Periodic Reporting for period 4 - NaviFlu (Navigating the evolutionary routes of influenza viruses)
Reporting period: 2024-01-01 to 2025-06-30
Seasonal influenza viruses re-infect us repeatedly, escaping antibody recognition, due to the evolution of the virus itself. Being able to predict when and how the virus will evolve would be transformative for influenza virus control. Problematically, we have only observed one of the likely many possible routes of virus evolution and we do not know how many viable routes may have existed or about the repeatability of the observed evolution. These knowledge gaps limit the predictability of influenza virus evolution. This project seeks to fill these gaps by rigorously assessing the repeatability of influenza virus evolution and the diversity of routes the virus can explore and by better understanding the processes that drive and limit this evolution.
Recent work has shown that prolonged influenza virus infections can result in substantial virus evolution and occasionally portend virus mutational patterns on a global scale. However, observing large numbers of such infections is challenging. In this project we are using an innovative ex-vivo human airway epithelium culture system to artificially create and study prolonged human infections. Together with cutting-edge next generation sequencing and new analysis tools, we will quantify the evolutionary landscape of seasonal influenza viruses.
The project has three objectives, each building in complexity:
1–Quantify the evolutionary dynamics of seasonal influenza viruses in the absence of antibody-mediated selection to understand the fitness tradeoffs the virus makes in order to escape immunity.
2–Determine how the antibody complexity of immune sera shape the evolutionary trajectories of virus antigenic evolution.
3–Quantify the impact of differences in selection pressures by site of infection and underlying host variation on virus evolution.
Through these objectives we will “play evolution forwards”, revealing the relative roles of different factors governing the mode and tempo of influenza virus evolution and quantify the predictability of virus evolution. This will improve the design of influenza vaccines, enhance prospects for influenza control, and lay new groundwork for exploring virus evolution.
Recent work has shown that prolonged influenza virus infections can result in substantial virus evolution and occasionally portend virus mutational patterns on a global scale. However, observing large numbers of such infections is challenging. In this project we are using an innovative ex-vivo human airway epithelium culture system to artificially create and study prolonged human infections. Together with cutting-edge next generation sequencing and new analysis tools, we will quantify the evolutionary landscape of seasonal influenza viruses.
The project has three objectives, each building in complexity:
1–Quantify the evolutionary dynamics of seasonal influenza viruses in the absence of antibody-mediated selection to understand the fitness tradeoffs the virus makes in order to escape immunity.
2–Determine how the antibody complexity of immune sera shape the evolutionary trajectories of virus antigenic evolution.
3–Quantify the impact of differences in selection pressures by site of infection and underlying host variation on virus evolution.
Through these objectives we will “play evolution forwards”, revealing the relative roles of different factors governing the mode and tempo of influenza virus evolution and quantify the predictability of virus evolution. This will improve the design of influenza vaccines, enhance prospects for influenza control, and lay new groundwork for exploring virus evolution.
Objective 1: No antibody-mediated selection
We developed a high throughput experimental system that would allow us to simulate long-term influenza virus infections in human tissue and virus genomic sequencing tools to quantify evolutionary dynamics in this system. As part of Objective 1, we established a near natural tissue system, culturing differentiated human nasal epithelial cells (HAEs) at an air liquid interface, mimicking the conditions of the human nose. We have five donors with three replicates of each for a total of 15 independent virus lines. Twelve of these reached a simulated infection period of one year (120 passages, and 1440 viral generations), the remaining three surpassed 50 passages (600 generations).
We have generated 360 GB of raw paired-end next generation sequencing (NGS) data from every 10th passage of the 15 independent lines. We found that influenza A/H3N2 remains genotypically plastic even in a near-neutral environment without any constraints of antibody-mediated selection. This suggests that the evolutionary space that the virus can explorate is large and remains relatively underexplored. Speaking to the original premise of the proposal, focused on the predictability of influenza virus evolution, the virus seems capable of many paths of optimization with little evidence for predictability.
This work has been presented at numerous symposiums and international conferences.
Objective 2: Antibody-mediated selection
Following the quantification of evolutionary dynamics in a near-neutral setting, Objective 2 introduced antibody-mediated selection pressure using serum from healthy donors. Even at high serum concentrations, the virus showed remarkable resilience, continuing to replicate without detectable mutations in the antigenic sites of the HA protein. This suggests that the within-host environment alone may not provide sufficient pressure to drive the emergence of escape mutations, and that broader population-level dynamics play a more prominent role in the creation of new variants.
Objective 3: Host variability
Objective 3 entailed adding an element of host variability, and investigating the impacts of underlying host respiratory health and variations in host tissue (e.g. nasal vs lung tissue) on influenza virus evolution.
To address this objective, we first established foundational experimental methods and analysis pipelines for studying host responses in our ex vivo tissue systems. We cultured three A/H3N2 and three A/H1N1 strains in differentiated human nasal epithelial cells at an air-liquid interface. We sampled the cultures at 24, 36, 48, 72 hours, and 7 days, and performed bulk RNA transcriptomics on these samples. This work was also accepted for an oral presentation at the 2025 ESWI conference in October 2025.
In order to investigate host tissue variations, particularly between the nasal cavity and the lungs, we needed a near natural tissue system of the lungs, to complement our near natural tissue system of the nose. To this end, we developed a complex lung organoid model with spatially organized fibroblasts, epithelial, endothelial, and immune cells from donated lung tissue, capable of productive influenza virus infection. This model is the first of its kind in the Netherlands, and only one similar model exists in the USA.
In addition to the wet lab work, we have also pursued extensive modelling work to explore the conceptual basis of virus evolution both in our experimental system and in the real world. This vein of research has included four studies: 1. theoretical modelling to explore within- and between- host evolution; 2. modelling based on virus samples from children to explore how duration of infection shapes virus evolution; 3. modelling based on human serological samples and historical epidemiological data to explore the impact of the COVID-19 pandemic on influenza virus epidemics; 4. theoretical modelling to explore diagnostic testing and genomic sequencing resource allocation to monitor the emergence of new virus variants. Each of these projects had revealed new aspects of influenza virus evolution and epidemiology. This wok was widely used during the pandemic including by the World Health Organization and its collaborative partners.
We developed a high throughput experimental system that would allow us to simulate long-term influenza virus infections in human tissue and virus genomic sequencing tools to quantify evolutionary dynamics in this system. As part of Objective 1, we established a near natural tissue system, culturing differentiated human nasal epithelial cells (HAEs) at an air liquid interface, mimicking the conditions of the human nose. We have five donors with three replicates of each for a total of 15 independent virus lines. Twelve of these reached a simulated infection period of one year (120 passages, and 1440 viral generations), the remaining three surpassed 50 passages (600 generations).
We have generated 360 GB of raw paired-end next generation sequencing (NGS) data from every 10th passage of the 15 independent lines. We found that influenza A/H3N2 remains genotypically plastic even in a near-neutral environment without any constraints of antibody-mediated selection. This suggests that the evolutionary space that the virus can explorate is large and remains relatively underexplored. Speaking to the original premise of the proposal, focused on the predictability of influenza virus evolution, the virus seems capable of many paths of optimization with little evidence for predictability.
This work has been presented at numerous symposiums and international conferences.
Objective 2: Antibody-mediated selection
Following the quantification of evolutionary dynamics in a near-neutral setting, Objective 2 introduced antibody-mediated selection pressure using serum from healthy donors. Even at high serum concentrations, the virus showed remarkable resilience, continuing to replicate without detectable mutations in the antigenic sites of the HA protein. This suggests that the within-host environment alone may not provide sufficient pressure to drive the emergence of escape mutations, and that broader population-level dynamics play a more prominent role in the creation of new variants.
Objective 3: Host variability
Objective 3 entailed adding an element of host variability, and investigating the impacts of underlying host respiratory health and variations in host tissue (e.g. nasal vs lung tissue) on influenza virus evolution.
To address this objective, we first established foundational experimental methods and analysis pipelines for studying host responses in our ex vivo tissue systems. We cultured three A/H3N2 and three A/H1N1 strains in differentiated human nasal epithelial cells at an air-liquid interface. We sampled the cultures at 24, 36, 48, 72 hours, and 7 days, and performed bulk RNA transcriptomics on these samples. This work was also accepted for an oral presentation at the 2025 ESWI conference in October 2025.
In order to investigate host tissue variations, particularly between the nasal cavity and the lungs, we needed a near natural tissue system of the lungs, to complement our near natural tissue system of the nose. To this end, we developed a complex lung organoid model with spatially organized fibroblasts, epithelial, endothelial, and immune cells from donated lung tissue, capable of productive influenza virus infection. This model is the first of its kind in the Netherlands, and only one similar model exists in the USA.
In addition to the wet lab work, we have also pursued extensive modelling work to explore the conceptual basis of virus evolution both in our experimental system and in the real world. This vein of research has included four studies: 1. theoretical modelling to explore within- and between- host evolution; 2. modelling based on virus samples from children to explore how duration of infection shapes virus evolution; 3. modelling based on human serological samples and historical epidemiological data to explore the impact of the COVID-19 pandemic on influenza virus epidemics; 4. theoretical modelling to explore diagnostic testing and genomic sequencing resource allocation to monitor the emergence of new virus variants. Each of these projects had revealed new aspects of influenza virus evolution and epidemiology. This wok was widely used during the pandemic including by the World Health Organization and its collaborative partners.
Our findings have challenged key assumptions in the field. Contrary to expectations of rigid fitness landscapes, we discovered unexpected genotypic plasticity in H3N2 viruses even in neutral environments, indicating a vast, underexplored evolutionary space with naturally low predictability. Furthermore, we demonstrated that within-host antibody pressure alone is surprisingly insufficient to drive typical antigenic escape mutations. This fundamentally shifts the state of the art by suggesting that population-level transmission dynamics, rather than just isolated within-host selection, are likely the primary drivers of global viral drift.
Technologically, we pushed the boundaries of standard tissue models by developing a complex, multi-cellular lung organoid model—the first of its kind in the Netherlands. Combined with our modelling work, which has already informed global health bodies, these tools significantly expand the capability of the field to study respiratory viral pathogenesis and evolution.
Technologically, we pushed the boundaries of standard tissue models by developing a complex, multi-cellular lung organoid model—the first of its kind in the Netherlands. Combined with our modelling work, which has already informed global health bodies, these tools significantly expand the capability of the field to study respiratory viral pathogenesis and evolution.