Pre-Print: A Cattle-derived Human H5N1 Isolate Suppresses Innate Immunity Despite Efficient Replication in Human Respiratory Organoids

#18,933

As the graphic above illustrates, over the past 20+ years we've seen a wide variance in the virulence - and CFR (case fatality rate) - of the H5N1 virus in countries around the globe. 

While 84% of confirmed Indonesian cases died, less than half that percentage succumbed in Egypt, and in Bangladesh, only one fatality out of 12 cases has been recorded. 

As we discussed more than a dozen years ago, in Differences In Virulence Between Closely Related H5N1 Strains, while there are likely many factors involved - including quality of, and delays in seeking medical care – it suggests that some HPAI H5N1 strains (clades, subclades, genotypes) are far more virulent than others.

Here in the United States, over the past couple of years, we've seen 70 confirmed and 7 probable H5N1 infections, and only 1 fatality. At the same time, in Cambodia nearly half of 33 recent cases (of a different clade; 2.3.2.1e) have died. 

While it may be obvious that some strains of H5N1 are deadlier than others, what isn't as apparent is what makes the difference.  We have seen some theories, however.

• In yesterday's mBio: Low levels of influenza H5N1 HA and NA antibodies in the human population are boosted by seasonal H1N1 infection but not by H3N2 infection or influenza vaccination, researchers suggest a similar NA gene in seasonal H1N1 may evoke some degree of immunity against H5N1. 

• It has also been suggested that the route of exposure (ocular or gastrointestinal) - or a lower viral load - may produce more localized, less severe infections.

• And unlike places like Cambodia, Egypt, and Indonesia - in the United States it is primarily adults occupationally exposed to H5N1 - and we already know that HPAI H5 has a predilection for infecting a younger cohort.

Of course, there may be multiple factors at work. 

Today, however, we have a preprint that offers another intriguing possibility; that the B3,13 genotype disables our innate immune response - suppressing cytokine production (and many symptoms) - but still replicates efficiently in human hosts. 

Using lab-created (influenza-naive) lung organoids, these researchers were able to study the response of the innate immune system.  How their results match up with humans typically saddled with a long and complex immune history remains to be seen. 

But they did find that the TX37-H5N1 human-bovine strain replicated quite efficiently in these human lung organoid tissues, but that the virus failed to trigger a robust immune response, which could help explain mild or asymptomatic infections. 

I've reproduced the abstract, and an excerpt from the discussion below. Follow the link to read it in its entirety.  I'll have a bit more when you return.

A cattle-derived human H5N1 isolate suppresses innate immunity despite efficient replication in human respiratory organoids
Shintaro Shichinohe, Hikaru Sugimoto, Masako Yamasaki, Rina Hashimoto, Tatsuru Morita, Daiki Kobayashi, Takahiro Hiono, Daisuke Motooka, Makoto Shimooka, Norikazu Isoda, Mai Thi Quynh Le, Ayato Takada, Yoshihiro Sakoda, Eiryo Kawakami,Kazuo Takayama, Tokiko Watanabe
doi: https://doi.org/10.1101/2025.11.02.684669
This article is a preprint and has not been certified by peer review

Preview PDF

Abstract

The H5N1 high pathogenicity avian influenza virus (HPAIV) of clade 2.3.4.4b, which spreads globally via wild birds, has become a major public health concern because it can infect a variety of mammals, including humans. In 2024, infection of dairy cattle with H5N1 HPAIV clade 2.3.4.4b was confirmed in the United States, and subsequent human cases were reported. Although these viruses are highly pathogenic in animal models, human infections have generally been mild, revealing a striking discrepancy. 

Here, we characterized the cattle-derived human H5N1 isolate A/Texas/37/2024 (TX37-H5N1) using three-dimensional human respiratory organoids derived from induced pluripotent stem (iPS) cells. Despite efficient replication, TX37-H5N1 induced minimal interferon and inflammatory cytokine responses.

 Bulk and single-cell RNA sequencing revealed reduced STAT1-mediated transcriptional activity in TX37-H5N1-infected organoids compared to the historic H5N1 human isolate A/Vietnam/1203/2004. These findings suggest that TX37-H5N1 fails to trigger the strong innate responses, including robust cytokine production, that are typically associated with severe H5N1 disease and are thought to contribute to cytokine storm-medicated pathogenesis. 

This attenuated response may help explain the discrepancy between the high pathogenicity of TX37-H5N1 in animal models and its mild clinical presentation in humans. While zoonotic influenza risk is often assessed using cell lines or animal models, our study highlights the value of using human respiratory organoids to evaluate human-specific virus-host interactions. This platform provides a complementary tool for assessing the risk of emerging avian influenza viruses.

        (SNIP)

Our study provides evidence that selective suppression of innate immunity may explain the mild clinical manifestations of B3.13 H5N1 viruses in humans and offers new perspectives for reconciling inconsistencies between animal models and clinical outcomes.

Viruses like TX37-H5N1 that replicate efficiently yet cause only mild disease in humans may evade detection if only traditional risk assessment strategies are employed (e.g., measuring LD₅₀ values and viral titers in mice and ferrets). 

Immune response profiling using human-derived organoids, as demonstrated here, offers a powerful complementary approach to capture human-specific innate immune responses that may not align with animal-model outcomes. Moreover, viruses with low-pathogenic but high-replication phenotypes pose unique challenges for public health. Such viruses may silently spread through symptomatic infection and prolonged viral shedding. 

While no sustained human-to-human transmission of B3.13 viruses has been reported to date, future mutations may alter transmissibility.

Because no single model can fully capture public health threats of emerging influenza viruses, an integrated framework is needed to evaluate viral replication efficiency, immune modulation, and transmissibility. Finally, the suppression of the STAT–IRF pathway identified in this study may offer new avenues for therapeutic intervention and have implications for pandemic preparedness and emerging infectious disease control. Future efforts should focus on characterizing host immune  responses in human cases and performing cross-strain comparisons to identify predictive markers of human pathogenicity.

        (Continue . . . )

If we could depend upon H5N1 to only produce mild symptoms going forward, then we could all relax - deal with a brief bout of conjunctivitis or the sniffles - and get on with life. 

But we are seeing an entirely different picture with other mammalian species, and there are no guarantees this virus won't evolve into a more virulent strain in humans.

Even if the B3.13 genotype can't make that leap, there are hundreds of other HPAI H5 genotypes currently circulating around the globe, and it only takes one to hit the genetic jackpot to plunge the world back into a pandemic crisis. 

Unfortunately, mild or asymptomatic individuals are the least likely to be tested for the virus - meaning that if  HPAI H5 were spreading stealthily in the community - we could easily be none the wiser. 

For more on the challenges in tracking the spread of this virus in the community, you may wish to revisit NAS : Diagnostic Tools, Gaps, and Collaborative Pathways in Human H5N1 Detection (Rapid Expert Consultation).