9JUN

B3.13 replicates better in human nasal tissue

4 min read

09:58UTC

NIH/NIAID researchers Flagg and Winski published peer-reviewed evidence on 12 May that the H5N1 strain dominating US dairy cattle replicates more efficiently in human nasal epithelium than any historical H5N1 strain and partly disables the immune alarm that would normally limit early replication.

← Back to Ebola outbreak gets an R0, and a fork Jump to analysis ↓

ScienceDeveloping

Key takeaway

B3.13 outperforms historical H5N1 in human nasal tissue and delays the immune alarm, while surveillance weakens.

NIH/NIAID researchers Flagg and Winski published a study in CDC Emerging Infectious Diseases (DOI: 10.3201/eid3205.260053) on 12 May showing that B3.13 (the clade 2.3.4.4b H5N1 genotype circulating in US dairy cattle since 2024) replicates more efficiently in human nasal epithelium cultures than historical H5N1 strains. ¹ The same paper found that B3.13 suppresses ISG (interferon-stimulated gene) responses, the innate immune system's first-line molecular alarm; a virus that triggers fewer ISGs buys more time to replicate before the body mounts a coordinated defence.

The Emory University aerosol study published in PLOS Biology on 1 May had already shown B3.13 present in submicron aerosol particles inside milking-parlour air across 14 California farms . Together, the two papers complete a mechanistic two-step pathway: B3.13 reaches the nasal epithelium via the aerosol route, and once there it replicates efficiently while partially evading innate immunity. Neither paper claims human-to-human transmission is imminent. What they jointly establish is a directional fitness signal that the 2024-era virological profile of B3.13 did not contain.

The EID paper also found that 66% of exposed dairy farm workers carried pre-existing neutralising antibodies against pdm09 H1N1 (the pandemic H1N1 strain from 2009, now endemic globally as a seasonal influenza). ² The researchers suggest this cross-immunity may explain why the CDC human case counter has stood at 71 since 6 March, even as the national dairy cattle total crossed 1,047 cases in 17 states. That explanation carries a corollary: workers WHO missed 2009 pandemic vaccination or WHO received only attenuated protection carry reduced cover if human spread accelerates.

The ISG suppression mechanism adds a clinical-detection wrinkle. ISG induction produces the fever and cytokine symptoms that typically prompt dairy workers to seek testing; a strain that delays ISG activation could circulate longer before self-reporting, compressing the window between exposure and detection. Bangladesh and Cambodia logged four human H5N1 cases by February, all from poultry contact , confirming the dairy-cattle route to human exposure remains less documented. The defensive side has one concrete advance: CEPI's (Coalition for Epidemic Preparedness Innovations) mRNA H5N1 vaccine candidate entered Phase 3 trial on 22 April , the furthest any such candidate has reached.

Deep Analysis

In plain English

**H5N1** is a bird flu virus that has been spreading through US dairy cattle since 2024. Scientists were worried about whether it could infect humans more easily, and this new study answers part of that question. Researchers at the **National Institute of Allergy and Infectious Diseases** grew human nasal cells in a lab and exposed them to the specific strain called **B3.13** that lives in US dairy herds. They found it copied itself more efficiently in those cells than older bird flu strains did, and it also partially switched off the cells' first alarm signal (the immune response that normally slows a virus down). The reassuring part: about two-thirds of dairy farm workers already have leftover immune protection from the 2009 swine flu pandemic. Scientists think this is probably why human cases have stayed rare even though the virus has been improving at getting into human nose cells. The worry is that this protection will not last forever, especially as the virus keeps spreading across more herds.

Deep Analysis

Root Causes

Three structural conditions combine to produce this finding. First, B3.13 evolved in the **mammary epithelium** of dairy cattle, a tissue environment that shares sialic acid receptor distribution with human nasal epithelium, creating inadvertent selective pressure for human nasal-compatible variants without any direct human infection required.

Second, the ISG suppression pathway B3.13 uses is the same interferon-evasion strategy seen in multiple pandemic-capable strains. The **NIAID/NIH** team's finding that B3.13 suppresses ISG responses below historical H5N1 levels is not a random mutation; it reflects cumulative selection across two years of continuous dairy-herd circulation.

Third, the surveillance architecture that might catch an early human adaptation event was partially dismantled on 1 May 2026 when **USDA APHIS** ended mandatory interstate testing, removing the movement-checkpoint layer that historically detected infections before they crossed state lines.

What could happen next?

Risk
If B3.13 acquires mutations that allow it to partially evade pdm09 antibodies, the 66% workforce-immunity buffer collapses and occupational exposure converts from rare to frequent.
Medium term · 0.65
Consequence
CDC and USDA face regulatory pressure to revise dairy-worker PPE guidance to include N95 respirators for milking operations, following the Emory aerosol evidence (ID:3109) and now nasal-fitness confirmation.
Short term · 0.75
Precedent
The EID paper establishes nasal-epithelium culture assays as the standard B3.13 fitness benchmark, meaning future variant comparisons will be assessed against this baseline.
Long term · 0.8

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: The 2009 pdm09 H1N1 emergence is structurally the more instructive case than the 1957 H2N2 shift. pdm09 was spreading undetected in humans for weeks before surveillance caught it, and prior partial immunity in older cohorts moderated case fatality relative to naive-population projections.

B3.13 sits in an analogous position: nasal fitness confirmed, ISG suppression confirmed, and a pre-existing partial-immunity buffer actively masking the true attack rate among the most exposed occupational group.

Counter-parallel: The 1997 Hong Kong H5N1 emergence had a higher initial nasal fitness signal but was contained by mass poultry culling before sustained human-to-human transmission emerged. That containment route is unavailable in the US dairy context: B3.13 is embedded in a 1,047-farm-herd system across 17 states, with no feasible mass culling option.

The dairy industry carries the immediate exposure. USDA compensates producers for depopulation at approximately $2,000-$4,000 per cow; a further expansion across the 17 affected states would accelerate indemnity claims that have already reached tens of millions of dollars since March 2024.

The structural dimension concerns the US raw milk market and export trust. B3.13 has been confirmed in raw milk from infected cows; any regulatory tightening on raw milk sales following publication of the nasal-fitness findings would affect approximately 3% of total US milk volume but carry outsized reputational cost to a politically protected sector.

International dairy importers from markets applying precautionary H5N1 standards, including several EU member states, will monitor the EID paper for guidance on import conditions.

Consensus view: Trevor Bedford (Bedford Lab, Fred Hutch) has consistently argued that mammalian-adaptive mutations in nasal-epithelium fitness matter more than raw case counts as a pandemic-risk signal. The Flagg and Winski's EID paper supports that framing: B3.13 now shows both improved nasal replication and partial ISG suppression, two of the three classical prerequisites for pandemic emergence, with only efficient human-to-human transmission remaining undemonstrated.

Counter-view: Florian Krammer (Mount Sinai) and colleagues have argued that pre-existing cross-reactive immunity in working-age populations, specifically the 66% pdm09 antibody prevalence among dairy workers documented in this study, creates a structural buffer that nasal fitness data alone cannot override. On this view, the fitness findings update the risk model but do not imply imminent pandemic transition without evidence of immune evasion against pdm09 antibodies.

Key tension: Whether the 66% pdm09 antibody prevalence represents a durable population-level brake or a temporary buffer that erodes as younger, less-exposed cohorts enter dairy work.

First Reported In

Update #2 · B3.13 gets better at humans as testing ends

CDC Emerging Infectious Diseases· 12 May 2026

Read original →

Causes and effects

Caused by

Emory aerosol study reframes dairy PPE

Emory PLOS Biology aerosol study established B3.13 reaches the nasal epithelium via aerosol; the EID May 2026 paper showed what happens once it arrives.

Occurred 1 May 2026

Read story →

This Event

B3.13 replicates better in human nasal tissue

The first mechanistic characterisation of B3.13 in human nasal tissue establishes a fitness trajectory that surveillance infrastructure is no longer fully equipped to track.

Different Perspectives

World Health Organization

WHO's DON606 recalibration to confirmed-only reporting gives the clean baseline the CDC model rests on, but the apparent fall from 1,040 to 534 carries misinterpretation risk WHO communications have not pre-empted. The PABS deadlock ahead of IGWG7 and continuing MBP134/remdesivir assessment without authorisation make WHO the body most able to accelerate the two decisions that could change the outbreak's trajectory.

European Union (ECDC)

ECDC's Week 23 CDTR tracked four simultaneous non-Ebola signals: the Dermatophilus congolensis novel-transmission cluster across France, Germany and Spain; a 4.2-fold malaria surge in Mayotte; the Salmonella ST2045 multi-country cluster; and two new Saudi MERS cases. The continental early-warning layer is carrying a full multi-pathogen picture while Bundibugyo dominates headlines.

Uganda

Uganda's 19 confirmed cases are concentrated in Kampala and Wakiso, an urban cluster that applied the 2022 Sudan-ebolavirus playbook; the Bwera border laboratory shortens cross-border confirmation to same-day. Uganda's regulatory authority must co-sign before MBP134 or remdesivir can dose any patient.

Democratic Republic of the Congo

Kinshasa shares Bundibugyo sequence data in real time with no treaty-guaranteed access to the vaccines that data informs, and its health minister called the US entry ban discriminatory while negotiating an early lift. DRC accounts for 515 of 534 confirmed cases and faces IS-controlled access blockades in Mambasa that health authorities cannot resolve.

United States (HHS/CDC)

HHS expanded the Ebola entry ban to green-card holders on 5 June, widening a restriction expiring around 17 June against WHO advice. The CDC simultaneously published the R0=2.51 modelling, the sharpest analytical contribution to the response, from a federal bench that holds the NIH and acting CDC director roles in one person.

Imperial College London / Cori and Ferguson

Anne Cori and Neil Ferguson place the case-fatality ratio at 30 to 40 per cent and read the 6.8-to-1 suspected-to-confirmed ratio as evidence that the laboratory figure understates true lethality. Many people die before a swab reaches them.