Fecal virome transplantation is sufficient to alter fecal microbiota and drive lean and obese body phenotypes in mice

Abstract

The gastrointestinal microbiome plays a significant role in modulating numerous host processes, including metabolism. Prior studies show that when mice receive fecal transplants from obese donors on high-fat diets (HFD) (even when recipient mice are fed normal diets after transplantation), they develop obese phenotypes, demonstrating the prominent role that gut microbiota play in determining lean and obese phenotypes. While much of the credit has been given to gut bacteria, the impact of gut viruses on these phenotypes is understudied. To address this shortcoming, we gavaged mice with viromes isolated from donors fed HFD or normal chow over a 4-week study. By characterizing the gut bacterial biota via 16S rRNA amplicon sequencing and measuring mouse weights over time, we demonstrate that transplanted viruses affect the gut bacterial community, as well as weight gain/loss. Notably, mice fed chow but gavaged with HFD-derived viromes gained more weight than their counterparts receiving chow-derived viromes. The converse was also true: mice fed HFD but gavaged with chow-derived viromes gained less weight than their counterparts receiving HFD-derived viromes. Results were replicated in two independent experiments and phenotypic changes were accompanied by significant and identifiable differences in the fecal bacterial biota. Due to methodological limitations, we were unable to identify the specific bacterial strains responsible for respective phenotypic changes. This study confirms that virome-mediated perturbations can alter the fecal microbiome in vivo and indicates that such perturbations are sufficient to drive lean and obese phenotypes in mice.

Keywords: High fat diet; bacteriophages; fecal microbiota; gut microbiome; obesity; virome.

Figure 1.

Study design. Recipient mice were fed chow or high-fat diet (HFD) for 14 weeks. After the first 10 weeks, fecal virome transplants (FVT) were administered via oral gavage every weekday for 4 weeks (red vertical bars) and feces were collected at regular intervals (dots). FVT were prepared in parallel from a separate group of donor mice on chow or high-fat diet. Gavage treatments consisted of chow-derived and HFD-derived viromes, as well as PBS controls. The study was conducted in 2 experimental replicates (i.e., trials). Trial 1 consisted of 4 mice per cage and 1 cage per treatment. Trial 2 consisted of 3 mice per cage and 3 cages per treatment. In Trial 2, we collected feces from 2 additional pre-gavage timepoints (purple dots). Five mice were omitted from the study due to death (one in Trial 1 HFD : PBS, one in Trial 2 HFD : PBS) or aggression toward other mice (one in each Trial 2 HFD : Chow, Chow : PBS, and Chow : HFD).

Figure 2.

Change in weight of individual mice in response to diet and FVT during 4 weeks of gavage. Percent change in weight is shown from initial to final weight of individual mice (a), as well as the percent change in weight over time for mice fed chow (b) or HFD (c). For mouse weights after 28 d (a), Diet and Gavage effects are significant (p = 2.2e-12, p = .0023, respectively) and effects of Trial, Cage, and Diet-by-Gavage interaction are not. For longitudinal analysis (b and c), Diet, Gavage, and Diet-by-Time interaction are significant (p = .028, p = .031, p < 2e-16, respectively) but Gavage-by-Time interaction is not. Data from PBS controls were omitted from statistical models.

Figure 3.

Bray-Curtis beta diversity of fecal microbiomes of FVT recipient mice visualized by Principal Coordinate Analysis (PCoA). Panel A: All recipient mice. Panels B and C: Trial 1 recipient mice fed chow (b) or HFD (c). Panels D and E: Trial 2 recipient mice fed chow (d) or HFD (e). Diet, Trial, Day, Gavage, and Day*Diet effects are significant when including all recipient mice (PERMANOVA with 999 permutations, R² and p-values reported in Panel f). Unfilled points indicate pre-gavage samples collected before day 0. Statistical tests were conducted on data from day 0–25.

Figure 4.

Shannon Index alpha diversity of fecal microbiomes during 4 weeks of FVT gavage from trial 1 (Panel A) and trial 2 (Panel B). Trial 1 (A) consisted of a single replicate per treatment, precluding statistical analyses. For trial 2, mice on HFD and receiving HFD-derived viromes (orange) had significantly lower alpha diversity between days 9–18 than mice on the same diet receiving chow-derived viromes (light blue) or PBS controls (green) (p < .05, spline regression and Wald test with p-values corrected for multiple comparisons). Overall, mice on a high fat diet had lower alpha diversity that mice on a chow diet.

Figure 5.

Aitchison compositional biplots of recipient mice visualized by principal component analysis (PCA). Panels a and B: Recipient mice from trial 1 fed chow (a) or HFD (b). Panels C and D: Recipient mice from trial 2 fed chow (c) or HFD (d). Color denotes treatment group (see Figure 1). Arrows denote important taxa in relation to sample clusters. Significance was determined by PERMANOVA with 999 permutations (Aitchison distances ~ Day * Gavage); the R² and p-value for the effect of gavage are indicated in respective panels.