Influenza A(H5N1) shedding in air corresponds to transmissibility in mammals

Abstract

An increase in spillover events of highly pathogenic avian influenza A(H5N1) viruses to mammals suggests selection of viruses that transmit well in mammals. Here we use air-sampling devices to continuously sample infectious influenza viruses expelled by experimentally infected ferrets. The resulting quantitative virus shedding kinetics data resembled ferret-to-ferret transmission studies and indicated that the absence of transmission observed for earlier A(H5N1) viruses was due to a lack of infectious virus shedding in the air, rather than the absence of necessary mammalian adaptation mutations. Whereas infectious human A(H1N1_pdm) virus was efficiently shed in the air, infectious 2005 zoonotic and 2024 bovine A(H5N1) viruses were not detected in the air. By contrast, shedding of infectious virus was observed for 1 out of 4 ferrets infected with a 2022 European polecat A(H5N1) virus and a 2024 A(H5N1) virus isolated from a dairy farm worker.

Fig. 1. Infectious virus in air samples and ferret nose swabs.

a–f, Infectious virus particles per air sample: A(H1N1_pdm) (a), A(H5N1_Indo/WT) (b), A(H5N1_Indo/AT) (c), A(H5N1_polecat) (d), A(H5N1_Texas) (e) and A(H5N1_bovine) (f). g–l, Infectious virus titres in ferret nose swabs: A(H1N1_pdm) (g), A(H5N1_Indo/WT) (h), A(H5N1_Indo/AT) (i), A(H5N1_polecat) (j), A(H5N1_Texas) (k) and A(H5N1_bovine) (l). Each differently coloured bar represents a single ferret (four ferrets per virus). The dotted horizontal lines represent the detection limit. The symbols above bars indicate the day of peak virus shedding in the air for A(H5N1)-inoculated animals. *This animal was euthanized at 72 hpi because of reaching predetermined humane endpoints; thus, no nose swabs were collected after the 72 hpi timepoint. Source data

Extended Data Fig. 1. Experimental air-sampling set-up.

For the collection of air-samples, the BioSpot-VIVASTM Series 315 bioaerosol sampler (Aerosol Devices, USA) was directly connected to a ferret cage holding an individually housed ferret, via a 1.4 m long tubing. The BioSpot was operated with a flowrate of 15 L min-1 under the following conditions: conditioner: 5 °C; initiator: 45 °C; moderator: 18 °C; nozzle: 27 °C; sample holder: between 13-16 °C. No extra airflow was supplied to the ferret cage, in addition to the 15L min-1 applied by the BioSpot sampler. With a cage volume of 45 L (30 ×30 x 50 cm) and a flow rate of 900L h-1 (60 x 15L min-1), this results in 20 (900L h-1 / 45 L) air changes per hour. The ferret cage and BioSpot were housed in separate, interconnected, negatively pressurized class 3 isolators. This separation was necessary because the BioSpot generated excessive heat, which led to an increased ambient temperature within the isolator.

Extended Data Fig. 2. Infectious virus particles per liter air.

A) A(H1N1_pdm), B) A(H5N1_Indo/WT), C) A(H5N1_Indo/AT), D) A(H5N1_polecat), E) A(H5N1_Texas), F) A(H5N1_bovine). Each differently colored bar represents an individual ferret (four ferrets per virus). *This animal was euthanized at 72 hpi because of reaching predetermined humane endpoints, thus no air-samples were collected after the 72 hpi timepoint. Source data

Extended Data Fig. 3. Viral RNA in air-samples.

A) A(H1N1_pdm), B) A(H5N1_Indo/WT), C) A(H5N1_Indo/AT), D) A(H5N1_polecat), E) A(H5N1_Texas), F) A(H5N1_bovine). Each symbol represents an individual ferret (four ferrets per virus) using the same colors as in Fig. 1. RNA was detected by qRT-PCR. *This animal was euthanized at 72 hpi because of predetermined humane endpoints, thus no air-samples were collected after the 72 hpi timepoint. Source data

Extended Data Fig. 4. Comparison of viral RNA load in air-samples across different viruses.

Four biological replicates were performed and are derived from individual ferrets, each point in the box plots represents an individual ferret. The box plots contain the 5th to 95th percentile, the whiskers mark the minimum and the maximum, and the black line in the middle marks the mean of the dataset. Statistical significance was assessed using a two-way ANOVA with Tukey’s multiple comparison test. *p = 0.02; **p = 0.004; ***p = 0.0005; ****p = 0.00002. Source data

Extended Data Fig. 5. Infection virus titers in ferret throat swabs.

A) A(H1N1_pdm), B) A(H5N1_Indo/WT), C) A(H5N1_Indo/AT), D) A(H5N1_polecat), E) A(H5N1_Texas), F) A(H5N1_bovine). Each differently colored bar represents a single ferret (four ferrets per virus). Dotted horizontal lines represent the detection limit. *This animal was euthanized at 72 hpi because of predetermined humane endpoints, thus no throat swabs were collected after the 72 hpi timepoint. Source data

Extended Data Fig. 6. Correlation analyses of the presence of infectious virus or viral RNA in ferret nose and throat swabs and air-samples.

Correlation between number of infectious virus in air-sample and infectious virus titer in A) nose swabs (R² = 0,1256), and B) throat swabs (R² = 0,0001). Correlation between amount of viral RNA in air-sample and viral RNA in C) nose swabs (R² = 0.3789), and D) throat swabs (R2 = 0.0222). Each symbol represents one of the six viruses tested. Each differently colored symbol represents a time point post inoculation in hours (hpi). The statistical analysis performed is simple linear regression. Source data