Pancreatic β‐cells express major histocompatibility complex class II: Do diabetic β‐cells have the capacity of antigen‐presenting cells?

Abstract

Russell et al. showed that major histocompatibility complex class II is expressed in pancreatic β‐cells by a highly accurate method called transcriptome analysis. Pancreatic β‐cells might have the capacity as antigen‐presenting cells.

Genetic factors, environmental factors (such as viral infections) and autoimmunity contribute to pancreatic β‐cell destruction in type 1 diabetes. As genetic factors, major histocompatibility (MHC) class II gene polymorphisms have been shown to contribute most strongly to susceptibility to type 1 diabetes, and disease susceptibility haplotypes for each race have also been identified. MHC class II molecules (human leukocyte antigen‐DR, DP, DQ isotypes) are selectively expressed on antigen‐presenting cells, such as dendritic cells and macrophages, and present the antigenic peptide to CD4⁺ T cells. On the contrary, MHC class I molecules, which are expressed on nearly all nucleated cells, present antigenic peptide to CD8⁺ T cells.

Immunohistochemical studies on pancreatic tissues from patients with type 1 diabetes obtained by autopsy or surgical biopsy have consistently shown overexpression of MHC class I on pancreatic β‐cells and infiltration of T cells into the islets. Therefore, at present, it is believed that type 1 diabetes is triggered by environmental factors, such as viral infection, resulting in local inflammation and antigen presentation by dendritic cells and macrophages through MHC class II, leading to the eventual destruction of pancreatic β‐cells by CD8⁺ T cells through MHC class I.

MHC class II expression is closely regulated by class II MHC transactivator (CIITA), and interferon gamma (IFNγ) has an important role in the upregulation of CIITA. It is considered that expression of MHC class II is limited to antigen‐presenting cells, such as dendritic cells, macrophages and B lymphocytes. However, it has been reported that non‐antigen‐presenting cells; for example, thyroid follicular cells, cardiomyocytes and vascular endothelial cells, can also express MHC class II by IFNγ in the local site of inflammation1.

Similarly, Bottazzo and Hanafusa et al.2 suggest that although MHC class II is not expressed on normal pancreatic β‐cells, it is expressed on β‐cells from patients with type 1 diabetes, and pancreatic β‐cells might have the ability to present antigens. However, because the presence of MHC class II on pancreatic β‐cells was shown by immunostaining in some patients, some scientists suggested the colocalization of insulin and MHC class Ⅱ might reflect pancreatic β‐cells being phagocytosed by macrophages.

Russell et al.3 approached this argument with a novel technique called transcriptome analysis using donor islets as samples. Pancreatic β‐cells were sorted by fluorescence‐activated cell sorting from pancreatic islets obtained from 12 non‐diabetic donors and four type 1 diabetes donors. Ribonucleic acid sequencing of bulk‐sorted pancreatic β‐cells showed expression of not only MHC class I gene, but also MHC class II and MHC class II‐related genes, such as CIITA, in pancreatic β‐cells from patients with type 1 diabetes3. Furthermore, islet cells from one non‐diabetic donor and one type 1 diabetes donor were separated, and single‐cell ribonucleic acid sequencing was carried out; and islet cells, including α‐cells or β‐cells, were distinguished by two‐dimensional plots using t‐distributed stochastic neighbor embedding. Furthermore, as a result of examining the gene profile of MHC class II or class II invariant chain CD74 in these pancreatic β‐cells, MHC class II expression was observed in 35.2% of pancreatic β‐cells from type 1 diabetes patients, but 7.8% in non‐diabetic pancreatic β‐cells3.

These studies show that pancreatic β‐cells from patients with type 1 diabetes have a greater then fivefold increase in MHC class II gene expression, IRF8 gene expression and SPI1 gene expression (positive regulators of MHC class II), and the expression of CTSS (gene for cathepsin S that plays an important role in antigen loading) compared with non‐diabetic pancreatic β‐cells.

Furthermore, the authors verified the surface antigens of these islet cells by flow cytometry and immunostaining. Flow cytometry using donor islets showed that the human leukocyte antigen‐DR‐positive rate in pancreatic β‐cells (insulin⁺ CD45⁻ cells) was >10‐fold in type 1 diabetes than in non‐diabetic islets. Immunostaining using pancreatic tissue showed that MHC class II and CIITA were not stained in non‐diabetic pancreatic islets, but were occasionally found in the islets patients with type 1 diabetes. Furthermore, CD68⁺ cells (macrophages) exist independently of CIITA⁺ β‐cells, and it was shown that in this study, CIITA and insulin double positive cells were not macrophages that phagocytosed β‐cells3.

This study raises the question of how MHC class II expression is controlled in pancreatic β‐cells. The authors co‐cultured non‐diabetic β‐cells with IFNγ, interleukin‐1 beta and tumor necrosis factor‐alpha for 48 h, and carried out flow cytometry. As a result, the expression of human leukocyte antigen‐DR was increased, as in the case of pancreatic β‐cells from type 1 diabetes patients. In other words, non‐diabetic pancreatic β‐cells also express MHC class II under certain conditions.

Their immunostaining also suggests that the level of MHC class II expression correlates with the degree of leukocyte infiltration around the islets, but is poorly associated with pancreatic β‐cell MHC class I expression or the presence of enterovirus capsid protein VP13.

Furthermore, it is known that in type 2 diabetes or autoimmune pancreatitis, pancreatic β‐cells do not express MHC class II4.

These data suggest that some local cytokine profiles; for example, the local increase of IFNγ, characteristic of type 1 diabetes, are associated with the expression of MHC class II in pancreatic β‐cells.

This study used a highly accurate experimental method to show that MHC class II is expressed in pancreatic β‐cells from patients with type 1 diabetes. However, the most important issue in this study is whether MHC class II positive pancreatic β cells have the ability to present antigen.

In this study, the expression of CD80/ 86, costimulatory molecules expressed on professional antigen‐presenting cells, necessary for effective antigen presentation to naive T cells, was not shown on the surface of pancreatic β‐cells. No cell‐to‐cell interaction between T cells and MHC class II⁺ β‐cells was shown. MHC class II generally binds to phagocytosed antigenic peptides, but it is not known if phagocytic ability of foreign antigens exists in pancreatic β‐cells.

Recently, however, it has been suggested that MHC class II might present endogenous peptides; for example, misfolded proteins5. MHC class II bound to the misfolded protein of pancreatic β‐cells might be expressed on the cell surface when inflammation occurs, such that the IFNγ level increases locally. In this case, antigen‐specific B cells and memory T cells might be activated, which is consistent with the production of various islet‐related autoantibodies in type 1 diabetes (Figure 1).

Figure 1.

The difference in T‐cell activation between professional antigen‐presenting cells and pancreatic β‐cells. Professional antigen‐presenting cells (APCs) present foreign antigens to naive T cells, which are activated by the interaction between CD80/86 and CD28. Pancreatic β‐cells express major histocompatibility complex (MHC) class II by interferon gamma (IFN), and might present endogenous peptides to CD4⁺ T cells.

Further investigation into these ideas will lead not only to a better understanding of the pathogenesis of type 1 diabetes, but also to the development of new treatments.

Disclosure

The authors declare no conflict of interest.