Hippo pathway signaling associated with immune cell trafficking in colorectal cancer.

Abstract

156 Background: Emerging evidence suggests subsets of the Hippo pathway have multiple functions of tumor development and immune response regulation. Increased understanding of the molecular characteristics of its signaling pathways and the impact on immune cell trafficking will be critical to develop colorectal cancer (CRC) therapies. Methods: A total of 13,008 CRC tumors were analyzed at Caris Life Sciences (Phoenix, AZ) with whole transcriptome and whole exome sequencing (NovaSeq). MSI-H/dMMR was tested by NGS (next generation sequencing), immunohistochemistry (IHC), and fragment analysis. Tumor mutational burden (TMB)-High was determined with a 10-mt/MB cutoff. RNA-deconvolution using QuantiSEQ was used to assess immune cell infiltration in the tumor microenvironment. Consensus molecular subtypes (CMS) were developed using RNA expression data. Gene expression was reported as transcripts per million. Z-score totals of 6 core Hippo genes were calculated in groups: MST1+ STK3 (G1), LATS1+ LATS2 (G2), YAP1+ WWTR1 (G3), and all 6 genes together (G4). Tumors with bottom quartile (QL) z-scores were compared with the top quartile (QH) using χ2/Fisher-Exact test and adjusted with the Benjamini-Hochberg method: adjusted p <.05 was considered significant; unadjusted p <.05 trending. Results: Gene expression levels were significantly positively correlated with each other (Spearman rho: 0.30–0.78). Highest expression levels of G1–4 were seen in CMS4 and lowest in CMS3 with significant differences. MSI-H/dMMR were significantly higher in QL than QH in all (G1: 7.4 vs 4.9%, G2: 7.4 vs 4.5, G3: 8.7 vs 4.1, G4: 7.5 vs 4.5). TMB-H prevalence (%) showed inverse relationships with MSI-H significantly in G1–3 (G1: 10.7 vs 7.5, G2: 10.5 vs 7.1, G3: 12.7 vs 6.2) and trending in G4 (11.0 vs 7.0, p =.006). Considering only MSS tumors, TMB-H trended more often in QL than QH in G1, 3, and 4 (3.6–3.7 vs 2.2–2.6). PD-L1 expression by IHC was significantly higher in QH than QL in all (5.3 vs 3.3%) and MSS tumors (2.2 vs 4.2) only in G3. Z-scores of G1–4 were all positively correlated with immune cell infiltrations. Significantly higher fractions of B cells, M2 macrophages, myeloid dendritic cells, NK cells, neutrophils, and CD8+ T cells were seen in QH than QL with a median fold change of 1.39. G1–4 z-scores all positively correlated with the expression of the analyzed immune-related genes. The highest Spearman rho averages were in HAVCR2 (0.54), CD86 (0.54), CD80 (0.53), PD-L2 (0.5), CD274 (0.46), and LAG3 (0.36). Significantly different mutation rates were seen in QH compared to QL in G1 ( TP53, KRAS, PIKCA, SMAD2, AMER1), G2 ( APC), G3 ( PIK3CA, APC), and G4 ( TP53, PIK3CA, KRAS). Conclusions: The Hippo pathway correlated with immune cell trafficking suggesting that YAP1/TAZ signaling may play a critical role in the immune responses. These findings may help develop novel therapeutic strategies targeting the Hippo pathway combined with immune therapies in CRC.