Repeated oral exposure to H5N1 influenza virus in pasteurized milk does not cause adverse responses to subsequent influenza infection

Abstract

In March 2024, a highly pathogenic avian influenza H5N1 (HPAI) clade 2.3.4.4b virus was identified in US dairy cows, with spillover to cats, poultry, and humans. Up to 30% of commercial pasteurized milk tested contained viral genome copies. The impact of residual viral remnants on host immunity is unknown. Orally ingested proteins can stimulate gut-associated lymphoid tissues, potentially inducing tolerance and altering responses to later infection. We found that milk pasteurization fully inactivated pandemic H1N1 and bovine H5N1 influenza viruses yet preserved hemagglutinin (HA) protein integrity. In mice, repeated oral exposure to inactivated virus did not alter mortality after H5N1 virus challenge. Preliminary data showed that naïve mice exposed to improperly pasteurized milk containing live H5N1 virus developed lethal infection, whereas prior H1N1 infection conferred protection. Mice with preexisting H1N1 immunity remained protected when challenged with bovine H5N1 virus after exposure to H5N1 pasteurized in milk. These findings suggest that pasteurized milk containing inactivated H5N1 virus poses minimal health risks.

Fig. 1.. Milk stabilizes influenza virus HA protein during pasteurization.

(A) Pasteurization of virus at either high temperature (72°C for 15 s) or low temperature (65°C for 30 min). H1N1 CA/09 or recombinant HA protein as a positive control was ran on a 4 to 20% gel and stained for total protein using Coomassie blue. (B) Samples were diluted (1:50) of pasteurized milk in PBS, live H1N1 virus in PBS, or H1N1 virus pasteurized in either PBS or milk for 72°C for 15 s and were incubated with a combination of a monoclonal antibody (mAb) and polyclonal antibody (pAb) against the H1N1 HA. (C) Analysis of protein expression relative to milk in PBS control. (D) HA amino acid sequence comparison between CA/09 H1N1 and A/bovine/Ohio/2024 (H5N1) with identical sequences highlighted in blue and additional similar residues highlighted in yellow. (E) In silico analyses of the H1N1 or H5N1 HA amino acid sequences that are important for protein stabilization and melting temperature. Statistical analysis includes (C) one-way ANOVA with Tukey’s multiple comparisons. *P < 0.05; ****P < 0.0001; n.s., not significant. Data are shown as mean with SD.

Fig. 2.. Orally administered live but not inactivated H1N1 virus protects against H1N1 virus challenge.

(A) Graphical summary of the experimental design. Created in BioRender. Brigleb, P. (2025) https://BioRender.com/btxaiap. (B) Weight change during and following oral gavage of pasteurized milk, 10⁶ TCID₅₀ live CA/09 H1N1, or virus pasteurized in milk. (C) Sera were collected from mice 21 days post–oral gavage start and before challenge. Antibodies against CA/09 H1N1 virus were assessed by ELISA. LOD, limit of detection. (D to F) Mice were rechallenged with 5× mLD₅₀ (5 × 10^3.3 TCID₅₀) 21 days poststart of oral gavage. (D) Survival. (E) Cumulative clinical scores. (F) Weight loss shown as the percent of starting weight. n = 5 to 15 mice per group. Statistical analyses include one-way ANOVA with multiple comparisons test (C), log-rank Mantel-Cox test (D), and two-way ANOVA with Dunnett’s multiple comparisons test [(E) and (F)]. *P < 0.05; **P < 0.01; ****P < 0.0001. Data are shown as mean with SD and compared to the naïve control.

Fig. 3.. Repeated ingestion of inactivated H5N1 in milk does not worsen disease after viral challenge.

(A) Graphical summary of the experimental design. Created in BioRender. Brigleb, P. (2025) https://BioRender.com/22e2lg3. (B) Weight change during and following oral gavage of pasteurized milk or virus pasteurized in milk. (C) Sera were collected from mice 21 days post–oral gavage start and antibodies against whole bovine H5N1 virus were assessed by ELISA. (D to H) Mice were rechallenged with 10× mLD₅₀ 21 days poststart of oral gavage. (D) Survival. (E) Cumulative clinical scores. (F) Weight change. (G) Lung and (H) brain titers at days 6 to 7 postinfection. n = 9 to 14 mice per group. Statistical analyses include two-way ANOVA (B), log-rank Mantel-Cox test (C), two-way ANOVA with multiple comparisons [(E) and (F)], and a one-way ANOVA with Tukey’s multiple comparisons test [(G) and (H)]. ***P < 0.001; ****P < 0.0001. Data are shown as mean with SEM.

Fig. 4.. H1N1 immune history protects against H5N1 challenge and is not impaired by ingestion of inactivated virus in milk.

(A) Graphical summary of the experimental design. Created in BioRender. Brigleb, P. (2025) https://BioRender.com/1h5i5fn. (B and C) Sera were collected before oral gavage and HAI titers to H1N1 PR8, and whole-virus ELISA titers to bovine H5N1 were assessed, shown as mean with SE. (D) Weight change during and following oral gavage of pasteurized milk or virus pasteurized in milk. (E) Sera were collected pre- and postgavage and postchallenge, and whole-virus ELISA titers were assessed to bovine H5N1, shown as mean with SE. (F to H) Mice were rechallenged with 10× mLD₅₀ 21 days poststart of oral gavage. (F) Survival. (G) Cumulative clinical scores. (H) Weight change. n = 4 to 6 mice per group. Statistical analyses include two-way ANOVA (D), log-rank Mantel-Cox test compared to naïve control (F), and a two-way ANOVA with multiple comparisons [(G) and (H)]. **P < 0.01; ****P < 0.0001. [(D), (G), and (H)] Data are shown as mean with SD and compared to the naïve control.