Genomic profiling of influenza infections to identify biomarkers of disease severity

National Institute Of Allergy And Infectious Diseases

Investigators

Linked publications & trials

Abstract

Host response to influenza infection is a complex trait that involves entire host-pathogen interaction networks of RNA transcripts, proteins, and metabolites that have an impact on cellular, tissue, and whole-organism behaviors, which ultimately define both the risk and severity of infection. This project could substantially broaden our understanding of severe influenza infection and help us make accurate predictions of influenza severity. As part of this project, we use an integrative systems-level approach to discover how host factors affect the evolution and transmission of influenza virus, and whether specific host factors could be leveraged as predictive markers of the responses to infection and to vaccination. We also set out to determine how disease severity is impacted by microbial communities in the respiratory tract, to reveal key signatures that could be targeted in novel therapeutics. As part of this project we tackle different types of studies that include: (a) Virus evolution and genetic diversity within infected hosts (b) Systems biology of infection and vaccination (c) Respiratory tract microbiome profiles in lung diseases (viral and non-viral) (d) Virus capture and detection using machine learning for novel diagnostic platforms (a) Virus evolution and genetic diversity within infected hosts: In humans and animal models, immunity against neuraminidase (NA) reduces disease severity and viral replication during influenza infection. However, we have a limited understanding of the impact of NA immunity on viral transmission. In FY25, we contributed to a manuscript published by our collaborator, Troy Sutton (PSU), on assessing if vaccine-induced NA immunity could disrupt transmission of the 2009 pandemic H1N1 virus [Septer et al, 2024, mBio]. The virus transmitted efficiently through chains of transmission in the presence of NA immunity, but NA vaccinated animals shed significantly less virus and had accelerated viral clearance. To determine if immune pressure led to the generation of escape variants, we sequenced viruses in ferret nasal wash samples. No mutations in NA were identified. These findings demonstrate vaccine-induced NA immunity is not sufficient to prevent infection via airborne exposure and onwards airborne transmission of the 2009 pandemic H1N1 virus. In FY25, we continued working with Dr. Sutton but this time on testing the effects of combined immunity to hemagglutinin (HA) and NA on airborne transmission of the 2009 pandemic H1N1 virus from pre-immune ferrets to naïve respiratory contacts. For all strategies used to induce immunity, combined immunity to HA and NA resulted in the largest reductions in viral shedding and transmission. Moreover, immunity to HA and NA conferred additive rather than synergistic reductions in transmission. Transmission was not associated with the emergence of escape variants, and logistical regression analyses showed the probability of transmission was less than 50% when viral titers were reduced below 101.5 TCID50/mL. These studies define the relationship between immunity to HA and NA on airborne transmission and identify a threshold viral titer indicative of decreased transmission in ferrets. A manuscript will be submitted to Science Advances in September 2025. Age is one of the most notable host risk factors associated with influenza infection. In collaboration with Dr. Ted Ross (UGA and Cleveland Clinic) and Dr. Lara Mahal (U Alberta), we established an age-specific ferret model to study the age-related differences in susceptibility and severity during influenza infection. Our previous findings detailed changes in viral kinetics, weight loss, disease symptoms, tissue damage, and glycomic profiles from ferret samples, where we confirmed that the different ferret age groups mimic age-related disease dynamics observed in humans. In FY25, we integrated multi-omics (transcriptomic, glycomic, 16s rRNA, and viral genomics) data collected from the young, adult, and aged ferrets infected with the 2009 pandemic H1N1 influenza A strain. We defined how age-specific differences impact the transmission bottleneck at infection and the rate of viral divergence. We also profiled how the tissue-related microenvironments, characterized by the host response and microbiome (nasal wash, trachea, soft palate, and lung), shape the intra-host evolution of the influenza virus. We are currently working on a manuscript for submission. In FY25, we started a new collaborative study with Stacey Schultz-Cherry (St. Jude) to characterize the genetic diversity and host adaptation of H5N1. In a recent outbreak, a highly pathogenic avian influenza A H5N1 strain has begun circulating in cattle, indicating a shift in its host and cellular tropism. Reports of H5N1 viruses isolated from humans with severe infections revealed multiple mutations that enabled the virus to bind a2,6 receptors. This underscores the importance of understanding viral tropism in mammalian-derived viruses. This project aims to define the genomic and structural variations of H5N1 influenza strains that contribute to shifts in viral tropism. Dr. Schultz-Cherry’s lab isolated primary cells from animals and differentiated them using an air-liquid interface cell model, including bovine tracheal and mammary epithelial cells, swine tracheal and nasal epithelial cells, and ferret nasal epithelial cells. Isolated H5N1 strains were used to infect the different cell lines, and supernatant was collected over time. We are conducting viral RNA sequencing using the supernatant collected and performing virus variant analyses. Comparing the longitudinal changes in the virus populations will allow us to characterize viral population shifts associated with cellular, host, and viral strain differences. Many RNA respiratory viruses, including influenza, generate defective viral genomes (DVGs), which can be packaged into defective interfering particles (DIPs). DVGs and DIPs have been shown to confer protection against disease severity by stimulating the host's innate immune response and decreasing the viral load of standard virions. We have shown that different DIPs vary in their protective ability. In FY25, we initiated a study to define the viral and host-specific mechanisms of DIP interference. We established a human airway epithelial cell model in our lab to interrogate the mechanistic effects of DIPs on influenza A virus infection. Immortalized basal progenitor cells collected from a healthy non-smoking adult are expanded and differentiated into a polarized, pseudostratified culture system that recapitulates the airway epithelial environment and yields more physiologically relevant data than traditional monolayer lung cell cultures. Infections are performed using standard wild-type (WT) viruses, purified DIPs previously characterized by our group, defective virus controls that lack interfering ability, and co-infections with WT stock and purified DIPs or WT and control viruses. As part of our analyses of the genetic diversity of influenza virus within the infected host, we have developed software (DiVRGE) that in FY25 we have made publicly available on our lab GitHub page (github.com/GhedinSGS). A manuscript is in preparation. (b) Systems biology of infection and vaccination Severity of influenza infection and effectiveness of the seasonal influenza vaccine vary between individuals and across populations. As part of the Collaborative Influenza Vaccine Innovation Centers (CIVIC) program, we have been collaborating on the analysis of the host systemic response to infection and vaccination using transcriptomic profiling of whole blood. The effectiveness of the seasonal influenza vaccine varies between individuals. Specific risk groups with comorbidities, such as the obese, have an increased risk of infection even when vaccinated. It has been proposed that a heightened baseline inflammatory state and alterations in adaptive immune cell populations in obese individuals may adversely impact immune response to pathogens and vaccines. Therefore, understanding differences in vaccine response in obese individuals is critical for the development of improved influenza vaccine design and strategies. Previously, we identified a transcriptomic signature prior to seasonal influenza vaccination that is predictive of an individual’s response to the influenza vaccine. In FY25, we published a study on the dynamics of the host response following vaccination. We analyzed systemic gene expression from whole blood collected from 163 vaccinated individuals on days 3, 7, and 28 post-vaccination. Gene expression profiles were compared between high and low vaccine responders in both obese and non-obese groups. We detected characteristic changes in gene expression over time, with a peak at day 7 post-vaccination. A large subset of genes associated with the adaptive immune response exhibited significantly different expression in high- and non-responders to the vaccine. Distinct gene sets presented differing temporal patterns in different weight and response groups, indicative of molecular processes impacting the immune response to vaccination. A manuscript is currently in preparation reporting on this study [Hockman et al, 2025, Microbiology Spectrum]. Indigenous populations across the world are at substantially higher risk of hospitalization and morbidity from influenza infection. They are also typically part of communities where vaccination coverage is low due to limited access by regular health systems. How such vaccine-naïve populations respond to influenza vaccination, and whether we could predict the efficiency of the response, could provide insight into the molecular mechanisms that underpin the diverse immune reaction, a critical step toward developing efficient vaccines. These studies could also help better understand how children respond to vaccination. In FY25, in a collaborative study with Stacey Schultz-Cherry (St-Jude Children’s Research Hospital) and Juan Dib (Tropical Health Foundation), who have established cohorts of indigenous populations in the Sierra Nevada mountains of Colombia, we are teasing out factors that help better understand why the response to influenza infection and vaccination varies amongst individuals as it is likely the result of both extrinsic (e.g., environment and lifestyle) and intrinsic (e.g., genetic) factors. Heritable variation in histo-blood group antigen (HBGA)-related genes, such as ABO, FUT2 (secretor), and FUT3 (Lewis A gene), were shown to be associated with the response to various virus infections and rotavirus vaccines. We have been testing the secretor status of this Colombian cohort with a genotyping assay targeting the known FUT2 gene mutations and doing correlation analyses with seroconversion in response to the vaccine. We are currently also evaluating the response by glycomic profiling (in collaboration with Lara Mahal at the University of Alberta) and transcriptomics. These studies highlight how various omics datasets and integrative modeling can extract new signatures of seroconversion from highly variable data. Non-secretor status determination is clinically actionable information as it would help predict the response to vaccination and inform the need to add an adjuvant to the vaccine for individuals who are non- or low-secretors. (c) Respiratory tract microbiome profiles in lung diseases (viral and non-viral) A factor that can also impact disease severity in respiratory conditions is the microbiome. In FY25, we collaborated with Leo Segal (NYU Langone) on a study to look at the metabolic microenvironment of the lower airways. Using samples of healthy individuals and via several mouse experiments we demonstrated by multi-omics that even with very dynamic clearance of transient microbes in the lower airways, the lung microbiome contributes to niche construction and to the metabolic microenvironment [Wong K.W. et al, 2025, Cell Host & Microbe]. Given the direct exposure of the upper airways to the environment, studying the respiratory microbiome and its interactions with environmental factors is key to understanding its complexity. In FY25, in collaboration with Prof. Luca Ferrari (University of Milan), we investigated the effects of seasonality and exposure to air pollutants, including microbes carried by air particles, on the nasal microbiome of health office workers. We observed an association between the relative abundance of respiratory bacteria identified in the indoor TSP and the upper respiratory microbiome, and that air pollution affected the relative abundance of specific microbial taxa. Furthermore, both indoor and outdoor TSP samples contained broad spectrum antibiotic resistance genes [Solazzo, G., …, Ferrari, L., Ghedin, E. (2025) Ecotoxicol Environ Saf. ] In FY25, we continued to quantify the effects of environmental exposure on the airway microbiome by contrasting populations with different lifestyles. Studying the microbiomes of communities with distinct lifestyles and exposure to environmental conditions represents an opportunity to identify factors that best explain variability across populations. Our focus in. a collaboration with Marcel Tongo (Institute of Medical Research and Study of Medicinal Plants (IMPM), Cameroon) has been on semi-nomadic hunter-gatherers in the rainforest of Cameroon (Baka pygmies) comparing their microbiome to individuals living in urban settings. We are characterizing the microbiomes of the upper airways, including the characterization of novel viruses and antibiotic resistance genes. In FY25 we have a new collaboration with Kevin Fennelly (NHLBI), Shamus Carr (NCI), and Adrian Zelazny (Clinical Center) where we are profiling the microbiome and antibiotic resistance in lung tissue from patients with pulmonary Mycobacterium avium complex (MAC). Opportunistic pathogens such as Aspergillus or Mycobacterium species (spp.) can cause severe pulmonary disease with the most common due to species in the MAC. The goals of the study, for which we applied to Bench-2-bedside funding, is to profile the microbiome from resected lung cavity and related biofilm created by either Mycobacterium spp. Or aspergilloma, isolate mycobacterial species for genomic analyses, and combine with imaging results to examine potential interactions of various species in the biofilm. These results may be the first step to identify a way to disrupt the biofilm and optimize therapy to the cavitary environment without resection. (d) Virus capture and detection for novel diagnostic platforms In collaboration with Mauricio Terrones group (PSU) and Shengxi Huang (Rice), in FY25, we continued our work on testing influenza virus on a microfluidics carbon nanotube platform (developed by Dr. Terrones) and Surface-enhanced Raman spectroscopy. The goal of this effort was to reach strain-level identification of viruses in real-time. Label-free detection is more powerful than targeted approaches that may fail to detect pathogens as they mutate, miss instances of coinfection, and only test for a limited number of pathogens within a single assay. Surface-enhanced Raman spectroscopy is a non-invasive, highly sensitive optical technique that characterizes molecules based on the vibrational modes of their bonds. Strain-level identification of viruses is critical for an effective public health response to potential outbreaks, yet current diagnostic methods often lack the necessary speed or sensitivity. AI assisted surface-enhanced Raman spectroscopy (SERS) offers great potential for fast and precise virus clarification through the unique vibrational fingerprints of biological components, but existing protocols typically operate outside of the tissue transparent near-infrared (NIR) window, limiting bio-applicability due to autofluorescence and photodamage. For this project, we contributed to a study that reports on an AI-empowered NIR-SERS platform that integrates machine learning with a rationally designed hybrid substrate—gold nanostars (AuNSs) coupled with gold-coated carbon nanotube arrays (AuCNT). This provided accurate classification of respiratory viruses—including influenza viruses and coronaviruses, at the type, subtype, and strain levels. This AI-driven diagnostics approach shows promise for enhancing viral surveillance of novel strains and outbreak response capabilities, thus potentially addressing critical challenges in global public health preparedness.