SARS-CoV-2 evolves increased infection elicited cell death and fusion in an immunosuppressed individual

The milder clinical manifestations of Omicron infection relative to pre-Omicron SARS-CoV-2 raises the possibility that extensive evolution results in reduced pathogenicity. To test this hypothesis, we quantified induction of cell fusion and cell death in SARS-CoV-2 evolved from ancestral virus during long-term infection. Both cell fusion and death were reduced in Omicron BA.1 infection relative to ancestral virus. Evolved virus was isolated at different times during a 6-month infection in an immunosuppressed individual with advanced HIV disease. The virus isolated 16 days post-reported symptom onset induced fusogenicity and cell death at levels similar to BA.1. However, fusogenicity was increased in virus isolated at 6 months post-symptoms to levels intermediate between BA.1 and ancestral SARS-CoV-2. Similarly, infected cell death showed a graded increase from earlier to later isolates. These results may indicate that, at least by the cellular measures used here, evolution in long-term infection does not necessarily attenuate the virus.

SARS-CoV-2 Omicron spike mediated immune escape, infectivity and cell-cell fusion

The SARS-CoV-2 Omicron BA.1 variant emerged in late 2021 and is characterised by multiple spike mutations across all spike domains. Here we show that Omicron BA.1 has higher affinity for ACE2 compared to Delta, and confers very significant evasion of therapeutic monoclonal and vaccine-elicited polyclonal neutralising antibodies after two doses. mRNA vaccination as a third vaccine dose rescues and broadens neutralisation. Importantly, antiviral drugs remdesevir and molnupiravir retain efficacy against Omicron BA.1. We found that in human nasal epithelial 3D cultures replication was similar for both Omicron and Delta. However, in lower airway organoids, Calu-3 lung cells and gut adenocarcinoma cell lines live Omicron virus demonstrated significantly lower replication in comparison to Delta. We noted that despite presence of mutations predicted to favour spike S1/S2 cleavage, the spike protein is less efficiently cleaved in live Omicron virions compared to Delta virions. We mapped the replication differences between the variants to entry efficiency using spike pseudotyped virus (PV) entry assays. The defect for Omicron PV in specific cell types correlated with higher cellular RNA expression of TMPRSS2, and accordingly knock down of TMPRSS2 impacted Delta entry to a greater extent as compared to Omicron. Furthermore, drug inhibitors targeting specific entry pathways demonstrated that the Omicron spike inefficiently utilises the cellular protease TMPRSS2 that mediates cell entry via plasma membrane fusion. Instead, we demonstrate that Omicron spike has greater dependency on cell entry via the endocytic pathway requiring the activity of endosomal cathepsins to cleave spike. Consistent with suboptimal S1/S2 cleavage and inability to utilise TMPRSS2, syncytium formation by the Omicron spike was dramatically impaired compared to the Delta spike. Overall, Omicron appears to have gained significant evasion from neutralising antibodies whilst maintaining sensitivity to antiviral drugs targeting the polymerase. Omicron has shifted cellular tropism away from TMPRSS2 expressing cells that are enriched in cells found in the lower respiratory and GI tracts, with implications for altered pathogenesis.

SARS-CoV-2 Omicron has extensive but incomplete escape of Pfizer BNT162b2 elicited neutralization and requires ACE2 for infection

The emergence of SARS-CoV-2 Omicron, first identified in Botswana and South Africa, may compromise vaccine effectiveness and the ability of antibodies triggered by previous infection to protect against re-infection (1). Here we investigated whether Omicron escapes antibody neutralization in South Africans, either previously SARS-CoV-2 infected or uninfected, who were vaccinated with Pfizer BNT162b2. We also investigated if Omicron requires the ACE2 receptor to infect cells. We isolated and sequence confirmed live Omicron virus from an infected person in South Africa and compared plasma neutralization of this virus relative to an ancestral SARS-CoV-2 strain with the D614G mutation, observing that Omicron still required ACE2 to infect. For neutralization, blood samples were taken soon after vaccination, so that vaccine elicited neutralization was close to peak. Neutralization capacity of the D614G virus was much higher in infected and vaccinated versus vaccinated only participants but both groups had 22-fold Omicron escape from vaccine elicited neutralization. Previously infected and vaccinated individuals had residual neutralization predicted to confer 73% protection from symptomatic Omicron infection, while those without previous infection were predicted to retain only about 35%. Both groups were predicted to have substantial protection from severe disease. These data support the notion that high neutralization capacity elicited by a combination of infection and vaccination, and possibly boosting, could maintain reasonable effectiveness against Omicron. A waning neutralization response is likely to decrease vaccine effectiveness below these estimates. However, since protection from severe disease requires lower neutralization levels and involves T cell immunity, such protection may be maintained.

SARS-CoV-2 cell-to-cell spread occurs rapidly and is insensitive to antibody neutralization

Viruses increase the efficiency of close-range transmission between cells by manipulating cellular physiology and behavior, and SARS-CoV-2 uses cell fusion as one mechanism for cell-to-cell spread. Here we visualized infection using time-lapse microscopy of a human lung cell line and used live virus neutralization to determine the sensitivity of SARS-CoV-2 cell-to-cell spread to neutralizing antibodies. SARS-CoV-2 infection rapidly led to cell fusion, forming multinucleated cells with clustered nuclei which started to be detected at 6h post-infection. To compare sensitivity of cell-to-cell spread to neutralization, we infected either with cell-free virus or with single infected cells expressing on their surface the SARS-CoV-2 spike protein. We tested two variants of SARS-CoV-2: B.1.117 containing only the D614G substitution, and the escape variant B.1.351. We used the much smaller area of single infected cells relative to infection foci to exclude any input infected cells which did not lead to transmission. The monoclonal antibody and convalescent plasma we tested neutralized cell-free SARS-CoV-2, with the exception of B.1.351 virus, which was poorly neutralized with plasma from non-B.1.351 infections. In contrast, cell-to-cell spread of SARS-CoV-2 showed no sensitivity to monoclonal antibody or convalescent plasma neutralization. These observations suggest that, once cells are infected, SARS-CoV-2 may be more difficult to neutralize in cell types and anatomical compartments permissive for cell-to-cell spread.

Escape of SARS-CoV-2 501Y.V2 variants from neutralization by convalescent plasma

SARS-CoV-2 variants of concern (VOC) have arisen independently at multiple locations and may reduce efficacy of current vaccines targeted at the spike glycoprotein. We re-cently described the emergence of VOC in South Africa (501Y.V2 or PANGO lineage B.1.351) with mutations in the spike receptor-binding domain (RBD) and N-terminal domain (NTD). Here, using a live virus neutralization assay (LVNA), we compared neutralization of a first wave virus (B.1.1.117) versus the 501Y.V2 variant using plasma collected from adults hospitalized with COVID-19 from two South African infection waves, with the second wave dominated by 501Y.V2 infections. Sequencing demonstrated that infections in first wave plasma donors were with viruses harbouring none of the 501Y.V2-defining RBD or NTD mutations, except for one with E484K. 501Y.V2 virus was effectively neutralized by plasma from second wave infections and first wave virus was effectively neutralized by first wave plasma. In cross-neutralization, 501Y.V2 virus was poorly neutralized by first wave plasma, with an 8.4-fold drop in neutralization relative to first wave virus and a 15.1-fold drop relative to 501Y.V2 neutralization by second wave plasma. In contrast, second wave plasma neutralization of first wave virus was more effective, showing 4.1-fold decline relative to 501Y.V2 virus neutralization and 2.3-fold decline relative to first wave plasma neutralization. While we only tested one plasma elicited by E484K alone, this potently neutralized both variants. The observed effective neutralization of first wave virus by 501Y.V2 infection elicited plasma provides preliminary evidence that vaccines based on VOC sequences could retain activity against other circulating SARS-CoV-2 lineages.