Determining factors of transmission and evolution of SARS-CoV-2 in populations at risk

National Institute Of Allergy And Infectious Diseases

Investigators

Linked publications & trials

Abstract

(a) Molecular epidemiology Whole genome sequencing of SARS-CoV-2 has become crucial for epidemiology studies and to determine how the virus sweeps through various populations, with the transmission potential of new emerging variants. In FY24, we had finalized genomic epidemiology studies in collaboration with clinical and public health teams in the Dominican Republic and Haiti. We combined phylogenetics with travel history data and used the viral diversity information to determine the number of independent introductions into each of these countries that shutdown their borders and airports early in the pandemic. From a public health perspective, this enables an evaluation of how efficient border closings are limiting the introduction of new variants. The Caribbean is a populous region (>40 million people in the island states) at the crossroads of many trade routes and visited by tourists from around the world yet is drastically under-sampled relative to Europe and North America. Under-surveilled regions pose a significant threat to global health security as novel viral variants can emerge and spread undetected. In FY25, we published the Haiti study where we identified the emergence of a SARS-CoV-2 lineage (B.1.478) in Haiti and documented its multiple introductions into the United States [Mushegian et al., 2024, PLOS Global Public Health]. We also finalized the phylogenetic analysis of SARS-CoV-2 from samples collected in the Dominican Republic in 2020 to study how travel and tourism can drive the spread of viruses. We showed patterns of repeated introductions and transmission chains between the Caribbean and major geographic regions early in the COVID-19 pandemic [Krietman et al., 2025, Submitted to Microbiology Spectrum]. (b) Virus evolution and genetic diversity within infected hosts While identification of new clades and lineages, and the associated viral consensus changes help in tracking spread of the virus, fewer studies have been done on the minority variants present in infected individuals. These minority variants could be seeding the emergence of new clades, thus identifying them early is of relevance for preparedness and to track transmission events. In FY25, we finalized 3 studies that focus on intra-host diversity. The first is a collaboration with Dr. James Musser (Houston Methodist), where we used deep sequence data of thousands of SARS-CoV-2 clinical samples to explore within-host diversity of the virus from a high-throughput viral surveillance program associated with a large hospital system. This project required a complete revision of how we tackle opportunistic datasets for deep sequencing analyses and the identification of all the artifacts that can be introduced. In FY25, we established a method for estimating viral loads using sequencing data outputs that would allow for the removal of low-quality samples, technical artifacts, and increase the accuracy when studying intra- and inter-host evolution and transmission of viruses. This novel method was developed to use raw sequencing metrics, particularly sequencing coverage unevenness, to train a simple and fast random forest regressor for predicting viral load from raw amplicon-based sequences. We also used precomputed variant calling files uploaded to NCBI to train the model. Accurate viral load data allows for increased confidence in the identification and abundance of intra-host variant analyses and can be used to establish models of selection and viral fitness. Viral load is typically measured using cycle threshold (Ct) values that are rarely reported in metadata provided on public repositories and databases. Using this method, we have achieved up to 92% accuracy when testing on SARS-CoV-2 datasets. The model is virus agnostic and can also trained be trained on influenza virus and Mpox datasets. The tool will be publicly available on our GitHub site and works in seconds when using the NCBI ACTIVE TRACE and the cloud infrastructure. We have a manuscript in preparation. For our second study, In FY25, we published a manuscript in Nature Communications in collaboration with Mirella Salvatore (Weill Cornell) Diego Diel (Cornell U). In this study we analyzed the within-host evolution of SARS-CoV-2 in immunocompromised patients with mostly B cell defects. We identified patients who had infections with SARS-CoV-2 virus carrying mutations leading to resistance against multiple antivirals. Sequence analysis showed that 9 of 15 patients analyzed carried viruses with mutations in the nsp12 (RNA dependent RNA polymerase), while four had viruses with nsp5 (3C protease) mutations. Infectious SARS-CoV-2 with a double mutation in nsp5 (T169I) and nsp12 (V792I) was recovered from respiratory secretions 77 days after initial COVID-19 diagnosis from a patient treated with remdesivir and nirmatrelvir-ritonavir. The Diel lab isolated the virus and confirmed in vitro decreased sensitivity to remdesivir and nirmatrelvir, which was overcome by combined antiviral treatment. Studies in golden Syrian hamsters in the Diel Lab demonstrated efficient transmission to contact animals. In a unique collaborative study with Dr. Daniel Chertow (VRC/NIAID), we determined the diversity of SARS-CoV-2 across tissues from an autopsy case from a patient with an inborn genetic disease that leads to immunocompromise. We discovered extensive spike protein adaptations that varied depending on the tissue site from which the genomes were isolated. These adaptations are predicted to have significant functional impacts on spike-ACE2 interactions. By linking viral population dynamics across the different tissue sites, we defined the level of tissue compartmentalization during the multi-organ spread of SARS-CoV-2. In FY25, we finalized our analyses and submitted a manuscript that is currently under review in the Journal of Virology. (c) Respiratory tract microbiome profiles and host response In our first study of the airway microbiome of COVID19 patients, we focused on the metagenomic and metatranscriptomic analysis of cross-sectional BAL samples collected from mechanically ventilated patients during the first wave of the pandemic in NYC. However, while these patients all had severe disease requiring intubation, the mortality rate was much lower in this cohort because only stable patients could undergo bronchoscopy. In FY25, in a follow-up study, we analyzed longitudinal samples from 70 patients for which 1-5 samples were collected up to 6 weeks follow-up. From the same cohort, we also did a focused study on patients who are immunocompromised. This cohort has rich metadata on therapeutic regimens and symptoms. We are currently drafting a publication that will be submitted by early September 2025. In FY24 we also published a study for the same cohort where we looked at immune responses in the airways and their association with poor outcome in critically ill COVID-19 patients. These data highlight the critical role of local adaptive immunity in the airways as a key defense mechanism against primary SARS-CoV-2 infection. In another collaborative study, we investigated the risk of ventilator-associated pneumonia (VAP) in mechanically ventilated patients with COVID-19. Nosocomial bacterial pneumonia following respiratory viral infections frequently occurs in patients requiring mechanical ventilation and it is associated with prolonged hospitalization and increased mortality. We hypothesized that alterations in the lower airway microbiome and host lung transcriptome contributed to infection risk. We examined the airway samples of a multi-center cohort of 245 participants admitted for respiratory failure and mechanical ventilation, of which 38% developed VAP. Specifically, unsupervised modeling of the microbiome identified an association between the presence of lower airway oral commensals and a reduced risk of VAP. Transcriptomic analysis showed altered inflammatory pathways prior to VAP diagnosis. In a preclinical model we showed that prior exposure with human oral commensals protects against pathogens. These findings suggest that host-microbiome signatures with lower airway oral commensals may prevent bacterial pneumonia. A manuscript will be submitted for publication in August 2025.