Cross-neutralization of Omicron BA.1 against BA.2 and BA.3 SARS-CoV-2

The Omicron SARS-CoV-2 has several distinct sublineages, among which sublineage BA.1 is responsible for the initial Omicron surge and is now being replaced by BA.2 worldwide, whereas BA.3 is currently at a low frequency. The ongoing BA.1-to-BA.2 replacement underscores the importance to understand the cross-neutralization among the three Omicron sublineages. Here we test the neutralization of BA.1-infected human sera against BA.2, BA.3, and USA/WA1-2020 (a strain isolated in late January 2020). The BA.1-infected sera neutralize BA.1, BA.2, BA.3, and USA/WA1-2020 SARS-CoV-2s with geometric mean titers (GMTs) of 445, 107, 102, and 16, respectively. Thus, the neutralizing GMTs against heterologous BA.2, BA.3, and USA/WA1-2020 are 4.2-, 4.4-, and 28.4-fold lower than the GMT against homologous BA.1, respectively. These findings have implications in COVID-19 vaccine strategy.

S ince the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in late 2019, the virus has evolved to increase viral transmission and immune evasion. The World Health Organization (WHO) has so far designated 5 variants of concern (VOC), including Alpha, Beta, Gamma, Delta, and Omicron. At the time of submitting this paper, the newly emerged Omicron variant had 3 distinct sublineages: BA.1, BA.2, and BA.3. BA.1 was first identified in South Africa in November 2021. BA.1 and its derivative BA.1.1 (containing an extra R346K substitution in the spike of BA.1) caused the initial surges of Omicron around the world. Subsequently, the frequency of BA.2 increased steeply, replacing BA.1 in many parts of the world. In the USA, the frequency of BA.2 increased from 0.4% to 54.9% between 22 January 2022 and 24 March 2022 (https://covid.cdc.gov/covid-data-tracker/#variantproportions). Compared with BA.1, BA.2 did not seem to cause more severe disease 1 , but may increase viral transmissible by~30% 2 . As of 30 March 2022, the frequency BA.3 remained low in the GISAID database (https://www.gisaid.org/). All three sublineages of Omicron could significantly evade vaccine-elicited neutralization, among which BA.3 exhibited the greatest reduction 3,4 . In addition, Omicron BA.1 could efficiently evade non-Omicron SARS-CoV-2 infection-elicited neutralization 5 . The increased transmissibility and immune evasion of the Omicron variant may be responsible for the replacement of VOC from the previous Delta to the current Omicron. Many unvaccinated individuals were infected by BA.1 during the initial Omicron surge 6 . Therefore, in this work, we examine the cross-neutralization of BA.1 infection against BA.2, BA.3, and other variants. Such laboratory information is essential to guide vaccine strategy and public health policy.

Results and discussion
To examine the cross-neutralization among the three Omicron sublineages, we collected 20 human sera from unvaccinated patients who were infected with Omicron BA.1 ( Table 1). The genotype of infecting virus was verified for each patient by Sanger sequencing. The sera were collected on day 8 to 62 after positive RT-PCR test. The serum panel was measured for neutralization against four recombinant SARS-CoV-2s (Fig. 1A): USA/WA1-2020 (wild-type) and three chimeric USA/WA1-2020 bearing the full-length spike protein from Omicron BA.1 (GISAID EPI_ISL_6640916), BA.2 (GISAID EPI_ISL_6795834.2), or BA.3 (GISAID EPI_ISL_7605591). The spike proteins of the three Omicron sublineages have distinct amino acid mutations, deletions, and insertions (Fig. 1A). To facilitate neutralization testing, an mNeonGreen (mNG) reporter was engineered into the four viruses, resulting in wild-type, BA.1-, BA.2-, and BA.3-spike mNG SARS-CoV-2s. The construction and characterization of the four mNG SARS-CoV-2s were recently reported 4 . Using an mNG-based fluorescent focus-reduction neutralization test (FFRNT), we determined the neutralizing geometric mean titers (GMTs) of the sera against wild-type, BA.1-, BA.2-, and BA.3spike mNG SARS-CoV-2s to be 16, 445, 107, and 102, respectively (Fig. 1B). Thus, the neutralizing GMTs against heterologous BA.2-spike, BA.3-spike, and wild-type viruses were 4.2-, 4.4-, and 28.4-fold lower than the GMT against the homologous BA.1spike virus, respectively (Fig. 1B). Consistently, all sera neutralized BA.1-spike virus at neutralizing titers of ≥80, whereas 13 out of 20 sera did not neutralize the wild-type USA/WA1-2020 (defined as 10 for plot and calculation purposes; Fig. 1B and Table 1). Notably, 2 sera neutralized BA.2-spike virus more efficiently than the BA.1-spike virus (indicated by symbol * in Fig. 1C). Collectively, the results support two conclusions. First, BA.1 infection elicited similar levels of cross-neutralization against BA.2 and BA.3, although at a decreased efficiency that was 4.2-to 4.4-fold lower than that against BA.1. This result is in contrast with the neutralization results from vaccinated sera (collected at 1 month after three doses of Pfizer/BioNTech's BNT162b2 vaccine) which neutralized BA.1 and BA.2 much more efficiently than BA.3 4 . Second, the neutralization of BA.1-infected sera against USA/WA1-2020 were 6.7-and 6.4-fold lower than that against Omicron BA.2 and BA.3, respectively. The results indicate the antigenic distinctions among different variant spikes, which must be carefully considered when deciding to switch the vaccine sequence to new variants 7 . If future variants are Omicron decedents, it would be conceptually attractive to switch vaccine sequence to an Omicron spike.
There are two limitations of the serum specimens used in this study. First, the sample size of serum specimen was relatively small. Second, all sera were collected on days 8 to 62 after positive RT-PCR test. The immune status of these specimens were heterogenous, with some sera collected at an acute plasma blast stage and other sera collected at a convalescent IgGdominant phase. Analysis of more specimens collected at later timepoints post-infection will substantiate the current observations.
Emerging evidence supports a vaccine booster strategy to minimize the health risk of the ongoing Omicron infection. First, 2 doses of BNT162b2 vaccine are inefficient to elicit robust neutralization against Omicron variant, whereas 3 doses of BNT162b2 produces robust neutralization against Omicron. Although Omicron-neutralizing activity remains robust for up to 4 months 3 , the durability of such neutralization beyond 4 months after dose 3 remains to be determined. The latter result, together with the real-world vaccine effectiveness, are required to guide the timing of dose 4 vaccine. Second, non-Omicron SARS-CoV-2 infection does not elicit robust neutralization against Omicron variant 5 , suggesting that previously infected individuals should be vaccinated to mitigate the health threat from Omicron. The crossneutralization of BA.1-infected sera against BA.2 and BA.3 suggests the recent BA.1-infected individuals are likely to be protected against the ongoing BA.2 surge. Third, vaccinemediated T cell immunity and non-neutralizing antibodies that mediate antibody-dependent cytotoxicity could also confer protection against severe COVID-19. After vaccination or infection, the majority of T cell epitopes are highly preserved against Omicron spikes 8 . In agreement with this notion, 3 doses of BNT162b2 conferred efficacy against Omicron disease, but the protection wanes over time, with overall efficacy remaining high up to 6 months after dose 3 9-13 . The real-world vaccine effectiveness and laboratory studies will guide vaccine booster strategy to achieve optimal breadth and duration of protection.   Table 1. Each data point represents the geometric mean FFRNT 50 obtained with a serum specimen against the indicated virus. The bar heights and the numbers above indicate geometric mean titers (GMTs). Error bars indicate the 95% confidence interval (CI) of the GMTs. Data are presented as GMT with 95% CI. Statistical analysis was performed using the Wilcoxon matched-pairs signed-rank test. Two-tailed P values of the GMT against BA.1-spike and USA-WA1/2020, BA.2-spike, or BA.3 spikes are all <0.0001. C FFRNT 50 values with connected lines for individual sera. Two sera exhibiting slightly higher FFRNT 50 s against BA.2 virus than that against BA.1-spike SARS-CoV-2 are indicated by symbol asterisk (serum ID 1 and 2 in Table 1).

Methods
Institutional Review Board (IRB number 20-0070). No informed consent was required since these deidentified sera were leftover specimens before being discarded. No diagnoses or treatment was involved either. cell monolayer in 96-well plates. After 1 h infection, the inoculum was aspirated and overlay medium (100 μl supplemented with 0.8% methylcellulose) was added to each well. After incubating the plates at 37°C for 16-18 h, raw images of mNG foci were acquired using Cytation TM 7 (BioTek) with Gene5 software. The foci in each well were counted and normalized to the no-serum-treated controls to calculate infection rates. The FFRNT 50 value was defined as the minimal serum dilution that suppressed >50% of fluorescent foci. The neutralization titer of each serum was determined in duplicates, and the geometric mean was presented. Figures were initially plotted using GraphPad Prism 9.0 software, and assembled in Adobe Illustrator. FFRNT 50 of <20 was treated as 10 for plot purpose and statistical analysis; in this way, we can differentiate between sera with FFRNT 50 of 20 and sera with no neutralization activity at 1:20 dilution (treated as FFRNT 50 of 10). Table 1 summarizes the FFRNT 50 results.
Statistics. The nonparametric Wilcoxon matched-pairs signed-rank test was used to analyze the statistical significance in Fig. 1B.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The raw data that support the findings of this study are shown in the Source data file. Source data are provided with this paper.