Gut Bacteria Erysipelatoclostridium and Its Related Metabolite Ptilosteroid A Could Predict Grade 2 Radiation-Induced Intestinal Injury During Pelvic Radiotherapy

Radiation-induced intestinal injury (RIII) is a common complication in cancer patients receiving pelvic and abdominal radiotherapy (RT). Although emerging evidence have confirmed that alteration of gut microbiome and metabonomics contribute to RIII, few prediction models based on them exist. Here, we attempted to establish a prediction model of RIII based on the gut microbiota and metabolite.Stage I to III cervical or endometrial cancer patients who received pelvic or abdominal RT for the first time in our department were enrolled, RIII was diagnosed and scored according to the RTOG criteria. Stool samples were collected before, 20-30 Gy and 45-50 Gy after RT, respectively. Gut microbiome were measured by 16S rRNA sequencing, and metabolites were detected by LC-MS analysis. Then, univariate logistic regression analysis was used to identify potential biomarkers, and their potential prediction ability of RIII was evaluated by receiver operator characteristic (ROC) curves.17 patients (16 cervical cancer and 1 endometrial cancer) were finally recruited. In patients who later developed grade 2 RIII, dysbiosis of gut microbiota was observed, which was characterized by significantly higher relative abundance such as Prevotella Erysipelatoclostridium, and Alloprevotella. Alteration of gut metabolites was also identified, which was reflected by the enrichment of ptilosteroid A. Univariate analysis showed that Erysipelatoclostridium and ptilosteroid A were related to the occurrence of grade 2 RIII. Notably, a strong positive correlation between gut bacteria Erysipelatoclostridium relative abundance and gut metabolite ptilosteroid A expression was found. Furthermore, combinations of Erysipelatoclostridium and ptilosteroid A could provide strong prediction value for grade 2 RIII (AUC value of 0.87).Dysbiosis of both gut microbiota and metabolome develops in patients with RIII. Gut bacteria Erysipelatoclostridium and its related metabolite ptilosteroid A may collaboratively predict grade 2 RIII, and could be used as a prediction model.