Safety and immunogenicity of SARS-CoV-2 variant mRNA vaccine boosters in healthy adults: an interim analysis

The emergence of SARS-CoV-2 variants of concern (VOCs) and variants of interest (VOIs) with decreased susceptibility to neutralization has generated interest in assessments of booster doses and variant-specific vaccines. Clinical trial participants who received a two-dose primary series of the COVID-19 vaccine mRNA-1273 approximately 6 months earlier entered an open-label phase 2a study (NCT04405076) to evaluate the primary objectives of safety and immunogenicity of a single booster dose of mRNA-1273 or variant-modified mRNAs, including multivalent mRNA-1273.211. As the trial is currently ongoing, this exploratory interim analysis includes preliminary descriptive results only of four booster groups (n = 20 per group). Immediately before the booster dose, neutralizing antibodies against wild-type D614G virus had waned (P < 0.0001) relative to peak titers against wild-type D614G measured 1 month after the primary series, and neutralization titers against B.1.351 (Beta), P.1 (Gamma) and B.1.617.2 (Delta) VOCs were either low or undetectable. Both the mRNA-1273 booster and variant-modified boosters were safe and well-tolerated. All boosters, including mRNA-1273, numerically increased neutralization titers against the wild-type D614G virus compared to peak titers against wild-type D614G measured 1 month after the primary series; significant increases were observed for mRNA-1273 and mRNA-1273.211 (P < 0.0001). In addition, all boosters increased neutralization titers against key VOCs and VOIs, including B.1.351, P.1. and B.1.617.2, that were statistically equivalent to peak titers measured after the primary vaccine series against wild-type D614G virus, with superior titers against some VOIs. This trial is ongoing.

S everal severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccines targeting the viral spike (S) protein have been developed 1 . One such vaccine, mRNA-1273 (Moderna), a lipid nanoparticle-encapsulated mRNA vaccine encoding the S protein of the Wuhan-Hu-1 isolate with two proline mutations introduced to stabilize the S protein into the prefusion conformation, had an acceptable safety profile and induced anti-SARS-CoV-2 immune responses in phase 1 (NCT04283461) 2,3 and phase 2 (NCT04405076) 4 trials in adults. In the phase 3 Coronavirus Efficacy (COVE) trial (NCT04470427), mRNA-1273 provided 94% efficacy against symptomatic COVID-19 disease in more than 30,000 participants 5 . Subsequently, mRNA-1273 received emergency use authorization from several global regulatory bodies, including the U.S. Food and Drug Administration [6][7][8] .
In a previous study, the neutralizing capacity of sera collected from participants 7 d after completion of the mRNA-1273 primary series against VOCs 17,18 was assessed using a previously described research grade vesicular stomatitis virus (VSV)-based SARS-CoV-2 pseudovirus neutralization (PsVN) assay 19 . Neutralizing antibody titers measured in sera from eight mRNA-1273 phase 1 trial participants were reduced 2.1-to 8.4-fold against the B.1.617.2, P.1 and B.1.351 variants 17,18 . Currently, a neutralizing antibody titer threshold predictive of protection from SARS-CoV-2 infection in humans is unknown. However, the reduction of in vitro neutralizing antibody titers against variants relative to the wild-type D614G virus raises the possibility of breakthrough infections and waning efficacy for the current SARS-CoV-2 vaccines.
To address this potential risk, modified versions of the prototype mRNA-1273 vaccine that contain the genetic sequence of the variant S protein continue to be developed. These variant vaccines are designed to stimulate an immune response against key sites of neutralization that have been altered on the S protein of variant viruses and, in the case of a multivalent vaccine, simultaneously against the wild-type strain. Here we report data from an exploratory interim analysis of the preliminary safety and immunogenicity of single booster doses of mRNA-1273 (50 µg), modified mRNA-1273.351

Results
Participants. Among 186 participants who received two primary doses of mRNA-1273 (100 µg) in the blinded phase of the P201 study 4 and subsequently received one booster dose of mRNA-1273 (50 µg) in the mRNA-1273 booster phase, 20 participants were randomly selected for inclusion in this interim analysis based on visit assessments completed and sample availability of pre-booster sera (Fig. 1).
Among 14,711 participants who received two primary doses of mRNA-1273 (100 µg) in the phase 3 COVE trial 5 , 60 participants were selected to enter the mRNA-1273 variant booster phase of the P201 trial and received a single booster dose of mRNA-1273.351 (50 μg; n = 20), mRNA-1273.211 (50 μg; n = 20) or mRNA-1273.351 (20 μg; n = 20). As one participant from the 20-μg mRNA-1273.351 group was lost to follow-up at day 29 and, thus, excluded, the analysis sample size for this cohort is 19.
Safety. The percentages of participants with solicited local and systemic adverse reactions (ARs) were generally similar among the booster groups ( Fig. 2 (Fig. 3a), whereas the neutralization titers for B.1.351 were low or non-detectable before the booster dose across all groups assessed (Fig. 3b).
Neutralizing antibody titers against the wild-type D614G and B.1.351 viruses increased after each booster dose compared to day 1 titers (P < 0.0001 for all booster groups) (Fig. 3). Specifically, on day 29, geometric mean titers (GMTs) against the wild-type D614G virus were 16.7-, 11.3-, 46.4-and 9.2-fold higher than day 1 (pre-booster) titers in the mRNA-1273 (50 µg), mRNA-1273.351 (50 µg), mRNA-1273.211 (50 µg) and mRNA-1273.351 (20 µg) booster recipients, respectively (Fig. 3a). Similarly, on day 29, GMTs  In the mRNA-1273 booster phase, 20 participants who had received two injections of 100-μg mRNA-1273 completed the blinded phase and who went on to receive a single booster dose of 50-μg mRNA-1273 were selected for this preliminary analysis, with selection based on completion of day 15 visit assessments and immunogenicity sample availability. In the mRNA-1273 variant booster phase, enrollment was site specific and was based on predefined inclusion/exclusion criteria. Administration of booster doses occurred in a sequential manner (mRNA-1273 variant booster phase, which included the mRNA-1273.351 (50 µg) cohort; the mRNA-1273.211 (50 µg) cohort; and the mRNA-1273.351 (20 µg) cohort). a Unblinded or not unblinded to assigned treatment in the blinded portion of the phase 2a mRNA-1273 trial. b Fifteen participants declined to receive a booster. c Thirteen participants declined to receive a booster. d Twenty-two participants declined to receive an mRNA-1273 primary vaccination. against the B.1.351 variant were 34.9-, 61.6-and 33.7-fold higher than day 1 (pre-booster) titers in the mRNA-1273.351 (50 µg), mRNA-1273.211 (50 µg) and mRNA-1273.351 (20 µg) booster recipients, respectively (Fig. 3b). These observations included participants who did not have measurable neutralizing antibodies against the wild-type D614G or B.1.351 virus before the booster dose but showed increases in their neutralizing antibody titers after the booster dose (Fig. 3); statistical analyses of this subgroup alone were not performed.

D614G and VOCs neutralization 1 and 6 months after primary series.
An exploratory analysis of the kinetics of the immune response at 1 and 6 months after the primary mRNA-1273 vaccination series was conducted across the four groups using the VSV-based PsVN assay. This assay was previously used to evaluate the neutralizing activity of serum from participants who received mRNA-1273 in a phase 1 trial against SARS-CoV-2 wild-type virus and variants 18 .
One month after the primary series, wild-type D614G neutralizing antibody GMT ranged from 1,210 to 2,213 across participants in the 50-µg booster groups (Fig. 4) and was 2,758 in the mRNA-1273.351 20-µg booster group ( Supplementary Fig. 1). B.1.351 and P.1 neutralizing antibody GMTs were 13-to 14-fold lower and 5-to 6-fold lower, respectively, compared to wild-type D614G at the same time point in the 50-µg groups.
Simultaneous analysis of samples using the VSV PsVN assay showed that, approximately 6 months after the mRNA-1273 primary vaccination series, neutralizing antibody levels decreased (P < 0.0001) compared to peak titers against wild-type D614G measured 1 month after the primary series (GMTs against wild-type D614G were 6-to 7-fold lower, and GMTs against B.1.351 and P.1 were 24-to 69-fold lower) (Fig. 4a-c). Neutralizing antibody levels against B.1.351 and P.1 were below the lower limit of quantification (LLOQ) of the assay in ~44% and 30% of samples, respectively. Neutralization of the B.1.617.2 variant was also reduced (P < 0.0001) 6 months after the completion of the primary series (Fig. 4d). Sera from a random subset of the 20 participants in the mRNA-1273 booster group was used to assess neutralization of B.1.617.1 and B.1.617.2 (n = 11 for both) 6 months after the primary series and showed a 33-to 40-fold reduction in neutralizing antibody titers against B.1.617.1 and B.1.617.2 in comparison to peak titers measured against wild-type D614G 1 month after the primary series (full mRNA-1273 interim analysis cohort; n = 20). Neutralizing antibody titers against B.1.617.2 fell below the LLOQ of the assay in five of 11 samples.
D614G and VOCs neutralization after booster. Neutralizing antibody titers against the wild-type D614G virus were measured with the VSV-based PsVN assay using samples collected 2 weeks after the booster dose and were compared against wild-type D614G GMT benchmarks from samples collected 1 month after the primary series vaccination in each group. These benchmarks were used to determine whether the boosters reached the same neutralization level shown in the pivotal study where efficacy was demonstrated (that is, levels seen for wild-type D614G where 94% efficacy was measured) 5 Supplementary Fig. 1). Compared to the wild-type D614G benchmarks for each group, the booster vaccines yielded superior (mRNA-1273 and mRNA-1273.211) or equivalent (mRNA-1273.351) GMTs against the wild-type D614G virus. Wild-type D614G neutralization was 3.8-fold (P < 0.0001), 1.7-fold (P value not significant (NS)) and 4.4-fold (P < 0.0001) higher 2 weeks after 50-µg booster doses of mRNA-1273, mRNA-1273.351 and mRNA-1273.211, respectively, compared to peak titers against wild-type D614G measured 1 month after the primary series (Fig. 4a-c). All three boosters, including mRNA-1273, increased neutralization against VOCs or VOIs to levels that were statistically equivalent to the wild-type D614G benchmarks, with superior titers measured versus some VOIs. Of the three booster vaccines assessed, the multivalent mRNA-1273.211 had the greatest increase in GMTs against all VOCs. Neutralization titers against B. 1.351, P.1, B.1.427/B.1.429, B.1.526, B.1.617.1 and B.1.617.2 were 1.1-fold (P value NS), 1.4-fold (P value NS), 2.7-fold (P < 0.0001), 2.2-fold (P < 0.0001), 1.2-fold (P value NS) and 1.2-fold (P value NS) higher, respectively, 2 weeks after the mRNA-1273.211 booster compared to peak titers against the wild-type D614G measured 1 month after the primary series (Fig. 4c).

Discussion
This preliminary evaluation describes the antibody persistence of mRNA-1273 and the safety and immunogenicity of a booster dose  18 . These results are consistent with those reported in a study using a lentiviral PsVN assay, in which monitoring of neutralizing antibody levels was performed up to 6 months after completion of the mRNA-1273 primary series 19 . Reduction of neutralizing antibody titers against B.1.351 and P.1 was evident 1 month after the primary series to a greater degree than that observed 7 d after the primary series 18 , likely due to further affinity maturation of B cells and alteration of the available antibody repertoire. Additional reduction or complete loss of detectible levels of neutralizing antibody ~6 months after the primary series was evident against B.1.351, P.1 and B.1.617.2. The safety profiles after single booster injections of mRNA-1273 (50 µg), mRNA-1273.351 (20 or 50 µg) and mRNA-1273.211 (50 µg) were generally similar to those observed after the mRNA-1273 primary series in the previously reported phase 2 and 3 studies 12,14 . The most common systemic ARs after the booster doses were fatigue, headache, arthralgia and myalgia, which occurred at similar-to-lower frequencies for the boosters than after receipt of the mRNA-1273 (100 μg) primary series.
Booster vaccination with mRNA-1273, mRNA-1273.351 and mRNA-1273.211 induced strong anamnestic responses, indicative of a robust B cell memory response 13 . Neutralizing antibody titers against the wild-type D614G virus after a booster dose were up to 4.4-fold higher than peak titers after the primary series. Neutralizing antibody titers against several VOCs (that is, B.1.351, P.1 and B.1.617.2) increased after the booster dose, with titers against several variants approaching or exceeding those measured after the primary series against the wild-type D614G virus (Fig. 4). Increased titers against the VOCs suggest that further maturation of antibodies is feasible after a two-dose primary series of mRNA-1273, regardless of the composition of the booster dose. Furthermore, boosting with mRNA-1273.351 and mRNA-1273.211 appeared to produce numerically greater neutralizing antibody titers against the B.1.351 variant than with mRNA-1273, although formal conclusions regarding the significance of these differences cannot be made. The multivalent mRNA-1273.211 (50 µg) booster yielded a GMT ratio rise ≥1 against all VOCs and VOIs 2 weeks after the booster dose versus peak wild-type D614G titers measured 1 month after the primary series vaccination (Fig. 4c). This rise was significant for B.1.427/B.1.429 and B.1.526, indicating that variant neutralization GMTs after the booster were higher than peak wild-type D614G virus GMTs after the primary series in the samples from this cohort, potentially increasing breadth of coverage against VOCs or VOIs.  **** **** **** **** **** **** **** **** **** **** **** **** ****  Data are presented as the geometric mean neutralizing antibody titers with 95% confidence intervals. The titers for individual participants are indicated with circles. The fold increases in titers measured at days 15 and 29 versus titers measured before the booster dose are shown. The horizonal dotted lines indicate the LLOQ. Generalized linear model was used to compare neutralization titers among groups; log 10 titer was regressed on group, and an individual-specific random effect was included to account for individual specific variability. Two-sided t-test was used for post hoc group comparisons. Sidak's method was used to adjust the P values for multiple comparisons. Statistical significance was determined at α < 0.01. ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05. NAb, neutralizing antibody.
There are some limitations related to this preliminary analysis. First, the results presented here are based on treatment groups that were not randomized. Instead, participants were assigned sequentially, given the different time frames of availability of the new vaccine formulations. The sample size was small (n = 20 per group) to facilitate rapid initiation of additional studies, which are needed to support the safe and effective use of a COVID-19 booster vaccine and to inform ongoing vaccination preparedness strategies regarding the need for booster doses as the pandemic evolves. Participants in this interim analysis were predominantly white and non-Hispanic or Latino (95-100% across booster groups), limiting generalizability to other races and ethnicities. Sex distributions and mean body mass index and age were not equivalent across the four booster groups; however, these differences are expected given the small cohort sizes. Although the lentiviral-based PsVN assay used in this evaluation is validated, the VSV-based PsVN assay used in the evaluation of samples against variants is a research grade assay that has not been validated. Nevertheless, the high correlation between the clinically validated lentivirus-based PsVN assay and research grade VSV-based PsVN assay provides support for the utility of the latter in improving the efficiency and time needed to perform such clinical analyses. Moreover, although these data are encouraging, in the absence of a correlate of protection it cannot be definitively The emergence of SARS-CoV-2 variants and the ability of the virus to partially overcome natural or vaccine-induced immunity has served as a call to action. Although a correlate of protection has not been established for SARS-CoV-2 infection or COVID-19 disease, lack of detectable neutralization against VOCs after ~6 months in some participants might be indicative of waning protection. However, it should be noted that an anamnestic response upon viral exposure is likely based on the induction of immune memory from the booster dose. The mRNA platform approach against SARS-CoV-2 VOCs in this trial appears to be effective in developing wild-type and variant-specific booster vaccines, with boosters increasing neutralizing titers against the wild-type D614G virus and against key VOCs and VOIs. Of note, significantly higher neutralizing titers against wild-type D614G, B.    In a-d, the GMTs against the wild-type D614G and variants measured in participants before the booster dose or 2 weeks after the booster dose were evaluated versus peak titers measured against the wild-type D614G 1 month after the primary vaccination series. Data are presented as the geometric mean neutralizing antibody titers with 95% confidence intervals. The GMT fold change versus the peak titers against the wild-type D614G virus after the primary vaccination series are shown, with red indicating fold drop and blue indicating fold rise. Results from individual participants are represented as dots on each figure. For all panels, blue colored dots indicate participants whose samples were tested (n = 11), and white dots indicate the remaining participants whose samples were not tested (n = 9). The horizonal dotted line indicates the LLOQ; the solid gray line indicates the within-cohort GMT benchmark. Generalized linear model was used to compare neutralization titers among groups; log 10 titer was regressed on group, and an individual-specific random effect was included to account for individual specific variability. Two-sided t-test was used for post hoc group comparisons. Sidak's method was used to adjust the P values for multiple comparisons. Statistical significance was determined at α < 0.01. ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05. NAb, neutralizing antibody; NS, not significant.
booster. Further research is needed to determine the clinical significance of these preliminary results. Although this trial evaluated the performance of booster vaccines that encode the original strain or the B.1.351 S protein, this strategy could be employed in the future to vaccinate against new VOCs through the development of new variant-specific vaccines.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41591-021-01527-y.