BackgroundImmune checkpoint inhibitors have emerged as a front-line treatment for cancer in multiple indications. Unfortunately, a majority of patients do not realize durable response as a result of primary resistance to the immunotherapy. We have previously described a novel 27-gene immuno-oncology assay and algorithm which demonstrated significant predictive value in both NSCLC and TNBC. This algorithm utilizes gene expression to assess the tumor immune microenvironment (TIME) by combining aspects of the immune response, surrounding stromal cell signaling, and tumor physiology. We hypothesized that features of this algorithm may not only identify responders to immunotherapy (immunomodulatory, IO subtype) but may better enrich for patients who would benefit from other targeted therapeutics that alter the tumor microenvironment such as VEGF or FAK inhibitors (mesenchymal, M subtype).MethodsPathway analysis was used on TNBC specimens representing both the IO and M subtypes as determined by the 27-gene immuno-oncology algorithm. Expression reads were scaled within each sample and the difference of the mean of expression of each gene within IO and M subtypes was determined to quantify relative expression within each pathway. Finally, the mesenchymal score obtained from the 27-gene immuno-oncology algorithm was used to stratify RNAseq expression data from xenograft models that were either sensitive or resistant to a FAK inhibitor (FAKi).ResultsPathway analysis identified stratification between the 27-gene immuno-oncology algorithm subtypes finding with the mesenchymal subtype is associated with higher WNT, TGF-B, and RAS pathways whereas the IO subtype was more highly associated with the JAK/STAT pathway. Additionally, the mesenchymal score from the 27-gene immuno-oncology algorithm was higher in the FAK inhibitor sensitive (0.36) xenograft models than the FAKi resistant (0.076) models (p = 0.025).ConclusionsThe 27-gene immuno-oncology algorithm assesses the TIME to account for the immune response, surrounding stromal cell signaling, and tumor physiology to provide both an immuno-oncology subtype and mesenchymal subtype. We have previously demonstrated improved ability of the IO subtype to predict response to ICIs over current gold standard biomarkers. These data suggest that the M subtype is a distinct feature of the IO subtype which may enrich for patients more likely to respond to targeted therapeutics that act upon the canonical tumor promoting signaling pathways.
Erratum to: Bacterial and Protozoan Lipoxygenases Could be Involved in Cell-to-Cell Signaling and Immune Response Suppression
