Our findings reveal trends of antibody reactivity against H5Nx HPAI and LPAI H5 among wild birds migrating through Alaska and Iceland, two previously described incursion points for virus spread into North America from both Asia and Europe [4, 8, 12, 17]. In this seroepidemiological study, we tested 1528 individual birds for antibody reactivity against H5Nx HPAI and H5 LPAI viruses, which revealed variability in seroprevalence by year, higher rates of exposure to H5 LPAI than H5Nx HPAI overall, and significantly more seropositive and suggestive exposure of birds to H5Nx HPAI in Alaska as compared to Iceland. These data likely reflect greater trans-Pacific movement of HPAI from East Asia, where clade 2.3.4.4 has been endemically circulating for the last several decades, to Alaska via avian migration [5, 8, 9, 11, 21, 36]. Additionally, our data demonstrate extensive evidence for exposure of wild birds to both H5Nx HPAI and H5 LPAI in both Alaska and Iceland, as demonstrated by dual antibody reactivity to LPAIpsv and HPAIpsv in HI assays. H5Nx HPAI antibodies were more frequently documented in sera from birds of the Order Anseriformes and among younger birds (1CY) among samples collected from Alaska. While we documented more exposure to H5Nx HPAI by year and across years among birds in and migrating through Alaska, our findings support that a significantly greater proportion of immuno-naïve birds to both H5Nx HPAI and H5 LPAI viruses were found during sampling in Iceland. These data demonstrate that a greater proportion of birds with exposure to H5Nx HPAI viruses migrate within and through Alaska, which may imply higher risk of H5Nx HPAI incursion to the rest of North America via the Pacific route. Additionally, given the high proportion of immunologically naïve birds to H5Nx HPAI viruses migrating within and through Iceland, there may be increased population-level susceptibility to future virus introduction events, increasing risks of outbreaks and onward spread to proximal neighboring regions.
The increase in seroprevalence documented in Alaska in 2014–2015 correlates with substantial outbreaks of clade 2.3.4.4 in Asia and North America in those same years [37]. Though these viruses were eradicated from domestic poultry in North America by June 2015, evidence suggests low-level maintenance in wild bird populations continued in North American birds through late 2016 [36, 38]. We detected no serological evidence of wild bird exposure to H5Nx HPAI in 2016, which may be due to a low sample size for that year, the low level of persistence in the population, or a combination of these factors. In Iceland, we detected increased trends of seropositivity for H5Nx HPAI viruses in 2015–2017 as compared to previous years. Our previous study revealed the most significant route of IAV introduction to Iceland is through the westward migration of wild birds originating from mainland Europe, therefore seroprevalence during 2015–2017 most likely corresponds to the widespread outbreaks of H5Nx HPAI viruses in wild and domestic birds in Eurasia during those same years [4, 21, 37]. The low level of seroprevalence during sampling years in Iceland compared to Alaska may be due to elevated migratory connectivity of Asia and North America, species diversity and abundance in Alaska (nearly 30% greater species abundance) permitting greater host distribution and availability for transmission, or factors yet to be distilled [5, 39–43].
Our findings demonstrate spatiotemporal and species differences in antibody titers to H5Nx HPAI and H5 LPAI viruses, including dual-reactivity, among wild birds in Alaska and Iceland. We documented serum samples with dual reactivity to HPAIpsv and LPAIpsv in HI assays, providing some evidence that aquatic birds may routinely be infected with both LPAI and HPAI H5 subtype viruses and mount antibody responses to each as a result. We have high assurance of this conclusion based on the following reasons. First, for our positive control and to detect the level of cross-reactivity between LPAIpsv and HPAIpsv, we used Emperor Goose serum that had been previously tested by another study for antibody reactivity to H5Nx HPAI and H5 LPAI antigens. We detected no cross-reactivity between HPAIpsv and LPAIpsv when separately incubated with serum samples with known antibody reactivity to H5Nx HPAI and H5 LPAI antigens. Second, HPAI viruses of clade 2.3.4.4 are antigenically distant from other H5Nx virus clades and LPAI H5 subtype viruses, therefore cross-reactivity of H5 LPAI antibodies with H5Nx HPAI virus or the HPAIpsv used in our HI assays is unlikely [44]. Third, pseudoviruses, as constructed in this study, have high specificity given they only contain the target surface antigen of interest, limiting non-specific interactions with other surface proteins that can occur when using live virus in HI assays [45]. The high proportion of dually exposed birds to both H5 LPAI and H5Nx HPAI viruses based on equal, onefold difference, and twofold difference antibody titers, also compels inquiry into the degree and effects of homo- and hetero-subtypic immunity in Arctic wild bird populations. Previous research has documented the protective effect of prior infection with homologous and heterologous IAV subtypes on subsequent duration of infection [46]. Additionally, previous infection with LPAI viruses has been shown experimentally to reduce H5Nx HPAI viral shedding [47, 48]. Data demonstrate that phylogenetic distances between HA sequences are inversely associated with immune protection in duck species [46]. Given the continued spread of H5Nx HPAI viruses in wild birds globally, and that prior exposure to LPAI may influence immunity and/or disease outcomes, future applications of novel serological assays such as those employed in this study, may help to provide additional insight into the role of population immunity in modulation of outbreak dynamics.
The majority of birds tested were found to be seronegative to both H5Nx HPAI and H5 LPAI viruses in Alaska and Iceland, which suggests that a large proportion of wild birds migrating through these regions are immunologically susceptible to H5 subtype virus infections. Though exposure to other H5 and non-H5 subtype IAVs may confer some cross-protection for H5 subtype viruses, this has not been systematically studied in nature to date. We found a greater than fivefold higher seroprevalence (12.4%) of H5Nx HPAI in birds migrating through Alaska compared to birds migrating through Iceland (2.3%), and the odds of immunological naivety (birds with no antibody reactivity to either HPAIpsv or LPAIpsv) among sampled birds was significantly higher in spring months. This accurately reflects and adds to current knowledge about the seasonality of IAV infections in wild birds in both regions. Generally, a decline in prevalence of IAVs are detected in the spring, when many species of wild birds migrate northward to breeding ranges in the Arctic [49, 50]. The summer breeding season coincides with increased prevalence of IAVs due to the high abundance of birds on breeding ranges within and proximal to the circumpolar Arctic, the introduction of immune-naïve juveniles to the population, and the environmental persistence and abiotic transmission of IAVs from the previous breeding season [49, 51]. The annual transmission cycle, therefore, begins in summer with infected adult and juvenile birds migrating southward in autumn, wintering at southern latitudes in winter, and starting the cycle again in spring. Data vary regarding the duration of antibody persistence following IAV infection in avian hosts from six months to much longer [52, 53], and have been shown to differ by species and sex [52]. Though antibody duration reactive to H5Nx HPAI viruses has not been systematically measured across species, it is possible that if reactivity to H5Nx HPAI viruses lasts less than a year, this would explain the decreased levels of pre-existing immunity among wild birds in spring. Future work to measure the immunity dynamics and the maintenance cycle will increase the reliability of models estimating infection rates and our general understanding of how immunity shapes outbreak dynamics of HPAI viruses [19, 54].
Our findings also quantified a significantly higher proportion of seronegative birds migrating through Iceland to H5Nx HPAI and H5 LPAI viruses, which demonstrates that birds in Iceland generally had low underlying immunity to H5 IAVs during our sampling timeframe. It has been suggested that the novel introduction of H5N1 to North America in 2021 originated in mainland Europe before migration across the Atlantic through Iceland [2, 20]. Our findings present one possibility of an ecological explanation for this never-before-documented event: low-levels of immunity to H5 subtype viruses among long-distance migratory birds in Iceland may have provided a suitable ecological niche for the intercontinental transport of H5N1 viruses across the Atlantic. Though our serosurveillance in Iceland ended in 2017, the massive geographic extent of the epizootic that resulted from the latter introduction event in 2021 to present has likely impacted serostatus trends among wild birds throughout the North Atlantic. Our findings suggest that continued serological surveillance in Iceland and other regions in the North Atlantic that connect migratory flyways between Europe and North America will provide useful information for inferring how population immunity may influence viral introduction events through the North Atlantic route.
Serosurveillance is a useful method for characterizing infection risk in wildlife populations, particularly in remote regions of the globe that are not easily accessible for routine virus surveillance. Data can be used to demonstrate virus exposure in a population, which is frequently underestimated by viral sampling, and provides estimates of population-level immunity against highly consequential zoonotic pathogens, as we demonstrate in this study for H5Nx HPAI viruses. Additionally, serosurveillance provides important data for susceptible-infected-recovered (SIR) and other population-level disease models of wild reservoir hosts, especially to evaluate the impact of HPAI introduction events on mortality, shedding, and enzootic persistence among avian populations. The development of pseudoviruses to study H5Nx HPAI and H5 LPAI virus exposure among wild birds in Alaska and Iceland enabled a streamlined pipeline for our two-stage screening and antibody quantification process on a large number of sera samples. Pseudoviruses are easily scalable, genetically stable, and an inherently safer alternative to using native highly pathogenic viruses for serological assays. In regions where highly biosecure laboratories (ex. BSL-3 and higher) are not available, pseudoviruses pose considerable promise for studying novel emerging viruses in wildlife, domestic animals, and humans [30].
Our seroprevalence findings differ from a study from 2015 which found limited evidence of specific antibodies of HPAI clade 2.3.4.4 in the Pacific (including birds sampled in Alaska), Central/Mississippi, and Atlantic flyways in the United States [55]. This study found 11% (n = 8/75), 26.4% (n = 24/91), and 13.4% (n = 28/209) of Snow Geese, Mallards, and Northern Pintails, respectively, with antibodies reactive to North American lineage H5 LPAI and no antibody reactivity to Gs/GD lineage H5Nx HPAI viruses. While our findings demonstrated 26.7% and 12.0% seroprevalence of H5 LPAI and H5Nx HPAI exposure in Alaska in 2015, respectively, multiple factors may have influenced these differing outcomes. First, the latter study targeted Anseriformes only, whereas our study sampled a wider range of species within and outside of this avian taxonomic order. Second, the latter study only sampled birds in August of 2015, which may have missed the significantly increased proportion of seropositive birds detected by our study from autumn and winter months. Third, there are still many gaps in knowledge about interpretation of serological data from wildlife exposed to viruses in natural systems, including varying durations of detectable antibody responses between virus subtypes, the impact of previous IAV infections on future serologic responses, and the potential for non-specific reactions which may affect HI or other serological assay results.
Limitations
We acknowledge limitations to the findings presented in this study. First, differences in sample sizes, species sampled, and locations by year, particularly in Alaska, could have introduced sampling bias. Species composition in each dataset may have reflected the specific avian taxonomic focus of each distinct sampling effort. For instance, during 2012–2014, only mallards were sampled, and in 2016 and 2018 only Glaucous-winged Gulls were sampled in Alaska. Though data in this paper are aggregated by taxonomic order (Anseriformes and Charadriiformes) to account for variation in sampling effort across species, more systematic and uniform sampling by species across years may provide additional insights into species susceptibility to HPAI. Similarly, in Alaska during two sampling years, 2010 and 2015, far more samples were collected from a greater diversity of species as compared to other years. These two years also comprised the highest seroprevalence rates of any of the ten sampling years. Years with fewer sample counts and less species diversity may have contributed to an underestimation of the true seroprevalence during those time points. Second, our use of a North American lineage H5 LPAI HA gene segment to construct LPAIpsv may have underestimated the true seroprevalence of H5 LPAI in wild birds in Alaska and Iceland. Birds in Alaska and Iceland may be exposed to Eurasian lineage H5 LPAI viruses with some frequency. A high proportion of Emperor Geese in Alaska, for example, spend part of the annual cycle in Eurasia where they may be exposed to diverse Eurasian lineage IAVs [23, 56]. Third, our data are reported as seropositive and suggestive seropositive for H5Nx HPAI and H5 LPAI viruses, and most of our analyses focused on seropositive results (twofold or more dilution differences between antibody titers for H5Nx HPAI and H5 LPAI). Though previous research studies have implemented the same standard twofold difference in antibody titer that we adopted to determine antibody presence between LPAI and HPAI H5 viruses [23, 57], we acknowledge that this determination may have introduced potentially arbitrary variability. We report onefold and twofold differences (suggestive and seropositive titers) to demonstrate the complexity of data interpretation related to categorizing true pathogen exposure among avian hosts.