a lobule adjacent to a lobule with two insulin⁻ islets in the patient with T1D (A and B, blue arrows). Three insulin⁺ islets had insulitis (B, black arrows), and both insulin⁻ islets did not have insulitis. One of the insulin⁺ islets with insulitis is shown at higher magnification (C and D) as well as an insulin⁻ insulitis⁻ islet (E and F). Few islet cells were Ki67⁺ (A and C), indicating no effect of insulitis on proliferating cell numbers. Scale bars: A and B, 500 μm; C–F, 50 μm.

Figure 2

Heterogeneity of islet β-cells and the numbers of CD3⁺ cells in insulitic islets in young donors at diabetes onset and after 5 years of disease duration. Representative islets were imaged for a 13-year-old at diabetes onset (nPOD 6228; A–F) and for another 13-year-old with T1D for 5 years (6243; G–L). Serial paraffin sections were stained using dual-IHC (Ki67 plus insulin [INS], CD3 plus glucagon [GCG]), as described in research design and methods. Decreasing proportions of β-cells are shown in different islets from both donors (A–C and G–I) as well as heterogeneity in numbers of CD3⁺ cells per islet (D–F and J–L). Heterogeneity in CD3⁺ counts between two insulin⁻ islets of similar sizes is also shown for one donor (K and L). Scale bars: A–F, H, and K, 50 μm; G, I, J, and L, 100 μm.

Figure 3

Insulitis frequency in relation to diabetes duration and donor ages at demise or disease onset. The insulitis frequency (%, total number of insulitic islets/total number of islets) is shown on the y-axis in comparison with years for diabetes duration (A), age at T1D onset (B), and age at demise (C). Data were displayed (pink) for the two AAb⁺ donors with insulitis for age at demise but were excluded from statistical analyses because of the small sample size. Insulitis frequency (%) had a low but significant correlation to diabetes duration, but not to donor age at demise or disease onset. The linear regression line with Spearman r and P values are shown.

Figure 4

Insulitis leukocyte subtyping by multiple immunofluorescence. Serial sections of an islet with insulitic aggregate from a 5-year-old donor with T1D for 4 months (nPOD 6209) were imaged. The numbers of leukocytes/islet were counted for total leukocytes (CD45, A), B and T lymphocytes (CD20 and CD3, B), T-lymphocyte subsets (CD8 and CD4, C), and dendritic cells and macrophages (CD11c and CD68, D) as described in research design and methods. Islet endocrine cells were identified using glucagon (GCG) and insulin (INS) stains. The merged images (A–D) include DAPI-stained DNA, and the individual fluorescence channels are shown below the merged images. Scale bars: 50 µm.

Figure 5

Leukocyte numbers increase in parallel with CD45⁺ and CD3⁺ numbers but are heterogeneous between islets. Numbers of CD45⁺, CD3⁺, CD20⁺, CD8⁺, and CD4⁺ cells were counted for each insulitic islet and averaged per donor (n = 13 donors with T1D, 2–10 islets/donor). Data were also displayed (pink) for each leukocyte marker for the two AAb⁺ donors with insulitis for comparison purposes but were excluded from statistical analyses because of the small sample size. Correlations between the numbers of leukocytes per islet were performed by Spearman analysis with linear regression lines shown. The numbers of CD45⁺ cells per islet were significantly correlated with the numbers of CD3⁺ and CD20⁺cells per islet (A). The numbers of CD3⁺ cells per islet were also significantly correlated with the numbers of CD8⁺ and CD4⁺ cells per islet (B). The numbers of leukocyte subtypes per islet were also analyzed to examine the variability in cell numbers among all islets for donors with T1D and two AAb⁺ donors with insulitis, with means depicted by box-and-whisker plots with Tukey test error bars (nonparametric) (C). Numbers in parentheses following each marker indicate the total number of individual islets analyzed for each leukocyte marker (CD45) or set of markers. Variability in the overall numbers of leukocytes per islet was observed. The numbers of leukocytes per islet had similar trends in AAb⁺ donors without diabetes.

Figure 6

Pancreas weight and β-cell area and mass are decreased in donors with T1D. Group means were depicted by box-and-whisker plots with Tukey test error bars (nonparametric, outliers shown as black circles). Data were displayed (pink) for the two AAb⁺ donors with insulitis and were excluded from statistical analyses because of the small sample size (n = 2). Pancreata weights were significantly lower in donors with T1D compared with AAb⁻ and AAb⁺ donors without diabetes, and no difference was detected between donors with T1D based on insulitis status (A). The β-cell area was significantly lower in both donor groups with T1D compared with AAb⁻ and AAb⁺ donors without diabetes (B). T1D donors with insulitis had a significantly higher β-cell area compared with T1D donors without insulitis (B). The β-cell mass was similar between AAb⁻ and AAb⁺ donors and was significantly different from that in groups of donors with T1D (C). T1D donors with insulitis had significantly higher insulin⁺ β-cell mass compared with T1D donors without insulitis (C). The α-cell area (D) and mass (E) were not significantly different between groups. *Significantly different from AAb⁻ and AAb⁺ donors (P < 0.001). **Significantly different from AAb⁻ and AAb⁺ donors (P < 0.001) and T1D donors without insulitis (P < 0.05).