Dysbiotic Fecal Microbiome in HIV-1 Infected Individuals in Ghana

Abstract

HIV-1 infected individuals under antiretroviral therapy can control viremia but often develop non-AIDS diseases such as cardiovascular and metabolic disorders. Gut microbiome dysbiosis has been indicated to be associated with progression of these diseases. Analyses of gut/fecal microbiome in individual regions are important for our understanding of pathogenesis in HIV-1 infections. However, data on gut/fecal microbiome has not yet been accumulated in West Africa. In the present study, we examined fecal microbiome compositions in HIV-1 infected adults in Ghana, where approximately two-thirds of infected adults are females. In a cross-sectional case-control study, age- and gender-matched HIV-1 infected adults (HIV+; n = 55) and seronegative controls (HIV-; n = 55) were enrolled. Alpha diversity of fecal microbiome in HIV+ was significantly reduced compared to HIV- and associated with CD4 counts. HIV+ showed reduction in varieties of bacteria including Faecalibacterium, the most abundant in seronegative controls, but enrichment of Proteobacteria. Ghanaian HIV+ exhibited enrichment of Dorea and Blautia; bacteria groups whose depletion has been reported in HIV-1 infected individuals in several other cohorts. Furthermore, HIV+ in our cohort exhibited a depletion of Prevotella, a genus whose enrichment has recently been shown in men having sex with men (MSM) regardless of HIV-1 status. The present study revealed the characteristics of dysbiotic fecal microbiome in HIV-1 infected adults in Ghana, a representative of West African populations.

Keywords: Ghana; HIV-1; dysbiosis; gut microbiome; sub‐Saharan Africa.

Figure 1

Top 10 abundant genera in fecal microbiome in all participants in the present study. Taxa bar plots showing the top 10 abundant genera in all participants (n = 110) are shown. Each bar represents frequencies of the genera in fecal microbiome of an individual. HIV-, HIV-1 uninfected individuals (n = 55); HIV+, HIV-1 infected individuals (n = 55).

Figure 2

Alpha diversity of fecal microbiome in HIV-1 uninfected and infected Ghanaian adults. (A) Comparison of Shannon diversities of fecal microbiome between HIV- (n = 55) and HIV+ (n = 55). HIV- showed significantly higher diversity than HIV+ (p < 0.001). (B) Comparison of Faith phylogenetic diversities between HIV- and HIV+. HIV- showed significantly higher diversity than HIV+ (p < 0.001). (C) Comparison of Shannon diversities among HIV-, HIV-1 infected controllers (viral load < 1,000 copies/ml) under ART for less than 2 years (ART<2 yrs, n = 10), HIV-1 infected controllers under ART for more than 2 years (ART>2 yrs, n = 22), and HIV-1 infected non-controllers (viral load > 1,000 copies/ml) under ART (Viremic, n = 20). HIV- showed significantly higher diversity than ART<2yrs (p < 0.01), ART>2yrs (p < 0.005), and Viremic (p < 0.01), respectively. (D) Comparison of Faith phylogenetic diversities among HIV-, ART<2 yrs, ART>2 yrs, and Viremic. HIV- showed significantly higher diversity than ART<2yrs (p < 0.005), ART>2yrs (p < 0.05), and Viremic (p < 0.05), respectively. ART<2yrs showed significantly lower diversity than Viremic (p < 0.05). (E) Comparison of Shannon diversities among HIV-, HIV+ on ART with Efavirenz (n = 42), and HIV+ on ART with Nevirapine (n = 10). (F) Comparison of Faith phylogenetic diversities among HIV-, HIV+ with Efavirenz, and HIV+ with Nevirapine. Significant difference was determined by Wilcoxon rank sum test (A, B) or Kruskal Wallis test with Benjamini, Krieger and Yekutieli FDR correction (C–F); *, **, ***, and **** indicate significant differences with p < 0.05, p < 0.01, p < 0.005, and p < 0.001, respectively.

Figure 3

Beta diversity of fecal microbiome in HIV-1 uninfected and infected Ghanaian adults. (A) Principal Coordinates Analysis (PCoA) of weighted Unifrac distances in fecal microbiome in HIV- and HIV+ individuals. (B) PCoA of Bray-Curtis distances in HIV- and HIV+. (C) PCoA of Aitchison distances in HIV- and HIV+. Seven representative amplicon sequence variants (ASV) influencing the observed segregation are shown. These analyses indicate segregation between HIV- and HIV+.

Figure 4

Difference in abundant taxa in fecal microbiome between HIV-1 uninfected and infected Ghanaian adults. (A) Data determined by linear discriminant analysis (LDA) effect size (LEfSe). Alpha value of 0.01 and 3.6 threshold on the logarithmic LDA score for discriminative features was used. P_, C_, O_, F_, G_ represent Phylum, Class, Order, Family, and Genus, respectively. (B) Cladogram. Green and Red indicate HIV- (Controls) and HIV+ (Cases), respectively.

Figure 5

Comparison of plasma markers between HIV-1 uninfected and infected Ghanaian adults. Comparisons of plasma I-FABP (A), sCD14 (B), and LBP (C) levels are shown. HIV+ (n = 55) showed significantly higher I-FABP and sCD14 levels than uninfected HIV- (n = 55) (p < 0.05 [*]; Wilcoxon rank sum test).

Figure 6

Analyses of correlation between fecal microbiome diversity and clinical markers. Asterisks indicate significant correlation determined by Spearman’s test; *, **, and *** indicate significant differences with p < 0.05, p < 0.01 and p < 0.005, respectively. Circle sizes are proportional to strength of correlation. Faith_pd, Faith’s phylogenetic diversity; shannon, Shannon’s index; ART_duration, duration under ART; Antibiotics_duration, duration since last Co-trimoxazole treatment; log_VL, viral load; sCD14, plasma soluble CD14 levels; IFABP, plasma intestinal fatty acid-binding protein levels; LBP, plasma lipopolysaccharide binding protein levels. n = 110 in faith_pd, shannon, sCD14, IFABP, and LBP; n = 55 (HIV+) in CD4 count, CD8 count, and CD4_CD8_ratio; n = 52 in ART_duration; n = 54 in Antibiotics_duration, n = 36 in log_VL (> 20 copies/ml).

Figure 7

Analyses of correlation of genera showing significant difference in abundance in HIV- and HIV+ with clinical and immunological markers and microbiome diversities. Faith_pd, Faith’s phylogenetic diversity (n = 110); shannon, Shannon’s index (n = 110); CD4 count (n = 55); CD8 count (n = 55); CD4_CD8_ratio (n = 55); ART_duration, duration under ART (n = 52); Duration since Co-trimoxazole use (n = 54); Log VL, viral load (> 20 copies/ml) (n = 36); sCD14, plasma soluble CD14 levels (n = 110); IFABP, plasma intestinal fatty acid-binding protein levels (n = 110); LBP, plasma lipopolysaccharide binding protein levels (n = 110). Asterisks indicate significant correlation determined by Spearman’s test; *, **, and *** indicate significant differences with p < 0.05, p < 0.01, and p < 0.005, respectively. Significantly abundant genera were determined by ANCOM and LEfSE. Genera significantly abundant in HIV+ are shown by red.