Genetic and Pathobiologic Characterization of Pandemic H1N1 2009 Influenza Viruses from a Naturally Infected Swine Herd

Abstract

Since its initial identification in Mexico and the United States, concerns have been raised that the novel H1N1 influenza virus might cause a pandemic of severity comparable to that of the 1918 pandemic. In late April 2009, viruses phylogenetically related to pandemic H1N1 influenza virus were isolated from an outbreak on a Canadian pig farm. This outbreak also had epidemiological links to a suspected human case. Experimental infections carried out in pigs using one of the swine isolates from this outbreak and the human isolate A/Mexico/InDRE4487/2009 showed differences in virus recovery from the lower respiratory tract. Virus was consistently isolated from the lungs of pigs infected with A/Mexico/InDRE4487/2009, while only one pig infected with A/swine/Alberta/OTH-33-8/2008 yielded live virus from the lung, despite comparable amounts of viral RNA and antigen in both groups of pigs. Clinical disease resembled other influenza virus infections in swine, albeit with somewhat prolonged virus antigen detection and delayed viral-RNA clearance from the lungs. There was also a noteworthy amount of genotypic variability among the viruses isolated from the pigs on the farm. This, along with the somewhat irregular pathobiological characteristics observed in experimentally infected animals, suggests that although the virus may be of swine origin, significant viral evolution may still be ongoing.

The zoonotic potential of swine influenza viruses is well recognized (18), and pigs have been considered a leading candidate for the role of intermediate host in the generation of reassortant influenza A viruses with pandemic potential. This has been largely based on genomic analysis of influenza A viruses isolated from swine and the fact that α2,3-linked sialic acid (avian-like) and α2,6-linked sialic acid (human-like) receptors are both abundant in the swine respiratory tract (12). Despite this, there is no direct evidence that the reassortment of the 1957 and the 1968 human pandemic viruses occurred in pigs (28). Furthermore, it is very likely that the 1918 pandemic virus was introduced to pigs from humans (8, 31). The origins of influenza A viruses that have been isolated from pigs include those that are wholly human or avian, as well as reassortants containing swine, human, and avian genes (2, 20, 29). Although there have been several instances of swine-to-human transmission, for example, that of triple-reassortant swine influenza (H1) viruses (rH1N1), which appeared after 1998, they did not lead to establishment of sustained transmission in the human population (23).

In the early spring of 2009, Mexico and the United States reported clusters of human pneumonia cases caused by a novel H1N1 influenza A virus. This virus subsequently spread across the globe at an unprecedented rate, prompting the WHO to declare a pandemic in June 2009. Phylogenetic analysis has inferred that the virus is likely a reassortant between a North American triple-reassortant swine H1N1 or H1N2 virus and a Eurasian lineage H1N1 swine influenza virus (7, 19). Bayesian molecular-clock analysis of each gene of this novel H1N1 virus (24) concluded that the mean evolutionary rate is typical of that of swine influenza viruses but that the duration of unsampled diversity for each gene segment had means that ranged from 9.24 to 17.15 years, suggesting that the proposed ancestors of this virus may have been circulating undetected for nearly a decade. Inadequate surveillance and characterization of influenza A viruses that circulate in swine have been blamed for this evolutionary gap.

On 28 April 2009 the Canadian Food Inspection Agency (CFIA) became involved in a suspected outbreak of swine influenza on a pig farm in Leslieville, Alberta, Canada. The farm was a 220-sow farrow-to-finish operation consisting of approximately 2,200 animals that ranged from newborn piglets to market weight pigs. The animals were not vaccinated against swine influenza, and although there had been prior problems with porcine reproductive and respiratory syndrome virus and Mycoplasma hypopneumoniae, two etiologic agents of the swine respiratory disease complex, the herd had been stable with respect to respiratory disease. Beginning 20 April, approximately 25% of the pregrower and grower pigs in two of the barns exhibited respiratory problems with clinical signs that included an acute onset of coughing, lethargy, and loss of appetite. These clinical signs were preceded by the hiring of a carpenter on 14 April to work on the ventilation system in the same two barns. This individual had been ill for 2 days after his return from Mexico on 12 April (10). Given the evolving situation in Mexico and the United States, the CFIA and Alberta Agriculture and Rural Development decided to place the herd under quarantine and to carry out a full epidemiological and laboratory investigation.

Here, we report on the characterization of the first pandemic H1N1 2009 viruses to be isolated from a naturally infected pig herd. Genetic sequence data from several viruses isolated from this outbreak have provided a glimpse into the mutation frequencies associated with replication of the virus in the swine host. Experimental infections of pigs comparing one of these swine isolates with the human isolate A/Mexico/InDRE4487/2009(H1N1) were also carried out and have provided insights into the pathobiological behavior of these viruses in pigs.

MATERIALS AND METHODS

Virus detection in field samples.

Total RNA was extracted from nasal-swab samples that were collected from individual pigs using an RNeasy Mini Kit (Qiagen). The resulting RNA was initially tested using a single-tube real-time RT-PCR assay that targets the M1 gene of influenza A viruses (25). This assay was modified to include detection of an armored RNA enterovirus (Asuragen) internal exogenous control, which was used to assess extraction efficiency and the presence of PCR inhibitors. The primers and probe used to detect the internal exogenous control were as follows: forward primer, 5′-CCTGTCGTAACGCGCAAGT-3′; reverse primer, 5′-CAGCCACAATAAAATAAAAGGAAACA-3′; probe, 5′-TET-CGTGGCGGAACCGACTACTTTGGG-BHQ-3.

Samples were run on a Cepheid SmartCycler using a OneStep reverse transcription (RT)-PCR kit (Qiagen). A conventional RT-PCR assay also targeting the influenza A virus M1 gene (6) and a conventional RT-PCR designed to detect the H1 gene of A/California/04/09-related viruses (Public Health Agency of Canada, National Microbiology Laboratory, unpublished method) were also used. Virus isolation was carried out concurrently by inoculating the allantoic cavities of 9- to 10-day-old embryonated chicken eggs (ECE) with antibiotic-treated, clarified nasal-swab samples. Allantoic fluid was harvested from any embryos that died after 24 h postinoculation and tested for the ability to hemagglutinate 0.5% suspensions of chicken red blood cells. Allantoic fluid was harvested from all surviving embryos at 6 days postinoculation at the conclusion of the first passage and tested for hemagglutinating agents, as well as for the presence of influenza A virus nucleic acid, by various RT-PCR assays.

ELISA.

Influenza group A-specific nucleoprotein (NP) antibodies were detected using a competitive enzyme-linked immunosorbent assay (cELISA), as described previously (32). The assay employed a positive cutoff of 30% inhibition, which was based on known chicken and turkey serum samples.

Nucleotide sequencing and phylogenetic analysis.

The coding regions for all genomic segments were amplified in a one-step RT-PCR using a universal primer set (9) and a high-fidelity RT-PCR kit (Invitrogen; SuperScript III One-Step RT-PCR System with Platinum TaqHigh Fidelity). RT-PCR amplicons were cloned into pCR4-Topo (Invitrogen) and used to transform OneShot TOP10m or MachTM-T competent E. coli. Consensus sequences for each gene were generated from three or more clones of each recombinant plasmid. The sequencing reactions employed BigDye Terminator chemistry, the products of which were run on an Applied Biosystems 3130xl Genetic Analyzer. A control sequence provided with the kit was included with each run. Overlapping sequences were generated for PB1, PA, hemagglutinin (HA), and neuraminidase (NA) gene segments. Partial overlapping sequences were generated for PB2, NP, M, and NS gene segments. Contiguous sequences were generated using the Lasergene software package.

Sequences were aligned with Clustal W (MEGA 4.0.2), and phylogenetic trees were inferred by the maximum-parsimony method utilizing the close-neighbor interchange search and 500 bootstrap replicates (27).

Viruses used for animal inoculation experiments.

A/swine/Alberta/OTH-33-8/2009 (H1N1) was isolated from a nasal-swab specimen collected from pig no. 8/barn D (Table 1 ) after one passage in ECE. This isolate had been randomly selected for full-genome sequencing and for that reason was also selected for inoculation of piglets. The inoculum consisted of a pool of allantoic fluid harvested from 2 ECE. A/Mexico/InDRE4487/2009 (H1N1) was isolated from a bronchoalveolar lavage fluid (BALF) sample that was submitted to the Public Health Agency of Canada's National Microbiology Laboratory using Madin-Darby canine kidney (MDCK) cells. This was followed by propagation in ECE prior to animal inoculation.

TABLE 1.

Summary of diagnostic test results for nasal-swab specimens

Sample	CT value		Conventional M gene RT-PCR assayb	Conventional H1 gene RT-PCR assay	Virus isolation	cELISA for Influenza A NP antibodiesc
	Spackman M gene rRT-RCR assaya	Modified Spackman M gene rRT-PCR assay
Pig no. 1, barn D	0	29.23	+	+	+	Negative
Pig no. 2, barn D	39.94	27.09	+	+	+	Negative
Pig no. 3, barn D	0	29.39	+	+	+	Negative
Pig no. 4, barn D	0	31.32	+	+	+	Negative
Pig no. 5, barn D	0	37.66	Weak +	−	+	Negative
Pig no. 6, barn D	0	34.41	Weak +	−	+	Negative
Pig no. 7, barn D	0	33.13	Weak +	−	+	Negative
Pig no. 8, barn D	36.94	26.26	+	+	+	Negative
Pig no. 9, barn D	0	27.74	+	+	+	Negative
Pig no. 10, barn D	0	33.99	+	Weak +	+	Negative
Pig no. 11, barn D	0	27.92	+	+	+	Negative
Pig no. 12, barn D	0	0	−	−	+	Positive
Pig no. 13, barn D	0	39.95	−	−	+	Positive
Pig no. 14, barn D	0	32.90	Weak +	−	+	Negative
Pig no. 15, barn D	0	0	−	−	−	Positive
Pig no. 16, barn D	0	38.96	−	−	+	Positive
Pig no. 17, barn D	0	31.14	Weak +	+	+	Negative
Pig no. 18, barn E	0	32.23	Weak +	Weak +	−	Positive
Pig no. 19, barn E	0	0	−	−	−	Positive
Pig no. 21, barn E	39.51	39.96	+	+	+	Positive
Pig no. 22, barn E	0	28.79	+	+	+	Negative
Pig no. 23, barn E	0	27.34	+	+	+	Negative
Pig no. 24, barn F	0	30.85	Weak +	+	+	Negative
Pig no. 25, barn F	0	30.40	+	+	+	Negative

^arRT-PCR, real-time RT-PCR (25).

^bFrom reference 6.

^cFrom reference 32.

Virus titration.

Virus titers were determined by endpoint titration on monolayers of MDCK cells grown in 96-well microtiter plates (Costar; Corning). Serial 10-fold dilutions of virus samples in α minimal essential medium (α-MEM)-0.3% bovine serum albumin (BSA) supplemented with 10 U/ml of tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK) trypsin (Worthington Biochemical Corporation, Lakewood, NJ) were used as previously described (31).

Animal experiments.

The absence of preexisting anti-influenza antibodies in 3-week-old Landrace piglets was confirmed using the HerdCheck Swine Influenza Virus (H1N1) Antibody Test Kit (IDEXX Laboratories) and an influenza A group-specific cELISA (32). A total of six animals were inoculated with A/swine/Alberta/OTH-33-8/2009, three by the intratracheal route and three by the intranasal route. Six animals were also inoculated with A/Mexico/ InDRE4487/2009, three intratracheally and three intranasally. All animals received 10^5.6 50% tissue culture infective doses (TCID₅₀) of the corresponding virus. Four control piglets were mock inoculated with phosphate-buffered saline (PBS). Rectal temperatures and clinical signs were recorded daily. Nasal and pharyngeal swabs were collected daily until 10 days postinoculation (p.i.) and at 13 days p.i. Rectal swabs and whole blood were collected at 1, 3, and 5 days p.i. Skeletal muscle from hind legs, BALFs, sections of individual lung lobes, turbinates, submandibular lymph nodes, and lung-associated lymph nodes were collected at euthanasia. Separate sterile instruments were used for sampling each tissue. All animal manipulations were in compliance with the Canadian Council on Animal Care guidelines (see Table 6).

TABLE 6.

Routes of inoculation for individual animals, schedule of euthanasia, and virus detection in lung-associated lymph nodes, turbinates, and lung samples

Isolate inoculateda	Pig no.	Route of virus inoculation	Euthanasia on day p.i.b:					rRT-PCRc		Virus isolation titerd
			3	5	7	9	13	BLN	RDL	BLN	Turbinates	RAL	RDL	BALF
Mexico	4	Tracheal	X					6.75	7	2.2	ND	5.66	4.7	3.1
	6	Nasal	X					6.3	6.85	2.2	5.6	4.94	4.3	4.05
	5	Nasal		X				6.34	7.83	-	4.2	5.55	4.7	4.8
	2	Tracheal			X			5.53	7.88	-	3	3.8	4.3	2.7
	3	Tracheal				X		4.88	4.45	-	-	-	-	-
	1	Nasal					X	3.8		-	-	-	-	-
Alberta	8	Nasal	X					6.6	4.74	1.9	2.7	-	-	-
	10	Tracheal		X				6.16	6.95	-	3.3	1.9	1.5	-
	12	Nasal			X			5.86	5.79	-	-	-	-	-
	9	Tracheal				X		4.7	4	-	-	-	-	-
	11	Nasal					X	4.2	3.45	-	-	-	-	-

^aMexico, isolate A/Mexico/InDRE4487/2009; Alberta, pig isolate A/swine/Alberta/OTH-33-8/2009 cultured from pig no. 8, barn D.

^bX indicates the day of euthanasia.

^crRT-PCR, real-time RT-PCR targeting the matrix protein. The results are given as log₁₀ copy numbers/0.1 g of tissue. RDL, right diaphragmatic lung lobe; BLN, bronchial lymph nodes.

^dVirus isolation in TCID₅₀/0.1 g of tissue; -, no virus isolation; ND, not done; RAL, right apical lung lobe.

Virus RNA detection in experimental samples.

Virus RNA was extracted from samples using TriPure Isolation Reagent (Roche Applied Science, Indianapolis, IN) according to the manufacturer's instructions. The forward and reverse primers of a real-time RT-PCR assay that targets the M1 gene (25) were modified (see Results) to increase the sensitivity of detection of this novel H1N1 virus. Samples were run on a Rotor-Gene 6000 thermocycler using the Qiagen One Step RT-PCR kit under the following reaction conditions: 30 min at 50°C and 15 min at 95°C, followed by 45 cycles of 5 s at 94°C and 60 s at 60°C. Full-length, in vitro-transcribed segment 7 RNA, serially diluted in buffer, was run with each assay in order to give a semiquantitative estimate of the viral load in each tissue. Samples with <750 copies/ml (0 to 25 copies/reaction) were considered negative. Armored RNA enterovirus (Asuragen) was used as an internal exogenous control.

Virus isolation from experimental samples.

Virus isolation from filter-sterilized fluid samples or from clarified 10% (wt/vol) tissue homogenates prepared in PBS was carried out on MDCK cells using a protocol similar to the one described above for virus endpoint titration.

For selected lung and BALF samples from pigs inoculated with the swine isolate, additional virus isolation attempts were carried out as follows. MDCK cell monolayers in the P12 plates were inoculated in triplicate with 10% (wt/vol) lung tissue homogenate diluted 1:100 in Dulbecco's modified Eagle's medium (DMEM) that was supplemented with 10 U/ml of TPCK-treated trypsin and penicillin-streptomycin. Following 1 h of incubation at 37°C in a humidified 5% CO₂ atmosphere, the cell monolayers were washed three times with DMEM, and then maintenance medium consisting of DMEM-0.3% (wt/vol) BSA supplemented with 10 U/ml of TPCK-treated trypsin and penicillin-streptomycin was added. After 48 h of incubation at 37°C in a humidified 5% CO₂ atmosphere, the culture supernatants and the washed cell monolayers were separately harvested into 0.9 ml of TriPure Isolation Reagent (Roche Applied Science) and analyzed by real-time RT-PCR for the presence of viral RNA.

Histopathology and IHC.

Immunohistochemistry (IHC) was performed as described previously (31), using the mouse monoclonal antibody F26NP9, specific for influenza A virus NP. The percentage of affected bronchioles in the lungs of piglets was calculated based on 25 to 34 fields per lung section at 10× magnification. Bronchioles with epithelial necrosis, attenuation of epithelial cells, acute inflammation, and lumenal exudate were considered affected. The IHC results were compared with findings in adjacent hematoxylin- and eosin-stained sections, allowing the correlation of the percentage of bronchioles with lesions and antigen presence.

RESULTS

Virus isolation and molecular characterization.

Twenty-four nasal swab and serum specimens collected from animals in the pregrower (pigs 10 to 14 weeks of age), grower (pigs 14 to 18.5 weeks of age), and grower/finisher (pigs 18.5 to 21 weeks of age) barns were analyzed at the National Centre for Foreign Animal Disease in Winnipeg. In light of the history of influenza-like illness in people and animals and the recent return from Mexico by the person hired to upgrade the ventilation system in the 2 most affected barns, the primers and probe of the real-time M1 gene RT-PCR assay (25) that was in routine use were compared with M1 gene sequences for several of the human swine origin H1N1 viruses that had been deposited in the Global Initiative on Sharing Avian Influenza Data (GISAID) open-access database. Based on this comparison, the primers were modified as follows to alleviate any potential loss of analytical sensitivity: M + 25, 5′-AGATGAGTCYTCTAACCGAGGTCG-3′; M − 124, 5′-TGCAAARACAYYTTCMAGTCTCTG-3′.

This modified assay detected 17 swab specimens as positive and 4 as suspicious for influenza A virus. In contrast, the real-time M1 gene RT-PCR assay as originally described (25) identified only 1 borderline positive and 2 suspicious samples. Ancillary testing by conventional RT-PCR assays specific for the M1 gene (6) and a conventional RT-PCR assay designed to detect the H1 gene of viruses related to A/California/04/2009 (National Microbiology Laboratory, unpublished method) identified 19 and 15 nasal-swab specimens as positive, respectively. Twenty-one of 24 nasal-swab samples yielded influenza A virus isolates from ECE. The majority of these were not associated with embryo mortalities and did not hemagglutinate chicken red blood cells. Virus detection relied on use of the modified M gene real-time RT-PCR assay, which was carried out on allantoic fluid harvested from live ECE at 4 and 6 days postinoculation. The 4- and 6-day p.i. isolates produced threshold cycle (C_T) values of 15.55 ± 4.06 (n = 10) and 18.64 ± 1.96 (n = 18), respectively, indicating that these viruses were capable of growing to high titers in ECE. This conclusion was supported by titration on MDCK cells, where titers of 10⁵ to 10⁶ TCID₅₀/ml were recorded. The H1 subtype of the isolates was confirmed by H1-specific conventional RT-PCR assay. Follow-up analysis showed that these virus isolates were capable of weakly hemagglutinating turkey red blood cells. The diagnostic test results are summarized in Table 1. Isolate A/swine/Alberta/OTH-33-8/2009 was randomly selected for full-genome sequencing. Phylogenetic analysis showed that the genome of this isolate was most similar to the novel H1N1 swine origin influenza A viruses that emerged in humans in the spring of 2009 (Fig. 1). Sequences for the H1 and NP genes obtained from 8 isolates and for M and N1 genes from 10 and 11 isolates, respectively, were very similar, but not identical, to each other (Tables 2 through 5). Analysis of the 8 H1 genes found 6 synonymous and 17 nonsynonymous nucleotide substitutions, with an estimated overall substitution frequency of 4 × 10⁻³. Analysis of the 11 N1 genes found 4 synonymous and 24 nonsynonymous nucleotide substitutions, also with an estimated overall substitution frequency of 4 × 10⁻³.

FIG. 1.

Phylogenetic trees for H1 and N1 gene sequences. (A) Unrooted nucleotide phylogenetic tree of H1 HA from selected H1N1 and H1N2 influenza A virus isolates of swine, human, and avian origin. The swine viruses were chosen to be representative of current, as well as historical, isolates from North America and Eurasia. The tree is one of a total of 19 equally parsimonious trees that were inferred using maximum parsimony, a close-neighbor interchange search, and 500 bootstrap replicates (bootstrap values are shown on the tree). The scale indicates the number of nucleotide changes. (B) Unrooted nucleotide phylogenetic tree of N1 NA from selected Eurasian and North American swine, human, and avian influenza A virus isolates. The swine isolates are representative of viruses that have been circulating in pigs in Eurasia since the early 1990s. The tree is one of a total of 59 equally parsimonious trees that were inferred using maximum parsimony, a close-neighbor interchange search, and 500 bootstrap replicates (bootstrap values are shown on the tree). The scale indicates the number of nucleotide changes.

TABLE 2.

Amino acid substitutions in the A/2009 H1N1 influenza virus hemagglutinin^a

GenBank accession no.	Isolate	Amino acid at positionb:
		HA1											HA2
		67	89	170	193	198	199	220	222	250	288	295	334	376	385	422
GQ303340	M 4487	C	T	G	Q	Y	V	V	D	V	T	I	A	T	K	N
GQ369403	Pig no. 1 (33-1)	R	-	-	-	-	-	-	-	-	A	-	-	-	-	-
GQ369406	Pig no. 2 (33-2)	-	-	-	-	N	-	-	-	-	-	-	V	-	-	-
GQ369409	Pig no. 3 (33-3)	-	-	E	-	-	-	-	V	I	-	-	-	A	-	-
GQ369412	Pig no. 7 (33-7)	-	A	-	-	-	-	-	-	-	-	-	-	-	-	T
GQ369417	Pig no. 21 (33-21)	R	-	-	-	-	A	M	-	-	-	V	-	-	-	-
GQ369422	Pig no. 23 (33-23)	-	-	-	-	-	-	-	-	-	-	-	-	-	E	-
GQ369425	Pig no. 24 (33-24)	-	-	-	H	-	-	M	-	-	-	-	-	-	-	-
GQ150328	Pig no. 8 (33-8)	-	-	-	-	N	-	-	-	-	-	-	V	-	-	-

^aAmino acid substitutions in the hemagglutinin protein of H1N1 viruses isolated from different pigs from barns D, E, and F as listed in Table 1 compared with the human isolate A/Mexico/InDRE/4487/2009 (H1N1). M 4487, isolate A/Mexico/InDRE/4487/09; isolates A 33-1, -2, -3, -7, -21, -23, and -24 are swine isolates from Alberta; A 33-8 is the isolate A/sw/Alberta/OTH-33-8/09. Isolates in the rows in boldface are those that were used in experimental-infection studies.

^b-, no substitution.

TABLE 5.

Amino acid substitutions in the A/2009 H1N1 influenza virus nucleoprotein^a

GenBank accession no.	Isolate	Amino acid at positionb:
		35	49	107	136	198	208	228	295	379	381
FJ998217	M 4487	V	D	D	T	E	N	M	H	T	D
GQ478577	Pig no. 1 (33-1)	-	-	-	-	-	-	-	-	I	N
GQ478578	Pig no. 2 (33-2)	-	E	-	I	-	-	-	-	I	-
GQ478579	Pig no. 3 (33-3)	-	-	-	I	-	-	-	L	I	-
GQ478580	Pig no. 14 (33-14)	-	-	-	-	-	-	-	-	I	-
GQ478581	Pig no. 22 (33-22)	-	-	G	-	-	-	-	-	I	-
GQ478582	Pig no. 23 (33-23)	-	-	-	-	-	-	-	-	I	-
GQ478583	Pig no. 24 (33-24)	-	-	-	-	-	-	-	-	I	-
GQ150329	Pig no. 8 (33-8)	A	-	-	-	K	D	T	-	I	Y

^aAmino acid substitutions in the neuraminidase proteins of H1N1 viruses isolated from different pigs from barns D, E and F as listed in Table 1 compared with the human isolate A/Mexico/InDRE/4487/2009 (H1N1). M 4487, isolate A/Mexico/InDRE/4487/09; isolates A 33-1, -2, -3, -14, -22, -23, and -24 are swine isolates from Alberta; A 33-8 is the isolate A/sw/Alberta/OTH-33-8/09. Isolates in the rows in boldface are those that were used in experimental-infection studies.

^b-, no substitution.

Progression of the infection on the farm.

Seven of the 24 sera that were submitted initially tested positive for influenza A virus NP antibodies. The nasal-swab samples from the corresponding animals gave either negative or weakly positive influenza M gene real-time PCR results (Table 1), indicating that these animals were likely in the early stages of recovery. Experimental-infection studies have shown that pigs develop IgM and IgG antibodies against influenza A virus HA and NP by 7 days p.i. (14), which is supported by our own findings (data not shown). Collectively, these results imply that the herd was initially infected 10 to 14 days earlier, which roughly coincides with the time when the carpenter began work in the barns. Follow-up testing using random sampling of the herd was carried out 13, 15, 27, 29, and 34 days later. All barns had pigs with antibodies to influenza A virus NP, with an increase over time in the proportion of pigs that were seropositive (10). Although a small number of nasal-swab specimens collected 13 and 15 days after the initial submission were PCR positive for influenza A virus nucleic acid, no live virus was isolated from them. The above time lines are consistent with swine influenza, with a predicted incubation period of 1 to 3 days followed by clinical recovery and a rapid cessation of shedding 4 to 8 days later. Based on these time lines, we estimate that the viruses isolated from the nasal-swab specimens may have undergone between 2 and 5 passages in the pigs on this farm.

While the herd was still under quarantine, a limited cull involving 475 grower/finisher pigs was undertaken on 8 May for animal welfare reasons. Necropsies performed on 10 of these animals revealed a lobular to coalescing bronchopneumonia macroscopically. Mild, chronic, nonspecific tracheitis; moderate bronchointerstitial pneumonia with perivascular and peribronchiolar lymphoid hyperplasia; and mild, multifocal necrotizing and suppurative alveolitis were evident on follow-up histopathologic examination. All pigs had lesions affecting bronchioles compatible with swine influenza in a subacute to chronic reparative stage.

Lastly, swine-to-human transmission was suspected in 2 individuals who were exposed to the infected pigs on 28 April. Pandemic H1N1 2009 was confirmed in both individuals, who became symptomatic within the expected incubation period after entering the infected barn (10).

Experimental-infection studies.

In order to gain additional insight into the interaction between this pandemic H1N1 2009 virus and the swine host, six 3-week-old piglets were inoculated with 10^5.6 TCID₅₀ of A/swine/Alberta/OTH-33-8/2009 per animal, three by the intratracheal route and three intranasally. This virus was used because it had been initially selected for full-genome sequencing. An additional six piglets were inoculated with the human isolate A/Mexico/InDRE4487/2009 using the same dose and routes of inoculation. The routes of inoculation appeared to have no significant effect on the outcome of the infection; therefore, the results are presented only in relation to the virus used.

The infected animals developed mild sneezing and a transient increase in rectal temperature (about 1°C above normal) for 2 to 3 days p.i., with no remarkable differences observed among the different inoculation groups. Most of the piglets started to shed virus oronasally, as determined by real-time RT-PCR and virus isolation, at 1 day p.i. Virus shedding could be detected until 8 days p.i. by virus isolation and until 9 days p.i. by real-time RT-PCR (Fig. 2), with no significant differences observed between the human and swine isolates. Virus titers in nasal swabs of some animals reached >10⁶ TCID₅₀/ml early postinoculation. Virus was recovered from the turbinates of piglets inoculated with A/Mexico/InDRE4487/2009 at 3, 5, and 7 days p.i., with the highest titer of 10^5.6 TCID₅₀/0.1 g observed at 3 days p.i. Virus was also recovered from the turbinates of piglets inoculated with A/swine/Alberta/OTH-33-8/2009 at 3 and 5 days p.i. (Table 6).

FIG. 2.

Detection of virus shedding in nasal and pharyngeal swabs. (A) Virus shedding in nasal swabs (solid lines) and pharyngeal swabs (dashed lines). The gray trend line joining the average values of the virus titers indicates the A/swine/Alberta/OTH-33-8/2009 isolate; the black trend line indicates the virus isolates from piglets infected with A/Mexico/InDre4487/2009. (B) Detection of viral RNA in nasal swabs (solid lines) and pharyngeal swabs (dashed lines). The gray trend line joins the average copy numbers/ml detected in piglets infected with the A/swine/Alberta/OTH-33-8/2009 isolate; the black line joins the average copy numbers/ml detected in piglets infected with the A/Mexico/InDre4487/2009 isolate. dpi (days p.i.).

Muscle, rectal swabs, blood, and all submandibular lymph nodes tested negative for the presence of viral RNA in both inoculation groups, as determined by real-time RT-PCR. However, high copy numbers of the viral RNA were recovered from lung-associated lymph nodes, where viral RNA persisted until 13 days p.i., although virus was isolated from only three animals at 3 days p.i. and the highest titer was 10^2.2 TCID₅₀/0.1 g (Table 6).

High copy numbers of viral RNA were detected in lungs and BALFs of piglets in both inoculation groups early in infection. Piglets inoculated with A/Mexico/InDRE4487/2009 had high levels of viral RNA beginning at 3 days p.i., which persisted until 5 and 7 days p.i. and then decreased rapidly thereafter. A/swine/Alberta/OTH-33-8/2009 RNA was detected until 7 days p.i. but at levels that were somewhat lower than those observed in pigs inoculated with the human isolate. The virus appeared to spread homogeneously throughout the lung, as viral RNA could be detected in the same animal in comparable amounts in different lung lobes (Table 6 and Fig. 3). At 13 days p.i., low copy numbers of viral RNA were still detectable in BALFs and lung tissue, suggesting that the virus may not have been completely cleared from the lungs (Fig. 3). Virus was isolated in relatively high titers from the lower respiratory tracts of piglets infected with A/Mexico/InDRE4487/2009 (Fig. 3) (up to 10^5.6 TCID₅₀/0.1 g in apical lobes and up to 10^4.7 TCID₅₀/0.1 g in diaphragmatic lobes) at 3, 5, and 7 days p.i. In sharp contrast, virus isolation attempts from lung tissue and BALFs of piglets inoculated with A/swine/Alberta/OTH-33-8/2009 yielded low virus titers of around 10² TCID₅₀/ml from only piglet no. 10 at 5 days p.i. (Fig. 3 and Table 6).

FIG. 3.

Virus and viral-RNA detection in the lower respiratory tract. (A) The black solid trend line indicates RNA detection (copy numbers per 0.1 g of tissues) in the tissue lung homogenates (right apical lobe) from piglets infected with the A/Mexico/InDRE4487/2009 isolate, and the gray line indicates RNA detection in the lung tissue homogenates from a pig infected with the A/swine/Alberta/OTH-33-8/2009 isolate. The black dashed trend line represents the virus isolation from the human isolate and the gray dashed line represents the virus isolation from the lungs of pigs infected with the swine isolate. (B) The black solid trend line indicates RNA detection (copy number per 1 ml) in the BALF from piglets infected with the A/Mexico/InDRE4487/2009 and the gray line indicates RNA detection in the BALF from a pig infected with A/swine/Alberta/OTH-33-8/2009. The black dashed trend line represents the virus isolation from the human isolate; no virus was isolated from BALF of the piglets infected with the A/sw/Alberta/OTH-33-8/2009 isolate.

Attempts to reisolate virus from clarified lung tissue homogenates and filter-sterilized BALFs of pigs infected with the A/swine/Alberta isolate were repeated a total of four times with no success, confirming our initial findings. When cell culture monolayers and supernatants were examined separately for the presence of viral RNA at 48 h following inoculation with lung homogenates or BALFs, only small amounts of viral RNA were detected in the cell monolayer for pig no. 10. No viral RNA or only negligible amounts were detected in any of the culture supernatants.

In addition to the presence of viral RNA, viral NP, detected by IHC, was also present in bronchiolar epithelial cells of both groups at 13 days p.i. (Fig. 4). Based on a total of 7 lung sections per pig examined, the largest amounts of antigen were detected in both experimental groups at 5 days p.i., but in general, A/Mexico/InDRE4487/2009-infected piglets had more antigen-positive bronchioles than A/swine/Alberta/OTH-33-8/2009-infected piglets: 75% positive bronchioles versus 55% positive bronchioles. This ratio remained until 7 days p.i., but at 9 days p.i., the swine isolate still affected ∼40% of the bronchioles compared with <30% of the bronchioles in piglets infected with the human isolate. In A/Mexico/InDRE4487/2009-infected animals, viral NP was detected mainly in bronchiolar epithelial cells, while viral NP of A/swine/Alberta/OTH-33-8/2009-infected animals was found quite extensively in lung macrophages (Fig. 4). Interestingly, no positive staining was observed in the alveoli of either group.

FIG. 4.

Histopathology and IHC detection of the nucleoprotein antigen in lungs. (a) IHC staining in the lung of a piglet inoculated with A/swine/Alberta/OTH-33-8/2009 at 3 days p.i. showing nucleoprotein-positive staining in epithelial cells lining the bronchiole, as well as in macrophages (arrowheads) in the luminal exudates and interstitium. Bar = 50 μm. (b) IHC staining in the lung of a piglet inoculated with A/Mexico/InDRE4487/2009. Specific immunolabeling for influenza A virus nucleoprotein is confined to bronchiolar epithelial cells. The arrow indicates the junction between the bronchiole and alveolus. Bar = 50 μm. (c) Piglet infected with A/swine/Alberta/OTH-33-8/2009, 5 days p.i. Shown is relatively prominent staining of the bronchiolar epithelial cells; the arrowhead indicates an example of staining in macrophages. Bar = 20 μm. (d) Strong prominent immunolabeling for the nucleoprotein in the epithelial cells of bronchioli, with some staining in macrophages (arrowhead) of a piglet infected with the A/Mexico/InDRE4487/2009 isolate, 5 days p.i. Bar = 20 μm. (e) Specific staining for the nucleoprotein in epithelial cells of a small terminal bronchiole and in macrophages in the interstitium (arrowhead) from a piglet infected with the A/swine/Alberta/OTH-33-8/2009 isolate, 7 days p.i. Bar = 50 μm. (f) Serial section of the bronchiole shown in panel e. The small terminal bronchiole is lined by hyperplasic epithelium and surrounded by a moderate number of inflammatory cells. Bar = 50 μm. (g) Small bronchioles with hyperplasic epithelium surrounded by a very intense inflammatory reaction in a piglet infected with A/Mexico/InDRE4487/2009, 7 days p.i. Bar = 50 μm. (h) Only a few individual bronchiolar epithelial cells were positively stained for the nucleoprotein at 13 days p.i. in both human and swine isolates. Shown is an example of the IHC labeling in the lung of a piglet inoculated with the A/Mexico/InDRE4487/2009 isolate. Bar = 20 μm.

DISCUSSION

Phylogenetic analyses performed in this study and by others (7, 19, 24) have shown that each segment of the pandemic H1N1 2009 influenza virus clusters with an established swine influenza lineage, suggesting that the progenitors of this virus likely originated in pigs. In previously documented instances of swine influenza virus infections of people, only limited human transmission was evident (22, 23, 30). The pandemic H1N1 2009 virus has demonstrated the ability for sustained human-to-human transmission, and the results of this and an associated study (10) confirm that the virus is also capable of human-to-swine, swine-to-swine, and swine-to-human transmission.

Based on a presumed single point source of infection for this swine herd (10), it is worth noting that none of the isolates, based on the sequences obtained for the H1, N1, NP, or M gene, were identical. This is in contrast to the apparent genetic stability of pandemic H1N1 2009 viruses that have been isolated from humans, as well as the relative genetic and antigenic stability of classical North American swine H1N1 viruses that circulated in pigs between 1930 and 1998 (29). The high proportion of nonsynonymous-to-synonymous nucleotide substitutions observed for both the H1 and N1 genes, along with the apparent overall high substitution frequencies (4 ×10⁻³) compared to those documented previously (1.69 ×10⁻⁴ to 8.89 × 10⁻⁴) (11), suggests that either a higher than normal mutation rate, strong positive selective pressure, or a combination of both may be at work. Only one amino acid substitution, V250I, identified in isolate A/swine/Alberta/OTH-33-3/09, lies within the phylogenetically important region (PIR) N (amino acids 249 to 252) of the H1 protein (5). PIR N, along with PIRs L and R, distinguishes swine viruses from human and avian viruses, implying they may be swine-specific receptor or antigenic sites (5). Three isolates were identified with substitutions in PIRs of the N1 protein (4). Isolate OTH-33-1 has an I263T substitution in region G (amino acids 263 to 264), isolate OTH-33-21 has a Q45R substitution in region B (amino acids 41 to 48) and an F387S substitution in region M (amino acids 385 to 399), and isolate OTH-33-22 has an N434S substitution in region N (amino acids 430 to 434). Region B is a glycosylation site in humans, while regions M and N are potential antigenic sites (4). The significance that any of the above substitutions have with respect to virus growth or pathogenesis in the swine host is not known; however, the diversity is a feature that is consistent with what has been observed for the recent triple-reassortant H1N1 and H1N2 swine influenza viruses that have emerged in North America since 1998 (reviewed in reference 29).

Whether the observed genetic heterogeneity reflects the true variability of virus circulating in this swine herd or is an artifact resulting from virus propagation and/or viral-RNA manipulation is of concern. All the swine viruses in the present study were passaged once in ECE prior to RT-PCR amplification, cloning, and sequencing. Although the error rates associated with the reverse transcription and PCR amplification steps have not been definitively determined, the conversion of viral RNA to cDNA and its subsequent amplification involved the use of a high-fidelity RT-PCR kit, which should have minimized any errors associated with viral-RNA manipulation. The reliability of the data was further enhanced by reading a minimum of 3 clones per gene segment in order to arrive at a consensus sequence. A limited comparison of directly sequenced RT-PCR product versus sequencing of RT-PCR product that was first cloned in Escherichia coli was carried out to determine if some of the sequence variation was the result of a cloning artifact. A single-nucleotide change was observed in 2 of the 5 genes that were compared (data not shown), suggesting that the cloning steps did not contribute significantly to the observed sequence variation. A previous study that addressed the within-host variation of avian influenza viruses in individually infected birds generated similar concerns regarding the error rates associated with the reverse transcription and PCR amplification steps (11). These authors utilized a two-step RT-PCR that targeted the HA and NS gene segments on RNA that was extracted directly from swab specimens collected from experimentally infected turkeys, chickens, and ducks. Several sites along the HA and NS gene segments were shown to have a high degree of heterogeneity between clones, with a small number of clones showing a marked increase in variation compared with the consensus. The authors concluded that these cDNA clones observed at low frequency with a high number of mutations were not artifacts but likely represented a minor species of virus in the population.

“Host-mediated” mutations that result from the propagation of human influenza A viruses in ECE have been recognized. For example, a set of 22 codons in the hemagglutinin gene of human H3N2 viruses known to undergo “host-mediated” mutations during propagation in ECE have been shown to affect inferences of viral evolution (3). In this regard, D225G and Q226R substitutions, which are associated with a disturbance in the α2,6-linked sialic acid binding preference of human H1N1 viruses, were detected following the propagation of human H1N1 viruses in ECE (17). However, the interpretation that specific amino acid substitutions involving the HA receptor binding site can only result from the selection of variants by the culture system being used for virus propagation should be viewed with caution. It is worth pointing out that glycine and arginine residues are present at positions 225 and 226 in all pandemic H1N1 isolates, regardless of the species of origin or the system used for virus isolation and propagation. Increased mutation rates of influenza A viruses have been demonstrated in laboratory experiments aimed at generating influenza A monoclonal-antibody-resistant mutants. In these experiments, Suárez et al. (26) observed higher-than-expected mutation rates for some viral clones. They concluded that an influenza virus population is heterogeneous with respect to its mutation rate and can include the contribution of “mutator mutants” that would enable the virus to quickly adapt to a changing host environment. Evidence for this occurring in nature was reported for a wholly avian origin H1N1 influenza virus that was introduced into the European pig population and for which an elevated rate of change for all genes examined was observed in the years immediately after the virus crossed the species barrier (16). This characteristic would be an obvious advantage to a virus adapting to a new host. Further sequence data from pig herds that become naturally infected with pandemic H1N1 2009 should eventually provide a clearer picture of the natural heterogeneity and evolution of this virus in pigs.

The courses of experimental infection, along with macroscopic and microscopic lesions in the lungs and virus shedding from the upper respiratory tract, resembled other influenza A virus infections in swine (21) and were similar in the two isolates that were used in this study. Furthermore, the results of this study are in general agreement with other studies that have looked at experimental pandemic H1N1 2009 infections in pigs (1, 15). The distribution of viral NP in the lungs of pigs in this study, as determined by IHC, was of interest, however, since the relative number of positive cells and the intensity of staining at 5 days p.i., especially for the human 2009 H1N1 isolate, was similar to the staining seen in the lungs of the piglets infected with the reconstructed human 1918 H1N1 pandemic influenza virus and dissimilar to that seen with the 1930 H1N1 classical swine influenza virus and other swine influenza viruses (13, 31). Additionally, the observed prolonged detection of viral antigen and delayed clearance of viral RNA and antigen in lung tissue until 13 days p.i. have not been reported previously and may be considered somewhat atypical.

Differences were observed between A/Mexico/InDRE4487/2009 and A/swine/Alberta/OTH-33-8/2009 in the recovery of infectious virus from the respiratory tract, especially from the lungs and to a lesser extent from the turbinates. Considering the relatively similar viral-RNA copy numbers present in corresponding lung samples, the low yield of infectious virus from lung tissue and BALFs of piglets infected with the swine isolate was unexpected. The differences observed in virus isolation may be related to the observed differences in the distribution of virus NP. While the human isolate showed preference for bronchiolar epithelial cells, the NP staining of the swine isolate was also prominent in lung macrophages. The possibility that lung macrophages may not have fully supported the replication of A/swine/Alberta/OTH-33-8/2009 is one explanation that will need to be examined further.

In summary, the genetic variability found among the isolates from pigs following this single introduction event, along with a phenotype that is not entirely consistent with other influenza infections in swine, suggests that the pandemic H1N1 2009 virus may also be a relatively new virus for pigs. Repeated introductions of this virus from humans to pigs may have a significant impact on virus evolution and subsequent epidemiology. Because of this, human-to-pig transmission of pandemic H1N1 2009 should be minimized. If infected pig herds are identified, changes in the viral genotype and phenotype should be closely monitored.

TABLE 3.

Amino acid substitutions in the A/2009 H1N1 influenza virus neuraminidase^a

GenBank acccession no.	Isolate	Amino acid at positionb:
		45	82	86	95	96	100	123	125	151	176	200	230	263	282	307	317	334	349	371	387	398	405	418	434
FJ998214	M 4487	Q	S	A	S	G	Y	S	S	D	S	N	E	I	Y	N	I	S	F	F	F	E	S	I	N
GQ369404	Pig no. 1 (33-1)	-	-	-	G	-	-	-	-	-	-	-	-	T	H	-	M	-	-	-	-	-	-	-	-
GQ369407	Pig no. 2 (33-2)	-	-	-	G	-	-	-	-	-	-	-	A	-	-	-	-	-	-	S	-	G	-	-	-
GQ369410	Pig no. 3 (33-3)	-	-	T	G	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ369413	Pig no. 7 (33-7)	-	-	-	G	-	-	-	-	G	-	S	-	-	-	I	-	-	-	-	-	-	-	M	-
GQ369415	Pig no. 14 (33-14)	-	-	-	G	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	G	-	-
GQ369418	Pig no. 21 (33-21)	R	-	-	G	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	S	-	-	-	-
GQ369420	Pig no. 22 (33-22)	-	-	-	G	-	-	-	-	-	P	-	-	-	-	-	-	-	-	-	-	-	-	-	S
GQ369423	Pig no. 23 (33-23)	-	T	-	G	-	-	-	-	-	-	-	-	-	-	-	-	-	L	-	-	-	-	-	-
GQ369426	Pig no. 24 (33-24)	-	-	-	G	-	-	-	P	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ369428	Pig no. 25 (33-25)	-	-	-	G	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ150330	Pig no. 8 (33-8)	-	-	-	G	E	H	L	-	-	-	-	-	-	-	-	-	C	-	-	-	-	-	-	-

^aAmino acid substitutions in the neuraminidase proteins of H1N1 viruses isolated from different pigs from barns D, E and F as listed in Table 1 compared with the human isolate A/Mexico/InDRE/4487/2009 (H1N1). M 4487, isolate A/Mexico/InDRE/4487/09; isolates A 33-1, -2, -3, -7, -14, -21, -22, -23, -24, and -25 are swine isolates from Alberta; A 33-8 is the isolate A/sw/Alberta/OTH-33-8/09. Isolates in the rows in boldface are those that were used in experimental-infection studies.

^b-, no substitution.

TABLE 4.

Amino acid substitutions in the A/2009 H1N1 influenza virus M1 and M2 proteins^a

GenBank accession no.	Isolate	Amino acid at positionb:
		M2					M1
		34	55	68	70	72	78	99	134	139	144	155	181	207	222	230
FJ998211	M 4487	G	F	V	E	M	R	L	R	T	F	A	L	N	H	K
GQ369405	Pig no. 1 (33-1)	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ369408	Pig no. 2 (33-2)	-	S	-	-	-	-	-	G	-	-	-	-	-	-	-
GQ369411	Pig no. 3 (33-3)	-	-	-	-	-	-	-	-	-	-	-	-	D	-	R
GQ369414	Pig no. 7 (33-7)	-	-	-	-	I	-	Q	-	-	-	V	-	-	-	-
GQ369416	Pig no. 14 (33-14)	-	-	-	G	-	-	-	-	-	-	-	-	-	-	-
GQ369419	Pig no. 21 (33-21)	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ369421	Pig no. 22 (33-22)	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
GQ369424	Pig no. 23 (33-23)	E	-	-	-	-	H	-	-	-	S	-	-	-	-	-
GQ369427	Pig no. 24 (33-24)	-	-	M	-	-	-	-	-	A	-	-	P	-	-	-
GQ150331	Pig no. 8 (33-8)	-	-	-	-	-	-	-	-	-	-	-	-	-	R	-

^aAmino acid substitutions in the neuraminidase proteins of H1N1 viruses isolated from different pigs from barns D, E and F as listed in Table 1 compared with the human isolate A/Mexico/InDRE/4487/2009 (H1N1). M 4487, isolate A/Mexico/InDRE/4487/09; isolates A 33-1, -2, -3, -7, -14, -21, -22, -23, and -24 are swine isolates from Alberta; A 33-8 is the isolate A/sw/Alberta/OTH-33-8/09. Isolates in the rows in boldface are those that were used in experimental-infection studies.

^b-, no substitution.