Emerging H5N1 mutations (PB2-D701N, PB2-E627K, HA-S110N, NA-V116X) in a recent outbreak in U.S. dairy cows: need for continuous monitoring

Mutations are considered the most critical and fundamental factor in the evolution of viruses^[1,2^]. In the H5N1 influenza virus, mutations can significantly alter its behavior, particularly to increase virulence, enhance receptor interaction for entry into the host, and increase resistance to antibodies and drugs^[3,4^]. However, several mutations have been noted in polymerase basic 2 (PB2), hemagglutinin (HA), and neuraminidase (NA). The question is: which mutations are essential among several mutations? Previously, we employed a combination of generative artificial intelligence (GenAI) (large language model or multimodal large language model) and innovative bioinformatics techniques to identify key antibody escape mutations in the S-protein of SARS-CoV-2^[5^]. Therefore, AI may be a crucial option for understanding the critical mutations in the H5N1 influenza virus. However, during the report on AI, researchers stated that transparency in AI reporting is an essential component^[6^]. Nevertheless, we have selected four mutations from the literature. Future researchers may utilize AI to select mutations in the H5N1 influenza virus.

At the beginning of 2024, highly pathogenic avian influenza (HPAI) caused by the H5N1 virus, strain A, spread among several humans, raising deep concern for human health in the USA. It is now an emerging threat to human health^[4,7-9^]. HPAI strain A (H5N1) has caused outbreaks in domestic and wild birds worldwide on occasion^[10^]. In 1996, H5N1 was first identified in China in an infected goose. The infection was noted in Southern China, and the virus strain was identified and named A/goose/Guangdong/1/1996^[11^]. In 1997, HPAI strain A of the H5N1 virus emerged in Hong Kong. Here, 18 human infection cases were found, which includes six deaths. During that time, widespread poultry mortality was noted^[11,12^]. Fatal cases of human and poultry birds indicated the infection was virulent during that time.

On the other hand, over the past 20 years, sporadic infections in humans with the HPAI A (H5N1) virus have been reported from 23 countries, exhibiting a broad spectrum of clinical severity. In these cases, more than 50% of a cumulative case fatality was noted^[8,13^]. In the USA, the virus was spread among cows on dairy farms, which is concerning for animal health; the dairy business is now facing challenges^[4^]. It was noted that the virus infected more than 875 cows^[7^]. In 2014, the virus spread from animals to humans. Uyeki et al reported that the virus had started to spread among dairy farm workers. They indicated that the onset of the infection occurred in late March 2024 in an adult dairy farm worker with discomfort in the right eye and redness^[8^]. Since then, genome sequence analysis by the US CDC has reported 70 confirmed cases of human infection^[14^]. Therefore, there is an urgent need to understand whether sporadic cases of human infection by HPAI strain A H5N1 will be fatal.

The cows in Texas first detected the H5N1 virus, which belongs to clade 2.3.4.4b. The genetic analysis was performed in the USA using the HA of H5N1 influenza, indicating that the virus belongs to the H5 clade 2.3.4.4b (A/Texas/37/2024, Texas)^[8,15^]. Researchers have reported several mutations of the H5N1 virus occasionally^[15,16^]. Due to the mutations, scientists have reported that the virus is changing its pattern. Plaza et al reported two PB2 mutations in avian H5N1, specifically E627K and D701N. These two mutations are found in the current panzootic and previous waves. However, they have concluded that the virus has changed its pattern. Therefore, continuous monitoring of emerging mutations in bovine and H5N1 viruses must be maintained with a focus on clade 2.3.4.4b^[17^]. Several scientists reported mutations in the HA, PB2, and NA segments. Recently, we reported mutations in HA, PB2, and NA segments. We reported one significant mutation, E627K, from PB2, which is responsible for the enhanced virulence of the virus. Similarly, we reported several mutations in the HA segment, such as E186D, Q222H, Q226L, T199I, T192I, and S137A. These mutations enhanced receptor binding interaction. On the other hand, mutations in NA (S137A, H275Y, and I222V) are responsible for drug resistance (oseltamivir resistance). Furthermore, we also noted one mutation in M2 (S31N), which confers the amantadine resistance. Overall, it has been noted that PB2 mutations are responsible for increasing virulence, HA mutations are responsible for receptor binding interactions, and mutations in NA are responsible for drug resistance^[3^]. Similarly, Lin et al reported a Q226L (Gln226Leu) substitution in the HA protein. This mutation helps in the interaction of a single mutation with human-type receptor specificity. This mutation in the H5N1 HA helps switch the receptor, i.e., from a bovine H5N1-type receptor to a human-type receptor specificity^[15^]. Similarly, Good et al reported that one single mutation (T199I) of HA in H5N1 dairy cows increases the breadth of that area and, thereby, receptor interaction. The mutation is responsible for distinct binding specificity to α2,3-linked sialic acid, indicating receptor specificity^[16^].

Again, in this study, we noted the recent PB2, HA, and NA mutations from GISAID. GISAID is a global resource and one of the significant influenza virus databases providing access to genomic data^[18^]. All Influenza data retrieved from the GISAID database revealed four mutations in three segments (PB2, HA, and NA) in dairy cows (Table 1). Here, we identified two mutations in PB2: D701N and E627K. Similarly, we identified one mutation in HA (S110N) and one in NA (V116X) (Fig. 1). In PB2, the D701N mutation plays several roles. Pardo-Roa et al stated that the D701N mutation is a mammalian-adaptation mutation^[19^]. Similarly, He et al illustrated that the absence of E627K and D701N mutations in the PB2 protein results in a virus with low pathogenicity^[20^]. The E627K and D701N mutations illustrate that these two mutations facilitate mammalian adaptation and can serve as markers of mammalian adaptation^[21^]. Therefore, these two mutations enable the virus to spread rapidly. Previously, we also demonstrated that the E627K mutation in PB2 enhanced the virulence properties of the H5N1 A-type virus^[3^].

Table 1.

Recently identified mutations in dairy cows in the PB2, HA, and NA segments. Retrieved from the GISAID database^[18^]

Sl No.	Segment	Mutations	Year of detection	Bovine H5N1 strain
1	PB2	D701N	2025	A/dairy_cow/USA/004223-001/2025
2	PB2	E627K	2025	A/dairy_cow/USA/004631-001/2025
3	HA	S110N	2025	A/dairy cow/USA/002093-001/2025
4	NA	V116X	2024	A/dairy_cow/USA/037573-002/2024

Figure 1.

Position of different significant mutations of H1N5 in PB2, HA, and NA segments identified in recent outbreak in U.S. dairy cows. (A) Significant mutation in PB2 in dairy cows, (B) significant mutations in HA in dairy cows, and (C) significant mutations in NA in dairy cows.

The H5N1 virus is a single-stranded RNA virus that evolves more rapidly compared to DNA viruses. Therefore, it is evident that the virus is also evolving within the host (dairy cows) (Fig. 2). During the evolution process, mutations in PB2, HA, and NA help the virus become more adaptable to its hosts and potentially lead to a more rapid-spreading virus. Therefore, proper surveillance of the virus is essential to understand its changing properties with the mutations.

Figure 2.

Illustration of the molecular phylogenetics of PB2, HA, and NA segments of the HPAI H5N1 virus in dairy cows. (A) Molecular phylogenetics of PB2 segment of the HPAI H5N1 virus in dairy cows, (B) molecular phylogenetics of HA segment of the HPAI H5N1 virus in dairy cows, and (C) molecular phylogenetics of NA segment of the HPAI H5N1 virus in dairy cows.

The article helps us understand the details of learning points, such as the need to focus on continuous monitoring for these four mutations and antigenic epitope selection for next-generation vaccine development considering those mutations, etc.

New strategies, including vaccination, must be implemented to control the recent spread of the H5N1 HAPI virus clade 2.3.4.4b. However, several seasonal influenza vaccines are available from four manufacturers: AstraZeneca, Sanofi Pasteur, Seqirus, and GlaxoSmithKline. These four vaccine producers produce ten various formulations of seasonal trivalent influenza vaccines, including an intranasal formulation. These vaccine manufacturers also contain eight diverse quadrivalent formulations. However, as part of pandemic preparedness and planning, the USA has licensed three prototypes of Influenza A H5N1 vaccines (monovalent virus vaccines) through the Food and Drug Administration (FDA) approval^[22^]. Yet, there are no H5N1 vaccines with HA segments of the virus that belong to the clade 2.3.4.4b. In this direction, scientists are attempting to develop H5N1 clade 2.3.4.4b using HA and/or NA antigenic epitopes. Recently, Li et al developed a recombinant nanoparticle vaccine using H5N1 clade 2.3.4.4b HA and/or NA-derived antigens^[23^]. The FDA must provide emergency authorization for those vaccines against the clade 2.3.4.4b.

Conversely, no oral H5N1 vaccine is available for mass administration to wildlife. Therefore, it is essential to prioritize oral vaccine development urgently. Available influenza vaccines should be licensed for poultry use, and they can be utilized to reduce the disease burden. However, they may not prevent infection and have not controlled the disease burden. China controls H5 and H7 through a large-scale national vaccination program for poultry. Influenza vaccines may also be used in dairy cows. However, the USA, Europe, and Brazil hesitate to utilize influenza vaccines in cattle or poultry due to international restrictions on trade in products from vaccinated animals.

On the other hand, vaccinated poultry mustn’t be a hindrance to safe trade. Soon, several new, more effective, and safe vaccine platforms will be available for animal influenza vaccine development, which will help ensure vaccines for cattle or poultry are not hindered by safe trade^[24^]. At the same time, the virus must be monitored at all levels to control its spread.

Ethical approval

This article does not require any human/animal subjects to acquire such approval.

Consent

Not applicable. No patients/animals used for this study. As it is a correspondence article, no consent is required.

Sources of funding

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author contributions

Chiranjib Chakraborty: Conceptualization, Data curation, Investigation, Writing – original draft, Writing – review & editing. Arpita Das: Validation, formal analysis. Manojit Bhattacharya: Validation formal analysis. Md. Aminul Islam: Validation, formal analysis. All authors critically reviewed and approved the final version of the manuscript.

Conflicts of interest disclosure

All authors report no conflicts of interest relevant to this article.

Research registration unique identifying number (UIN)

Name of the registry: Not applicable. UniqueiIdentifying number or registration ID: Not applicable. Hyperlink to your specific registration (must be publicly accessible and will be checked): Not applicable.

Guarantor

Md. Aminul Islam, COVID-19 Diagnostic Lab, Department of Microbiology, Noakhali Science and Technology University, Noakhali-3814, Bangladesh E-mail: aminulmbg@gmail.com

Provenance and peer review

Not commissioned, internally peer-reviewed.

Data statement

The data in this correspondence article is not sensitive in nature and is accessible in the public domain. The data are therefore available and not of a confidential nature.

Declaration of whether any AI was used in the research and manuscript development

The authors declare no AI was used in the research and manuscript development.