• COVID-19
• ENTERIC BACTERIAL MICROFLORA
• IMMUNE RESPONSE

As reviewed by us recently,1 an increasing number of studies suggest that the gut microbiota is an important regulator of immune responses to vaccination. Consistent with these data, several recent clinical studies,2–4 including three published in Gut,5–7 uncover correlations between the composition of the faecal/gut microbiota and antibody responses to different COVID-19 vaccines. Additionally, recent antibiotic usage has also been associated with a lower seroconversion rate following BNT162b2 vaccination.8 These studies suggest that the gut microbiota plays a significant role in regulating optimal immune responses to COVID-19 vaccines. Despite these important studies, the causal and mechanistic links between the gut microbiota and responses to COVID-19 vaccines remain to be elucidated.

To investigate these links, we turned to two well-established models used previously to assess the relationship between the gut microbiota and responses to vaccination; antibiotic-treated and germ-free (GF) mice.9–11 We first assessed Spike-specific and receptor-binding domain (RBD)-specific IgG responses (online supplemental material 1) in BNT162b2-vaccinated antibiotics treated (ABX) mice, relative to vaccinated untreated (No ABX) SPF mice (two 3 µg doses intramuscular injection (i.m.) 2 weeks apart). Despite significant depletion of the gut microbiota at the time of primary vaccination (online supplemental figures S1A), Spike/RBD-specific IgG responses were not significantly different between ABX and No ABX mice before, or after, the secondary vaccination (figure 1A-B, online supplemental figures S1B). Next, we assessed T cell cytokine responses following in vitro stimulation with an overlapping Spike peptide pool at 6 weeks postboost. While CD4⁺ T cells demonstrated minimal cytokine secretion (online supplemental figures S1C-D), CD8⁺ T cells mounted robust Spike-specific recall responses, however, there was no significant difference in cytokine secretion between ABX and No ABX mice (figure 1C-D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Depletion of the microbiome in antibiotic-treated (ABX) or germ-free (GF) mice does not impair adaptive immune responses to BNT162b2 vaccination. C57BL/6J mice were untreated (No ABX) or treated (ABX) with vancomycin in their drinking water for 7 days prior to administration of PBS (Mock) or two doses of 3 µg of BNT162b2 by intramuscular injection, 2 weeks apart. (A) S1-specific and (B) receptor-binding domain (RBD)-specific IgG antibodies in the serum of mock and vaccinated mice were assessed in the weeks following vaccination by ELISA. (C) Representative flow cytometry analysis of IFNγ and TNFα secretion by CD8⁺ T cells from ABX and No ABX mice after in vitro stimulation (4 hours) of splenocytes with an overlapping Spike peptide pool. Splenocytes were collected 6 weeks postboost. (D) The frequency of IFNγ- and TNFα-secreting CD8⁺ T cells among CD3⁺CD8⁺ cells. (E) S1-specific and (F) RBD-specific IgG antibodies in the serum of BNT162b2-vaccinated conventional (SPF), faecal microbiota transplant (FMT) and germ-free (GF) mice in the weeks following vaccination. The number of Spike-specific IgG antibody secreting cells in the (G) spleen and (H) bone marrow as determined by ELISpot assay. (I) Representative flow cytometry analysis of IFNγ and TNFα secretion by CD8⁺ T cells from SPF, FMT and GF mice after in vitro stimulation (4 hours) of splenocytes with an overlapping Spike peptide pool. Splenocytes were collected 6 weeks postboost. (J) The frequency of IFNγ-secreting and TNFα-secreting CD8⁺ T cells among CD3⁺CD8⁺ cells. (K) Representative flow cytometry analysis of Spike_539-546-specific CD8⁺ T cells in the spleens of the indicated mice at 6 weeks postboost. (L) The frequency and total number of Spike_539-546-specific CD8⁺ T cells in the spleens of the indicated mice. n=8 mice/group. Data are represented as mean±SEM. One-way ANOVA was used to assess statistical significance. ANOVA, analysis of variance; ns, not significant.

We next considered that complete depletion of the gut microbiota in GF mice might have a greater impact on immune responses to the BNT162b2 vaccine. We therefore assessed Spike-specific IgG responses in BNT162b2-vaccinated GF mice in comparison to conventional SPF mice and GF mice that were recolonised via a faecal microbiota transplant (FMT). As we observed in ABX mice, total IgG responses to vaccination were not significantly impaired in GF mice (figure 1E-F). There was also no significant difference in the number of Spike-specific IgG antibody secreting cells recovered from the spleen and bone marrow at 6 weeks post-boost (figure 1G-H). Additionally, there was no significant difference in the pseudovirus (Wuhan-Hu-1 and Omicron) neutralising capacity of BNT162b2-induced antibodies at 2 weeks postboost (online supplemental figures S1E-F). Interestingly, there may be some differences in class switching in GF relative to SPF mice, as IgG1 and IgG2c responses were significantly higher and lower, respectively, in GF mice at specific time points postvaccination (online supplemental figures S1G-H). At most time points, however, there were no significant differences between GF and SPF mice.

The number and proportion of germinal centre B (GCB) and T follicular helper (Tfh) cells in the vaccine-draining iliac lymph node (iliLN) were also not significantly different in GF, compared with SPF mice, at 2 weeks postboost (online supplemental figures S1I-L). In the spleens of GF mice, however, there was a significant increase in the frequency and number of total GCB but not Tfh cells, relative to SPF mice (online supplemental figures S1M-N). Spike-specific CD8⁺ T cell responses were also assessed by intracellular cytokine staining and were not significantly different in GF mice (figure 1I-J). Finally, the magnitude of the CD8⁺ T cell response to vaccination was assessed using Spike_539-546 tetramers. This analysis revealed no significant differences in the frequency and number of Spike-specific T cells in the spleen 6 weeks postboost (figure 1K-L), or in the number of Spike-specific CD8⁺ T cells recovered from the peripheral blood, lungs, or vaccine-draining lymph nodes (online supplemental figure S1O).

We hypothesised that the microbiome might play a more important role in immune responses to lower doses of BNT162b2, so we repeated the experiments vaccinating mice with a lower dose (two 1 µg doses i.m., 2 weeks apart) (online supplemental figures S2A-B). Spike/RBD-specific IgG titers in GF mice were not significantly different compared with SPF mice, though interestingly IgG responses were impaired in FMT mice at 6 weeks postboost, but not at earlier time points (online supplemental figures S2C-D). There were no significant differences in the total number or proportion of GCB or Tfh cells in the vaccine-draining lymph nodes (online supplemental figures S2E-F). Furthermore, there was no significant difference in Spike-specific T cell cytokine recall responses (online supplemental figures S2G-H) or in the number/proportion of Spike_539-546-specific CD8⁺ T cells recovered from the spleen, lungs, or vaccine-draining lymph nodes (online supplemental figures S2I-K). Similar results were observed when ABX and No ABX mice were vaccinated with two 0.2 µg doses (data not shown).

In summary, our study comprehensively assesses antibody and T cell responses to the BNT162b2 mRNA vaccine in ABX-treated, GF, FMT and conventional SPF mice. Surprisingly, given both preclinical data showing impaired responses to other vaccines in similar models9–11 and the recent clinical studies outlined above, antibody and T cell responses to the BNT162b2 mRNA vaccine were not significantly impaired in either GF or antibiotic-treated mice. While our data demonstrate that, in mice, the microbiota is not required for optimal B or T cell responses to the BNT162b2 vaccine they do not preclude the possibility that specific microbes in the human microbiota could influence immune responses to this vaccine in humans. Our data do, however, emphasise the need for further research to prove that associations between the gut microbiota and immune responses to COVID-19 vaccines are causal and not simply correlative.