Longitudinal gut microbiota composition of South African and Nigerian infants in relation to tetanus vaccine responses

Abstract

Infants who are exposed to HIV but uninfected (iHEU) have higher risk of infectious morbidity than infants who are HIV-unexposed and uninfected (iHUU), possibly due to altered immunity. As infant gut microbiota may influence immune development, we evaluated the effects of HIV exposure on infant gut microbiota and its association with tetanus toxoid vaccine responses. We evaluated the gut microbiota of 82 South African (61 iHEU and 21 iHUU) and 196 Nigerian (141 iHEU and 55 iHUU) infants at <1 and 15 weeks of life by 16S rRNA gene sequencing. Anti-tetanus antibodies were measured by enzyme-linked immunosorbent assay at matched time points. Gut microbiota in the 278 included infants and its succession were more strongly influenced by geographical location and age than by HIV exposure. Microbiota of Nigerian infants, who were exclusively breastfed, drastically changed over 15 weeks, becoming dominated by Bifidobacterium longum subspecies infantis. This change was not observed among South African infants, even when limiting the analysis to exclusively breastfed infants. The Least Absolute Shrinkage and Selection Operator regression suggested that HIV exposure and gut microbiota were independently associated with tetanus titers at week 15, and that high passively transferred antibody levels, as seen in the Nigerian cohort, may mitigate these effects. In conclusion, in two African cohorts, HIV exposure minimally altered the infant gut microbiota compared to age and setting, but both specific gut microbes and HIV exposure independently predicted humoral tetanus vaccine responses.IMPORTANCEGut microbiota plays an essential role in immune system development. Since infants HIV-exposed and uninfected (iHEU) are more vulnerable to infectious diseases than unexposed infants, we explored the impact of HIV exposure on gut microbiota and its association with vaccine responses. This study was conducted in two African countries with rapidly increasing numbers of iHEU. Infant HIV exposure did not substantially affect gut microbial succession, but geographic location had a strong effect. However, both the relative abundance of specific gut microbes and HIV exposure were independently associated with tetanus titers, which were also influenced by baseline tetanus titers (maternal transfer). Our findings provide insight into the effect of HIV exposure, passive maternal antibody, and gut microbiota on infant humoral vaccine responses.

Keywords: HIV-exposed uninfected infants; Nigeria; South Africa; gut microbiota; tetanus toxoid; vaccine response.

Fig 1

Geographical location strongly affects gut microbiota among African infants in the first week of life. (A) Heatmap of the top 20 taxa in the gut microbiota of South African (n = 63) and Nigerian (n = 141) infants in the first week of age. Study site, HIV exposure status, and community cluster types (based on PAM clustering; k = 3) are shown in annotation bars. (B) Comparison of α-diversity (Shannon index) between South African (n = 63) and Nigerian (n = 141) infants during the first week of life. (C) PCoA and PERMANOVA (Bray-Curtis dissimilarity) of gut microbiota during the first week of age (South African, n = 63; Nigerian, n = 141), colored by study site and shaped by community groups, based on PAM clustering (k = 3). ****P < 0.0001.

Fig 2

Longitudinal transition of gut microbiota is distinct among infants in South Africa and Nigeria. (A) The transition of α-diversity (Shannon index) of infant gut microbiota over the first 15 weeks of age in South Africa (n = 82) and Nigeria (n = 196). (B) PCoA and PERMANOVA (Bray-Curtis dissimilarity) of gut microbiota at 1 week and 15 weeks of age, colored by study site and shaped by visit (South African, n = 82; Nigerian, n = 196). (C) Alluvial plot showing the transition of cluster groups from week 1 to week 15 at each study site (South African, n = 82; Nigerian, n = 196). Samples were grouped according to PAM clustering (k = 3), indicated by color. *P < 0.05; ****P < 0.0001.

Fig 3

Infants’ gut microbial succession over the first 15 weeks differs substantially between the study sites. Relative abundance plot of most abundant 30 taxa of South African (n = 82) and Nigerian (n = 196) infants at the 1 week and 15 weeks of age. Each column represents individual participants. B. longum subspecies infantis and B. longum subspecies longum are indicated as the same color (green).

Fig 4

Passive maternal antibody and HIV exposure are both associated with infant TT vaccine response. (A) Scatter plot and Spearman’s rank correlation coefficients (R) of anti-tetanus IgG titers (IU/mL) between Nigerian mothers (y-axis; n = 191) and their infants at week 1 (x-axis; n = 191). Dots and lines of best fit are colored by HIV exposure status. (B) Comparison of anti-tetanus IgG titers between iHEU (n = 197) and iHUU (n = 72) at week 1 and week 15. P-values comparing anti-tetanus IgG titers were adjusted for multiple comparisons using the Benjamini-Hochberg method. W1, 1 week of age; ns, not significant. *P < 0.05.

Fig 5

HIV exposure status and gut microbiota are independently associated with TT vaccine response. (A) Correlation analysis of infants’ anti-tetanus IgG titers (IU/mL) measured at 15 weeks of age and α-diversity (Shannon index) at each study site and visit (South African, n = 65; Nigerian, n = 170). Spearman’s rank correlation coefficients (R) are indicated on each panel. (B) Rank-transformed top 50 ASVs (at either week 1 or week 15), HIV exposure status, and anti-tetanus IgG titer data at week 1 were used as explanatory variables for the LASSO regression to assess the association with TT vaccine response at 15 weeks of age. Each model was constructed separately based on geographical location and time point. The optimal coefficient tuning parameter (lambda.min) was chosen using 10-fold cross-validation. Selected variables and their glmnet coefficients were plotted. Color of the bars represents taxonomy at the family level. Week 1 ASVs and HIV exposure status were associated with week 15 TT vaccine response among South African infants. No variables were selected for the Nigerian cohort. W15, 15 weeks of age.