Prognostic role of molecular subclassification based on mutational profiling, microsatellite instability, and tumor-associated neutrophil status in stage III colon cancer

Abstract

Colorectal cancer (CRC) was the third most diagnosed cancer and the second leading cause of death worldwide in 2020. The underlying molecular pathogenesis of CRC is a complex heterogeneous process. The latest molecular subclassification of CRC, named consensus molecular subtypes (CMSs), consider not only the mutation burden, but also metabolism, epigenetics, and immune response. However, there are still 13% of CRC patients, who fail to fit in any CMS due to tumor heterogeneity.

This project aimed to stratify patients into different molecular subclasses (MSCs) to predict the relapse and metastatic risks by utilizing polymerase chain reaction (PCR)-based microsatellite instability (MSI) assay, elevated microsatellite alteration in selected tetranucleotides (EMAST) essays, quantification of tumor-associated neutrophils (TAN) infiltration by immunohistochemical staining with the surface marker, CD66b, to represent the inflammatory status, and genetic profiling with a 77-gene panel specific for colorectal cancer by next-generation deep sequencing.

Formalin-fixed, paraffin-embedded (FFPE) blocks from 152 colon stage III patients were utilized for the risk stratification of MSCs based on the MSI/EMAST combinations and TAN CD66b level. For a subset of patients with genetic mutational profiling data (69, 45.4%), the usefulness of KRAS/BRAFV600E status was used for further classification of the CMS is based on MSI/EMAST/CD66b level. Although the prognostic role of MSI-H phenotype is well-known, MSI-L/MSS and EMAST status remains unclear. Patients with MSI-L or EMAST-positive status were defined as MSI-moderate (MSI-M). The underlying molecular mechanism of MSI-M phenotype was suggested to be due to inflammation-mediated mis-translocation of MSH3 resulting in deficiency of repair of microsatellites. Preliminary data by multivariate COX analysis suggested the MSI/EMAST combination was an independent prognostic and colon stage III patients with MSI-M phenotype had a worse prognosis compared to MSI-H phenotype. TANs are innate immune cells frequently found in the tumor microenvironment (TME). TANs associate with adverse prognosis in many cancers, but its role in CRC remains contradictory. My preliminary observation suggested positive CD66b-TAN infiltration in peripheral/invasive front region (PFS p = 0.014; OS p = 0.022) was an independent prognostic factor for PFS and OS and associated with improved survival. The top three frequently mutated genes included TP53 (63.7%), APC (51.6%) and KRAS (46.2%). The KRAS mutants showed significant but marginal association with poor prognosis in PFS and OS respectively. Among 109 patients, MSC classification based on the MSI/EMAST combinations and CD66b TAN status identified MSC2, patients with MSI-M/CD66b-L status, showed the worst prognosis and highest risks of shorter PFS (HR = 10.87, 95% CI = 2.43 – 48.6, p = 0.002) and OS (HR = 13.45, 95% CI = 2.96 – 61.25, p = 0.001) compared to MSC1 (MSI-H). Patients with MSC3 (MSI-M/CD66b-H), MSC4 (MSS/CD66b-L) and MSC5 (MSS/CD66b-H) also had poorer prognosis compared to MSC1.

Further molecular subclassification in a subset of 106 patients with KRAS mutation status resulted in 6 MSCs with independent prognostic roles. The current findings provide novel insights of combining biomarkers to reflect the complex TME and immune status to the tumor genetic profile for disease risk stratification and treatment guidance for personalized precision medicine.