Altered Virome and Bacterial Microbiome in Human Immunodeficiency Virus-Associated Acquired Immunodeficiency Syndrome

Abstract

Human immunodeficiency virus (HIV) infection is associated with increased intestinal translocation of microbial products and enteropathy as well as alterations in gut bacterial communities. However, whether the enteric virome contributes to this infection and resulting immunodeficiency remains unknown. We characterized the enteric virome and bacterial microbiome in a cohort of Ugandan patients, including HIV-uninfected or HIV-infected subjects and those either treated with anti-retroviral therapy (ART) or untreated. Low peripheral CD4 T cell counts were associated with an expansion of enteric adenovirus sequences and this increase was independent of ART treatment. Additionally, the enteric bacterial microbiome of patients with lower CD4 T counts exhibited reduced phylogenetic diversity and richness with specific bacteria showing differential abundance, including increases in Enterobacteriaceae, which have been associated with inflammation. Thus, immunodeficiency in progressive HIV infection is associated with alterations in the enteric virome and bacterial microbiome, which may contribute to AIDS-associated enteropathy and disease progression.

Keywords: AIDS; AIDS enteropathy; HIV; adenovirus; microbiome; systemic inflammation; virome.

Figure 1. Study Design and Cohort Characteristics

(A) Subject cohort and study design. Correlation between CD4 T cell count (y-axis) compared to (B) HIV viral load (p < 0.0001); (C) weight (p = 0.0037); and (D) body mass index (p= 0.0093). Circulating levels of sCD14 were graphed by (E) HIV status and ART treatment group and (F) CD4 T cell count. (G) Correlation between sCD14 levels and HIV viral load (p =0.002). p ≤ 0.01 = **, p ≤ 0.0001 = ****. Bars indicate median ± interquartile range (IQR). See also Table S1.

Figure 2. Eukaryotic viruses identified in the fecal samples

Alignment (red bars) of (A) two near full-length adenovirus contigs (34,669 and 32,313 bp) and (B) two representative Human Papillomavirus (HPV) contigs (7,362 and 7,857 bp) to the most closely related virus genome (blue bars). Percentages indicate nucleotide identity over the length of the best-aligned homologous region compared to the reference genome. C) Correlation of Alphatorquevirus copies/ml as measured by quantitative real-time qPCR with number of unique Anelloviridae-assigned sequences (p < 0.0001). (D) Phylogenetic distance of anellovirus contig sequences (colored circles) with anellovirus reference genomes (white circles). Colored circles represent viruses identified in samples from: blue, HIV-negative; green, CD4 >200; and red, CD4 <200 subjects. Bar indicates bootstrap distance. See also Figure S1, Table S2, S3.

Figure 3. Comparison of enteric eukaryotic virus sequences and CD4 T cell count

Abundance of (A) Adenoviridae-, (B) Anelloviridae-, (C) Circoviridae- and (D) Papillomaviridae-assigned sequences in samples collected from HIV-negative subjects and subjects with CD4 >200 and <200. Sequences were normalized by dividing by the number of dereplicated (<95% identical), high-quality sequences. p ≤ 0.05 = *, p ≤ 0.01 = **. Bars indicate median ± interquartile range (IQR). Y-axis standardized by taking square root of the normalized value. See also Figures S2, S5, Table S4.

Figure 4. Bacterial community profiling

(A) Heatmap showing relative abundance of the 20 most frequent bacterial families (y-axis) by sample (x-axis), grouped by HIV status and CD4 T cell count. Percent abundance is indicated by the gradient key. (B) Chao1 rarefied bacterial richness grouped by CD4 T cell count. (C) Comparison of bacterial Faith’s phylogenetic diversity in HIV-negative and HIV-positive subjects by CD4 T cell count. Statistical analysis was performed in QIIME using two-sample, non-parametric t-tests with Monte Carlo permutations. Error bars indicate SEM. p ≤ 0.05 = * Principle Coordinate Analysis (PCoA) plots of the weighted UniFrac distances colored by HIV status and (D) CD4 T cell count or (E) ART treatment. See also Figure S3, S4, Table S7.

Figure 5. Differentially abundant bacterial taxa

Variance stabilization techniques were utilized to determine discriminant bacterial OTUs. Dots represent one OTU assigned to the indicated bacterial family (column label). The position of the dot on the y-axis indicates the abundance of that OTU (positive or negative) in relation to the opposing group. Bacterial OTUs with significant differences between (A) HIV-negative and CD4 <200 as well as (B) CD4 >200 versus <200 were graphed by log2 fold-change (y-axis) and grouped by family association (x-axis). Coloring indicates bacterial phyla to which the OTUs belong. See also Tables S5, 6.