SARS-CoV-2 evolution and patient immunological history shape the breadth and potency of antibody-mediated immunity (preprint)

Since the emergence of SARS-CoV-2, humans have been exposed to distinct SARS-CoV-2 antigens, either by infection with different variants, and/or vaccination. Population immunity is thus highly heterogeneous, but the impact of such heterogeneity on the effectiveness and breadth of the antibody-mediated response is unclear. We measured antibody-mediated neutralisation responses against SARS-CoV-2 Wuhan, SARS-CoV-2a, SARS-CoV-2d and SARS-CoV-2o pseudoviruses using sera from patients with distinct immunological histories, including naive, vaccinated, infected with SARS-CoV-2 Wuhan, SARS-CoV-2a or SARS-CoV-2d, and vaccinated/infected individuals. We show that the breadth and potency of the antibody-mediated response is influenced by the number, the variant, and the nature (infection or vaccination) of exposures, and that individuals with mixed immunity acquired by vaccination and natural exposure exhibit the broadest and most potent responses. Our results suggest that the interplay between host immunity and SARS-CoV-2 evolution will shape the antigenicity and subsequent transmission dynamics of SARS-CoV-2, with important implications for future vaccine design.

A Bayesian approach to incorporate structural data into the mapping of genotype to antigenic phenotype of influenza A(H3N2) viruses (preprint)

Surface antigens of pathogens are commonly targeted by vaccine-elicited antibodies but antigenic variability, notably in RNA viruses such as influenza, HIV and SARS-CoV-2, pose challenges for control by vaccination. For example, influenza A(H3N2) entered the human population in 1968 causing a pandemic and has since been monitored, along with other seasonal influenza viruses, for the emergence of antigenic drift variants through intensive global surveillance and laboratory characterisation. Statistical models of the relationship between genetic differences among viruses and their antigenic similarity provide useful information to inform vaccine development though accurate identification of causative mutations is complicated by highly correlated genetic signals that arise due to the evolutionary process. Here, using a sparse hierarchical Bayesian analogue of an experimentally validated model for integrating genetic and antigenic data, we identify the genetic changes in influenza A(H3N2) virus that underpin antigenic drift. We show that incorporating protein structural data into variable selection helps resolve ambiguities arising due to correlated signals, with the proportion of variables representing haemagglutinin positions decisively included, or excluded, increased from 59.8% to 72.4%. The accuracy of variable selection judged by proximity to experimentally determined antigenic sites was improved simultaneously. Structure-guided variable selection thus improves confidence in the identification of genetic explanations of antigenic variation and we also show that prioritising the identification of causative mutations is not detrimental to the predictive capability of the analysis. Indeed, incorporating structural information into variable selection resulted in a model that could more accurately predict antigenic assay titres for phenotypically-uncharactrised virus from genetic sequence. Combined, these analyses have the potential to inform choices of reference viruses, the targeting of laboratory assays, and predictions of the evolutionary success of different genotypes, and can therefore be used to inform vaccine selection processes.

The hyper-transmissible SARS-CoV-2 Omicron variant exhibits significant antigenic change, vaccine escape and a switch in cell entry mechanism (preprint)

Vaccination-based exposure to spike protein derived from early SARS-CoV-2 sequences is the key public health strategy against COVID-19. Successive waves of SARS-CoV-2 infections have been characterised by the evolution of highly mutated variants that are more transmissible and that partially evade the adaptive immune response. Omicron is the fifth of these Variants of Concern (VOCs) and is characterised by a step change in transmission capability, suggesting significant antigenic and biological change. It is characterised by 45 amino acid substitutions, including 30 changes in the spike protein relative to one of the earliest sequences, Wuhan-Hu-1, of which 15 occur in the receptor-binding domain, an area strongly associated with humoral immune evasion. In this study, we demonstrate both markedly decreased neutralisation in serology assays and real-world vaccine effectiveness in recipients of two doses of vaccine, with efficacy partially recovered by a third mRNA booster dose. We also show that immunity from natural infection (without vaccination) is more protective than two doses of vaccine but inferior to three doses. Finally, we demonstrate fundamental changes in the Omicron entry process in vitro, towards TMPRSS2-independent fusion, representing a major shift in the replication properties of SARS-CoV-2. Overall, these findings underlie rapid global transmission and may alter the clinical severity of disease associated with the Omicron variant.

Mutations that adapt SARS-CoV-2 to mustelid hosts do not increase fitness in the human airway. (preprint)

SARS-CoV-2 has a broad mammalian species tropism infecting humans, cats, dogs and farmed mink. Since the start of the 2019 pandemic several reverse zoonotic outbreaks of SARS-CoV-2 have occurred in mink, one of which reinfected humans and caused a cluster of infections in Denmark. Here we investigate the molecular basis of mink and ferret adaptation and demonstrate the spike mutations Y453F, F486L, and N501T all specifically adapt SARS-CoV-2 to use mustelid ACE2. Furthermore, we risk assess these mutations and conclude mink-adapted viruses are unlikely to pose an increased threat to humans, as Y453F attenuates the virus replication in human cells and all 3 mink-adaptations have minimal antigenic impact. Finally, we show that certain SARS-CoV-2 variants emerging from circulation in humans may naturally have a greater propensity to infect mustelid hosts and therefore these species should continue to be surveyed for reverse zoonotic infections.

Reduced neutralisation of the Delta (B.1.617.2) SARS-CoV-2 variant of concern following vaccination (preprint)

Vaccines are proving to be highly effective in controlling hospitalisation and deaths associated with SARS-CoV-2 infection but the emergence of viral variants with novel antigenic profiles threatens to diminish their efficacy. Assessment of the ability of sera from vaccine recipients to neutralise SARS-CoV-2 variants will inform the success of strategies for minimising COVID19 cases and the design of effective antigenic formulations. Here, we examine the sensitivity of variants of concern (VOCs) representative of the B.1.617.1 and B.1.617.2 (first associated with infections in India) and B.1.351 (first associated with infection in South Africa) lineages of SARS-CoV-2 to neutralisation by sera from individuals vaccinated with the BNT162b2 (Pfizer/BioNTech) and ChAdOx1 (Oxford/AstraZeneca) vaccines. Across all vaccinated individuals, the spike glycoproteins from B.1.617.1 and B.1.617.2 conferred reductions in neutralisation of 4.31 and 5.11-fold respectively. The reduction seen with the B.1.617.2 lineage approached that conferred by the glycoprotein from B.1.351 (South African) variant (6.29-fold reduction) that is known to be associated with reduced vaccine efficacy. Neutralising antibody titres elicited by vaccination with two doses of BNT162b2 were significantly higher than those elicited by vaccination with two doses of ChAdOx1. Fold decreases in the magnitude of neutralisation titre following two doses of BNT162b2, conferred reductions in titre of 7.77, 11.30 and 9.56-fold respectively to B.1.617.1, B.1.617.2 and B.1.351 pseudoviruses, the reduction in neutralisation of the delta variant B.1.617.2 surpassing that of B.1.1351. Fold changes in those vaccinated with two doses of ChAdOx1 were 0.69, 4.01 and 1.48 respectively. The accumulation of mutations in these VOCs, and others, demonstrate the quantifiable risk of antigenic drift and subsequent reduction in vaccine efficacy. Accordingly, booster vaccines based on updated variants are likely to be required over time to prevent productive infection. This study also suggests that two dose regimes of vaccine are required for maximal BNT162b2 and ChAdOx1-induced immunity.

Recurrent independent emergence and transmission of SARS-CoV-2 Spike amino acid H69/V70 deletions (preprint)

SARS-CoV-2 Spike amino acid replacements in the receptor binding domain (RBD) occur relatively frequently and some have a consequence for immune recognition. Here we report recurrent emergence and significant onward transmission of a six-nucleotide deletion in the S gene, which results in loss of two amino acids: H69 and V70. Of particular note this deletion, 𝚫H69/V70, often co-occurs with the receptor binding motif amino acid replacements N501Y, N439K and Y453F. One of the 𝚫H69/V70+ N501Y lineages, B.1.1.7, is comprised of over 4000 SARS-CoV-2 genome sequences from the UK and includes eight other S gene mutations: RBD (N501Y and A570D), S1 (𝚫H69/V70 and 𝚫144/145) and S2 (P681H, T716I, S982A and D1118H). Some of these mutations have presumably arisen as a result of the virus evolving from immune selection pressure in infected individuals and at least one, lineage B.1.1.7, potentially from a chronic infection. Given our recent evidence that 𝚫H69/V70 enhances viral infectivity (Kemp et al. 2020), its effect on virus fitness appears to be independent to the RBD changes. Enhanced surveillance for the 𝚫H69/V70 deletion with and without RBD mutations should be considered as a priority. Permissive mutations such as 𝚫H69/V70 have the potential to enhance the ability of SARS-CoV-2 to generate new variants, including vaccine escape variants, that would have otherwise significantly reduced viral fitness.

Decomposing the sources of SARS-CoV-2 fitness variation in the United States (preprint)

The fitness of a pathogen is composite phenotype determined by many different factors influencing growth rates both within and between hosts. Determining what factors shape fitness at the host population-level is especially challenging because both intrinsic factors like pathogen genetics and extrinsic factors such as host behaviour influence between-host transmission potential. These challenges have been highlighted by controversy surrounding the population-level fitness effects of mutations in the SARS-CoV-2 genome and their relative importance when compared against non-genetic factors shaping transmission dynamics. Building upon phylodynamic birth-death models, we develop a new framework to learn how hundreds of genetic and non-genetic factors have shaped the fitness of SARS-CoV-2. We estimate the fitness effects of all amino acid variants and several structural variants that have circulated in the United States between February and September 2020 from viral phylogenies. We also estimate how much fitness variation among pathogen lineages is attributable to genetic versus non-genetic factors such as spatial heterogeneity in transmission rates. Up to September 2020, most fitness variation between lineages can be explained by background spatial heterogeneity in transmission rates across geographic regions. Furthermore, no genetic variant including the Spike D614G mutation has had a significant effect on population-level fitness. Instead, the rapid increase in the frequency of the Spike D614G can be explained by the variant having a spatial transmission advantage due to first establishing in regions with higher transmission rates during the earliest stages of the pandemic.

Recurrent emergence and transmission of a SARS-CoV-2 Spike deletion ΔH69/V70 (preprint)

SARS-CoV-2 amino acid replacements in the receptor binding domain (RBD) occur relatively frequently and some have a consequence for immune recognition. Here we report recurrent emergence and significant onward transmission of a six-nucleotide out of frame deletion in the S gene, which results in loss of two amino acids: H69 and V70. We report that in human infections {Delta}H69/V70 often co-occurs with the receptor binding motif amino acid replacements N501Y, N439K and Y453F, and in the latter two cases has followed the RBD mutation. One of the {Delta}H69/V70+ N501Y lineages, now known as B.1.1.7, has undergone rapid expansion and includes eight S gene mutations: RBD (N501Y and A570D), S1 ({Delta}H69/V70 and {Delta}144) and S2 (P681H, T716I, S982A and D1118H). In vitro, we show that {Delta}H69/V70 does not reduce serum neutralisation across multiple convalescent sera. However, {Delta}H69/V70 increases infectivity and is associated with increased incorporation of cleaved spike into virions. {Delta}H69/V70 is able to compensate for small infectivity defects induced by RBD mutations N501Y, N439K and Y453F. In addition, replacement of H69 and V70 residues in the B.1.1.7 spike reduces its infectivity and spike mediated cell-cell fusion. Based on our data {Delta}H69/V70 likely acts as a permissive mutation that allows acquisition of otherwise deleterious immune escape mutations. Enhanced surveillance for the {Delta}H69/V70 deletion with and without RBD mutations should be considered as a global priority not only as a marker for the B.1.1.7 variant, but potentially also for other emerging variants of concern. Vaccines designed to target the deleted spike protein could mitigate against its emergence as increased selective forces from immunity and vaccines increase globally. HighlightsO_LI{Delta}H69/V70 is present in at least 28 SARS-CoV-2 lineages C_LIO_LI{Delta}H69/V70 does not confer escape from convalescent sera C_LIO_LI{Delta}H69/V70 increases spike infectivity and compensates for RBD mutations C_LIO_LI{Delta}H69/V70 is associated with greater spike cleavage C_LIO_LIB.1.1.7 requires {Delta}H69/V70 for optimal spike cleavage and infectivity C_LI

The Potential for SARS-CoV-2 to Evade Both Natural and Vaccine-induced Immunity (preprint)

SARS-CoV-2 attaches to the surface of susceptible cells through extensive interactions between the receptor binding domain (RBD) of its spike protein and angiotensin converting enzyme type 2 (ACE2) anchored in cell membranes. To investigate whether naturally occurring mutations in the spike protein are able to prevent antibody binding, yet while maintaining the ability to bind ACE2 and viral infectivity, mutations in the spike protein identified in cases of human infection were mapped to the crystallographically-determined interfaces between the spike protein and ACE2 (PDB entry 6M0J), antibody CC12.1 (PDB entry 6XC2), and antibody P2B-2F6 (PDB entry 7BWJ). Both antibody binding interfaces partially overlap with the ACE2 binding interface. Among 16 mutations that map to the RBD:CC12.1 interface, 11 are likely to disrupt CC12.1 binding but not ACE2 binding. Among 12 mutations that map to the RBD:P2B-2F6 interface, 8 are likely to disrupt P2B-2F6 binding but not ACE2 binding. As expected, none of the mutations observed to date appear likely to disrupt the RBD:ACE2 interface. We conclude that SARS-CoV-2 with mutated forms of the spike protein may retain the ability to bind ACE2 while evading recognition by antibodies that arise in response to the original wild-type form of the spike protein. It seems likely that immune evasion will be possible regardless of whether the spike protein was encountered in the form of infectious virus, or as the immunogen in a vaccine. Therefore, it also seems likely that reinfection with a variant strain of SARS-CoV-2 may occur among people who recover from Covid-19, and that vaccines with the ability to generate antibodies against multiple variant forms of the spike protein will be necessary to protect against variant forms of SARS-CoV-2 that are already circulating in the human population.

Planning as Inference in Epidemiological Models (preprint)

In this work we demonstrate how existing software tools can be used to automate parts of infectious disease-control policy-making via performing inference in existing epidemiological dynamics models. The kind of inference tasks undertaken include computing, for planning purposes, the posterior distribution over putatively controllable, via direct policy-making choices, simulation model parameters that give rise to acceptable disease progression outcomes. Neither the full capabilities of such inference automation software tools nor their utility for planning is widely disseminated at the current time. Timely gains in understanding about these tools and how they can be used may lead to more fine-grained and less economically damaging policy prescriptions, particularly during the current COVID-19 pandemic.