Most people who breathe in tuberculosis win the fight and never know it happened. Roughly a quarter of everyone alive is carrying Mycobacterium tuberculosis right now, and only about 5 to 10 percent will ever get sick from it. The bug is in there. The immune system just handles it. For a long time the easy assumption was that the people who lose that fight lose it because their T cells are too weak to finish the job. A new study out of the Francis Crick Institute went down into the actual lung to watch, and what it found flips that assumption on its head.

The people who progress to disease do not have too few T cells doing too little. They have too many neutrophils doing far too much, and the neutrophils are crowding the T cells out. I have read a lot of TB immunology, and this one rearranged how I think about who survives it.

Going to the scene of the crime

What makes this paper different is where the researchers looked. Almost everything we know about who progresses to active TB comes from blood, because blood is easy: you draw a tube, you run the genes. The Crick team, led by the immunologist Anne O’Garra, went to the place TB actually happens. Working with the University Hospitals of Leicester NHS Trust, they collected bronchoalveolar lavage, fluid washed out of the deep airways, from recent household contacts of people with pulmonary TB, the population that just shared a kitchen and a bedroom with an infectious case. Some of those contacts would go on to control the infection. Some would progress to disease. The samples were read three ways: bulk RNA sequencing, single-cell RNA sequencing, and flow cytometry, cell by cell.

This is not a casual sample to collect. Bronchoscopy is invasive, and you are doing it on people who are, for the moment, perfectly well. That is exactly why a dataset like this barely exists, and it is also the catch I will come back to.

The airway that loses

When the team sorted the airways by what happened next, a split fell out that was almost embarrassingly clean.

In the people heading toward active disease, and in patients already diagnosed, the airway was a neutrophil town. Neutrophils are the immune system’s blunt instrument, the first responders that show up to a wound and start burning the place down to deny it to bacteria. Useful in a cut on your hand. Catastrophic, it turns out, in the delicate gas-exchanging tissue of your lung. About half of those neutrophils were lit up with type I interferon-linked genes, and they were pumping out high levels of CXCL8, the chemical flare that calls more neutrophils in. You can see the trap from here: a flood that summons its own reinforcements, an inflammatory loop tightening on itself.

And the T cells in that airway, the cells that are supposed to be quarterbacking the whole defense? They were exhausted. The single-cell data caught them mid-collapse, carrying signatures of exhaustion, of cytotoxicity firing without control, and of outright cell death. The team found a clean inverse relationship: the more the neutrophils filled the airway, the fewer functioning T cells were left standing. The blunt instrument was not just failing to clear the bacterium. It was muscling out the cells that could have.

The airway that wins

Now flip it. The contacts who controlled their infection had a lung that reads, honestly, like restraint. Lower T cell activation, not higher. Their cells were expressing regulatory genes and even stem-like genes, the program that lets a population of immune cells renew itself and keep answering the bell for the long haul. That is the detail I keep turning over, because it runs against instinct. TB is a slow, grinding siege of an infection. The winning immune response is not the one that screams loudest on day one. It is the one that can still show up in month six.

We tend to root for the aggressive immune response, the big inflammatory swing. This says the people who beat TB are the ones whose lungs keep their cool, hold the T cells in reserve, and never let the neutrophil brawl get started.

The thread O’Garra has been pulling for sixteen years

The villain in this story, type I interferon driving a neutrophil program, is the same villain O’Garra named back in 2010, when her group published the landmark interferon-inducible, neutrophil-driven blood signature of active TB in Nature. That paper rewrote how the field thinks about TB and put type I interferon, long assumed to be antiviral and helpful, on the suspect list. But it was a blood signature. It told you something systemic was happening without telling you where, or why it mattered, in the organ that counts.

Sixteen years later, the same lab has caught the same fingerprint at the actual scene, in the airway, cell by cell, and shown what it is doing: recruiting neutrophils and choking off the T cell-macrophage interactions the lung needs to contain the bug. A companion thread in mice spells the mechanism out even more bluntly. Type I interferon drives neutrophils to swarm, and when researchers blocked the interferon receptor early, protective CD4 T cells and their contacts with macrophages flowed back into the lesions and the lung bacterial burden dropped. A decade and a half of work, converging on one idea: in tuberculosis, the interferon-neutrophil axis is not the cavalry. It is the friendly fire.

The truth is in the lung. The test has to come from blood.

So can we act on it? Here I make myself slow down, because the press-release version runs straight to “new treatments,” and the data is not there yet.

What the team has is a clean mechanistic map and a real hypothesis: that the balance of neutrophils and T cells in the airway is what tips a quiet infection into a deadly one. O’Garra frames the payoff as a maybe, that “if we can find a way to identify these immune signatures early”, you might predict who is at risk and intervene before they get sick. The candidate interventions she points to, CXCR2 inhibitors that block the neutrophil-recruiting signal and are already in trials for other lung diseases, and vaccines aimed at those stem-like T cells, are promising and untested in this setting. Nobody has shown that damping the neutrophil flood in an exposed human prevents progression. That is the trial that has to happen next, and it has not.

Then there is the brute practical problem: you cannot bronchoscope a quarter of the planet. The whole appeal of the older blood work, including newer blood transcriptomic scores being tested in household contacts in Brazil, is that a blood draw scales and a lung washout never will. The airway is where the truth lives. The blood is where any usable test will have to read it, and bridging those two is the unglamorous work that turns a stunning Nature Immunology figure into something that reaches a kid in a crowded house in Jakarta.

TB is not a historical footnote. In 2024 the WHO confirmed it had reclaimed its title as the world’s single deadliest infectious disease, back above COVID, with 1.25 million people dead in 2023 and 10.8 million newly sick. The same report put the global TB response at barely a quarter of its target, $5.7 billion against a $22 billion need. We spent the back half of the pandemic treating one respiratory pathogen as the only one that mattered, and the actual record-holder kept killing on schedule, mostly poor people, mostly far away, mostly unwatched. A study that finally explains why some of them get sick should command attention and money. Whether it does is a test of priorities, not of science.

What I make of it: this is the most convincing mechanistic story I have seen for the oldest question in TB, and it is the same answer the same scientist has been circling for sixteen years, now nailed down where it counts. I would watch the CXCR2 trials closely, and I would watch even harder to see whether anyone funds the blood test that makes this finding portable. The lung told us the truth. The question is whether we bother to listen at scale.

Sources

  1. Nature Immunology – O’Garra et al., airway immune signatures of protection and progression in recent TB household contacts (2026)
  2. Francis Crick Institute / Medical Xpress – early immune clues that determine who develops TB, with study detail and researcher quotes (2026)
  3. Nature – Berry et al., an interferon-inducible neutrophil-driven blood transcriptional signature in human tuberculosis (2010)
  4. PMC – Type I IFN drives neutrophil swarming, impeding lung T cell–macrophage interactions and TB control (mouse mechanistic study)
  5. WHO – Tuberculosis resurges as top infectious disease killer, Global TB Report 2024
  6. medRxiv – Combining blood transcriptomic signatures to predict progression to TB among household contacts in Brazil (2025)