Immunotherapy resistance in non-small cell lung cancer (NSCLC) may be mediated by an immunosuppressive microenvironment, which can be shaped by the mutational landscape of the tumor. Here, we observed genetic alterations in the PTEN/PI3K/AKT/mTOR pathway and/or loss of PTEN expression in >25% of patients with NSCLC, with higher frequency in lung squamous carcinomas (LUSC). Patients with PTEN-low tumors had higher levels of PD-L1 and PD-L2 and showed worse progression-free survival when treated with immunotherapy. Development of a Pten-null LUSC mouse model revealed that tumors with PTEN loss were refractory to antiprogrammed cell death protein 1 (anti-PD-1), highly metastatic and fibrotic, and secreted TGFß/CXCL10 to promote conversion of CD4+ lymphocytes into regulatory T cells (Treg). Human and mouse PTEN-low tumors were enriched in Tregs and expressed higher levels of immunosuppressive genes. Importantly, treatment of mice bearing Pten-null tumors with TLR agonists and anti-TGFß antibody aimed to alter this immunosuppressive microenvironment and led to tumor rejection and immunologic memory in 100% of mice. These results demonstrate that lack of PTEN causes immunotherapy resistance in LUSCs by establishing an immunosuppressive tumor microenvironment that can be reversed therapeutically. Significance: PTEN loss leads to the development of an immunosuppressive microenvironment in lung cancer that confers resistance to anti-PD-1 therapy, which can be overcome by targeting PTEN loss-mediated immunosuppression.

Brain radiation necrosis (RN) is a subacute or late adverse event following radiotherapy, involving an exacerbated inflammatory response of the brain tissue. The risk of symptomatic RN associated with stereotactic radiosurgery (SRS) as part of the treatment of brain metastases (BMs) has been a subject of recent investigation. The activation of the signal transducer and activator of transcription 3 (STAT3) was shown in reactive astrocytes (RA) associated with BMs. Given that the pathophysiological mechanisms behind RN are not fully understood, we sought to investigate the role of STAT3 among other inflammatory markers in RN development. A mouse model of RN using clinical LINAC-based SRS was designed to induce brain necrosis with the administration of 50 Gy in a single fraction to the left hemisphere using a circular collimator of 5 mm diameter. Immunohistochemistry and multiplex staining for CD4, CD8, CD68, GFAP, and STAT3 were performed. For validation, eleven patients with BMs treated with SRS who developed symptomatic RN and required surgery were identified to perform staining for CD68, GFAP, and STAT3. In the mouse model, the RN and perinecrotic areas showed significantly higher staining for F4/80+ and GFAP+ cells, with a high infiltration of CD4 and CD8 T-lymphocytes, when compared to the non-irradiated cerebral hemisphere. A high number of GFAP+pSTAT3+ and F4/80+pSTAT3+ cells was found in the RN areas and the rest of the irradiated hemisphere. The analysis of human brain specimens showed that astrocytes and microglia were actively phosphorylating STAT3 in the areas of RN and gliosis. Phosphorylated STAT3 is highly expressed in the microglia and RA pertaining to the areas of brain RN. Targeting STAT3 via inhibition represents a promising strategy to ameliorate symptomatic RN in BM patients undergoing SRS.