Loss of IDO1 Expression From Human Pancreatic β-Cells Precedes Their Destruction During the Development of Type 1 Diabetes

Type 1 diabetes (T1D) results from a breakdown of immune tolerance that leads to the selective destruction of β-cells in the pancreas, but the circumstances driving this dysfunction remain unclear. Indoleamine 2,3-dioxygenase-1 (IDO1) is a metabolic enzyme that catalyzes the first rate-limiting step of tryptophan catabolism, ultimately leading to the production of immunoregulatory molecules known as kynurenines. Its catalytic and noncatalytic effects are involved in the regulation of immunity (1), including the induction of tolerogenic dendritic cells (2) and regulatory T cells (3). However, IDO appears to be involved in selective immune regulation mechanisms, as IDO knockout mice do not develop a fulminant autoimmune phenotype (4). Interestingly, the dysregulation of the tryptophan metabolic pathway was suggested to contribute to the development of T1D in NOD mice (5–7).

A recent report from Orabona et al. (8) reveals that the majority of children with T1D have a defect in IDO1 expression in peripheral blood mononuclear cells. This defect is characterized by very low or absent levels of the protein IDO1. The same study reports that tocilizumab, a humanized interleukin-6 (IL-6) receptor antibody that blocks the IL-6 receptor, reverses this phenotype and controls hyperglycemia in NOD mice with overt diabetes (8). Therefore, the restoration of IDO1 immunoregulatory mechanisms may also be clinically beneficial in patients with T1D.

In light of these promising results, we investigated IDO1 expression in pancreata of individuals with T1D. We obtained pancreatic tissue sections from donors without diabetes and with diabetes collected by the Network for Pancreatic Organ Donors with Diabetes (nPOD) and from live patients with recent-onset T1D included in the Diabetes Virus Detection study (DiViD) (9) and systematically analyzed IDO1 and insulin expression by immunofluorescence assay. Although IDO1 was constitutively expressed in β-cells from donors without diabetes, it was nearly absent in insulin-deficient islets. Moreover, we observed that IDO1 was seldom to not expressed in certain insulin-containing islets from donors with multiple positive autoantibodies (AAb⁺) or with T1D, suggesting an impairment of IDO1 in the early stages of islet dysfunction. These findings could have important implications for the development of drugs able to target IDO1 expression in β-cells.

Pancreata were collected and processed by the nPOD and DiViD as previously described (9,10). Forty pancreatic sections from the tail region were analyzed: 8 donors without diabetes, 10 AAb⁺ donors with prediabetes, 6 patients with recent-onset T1D, 11 donors with T1D of longer duration, and 5 donors with type 2 diabetes (T2D) (Table 1). Tonsil control tissues were provided by the laboratory of Shane Crotty at La Jolla Institute for Allergy and Immunology. The La Jolla Institute for Allergy and Immunology Institutional Review Board approved all experimental procedures (protocol #DI3-054-0315).

Pancreas sections were subjected to a double-indirect immunofluorescence staining for IDO1 (clone 4.16 H1 [11]) and insulin (clone ICBTACLS). A detailed protocol is provided in the Supplementary Data. Alternatively, pancreas sections were subjected to a double-indirect immunofluorescence staining for IDO1 and CD11c (clone 2F1C10; 1:100, 1 h; Proteintech) or MHC class I (MHCI; clone EMR8-5; 1:200; Abcam).

All sections were scanned with an Axio Scan.Z1 slide scanner (Carl Zeiss), and images were acquired with the ZEN 2 slidescan module (Carl Zeiss).

Blinded samples were evaluated by two investigators (F.A. and N.G.). Thirty islets were randomly selected to account for heterogeneity of the sections (12,13). IDO1-positive area, colocalization of IDO1, and insulin-positive area or the percentage of IDO1-positive areas within the insulin-positive area were quantified with Image-Pro Premier software 9.1 (Media Cybernetics, Inc.). Additional details can be found in the Supplementary Data.

The ZEN 2 analysis module was used to determine IDO1- and insulin-positive areas in all of the islets of four case subjects (6073, 6267, 6247, and 6). Islet location and IDO1 percentage of positive area were then plotted as heat maps using MATLAB (MathWorks).

Data are presented as mean ± SD and analyzed using a one-way ANOVA or a two-tailed unpaired Student t test. P values were adjusted for multiple comparisons using the Bonferroni correction. Analyses were performed using Prism version 7 (GraphPad Software). A value of P < 0.05 was considered significant.

We selected a cohort that featured the different stages of the disease (i.e., prediabetes, recent-onset T1D, and T1D of longer duration). Mean age at time of tissue collection was not different between groups. As expected, the mean BMI of case subjects with T2D (35.4 ± 6.2) was higher than the mean BMI of those without diabetes, those who were AAb⁺, or those with recent-onset and longer-duration T1D (26.6 ± 4.5 vs. 25.8 ± 5.8 vs. 26.2 ± 3.2 vs. 24.4 ± 3.1, respectively) (Table 1). Among 17 case subjects with T1D, age at onset was heterogeneous (14–35 years old), 7 out of 15 donors were C-peptide negative (2 C-peptide values were unknown), and 6 out of 17 had no remaining insulin-containing islets.

IDO1 was detected in both exocrine and endocrine pancreas. We first investigated the localization of IDO1 in the endocrine pancreas. Insulin and IDO1 signals mostly overlapped, indicating that IDO1 was constitutively expressed by β-cells (Fig. 1A). IDO1 localization was confirmed with a second commercially available antibody (clone 10.1) (Supplementary Fig. 1). Furthermore, there were no statistical differences between groups in the localization of IDO1 (Fig. 1B), which confirmed that IDO1 was consistently expressed in β-cells, independently of the status of diabetes.

In the exocrine pancreas, IDO1 staining was notably dimmer than in the endocrine tissue. IDO1-positive cells were found at a very low density (≤1 cell/cm²) (Fig. 1C) and identified as CD11c-positive cells (Fig. 1D), presumably dendritic cells.

Next, the percentage of IDO1-positive area present in the islets was assessed. We observed that IDO1 was significantly less expressed in insulin-containing islets from donors with T1D (6.6 ± 4.5%) regardless of disease duration than in islets from donors without diabetes (18.7 ± 2.3%) or donors with T2D (15.0 ± 3.2%). Moreover, in insulin-deficient islets from donors with T1D, IDO1 was mostly absent (1.4 ± 1.5%) (Fig. 1E).

Next, we specifically assessed the expression of IDO1 in insulin-containing islets (percentage of IDO1 in the insulin-positive area) and discovered that IDO1 was heterogeneously expressed in insulin-containing islets (representative examples) (Fig. 2A). In order to visualize the distribution of IDO1 expression in the islets, heat maps showing insulin-deficient islets (purple dots) and the percentage of IDO1 in insulin-containing islets (gradient green to red) were created. In donors without diabetes and single AAb⁺ donors, IDO1 expression was high (>50%), whereas it was markedly reduced in individuals with T1D of longer duration (<20%). Interestingly, double AAb⁺ donors and patients with recent-onset T1D presented higher heterogeneity in IDO1 distribution. Both of these groups showed lobe-specific impairment of IDO1 expression (Fig. 2B), similar to the lobular pattern of β-cell loss in T1D.

Finally, in order to clarify at which stage of T1D IDO1 expression was impaired, the percentage of insulin-positive area and percentage of IDO1 in β-cells from all of the cases were quantified, and the results were displayed as scatter plots (Fig. 3A). The heterogeneity of IDO1 expression in double AAb⁺ donors and patients with recent-onset T1D (8–89 and 0–88% percentage of positive insulin area, respectively) was found to be substantially higher than in donors without diabetes, single AAb⁺ donors, donors with T1D of longer duration, or donors with T2D (48–98, 43–90, 1–58, and 40–88%, respectively), confirming observations from the heat maps. Moreover, we observed major differences in IDO1 expression depending on the antibody status and stage of disease. In islets from double AAb⁺ donors and patients with recent-onset T1D, a higher percentage (30.5 and 42%, respectively) of IDO1^low islets was observed when compared with donors without diabetes, single AAb⁺ donors, or donors with T2D (0, 10.6, and 16%, respectively). Interestingly, the scatter plots suggested that the loss of IDO1 occurred before T1D onset (Fig. 3A, middle left panel), whereas notably less insulin was expressed around the time of diagnosis (Fig. 3A, middle right panel). In donors with T1D who still had remaining insulin-containing islets, IDO1^negInsulin^pos islets were found, whereas IDO1^posInsulin^neg islets were not, which supported the idea of an early IDO1 loss. Finally, we compared MHCI hyperexpression in islets with IDO1 expression. We observed that although IDO1^low islets are more likely to hyperexpress MHCI, not all IDO1^low islets displayed MHCI hyperexpression (Fig. 3B).

IDO1, which leads the catabolism of tryptophan, is known to play multiple roles in the regulation of immunity through its antimicrobial effects and its activation of regulatory immune responses promoting immune tolerance (14). The enzyme therefore plays a role in controlling autoimmunity (15) and appears to be involved in several pathophysiological conditions, including autoimmune diseases (16). Interestingly, Orabona et al. (8) described a defect of IDO1 at the peripheral level in children with T1D. In light of these findings, we systematically investigated the pancreatic expression of IDO1 in patients with T1D.

IDO1 is expressed in various human tissues and cells, including antigen-presenting cells and regulatory T cells (11). In isolated rat islets, IDO1 mRNA was not constitutively expressed, and its transcription was only activated by interferon-γ and IL-1β in β-cells (17). In isolated human islets, PDX1-positive cells (presumably β-cells) and other endocrine cells showed a strong immunoreactivity to IDO1, which was enhanced when the islets were treated with interferon-γ (18). In this study, we report for the first time, using two antibodies specific for IDO1, that human endocrine tissue expresses IDO1 primarily in β-cells. Moreover, we described the presence of scarce IDO1-positive cells in the exocrine pancreas that are likely to be tolerogenic dendritic cells (19). Previous studies have described low plasma levels of tryptophan catabolites in NOD mice (5) and patients with T1D (20,21). In this study, for the first time, we show that the peripheral deficiency of IDO1 in human T1D is concomitant with low expression of IDO1 in insulin-containing islets and its quasi-absence in insulin-deficient islets in the pancreas.

These major findings call into question whether the absence of IDO1 is a cause or a result of β-cell dysfunction. We therefore investigated IDO1 expression in β-cells only and discovered major differences depending on the stage of disease. Indeed, in islets from donors without diabetes, single AAb⁺ donors, and donors with T2D, IDO1 expression was consistently high, whereas in islets from double AAb⁺ donors and case subjects with recent-onset T1D, heterogeneity was notably higher, indicating a shift in IDO1 expression around the time of T1D diagnosis. Our observations imply that IDO1 decay may occur in the preclinical phases of T1D and might precede the time of β-cell destruction. Thus, reverting IDO1 loss might prevent or delay T1D outcome, as reported by Zhang et al. (22) in a NOD mice model in which fibroblasts overexpressing IDO1 protected β-cells from destruction and reversed hyperglycemia. Moreover, Mondanelli et al. (23) have reported a protective and therapeutic effect of bortezomib, a proteasomal inhibitor that attenuates IDO1 proteasomal degradation, in NOD mice.

By nature, any human histopathological investigation using tissues from deceased organ donors will be cross-sectional. However, because T1D is pathologically a highly heterogeneous disease that gradually affects selected lobes of the pancreas, all stages of T1D can essentially be observed in a single organ section (heat maps in Fig. 2B). This allowed us to conclude that IDO1 was lost before the decline of insulin secretion.

Our observations raise important questions for the role of IDO1 in β-cells. Previous studies have shown that the enzyme can be involved in either immune or nonimmune events (24,25). In the pancreatic islets, it may be that the loss of IDO1, and thus tryptophan metabolites, weaken the immunomodulatory microenvironment and make the β-cells more prone to immune attacks by activating resident or infiltrating immune cells. Alternatively, the fact that IDO1 is constitutively expressed in β-cells could suggest that the enzyme has a prominent role in their physiology. These theories will need to be developed in further studies using pancreatic islet models.

Clinical trials to reverse T1D or prevent loss of residual β-cell function have had limited success so far. One reason could be that β-cell dysfunction contributes more to the disease (especially early on) than autoimmune attacks. A striking finding of the current study is the early impairment (prior to insulin decline) of intraislet expression of IDO1 within the pancreata of donors with prediabetes and donors with T1D. Considering the potential role of IDO1 in immune and nonimmune events, its impairment might be involved in the cascade, which leads to β-cell dysfunction. Future studies should use isolated human islets to better understand the role of IDO1, which will also aide the development of future targeted therapies.

Acknowledgments. The authors thank Zbigniew Mikulski, Sara McArdle, Yasaman Lajevardi, Ericka Castillo, and Priscilla Colby of La Jolla Institute for Allergy and Immunology for help with image acquisition, analysis, and administrative assistance, respectively. IDO1 mouse anti-human IDO1 antibody was obtained from Prof. Benoit Van den Eynde (Ludwig Institute for Cancer Research) through a material transfer agreement. This research was performed with the support of nPOD, a collaborative T1D research project sponsored by JDRF. Organ Procurement Organizations partnering with nPOD to provide research resources are listed at http://www.jdrfnpod.org//for-partners/npod-partners/. The authors also thank Ellie Ling (Eleanor Ling Medical Writing Services) for editorial assistance.

Funding. This research was performed with the support of nPOD, sponsored by JDRF International grant 25-2013-268. This study was also supported by National Institutes of Health/National Institute of Allergy and Infectious Diseases grant U01-AI102370-08.

Duality of Interest. B.V.d.E. is co-founder of and consultant for iTeos Therapeutics, a company involved in the development of IDO and tryptophan-2,3-dioxygenase inhibitors. M.G.v.H. is an employee of Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. F.A. designed, performed experiments, interpreted data, and wrote the manuscript. G.M. performed experiments and revised the manuscript. N.G. performed experiments and helped with the analyses. T.R.C. interpreted data and revised the manuscript. J.Z.G. assisted with the statistical analysis. L.K. and K.D.-J. collected patient material and revised the manuscript. K.D.-J. is principal investigator of the DiViD study. B.V.d.E. characterized and provided the IDO1 antibody. C.O. and U.G. revised the manuscript. M.G.v.H. designed experiments, interpreted data, and wrote the manuscript. M.G.v.H. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented at the 15th International Congress of the Immunology of Diabetes Society, San Francisco, CA, 19–23 January 2017, and the JDRF nPOD 9th Annual Scientific Meeting, Fort Lauderdale, FL, 19–22 February 2017.