Transmission of Human Influenza A Virus in Pigs Selects for Adaptive Mutations on the HA Gene

ABSTRACT

Influenza A viruses (FLUAV) cause respiratory diseases in many host species, including humans and pigs. The spillover of FLUAV between swine and humans has been a concern for both public health and the swine industry. With the emergence of the triple reassortant internal gene (TRIG) constellation, establishment of human-origin FLUAVs in pigs has become more common, leading to increased viral diversity. However, little is known about the adaptation processes that are needed for a human-origin FLUAV to transmit and become established in pigs. We generated a reassortant FLUAV (VIC11pTRIG) containing surface gene segments from a human FLUAV strain and internal gene segments from the 2009 pandemic and TRIG FLUAV lineages and demonstrated that it can replicate and transmit in pigs. Sequencing and variant analysis identified three mutants that emerged during replication in pigs, which were mapped near the receptor binding site of the hemagglutinin (HA). The variants replicated more efficiently in differentiated swine tracheal cells compared to the virus containing the wildtype human-origin HA, and one of them was present in all contact pigs. These results show that variants are selected quickly after replication of human-origin HA in pigs, leading to improved fitness in the swine host, likely contributing to transmission.

IMPORTANCE Influenza A viruses cause respiratory disease in several species, including humans and pigs. The bidirectional transmission of FLUAV between humans and pigs plays a significant role in the generation of novel viral strains, greatly impacting viral epidemiology. However, little is known about the evolutionary processes that allow human FLUAV to become established in pigs. In this study, we generated reassortant viruses containing human seasonal HA and neuraminidase (NA) on different constellations of internal genes and tested their ability to replicate and transmit in pigs. We demonstrated that a virus containing a common internal gene constellation currently found in U.S. swine was able to transmit efficiently via the respiratory route. We identified a specific amino acid substitution that was fixed in the respiratory contact pigs that was associated with improved replication in primary swine tracheal epithelial cells, suggesting it was crucial for the transmissibility of the human virus in pigs.

INTRODUCTION

Influenza A viruses (FLUAV) are enveloped viruses known to cause respiratory disease in several species, including humans and pigs. FLUAV has a wide host range and can adapt and cross between species (1). FLUAV circulates worldwide in pigs, with various strains and subtypes prevalent among swine populations. The major subtypes of FLUAV of swine include H1N1, H3N2, and H1N2, and considerable genetic and antigenic diversity exists within these subtypes (2).

Hemagglutinin (HA) is the major envelope glycoprotein of FLUAV and is a critical component in determining host specificity due to its role in initial binding and entry into the host cell (1, 3, 4). The HA binds via its receptor binding site (RBS) to sialic acids (SA) on the host cell in the conformation of either SA-α-2,3 linkage or SA-α-2,6 linkage. The SA-α-2,3 conformation is dominantly expressed in avian species, particularly in the respiratory and intestinal epithelial cells. SA-α-2,6 conformation is dominantly expressed in humans, usually in the upper respiratory tract (5). Thus, the host-specificity of the HA is affected by its ability to bind to these receptors (6). Amino acid substitutions in or near the RBS can change receptor-binding preference and facilitate transmission between host species (4). Specific amino acid substitutions have been determined to be important in defining the receptor binding profile and enabling species crossing, such as the 226 position from H3 and H9 FLUAVs (7, 8).

Pigs are a unique host for FLUAV as their respiratory epithelial cells are known to express both types of SA linkages, allowing opportunity for infection with both avian and human origin FLUAV strains (9–11), increasing the likelihood of reassortment to occur. One example was the generation of the triple-reassortant internal gene constellation (TRIG) that emerged in the late 90s (12). The TRIG cassette contains internal gene segments derived from swine (matrix, non-structural, and nucleoprotein), human (polymerase basic 1), and avian (polymerase acidic [PA] and PB2) FLUAVs forming a gene constellation that has remained well conserved. Previous studies have shown that the TRIG cassette is prone to stably incorporate HAs and NAs from various origins, potentially explaining the increased evolutionary rate of swine FLUAV and increased diversity after its introduction (13). The TRIG cassette also contributed to the generation of the pandemic H1N1 virus (H1N1pdm09) (14). Since then, genes from H1N1pdm09 have been incorporated into the backbone of swine FLUAVs in North America (15), and most of the swine strains circulating in the US now contain a combination of internal genes of the TRIG and H1N1pdm09 lineages, with most of the genetic diversity arising from mutations in the HA and NA (16).

Frequent interspecies transmission is known to occur between humans and pigs. Numerous introductions of human FLUAV to pigs were identified from 1990 to 2011, including seasonal H1 and H3 viruses and continual detection of H1N1pdm09 (2, 17, 18). These human-to-swine spillover events have resulted in the establishment of many novel virus lineages circulating in pigs globally, contributing to the great genetic and antigenic diversity of swine FLUAV (17, 19). One such lineage (2010.1 lineage) resulted from a recent spillover from a human H3N2 seasonal virus which became one of the predominant H3N2 lineages in US swine and continues to evolve since its first detection in 2012 (19, 20). Although FLUAV are frequently exchanged between humans and pigs, most human-origin viruses reassort with swine strains after the spillover and the persisting human-origin genes show significant changes compared to the human FLUAV ancestors, particularly the envelope genes. However, the molecular determinants enabling the adaptation of human-origin FLUAVs to pigs are still not clear.

Previous work has shown that wholly human H3N2 viruses do not typically replicate efficiently in pigs, and transmission is rarely observed (19, 21). Thus, to evaluate the molecular changes during replication and transmission of human-origin H3N2 FLUAV surface genes in pigs, we generated a reassortant strain via reverse genetics containing human seasonal HA and NA FLUAV genes in a common internal gene constellation currently found in U.S. swine FLUAV. We showed that this reassortant virus resulted in transmission in pigs and identified mutations in the HA gene segment of the resultant virus that were associated with improved transmission and replication in primary swine tracheal epithelial cells (pSTECs). Our results suggest that advantageous mutations are selected quickly in the HA gene of human seasonal FLUAV during replication in swine.

RESULTS

The reassortant VIC11pTRIG replicated in pigs.

Pigs (n = 10/virus) were challenged with A/Victoria/361/2011 (A/VIC/11), VIC11rg, VIC11p, VIC11pTRIG, and A/swine/Missouri/A01410819/2014 (sw/MO/14). Bronchoalveolar lavage fluid (BALF) was collected from directly infected pigs at 5 days postinfection (dpi), when the percentage of lung lesions was assessed (Fig. 1A and B). Confirming our previous results (19), viral titers were only detected in BALF from 2 pigs infected with the A/VIC/11 and VIC11rg wholly human seasonal viruses and only 1 pig infected with VIC11p. In contrast, 7 pigs from the VIC11pTRIG-infected group had detectable virus titers in the BALF, while virus was detected in all pigs from the group infected with the fully swine-adapted sw/MO/14 (Fig. 1A). Pigs infected with the sw/MO/14 showed significantly higher titers compared with all other groups. Although replication was confirmed in a few pigs infected with A/VIC/11, VIC11rg, or VIC11p, titers were low (below 10^2.5 TCID₅₀ [50% tissue culture infective dose]/mL), bringing the group average to below the limit of detection (Fig. 1A). Similarly, FLUAV-specific antigen staining was detected by immunohistochemistry (IHC) in sw/MO/14-challenged pigs, but signals were not observed in any of the other challenged pigs (Table 1).

FIG 1.

Viral titers in the lungs and macroscopic lung lesions of pigs directly infected with reassortant viruses. (A) Viral titers in bronchoalveolar lavage fluid (BALF) collected at 5 days postinfection (dpi) from pigs directly infected with A/VIC/11, VIC11rg, VIC11p, VIC11pTRIG, or sw/MO/14. Values are shown as mean 50% tissue culture infective dose (TCID₅₀)/mL titers ± standard error of the mean, with scattered dots representing individual pigs. (B) Percentage of the lungs affected with purple-red consolidation at 5 dpi. Numbers indicated above the error bar indicate the number of positive or affected pigs/total number of pigs in the group. Different lowercase letters indicate significant difference at P < 0.05. NC, negative control.

TABLE 1.

Lung and trachea microscopic pathology and virus detection in pigs challenged with reassortant viruses and negative controls^a

Virus	Scores
	Microscopic lung lesionsb	Lung IHCc	Microscopic trachea lesionsc	Trachea IHCd
NC	0.25 ± 0.16 A	0.0 ± 0.0 A	0.20 ± 0.20 A	0.0 ± 0.0 A
A/VIC/11	0.78 ± 0.19 A	0.0 ± 0.0 A	0.75 ± 0.22 A	0.0 ± 0.0 A
VIC11rg	0.60 ± 0.18 A	0.0 ± 0.0 A	0.20 ± 0.15 A	0.0 ± 0.0 A
VIC11p	0.58 ± 0.15 A	0.0 ± 0.0 A	0.05 ± 0.05 A	0.0 ± 0.0 A
VIC11pTRIG	0.48 ± 0.21 A	0.0 ± 0.0 A	0.65 ± 0.18 A	0.02 ± 0.02 A
sw/MO/14	9.43 ± 0.61 B	5.3 ± 0.4 B	2.13 ± 0.29 B	2.58 ± 0.23 B

^aResults are shown as means ± standard errors of the means. Different uppercase letters within the same column indicate significant difference at P < 0.05. NC, negative control; A/VIC/11, A/Victoria/361/2011; sw/MO/14, A/swine/Missouri/A01410819/2014; IHC, immunohistochemistry.

^bThe range of possible scores is 1 to 22.

^cThe range of possible scores is 1 to 8.

^dThe range of possible scores is 1 to 4.

As expected, sw/MO/14, which contained all gene segments well-adapted to swine, induced a high percentage of pneumonia in all pigs; significantly higher compared to all other groups, which contained different combinations of human-origin gene segments (Fig. 1B). The average percentage of macroscopic lesions was very low in the groups challenged with A/VIC/11, VIC11rg, VIC11p, or VIC11pTRIG, with no significant differences compared with the negative (non-infected) group. Similarly, pigs infected with sw/MO/14 showed the highest average microscopic lung and trachea lesion scores (Table 1), significantly higher than the other groups. Although some pigs infected with A/VIC/11 and VIC11pTRIG showed slightly higher trachea lesion scores compared to the other groups challenged with viruses containing the A/VIC/11 HA (Table 1), no significant differences were observed among these groups nor compared to the negative controls.

The reassortant VIC11pTRIG transmitted between pigs.

Nasal swab viral titers were assessed in directly inoculated pigs from 1 to 5 dpi and in respiratory contacts from 1 to 5, 7, and 9 days post-contact (dpc). Consistent with replication in the lungs, only a small number of pigs directly infected with the A/VIC/11 or VIC11rg shed low virus titers, with only 3 and 2 pigs, respectively, testing positive by virus isolation at any time point (Fig. 2A). Compared to the pigs infected with wholly human seasonal virus (A/VIC/11 or VIC11rg), more of the pigs directly infected with VIC11p shed virus, with almost all pigs (7 out of 10) being positive at least in one time point. However, no statistical difference was observed compared to the A/VIC/11 or VIC11rg groups. In contrast, virus was detected in the nasal swabs of all pigs infected with VIC11pTRIG and sw/MO/14 at all time points, and viral titers were significantly higher in comparison with the other three groups for most time points (Fig. 2A). Consistent with the little to no shedding observed in A/VIC/11-, VIC11rg-, and VIC11p-directly infected pigs, none of the respiratory contact pigs (n = 5/group) which were in the same room as these groups tested positive at any time point (Fig. 2B). In contrast, all pigs in contact with pigs infected with VIC11pTRIG or sw/MO/14 were positive in at least one time point, starting mostly at 4 dpc (Fig. 2B). Only one pig in the VIC11pTRIG contact group was positive at 3 dpc (data not shown). Consistent with these results, all pigs in these two groups were seropositive by hemagglutination inhibition (HI) assay at 15 dpc (Fig. 2C).

FIG 2.

Viral titers in nasal swabs of directly infected and respiratory contact pigs after infection with human-origin reassortant viruses and seroconversion of respiratory contact pigs. Nasal swabs samples were collected from (A) directly infected pigs daily from 1 to 5 days postinfection (dpi) and from (B) respiratory contact pigs daily from 1 to 5, then 7 and 9 days post-contact (dpc). (C) Antibody response in respiratory contact pigs at 15 dpc measured by hemagglutination inhibition (HI) assay. Groups were infected with A/VIC/11, VIC11rg, VIC11p, VIC11pTRIG, or sw/MO/14. Values are shown as mean TCID₅₀/mL titers or geometric mean titers ± standard error of the mean. Numbers indicated above the error bar depict the number of positive pigs/total number of pigs in the group. Dotted line indicates limit of detection. Different lowercase letters indicate significant difference at P < 0.05.

Minor variants are selected quickly after replication of the reassortant VIC11pTRIG in pigs.

Nasal swab samples from 5 directly inoculated pigs collected at 1, 3, and 5 dpi, and 5 respiratory contacts collected at 5, 7, and 9 dpc in the VIC11pTRIG group were selected for NGS. Samples from directly infected pigs with highest TCID₅₀/mL titers in any given time point (>10^2.75 TCID₅₀/mL) were selected for sequencing, and all respiratory contacts. Of the 30 NS samples selected for NGS sequencing, 9 samples from directly infected pigs (pigs 375, 378, 340 in all selected time points) and 8 samples from respiratory contact pigs (pigs 407, 408, 409, 410, at least two animals per time point) were amplified by multi-segment real-time PCR (MS-RT-PCR), and 17 produced complete genome assemblies by NGS. Viral genomes were analyzed for HA and NA variants. A total of 3 dominant variants were discovered in the HA: A138S, V186G, and F193Y. A138S emerged in two directly infected animals, becoming dominant (77% of sequences) in one animal by 5 dpi (Fig. 3A). A138S was fixed (99% of sequences) in all 4 contact pigs after transmission (Fig. 3B). The V186G substitution was present in 2.6% of sequences in one directly infected animal at 5 dpi and was not detectable in the contact pigs soon after transmission but became dominant in one animal on 9 dpc with a frequency of 60% (Fig. 3C and D). The F193Y substitution was only detected in one directly inoculated pig at a frequency of 56% by 3 dpi, declining on 5 dpi to 27% (Fig. 3E). F193Y was transmitted to one pig with a frequency of 9% at 5 dpc and with a frequency of 4% in another animal by 9 dpc (Fig. 3F). In addition to the major variants, several low-frequency variants were identified along the HA, in a total of 46 positions (Table 2). Two substitutions were detected in the PB2 of two respiratory contact pigs, I289V in pig 409 at 7 dpc and S641R in pig 410 at 9 dpc, and they were not associated with the two HA mutations detected sporadically (V186G and F193Y). No dominant variants were observed in the NA or other viral gene segments.

FIG 3.

Variants identified in directly infected and respiratory contact pigs in the group infected with the VIC11pTRIG reassortant. Dominant variants detected in each directly infected and respiratory contact pigs are shown. Variants with a frequency higher than 50% that were detected in nasal swabs in at least one occasion were considered dominant. (A and B) Substitution A138S. (C and D) Substitution V186G. (E and F) Substitution F193Y. All positions are based on H3 numbering.

TABLE 2.

Variants identified in directly infected and respiratory contact pigs infected with VIC11pTRIG^a

Characteristic	Nucleotide position
	89	199	256	286	296	489b	553	563	585	634b	640	643	655b	711	730	742
VIC11 HA	T	A	A	T	C	G	T	A	A	T	A	A	T	G	G	G
Variant sequence	C	G	G	C	T	T	A	G	G	G	G	G	A	T	A	A
Directly infected		1			2	5	1	1	1	2			2	2
Contact	1		1	1		8		1		2	1	1	2		1	1
	748	751	757	762	802	959	997	1025	1060	1109	1112	1144	1196	1244	1264	1271
VIC11 HA	G	A	C	A	T	C	G	G	C	G	A	A	A	C	A	A
Variant sequence	A	C	A	G	C	T	A	A	T	A	G	G	T	T	G	G
Directly infected			1	1	2	1	1	1		1			1	1	1	1
Contact	1	1							1		1	1
	1284	1286	1335	1412	1418	1449	1487	1536	1601	1648	1686	1692	1722	1748
VIC11 HA	G	A	C	C	G	A	C	G	G	G	T	T	T	A
Variant sequence	A	G	T	T	A	G	T	A	A	A	G	C	C	C
Directly infected	1	1	1			1	1	1	1	1	1		1
Contact				1	1							1		1

^aList of variants in nasal swab samples from directly infected and respiratory contact pigs in this study. VIC11 HA represents the HA sequence of A/Victoria/361/2011 H3N2, which is the same as VIC11pTRIG. Substitutions at the positions relative to the reference sequence that are present in ≥2% of variant viruses are shown. Numbers of samples containing each variant for either directly infected or respiratory contacts are listed.

^bPositions 489, 634, and 655 correspond to amino acid residues 138, 186, and 193, respectively.

We plotted the location of the 3 substitutions on the HA head and found all residues were located close to H3 antigenic and binding sites (Fig. 4). The A138S substitution is located under the 220 loop, which is one of the major structures forming the receptor binding site (RBS). The F193Y and V186G are located next to the 190 helix, another structure that forms the RBS.

FIG 4.

Location of substitutions identified on the Vic11pTRIG HA. Dominant variants identified in this study on the hemagglutinin (HA) of the Vic11pTRIG are highlighted on the structure of the human A/Victoria/361/2011 H3. Cyan, residue A138S; light green, residue V186G; orange, residue F193Y. The 190 helix, 130 loop, and 220 loop of the receptor binding site (RBS) are highlighted.

Amino acid substitutions A138S, V186G, and F193Y improve binding and replication of the human seasonal HA of VIC11pTRIG in swine cells.

Viral binding was evaluated by infecting pSTECs at a high multiplicity of infection (MOI) under restricting conditions for virus entry comparing the mutated HAs with the VIC11pTRIG wild-type (wt) HA (Fig. 5). ty/OH/04p was used as a control and contains the same backbone as the other four viruses. All variants showed increased binding to swine tracheal cells compared to VIC11pTRIG, reaching similar intensity and distribution of binding to that observed for ty/OH/04p (Fig. 5). Viral growth kinetics were evaluated in pSTECs and Madin-Darby canine kidney cells (MDCKs) to compare the replication of the A138S, V186G, and F193Y variants (VIC11pTRIG_A138S, VIC11pTRIG_V186G, and VIC11pTRIG_F193Y) to VIC11pTRIG (Fig. 6). No significant difference in replication kinetics was observed in MDCK cells for any of the viruses (Fig. 6A). The three variants showed more efficient replication in pSTECs at most time points (Fig. 6B) compared to VIC11pTRIG containing the human seasonal HA, with the highest difference at 72 h postinfection (hpi). Replication of VIC11pTRIG_A138S and VIC11pTRIG_V186G was similar to that of ty/OH/04p, while replication of VIC11pTRIG_F193Y was intermediary until 48 hpi, reaching similar titers as ty/OH/04p at 72 hpi.

FIG 5.

In vitro binding of mutants containing A138S, V186G, and F193Y substitutions in primary swine tracheal epithelial cells (pSTECs). Immunofluorescence microscopy of VIC11pTRIG reassortant, A138S mutant virus (VIC11pTRIG_A138S), V186G mutant virus (VIC11pTRIG_V186G), F193Y mutant virus (VIC11pTRIG_ F193Y), and ty/OH/04p binding to pSTECs. An uninfected control was included (negative). pSTECs were infected with each virus at a multiplicity of infection (MOI) of 100 at 4°C and fixed after 30 min incubation. H3N2 HA proteins (red) were detected on the cell surface using monoclonal antibody. The nuclei were stained with DAPI (4′,6-diamidino-2-fenilindol) (blue). The merged images are shown. Scale bar = 100 μm. Insets represent 2× magnification of boxes (scale bar of insets, 50 μm). Each experiment was performed two times in triplicate.

FIG 6.

In vitro replication kinetics of mutants containing A138S, V186G, and F193Y substitutions in primary swine tracheal epithelial cells (pSTECs) and Madin-Darby canine kidney (MDCK) cells. Viral growth kinetics of A138S mutant virus (VIC11pTRIG_A138S), V186G mutant virus (VIC11pTRIG_V186G), F193Y mutant virus (VIC11pTRIG_F193Y), VIC11pTRIG reassortant and ty/OH/04p in (A) MDCK cells and (B) pSTECs. pSTECs and MDCK cells were infected at an MOI of 0.1, and supernatants were collected at 0, 12, 24, 48, and 72 h postinfection (hpi). Viral titers were quantified by real-time PCR with a TCID₅₀/mL equivalent. Values are shown as mean TCID₅₀/mL titers ± standard error of the mean.

DISCUSSION

Determining the evolutionary processes of influenza virus cross-species transmission is key to understanding the mechanisms which control the emergence of new influenza viruses and establishment in a new host population. Spillover events of FLUAV between humans and swine are common and contribute to the extensive diversity of FLUAV circulating in pigs. The TRIG backbone seems to have an important part in expanding this diversity as it has the capability to incorporate HA and NA segments without destabilizing the overall fitness of the resulting reassortant viruses. This constellation seems to present advantageous fitness in pigs as it continues to replace other constellations that are introduced. This has been the case for all human-origin FLUAV established in pigs in the United States since the TRIG backbone was introduced; the only difference is that more recently different combinations with H1N1pdm09 internal genes are being detected (2, 4, 19, 20, 22). While acquiring an efficient and permissive internal gene constellation seems to be a crucial step in the adaptation of FLUAV to a new species, the molecular mechanisms that drive viral evolution after transmission of human-origin viruses in swine is not well understood. Thus, we generated a novel reassortant virus (VIC11pTRIG) that contains human seasonal HA and NA in a backbone highly adapted to swine and assessed its evolution during replication and transmission in pigs. Our results show that this ideal backbone significantly increases fitness of human seasonal H3N2 in pigs and is crucial for transmissibility in pigs, allowing for the further evolution and selection of HA variants.

In this study, we have shown that acquiring the ideal combination of genes is likely the first step in the adaptation of human-origin FLUAV to successfully replicate and transmit in pigs. The VIC11pTRIG group showed significantly higher viral titers in nasal swabs and BALF compared to groups with a full constellation of human-origin internal genes or the addition of the H1N1pdm09 M gene. More importantly, VIC11pTRIG was the only VIC11 reassortant able to transmit among pigs. The fitness advantage of the TRIG cassette and the importance of an ideal gene constellation had been confirmed previously in co-infection studies, in which only an H3N2 virus with a particular constellation containing the TRIG cassette was able to transmit between pigs (12). Furthermore, the addition of the H1N1pdm09 M gene conferred a replication advantage with the human-origin internal gene constellation, as a higher number of infected pigs in this group shed higher viral titers compared to other groups with human seasonal internal genes (A/VIC/11 and VIC11rg), although differences were not statistically significant. However, although directly infected pigs shed virus, there was no transmission. Interestingly, while transmission was confirmed in the VIC11pTRIG reassortant group to similar levels as in the swine-adapted sw/MO/14 group, infection resulted in little to no lung lesions and significantly lower titers in the lungs compared to the swine-adapted virus. These findings are consistent with previous studies which demonstrated that reassortant viruses carrying the A/VIC/11 HA showed replication in the upper respiratory tract and limited replication in the lungs (19).

HA is a major factor for host specificity (23, 24) and a prime target for evolution as a human seasonal virus adapts to the swine host. Specific amino acid substitutions at or near the RBS have been shown to alter the receptor binding or antigenicity of influenza viruses (5, 25, 26), which likely affects host specificity. Hence, it is reasonable to infer that once the virus acquires a gene constellation that allows effective replication, in this case the pTRIG backbone, substitutions that favor receptor binding in the new host may be an additional step to an efficient transmission. Here, the 3 dominant substitutions that arose, A138S, V186G, and F193Y, were all located within the antigenic site B of H3N2 (27), near or within the RBS. Thus, it is possible that these substitutions altered the conformation of the HA and its interaction with the host cell. We confirmed that all three substitutions improved binding to swine cells in comparison to the A/VIC/11 HA. Interestingly, the 138 position was shown previously to affect infectivity (28–30). The advantageous effect of the A138S substitution for replication in swine was confirmed in growth kinetic assays in pSTECs, a system that mimics the swine respiratory tract, which suggests that this substitution was crucial for the transmissibility of the VIC11pTRIG reassortant virus in our study. Although this substitution was the only one fixed in respiratory contact animals after transmission, the other two substitutions also showed improved replication compared to the virus containing wild-type human FLUAV HA and NA (VIC11pTRIG). The V186G substitution has been detected in surveillance studies as a variant of A/VIC/11-like strains (31) and was associated with replication and evolution in an immunocompromised human (32). Although this variant was not transmitted as a major variant in pigs, it subsequently became dominant in a contact animal, and it may represent one of many different evolutionary pathways that could later be selected in swine as an advantageous variant, as previously proposed for transmissibility of avian H1N1 in mammals (33). The F193Y substitution was dominant in one directly infected pig and was potentially transmitted at a low frequency. However, it did not become dominant in contact pigs, suggesting that this evolutionary pathway did not present an advantage in the swine host.

Although the reassortant virus that we studied here contains human seasonal HA and NA, the rest of the genome was formed by well-adapted swine FLUAV internal genes, limiting our ability to evaluate evolution outside of the surface genes. Nevertheless, using an artificial virus with a background known to be effective in pigs allowed us to examine the immediate virus evolution after initial replication in the swine host. As expected, the diversity of the HA was significantly higher after replication in pigs compared to other segments, including the NA, and no substitutions were fixed in any other segments (data not shown). However, it is unknown whether the A138S HA substitution would have become fixed if a wholly human virus was tested.

Overall, this study suggests that for a human-origin virus to become adapted to pigs, it needs an efficient internal gene constellation capable of enhancing replication, allowing for HA diversification and selection of advantageous variants that will be transmitted between pigs. However, our study is limited to a single transmission event and further studies are needed to accurately confirm these evolutionary processes and the molecular mechanisms required for the adaptation of human-origin viruses in swine.

MATERIALS AND METHODS

Viruses.

The wild-type H3N2 human isolate A/Victoria/361/2011 (A/VIC/11) was obtained from St. Jude Children’s Research Hospital (kindly provided by Richard Webby). This virus was incorporated into the 2012 to 2013 human influenza vaccine for the Northern hemisphere (34). The wild-type H3N1 swine isolate A/swine/Missouri/A01410819/2014 (sw/MO/14) was obtained from the USDA National Veterinary Service Laboratories Swine FLUAV repository (19). This virus was one of the first strains of the 2010.1 lineage detected in pigs (19) and was used as a control since it represents a human-origin virus that had already become adapted to swine. The 2010.1 swine FLUAV lineage closest human seasonal HA ancestor is similar to the A/VIC/11 strain. Therefore, A/VIC/11 and sw/MO/14 are of similar ancestry. The swine-origin H3N2 virus A/turkey/Ohio/313053/2004 (ty/OH/04), which contains the triple reassortant internal gene constellation, was detected in turkeys but contained all genes of swine-origin (including TRIG backbone) (35). This virus was shown to replicate efficiently in pigs and to be highly efficient for generation of recombinant viruses via reverse genetics (36–38). It was used as the source for TRIG genes and to generate one of the control viruses for in vitro studies. Four viruses were generated by reverse genetics using cloned cDNAs: VIC11rg encodes all genes from A/VIC/11; VIC11p virus encodes the M gene from A/California/04/2009 (H1N1pdm09) and the remaining seven genes from A/VIC/11; VIC11pTRIG virus contains the HA and NA from A/VIC/11, the M from H1N1pdm09, and the remaining five genes of ty/OH/04; and ty/OH/04p virus encodes the M from H1N1pdm09 and the remaining seven genes of ty/OH/04. Three mutant viruses were generated, VIC11pTRIG_A138S, VIC11pTRIG_V186G, and VIC11pTRIG_F193Y, which encode the same genes as VIC11pTRIG except for the A138S, V186G, or F193Y substitutions in the HA respectively (H3 numbering). All viruses used in this study are described in Fig. 7. Viruses were generated using an eight-plasmid reverse genetics system based on the bidirectional plasmid vector pDP2002, as previously described (39, 40). The mutants were generated by using the Phusion Site Directed Mutagenesis kit (Thermo Fisher Scientific, Waltham, MA) with specific mutagenesis primers. Full-length sequencing of viral stocks was performed to verify gene combinations either by next-generation sequencing (NGS) or Sanger sequencing. Viruses were propagated in Madin-Darby canine kidney (MDCK) cells for stock preparation.

FIG 7.

Viruses used for this study. Specification of origin of gene segments in each virus. A/VIC/11, A/Victoria/361/2011 H3N2; ty/OH/04, A/turkey/Ohio/313053/2004 H3N2; A/CA/09, A/California/04/2009 H1N1; sw/MO/14, A/swine/Missouri/2014 H3N1. A138S, V186G, or F193Y, A/VIC/11 HA containing the A138S, V186G, or F193Y substitutions, respectively.

Cell lines.

Human embryonic kidney 293T cells (HEK293T) and MDCK cells were used for transfection of plasmids for generation of viruses by reverse genetics. MDCK cells were also used for viral growth kinetics. Both cell lines were cultured at 37°C with 5% CO₂ using Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% l-glutamine and 1% antibiotics/antimycotics (Sigma-Aldrich, St. Louis, MO). Primary swine tracheal epithelial cells (pSTECs) were used for viral binding and growth kinetic studies. Briefly, trachea samples were collected from 7-week-old pigs obtained from a high health status herd and humanely euthanized with a lethal dose of pentobarbital (Euthasol, Virbac, Carros, France). Animals were cared for in compliance with the Institutional Animal Care and Use Committee of the University of Georgia. pSTECs were harvested from trachea samples by trimming and digesting the tissue in DMEM/F12 media containing 1.5 mg/mL of pronase (Sigma-Aldrich, St-Louis, MO) with antibiotic/antimycotic (Life Technologies, Waltham, MA) at 4°C for 24 h. The harvested pSTECs were then cultured and differentiated under air-liquid interface (ALI) conditions in collagen-coated transwell inserts (Corning, New York, NY) using TEC plus and TEC plus ALI media (respectively), as previously described (41). The pSTECs were cultured for a minimum of 3 weeks at 37°C with 5% CO₂. The basal media of the transwell inserts were replaced every 48 h. Cilia activity was checked every 2 to 3 days and trans-epithelial electrical resistance was measured to ensure confluence of the cells using an EVOM meter (World Precision Instruments, Sarasota, FL).

In vivo study.

Eighty 3-week-old cross-bred healthy pigs were obtained from a herd free of FLUAV and porcine reproductive and respiratory syndrome virus (PRRSV). Prior to the start of the study, pigs were treated with ceftiofur crystalline-free acid (Zoetis Animal Health, Parsippany, NJ) and enrofloxacin (Elanco Animal Health, Greenfield, IN) to reduce bacterial contaminants. Animals were demonstrated to be free of other respiratory pathogens by testing bronchoalveolar lavage fluid (BALF) at the end of the study for PRRSV, porcine circovirus type 2 (PCV2), and Mycoplasma hyopneumoniae nucleic acid by RT-PCR (VetMax, Life Technologies, Carlsbad, CA) and shown to be seronegative to FLUAV antibodies by a commercial enzyme-linked immunosorbent assay kit (AI MultiS-Screen kit, IDEXX, Westbrook, ME). Pigs were divided into six groups, housed in biosafety level 2 containment, and cared for in compliance with the Institutional Animal Care and Use Committee of the National Animal Disease Center (USDA-ARS).

Pigs (n = 10/group) were inoculated intranasally (1 mL) and intratracheally (2 mL) with 10⁵ TCID₅₀ per mL of each assigned virus: A/VIC/11, VIC11rg, VIC11p, VIC11pTRIG, and sw/MO/14. Five pigs were assigned as negative controls (NC). Inoculation was performed under anesthesia, using an intramuscular injection of a cocktail of ketamine (8 mg/kg of body weight), xylazine (4 mg/kg), and Telazol (6 mg/kg) (Zoetis Animal Health, Parsippany, NJ). Five contact pigs were placed in separated raised decks in the same room as each inoculated group at 2 dpi to evaluate respiratory (airborne) transmission. Pigs were observed daily for clinical signs of respiratory disease. Nasal swabs (FLOQSwabs, Copan Diagnostics, Murrieta, CA) were collected from 0 to 5 dpi for directly inoculated pigs and from 0- to 5-, 7-, and 9-dpc for respiratory contacts, placed in 2 mL minimal essential medium (MEM), and frozen at −80°C until used. Two pigs died from causes unrelated to FLUAV, leaving eight pigs in the VIC/11 wt group. Primary pigs were humanely euthanized with a lethal dose of pentobarbital (Fatal Plus, Vortech Pharmaceuticals, Dearborn, MI) and necropsied at 5 dpi. Postmortem samples included BALF, trachea, and right cardiac or affected lung lobe. Indirect contact pigs were humanely euthanized at 15 dpc.

Virus titers in nasal swabs and lungs.

For virus isolation, nasal swab (NS) samples were filtered (0.45 μm) and 200 μL was plated in 24-well plates onto confluent MDCK cells washed twice with phosphate-buffered saline (PBS), as previously described (42). Tenfold serial dilutions in serum-free Opti-MEM (Gibco, Life Technologies, Carlsbad, CA) supplemented with 1 μg/mL tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin and antibiotics were made with each BALF and virus isolation-positive NS sample. Each dilution was plated in triplicate onto PBS-washed confluent MDCK cells in 96-well plates. At 48 h, plates were fixed with 4% phosphate-buffered formalin and stained using immunocytochemistry with an anti-influenza A virus nucleoprotein monoclonal antibody as previously described (43). TCID₅₀/mL virus titers were calculated for each sample according to the methods of Reed and Muench (44).

Pathological examination of lungs.

At necropsy, lungs were removed and evaluated for the percentage of the lung affected with purple-red consolidation typical of FLUAV infection. The percentage of the surface affected by pneumonia was visually estimated for each lung lobe, and a total percentage for the entire lung was calculated based on weighted proportions of each lobe to the total lung volume (45). Tissue samples from trachea and lung were fixed in 10% buffered formalin for histopathologic examination. Tissues were routinely processed and stained with hematoxylin and eosin. Microscopic lesions were evaluated by a veterinary pathologist blinded to treatment groups and scored according to previously described parameters (46). Trachea and lung tissues were assessed for FLUAV-specific antigen using immunohistochemistry and scored as previously described (46). Individual scores were summed and a composite score for each pig was computed for lung and trachea microscopic lesions.

Serology.

Serum samples were collected from indirect contact pigs at 15 dpc to check for seroconversion using an HI assay. Prior to HI, sera were treated with receptor-destroying enzyme (Sigma-Aldrich, St. Louis, MO), heat-inactivated at 56°C, and adsorbed with 50% turkey red blood cells (RBC). HI assays were performed with either A/VIC/11 or sw/MO/14 as antigens and 0.5% turkey RBCs using standard techniques (47). Reciprocal titers were divided by 10 and log₂-transformed and reported as the geometric mean.

Next-generation sequencing and variant analysis.

Viral RNA was extracted from NS samples and tissue culture supernatants via the MagMAX AI/ND extraction kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. After extraction, viral RNA was amplified by a one-step multi-segment RT-PCR using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA) as previously described (48). PCR conditions were 55°C for 2 min, 94°C for 2 min, followed by 35 cycles at 94°C for 30 sec, 50°C for 30 sec, and 68°C for 3 min, and a final elongation of 4 min at 68°C. Influenza whole-genome sequencing libraries were prepared using the Nextera XT DNA library preparation kit (Illumina, San Diego, CA). Libraries were selected and purified using 0.7× Agencourt AMPure XP Magnetic Beads (Beckman Coulter Life Sciences, Indianapolis, IN) and samples were normalized to 4 nM and pooled. Fragment size distribution was analyzed using the High Sensitivity DNA kit (Agilent, Santa Clara, CA). Pooled libraries were sequenced via the high-throughput MiSeq platform (Illumina, San Diego, CA) using the 300-cycle MiSeq reagent kit in a paired end format (Illumina, San Diego, CA). Full genome assembly was performed using a pipeline from previously published methods (48). Variant calling was performed using LoFreq v2.1.3.1 (49) with a frequency threshold of 0.02 based on replicate sequence runs for the same control samples, a minimum depth of coverage of 100, and a central base quality score of Q30 or higher. Quantification of viral diversity within and between hosts was calculated using the consensus sequences, and single nucleotide variants within all hosts were derived from the variant call data utilizing the DNASTAR Lasergene version 17 (DNASTAR, Madison, WI).

Mapping of variants on the HA protein.

Three-dimensional (3D) analysis was conducted on the HA protein of the VIC11pTRIG variants identified in the study. Protein sequences were checked using DNASTAR Lasergene Version 17 (DNASTAR, Madison, WI) and ExPASy (Swiss Institute of Bioinformatics resources) (50), submitted to the I-TASSER (University of Michigan, Ann Arbor, MI), and visualized with CHIMERA v1.141 (University of California, San Francisco, CA). Only protein structures with a positive C-score (confidence score) were selected for analysis.

Viral binding assay.

The binding profiles of VIC11pTRIG, VIC11pTRIG_A138S, VIC11pTRIG_V186G, VIC11pTRIG_F193Y, and ty/OH/04p were evaluated in pSTECs. pSTECs were seeded and differentiated as mentioned above. Before infection cells were incubated for 10 min at 4°C and subsequently infected at a MOI of 100 using infectious medium consisting of F12/DMEM supplemented with 0.075% bovine serum albumin (BSA; Sigma-Aldrich St. Louis, MO) and 200 nm GlutaMax (Thermo Fisher Scientific, Waltham, MA). Plates were placed at 4°C for 30 min to allow virus binding, but not internalization, and then cells were fixed with 4% paraformaldehyde for 30 min. For these experiments, cells were not permeabilized. After fixation, cells were incubated for 1 h in blocking buffer (5% BSA in PBS). Cells were washed three times with PBS, and a primary anti-multi-hemagglutinins H3N2 antibody (eEnzyme, Gaithersburg, MD) diluted 1:500 in blocking buffer was added. After 1 h of incubation, cells were washed three times with PBS and incubated for 1 h with an Alexa Fluor 594-conjugated secondary antibody (Thermo Fisher Scientific, Waltham, MA) diluted 1:1,000 in blocking buffer containing 0.5 μg/mL 4′,6-diamidino-2-fenilindol (DAPI; Thermo Fisher Scientific, Waltham, MA). Finally, cells were washed five times with PBS and mounted on glass slides with VECTASHIELD PLUS mounting medium (Vector Laboratories, Newark, CA). Imaging was performed using an Olympus IX51 fluorescence microscope.

Viral growth kinetics.

Viral growth kinetics were performed for strains VIC11pTRIG, VIC11pTRIG_A138S, VIC11pTRIG_V186G, VIC11pTRIG_F193Y, and ty/OH/04p in MDCKs and pSTECs. Both cells were infected at an MOI of 0.01. Infection media for MDCK consisted of Opti-MEM (Thermo Fisher Scientific, Waltham, MA) containing 1 μg/mL TPCK trypsin (Worthington Biochemicals, Lakewood, NJ) with 1% antibiotic/antimycotic solution (Sigma-Aldrich St. Louis, MO). Infectious media for pSTECs was the same as described above. Cells were infected in triplicates and supernatant samples were collected at 0, 12, 24, 48, and 72 hpi. Virus titers were quantified and titrated by real-time PCR (QuantaBio ToughMix, VWR International, Radner, PA) with a TCID₅₀/mL equivalent as previously described (51). Results from two independent experiments were compiled for analysis.

Statistical analysis.

All statistical analyses were conducted by using GraphPad Prism version 9.3.1 (GraphPad Software, San Diego, CA) with P < 0.05 considered significant. Statistical methods include nonparametric one-way analysis of variance (ANOVA), two-way ANOVA, and nonparametric t tests.