CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

311

Evolving of Immune Evasion During a SARS-CoV-2 Chronic Infection: A Pathway to Variants of Concern Suzanne Pickering 1 , Luke Blagdon Snell 1 , Xiaohao Xu 1 , Helena Winstone 1 , Jeffrey Seow 1 , Carl Graham 1 , Rahul Batra 2 , Michael H. Malim 1 , Katie Doores 1 , Jonathan Edgeworth 2 , Gaia Nebbia 2 , Stuart Neil 1 , Rui Galao 1 1 King's College London, London, UK, 2 Guy's and St Thomas' NHS Foundation Trust, London, UK Background: Acute SARS-CoV-2 infections are mostly cleared within a few days but can become chronic in immunocompromised individuals. Impaired immune clearance is also thought to be a major factor in the evolution of new SARS-CoV-2 variants of concern(VOC). However, it is unclear how intra-host variation accumulates over the course of very long-term chronic SARS-CoV-2 infection, and how this reflects evolving mechanisms of adaptive and innate immune evasion. Methods: A chronic infection (505 days) with SARS-CoV-2 B.1 virus of a 23yo male with an unknown immunodeficiency was followed in this study. Longitudinal samples of nasopharyngeal swabs, and corresponding full-length spike haplotypes sequencing, were obtained together with contemporaneous autologous sera. A live virus was isolated for day 329 of infection. Neutralisation assays were performed using FL-virus and lentiviral particles pseudotyped with spike proteins representative of viral quasispecies haplotypes, to assess evolution of immune evasion from autologous and heterologous sera, and a panel of mAb. Calu3 cells were infected with isolated FL-virus or VOCs for comparative studies on innate immune evasion. Results: We observed that original B.1 virus carrying a D614G mutation in spike gradually accumulated 55 further spike mutations during the course of infection, mostly in key antigenic regions. Interestingly, 23 of those are defining of different VOCs such as Alpha and Omicron. Although this patient was not vaccinated or treated with antibody-based therapies, we demonstrated that accumulating mutations led to a virus universally resistant to wave 1 neutralising responses. We also observed that the virus isolate induced significantly lower proinflammatory gene expression in Calu3 cells than B.1 D614G virus. This correlated with a non-spike adaptation that appeared early in the infection, leading to higher expression of the viral innate antagonist protein orf9b. This same mutation has arisen independently in major VOC such as Alpha and Delta. Conclusions: We characterised a pathway by which a wave 1 SARS-CoV-2 virus gradually evolved mechanisms of evasion from both adaptive and innate immune responses redolent of major VOC during a single chronic infection. This provides one of the strongest pieces of evidence for the involvement of long term infections in the evolution of new variants that resist antibody-mediated immunity. We also suggest adaptations that resist innate immunity occur early in a chronic infection to promote viral persistence. Karen Zafilaza 1 , Antoine Fauchois 1 , Valentin Leducq 1 , Quentin Le Hingrat 2 , Stephane Marot 1 , Theophile Cocherie 3 , Romane Guilbaud 2 , Valentine Marie Ferré 1 , Charlotte Charpentier 2 , Diane Descamps 2 , Eve Todesco 1 , Anne-Geneviève Marcelin 1 , Vincent Calvez 1 , Cathia Soulié 4 1 Assistance Publique – Hôpitaux de Paris, Paris, France, 2 Hôpital Bichat-Claude-Bernard, Paris, France, 3 Pitie Salpetriere Hospital, Paris, France, 4 Institut national de la santé et de la recherche médicale (Inserm), Paris, France Background: Although documented, the mutational dynamics of SARS-CoV-2 in immunocompromised patients remain relatively understudied. This study examined a cohort of immunocompromised patients hospitalized with SARS CoV-2 infection to assess the distinct evolutionary trajectories of SARS-CoV-2 in these individuals compared to their immunocompetent counterparts. We compared the genetic diversity of SARS-CoV-2 between these groups over time. Methods: 87 immunocompromised hospitalized patients and 52 non immunocompromised subjects from Pitié-Salpêtrière Hospital (Paris, France) diagnosed with SARS-CoV-2 infection by RT-PCR were enrolled. Samples were collected at the time of infection (D0) and during their follow-up. Whole genome sequencing was performed using Oxford Nanopore technology (ONT). Consensus sequences were generated using a pipeline including minimap2 and DeepVariant. We performed a sparse principal component analysis (sPCA) and a t-student stochastic neighbor embedding (t-SNE) to identify trends in longitudinal data. We used univariate analyses to catch differentially detected single mutations. We built a global and specific models on isolated mutation to measure the association with follow-up time and immunosuppression. Genetic Mutation Differences Patterns of SARS-CoV-2 in Immunocompromised Patients Versus Controls

profiles of single/co-infected cells were used to identify specific cell markers and make functional inferences about the different cell states. Results: ScRNA-seq data identified four distinct cell clusters: 1) uninfected, 2) SARS-CoV-2-infected, 3) HRSV-infected, and 4) SARS-CoV-2/HRSV co-infected cells. Within Cluster 4, two sub-clusters were further distinguished based on HRSV gene expression levels: 4a (low) and 4b (high). Marker analysis revealed molecular signatures of both protein-coding and long non-coding (lnc) RNA characterizing each of the four cell clusters. Specifically, co-infected cells showed altered expression of genes linked to viral defense (n=27; P=6.2×10 - ⁸), innate immunity (n=30; P=2.5×10 - ⁵) and type I IFN regulation (n=8; P=3.4×10 - ⁴). When analyzing sub-clusters, JAK-STAT (n=8; P=5.7×10 -5 ) and ERK1/ERK2/MAPK (n=15; P=9.3×10 -5 ) signaling pathways were specifically up regulated in sub-cluster 4a and 4b, respectively. Co-infected cells also showed higher expression of at least 16 lncRNAs, some not yet functionally annotated, while others ( e.g. AC011287.1) may be involved in trained immunity. As for the sub-clusters, lncRNAs NEAT1 and MALAT1 were downregulated in 4a compared to 4b (P=1.1x10 -9 and P=2.3 x 10 -16 , respectively). Conclusions: ScRNA-seq revealed transcriptionally distinct subpopulations of lung epithelial co-infected cells. Two sub-clusters of coinfected cells, characterized by different HRSV expression levels, are presumably associated with a competitive/cooperative interaction between HRSV and SARS-CoV-2 within the same cell. Further analyses will be mandatory to ascertain the functional role exerted by lncRNAs that may have therapeutic relevance in the context of HRSV/SARS-CoV-2 coinfection. Emergence of Divergent Spike Haplotypes in Individuals With Persistent SARS-CoV-2 Infections Luke Blagdon Snell 1 , Suzanne Pickering 1 , Adela Alcolea-Medina 1 , Rahul Batra 2 , Michael H. Malim 1 , Gaia Nebbia 2 , Jonathan Edgeworth 2 , Stuart Neil 1 , Rui Galao 1 1 King's College London, London, UK, 2 Guy's and St Thomas' NHS Foundation Trust, London, UK Background: Persistent infection with SARS-CoV-2 is hypothesised to give rise to new variants of concern (VoCs). However, sequencing methods to investigate intrahost variation are limited: short-read technology precludes haplotyping, and long-read technology has high error rates. We describe a new method for investigating SARS-CoV-2 quasispecies and apply it to identify divergent spike haplotypes arising during persistent infection. Methods: A long-read nanopore sequencing workflow was developed to generate high-quality reads (99.9% accuracy) covering the entire spike gene sequence, along with a bioinformatic tool (HaploVar) to identify and quantify unique spike haplotypes at ≥1% frequency in samples. Respiratory samples were processed from patients with acute and persistent infection lasting longer than 4 weeks. Molecular-clock and divergence phylogenies were constructed using TreeTime v0.11.3 and IQTree v2.3.0, then displayed in ete v3.1.3. Haplotypes were screened for recombination using GARD v0.2. Results: Sixteen patients with persistent infection were identified, ranging from 31-505 days and with a median of 7 longitudinal samples per patient. Infecting variants spanned the pandemic from B.1 until JN.1. Eighty-one samples from persistent infection and 75 samples from acute infection were successfully processed. Number of spike haplotypes identified in each sample was related to days post onset of symptoms (p<0.001) and being a persistent infection (p<0.001), but not viral load (Ct value) nor sequencing data yielded. During acute infection the median number of spike haplotypes was 1 (IQR 1-2). In persistent infection after 28 days the median number of spike haplotypes was 3 (IQR 2-7), compared to 7 (IQR 2-11) after 100 days. Many mutations were identical or at the same position to those seen in subsequent VoCs, including L18K and E484Q (Beta); K417T (Gamma); T19S, L452R (Delta); A570T (Alpha); H655Y (Omicron); G446S, Q954K (BA.1); S477N, D796H (BA.2); N460K (BA.2.75); Q493E (KP.3); and Y453F (Mink). Individuals treated with sotrovimab showed evidence of quasispecies with resistance mutations. Putative intra-spike recombination between haplotypes was identified in 3 persistent infections. Conclusions: This novel workflow identifies intrahost spike haplotypes, and provides strong evidence for the emergence of variants of concern from persistent infection. We provide the first evidence of intrahost recombination from within a single individual as a mechanism of generating diversity of SARS-CoV-2. The figure, table, or graphic for this abstract has been removed.

Poster Abstracts

310

312

The figure, table, or graphic for this abstract has been removed.

63

CROI 2025