#19,354
In November of 2024, we learned (see Canada: PHAC Confirms HPAI H5N1 Genotype D1.1 In B.C. Human Infection) of a severe H5N1 influenza infection involving a teenage girl in British Columbia which was caused by the recently emerged D1.1 genotype.Despite an intense investigation, the source of her infection was never determined.
In late December, it was disclosed (see NEJM: Critical Illness in an Adolescent with Influenza A(H5N1) Virus Infection) that the patient was a 13-year-old girl, and that after several weeks of intensive care she was able to come off ECMO (November 22nd), and was extubated on November 28th.
In Referral: MedCram On Avian Flu Mutations That Favor Human Transmission, we looked at early reports of ambiguous mutations at several key sites (Q226 and E190 (H3 numbering)) in the HA gene. The NEJM report also mentioned the PB2-E627K mutation was detected (52% allele frequency).
All of which raised concerns that this particular patient might have been infected with a more efficient human-adapted virus. But, as we've seen with most H5 human infections, there was no indication of human-to-human transmission.
Since then we've also seen studies suggesting that Avian influenza virus A(H5N1) genotype D1.1 is better adapted to human nasal and airway organoids than genotype B3.13, along with at least two North American deaths (here & here).
In fact, they did the opposite.
These mutations were detected 8 days post symptom onset in a minority (≤ 35%) of viral sequences detected in the patient's lower respiratory tract, and may have arisen after the patient was infected. Together, and/or separately, they produced almost-undetectable binding to α2,3 (avian-like) and α2,6 (human-like) sialoside receptors.
Despite this, the patient's lungs were clearly infected, producing severe - even life-threatening - illness.
This is a lengthy, highly technical, and (for me, at least) daunting research paper which raises a lot of unanswered questions.
While that may seem like `good news', it doesn't tell us much about the wild-type virus that originally caused this girl's infection, or where or how that virus entered the body.
This is a sample of one, which suggests the Q226H and E190D mutations appear less alarming than originally thought. It does not, however, change the evaluation of the Q226L mutation, which has been shown to enhance human-type (α2,6) receptor binding.
Less reassuringly, the authors note that severe disease occurred despite the lack of α2,6 receptor binding signatures. They note that human infection "is not always dependent" on α2,6 binding.
One potential alternative the authors propose is the virus may still attach well enough to fuse, through many weak contacts or through receptors they didn't test, though this is only a hypothesis.
Diminished sialoside binding in novel H5N1 influenza hemagglutinin variants identified in a human patient
John H. Ni, Saeid Malek Zadeh, Alison M. Berezuk, Ryan Lynam, Peter Axerio-Cilies, Xing Zhu, Katharine S. Tuttle, Gethin Rh. Owen, Maria Tokuyama & Sriram Subramaniam
Nature Communications volume 17, Article number: 9972 (2026)
Abstract
In 2024, an adolescent female in British Columbia was hospitalised presenting with severe symptoms including respiratory failure due to infection with a novel H5N1 influenza strain (BC24). Using cryogenic electron microscopy, we show here that the N169 α2,3-linked auto-glycan that is found in the sialic acid binding site of previously studied H5 hemagglutinin (HA) proteins is absent in purified BC24 HA protein, suggesting greatly reduced affinity for α2,3-linked sialosides.
Glycan microarray and enzyme-linked immunosorbent assay analyses show that HA variants identified in the BC24 case display severely reduced or no binding to both α2,3-linked sialosides and α2,6-linked sialosides. Full-length BC24 HA expressed in A549 alveolar carcinoma cells drives membrane fusion, albeit at lower levels than previous H5 HA proteins, and post-infection sera from the patient display strong binding to BC24 HA and HA proteins from other influenza subtypes.
As each of the two mutations of interest, independently and in conjunction, severely reduce sialoside binding, there appears to have been in this case multiple populations of virus with the diminished receptor binding phenotype. The substantial minority prevalence of weakly binding HA variants in this BC24 case reveals further complexity in the factors that may be present in severe avian influenza infection.
(SNIP)
The cryo-EM structural analyses demonstrate that the combined effect of the E190D and Q226H mutations sequenced from BC24 HA alters the glycan-binding behavior of this HA. Subsequent experiments through glycan microarray analyses, ELISA, and fusion assays in A549 cells support the hypothesis that the BC24 HA has severely reduced affinity for both α2,3-linked and α2,6-linked sialosides. In particular, ELISA experiments show that both the E190D and Q226H mutations reduce sialoside affinity in isolation, as well. Therefore, multiple distinct minority variants in the BC24 case are expected to have a reduced binding phenotype.
We show that the virus likely retains the ability to drive membrane fusion, adding complexity to our understanding of avian influenza’s paths to generating severe disease in humans. Several important conclusions are suggested by our analyses. The lack of binding to α2,6-linked sialosides by HA variants in the BC24 case implies that human infection is not always dependent on such binding. Moreover, generally reduced binding of HA to sialosides, including both α2,3- and α2,6-linked sialosides, can be a phenotype of fit viruses in a severe human disease context.
We speculate that such a reduction in binding may enable deeper penetration into the lung and potentially contribute to a more severe infection; however, further study will better elucidate how differences in receptor-binding sites influence the infectivity of this influenza strain.
While the original reports of two mutations - E190D and Q226H - were seen as potential red flags for human adaptation, today's report found the opposite; these mutated HA proteins barely bound to either avian-type or human-type receptors - yet the patient became critically ill.
The authors speculate that weak binding may have helped these variants reach the lower airway, but the belated retrieval of viral samples can't show when the mutations arose or what the original virus looked like.
All of which makes these findings only partly reassuring. These two mutations now look less alarming than originally feared, but the patient's infection still flourished, causing severe, life-threatening illness.
A reminder that - even after years of research - there's still a lot we don't know about how viruses in general - and H5N1 in particular - interact with their hosts.