Tumoral microbial and pathway alterations associated with young-onset rectal cancer and its response to therapy.

Abstract

160 Background: Given the increasing incidence of young onset rectal cancer (YORC) in recent decades, we examined the potential of tumor and fecal microbiome pathways in YORCs as compared to late-onset rectal cancers (LORCs) while also exploring markers of response to neoadjuvant therapy (NAT). We hypothesize microbial markers reflect unique and potentially causative biology in YORC, and variability in response to NAT. Methods: Treatment-naïve patients with YO- (<50 years old) or LO- ≥50 years old) rectal cancers were prospectively enrolled. Tumor and tumor-adjacent normal tissues (TANs) were collected via endoscopic biopsy, as well as stool at baseline. After whole genome sequencing, metagenomic analysis was performed using MetaPhlAn4, VirMAP and HUMAnN3 for taxonomical and metabolomic profiling. Association of microbial and immune signatures were examined comparing tumors vs. TANs, and YO- vs. LORC tumors. Differential abundance analyses were performed using paired Wilcoxon test and MaAsLin2. Molecular subtype defined for tissues with whole transcriptomic sequencing available. Results: 162 tumors and paired normal tissues, 73 (45.1%) YO- and 89 (54.9%) LORC were analyzed. Additional whole transcriptomic sequencing was available for 88 patients. The majority 145 (89.5%) of patients received neoadjuvant therapy. Observed bacterial diversity was higher in tumors as compared to tumor associated normal (TAN) tissue (p<0.001). Moreover, bacterial diversity also increased linearly with age of diagnosis in tumors (p=0.01), but not TAN (p=NS). Oral bacteria previously associated with colorectal cancer were noted in these rectal cancers including Fusobacterium nucleatum (q<0.001) and Parvimonas micra (q<0.001). Escherichia coli, previously associated with colorectal cancer and development of metastatic disease was associated with YORC (q<0.001) but not LORC. For NAT treated patients, major pathologic response in YORC was associated with higher bacterial diversity (p=0.03) but not in LORC (p=NS). Similarly, no significant bacterial marker of response to neoadjuvant therapy was noted in LORC. While Butyricimonas virosa (q<0.001), Alistipes shahi (q=0.01), and Parabacteroides merdae (q=0.01) were all associated with a diminished pathologic response. Conclusions: Oral bacterial signatures previously associated with colorectal cancer were noted in both YORC and LORC when comparing matched tumor and TAN. Importantly, several unique bacterial associations were identified with YORC, including Escherichia Coli – previously associated with development of metastases in CRC. We also noted distinct microbial associations related to response to therapy in YORC suggesting the unique biology of these tumors. Functional bacterial pathway associations with response suggest a potential to augment NAT with existing or next-generation microbial modulation strategies.