Nirmatrelvir-resistant SARS-CoV-2 variants with high fitness in an infectious cell culture system

Abstract

The oral protease inhibitor nirmatrelvir is of key importance for prevention of severe coronavirus disease 2019 (COVID-19). To facilitate resistance monitoring, we studied severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) escape from nirmatrelvir in cell culture. Resistant variants harbored combinations of substitutions in the SARS-CoV-2 main protease (Mpro). Reverse genetics revealed that E166V and L50F + E166V conferred high resistance in infectious culture, replicon, and Mpro systems. While L50F, E166V, and L50F + E166V decreased replication and Mpro activity, L50F and L50F + E166V variants had high fitness in the infectious system. Naturally occurring L50F compensated for fitness cost of E166V and promoted viral escape. Molecular dynamics simulations revealed that E166V and L50F + E166V weakened nirmatrelvir-Mpro binding. Polymerase inhibitor remdesivir and monoclonal antibody bebtelovimab retained activity against nirmatrelvir-resistant variants, and combination with nirmatrelvir enhanced treatment efficacy compared to individual compounds. These findings have implications for monitoring and ensuring treatments with efficacy against SARS-CoV-2 and emerging sarbecoviruses.

Nirmatrelvir Resistance in SARS-CoV-2 Omicron_BA.1 and WA1 Replicons and Escape Strategies

“…We hypothesize that the apparent strain-specific differences in replicon fitness of E166V NIR-resistant replicons may provide insight into the barrier to resistance among the studied strains. The poor fitness of E166VWA1 combined with the need for secondary mutations (L50F or T21I [26][27][28]) for sufficient viral growth is consistent with a relatively higher barrier to resistance compared to E166BA.1, which appears to have a significantly lower replicon fitness cost while maintaining NIR resistance. Consistent with the lower fitness of E166V-carrying viruses, this mutation has been reported only 10 times in 13,313,267 sequences submitted at the GISAID database (~0.000075%).…”

“…Since N142L and E166V did not eliminate nsp5 activity, we selected these for further studies using Omicron and WA1 replicon systems. For comparison purposes we later expanded the ongoing analysis of replicon mutations to include L50F, E166A, and L167F that were listed in preprint studies that appeared at the later stages of our study [26][27].…”

“…All these H-bonds are lost by the E166V mutation. Indeed, recent MD studies suggested that V166 would destabilize nsp5 dimerization through disruption of dimer interactions S1-F140 and R4-E290 [27]. Also, biochemical experiments in a preprint sugest interference of A166 on nsp5 dimerization [26].…”

“…However, resistance could emerge if mutations at or near the active site of nsp5 impair NIR binding while still permitting cleavage of the natural substrates. Indeed, virological passaging data with NIR have already led to the investigation of various NIR-resistant mutations [26][27][28][29] [59]. In this report, we test several possible mutations derived either from analysis of the structures of nsp5 bound to NIR or natural substrates [30,31],…”

“…or from reported data derived from passage of SARS-CoV-2 in the presence of NIR [26,27] [28]. We describe the effects of these mutations on fitness in replicon systems of Omicron and WA1 SARS-CoV-2 strains and test for drug resistance to NIR, GC376, and PF-00835231 [19,32].…”