A Clinically-Mimetic Focal Bowel Irradiation Model to Assess Spatial Microbiome Perturbations after Duodenal Radiotherapy

• Published in: International journal of radiation oncology, biology, physics Vol. 120; no. 2; p. e365
• Main Authors: Habimana-Griffin, L., Prusa, J., Wang, B., Strong, L., Fishbein, S., Ning, J., Reynoso, F.J., Markovina, S., Ciorba, M., Dantas, G.
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.10.2024
• ISSN: 0360-3016
• DOI: 10.1016/j.ijrobp.2024.07.806

The duodenum is a key organ at risk during stereotactic ablative radiotherapy (SABR) and understanding mechanisms of duodenal radiotoxicity may mitigate side effects of treatment and expand the therapeutic window of SABR. The gut microbiome is thought to play an important role in bowel radiotoxicity, however existing preclinical models either require surgical manipulation or do not recapitulate high dose conformal treatment fields used during SABR, which can confound microbiome studies. We sought to develop a non-invasive clinically-mimetic focal bowel irradiation model to assess the role of the microbiome after high-dose radiation to the duodenum. Male and female C57BL/6 mice (N = 8-10/group/time point) underwent sham-CT or focal irradiation to 12 or 18 Gy delivered with conformal arcs to the proximal duodenum with a small animal irradiator using CT-guidance with oral iodine contrast for bowel visualization. Stool samples were collected prior to treatment and at days 4, 14, or 91 post-treatment. Subsets of mice were sacrificed at days 4, 14 or 91 and duodenal samples were harvested. Genomic DNA was extracted from stool and duodenal tissue homogenates and processed for 16S rRNA sequencing. Reads were processed and taxonomy was assigned using the DADA2 package in R. The Vegan package in R was used to calculate diversity metrics. Mean diversity scores were compared using pairwise Wilcoxon tests with Benjamini-Hochberg (BH) correction. MaAsLin2 was used to identify bacterial genera associated with radiotherapy treatment. One hundred percent of mice survived until the time of planned sacrifice. There were no significant differences in Shannon Diversity index or Bray-Curtis dissimilarity in the stool of irradiated mice compared to sham controls at any timepoint (ns, p > 0.05). Duodenal samples at Day 4 showed no difference in Shannon Diversity index (ns, p > 0.05), but significantly lower Bray-Curtis dissimilarity in the 12 Gy group compared to sham controls (BH adjusted p<0.01). Multivariable association modeling with MaAsLin2 identified 9 genera that were differentially abundant in both the 12 Gy and 18 Gy treatment groups compared to sham controls in the duodenum at day 4 (BH adjusted p<0.05). These genera included Akkermansia, Lactobacillus, Turicibacter, Pseudomonas, Paracoccus, A2, Intestinimonas, Clostridium.sensu.stricto.1, and Lachnospiraceae.NK4A136.group. In contrast, there were no differentially abundant genera in the radiation groups compared with sham controls in the stool at any time point (ns, p > 0.05). Our clinically-mimetic focal bowel irradiation model reveals biogeographically distinct microbiome perturbations in the duodenum after duodenal radiotherapy, while stool samples show no change relative to controls. Gut biogeography should be considered in future studies examining focal bowel radiation injury as stool alone may not accurately represent community dynamics at the site of injury.