Objectives Immune checkpoint inhibitors have achieved clinical success in cancer treatment, but this treatment causes immune‐related adverse events, including type 1 diabetes (T1D). Our aim was to test whether a JAK1/JAK2 inhibitor, effective at treating spontaneous autoimmune diabetes in nonobese diabetic (NOD) mice, can prevent diabetes secondary to PD‐L1 blockade. Methods Anti‐PD‐L1 antibody was injected into NOD mice to induce diabetes, and JAK1/JAK2 inhibitor LN3103801 was administered by oral gavage to prevent diabetes. Flow cytometry was used to study T cells and beta cells. Mesothelioma cells were inoculated into BALB/c mice to induce a transplantable tumour model. Results Anti‐PD‐L1‐induced diabetes was associated with increased immune cell infiltration in the islets and upregulated MHC class I on islet cells. Anti‐PD‐L1 administration significantly increased islet T cell proliferation and islet‐specific CD8 + T cell numbers in peripheral lymphoid organs. JAK1/JAK2 inhibitor treatment blocked IFNγ‐mediated MHC class I upregulation on beta cells and T cell proliferation mediated by cytokines that use the common γ chain receptor. As a result, anti‐PD‐L1‐induced diabetes was prevented by JAK1/JAK2 inhibitor administered before or after checkpoint inhibitor therapy. Diabetes was also reversed when the JAK1/JAK2 inhibitor was administered after the onset of anti‐PD‐L1‐induced hyperglycaemia. Furthermore, JAK1/JAK2 inhibitor intervention after checkpoint inhibitors did not reverse or abrogate the antitumour effects in a transplantable tumour model. Conclusion A JAK1/JAK2 inhibitor can prevent and reverse anti‐PD‐L1‐induced diabetes by blocking IFNγ and γc cytokine activities. Our study provides preclinical validation of JAK1/JAK2 inhibitor use in checkpoint inhibitor‐induced diabetes.

EZH2 is mutated in nearly 25% of follicular lymphoma (FL) cases. Little is known about how EZH2 affects patients’ response to therapy. In this context, the aim of this study was to retrospectively analyze the frequency of mutations in EZH2 at diagnosis in tissue and ctDNA in patients with FL and to assess the patients’ outcomes after receiving immunochemotherapy, depending on the EZH2 mutation status. Among the 154 patients included in the study, 27% had mutated EZH2 (46% with high-grade and 26% with low-grade FL). Of the mutated tissue samples, the mutation in ctDNA was identified in 44% of cases. EZH2 mutation in ctDNA was not identified in any patient unmutated in the tissue.Unmutated patients who received R-CHOP had significantly more relapses than patients who received R-Bendamustine (16/49 vs. 2/23, p = 0.040). Furthermore, our results show that patients with mutated EZH2 treated with R-CHOP vs. those treated with R-Bendamustine present a lower incidence of relapse (10% vs. 42% p = 0.09 at 4 years), a higher PFS (92% vs. 40% p = 0.039 at 4 years), and higher OS (100% vs. 78% p = 0.039 at 4 years). Based on these data, RCHOP could be a more suitable regimen for mutated patients, and R-bendamustine for unmutated patients. These findings could mean the first-time identification of a useful biomarker to guide upfront therapy in FL.

No abstract

No abstract