Ocular infectivity and replication of a clade 2.3.4.4b A(H5N1) influenza virus associated with human conjunctivitis in a dairy farm worker in the USA: an in-vitro and ferret study.

backgroundThe human eye represents a potential site of influenza A virus (IAV) replication, and an entry point for the virus to reach the respiratory tract. The frequent detection of conjunctivitis among farm workers with confirmed infection with clade 2.3.4.4b A(H5N1) IAV from this ongoing outbreak represents an atypical disease presentation for this virus subtype. We aimed to investigate whether the occurrence of ocular complications reported following clade 2.3.4.4b A(H5N1) virus infection was associated with an enhanced capacity of this virus to replicate in mammalian ocular tissue and cause infection following ocular exposure.

methodsPrimary human nasal and corneal tissue constructs were infected with A(H5N1) A/Texas/37/2024 (Texas/37), A(H1N1)pdm09 A/Nebraska/14/2019 (Neb/14), and A(H7N7) A/Netherlands/219/2003 (NL/219) viruses (multiplicity of infection [MOI] of 0·01-0·02, 33°C). Corneal tissue constructs were also infected with an expanded panel of IAVs (Texas/37, A[H5N1] A/Michigan/90/2024 [MI/90], A[H5N1] A/Chile/25945/2023 [Chile/25945], NL/219, A/Netherlands/230/2003 [NL/230], and Neb/14; MOI of 0·01, 37°C). In-vitro infections of tissue constructs were used to assess replication kinetics by infectious virus titration. Induction of innate host antiviral responses in infected corneal tissue constructs was assessed by PCR array (MOI of 2·00, 37°C). Ferrets (serologically naive or pre-immune to A[H1N1]pdm09 virus) were inoculated by the ocular route with Texas/37 A(H5N1) virus-using a liquid inoculum (10⁶ plaque forming units [PFU]), aerosol inhalation (15-16 PFU), or ocular-only aerosol exposure (18-132 PFU)-to assess pathogenicity and tropism of the virus following different exposure routes. Transmissibility was assessed by placing serologically naive or pre-immune ferrets inoculated by ocular-only aerosol exposure in direct contact with serologically naive ferrets, monitoring pathogenicity in contact animals, and measuring viral titres in nasal washes of both inoculated and contact ferrets.

findingsNasal and corneal tissue constructs supported replication of all IAVs tested. In corneal tissue constructs, A(H7N7) and A(H1N1)pdm09 viruses reached 10-fold higher overall titres than A(H5N1) isolates. Relatively few genes (n=13) related to antiviral responses were significantly differentially expressed in corneal tissue constructs infected with IAV, with no consistent differential expression among clade 2.3.4.4b A(H5N1) viruses associated with either conjunctivitis or severe respiratory disease, although strain-specific differences were observed. Serologically naive ferrets inoculated by liquid ocular, aerosol inhalation, or aerosol-only ocular routes with Texas/37 virus exhibited a systemic and fatal infection in all animals, transmitting the virus to naive cagemates. By contrast, reduced disease severity following ocular-only aerosol inoculation was observed in animals with pre-existing heterosubtypic immunity. No serologically naive ferrets placed in direct contact with pre-immune ferrets inoculated with Texas/37 virus by the ocular-only aerosol route became infected.

interpretationA clade 2.3.4.4b A(H5N1) virus from the dairy cattle outbreak in the USA that was first detected in March, 2024, does not appear to possess features indicative of an ocular tropism. However, this virus can maintain a virulent and transmissible phenotype in ferrets following ocular exposure, highlighting the importance of ocular protection.

fundingUS Centers for Disease Control and Prevention.