Long-Term Epidemiology and Evolution of Swine Influenza Viruses, Vietnam

Influenza A viruses are a One Health threat because they can spill over between host populations, including among humans, swine, and birds. Surveillance of swine influenza virus in Hanoi, Vietnam, during 2013–2019 revealed gene pool enrichment from imported swine from Asia and North America and showed long-term maintenance, persistence, and reassortment of virus lineages. Genome sequencing showed continuous enrichment of H1 and H3 diversity through repeat introduction of human virus variants and swine influenza viruses endemic in other countries. In particular, the North American H1-δ1a strain, which has a triple-reassortant backbone that potentially results in increased human adaptation, emerged as a virus that could pose a zoonotic threat. Co-circulation of H1-δ1a viruses with other swine influenza virus genotypes raises concerns for both human and animal health.

through reassortment of classical swine, avian, and human influenza viruses in the 1990s (1,5), European avian-like H1N1 swIV introduced into pigs during the 1970s, and several human H1N1-and H3N2derived viruses, including pH1N1, pre-2009 H1N1, and H3N2-variant viruses (4). The increasing genetic diversity of IAVs in swine presents a pandemic concern (4,6). Human infections of emerging swIV are most frequently reported in the United States, where all nonhuman influenza viruses are nationally notifiable, with recurring outbreaks at agricultural fairs (7,8); however, sporadic human cases due to swIVs have been recently reported in Asia (9,10), Europe (11), and Australia (12).
In 2013, swIV surveillance was established at a collective slaughterhouse in Hanoi, Vietnam, that sourced pigs from 23 local provinces (13) (Table 1). During 2013 and 2014, cocirculation of multiple swIVs was detected, including viruses originating from pH1N1, H1N2 (with pre-2009 seasonal influenza-derived H1 hemagglutinin [HA]), H3N2 derived from human seasonal influenza, and TR H1N2 and H3N2 viruses (13). For this study, we extended virologic surveillance to 2019 and serologic surveillance from 2013 to 2019 in this slaughterhouse to characterize swIV evolution.

Surveillance and Virus Isolation
We examined 150 paired serum and nasal swab specimens collected from monthly convenience sampling at the Van Phuc slaughterhouse, the main collective swine slaughterhouse in Hanoi, Vietnam. We recorded the province of origin for each sampled pig. We report, in this study, findings from serum samples collected May 7, 2013-August 28, 2019. Virus isolation and sequence data from swab samples collected May 7, 2013-May 19, 2016, were reported previously (13); virus isolation and sequence data from swab samples collected June 30, 2016-August 28, 2019, are reported in this study. We cultured samples for virus isolation in MDCK cells at the National Institute of Veterinary Research Laboratory in Hanoi using described methods (13).

Phylogenetic Analysis and Genotypic Diversity
We aligned gene segments individually by using MAFFT version 7.490 (18) and constructed maximumlikelihood phylogenetic trees by using IQ-TREE version 2.1.4 on the basis of the best-fit nucleotide substitution model determined by ModelFinder (IQ-TREE, http://www.iqtree.org), estimating branch supports by using an approximate likelihood ratio (SH-like) test (19). We determined genotypes of swIVs by assigning each segment to specific lineages on the basis of maximum-likelihood phylogenies and characterizing the genotype based on the clade distribution of its internal segments (20). We inferred time-scaled phylogenies calibrated by sample collection dates by using the maximum-likelihood method in IQ-TREE version 2.1.4, implementing a least-square dating algorithm (21). We visualized trees in FigTree version 1.4.4. (http://tree.bio.ed.ac.uk/software/figtree).

Serologic Assays
We randomly selected 10 pig serum samples from each monthly sampling visit to the same slaughterhouse (13). In total, we tested 760 pig serum samples collected during May 2013-August 2019 by using the hemagglutination inhibition (HI) assay against 3 swine H1 subtype influenza virus strains and 1 swine H3-subtype influenza virus strain isolated in this study, as well as pH1N1 virus. We grew the viruses in MDCK cells and used them as antigens. The HI tests were carried out according to the World Health Organization's standard protocol for animal influenza diagnosis (22), with details as previously described (13).

Incidence of swIVs and Subtype Diversity
No clear pattern of seasonality was detected from the 150 paired swab and serum specimens collected during June 2016-August 2019, nor from previous   (Table 1; Figure 1, panel C). There were 114 influenza-positive nasal swab samples with 1 mixed-subtype infection among the 5,850 swabs tested, yielding 115 viruses for study, for an isolation rate of 1.95%. Subtyping of the 115 swIV isolates showed the circulation of 3 virus subtypes, H1N1 (n = 16), H1N2 (n = 46), and H3N2 (n = 53); 1 H1N2 and H3N2 co-infection was detected. Our previous study in the same abattoir during 2013-2014 yielded the same 3 virus subtypes, H1N1 (n = 16), H1N2 (n = 35), and H3N2 (n = 26) (13). Of note, we observed a high genetic similarity of viruses collected during each sampling at the Van Phuc slaughterhouse, suggesting that viruses were not maintained in the abattoir but were repeatedly introduced from sourced populations.
In each year, we observed the cocirculation of multiple genotypes; few genotypes were maintained across multiple years ( Figure 3 We found that pH1N1 viruses with PB2, PB1, nucleoprotein (NP), and nonstructural genes from TR viruses and N2-NA genes from human H3N2-derived swIV (genotypes 2, 3, and 13) were detected sporadically (Figure 3). The NA and internal genes of swine pH1N1 viruses were not monophyletic but interspersed between human-origin virus sequences. All potential spillover events were closely related to human viruses, not limited to human viruses in Vietnam.
In contrast to the predominance of a single pH1N1 virus genotype, the H1N2 TR viruses repeatedly acquired internal genes from pH1N1 viruses (genotypes 5, 6, 8, 9, and 14); the most recent H1N2 TR viruses contained all internal genes of pH1N1 viruses. Two H1N2 swIV sequences (genotype 16) contain Eurasian avian-like surface proteins and pH1N1 internal genes. Similarly, the genotype 8 H1N2 viruses, with H1-δ HAs and all other segments from TR viruses, repeatedly gained pH1N1 genes to form genotypes 7, 12, 15, and 20. The most recent H1N2 TR viruses collected in 2019 contain TR N2 and pH1N1 internal segments (genotype 7). However, H1-δ genotype 4 viruses with no pH1N1 genes continued to be detected until 2019. Furthermore, H1-δ1a viruses (genotype 10) isolated in August 2019 contained North American TR internal and NA genes with only the matrix protein (M) gene from pH1N1. Those viruses have not been seen anywhere other than the United States and now Vietnam, suggesting that those swIVs likely spread to Vietnam directly from the United States.

Discussion
Our study, conducted during 2016-2019, provides insights into the evolution and epidemiology of swIV in Vietnam, a major pork-producing country in Asia. Through longitudinal surveillance at a central slaughterhouse in Hanoi, which sourced pigs from across the country, we found H1N1, H1N2, and H3N2 swIVs co-circulating. This finding is consistent with surveillance conducted in 2013-2014 (13,26) and with other studies in Vietnam (23,27,28), indicating that the swIV subtypes circulating in Vietnam are similar to those found across Asia and globally (31)(32)(33)(34).
We found that the genetic diversity of swIV in Vietnam since 2010 is attributable to the persistence of several swine-origin H1N2 and H3N2 viruses, first reported in other countries in Asia and North America and likely imported via swine trade. We also identified repeat introductions of human pH1N1 viruses through reverse zoonosis. We found extensive reassortment of major swIV lineages, including H1N2 and H3N2 frequently acquiring pH1N1 internal genes and only rare acquisition of other swIV internal genes by pH1N1 lineage viruses. As a result, most recent swIVs from Vietnam contain pH1N1 internal genes. However, 2 divergent H1-δ virus lineages (originally derived from pre-2009 seasonal H1N1 viruses), detected in our most recent sample from 2019, maintained the TR lineage NA and internal genes (genotypes 10, 12, and 20) (Figure 3, panel B).
Although repeated introductions of pH1N1 into swine is consistent with previous swine surveillance from Vietnam (13) and southern China (35,36), the frequency of detections of new human pH1N1 virus lineages in swine in Vietnam has tapered off since 2015. Although onward transmission of pH1N1 was not sustained for most introductions, 1 lineage, first detected in northern Vietnam during 2013-2014 (13), persisted in swine for >4 years and reassorted with other prevailing swIV lineages. When introduced as components of reassortants, pH1N1-origin gene segments tend to be maintained in pigs (27), whereas viruses whose genomes are entirely derived from pH1N1 tend not to be sustained in the pig population. This phenomenon has been observed globally and in southern China and Hong Kong surveillance studies, which showed that viruses from purely pH1N1 failed to sustain the pig population after each introduction (36). Similar evidence of pH1N1 being sustained within swine populations has been reported from the United States (37) and Australia (38).
Pre-2009 seasonal H1N1 influenza-derived swine viruses (classified under H1-δ) are potentially gaining predominance in swine in Vietnam. Of note, an H1N2 H1-δ1a virus lineage with limited cross-reactivity to other H1 swIV lineages was detected in 2019, and serosurveillance showed evidence of circulation since March 2016. The increasing prevalence of H1-δ-like lineages increases the risk for zoonotic transmission. Previous studies demonstrated significant antigenic distance between pH1N1-like viruses and H1-δ cluster viruses (H1-δ-like and H1-δ1a) (29,30), and current seasonal influenza vaccines do not elicit protection against H1-δ swIVs (39)(40)(41). It is therefore important to ascertain whether current diagnostics can distinguish pre-2009 H1 viruses from currently circulating human seasonal H1 strains.
H1-δ1a viruses likely entered Vietnam via imported swine, demonstrating the role of trade in the global dissemination of swIVs (42). East and Southeast Asia countries, including China, Thailand, and Vietnam, which produce >50% of pork globally, are hotspots for emerging infectious diseases from swine (43,44). Vietnam ranks second in Asia for pork production, producing 19.62 million heads in 2019 (45). In addition, pork consumption in Vietnam has risen rapidly, from 12.8 kg/head/year in 2001 to 31.4 kg/head/year in 2018 (46), mostly in the form of fresh pork. Vietnam has also been importing breeding hogs from the United States since 1996. On average, 553 breeding hogs were imported each year from the United States during the 2010s, with a peak in 2014-2015 (Appendix 2 Figure 2) (47). The H1-δ1a lineage may have been introduced to Vietnam by imported breeding hogs in early 2016, as was seen in mainland China in the early 1990s (48).
In each of the swIV lineages detected in Vietnam, we discovered evidence of localized circulation, as evidenced by periods of unsampled diversity and long phylogenetic branches, likely at the provincial level. Despite the limited mixing of swine populations in Vietnam before slaughter (44), the infrequent detection of persistent lineages in the central slaughterhouse suggests that the genetic diversity of swIV in Vietnam may be high. This diversity is likely driven by factors such as livestock density and turnover. In comparison, swine in the United States are exposed to a greater degree of mixing during their lifespan because they are transported across long distances for feeding and fattening, which results in replacement by advantageous swIV lineages (37).
Our findings indicate that the swIV gene pool in Vietnam is continually enriched by importations from North America and from other countries in Asia. Those viruses include a novel cluster of H1-δ1a (genotype 10) viruses, which may pose a zoonotic threat (49). As the H1-δ1a virus was found only during the last sample collection, its persistence is uncertain. Recurrent transmission of pH1N1 viruses from humans to swine and reassortment with other swIV lineages have increased genetic diversity. Hence, to limit further introductions and diversification of the swIV gene pool in Vietnam, it is important to actively monitor both local swine herds and imported swine. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 29, No. 7, July 2023 *Seropositivity defined as hemagglutination inhibition assay titer ≥40. HA, hemagglutinin; TR, triple reassortant. †A total of 30 serum samples were positive for both A/California/04/2009 (H1N1) and A/swine/Hanoi/11-260/2019 (H1N2). Because cross reactivity between these viruses is low, this result likely indicates sequential infections.