By March 2022, ≈68% of the population of Colombia had been vaccinated, predominantly with spike-based mRNA (BNT162b2), vectored (AZD1222; AstraZeneca, https://www.astrazeneca.com), and chemically inactivated whole virus–based vaccines (CoronaVac) (Appendix Figure 1). To investigate the potency of natural and vaccine-derived immunity, we tested and compared the neutralization activity in 49 serum samples from vaccinated and naturally infected persons in Colombia. Among vaccinated persons, we tested serum from 32 persons sampled in October 2021. Of those, 10 vaccinated with BioNTech-Pfizer were tested a median 99.5 d (range 65–170) after completing vaccination, 7 vaccinated with AstraZeneca were tested a median 146.0 d (range 129–173) after completing vaccination, and 15 vaccinated with CoronaVac were tested a median 46.0 d (range 28–131) after completing vaccination. We tested serum samples from 17 persons who tested positive for SARS-CoV-2 antibodies (MAGLUMI 2019-nCoV IgG; Snibe Diagnostic, https://www.snibe.com) (Table 1; Appendix Table 1) during a seroprevalence study conducted in November 2020. To control whether persons vaccinated with spike-based vaccines were not previously infected, serum samples were tested for IgG against the SARS-CoV-2 nucleocapsid protein by ELISA (SARS-CoV-2 NCP kit; Euroimmun, https://www.euroimmun.com) (Table 2). We used 50% plaque reduction neutralization tests to obtain neutralizing titers against an early isolate and the Alpha, Beta, Delta, Gamma, Omicron BA.1, and Mu variants (Appendix).

Figure 2

Figure 2. Comparative neutralization of the Mu SARS-CoV-2 variant in Colombia. A–C) Neutralization of SARS-CoV-2 variants from serum samples from persons fully immunized with BNT162b2 (Pfizer-BioNTech, https://www.pfizer.com) (A), AZD1222 (AstraZeneca, ...

Neutralizing antibody titers against Mu were significantly lower than those against the parental isolate (p<0.0001 by Wilcoxon matched-pairs signed-rank test) in all serum samples tested in this study, irrespective of whether immune responses were elicited by vaccination or by natural infection. Vaccine-derived antibodies neutralized Mu on average 8.1-fold (p<0.0001 by Wilcoxon test) less than the parental strain resembling the vaccine backbones (Figure 2, panels A–C; Appendix Figure 2). We found a similar 8.0-fold reduced neutralization of Mu (p<0.0001 by Wilcoxon test) for the group of naturally infected persons (Figure 2, panel D). Despite the relatively lower neutralization potency observed in serum samples from persons immunized with the inactivated full virus-based vaccine Sinovac, observed differences in the ability to neutralize Mu compared with the parental strain among the 3 vaccine groups were not statistically significant (range 7.7–11.4-fold; p = 0.8298 by Kruskal-Wallis test) (Figure 2).

Compared with other variants, neutralizing antibody titers from serum samples of both naturally infected persons and vaccinees were lower against Mu than against all VOCs except for Omicron (Figure 2, panels A and B). Therefore, our results provide strong evidence for immune evasion of the Mu VOI on the basis of results from robust neutralization testing using full viral isolates. Neutralization of Mu by vaccine-induced antibodies was significantly lower than for Beta (p = 0.0083 by Wilcoxon text), for which immune evasion properties led to the suspension of AstraZeneca usage in South Africa (6), and Gamma, which resulted in breakthrough infections in Latin America (7). Immune evasion of Mu is consistent with shared mutations in spike protein residues associated with immune evasion in Beta and Gamma, such as E484K (8). In addition, the mutation leading to the amino acid exchange R346K in Mu is known to be involved in the evasion of monoclonal antibody–mediated neutralization (9), and genomic exchanges occurring at 3 adjacent sites (Y144T, Y145S, and insertion of the amino acid asparagine [N] between spike residues 145 and 146) have been associated with the immune escape properties of Mu (10,11).

Antigenic cartography was recently employed to map the antigenic relationship between the SARS-CoV-2 Omicron and Delta VOCs and other previously circulating VOCs and VOIs (S.H. Wilks et al., unpub. data, https://www.biorxiv.org/content/10.1101/2022.01.28.477987v1). Among the serum samples from Colombia vaccinees, there was a high antigenic distance between Mu and most variants from other serum samples, which clustered together with the parental strain and Alpha (Appendix Figure 3). Of note, antibody responses in naturally infected persons supported past infection with strains bearing similarities to early SARS-CoV-2 isolates and the Gamma variant (Figure 2, panel D). Antibody reactivity in naturally infected persons was thus in concordance with the circulation of SARS-CoV-2 variants in South America during the time of sampling in late 2020 (12), supporting the robustness of our data.

Our study was limited by different time points for sampling of vaccinees and the lack of information on natural infections altering immune responses in vaccinees. However, lack of detectable N-protein antibody responses and the absence of clinical records suggestive of COVID-19 infection in vaccinees immunized with spike-based vaccines supports the robustness of our data despite the vaccinees’ unclear infection histories.

Our data highlight the importance of continuous monitoring for the emergence of new SARS-CoV-2 variants and strains and the timely identification of those variants with potential to evade naturally elicited and vaccine-derived immune responses, using local sampling specimens in the context of regional epidemiologic conditions. Moreover, our data confirmed the potential of Mu to partially evade immune responses, which may affect the efficacy of vaccination programs in southern America and other areas (7,13). Further studies are warranted to evaluate the pathogenicity of and cell-mediated immunity against Mu and the ability of immune responses associated with Mu to neutralize other SARS-CoV-2 variants. However, because vaccination boosters still provide some degree of protection against severe disease from Omicron (3,14), which shows more immunity evasion than Mu, vaccination will likely still provide protection against severe disease from Mu.