key: cord-0020494-vncwp4xs
authors: Fraiha, Ana Luiza Soares; Matos, Ana Carolina Diniz; Cunha, João Luis Reis; Santos, Beatriz Senra Álvares da Silva; Peixoto, Maria Vitória Chaves; Oliveira, Anna Gabriella Guimarães; Galinari, Grazielle Cossenzo Florentino; Nascimento, Heloíza Irtes de Jesus; Guedes, Maria Isabel Maldonado Coelho; Machado, Alexandre de Magalhães Vieira; Costa, Erica Azevedo; Lobato, Zélia Inês Portela
title: Swine influenza A virus subtypes circulating in Brazilian commercial pig herds from 2012 to 2019
date: 2021-08-28
journal: Braz J Microbiol
DOI: 10.1007/s42770-021-00550-y
sha: d2dd540d53ccc1e03b53723409664a8a60e9d01c
doc_id: 20494
cord_uid: vncwp4xs

The swine influenza A virus (SIAV) subtypes/lineages H1N1pdm09, H3N2, H1N2, and H1N1 of seasonal human origin are widespread in Brazilian swine herds. A monovalent inactivated H1N1pdm09 vaccine was licensed in Brazil in 2014. However, there are concerns about its efficacy due to the limited vaccine cross-protection against heterologous viruses and the potential for exacerbated reactions against vaccine strains. Thus, monitoring SIAVs subtypes/lineages that are circulating in the Brazilian swine population is important, by applying a fast and efficient diagnostic test in herd field samples. A RT-PCR assay was developed, using primers specific for HA subtyping of Brazilian SIAV, and was used to evaluate the occurrence of subtypes from samples collected between 2012 and 2019. From 167 field samples positive for influenza A, 117 were subtyped by nested RT-PCR assay. A higher occurrence of H1N1pdm was observed from 2012 to 2015, H3N2 in 2017, and H1hu in 2017 to 2019. A hemagglutination inhibition test was performed in serum samples received from 2017 to 2019, confirming these data. The molecular data highlights the importance of H1hu and H3N2 detection since there are no vaccines available for the subtypes/lineages and raises an alert of H1hu for its potential to infect humans. Serological data suggest a cyclical profile of occurrence between the H3N2 and H1N1pdm over time. Monitoring SIAVs circulating in Brazilian swine herds is necessary, which provides the relevant information for field veterinarians to apply effective control measures on the properties. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s42770-021-00550-y.

Among the respiratory diseases that affect swine production, infection with swine influenza A virus (SIAV) is one of the most striking since it may cause explosive outbreaks of acute respiratory disease, affecting almost 100% of the herd, and its clinical signs can be exacerbated by coinfections with bacteria and other viruses [1] . The disease leads to a reduction in feed intake and weight loss, causing substantial economic losses to the pork industry. Moreover, swine influenza is a major concern to humans due to the risk that these viruses can cross the species barrier and be transmitted to humans [2, 3] .

The subtypes of SIAV that infects pigs worldwide belong to the species Influenza A virus, genus Alphainfluenzavirus, family Orthomyxoviridae, and its genome is composed of eight negative-polarity single-stranded segments that encode at least Responsible Editor: Fernando R. Spilki 10 viral proteins. Due to the segmented genotype and the lack of proofreading by the RNA polymerase during virus replication, reassortments and mutations are common, contributing to influenza A virus (IAV) variability. The hemagglutinin (HA) and neuraminidase (NA) glycoproteins are the main targets of the host immune response and are highly subject to mutations, contributing to virus evolution and evasion of the host immune response [4] .

Worldwide, different subtypes of SIAV circulate in the swine population: H1N1, H1N2, H3N2. In Brazil, SIAV has been a major threat to the pork industry since the emergence of the 2009 pandemic influenza virus in humans, when several outbreaks associated with the H1N1pdm09 virus were described in swine farms [5] . Since then, this virus seems to have become endemic in Brazilian pig herds, and its emergence favors reassortments with other circulating SIAVs, producing multiple new strains and increasing its genetic diversity [3] . Currently, the subtypes/lineages that are present and widespread in Brazil include H3N2, H1N2, and H1N1 seasonal human origin and the 2009 H1N1 pandemic virus [6, 7] .

In 2014, a monovalent inactivated vaccine encoding the hemagglutinin protein of the 2009 pandemic H1N1 lineage was licensed in Brazil [8] , and so far, it is the only one commercially vaccine available for use in Brazil, in addition to autogenous vaccines that are also allowed for use only to the target herd or adjacent properties [9] . Although there are no studies about the coverage of vaccine use in Brazilian herds, there are concerns about the use of the monovalent commercial vaccine, whether it provides sufficient protection to herds since studies have indicated that the vaccine elicited poor cross-protection against heterologous viruses [8, 9] and there are different strains circulating in herds, reducing the usefulness of vaccination for Brazilian herds.

Because of the challenges involving viral genetic variability, such as coinfection, multiple strains circulating in pigs, and poor vaccine cross-protection, the monitoring of viral subtypes/ lineages circulating in the swine population is very relevant to animal health. The application of rapid and efficient diagnostic tests in affected herds allows for the monitoring of viral subtypes in the country and provides the necessary information, allowing veterinarians to undertake efficient control measures on farms. Efficient control measures by vaccination can only be performed by knowing the HA subtypes/lineages that are circulating in the swineherds, such as H1pdm09, H1hu (seasonal human origin), and H3. Moreover, since the majority of available swine influenza vaccines are inactivated, including the only one commercially available in Brazil, knowing the circulating SIAVs in Brazilian herds is essential. This information is important, as vaccine protection is due mainly to humoral immune response and it is deeply impacted by antigenic mismatches among circulating strains and vaccine formulations.

The aim of this work was to evaluate the occurrence of different subtypes/lineages of the swine influenza virus circulating in Brazilian herds, by the analysis of clinical samples collected between 2012 and 2019, using a nested RT-PCR assay developed and validated in this study. Serological analysis was also performed in serum samples collected from 2017 to 2019.

Nested RT-PCR for SIAV detection of HA gene subtypes/lineages

To increase the sensitivity of the nested RT-PCR assay for detection of different SIAV subtypes and lineages for Brazilian field samples, primers were designed for the partial detection of the HA gene in the H1pdm09, H3, and H1hu sequences. The sequences of the HA glycoprotein gene from Brazilian swine influenza viruses were selected from the GenBank Database (Supplementary Table 1 ) and aligned by ClustalW multiple sequence alignment [10] . The primers were designed using Primer-BLAST software. The sequences were aligned separately by HA subtype (H1 or H3), and conserved regions of H1 and H3 were used to design the external primers. Using the selected regions, internal conserved regions of H1pdm, H1hu, and H3 were chosen to design the primers for nested RT-PCR. Information about the primers used for PCRs is shown in Table 1 , as well as the primer positions, amplified base pair fragments, and respective melting temperatures.

The H1N1pdm09, H3N2, and H1N1hu viruses used as positive controls for nested RT-PCR assays were previously characterized by RT-PCR and sequencing [11, 12] (Table 2) . Following the guidelines of the OIE Manual of Diagnostic Tests and Vaccines for Terrestrial Animals [13], the strains were isolated by sequenced passages in Madin-Darby canine kidney (MDCK) cell culture from lung fragments of animals that presented respiratory clinical signs on Brazilian pig farms.

Viral RNA was extracted using TRIzol® reagent (Thermo Fisher, USA) according to the manufacturer's instructions. RNA was stored at − 80°C.

For the first reaction, the viral genome was reverse transcribed and amplified using a Superscript One-Step RT-PCR System with Platinum Taq DNA polymerase (Life Technologies, Carlsbad, CA, USA). All samples were subjected to a standardized PCR containing 1× PCR buffer, 0.4 μM of each primer, 200 μM of each dNTP, varying concentrations of MgSO 4 (1.5 mM and 2 mM), 1.5 U of Taq DNA polymerase, and 2.5 μL of the extracted RNA, for a final volume of 25 μL. The cycling conditions included an initial reverse transcription at 48°C or 50°C for 30 min, followed by a denaturation step at 95°C for 2 min and 40 cycles of denaturation at 95°C for 30 s, an annealing step varying from 50 to 57°C, and an extension step at 64°C for 1 min, followed by a final extension step at 68°C for 5 min.

In the second reaction, 1 μL of an amplicon was amplified in a 25 μL reaction containing 1× PCR buffer, 0.4 μM of each primer, varying concentrations of MgSO 4 (1.5 mM and 2 mM), 200 μM of each dNTP, 1 U of Taq DNA polymerase (Phoneutria Biotecnologia, Brazil), and ultrapure nucleasefree water to the final volume.

The nested reaction started with a denaturation temperature of 95°C for 5 min followed by 40 cycles of a denaturation step at 95°C for 30 s, an annealing step varying from 52 to 57°C for 30 s, and an extension step at 68°C for 1 min, followed by a final extension step at 68°C for 5 min. The PCR amplicons were identified by electrophoresis on a 1% agarose gel and stained with ethidium bromide.

The method of limiting dilution was used to evaluate the analytic sensitivity of the nested and one-step RT-PCR assays. The primer sets were tested for the detection of the three reference swine influenza viruses. The H1N1pdm09 and H3N2 isolates, with titers of 4 × 10 5 plaque-forming units (PFU)/mL, were serially diluted 10-fold (10 −1 up to 10 −14 ), and RNA was extracted from 300 μL of each dilution. It was not possible to titrate H1N1hu isolate because its RNA was already extracted. From the extracted RNA, a 10-fold serial dilution was performed to be used in the RT-PCR assay [14, 15] .

To test the specificity of the nested RT-PCR assay, all the primer sets from the second amplification reaction were tested for each reference sample. Cross-reactivity with DNA viruses and bacteria that infect swine, such as porcine circovirus type 2 and Mycoplasma hyopneumoniae were evaluated as well as for RNA virus such as Senecavirus A and rotavirus A, B, and C.

Samples received for the diagnosis of SIAV at the Laboratório de Pesquisa em Virologia Animal (LPVA), Department of Preventive Veterinary Medicine, School of Veterinary Medicine of the Universidade Federal de Minas Gerais (UFMG), were tested for the influenza A matrix gene by RT-PCR [5] . A total of 167 field samples (146 lung tissues and 21 nasal swabs), positive for the influenza A matrix gene, were then analyzed for SIAV differentiation by nested RT-PCR. These samples were collected from animals presenting respiratory clinical signs from non-vaccinated commercial pig Table 2 ) in the years 2012-2013, 2014-2015, 2017-2018, and 2019. As these samples were received for a diagnosis, much information related to the age of the sampled animals and the exact location of the herd's properties were not sent by the field veterinarians that requested the exams. These data were then missed, which made it impossible for us to trace more details of the received samples, with the exception of the year of collection, city, and state of origin.

For NA gene (N1 and N2) subtyping, specific primers and PCR protocols were used according to Choi et al. [14] .

The resulted RT-PCR amplicons from the first reaction of positive H1pdm samples and from the second reaction of H3 and H1hu positive samples were purified using a DNA Clean and Concentrator kit (Zymo Research, USA) and subsequently dosed by a NanoDrop spectrophotometer (Thermo Fisher, USA). Samples with more than 40 ng were selected for sequencing by the BigDye Terminator v3.1 Cycle Sequencing kit (Thermo Fisher, USA). Sequencing reactions were performed by capillary electrophoresis on an ABI 3730 instrument (Applied Biosystems, USA). The nucleotide sequence data were analyzed and assembled using SegMan Software 7.1.0 (DNAStar Inc.). Consensus sequences were compared with sequences reported in GenBank (Supplementary  Table 1) , and those with an identity percentage higher than 95% using BLAST (Basic Local Alignment Search Tool) from the NCBI (National Center for Biotechnology) were selected. Phylogenetic analyses of the HA gene fragments of the Brazilian influenza virus samples were performed by Bayesian inference. For each influenza subtype/lineage H1pdm, H1hu, and H3, newly generated sequences were aligned with references obtained from the GISAID and GenBank databases (Supplementary Table 4 using MAFFT v7.471 [16] with global pair, adjustment direction, and 1000 iterations. The alignment was manually polished in AliViewV1.2.6 [17] (Larsson, 2014) to remove gaps that were present in less than 10% of the isolates and overhangs, resulting in final alignment sequences of 319, 195, and 156 nucleotides for H1pdm09, H1hu, and H3, respectively. The phylogeny was estimated using MrBayesV3.2.7a [18] , with 3 million MCMC replicates, 25% burnin, and the HKY + I nucleotide substitution model for H3N2 and GTR + I for H1N1 and H1hu, which was selected using JmodelTestV2.1.10 based on the AIC approach [19] (Posada, 2008) . The tree image was generated with the ITOL online tool [20] (Letunic and Bork, 2019).

Between 2017 and 2019, a total of 1631 serum samples from non-vaccinated pig farms located in different Brazilian states were received at the LPVA for diagnosis (Supplementary Table 3 ). In this period, 263 samples were obtained from 2017, 677 from 2018, and 691 from 2019. A hemagglutination test for detection of the viral titers in hemagglutinating units (UHA) was performed using the H1N1pdm09 and H3N2 reference viruses ( Table 2 ) according to the protocol described by WHO (2002) [21] . Due to the low hemagglutination titer of the H1N1hu isolated reference strain, it was not possible to use it in serological analyzes. Samples with an HI titer ≥ 40 were considered positive.

The concentration of MgCl 2 for the one-step RT-PCR and second round of PCR was 2 mM, which minimized the occurrence of nonspecific bands. The annealing temperatures that minimized the occurrence of nonspecific bands were 50°C or 52°C (for H1 and H3, respectively) for one-step RT-PCR and 52°C, 52°C, and 55°C (for H1N1pdm09, H3N2, and H1hu, respectively) for nested RT-PCR. Therefore, the cycling conditions for the first RT-PCR were an initial reverse transcription at 48°C or 50°C (for H1 and H3, respectively) for 30 min, followed by a denaturation step at 95°C for 2 min and then 40 cycles of denaturation at 95°C for 30 s, an annealing step at 50°C or 52°C (for H1 and H3, respectively) for 30 s, and an extension at 64°C for 1 min, followed by a final extension step at 68°C for 5 min. The second round of PCR started with a denaturation temperature of 95°C for 5 min followed by 40 cycles of a denaturation step at 95°C for 30 s, annealing at 52°C, 52°C, or 55°C (for H1N1pdm09, H3N2, and H1hu, respectively) for 30 s, and extension at 68°C for 1 min, followed by a final extension step at 68°C for 5 min.

To assess the sensitivity of nested RT-PCR, a 15-fold dilution for H1N1pdm09 and H3N2 and a 10-fold dilution for H1N1hu were performed. The sensitivity results of the assay are shown in Supplementary Figure 1 . The limit of detection of the nested RT-PCR for H1N1pdm09 was 10 −11 , a 10 8 higher than that of the one-step RT-PCR, which used the first set of primers. For H3N2, the limit of detection was 10 −9 , which was different from the first reaction, that was only 10 −5 . The detection limit for H1N1hu was lower, corresponding to 10 −4 , near the limit of the first reaction, which was 10 −3 .

To determine the specificity of the assay, the three SIAV reference samples (H1N1pdm09, H1N1hu, and H3N2) were tested with all three sets of primers used in the nested RT-PCR. All three subtypes/lineages yielded the expected base pair fragments, which were 159 bp, 303 bp, and 279 bp for H1pdm09, H1hu, and H3, respectively. It should be noted that when the H1N1pdm09 sample was tested with the H1hu primers, a band near 600 bp was detected representing amplification of the fragment related to the first RT-PCR (616 bp for H1), but there was no amplification of the 303-bp band, which is specific for H1hu (Supplementary Figure 2) . When crosstested against nonspecific primers for each viral subtype/lineage, none of the samples yielded any amplification product (Supplementary Figure 2) . Cross-reactivity with other swine RNA and DNA swine pathogens was not observed confirming the specificity of the test

In addition, to confirm the specificity of the nested RT-PCR, the amplicons were sequenced and analyzed by molecular phylogeny. The partial genome sequence of field samples indicated the specific subtype/lineage in accordance with the nested RT-PCR results.

Of the 167 positive samples for the influenza A matrix gene, the detection of the HA subtype/lineage was done in 117 (117/ 167-70%) by nested RT-PCR. Fifty-five samples were positive for H1N1pdm (55/117-47%), 29 for H1hu (29/117-24.8%), and 10 for H3N2 (10/117-8.5%) ( Table 3) .

Mixed infections with two subtypes/lineages (H1N1pdm + H1hu; H1hu + H3N2; H1N1pdm09 + H3N2) were 10/117-8.5%, 6/117-5.1%, and 4/117-3.4%, respectively, and those with all subtypes/lineages (H1N1pdm + H1hu + H3N2) were 3/117-2.6%.

Analyzing showed that 77.1% of the samples were positive for H1N1pdm, and only 2.9% and 11.4% were positive for H3N2 and H1hu, respectively. In these years, coinfections were observed only between H1N1pdm and H1hu, corresponding to 8.6% of the samples.

Analyzing the occurrence of different SIAVs in 2017-2018, the percentage of samples positive for H1N1pdm was 22.7%, which was lower than that in the other years. However, an increase in H3N2-positive samples to 22.7% (5/22) was observed, equaling the percentage obtained for H1N1pdm. Samples positive for H1hu were found at a higher percentage than H1N1pdm and H3N2, corresponding to 27.2%. Coinfections between the three subtypes/lineages were detected in 13.6% of the samples. In addition, coinfections between the pairs H1N1pdm + H3N2, H1N1pdm + H1hu, and H3N2 + H1hu were observed at 4.6% for each combination. In 2019, the occurrence of H1hu was the highest (34.8%), followed by H1N1pdm (30.4%). Coinfections between H1N1pdm + H1hu also increased, with an occurrence of 13%. A significant decrease in H3N2 in single infections (8.7%) and coinfections of H3N2 + H1N1pdm (2.2%) occurred, although infections with H1hu increased (10.9%).

The relative occurrence of H1N1pdm, H3N2, and H1hu by year, considering the total number of performed subtyping analyses and not considering coinfections, is shown in Figure 1 . In 2012-2013, the most detected lineage was H1N1pdm (68.8%-11/16), followed by H1hu (18.8%-3/16) and H3N2 (12.5%-2/16). In 2014-2015, H1N1pdm remained the most detected (78.9%-30/38), followed by H1hu (18.4%-7/38) and H3N2 (2.6%-1/38).

However, in 2017-2018, there was a balance among the assessed influenza subtypes/lineages. Indeed, an increase in the occurrences of the H3N2 and H1hu was observed, with the H3N2 and H1N1pdm showing equivalence of 32.3% (10/ 31) and H1hu demonstrating 35.5% (11/31), the highest occurrence among the analyzed samples. Increased occurrences of H1N1pdm and H1hu were also observed in 2019 compared to 2017-2018, with 36.2% (21/58) and 46.6% (27/58) in the subtypes sampled, respectively. In 2019, the occurrence of H1hu remained higher than that of H1N1pdm09, while the detection of H3N2 was 17.2% (10/58).

Of the 1631 serum samples analyzed by the HI test, 1220 (75%) were positive for IAV. In the sera collected in 2017, HI analysis detected the highest occurrence of H1N1pdm + H3N2 coinfections, 42.6% (112/263), followed by specific antibodies for single infections of H3N2 and H1N1pdm in 32.7% (86/263) and 10.6% (28/263) of serum samples, respectively. In 2018, serum samples that were only positive for H3N2 (39.9%-270/677) were also more prevalent than those that were only positive for H1N1pdm (3%-23/677). A large occurrence of H1N1pdm + H3N2 (27.8%-188/677) was observed at that time. In 2019, in sharp contrast to our previous finding in 2017 and 2018, we detected the highest occurrence of specific antibodies for H1N1pdm (43.8%-303/691), followed by coinfections at 27.1% (187/691) and a low occurrence of specific antibodies to H3N2 (3.3%-23/691) ( Figure 2 ).

Serological data from received samples suggest that there is a cyclical profile of occurrence between the H3N2 and H1N1pdm in Brazilian swine herds each year and that H1N1pdm is not always the most prevalent subtype/lineage.

Thirty-nine amplicons were sequenced: twenty-two amplicons from H1pdm, ten from H3, and seven from H1hu. The nucleotide sequences of the HA nucleotide sequences were deposited in the GISAID database, and information regarding the subtype/ lineage, state, year of collection, and GISAID accession numbers is listed in Supplementary Table 5 . As different genetic regions were amplified between subtypes/lineages, phylogenetic analyses were performed separately. The use of short sequences could reduce phylogenetic reliability and that the use of larger sequences, as the complete HA region, could be more informative. However, there are still some relevant differences in the evaluated fragments to allow potential phylogenetic estimations by robust methods (Supplementary Figure 3) . Hence, we used Bayesian inference, one of the best methods to estimate phylogeny, to try to obtain some reliable phylogenetic signal that could be lost in other simpler approaches, as Neighbor Joining and Maximum Parsimony. For this reason, there were several branches that were strongly supported in the threes, with close to 100 posterior probability values. These branches could be at least considered, even with the short sequences. The sequence variability and phylogenetic analysis of the HA partial gene can be seen in Supplementary Figures 4, 5 and 6 . 

Monitoring circulating SIAV subtypes/lineages is necessary and essential for epidemiological surveillance, helping veterinarians to establish prevention and treatment measures in affected farms, as well as for vaccine development. Rapid, sensitive, and specific diagnostic tests are important to monitor circulating viruses. In Brazil, no commercial tests are available to subtype SIAV. Some studies used next-generation sequencing (NGS) techniques to obtain a complete sequence of the viral genome from field samples [5, 7] , which is very efficient in obtaining genetic information but is not a feasible large-scale strategy since it is laborious, time-consuming, and expensive. Some studies using multiplex real-time RT-PCR (RT-qPCR) for the detection of H3 and different H1 lineages circulating in Europe, subtyped between 89 and 94% of SIAV-positive clinical samples [15, 22, 23] . Another study using a one-step multiplex RT-qPCR assay for the identification of different subtypes and lineages of SIAV circulating in Brazil demonstrated high sensitivity and specificity when applied for the diagnostic of IAV in clinical samples [24] . However, real-time PCR requires a greater investment and qualified technicians, which could be limitations for diagnostic labs and customers. High sensitivity in subtyping multiplex RT-PCR was demonstrated mostly when applied for SIAV isolates [25, 26] . Other studies have developed multiplex RT-PCR tests using primer sets that distinguish H1 and H3 but do not cover the differentiation of H1 lineages [14, [27] [28] [29] . In our study, we highlighted a highsensitivity subtyping diagnostic technique in clinical samples.

Using the developed nested RT-PCR and applying it to clinical samples, it was possible to release results in less than 24 h to field veterinarians, making viral isolation to increase the viral load for posterior subtyping, which takes 2-3 weeks, unnecessary. Moreover, the developed test was standardized using primers specifically delineated from HA Brazilian swine influenza virus sequences, increasing its specificity and importance of use to subtype field samples for monitoring SIAV in Brazil. Furthermore, and with greater importance, using the test to analyze clinical samples collected in 2012-2013, 2014-2015, and 2017-2019, it was possible to detect the four SIAV subtypes/lineages: H1N1pdm09, H1hu (H1N1 and H1N2), and H3N2. Remarkably, our technique enabled us to demonstrate by nested PCR that between 2012-2013 and 2014-2015, a higher detection of H1N1pdm was observed, followed by H1hu, mixed infections between H1hu + H1N1pdm and H3N2 + H1N1pdm, and H3N2. The subtyping data corroborate with previously published serological studies, which have shown that H1N1pdm is endemic in Brazil. A study performed in 2011 in the RS, PR, SC, SP, MG, and MS states showed that the H1N1pdm lineage was more prevalent in animal herds, followed by H1N2 and H3N2, and a high prevalence of coinfections between the lineages with the involvement of H1N1pdm was also observed [7] . In 2012, another study performed in swine farms in the state of MG indicated that 26.2% of the tested animals had antibodies against H1N1pdm and only 1,57% had antibodies against H3N2 [6] . In 2014-2015, a serological study performed by our group in unvaccinated farms from several Brazilian states showed that 96% of the herds were positive for H1N1pdm and 68% were positive for H3N2, of which 36% presented coinfections with H1N1pdm [30] . With the serological studies already published, the PCR subtyping data corroborate the higher occurrence of H1N1pdm in the 2012-2015 period. Coinfections between different SIAVs were also observed, calling attention to the higher probability of the occurrence of reassortments between different subtypes/lineages. The higher detection of H1hu, especially in the PCR subtyping study, reinforces that this human origin lineage circulates in Brazilian swine herds and seems to have settled in this population [3, 31, 32] . 14 Compared to previous years, in 2017-2018, serological and PCR subtyping data from clinical samples revealed a significant increase in the occurrence of H3N2, as well as in H1N1pdm + H3N2, and a decrease in H1N1pdm. The results of molecular analysis corroborate with data previously published demonstrating a higher detection of H3N2 in clinical samples in this period [24] . The increasing occurrence of H3N2 in swine may be related to the occurrence of this IAV subtype in the human population. According to human serological data published by the WHO (2018) [33] in 2017, a greater and significant occurrence of H3N2 was observed in the Brazilian human population, while the occurrence of H1N1pdm was not expressive, mostly because of the introduction of a new H1N1pdm vaccine strain into the human vaccine formulation. Data from swine and human IAV show a parallel occurrence of the virus subtypes in both species and monitoring IAV circulating in the human population is a tool that may help forecast possible outbreaks of influenza in swine. The increase of H3N2 occurrence in pigs in 2018, evidenced mainly by serological data of samples collected from non-vaccinated farms, may be related to the lower contact of pigs with H1N1pdm since this lineage circulated with a low prevalence in the human population. It is worth mentioning that transmission of influenza from humans to pigs, mainly of the H1N1pdm, has been shown to be more frequent and constant than the reverse route [34] , and biosafety measures are important to prevent this transmission.

In contrast, in 2019, the H1hu lineage was the most prevalent when considering single inflections or coinfections in nested PCR analysis, followed by H1N1pdm, which had the highest incidence in serology. This result suggests a cyclic profile of the influenza virus and the importance of human seasonal subtype/lineage that are constantly circulating in swine herds. There are no H1hu and H3N2 vaccines available in Brazil. Vaccination of swineherds, with H1N1pdm09 vaccines, which are constantly challenged with H1hu and H3N2, in addition to not offering efficient cross-protection, may lead to vaccine-associated enhanced respiratory disease (VAERD) [31] , which may occur when vaccinated pigs are challenged with heterologous viruses. Piglets vaccinated with delta-1 cluster H1N2 (H1hu) and challenged with H1N1pdm09 virus have demonstrated enhanced clinical signs of the disease and higher lung lesion scores than non-vaccinated/challenged animals [35] . Furthermore, the high occurrence of H1hu in the swine population raises an alert for its potential to cause massive infection in humans that are susceptible to the lineage. To date, in Brazil, only occasional cases of H1N1hu and H1N2hu have been reported [31, 32, 36, 37] , and those have been found in people who worked and had contact with sick pigs. Nonetheless, our data show that the circulation of influenza viruses of variable strains and subtypes/lineages poses an unneglectable risk both to pig farmers and the general population. Therefore, strict SIAV surveillance and the development of safer, more efficient vaccines should be developed to solve this problem.

This study had some limitations related to the sampling. As all samples of the study were received in our laboratory for the diagnosis of respiratory diseases, we could not determine the prevalence of SIAV subtypes/lineages in Brazilian herds, but we were able to estimate their occurrence by demonstrating a profile of which subtypes/ lineages were circulating in the herds in the analyzed period. Due to missed data about the samples, we were unable to analyze the occurrence of the strains by age or category as well as their frequency of occurrence in the herds, in both serological and molecular analyses. The sequencing of the subtyped samples by the RT-PCR was able to confirm the specificity of the standardized molecular test; however, we could not robustly discuss the phylogenetic analysis due to the small size of the sequences In summary, it was possible to subtype and differentiate SIAVs that are circulating in Brazilian swine herds by the RT-PCR developed in the present study, without the need for viral isolation, demonstrating the importance of monitoring viruses in the field samples and contributing to epidemiological surveillance. The data presented in this study suggest that the occurrence of different subtypes/lineages, demonstrated by both RNA detection and serology, varies over time within an annual cyclical period. Surveillance of influenza virus in humans and pigs using genetic characterization is an important auxiliary tool to prevent possible influenza outbreaks in both species and to establish updates for vaccines. To obtain more consistent data and accurate analyses, further studies analyzing the complete HA gene of SIAVs circulating in Brazil from different years must be performed. 

Ethics approval Not applicable.

Consent to participate Not applicable.

Competing interests The authors declare no competing interests.

Coinfections and their molecular consequences in the porcine respiratory tract

Review of influenza A virus in swine worldwide: a call for increased surveillance and research

Influenza A viruses of human origin in swine

Evolution of influenza A virus by mutation and re-assortment

Isolation and characterization of a pandemic H1N1 influenza virus in pigs in Brazil

Distribution of antibodies against influenza virus in pigs from farrow-to-finish farms in Minas Gerais state

Influenza A virus infection in Brazilian swine herds following the introduction of pandemic 2009 H1N1

Swine influenza virus vaccines -to change or not to change -That's the question

Vaccine development for protecting swine against influenza virus

CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice

Genetic characterization of influenza virus circulating in Brazilian pigs during 2009 and 2010 reveals a high prevalence of the pandemic H1N1 subtype

Genetic characterization of influenza virus isolated from commercial herds pigs in southern and southeastern Brazil in 2014. In: XXVI Congresso de Virologia, 11 a 14 de Outubro, Costão do Santinho, Florianópolis, SC 13. OIE. World Organization for Animal Health. Manual of diagnostic tests and vaccines for terrestrial animals

Detection and subtyping of swine influenza H1N1, H1N2 and H3N2 viruses in clinical samples using two multiplex RT-PCR assays

Molecular subtyping of European swine influenza viruses and scaling to high-throughput analysis

MAFFT Multiple Sequence Alignment Software Version 7: improvements in performance and usability

AliView: a fast and lightweight alignment viewer and editor for large datasets

MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space

jModelTest: phylogenetic model averaging

Interactive Tree Of Life (iTOL) v4: recent updates and new developments

Manual on animal influenza diagnosis and surveillance

Rapid detection and subtyping of European swine influenza viruses in porcine clinical samples by haemagglutinin-and neuraminidase-specific tetra-and triplex real-time RT-PCRs

Subtyping of swine influenza viruses using a high-throughput real-time PCR platform

Evaluation of two multiplex RT-PCR assays for detection and subtype differentiation of Brazilian swine influenza viruses

One-step multiplex RT-qPCR for the detection and subtyping of influenza A virus in swine in Brazil

Multiplex RT-PCR assay for differentiating European swine influenza virus subtypes H1N1, H1N2 and H3N2

Real-time reverse transcription-polymerase chain reaction assays for the detection and differentiation of North American swine influenza viruses

One-step multiplex RT-PCR for detection and subtyping of swine influenza H1, H3, N1, N2 viruses in clinical samples using a dual priming oligonucleotide (DPO) system

Detection and subtyping of swine influenza viruses in clinical samples by the mean of developed multiplex polymerase chain reaction assays

Estudos sorológicos dos subtipos do vírus influenza a suíno circulantes nas regiões sul e sudeste do brasil no período de

A human-like H1N2 influenza virus detected during an outbreak of acute respiratory disease in swine in Brazil

Whole-genome characterization of a novel human influenza A(H1N2) virus variant

Reassortant between swine H3N2 and 2009 pandemic H1N1 in the United States resulted in influenza A viruses with diverse genetic constellations with variable virulence in pigs

Kinetics of lung lesion development and pro-inflamatory cytokine response in pigs with vaccine-associated enhanced respiratry disease induced by challenge with pandemic (2009) A/H1N1 influenza vírus

Novel human-like h1n1 swine influenza virus with the six internal genes derived from h1n1/2009 virus

Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Acknowledgements We are thankful to our colleagues from the Laboratório de Pesquisa em Virologia Animal (Escola de Veterinária-UFMG) for their excellent technical support.