Alterations in the gut microbiota associated with HIV-1 infection

Abstract

Understanding gut microbiota alterations associated with HIV infection and factors that drive these alterations may help explain gut-linked diseases prevalent with HIV. 16S rRNA sequencing of feces from HIV-infected individuals revealed that HIV infection is associated with highly characteristic gut community changes, and antiretroviral therapy does not consistently restore the microbiota to an HIV-negative state. Despite the chronic gut inflammation characteristic of HIV infection, the associated microbiota showed limited similarity with other inflammatory states and instead showed increased, rather than decreased, diversity. Meta-analysis revealed that the microbiota of HIV-infected individuals in the U.S. was most similar to a Prevotella-rich community composition typically observed in healthy individuals in agrarian cultures of Malawi and Venezuela and related to that of U.S. individuals with carbohydrate-rich, protein- and fat-poor diets. By evaluating innate and adaptive immune responses to lysates from bacteria that differ with HIV, we explore the functional drivers of these compositional differences.

Fig. 1. Unweighted UniFrac clustering of HIV dataset

Each point represents a single fecal sample. Panels A and B represent the same PCoA analysis but for clarity, Panel A only shows HIV positive individuals not on ART at the time of initial sample collection and HIV-negative controls, and Panel B shows only HIV positive individuals on ART. Points are labeled with ART duration at the time of sample collection in months. Samples from the same individual at two timepoints are joined with a line. Bacterial families are plotted as a weighted average of the coordinates of all samples, where the weights are the relative abundances of the taxon in the samples (grey circles). The size of the sphere representing a taxon is proportional to the mean relative abundance of the taxon across samples. See Fig. S1 for results with Weighted UniFrac.

Fig. 2. Taxonomy bar charts showing dominance of Prevotellaceae in HIV infected subjects

Families significantly different between ART naïve individuals chronically infected with HIV and HIV-negative controls are marked with a star (p<0.05 ANOVA with FDR correction, only the first time point was used for individuals who had multiple time points). Those enriched in HIV-negative have a blue star and in HIV (Chronic) in green. Samples from the same individual at two timepoints are adjacent and joined with a bridge. See Tables S2A, S2B, and S2C for lists of discriminatory genera, families, and 97% OTUs respectively and Fig. S2 for a phylogenetic tree relating discriminatory 97% OTUs.

Fig. 3

Unweighted UniFrac PCoA plot comparing data from this study with a global survey of fecal microbial community composition conducted in individuals from Malawi, the Amazonas State of Venezuela, and the US (Yatsunenko et al., 2012) (Global Gut) and 2) a study linking long-term dietary patterns with gut microbial enterotypes (Wu et al., 2011) (Long term diet), using a reference mapping protocol. A) Points colored by study, B) Same plot but with points colored across a red-blue spectrum from low to high relative abundance of the genus Bacteroides, C) Same as for panel B but for the Prevotella genus, D) Points colored by country of residence and HIV status. See Fig. S3 for an additional meta-analysis exploring the relationship of HIV infection with Crohn’s disease.

Fig. 4. Characterization of immune responses to selected bacteria in untreated HIV positive and HIV negative subjects

A) TNF-α to IL-10 ratio (as measured using a bead array assay) after a 24 hour stimulation. An asterisk indicates that E. biforme had a significantly higher TNF-α/IL-10 ratio compared to the other bacterial isolates with the Kruskal-Wallis ANOVA and Dunns test. Red signifies HIV positive and blue HIV-negative subjects. B) Percentage of divided (Cell Trace low) CD4+ T cells enumerated by flow cytometry after a 6 day stimulation. Statistical significance was determined using Mann-Whitney T-tests. Bacteria that decreased in relative abundance with chronic, untreated HIV infection are red, those that increased are blue, and no difference is black. Points in the scatter plots depict each individual and the bar is the median. See also Fig. S4.