Influence of the intestinal microbiota on the immunogenicity of oral rotavirus vaccine given to infants in south India

Abstract

Oral rotavirus vaccines have consistently proven to be less immunogenic among infants in developing countries. Discrepancies in the intestinal microbiota, including a greater burden of enteropathogens and an altered commensal community composition, may contribute to this trend by inhibiting the replication of vaccine viruses. To test this possibility, we performed a nested case-control study in Vellore, India, in which we compared the intestinal microbiota of infants who responded serologically or not after two doses of Rotarix delivered at 6 and 10 weeks of age as part of a clinical trial (CTRI/2012/05/002677). The prevalence of 40 bacterial, viral, and eukaryotic pathogen targets was assessed in pre-vaccination stool samples from 325 infants using singleplex real-time PCR on a Taqman array card (TAC). In a subset of 170 infants, we assessed bacterial microbiota composition by sequencing the 16S rRNA gene V4 region. Contrary to expectations, responders were more likely than non-responders to harbor ≥1 bacterial enteropathogen at dose 1 (26% [40/156] vs 13% [21/157] of infants with TAC results who completed the study per protocol; χ², P = .006), although this was not apparent at dose 2 (24% [38/158] vs 23% [36/158]; P = .790). Rotavirus shedding after dose 1 was negatively correlated with the replication of co-administered oral poliovirus vaccine (OPV). We observed no consistent differences in composition or diversity of the 16S bacterial microbiota according to serological response, although rotavirus shedding was associated with slightly more bacterial taxa pre-vaccination. Overall, our findings demonstrate an inhibitory effect of co-administered OPV on the first dose of Rotarix, consistent with previous studies, but in the context of OPV co-administration we did not find a strong association between other components of the intestinal microbiota at the time of vaccination and Rotarix immunogenicity.

Keywords: Enteropathogens; Immunogenicity; Microbiota; Rotarix; Rotavirus.

Fig. 1

Association between concurrent pathogens and seroconversion after two doses of Rotarix. Prevalence of concurrent pathogens at (A) 6 weeks and (B) 10 weeks of age by seroconversion status. Pathogens present in at least 1% of the study population are included. (C, D) Pathogen count and mixed infection prevalence at (C) 6 weeks and (D) 10 weeks of age by seroconversion status. Mean pathogen counts are indicated by dotted lines. (E) Impact of concurrent enteropathogens at 6 and 10 weeks of age on the odds of seroconversion. Rotaviruses were excluded from analyses of pathogen groups, mixed infections, and pathogen count. ^*P < .05. Abbreviations: Bac, bacteria; EAEC, enteroaggregative Escherichia coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; EV, enterovirus; Euk, eukaryote; OR, odds ratio; Vir, virus; w, weeks.

Fig. 2

Association between concurrent pathogens and Rotarix replication. (A) Prevalence of concurrent pathogens at 6 weeks of age according to shedding status at 4 and/or 7 days after the first dose of RV1. Pathogens present in at least 1% of the study population are included. (B, C) Prevalence of Sabin viruses and NPEVs at 6 and 10 weeks of age according to shedding status. (D) Prevalence of enteroviruses at 10 weeks of age according to the shedding of ≥ 1 Sabin virus at 6 weeks. The shedding of Sabin viruses at 10 weeks of age was used as an indicator of take following the OPV dose administered at 6 weeks. ^*P < .05; ^**P < .005. Abbreviations: Bac, bacteria; EAEC, enteroaggregative Escherichia coli; EPEC, enteropathogenic E. coli; ETEC, enterotoxigenic E. coli; Euk, eukaryote; EV, enterovirus; NPEV, non-polio enterovirus; RV, rotavirus; Sabin+, positive for ≥1 Sabin serotype; Sabin-, negative for all Sabin serotypes; STEC, Shiga toxin-producing E. coli; Vir, virus; w, weeks.

Fig. 3

Association between microbiota composition and Rotarix response. (A) Phylum- and genus-level composition of the bacterial microbiota at 6 and 10 weeks of age. (B) OTU count and Shannon index (mean ± standard error) by rotavirus seroconversion status. (C) Unweighted Unifrac distances between 6-week samples, visualized via principal coordinates analysis. (D, E) Equivalent alpha and beta diversity plots are displayed with respect to shedding status after the 6-week RV1 dose. ^*P < .05. Abbreviations: OTU, 97%-identity operational taxonomic unit; PC, principal coordinate; RV, rotavirus; RV1, Rotarix; w, weeks.

Fig. 4

Impact of probiotic supplements on the bacterial microbiota. (A) Receipt of probiotics resulted in enrichment of a single OTU (classified as L. zeae). Mean relative abundance of this OTU in each study arm is indicated by a horizontal line, while prevalence is indicated by a cross. (B) OTU count and Shannon index (mean ± standard error) by study arm. (C) Unweighted Unifrac distances between 6-week samples, visualized via principal coordinates analysis. (D) Mean accuracy (±SD) across 100 iterations of the Random Forest algorithm for models predicting receipt of probiotics-only (upper) or zinc and probiotics (lower). OTU 21300 corresponds to the Lactobacillus strain that was enriched among probiotics recipients. (E, F) Highest-ranking taxa (and corresponding OTU IDs) by Random Forest importance score (mean decrease in accuracy ± SD) for models predicting receipt of (E) probiotics-only and (F) zinc and probiotics. ^*P < .05; ^**P < .005. Abbreviations: LGG, probiotics (Lactobacillus rhamnosus GG); OTU, 97%-identity operational taxonomic unit; PC, principal coordinate; w, weeks; Zn, zinc.