Ipsilateral or contralateral boosting of mice with mRNA vaccines confers equivalent immunity and protection against a SARS-CoV-2 Omicron strain

ABSTRACT Boosting with mRNA vaccines encoding variant-matched spike proteins has been implemented to mitigate their reduced efficacy against emerging SARS-CoV-2 variants. Nonetheless, in humans, it remains unclear whether boosting in the ipsilateral or contralateral arm with respect to the priming doses impacts immunity and protection. Here, we boosted K18-hACE2 mice with either monovalent mRNA-1273 (Wuhan-1 spike) or bivalent mRNA-1273.214 (Wuhan-1 + BA.1 spike) vaccine in the ipsilateral or contralateral leg after a two-dose priming series with mRNA-1273. Boosting in the ipsilateral or contralateral leg elicited equivalent levels of serum IgG and neutralizing antibody responses against Wuhan-1 and BA.1. While contralateral boosting with mRNA vaccines resulted in the expansion of spike-specific B and T cells beyond the ipsilateral draining lymph node (DLN) to the contralateral DLN, administration of a third mRNA vaccine dose at either site resulted in similar levels of antigen-specific germinal center B cells, plasmablasts/plasma cells, T follicular helper cells, and CD8+ T cells in the DLNs and the spleen. Furthermore, ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214 vaccines conferred similar homologous or heterologous immune protection against SARS-CoV-2 BA.1 virus challenge with equivalent reductions in viral RNA and infectious virus in the nasal turbinates and lungs. Collectively, our data show limited differences in B and T cell immune responses after ipsilateral and contralateral site boosting by mRNA vaccines that do not substantively impact protection against an Omicron strain. IMPORTANCE Sequential boosting with mRNA vaccines has been an effective strategy to overcome waning immunity and neutralization escape by emerging SARS-CoV-2 variants. However, it remains unclear how the site of boosting relative to the primary vaccination series shapes optimal immune responses or breadth of protection against variants. In K18-hACE2 transgenic mice, we observed that intramuscular boosting with historical monovalent or variant-matched bivalent vaccines in the ipsilateral or contralateral limb elicited comparable levels of serum spike-specific antibody and antigen-specific B and T cell responses. Moreover, boosting on either side conferred equivalent protection against a SARS-CoV-2 Omicron challenge strain. Our data in mice suggest that the site of intramuscular boosting with an mRNA vaccine does not substantially impact immunity or protection against SARS-CoV-2 infection.

vaccines (e.g., mRNA, viral-vectored, inactivated, nanoparticle, and protein subunitbased) targeting the SARS-CoV-2 spike protein were deployed with billions of doses now administered globally.Many of these vaccines are effective against the severe disease, hospitalization, and death caused by SARS-CoV-2, with an estimated 20 million lives saved during the first year of vaccine rollout (1).Despite the success of the COVID-19 vaccines, multiple waves of SARS-CoV-2 variants [e.g., B.1.1.7 (Alpha), B.1.351(Beta), B.1.617.2 (Delta), BA.1, BA.5, XBB.1.5,and currently circulating KP.3 (Omicron)] continue to emerge, each with different constellations of amino acid substitutions, deletions, and insertions in the spike protein and elsewhere in the genome.As these successive variants often show increased transmissibility and greater antibody escape potential, they pose challenges to the efficacy of existing vaccines.Indeed, the antigen-shifted Omicron variants compromised the protective efficacy of immune responses elicited by vaccination or prior infection with non-Omicron strains (2)(3)(4)(5).
Beyond changes in the spike protein of variants, waning immunity poses a separate challenge to vaccine efficacy.Although mRNA vaccines (Moderna mRNA-1273; Pfizer-BioNTech BNT162b2) induce durable germinal center B cells (GCB) responses for at least 6 months (6)(7)(8)(9), the serum neutralizing antibody responses wane within 3-6 months and decline further by 8 months, with an estimated half-life of approximately 60 days (10)(11)(12).Waning humoral antibodies and the emergence of resistant, highly transmissible variants correlate with an increased frequency of breakthrough infections in vaccinated individuals (13,14).
To address concerns arising from reduced efficacy of vaccines against current and future variants, third, fourth, fifth, and even sixth doses (herein, boosters) of monovalent (Wuhan-1 or more recently, XBB.1.5) or bivalent mRNA vaccines encoding the historical (Wuhan-1) and variant-matched (BA.1 or BA.5) spike protein have been implemented.Indeed, mRNA booster vaccines have been shown to increase the magnitude and breadth of neutralizing antibodies against SARS-CoV-2 (15)(16)(17)(18)(19)(20).One frequent question regarding boosting is which arm should be used and whether the site injected affects immunogenicity and protection with mRNA or other COVID-19 vaccines.A study in mice with alum-adjuvanted influenza protein antigens showed that ipsilateral boost ing with a homologous vaccine induced recall germinal center responses with B cells having greater levels of somatic hypermutation and antibodies with higher avidity and cross-reactivity than boosting at a remote site (21).In comparison, data from human studies with Pfizer-BioNTech BNT162b2 mRNA vaccine are less certain, as two studies reached different conclusions.Whereas Ziegler et al. reported that ipsilateral boosting with homologous BNT162b2 mRNA vaccine elicited higher serum neutralizing titers 2 weeks later compared to contralateral boosting (22), Fazli et al. showed that contralateral boosting with a homologous mRNA vaccine was associated with higher binding and neutralizing antibody titers at 8 and 14 months (23).Moreover, a study in mice with a two-dose regimen of BNT162b2 mRNA vaccine reported that ipsilateral compared to contralateral leg boosting produced higher numbers of GCBs that bound the receptorbinding domain (RBD) of an ancestral strain in the draining lymph node (DLN) and higher levels of antibody with enhanced affinity against RBD of ancestral and Omicron strains but only at the early (day 9) but not late (week 19) time points (24).However, the impact of ipsilateral vs contralateral boosting on T cell responses or protective immunity in the context of homologous or heterologous SARS-CoV-2 challenge has not been tested.
Here, we evaluated the impact of ipsilateral and contralateral leg boosting with preclinical versions of historical (mRNA-1273, Wuhan-1 spike) and bivalent (mRNA-1273.214,Wuhan-1 + BA.1 spike) mRNA vaccines on adaptive B and T cell responses and protection against challenge with a BA.1 variant in susceptible K18-hACE2 transgenic mice.We observed that ipsilateral and contralateral boosting elicited comparable levels of serum spike-specific antibody as well as antigen-specific B and T cell responses.The differential impact of boosting site on protective response in the context of virus challenge was limited, as ipsilateral and contralateral boosting with either mRNA-1273 or mRNA-1273.214vaccines conferred similar levels of virological protection against BA.1 infection.

Antibody responses after boosting with mRNA vaccines
To compare serum antibody responses following vaccine boosting on the ipsilateral or contralateral side, cohorts of 7 to 9-week-old female K18-hACE2 transgenic mice were immunized intramuscularly in the left hind leg twice over a 3-week interval with a preclinical version of mRNA-1273 (Fig. 1A) that encodes for the prefusion-stabilized spike protein of SARS-CoV-2 Wuhan-1 strain (25).Animals were boosted 12 weeks later in the corresponding site in the left (ipsilateral) or right (contralateral) leg with 0.25 µg of mRNA-1273 or a preclinical bivalent mRNA-1273.214vaccine composed of a 1:1 mixture of mRNA-1273 and the BA.1-matched mRNA-1273.529vaccine (Fig. 1A).Boosting with these two different vaccines enabled analysis of the effects of immunization site on the breadth of the immune response.Serum was collected 1 day before and 4 weeks after boosting, and IgG responses against recombinant Wuhan-1 and BA.1 pre-fusion-stabi lized spike (S-2P; Fig. 1B through E) or RBD (Fig. 1F through I) proteins were determined by enzyme-linked immunosorbent assay (ELISA).Prior to boosting, as expected, similar serum IgG titers were detected among all four groups against Wuhan-1 spike (Fig. 1B), and the levels against BA.1 spike were approximately seven-to eightfold lower (P < 0.01; Fig. 1C).Boosting with either mRNA-1273 or mRNA-1273.214elicited enhanced IgG responses against the spike (Fig. 1B through E) and RBD (Fig. 1F through I) of Wuhan-1 and BA.1.For both mRNA-1273 and mRNA-1273.214boosts, lower serum IgG titers were observed against the BA.1 spike (4-5-fold, P < 0.01) and RBD (17-20-fold, P < 0.0001) than the corresponding Wuhan-1 proteins.In pre-and post-boost comparisons, boosting with mRNA-1273 increased IgG binding titers against Wuhan-1 spike (8-10-fold, P < 0.01, Fig. 1D) or RBD (six-to eightfold, P < 0.0001, Fig. 1H) with smaller increases against BA.1 RBD.In comparison, mRNA-1273.214induced larger increases against the BA.1 RBD (11-15-fold, P < 0.01, Fig. 1I).In the context of the booster location, ipsilateral and contralateral injections with mRNA-1273 or mRNA-1273.214induced similar IgG binding titers against either Wuhan-1 or BA.1 protein.
We next characterized the inhibitory effects of pre-and post-boost serum on SARS-CoV-2 infectivity using an established focus-reduction neutralization test (FRNT) (26) and authentic SARS-CoV-2 WA1/2020 D614G and BA.1 viruses (Fig. 2A through D; Fig. S1).Because of limited amounts of sera, we started dilutions at 1/60, which is just above the estimated threshold level of neutralization associated with protection in humans (27).As expected, 12 weeks after the primary immunization series with mRNA-1273 and before boosting, the serum neutralizing titers against WA1/2020 D614G were equivalent among all four groups (Fig. 2A).However, as observed in previous studies (16,17,28), serum from mice receiving two doses of mRNA-1273 showed poor neutralizing activity against BA.1, with most falling below the limit of detection of the assay (Fig. 2B).One month after boosting with either mRNA-1273 or mRNA-1273.214,serum neutralizing titers against WA1/2020 D614G rose five-to eightfold (P < 0.001; Fig. 2A and C) with minimal differences observed after ipsilateral or contralateral leg boosting.In comparison, ipsilateral or contralateral boosting with mRNA-1273 showed marginally increased neutralizing titers against BA.1, which did not achieve statistical significance (Fig. 2B and D).As expected for a variant-matched vaccine, boosting with mRNA-1273.214induced three-to fivefold higher neutralizing responses against BA.1 (P < 0.001), with ipsilateral boosters showing slightly higher titers than contralateral boosters, although these differences did not attain statistical significance (Fig. 2B and D).Overall, only small differences in neutralizing activity were observed when comparing the site (ipsilateral vs contralateral) of mRNA vaccine boosting.

B cell responses after boosting with mRNA vaccines
To assess further the impact of the site of boosting on immune responses, we interro gated B cells in the DLN and spleen.In this set of experiments, we applied the same vaccination scheme (Fig. 1A) but boosted with a higher 1 µg dose of mRNA-1273 or bivalent mRNA-1273.214either on ipsilateral (left) or contralateral (right) side; a higher booster dose was used to enhance our ability to detect spike-specific B cell responses by flow cytometry and to overcome the strong imprinting effects of mRNA vaccines (29)(30)(31), so we could assess effects on de novo responses.Seven days after boosting, we harvested the left and right inguinal DLNs as well as the spleen to characterize GCBs (CD19 + IgD low GL7 + Fas + ) and plasmablast/plasma cell (PB/PC: CD19 + IgD low TACI + CD138 + ) responses (Fig. 3 and 4; Fig. S2 to S7).Ipsilateral and contralateral leg boosting with mRNA-1273 or mRNA-1273.214vaccines induced seven-to eightfold higher frequencies (P < 0.0001) and 100-200-fold higher numbers (P < 0.0001) of GCBs in their respective DLN (ipsilateral, left; contralateral, right) than in naïve mice (Fig. 3A; Fig. S3A).Boosting with either vaccine induced lower GCB responses in the non-draining, contralateral LN with only two-to fivefold higher total numbers of GCB compared to unvaccinated mice (Fig. 3A).These data suggest that mRNA vaccine boosting preferentially induces GC reactions in the DLN, which is consistent with data from human studies (7,9).Nonetheless, mRNA-1273 and mRNA-1273.214vaccines generally elicited similar overall GCB responses after ipsilateral and contralateral boosts.We next profiled antigen-specific GCB responses using fluorescently labeled Wuhan-1 and BA.1 spike protein and flow cytometry (Fig. 3B through G; Fig. S3B and C).Ipsilat eral and contralateral boosting with mRNA-1273 or mRNA-214 vaccines both elicited robust GCB responses in DLNs against Wuhan-1 (Fig. 3B through D) and BA.1 (Fig. 3E through G) spike proteins.Regardless of the boost location or mRNA vaccine used, higher numbers and frequencies of spike-specific GCBs were detected in the DLN on the side of the boost.When we compared the mRNA-1273 and mRNA-1273.214response in the DLN where the dominant GC reaction occurred, mRNA-1273 elicited approximately twofold higher numbers (Fig. 3C) and frequency (Fig. S3B) of GCBs that bound Wuhan-1 spike than mRNA-1273.214.In comparison, mRNA-1273.214elicited approximately two-to threefold higher numbers (Fig. 3F) and frequency (Fig. S3C) of GCBs that bound BA.1 spike than mRNA-1273.Ipsilateral and contralateral leg boosting elicited relatively equivalent overall numbers of Wuhan-1 or BA.1 spike reactive GCBs in the DLNs regardless of vaccines being used (Fig. 3D and G), with substantially reduced responses observed in LNs on the side opposite of boosting (Fig. 3C and F).
We further gated the spike-binding GCB populations (Fig. S4A and B) to determine whether GCBs responses were Wuhan-1 spike-specific (Fig. S4C and D), cross-reactive against Wuhan-1 spike and BA.1 spike (Fig. S4E and F), or BA.1 spike-specific (Fig. S4G  and H).We observed two-to threefold higher numbers of Wuhan-1 spike-specific GCBs after mRNA-1273 boosting than with mRNA-1273.214boosting in the corresponding draining LN (Fig. S4C and D).Whereas the BA.1 spike-binding GCBs in the draining LN of mice boosted with mRNA-1273 were overwhelmingly cross-reactive against Wuhan-1 spike (Fig. S4F through H, left graph), mice boosted with mRNA 1273.214 elicited both cross-reactive and BA.1-specific GCBs (Fig. S4F and H, right graph).Regardless of the vaccines used, ipsilateral and contralateral leg boosting elicited relatively equivalent overall numbers of type-specific and cross-reactive GCBs in the DLNs (Fig. S4D, F and H).We also assessed total and antigen-specific PB/PC responses in the DLNs (Fig. 4; Fig. S3D through F) and observed a pattern like the GCB response.Ipsilateral and contrala teral leg boosting with mRNA-1273 or mRNA-1273.214vaccines elicited greater total (Fig. 4A; Fig. S3D) and spike-specific (Fig. 4B through G; Fig. S3E and F) PB/PC responses in the DLN on the same side of the boost.mRNA-1273 elicited a higher frequency (Fig. S3E) and number of PB/PCs that bound to Wuhan-1 spike than mRNA-1273.214(Fig. 4C), and reciprocally, mRNA-1273.214elicited a greater frequency (Fig. S3F) and number of PB/PCs that bound to BA.1 spike (Fig. 4F).In terms of the effect of the boosting site, ipsilateral and contralateral leg boosting elicited relatively similar numbers of Wuhan-1-or BA.1-spike reactive PBs/PCs in the respective DLNs regardless of the vaccine used (Fig. 4D and G), with lower responses in LNs achieved at the site opposite of boosting (Fig. 4C and F).We performed additional gating (Fig. S5A and B) to determine whether PBs/PCs were Wuhan-1 spike-specific (Fig. S5C and D), Wuhan-1 and BA.1 spike cross-reactive (Fig. S5E and F), or BA.1 spike-specific (Fig. S5G and H).Boosting with mRNA-1273 or mRNA-1273.214elicited comparable numbers of Wuhan-1 spike-specific (Fig. S5C and D) or cross-reactive PBs/PCs in the corresponding draining LN (Fig. S5E and  F).Whereas the BA.1 spike-binding PBs/PCs of mice that received mRNA-1273 boosting were cross-reactive against Wuhan-1 spike (Fig. S5F through H, left graph), mice receiving the mRNA-1273.214booster had both cross-reactive and BA.1 spike-specific PBs/PCs (Fig.

S5F through H, right graph).
We also profiled the total antigen-specific B cell response in the LN by gating on activated B cells (CD19 + IgD low ; Fig. S6A and D).Ipsilateral or contralateral leg boosting elicited relatively equivalent numbers of Wuhan-1 or BA.1 spike-reactive B cells in the DLNs (Fig. S6C and F), with reduced responses observed in LNs on the side opposite of boosting (Fig. S6B and E).Overall, total antigen-specific B cell responses against homologous or heterologous spike antigens in the DLN were comparable regardless of whether the mRNA vaccine booster occurred ipsilateral or contralateral to the site of the primary vaccination series.
We also measured the total and spike-specific PB/PC response in the spleen 7 days after boosting (Fig. S7).Ipsilateral or contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines induced 5-10-fold higher numbers (P < 0.0001) of total PB/PC in the spleen than in unvaccinated mice (Fig. S7A).Mice boosted with either mRNA-1273 or mRNA-1273.214vaccine had six-to ninefold higher numbers (P < 0.05) of antigen-spe cific PB/PCs that bound spike of Wuhan-1 compared to BA.1 (Fig. S7B and C).However, as we observed in the DLN, ipsilateral or contralateral boosting with mRNA-1273 or mRNA-1273.214induced equivalent levels of PB/PCs in the spleen that bound Wuhan-1 (Fig. S7B) and BA.1 (Fig. S7C) spike proteins.

T cell responses after boosting with mRNA vaccines
Since T cell immunity also contributes to protection against SARS-CoV-2 infection and disease (8,(32)(33)(34)(35), we interrogated their responses after mRNA vaccine boosting (Fig. 5; Fig. S8 and S9).At 7 days after ipsilateral or contralateral leg boosting with mRNA-1273 or mRNA-1273.214vaccines, similar or slightly higher numbers of total CD3 + T cells were present in the respective DLN (ipsilateral, left; contralateral, right) compared to non-boosted mice (Fig. 5A).We next evaluated spike-specific CD8 + T responses using class I MHC tetramers displaying a conserved immunodominant peptide (S 539-546 ; VNFNFNGL).Ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines elicited higher frequencies and numbers of tetramer + CD8 + T cells in the DLN, with reduced responses observed on the side opposite of boosting (Fig. 5B through D).Ipsilateral and contralateral mRNA vaccine boosting elicited comparable numbers of tetramer + CD8 + T cells in the DLNs regardless of the vaccine used (Fig. 5E).Ipsilateral or contralateral boosting with mRNA-1273 or mRNA-1273.214mRNA also elicited comparable frequencies and numbers of tetramer + CD8 + T cells in the spleen (Fig. 5F).Follicular T helper cells (T FH : CD4 + CD45 + CXCR5 + PD-1 + ) shape B cell responses by instructing GC reactions including the process of somatic hypermutation and affinity selection (36).We assessed the impact of ipsilateral and contralateral boosting on T FH cell responses in the LNs at 7 days post boosting (Fig. S9A through D).Like that observed with CD8 + T cell responses, ipsilateral and contralateral leg boosting with mRNA-1273 or mRNA-1273.214vaccines elicited higher frequencies (Fig. S9A and B) and numbers (Fig. S9C, 3-10-fold, P < 0.05) of T FH cells in the respective DLN, with reduced responses observed in the LN on the side opposite of boosting (Fig. S9B and C).Ipsilateral and contralateral boosting elicited equivalent numbers of T FH cells in the DLNs regardless of the vaccine used (Fig. S9D).Likewise, ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214mRNA elicited comparable frequencies and numbers of T FH cells in the spleen (Fig. S9E).

Effect of ipsilateral or contralateral boosting on protection against BA.1 challenge
To evaluate the protective activity of the mRNA vaccines following ipsilateral or contralateral boosting, the cohorts of K18-hACE2 mice were challenged 9 weeks later via an intranasal route with 10 4 focus-forming units (FFUs) of BA.1 (Fig. 6A).One limita tion of these studies is that BA.1 is inherently less pathogenic in rodents (37)(38)(39)(40) and replicates to lower levels in both the upper and lower respiratory tracts of K18-hACE2 mice without the development of clinical signs of disease compared to earlier variants (16,37); as such, we focused our analysis on virological endpoints.Nasal turbinates and lungs were harvested on day 4 after BA.1 infection and assayed for viral RNA levels by quantitative reverse transcription-PCR (qRT-PCR) (Fig. 6B and C).In the nasal turbinates of control mRNA vaccine immunized mice, although some variability was observed, moderate amounts of BA.1 viral RNA were detected (approximately 5 × 10 4 copies of N transcript per milligram; Fig. 6B).Ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines conferred protection in the nasal turbinates, with 100-1,000-fold reductions (P < 0.05) in viral RNA compared to the control vaccina ted mice.In the lungs of control vaccinated K18-hACE2 mice, 10 5 -10 7 copies of viral RNA per milligram of tissue were detected (Fig. 6C).Ipsilateral and contralateral boosting with either mRNA-1273 or mRNA-1273.214substantially reduced (mRNA-1273, ipsilateral or contralateral, 90-190-fold; mRNA-1273.214,ipsilateral or contralateral 370-1,000-fold, P < 0.0001) levels of viral RNA in the lung compared to the control vaccinated mice with no statistically significant differences conferred by the site of boosting (P > 0.05).Consistent with these data, no differences in infectious BA.1 were detected by plaque assay after ipsilateral or contralateral boosting with mRNA-1273 or mRNA-1273.214,with each vaccine and immunization site resulting in significant reductions (14-26-fold, P < 0.0001) compared to control mRNA vaccinated and challenged controls (Fig. 6D).

DISCUSSION
One of the key immunization strategies to mitigate waning immunity and reduced efficacy against SARS-CoV-2 variants that escape neutralization is boosting with vaccines that encode variant spike proteins.A frequent question regarding the clinical practice of boosting is which arm should be used, and whether boosting in the alternating arm impacts immunogenicity and protection against SARS-CoV-2.In this study, we evaluated in mice the impact of ipsilateral or contralateral site boosting of mRNA vaccines on the adaptive immune responses and protection against SARS-CoV-2 challenge.We observed that ipsilateral and contralateral boosting elicited comparable overall levels of serum spike-specific antibody as well as antigen-specific B and T cell responses.The differential impact of boosting site on protective response in the context of virus challenge was limited, as ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines both conferred similar levels of virological protection against BA.1 infection in the respiratory tract.
Our data showed that ipsilateral and contralateral boosting elicited comparable levels of anti-spike IgG in serum at 4 weeks after immunization.Functional analysis of antibody responses revealed similar neutralizing activities against SARS-CoV-2 WA1/2020 D614G and BA.1 viruses with small differences that did not achieve statistical significance between booster site regimens.Our results in K18-hACE2 mice with COVID-19 mRNA vaccines generally are consistent with a two-dose vaccination study in conven tional C57BL/6 mice using homologous influenza hemagglutinin (H1-HA) antigen, which showed comparable levels of serum anti-HA IgG at 8 days post boost, although that study did not examine neutralizing antibody activities (21).Our results contrast with this study when analyzing the effects of a heterologous antigen boost.In the influenza vaccine study, ipsilateral, but not contralateral, boosting with heterologous H3-HA enhanced H1-binding serum IgG levels, whereas contralateral boosting with H3-HA did not.While we did not observe such effects when comparing ipsilateral and contrala teral administration of monovalent (mRNA-1273) and bivalent (mRNA-1273.214) mRNA vaccine boosters, the bivalent vaccine contains the original mRNA-1273, which likely affects recall responses against the original Wuhan-1 spike antigen.Future studies with a fully heterologous monovalent boost [e.g., mRNA-1273.529(BA.1) or mRNA-1273.815(XBB.1.5)]may allow for more direct comparisons to the influenza immunization experiments.
Our serum antibody data showed some similarities and differences with a recent study in C57BL/6 mice that used a two-dose homologous BNT162b2 (Pfizer-BioNTech) mRNA vaccination regimen (24).That study showed that ipsilateral boosting with BNT162b2 mRNA vaccine produced higher titers of serum antibodies with enhanced affinity against RBD of Wuhan-1 and Omicron strains at early (day 9) but not the late (week 19) timepoints post boost.However, like our results, no difference was observed in overall serum IgG levels after ipsilateral or contralateral boosting when the analysis was performed with the spike proteins at either the early or later timepoints.Beyond pre-clinical studies in small animals, the data on the impact of ipsilateral vs contrala teral boosting on antibody responses in humans remain inconclusive.Two observatio nal studies conducted among SARS-CoV-2-uninfected individuals who received the BNT162b2 mRNA vaccine reached different conclusions.Ziegler et al. reported that spike-specific IgG levels did not differ between two groups that received a second dose of BNT162b2 on the ipsilateral (n = 147) or contralateral side (n = 156), although inhibitory activity was higher in the group boosted on the ipsilateral side at 2 weeks after the second dose based on an assay that measured inhibition of RBD binding to hACE2 (22).In contrast, Fazli et al. reported that contralateral boosting (n = 440) with BNT162b2 mRNA vaccine was associated with slightly higher spike-and RBD-specific serum IgG titers compared to ipsilateral boosting (n = 507), and this effect increased over time from a 1.1-fold to a 1.4-fold effect by 14 months.Contralateral boosting also resulted in increases (1.3-fold to 4.0-fold) in neutralizing antibody titers against B.1.1.529at 8 and 14 months, with lesser effects against the WA1/2020 D614G strain (no difference at 8 months and twofold difference at 14 months) (23).
To investigate the mechanisms by which boosting site selection might influence humoral immunity, we interrogated B cells in the LNs and spleen.Ipsilateral and contralateral leg boosting with mRNA-1273 or mRNA-1273.214vaccines elicited robust GCB responses in the DLNs against Wuhan-1 and BA.1 spike proteins at 7 days post boost, with reduced responses observed in the LNs on the site opposite of boosting.As expected, bivalent mRNA-1273.214elicited higher numbers of GCBs that specifically bound BA.1 spike than mRNA-1273.Whereas contralateral boosting extended the GCB response to the contralateral DLN, ipsilateral or contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines overall induced similar levels of GCBs that bound Wuhan-1 or BA.1 spike in the DLNs.Our results are consistent with fate-mapping studies in mice, which showed that recruitment of antigen-experienced memory B cells to secondary GCs is inefficient (41,42), as the progeny of primed GC B cells account for only a subset of total GCBs.Our analysis of PBs/PCs also revealed that boosting with the bivalent mRNA-1273.214mRNA vaccine elicited higher numbers of PBs/PCs that bound specifically to BA.1 spike than the mRNA-1273 vaccine.Ipsilateral and contralateral boosts elicited equivalent PBs/PCs responses in the respective DLNs and spleen, which is consistent with the similar spike-and RBD-specific IgG responses in serum by the two booster regimens.However, our results both agree and contrast with the two-dose influenza immunization study in mice (21), which showed that ipsilateral and contrala teral boosting with homologous influenza hemagglutinin H1 antigen elicited compara ble total GCB but better quality of H1-reactive GCBs in the DLN; in that study, ipsilateral boosting with the heterologous H3-HA antigen increased the breadth of GCBs that bound to H1-HA.Our GCB results also differ from the study with two homologous doses of BNT162b2 mRNA in mice (24), which showed that ipsilateral boosting induced a higher number of GCBs that bound RBD of the Wuhan-1 strain in the ipsilateral DLN than that contralateral boosting elicited in the contralateral DLN.
Despite increasing data on antibody and B cell responses induced after immunization at different boosting sites in mice and humans, the results remain inconsistent and challenging to develop paradigms.Multiple factors in study design likely contribute to discrepant results among studies.(i) Vaccine platforms could impact responses.Whereas the influenza study used HA protein-based vaccines to interrogate the impact of ipsilateral or contralateral boosting, we used monovalent and bivalent mRNA vaccines.Vaccine platforms can affect the extent of GC responses in the DLN, as mRNA vaccines can induce persistent GC responses that last months compared to traditional proteinbased platforms (9,(43)(44)(45)(46). (ii) The frequency of boosters and interval between primary immunization could change the dynamics of B cell and antibody responses generated from ipsilateral and contralateral boosting.Whereas the previous studies in mice boosted after one-dose prime immunization, our study evaluated the impact of boost location following a two-dose primary vaccination series.Indeed, a third COVID-19 mRNA vaccine dose in humans has been associated with an increase in RBD-specific memory B cells both from expanded clones present after the second dose and the emergence of new clones (47,48).The boosting interval also differs between our study (12 weeks) and the prior BNT162b2 mRNA vaccine study (3 weeks) in mice (24), which has been shown to independently impact antibody responses (49,50).(iii) The relative dosing of the booster could impact responses.Higher doses could result in prolonged antigen retention in the follicular dendritic cells in the DLN that differentially shape B cell and antibody respon ses (51,52).(iv) The timing of assessment.The two human studies on ipsilateral and contralateral boosting may have reached different conclusions because they evaluated antibody responses at different time points, ranging from 2 weeks to 14 months (22,23).Analogously, the study in C57BL/6 mice with two homologous doses of BNT162b2 mRNA vaccine showed different affinity measurements at early and late time points after ipsilateral and contralateral boosting (24).Together, these data suggest that at the early phase after boosting, memory B cells in the DLN on the ipsilateral side with respect to priming likely differentiate into PB/PC rapidly after ipsilateral boosting, resulting in earlier serum antibody production than contralateral boosting.In comparison, after contrala teral boosting, in the contralateral DLN, newly recruited antigen-specific B cells from the naive pool expand, affinity mature, and differentiate into plasmablasts to produce antibodies.(v) Antigenic distance between primary and boost vaccines could impact the responses.As shown in the influenza study (21), ipsilateral boosting with a heterologous antigen increased cross-reactive antibody responses more than contralateral boosting, which might be better at overcoming immune imprinting.The bivalent vaccine in our study contains the original mRNA-1273, which likely affects recall responses against the original Wuhan-1 spike antigen.For mRNA vaccines, more study of the effects of ipsilateral and contralateral boosting is likely needed for evaluating fully heterologous monovalent vaccines.
T cell immune responses contribute to protection against SARS-CoV-2 infection and disease, especially in the setting of poor neutralizing antibody responses against variants (8,(32)(33)(34)(35).We interrogated spike-specific CD8 and total T FH cell responses locally and systemically.Ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines both elicited enhanced spike-specific CD8 + T cell and overall T FH responses in the spleen and DLN at 7 days post boosting, with reduced responses observed in LN at the site opposite of boosting.Notwithstanding these data, ipsilateral and contralateral boosting with mRNA-1273 or mRNA-1273.214vaccines elicited comparable spike-specific CD8 + T cells and overall T FH responses in the spleen and respective DLN.These results are consistent with a study in C57BL/6 mice that used self-amplifying mRNA vaccine encoding the influenza A virus nucleoprotein (53) and showed compara ble antigen-specific CD4 + and CD8 + T cells in the spleen 28 days after ipsilateral and contralateral boosting.Our results with T FH cells differ from this study, as we charac terized these cells in the spleen and respective DLNs at 7 days after ipsilateral and contralateral boosting.Our CD8 + T cell results also differ from a study in humans, which showed that frequencies of spike-specific CD8 + T cell in the blood were higher after ipsilateral than contralateral boosting (22).The reasons for the differences are not clear beyond the different anatomic compartments (spleen vs blood) sampled.
We uniquely assessed the impact of ipsilateral and contralateral boosting on vaccine-mediated protection against virus challenge.Boosting with either mRNA-1273 or mRNA-1273.214efficiently reduced BA.1 viral RNA levels in the upper and lower respiratory tracts with a trend toward a higher level of protection in animals boosted with mRNA-1273.214,as observed previously (16).Notably, our data showed that the site of boosting with mRNA-1273 or mRNA-1273.214vaccines did not affect protec tion against BA.1 as judged by measurement of viral RNA and infectious virus in the nasal turbinates and lungs.Boosting at either site with mRNA-1273 or mRNA-1273.214reduced the levels of viral infection, with each vaccine and immunization site resulting in significant reductions compared to unvaccinated, challenged controls.Overall, these data in mice show that the limited differences in overall immune response induced by ipsilateral or contralateral boosting with historical (mRNA-1273) or variant-matched bivalent (mRNA-1273.214)vaccines do not substantively impact virological protection against challenge by a SARS-CoV-2 Omicron strain.
We acknowledge several limitations in our study.(i) We evaluated humoral or cellular immune responses at only one time point following boosting.Longitudinal studies after ipsilateral and contralateral boosting are needed to assess whether the dynamics of immune responses change over time.(ii) A higher 1 mg booster dose was used in our analysis of B and T cell responses.In pilot experiments, the response after a lower 0.25 mg dose was detectable in most but not all mice.We selected the higher booster dose to discern statistical differences in immune cell populations and also because it enables mice to overcome the strong imprinting effects of mRNA vaccines (29)(30)(31), so de novo BA.1 spike-specific responses could be measured.(iii) Our study evaluated immune responses after one ipsilateral or contralateral boost after a primary vaccination series.How additional boosting on the contralateral side shifts the recall responses remains to be determined.(iv) We used monovalent and bivalent mRNA vaccine formulations in our boosting schemes in part because of when the studies were initiated.Experiments that test monovalent heterologous boosting with more contemporary vaccines (encoding spikes from BA.1, BA.5, XBB.1.5,or emerging KP.2/KP.3strains) are needed to determine how the boosting site affects or potentially overcomes immune imprinting.(v) We performed studies only in uninfected mice receiving vaccines.Future studies on boost location are needed in the setting of antecedent SARS-CoV-2 infection to extrapolate our findings to conditions of hybrid immunity.(vi) We did not analyze the impact of boost location on bone marrow-derived long-lived plasma cells (LLPCs).LLPCs constitutively secrete high levels of antibodies for extended periods of time and dominantly contribute to serum antibody levels at times remote from vaccination or infection.(vii) Finally, studies with different vaccine platforms, as well as other animal models and ultimately humans, are required for corroboration.
Our studies in K18-hACE2 mice provide evidence that ipsilateral and contralateral boosting with mRNA vaccines elicit comparable immune responses, and this was associated with equivalent control of infection by a SARS-CoV-2 Omicron challenge strain.Thus, at least in mice, the boost site location for mRNA vaccines appears to have limited impacts on vaccine immunogenicity and protection against SARS-CoV-2.

Mice
Heterozygous K18-hACE2 C57BL/6J mice (strain: 2B6.Cg-Tg(K18-ACE2)2Prlmn/J, Cat # 034860) were obtained from The Jackson Laboratory.Virus inoculations were performed under anesthesia that was induced and maintained with ketamine hydrochloride and xylazine, and all efforts were made to minimize animal suffering.Experiments were neither randomized nor blinded.

mRNA vaccine and lipid nanoparticle production process
Sequence-optimized mRNA encoding prefusion-stabilized Wuhan-1 (mRNA-1273) and BA.1 (mRNA-1273.529, the Omicron gene component in mRNA-1273.214)SARS-CoV-2 S-2P proteins were synthesized in vitro using an optimized T7 RNA polymerase-mediated transcription reaction with complete replacement of uridine by N1m-pseudouridine (58).Bivalent mRNA-1273.214vaccine is a 1:1 bench side mix of separately formula ted mRNA-1273 and mRNA-1273.529vaccines.A non-translating control mRNA was synthesized and formulated into lipid nanoparticles as previously described (59).The transcription reaction included a DNA template containing the immunogen open-read ing frame flanked by 5' untranslated region (UTR) and 3' UTR sequences and was terminated by an encoded poly A tail.After RNA transcription, the cap-1 structure was added using the vaccinia virus capping enzyme and 2ʹ-O-methyltransferase (New England Biolabs).The mRNA was purified, sterile filtered, and kept frozen at -20°C until further use.
The mRNA was encapsulated in a lipid nanoparticle as described previously (60).Vials were filled with formulated lipid nanoparticle and stored frozen at -20°C until further use.The vaccine product underwent analytical characterization, which included the determination of particle size and polydispersity, encapsulation, mRNA purity, doublestranded RNA content, osmolality, pH, endotoxin, and bioburden, and the material was deemed acceptable for in vivo study.

Viral antigens
Recombinant soluble spike and RBD proteins from Wuhan-1 and BA.1 SARS-CoV-2 strains were expressed as described (61,62).Recombinant proteins were produced in Expi293F cells (Thermo Fisher) by transfection of DNA using the ExpiFectamine 293 Transfection Kit (ThermoFisher).Supernatants were harvested 3 days post-transfec tion, and recombinant proteins were purified using Ni-NTA agarose (Thermo Fisher).The proteins were then buffer exchanged into phosphate-buffered saline (PBS) and concentrated using Amicon Ultracel centrifugal filters (EMD Millipore).

ELISA
Purified recombinant Wuhan-1 or BA.1 spike or RBD proteins were coated onto 96-well Maxisorp clear plates at 2 mg/mL (spike) or 4 µg/mL (RBD) in 50 mM Na 2 CO 3 pH 9.6 (50 μL) overnight at 4°C.Coating buffers were aspirated, and wells were blocked with 200 µL of 1× PBS + 0.05% Tween-20 (PBST) +2% bovine serum albumin (BSA) +0.02% NaN 3 (Blocking buffer, PBSTBA) overnight at 4°C.Sera were serially diluted in blocking buffer and added to the plates.Plates were incubated for 1 h at room temperature and then washed three times with PBST, followed by the addition of 50 µL of 1:2,000 dilution of horeseradish peroxidase (HRP)-conjugated anti-mouse IgG (Southern Biotech Cat. #1030-05) in PBST.Following a 1 h incubation at room temperature, plates were washed three times with PBST, and 50 µL of 1-Step Ultra TMB-ELISA was added (Thermo Fisher Cat.#34028).Following a 2-5 min incubation, reactions were stopped with 50 µL of 2 M H 2 SO 4 .The absorbance of each well at 450 nm was determined using a microplate reader (BioTek) within 5 min of the addition of sulfuric acid.The endpoint serum dilution was calculated with curve fit analysis of optical density values for serially diluted sera with a cut-off value set to six times the mean of the background signal.

Mouse experiments
Seven-to 9-week-old female K18-hACE2 C57BL/6 mice were immunized 3 weeks apart with 0.1 μg of mRNA-1273 vaccine in 50 µL of PBS via intramuscular injection in the left hind leg.Animals were boosted 12 weeks later in the left (ipsilateral) or right (contrala teral) leg with 0.25 µg of mRNA-1273 or bivalent mRNA-1273.214vaccine (1:1 mixture of mRNA-1273 and the BA.1-matched mRNA-1273.529).Animals were bled 1 day before and 4 weeks after boosting for antibody analysis.
To evaluate the protective activity of the mRNA vaccines 9 weeks after boosting, the cohorts of K18-hACE2 mice were challenged via an intranasal route with 10 4 FFUs of BA.1.Animals were euthanized at 4 dpi.Nasal turbinates and lungs were harvested for virological and immunological analyses.For B and T cells analysis, the same vaccination scheme was applied, but animals were boosted with 1 µg of mRNA-1273 or bivalent mRNA-1273.214 in the ipsilateral (left) or contralateral (right) leg.Seven days after boosting, the left and right inguinal LN and spleen were harvested for immune cells analysis.

Measurement of viral RNA
Tissues were weighed and homogenized with zirconia beads in a MagNA Lyser instrument (Roche Life Science) in 1 mL of DMEM medium supplemented with 2% heat-inactivated FBS.Tissue homogenates were clarified by centrifugation at 10,000 rpm for 5 min and stored at −80°C.RNA was extracted using the MagMax mirVana.Total RNA isolation kit (Thermo Fisher) on the Kingfisher Flex extraction robot (Thermo Fisher).RNA was reverse transcribed and amplified using the TaqMan RNA-to-CT 1-Step Kit (Thermo Fisher).Reverse transcription was carried out at 48°C for 15 min followed by 2 min at 95°C.Amplification was accomplished over 50 cycles as follows: 95°C for 15 s and 60°C for 1 min.Copies of SARS-CoV-2 N gene RNA in samples were determined using a published assay (64).

DATA AVAILABILITY
All data supporting the findings of this study are available within the paper and from the corresponding author upon request.All requests for resources and reagents should be directed to and will be fulfilled by the corresponding author.All reagents will be made available on request after completion of a Materials Transfer Agreement.Preclinical mRNA vaccines can be obtained under an MTA with Moderna (contact: Darin Edwards, darin.edwards@modernatx.com).

ETHICS APPROVAL
All virus experiments were performed with Institutional Biosafety Committee approval in an approved biosafety level 3 (BSL-3 or A-BSL3) facility at Washington University School of Medicine.Animal studies were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.The protocols were approved by the Institutional Animal Care and Use Committee at the Washington University School of Medicine (assurance number A3381-01).

ADDITIONAL FILES
The following material is available online.

FIG 1
FIG 1 Serum IgG responses in K18-hACE2 mice following boosting.Seven-to 9-week-old female K18-hACE2 mice were immunized with a primary two-dose vaccination series spaced 3 weeks apart in the left hind leg with 0.1 μg of mRNA-1273.Animals were boosted 11-12 weeks later in the left (ipsilateral) or right (contralateral) leg with 0.25 μg of mRNA-1273 or bivalent mRNA-1273.214.One day before (pre-boost) or 4 weeks after boost (post-boost), serum was collected (Continued on next page)

FIG 1 (Full 5 FIG 2
FIG 1 (Continued) and IgG titers were determined by ELISA.(A) Scheme of immunizations and blood collection.(B and C) Serum IgG responses against the spike of Wuhan-1(B) or BA.1 (C).(D and E) Paired analysis of serum IgG titers of pre-and post-boost against WA1/2020 D614G (D) and BA.1 (E).(F and G) Serum IgG responses against RBD domain of Wuhan-1 (F) or BA.1 (G).(H and I) Paired analysis of serum IgG titers of pre-and post-boost against WA1/2020 D614G (H) and BA.1 (I) [n = 15-18, two experiments, column heights indicate geometric mean titers (GMT), and dotted lines show the limit of detection (LOD)].GMTs or fold-changes are indicated above the corresponding graphs.Statistical analyses: B, C, F, and G. Mann-Whitney test: D, E, H, and I. Wilcoxon matched signed-rank test.*P < 0.05; **P < 0.01, P < 0.001, and P < 0.0001.

FIG 3
FIG 3 Germinal center B cell responses in the lymph nodes following boosting with mRNA-1273 or mRNA-1273.214.Seven-to 9-week-old female K18-hACE2 mice were immunized with a primary two-dose vaccination series spaced 3 weeks apart in the left hind leg with mRNA-1273.Animals then were boosted 11-12 weeks later in the left (ipsilateral) or right (contralateral) leg with 1 µg of mRNA-1273 or mRNA-1273.214.Seven days after boosting, inguinal LNs from the left and right sides were analyzed for GCB responses by flow cytometry.(A) Quantification of total number of CD19 + IgD low GL7 + Fas + GCBs in respective LNs.Note that the number of GCBs in the LNs on the respective side (L-LN or R-LN) from unvaccinated mice are shown in each graph with the different vaccines (mRNA-1273 or mRNA1273.214)for comparison purposes.(B) Representative flow cytometry scatter plots of Wuhan-1 spike-reactive GCBs.(C) Total numbers of Wuhan-1 spike-reactive GCBs in the respective LNs.(D) Comparison of numbers of Wuhan-1 spike-reactive GCBs in the DLNs.(E) Representative flow cytometry scatter plots of BA.1 spike-reactive GCBs.(F) Total number of BA.1 spike-reactive GCBs in respective LNs.(G) Comparison of numbers of BA.1 spike-reactive GCBs in the DLNs.Data are from two experiments (n = 7-8, each data point represents an individual mouse, column heights indicate geometric mean values, and dotted lines show the LOD).Data in D and G correspond to data in C and F, respectively, and are replotted for direct statistical comparison.Statistical analyses: A; one-way analysis of variance (ANOVA) with Tukey's post-test: C, D, F, and G; unpaired two-tailed Mann-Whitney test: ns, not significant; *P < 0.05, P < 0.01, P < 0.001, and P < 0.0001.

FIG 4 9 FIG 5
FIG 4 Plasmablast and plasma cell responses in the lymph nodes following boosting with mRNA-1273 or mRNA-1273.214.K18-hACE2 mice were immunized and boosted as described in Fig. 3. Seven days after boosting, inguinal LNs from the left and right sides were analyzed for PB/PC responses by flow cytometry.(A) Total number CD19 + IgD low CD138 + TACI + PBs/PCs.Note that the number of PB/PCs in the LNs on the respective side (L-LN or R-LN) from unvaccinated mice are shown in each graph with the different vaccines (mRNA-1273 or mRNA-1273.214)for comparison purposes.(B) Representative flow cytometry scatter plots of Wuhan-1 spike-reactive PBs/PCs.(C) Total number of Wuhan-1 spike-reactive PBs/PCs in respective LNs.(D) Comparison of numbers of Wuhan-1 spike-reactive PBs/PCs in the DLNs.(E) Representative flow cytometry scatter plots of BA.1 spike reactive PBs/PCs.(F) Total number of BA.1 spike reactive PB/PC cells in respective LNs.(G) Comparison of numbers of BA.1 spike reactive PBs/PCs in the DLNs.Data are from two independent experiments (n = 7-8, each data point represents an individual mouse, column heights indicate geometric mean values, and dotted lines show the LOD).Data in D and G correspond to data in C and F, respectively, and are replotted for direct statistical comparison.Statistical analyses: A; one-way ANOVA with Tukey's post-test: C, D, F, and G; unpaired two-tailedMann-Whitney test: ns, not significant; *P < 0.05, P < 0.01, P < 0.001, and P < 0.0001.

FIG 6
FIG 6 Protection against BA.1 infection in K18-hACE2 mice following boosting with mRNA-1273 or mRNA-1273.214.(A) Scheme of immunizations and virus challenge.Seven-to 9-week-old female K18-hACE2 mice were immunized with a primary two-dose vaccination series spaced 3 weeks apart in the left hind leg with 0.1 μg of mRNA-1273.Animals were boosted 11-12 weeks later in the left (ipsilateral) or right (contralateral) leg with 0.25 μg of mRNA-1273 or bivalent mRNA-1273.214.Mice immunized with mRNA-control (left hind leg) were used as a negative control group.Nine weeks after boosting, mice were challenged via an intranasal route with 10 4 FFUs of SARS-CoV-2 BA.1.Viral RNA levels were determined at 4 dpi in the nasal turbinates (B) and lungs (C).(D) Infectious virus in the lungs.Data are from two independent experiments (n = 15-18, each data point represents an individual mouse, and column heights indicate median values).Note that the viral yield in the nasal turbinates or lungs (B-D) after immunization with mRNA-control is shown in each graph with the different vaccines (mRNA-1273 or mRNA1273.214)for comparison purposes.Statistical analyses: one-way ANOVA with Tukey's post-test: ns, not significant; *P < 0.05, P < 0.01, P < 0.001, and P < 0.0001.