Tumor gene patterns predict chemoimmunotherapy response in dogs

Abstract

Pet dogs develop spontaneous diffuse large B cell lymphoma (DLBCL), and veterinary clinical trials have been employed to treat canine DLBCL and to inform clinical trials for their human companions. A challenge that remains is selection of treatment to improve outcomes. The dogs in this study were part of a larger clinical trial evaluating the use of combinations of doxorubicin chemotherapy, anti-CD20 monoclonal antibody, and one of three small molecule inhibitors: KPT-9274, TAK-981, or RV1001. We hypothesized that significant differential expression of genes (DEGs) in the tumors at baseline could help predict which dogs would respond better to each treatment based on the molecular pathways targeted by each drug. To this end, we evaluated gene expression in lymph node aspirates from 18 trial dogs using the NanoString nCounter Canine Immuno-oncology (IO) Panel. We defined good responders as those who relapsed after 90 days, and poor responders as those who relapsed prior to 90 days. We analyzed all dogs at baseline and compared poor responders to good responders, and found increased CCND3 correlated with poor prognosis and increased CD36 correlated with good prognosis, as is observed in humans. There was minimal DEG overlap between treatment arms, prompting separate analyses for each treatment cohort. Increased CREBBP and CDKN1A for KPT-9274, increased TLR3 for TAK-981, and increased PI3Kδ, AKT3, and PTEN, and decreased NRAS for RV1001 were associated with better prognoses. Trends for selected candidate biomarker genes were confirmed via qPCR. Our findings emphasize the heterogeneity in DLBCL, similarities and differences between canine and human DLBCL, and ultimately identify biomarkers that may help guide the choice of chemoimmunotherapy treatment in dogs.