How flu hijacks the cell: a Hamburg team maps the moment of takeover

At 09:00 UTC on 20 July 2026, a research consortium based at EMBL Hamburg and the Leibniz Research Institute for Molecular Pharmacology in Berlin released the most detailed map yet of what happens inside a human cell in the first hours of an influenza A infection. Using a combination of cryo-electron tomography, proximity labelling and quantitative mass spectrometry, the team tracked how viral proteins dismantle the host's normal housekeeping and turn it into a manufacturing line for new virus particles. The dataset, made openly available through EMBL's imaging and proteomics repositories, lets virologists see the infection not as a single event but as a sequence of takeovers, each one timed to a different stage of the viral life cycle.

The map reframes a basic question in virology. Influenza does not bring its own machinery; it borrows ours. For years, researchers knew the virus hijacked the cell's ribosomes to make viral proteins and the nucleus to replicate its RNA. What this study adds is a time-stamped inventory of which human proteins are seized, when, and in what order, down to a resolution that approaches individual molecular complexes. That level of detail is closer to what structural biologists have achieved for isolated viruses, and it now extends to the messier, more consequential question of how the virus actually operates inside living tissue.

What the team actually saw

The infection proceeds, the data show, in waves. Within the first two hours, viral polymerase complexes dock onto the host's RNA processing machinery in the nucleus, commandeering the splicing apparatus that the cell normally uses to mature its own messenger RNAs. By the fourth hour, the endoplasmic reticulum is reshaped into membranous webs that the virus uses as scaffolding for assembling its ribonucleoproteins. By the eighth hour, mitochondrial fragments are being repositioned around the assembly sites, apparently to manage the calcium and ATP demands of mass-production.

The most striking finding concerns a cluster of human proteins involved in nuclear export. Influenza cannot complete its life cycle unless newly synthesised viral ribonucleoproteins reach the cytoplasm, and the team identified the specific host factor, CRM1, with supporting roles from the THOC complex, that the virus coopts to move them. Disrupting that handoff in cell culture sharply reduced viral yield, confirming the step as a bottleneck rather than a side route. The implication is therapeutic: a drug that blocks the viral protein from binding CRM1, without interfering with the host protein's normal duties, would in principle stop the infection without killing the cell.

A different kind of antiviral target

Most existing flu drugs hit viral proteins directly. Oseltamivir and zanamivir inhibit neuraminidase, the surface enzyme the virus uses to escape an infected cell. Baloxavir targets the cap-snatching endonuclease of the polymerase. Both classes are vulnerable to resistance: a single mutation in the viral target can erode their effectiveness, which is why seasonal surveillance routinely flags variants with reduced susceptibility.

The Hamburg map suggests a different strategy. Instead of aiming at viral proteins, which mutate quickly, a next-generation drug could aim at the contact points between viral and human proteins, the docking sites that the virus cannot easily change without losing function. This host-directed approach has a longer history in HIV and hepatitis C research, where it has produced some clinical candidates, though none yet in routine use for influenza. The advance here is the precision of the contact map: for the first time, virologists can see which surfaces of which human proteins are touched, and how much of each is buried in the interaction.

That granularity matters because the great risk of host-directed antivirals is collateral damage. CRM1, for example, is essential for the normal export of many cellular RNAs, not just viral ones. A blunt inhibitor would harm the patient along with the virus. The map gives chemists a structural basis for designing a molecule that blocks the viral surface without sealing the whole export channel.

What the map does not yet show

Two limitations sit plainly inside the published work. The cells in the imaging experiments were laboratory lines, not the airway epithelial cells that flu actually infects in a human lung. The team's collaborators at FMP are now repeating parts of the protocol in primary bronchial cells, which carry a different architecture of cilia, mucus secretion and immune signalling. Influenza behaves differently in those cells, and any drug target validated only in transformed lines will need revalidation in tissue that resembles the human airway.

A second caveat is temporal. The map covers roughly the first twelve hours of infection, when viral gene expression is at its peak and structural proteins are being assembled. Later stages, including the packaging of viral genomes into new particles and the final budding event at the plasma membrane, remain less resolved. The consortium has signalled that a follow-up dataset on late-stage assembly is in preparation, but it is not yet in the public archives.

Why a structural map matters outside the lab

Influenza is one of the few infectious diseases that returns every year with enough genetic drift to escape prior immunity, and occasionally with enough shift to cause a pandemic. The 2024–25 northern-hemisphere season produced above-baseline hospitalisation rates in several European countries, and southern-hemisphere surveillance for 2026 is already showing mixed lineages of H1N1 and H3N2 co-circulating. Vaccines remain the first line of defence, but their production cycle is long and their effectiveness varies year to year. Antivirals fill the gap, and the resistance clock on each class is measured in seasons rather than decades.

The Hamburg work does not produce a new drug. It produces the structural map on which a new class of drugs could be designed. That distinction is worth holding onto, because the gap between a contact map and a clinical candidate is measured in years of medicinal chemistry, animal studies and human trials. But for a virus that has outlasted every drug thrown at it so far, knowing precisely where the wrench enters the cellular machine is the first condition for jamming it.

How Monexus framed this: the wire coverage has emphasised the technical novelty of the dataset. The angle here is what the map changes about the antiviral pipeline, and where the honest limits of the result still sit.