The resident gut microbiome modulates the effect of synbiotics on the immunogenicity after SARS-COV-2 vaccination in elderly and diabetes patients

Abstract

The study aims to tackle the seed and soil microbiome and mechanisms that contribute to the effect of synbiotics in enhancing immunogenicity after SARS-CoV-2 vaccination in elderly and diabetic patients. Among 369 subjects who received 3 months of SIM01, a gut microbiota-derived synbiotic formula of three Bifidobacterium strains (B. adolescentis, B. bididum, and B. longum) or a placebo after the SARS-CoV-2 vaccines (mRNA vaccine BNT162b2 (Pfizer-BioNTech) or the inactivated vaccine Sinovac-CoronaVac), we performed metagenomic sequencing in stool samples of 280 vaccinees collected at baseline and 3-month postvaccination and metabonomic sequencing in 276 vaccinees collected at baseline and 1-month postvaccination. The open niche of autochthonous gut microbiota (lower levels of Bifidobacterium and decreased functional potential for carbohydrate metabolism) was associated with enhancing SIM01-contained species. The enrichment of three bifidobacterial species after 3 months of SIM01 intervention (BABBBL_fc) was positively correlated with the level of neutralizing antibodies to the BNT162b2 vaccine at 6-month postvaccination. The fold change of benzoic acid was positively correlated with BABBBL_fc in the BNT162b2 vaccinees, which was also implicated with SARS-CoV-2 surrogate virus neutralization test (sVNT)% levels at 1-month postvaccination. Importantly, SIM01 strain engraftment assessed by StrainPhlAn (A metagenomic strain-level population genomics tool) was associated with a higher fold change of three bifidobacterial species and could be predicted based on the baseline gut microbiome. Therefore, the resident gut microbiome affected the SIM01 engraftment, which was associated with the immunogenicity of SARS-CoV-2 BNT162b2 vaccines.

Fig. 1. Study design and change in gut community cluster from baseline to 3-month after the second vaccine dose (postvaccination, p.v.) after 3 months of SIM01 intervention.

A Study design. Eligible subjects received a gut microbiota-derived synbiotic formula (SIM01) or placebo for 3 months in all subjects aged ≥65 years or with type 2 diabetes mellitus (T2DM). All subjects received the COVID-19 vaccine, including either the mRNA vaccine BNT162b2 (Pfizer-BioNTech) or the inactivated vaccine Sinovac-CoronaVac, within 1 week of intervention. Study visits were arranged at 1-month, 3-month, and 6-month after completion of the second dose of the COVID-19 vaccine. B The Sanky diagram shows the transition of three gut community clusters between baseline and 3-month postvaccination (p.v.) following a 3-months SIM01 intervention in both placebo and SIM01 groups. The width of each flow corresponds to the number of subjects undergoing an intra- or inter-cluster transition. The height of each bar is proportional to the number of subjects in each group. p values were given by the Chi-square test. The first line of p values indicated the overall comparison (transition in three clusters before and after treatment). The second line of p values indicates the subgroup comparison transition between cluster 2 and other clusters (combined C1 and C3). C cluster. C The relative abundance of the Bifidobacterium genus in the samples with distinct gut community clusters. p values were given by the Wilcoxon test (two-sided). D, E The gut microbiome Shannon diversity (D) and the richness of observed species (E) in the samples with distinct gut community clusters. p values were given by the Wilcoxon test (two-sided). Elements on boxplots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5×IQR; points, samples. IQR interquartile range. * indicated p < 0.05, ** indicated p < 0.01, *** indicated p < 0.001, **** indicated p < 0.0001 after adjustment for multiple comparisons.

Fig. 2. The autochthonous gut microbiota affects the enhancement of SIM01-contained species.

A, B The changes in the sum relative abundance of three Bifidobacterium (A) or B. adolescentis (B) between baseline and 3-month postvaccination (p.v.) after subgrouping based on the increase or non-increase of SIM01-contained species within the SIM01 arm. Non-Increase, N = 34 each time point. Increase, N = 110 at each time point. C The depleted species at baseline in the gut of subjects who were characterized by an increase in SIM01-contained species after 3 months of treatment. Maaslin2 was used for identification after adjusting the gender, age, and T2D status. p < 0.05, FDR <0.25. D Correlation between fold change of SIM01 species and the overlap number of the KO-encoded in SIM01 strains in the baseline samples. Coefficients and p values of the correlations were given by Spearman’s correlation tests. E The heatmap showed the presence and absence of baseline KOs that were associated with the enhancement of SIM01-related species after 3 months of SIM01 intervention. Each row represents the baseline sample of each subject.

Fig. 3. SIM01-induced Bifidobacterium increase in the gut after 3 months of treatment is associated with higher neutralizing antibodies at 6 months after BioNTech mRNA vaccination.

A, B The correlation between the fold change of three Bifidobacterium in the gut after 3 months of SIM01 treatment and the anti-spike receptor-binding domain (RBD) (A) or the neutralizing antibody sVNT (%) (B) at 1 month after the second dose of BioNTech was examined using Spearman’s correlation test. C The correlation between the fold change of three Bifidobacterium in the gut after 3 months of SIM01 treatment and the neutralizing antibody sVNT (%) at 6 months after the second dose of BioNTech was examined using Spearman’s correlation test. N = 75. D The correlation between the anti-spike receptor-binding domain (RBD) at 1 month after the second dose of BioNTech and the neutralizing antibody sVNT (%) at 6 months after the second dose of BioNTech was examined using Spearman’s correlation test. Regression lines with 95% CI (orange area) were shown on scatter plots. E Logistic regression based on fold change of individual Bifidobacterium species or a combined model based on fold change of all three Bifidobacterium species adjusting age, gender, and T2D status was used to calculate the odds ratio (OR, 95% Confidence interval) values for the for high (≥60%, n = 32) vs. low sVNT levels (<60%, n = 43) responders among BioNTech (BNT162b2) vaccines at 6 month after second dose within SIM01 arm. Each OR was presented as a dot with a bar showing the 95% CI. p < 0.05 was presented as an orange dot. OR, odds ratio; T2D, type 2 diabetes mellitus; fc fold change. F AUROC (95% CI) values of models based on combined fold change of three species contained in SIM01 for high (≥60%) vs. low sVNT levels (<60%) at 6-month postvaccination adjusting age, gender, and T2D status. p.v., postvaccination. G The volcano plot shows the differential pathway between two groups of subjects with the non-increase or increase SIM01 species in the gut after 3 months of treatment within the SIM01 arm. Maaslin2 was used for identification after adjusting the gender, age, and T2D status. p < 0.05, FDR < 0.25. H Correlations between the fold change of the relative abundance of SIM01-contained three species and the fold change of differential metabolites between baseline and 1-month postvaccination in BioNTech vaccinees. The label within each frame: · indicated p < 0.1, * indicated p < 0.05, ** indicated p < 0.01 based on Spearman correlation analysis. The color indicates the correlation Rho value. I AUROC (95% CI) values of models based on the combined fold change of three species contained in SIM01 and the fold change of three metabolites between baseline and 1-month postvaccination for high (≥60%) vs. low sVNT levels (<60%) at 6-month postvaccination, adjusting age, gender, and T2D status. p.v. postvaccination, T2D type 2 diabetes mellitus.

Fig. 4. The engraftment of the SIM01 strain based on StrainPhlAn.

A Phylogenetic tree of dominant haplotypes of B. adolescentis per sample. B Subset of the phylogenetic tree of dominant haplotypes of B. adolescentis of the samples with SIM01 B. adolescentis strain as the dominant strain after 3 months of treatment. Orange circles and text represent the reference genomes of the SIM01 B. adolescentis strain. C The fold change of relative abundance of SIM01-contained three species in three groups of subjects (None: without any Bifidobacterium species contained in SIM01 that involved in the StrainPhlAn construction, None SIM01 strain: with Bifidobacterium species involved in the StrainPhlAn construction but not the same as the SIM01 strain, Any SIM01 strain: with at least one Bifidobacterium strain contained in SIM01 that involved in the StrainPhlAn construction). p values were given by the Wilcoxon test (two-sided). Points, samples. Elements on boxplots: center line, median; box limits, upper and lower quartiles; whiskers, 1.5×IQR. D AUROC (95% CI) values of models about SIM01 strain engraftment based on baseline differential gut microbial species. E The variable importance of features in the prediction model. F Correlation between the relative abundance of baseline SIM01-contained three Bifidobacterium species with the relative abundance of Anaerostipes hadrus. Coefficients and p values of the correlations were given by Spearman’s correlation tests.