481 Individualized Responses of Gut Microbiota to Dietary Intervention Modeled in Humanized Mice

Abstract

Background and Aim: Several studies have looked at the effects of different types of fiber (i.e., prebiotics) on gut microbiota and symptoms in patients with IBS; however there is lack of consistent response to prebiotics among patients. In order to determine the reason for lack of predictable responses in patients, here we investigate the effect of a specific dietary change on gut microbial composition using humanized mice. Methods: Germ free (GF) mice were colonized with human feces (humanized) by gavage of 200μl of frozen feces mixed with an equal volume of pre-reduced PBS. Microbial composition was assessed by 16S rRNA based pyrosequencing using the 454-titanium platform. Non-targeted fecal metabolomics was performed using UPLC-MS. Data analysis was performed using QIIME and MetaboAnalyst. Results: GF mice were humanized using feces from three healthy human donors: two males and one female (n=4-8/donor). Four weeks following humanization, fecal samples were collected from each mouse. Un-weighted UniFrac based PCoA plots of the 16S rRNA microbial composition data demonstrate reproducible reconstitution of the corresponding human fecal inoculum in humanized mice. The metabolomic profile of each sample was used as a read-out of the microbiota functional status. PCoA of these metabolomicsbased fingerprints revealed that reproducible aspects of microbiota functional status are dictated by the human fecal inoculum. Together these data show that individual aspects of the human gut microbiota's composition and metabolomic activity can be reconstructed within humanized mice. The presence of most fecal metabolites was independent of the fecal source (human or humanized mouse) suggesting that much of the microbiota functional status is resilient to modification by the host species and largely shared between individuals. In order to determine if dietary intervention impacts all three groups of humanized mice similarly, diet was switched from standard diet to FOS-enriched diet (fructo-oligosaccharide, commonly used prebiotic; 10%w/v). Fecal samples were collected before and after dietary intervention. Un-weighted UniFrac based PCoA plots of the 16S rRNA microbial composition data show distinct diet-induced changes in microbial composition between the three groups of humanized mice, with only a small effect seen in one group. Conclusion: Gut microbial composition and function can be reliably reconstructed in humanized mice making them a good model to study effects of environmental factors on the human gut microbiota. The effect of prebiotics/dietary intervention on gut microbial composition depends on the initial microbial community composition. This study may therefore provide new insight into the observed differences in patient responses to prebiotics in previous studies.