Virological characteristics of the SARS-CoV-2 KP.3, LB.1 and KP.2.3 variants

Abstract

The SARS-CoV-2 JN.1 variant, arising from BA.2.86.1 with a substitution in the spike (S) protein, S:L455S, exhibited increased fitness and outcompeted the previously predominant XBB lineages by the beginning of 2024. Subsequently, JN.1 subvariants including KP.2 and KP.3, which convergently acquired S protein substitutions such as S:R346T, S:F456L, and S:Q493E, have emerged concurrently. Furthermore, JN.1 subvariants such as LB.1 and KP.2.3, which convergently acquired S:S31del in addition to the above substitutions, have emerged and spread as of June 2024. Here we investigated the virological properties of KP.3, LB.1 and KP.2.3. We estimated the relative effective reproduction number (Re) of KP.3, LB.1, and KP.2.3 using a Bayesian multinomial logistic model based on the genome surveillance data from Canada, the UK, and the USA, where these variants have spread from March to April 2024. The Re of KP.3 is more than 1.2-fold higher than that of JN.1 and higher than or comparable to that of KP.2 in these countries. Importantly, the Re values of LB.1 and KP.2.3 are even higher than those of KP.2 and KP.3. These results suggest that the three variants we investigated herein, particularly LB.1, and KP.2.3, will become major circulating variants worldwide in addition to KP.2 and KP.3. The pseudovirus infectivity of KP.2 and KP.3 was significantly lower than that of JN.1. On the other hand, the pseudovirus infectivity of LB.1 and KP.2.3 was comparable to that of JN.1. Neutralization assay was conducted by using four types of breakthrough infection (BTI) sera with XBB.1.5, EG.5, HK.3 and JN.1 infections as well as monovalent XBB.1.5 vaccine sera. In all four groups of BTI sera tested, the 50% neutralization titers (NT50) against LB.1 and KP.2.3 were significantly lower than those against JN.1 (2.2-3.3-fold and 2.0-2.9-fold) and even lower than those against KP.2 (1.6-1.9-fold and 1.4-1.7 fold). Although KP.3 exhibited neutralization resistance against all BTI sera tested than JN.1 (1.6-2.2-fold) with statistical significance, there were no significant differences between KP.3 and KP.2. In the case of infection-naive XBB.1.5 vaccine sera, the NT50 values of JN.1 subvariants were very low. In the case of XBB.1.5 vaccine sera after natural XBB infection, the NT50 values against KP.3, LB.1 and KP.2.3 were significantly lower than those of JN.1 (2.1-2.8-fold) and even lower than KP.2 after infection (1.4-2.0-fold). Overall, our results suggest that the S substitutions convergently acquired in the JN.1 subvariants contribute to immune evasion, and therefore, increase their Re when compared to parental JN.1. More importantly, LB.1 and KP.2.3 exhibited higher pseudovirus infectivity and more robust immune resistance than KP.2. These data suggest that S:S31del is critical to exhibit increased infectivity, increased immune evasion, and therefore, potentially contributes to increased Re.