Human adipose-derived mesenchymal stem cells prevent type 1 diabetes induced by immune checkpoint blockade.

Aims/hypothesis: Immunomodulators blocking cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed cell death protein 1 (PD-1) or programmed death-ligand 1 (PD-L1) have improved the treatment of a broad spectrum of cancers. These immune checkpoint inhibitors (ICIs) reactivate the immune system against tumour cells but can also trigger autoimmune side effects, including type 1 diabetes. Mesenchymal stem cell (MSC) therapy is the most prevalent cell therapy, with tissue-regenerating, anti-fibrosis and immunomodulatory functions provided by the secretome of the cells. Here, we examined whether systemic MSC treatment could prevent the development of type 1 diabetes in a NOD mouse model.

Methods: The purified PD-L1 monoclonal antibody was administered to induce diabetes in male NOD mice which normally do not develop diabetes. Human adipose-derived MSCs were administered by tail vein injections. T cells, macrophages and monocyte-derived macrophages expressing C-X-C motif chemokine ligand 9 (CXCL9) in pancreatic sections of NOD mice and a cancer patient who developed diabetes following the ICI treatments were analysed by immunofluorescence. Tissue localisation of the injected MSCs, plasma exosome levels and plasma cytokine profiles were also investigated.

Results: PD-1/PD-L1 blockade induced diabetes in 16 of 25 (64%) NOD mice which received anti-PD-L1 mAb without hMSCs [MSC(-)], whereas MSC administration decreased the incidence to four of 21 (19%) NOD mice which received anti-PD-L1 mAb and hMSCs [MSC(+)]. The PD-1/PD-L1 blockade significantly increased the area of CD3-positive T cells (6.2-fold) and macrophage-2 (Mac-2) antigen (2.5-fold)- and CXCL9 (40.3-fold)-positive macrophages in the islets. MSCs significantly reduced T cell (45%) and CXCL9-positive macrophage (67%) accumulation in the islets and the occurrence of diabetes. The insulin content (1.9-fold) and islet beta cell area (2.7-fold) were also improved by MSCs. T cells and CXCL9-positive macrophages infiltrated into the intricate gaps between the beta cells in the islets by PD-1/PD-L1 blockade. Such immune cell infiltration was largely prevented by MSCs. The most striking difference was observed in the CXCL9-positive macrophages, which normally did not reside in the beta cell region in the islets but abundantly accumulated in this area after PD-1/PD-L1 blockade and were prevented by MSCs. The CXCL9-positive macrophages were also observed in the islets of a cancer patient who developed diabetes following the administration of ICIs but few CXCL9-positive macrophages were observed in a control patient. Mechanistically, the injected MSCs accumulated in the lung but not in the pancreas and strongly increased plasma exosome levels and changed plasma cytokine profiles.

Conclusions/interpretation: Our results suggest that MSCs can prevent the incidence of diabetes associated with immune checkpoint cancer therapy and may be worth further consideration for new adjuvant cell therapy.