P301 - Nutritional Prehabilitation in Mice Prevents Colorectal Cancer Recurrence via Modulation of the Microbiome

Purpose/Background: Up to 30% of patients undergoing surgery for colorectal cancer (CRC) will develop a postoperative local recurrence or distal metastasis. While a high-fat ‘Western’ diet is a risk factor, how it drives postoperative tumor formation is unclear. We have previously shown that a high-fat diet enriches the microbiome of mice with pro-tumorigenic bacteria that produce collagenase. When present in mice undergoing surgery to mimic human CRC resection, these collagenolytic bacteria create a microenvironment enriched with signals that induce a metastatic phenotype in cancer cells that subsequently drive postoperative tumor formation. Given that diet is modifiable, preoperative nutritional prehabilitation targeted to influence the gut microbiome represents a novel strategy to decrease postoperative CRC recurrence. Therefore, here we test the hypothesis that short-term preoperative nutritional prehabilitation that decreases the microbiome’s collagenolytic activity will prevent postoperative tumor formation.

Methods/Interventions: We utilized a modified version of our validated model of CRC recurrence, where mice eat a high-fat/no-fiber control diet for 12 weeks and then undergo a distal colon resection followed by an enema of luminescent AKPT tumoroids on postoperative day (POD) 2. To test how preoperative nutritional prehabilitation influences postoperative tumor formation, we designed 4 custom diets that varied solely in their composition of fiber and fat. After 6 weeks on the control high-fat/no-fiber diet, experimental cohorts were created by switching mice to one of the 4 custom diets for an additional 6 weeks, and then underwent surgery and tumoroid enema as above. Mice were imaged weekly to measure tumor burden and were sacrificed on POD 56 to assess for gross tumors. Stool recovered on the day of surgery (6 weeks post-nutritional intervention) was assayed for bacterial composition and function via shotgun metagenomics and in vitro collagenase activity.

Results/Outcomes: High-fat/no-fiber fed mice switched preoperatively to a low-fat/high-fiber diet demonstrated significantly less tumor incidence (7% vs. 46%; p=.02) and tumor burden (1.1x10^7 A.U. vs. 1.7x10^8 A.U.; p=.03) compared to control mice that remained on a high-fat/no-fiber diet throughout the entire experiment (Fig1A). There were no significant differences in tumor formation or burden between the control mice and the other 3 experimental diets. Metagenomic analysis showed that high-fat/no-fiber fed mice prehabed with the low-fat/high-fiber diet drove a distinct bacterial composition (Fig1B). Metagenomic analysis demonstrated that the number of collagenase genes was significantly decreased in mice that were prehabed with a low-fat/high-fiber diet compared to mice that either remained on the high-fat/no-fiber diet or were switched to one of the other experimental diets (Fig1C). To determine that these differences translated to phenotypic changes, functional collagenase assays were performed on cultivated bacteria. Similarly, mice prehabed with a low-fat/high-fiber diet had significantly less bacterial collagenolytic activity than any of the other cohorts (Fig1D).

Conclusion/Discussion: Our findings indicate that nutritional prehabilitation with a low-fat/high-fiber diet prevents postoperative tumor formation via its influence on the microbiome. Our data provide scientific justification to immediately test how a preoperative low-fat/high-fiber diet can improve oncological outcomes in humans.

Figure 1
Fig 1. (A) Tumor incidence (%) by diet. (B) Principal Component Analysis (PCoA) plot showing composition differences in the intestinal microbiome by diet type. (C) Metagenomic analysis showing the number of collagenase genes detected within the intestinal microbiome of the mice from each diet group. (D) Functional collagenase assays assessing the percentage of collagenolytic colonies within the intestinal microbiome from mice within each diet group. N=15 mice per cohort. *p < 0.05 compared to all other groups.