Pandemic Seasonal H1N1 Reassortants Recovered from Patient Material Display a Phenotype Similar to That of the Seasonal Parent

ABSTRACT

We have previously shown that 11 patients became naturally coinfected with seasonal H1N1 (A/H1N1) and pandemic H1N1 (pdm/H1N1) during the Southern hemisphere winter of 2009 in New Zealand. Reassortment of influenza A viruses is readily observed during coinfection of host animals and in vitro; however, reports of reassortment occurring naturally in humans are rare. Using clinical specimen material, we show reassortment between the two coinfecting viruses occurred with high likelihood directly in one of the previously identified patients. Despite the lack of spread of these reassortants in the community, we did not find them to be attenuated in several model systems for viral replication and virus transmission: multistep growth curves in differentiated human bronchial epithelial cells revealed no growth deficiency in six recovered reassortants compared to A/H1N1 and pdm/H1N1 isolates. Two reassortant viruses were assessed in ferrets and showed transmission to aerosol contacts. This study demonstrates that influenza virus reassortants can arise in naturally coinfected patients.

IMPORTANCE Reassortment of influenza A viruses is an important driver of virus evolution, but little has been done to address humans as hosts for the generation of novel influenza viruses. We show here that multiple reassortant viruses were generated during natural coinfection of a patient with pandemic H1N1 (2009) and seasonal H1N1 influenza A viruses. Though apparently fit in model systems, these reassortants did not become established in the wider population, presumably due to herd immunity against their seasonal H1 antigen.

INTRODUCTION

Pandemic H1N1 (2009) (pdm/H1N1) was first recognized in California in April 2009 and subsequently spread globally (1–3) and had evolved by several reassortment events most recently between a North American triple reassortant swine influenza virus and a Eurasian avian-like swine H1N1 virus (4, 5).

In New Zealand, seasonal H3N2 (A/H3N2) and oseltamivir-resistant seasonal H1N1 (A/H1N1) were circulating in 2009 at the time of introduction of pdm/H1N1. The first recognized cases of pdm/H1N1 in New Zealand occurred in late April 2009, and the outbreak reached sustained community transmission by mid-June 2009 (6, 7). During a 2-week period of considerable cocirculation of A/H1N1 and pdm/H1N1 viruses in June 2009, 11 individuals were known to have acquired natural coinfections with A/H1N1 and pdm/H1N1 viruses (8).

Few investigations have addressed influenza coinfection in humans: in Japan coinfections with A/H3N2 or A/H1N1 and influenza B viruses were detected in 2004/2005 (9, 10), and in Corsica, France, in 2007 (11). In September 2009, six patients in Beijing, China, became coinfected with pdm/H1N1 and A/H3N2, but reassortment between the two viruses was not detected (12). In October 2009, two patients in Cambodia were also found to be coinfected with pdm/H1N1 and A/H3N2 (13). Patient material was plaque purified on MDCK cells, and six virus plaques were examined: no reassortment between the coinfecting viruses was detected (13). Recently, Rith et al. (14) reported reassortment between pdm/H1N1 and A/H3N2 in a naturally coinfected patient, and several groups have detected coinfection of avian H7N9 and human H1N1, H3N2, and influenza B viruses in human patients (14–17).

In our study, we sought to (i) determine whether reassortant viruses could be recovered from clinical material from individuals naturally coinfected with different H1N1 viruses and (ii) assess the relative fitness of these reassortants compared to parental A/H1N1 and pdm/H1N1.

MATERIALS AND METHODS

Cell culture.

Madin-Darby canine kidney (MDCK) cells were obtained from the American Type Culture Collection and maintained in minimal essential medium (MEM; Gibco, Grand Island, NY) plus 5% fetal calf serum, antibiotic-antimycotic medium (Gibco), vitamins (Gibco), and l-glutamine (Gibco). MDCK-SIAT1 cells were maintained as described previously (18). To differentiate normal human bronchial epithelial (NHBE) cells, inserts were coated with rat tail collagen type I (BD Bioscience, San Diego, CA) as previously described (19). NHBE cells from one donor and purchased from Lonza (Atlanta, GA) were expanded, frozen, resurrected, and then seeded onto 12-well inserts (PET membranes; Corning, Lowell, MA) and cultured in bronchial epithelial basal medium (Lonza) plus SingleQuots (Lonza) additives, with freshly prepared l-retinoic acid (Sigma, St. Louis, MO) at a final concentration of 5 × 10⁻⁸ M for 2 weeks until complete confluence as previously described (19). The cells were then removed to the air-liquid interface until the production of mucin and appearance of ciliated cells. Cells were washed on the apical side with phosphate-buffered saline (PBS; Corning Cellgro, Manassas, VA) every 2 to 4 days to remove mucin.

Recovery of reassortants.

Clinical material from patient 1 (8) was plaqued directly onto MDCK cells as described previously (20) with modifications. Briefly, MDCK cells were grown to confluence in six-well tissue culture plates and washed twice with warm PBS. Clinical material was serially diluted in infection medium (MEM [Gibco], 0.3% bovine serum albumin, antibiotic-antimycotic medium, vitamins, and l-glutamine), and 500 μl was added to the cells. The cells were incubated at 37°C and 5% CO₂ for 1 h and inoculum aspirated, and an agar overlay was applied (SeaPlaque agarose [Lonza] in infection medium plus 1 μg/ml TPCK [tolylsulfonyl phenylalanyl chloromethyl ketone]-trypsin [Worthington Biochemical Corp., Lakewood, NJ]). Plaques were picked after 72 h, placed into infection media, and propagated for one passage in MDCK-SIAT cells. M-gene-positive specimens (as determined by real-time reverse transcription-PCR [rRT-PCR] test for M gene using World Health Organization [WHO]-recommended primers and probe) were further screened for the presence of seasonal H1 or pandemic H1 (21). The plaque-passage cultures of either pandemic or seasonal hemagglutinin (as determined by rRT-PCR) were genotyped using a PCR-based assay previously described (22). Six reassortants of different genotypes were selected for further study and underwent two further plaque purifications using MDCK cells. The rRT-PCR and genotyping tests were repeated, and complete sequences were generated using an Illumina platform and Sanger sequencing (23–25). In addition, one A/H1N1 and one pdm/H1N1 virus were isolated from New Zealand patients in 2009 during the same study. The following virus virus abbreviations are used throughout the present study: A/New Zealand/3958/2009 (A/H1N1/3958), A/New Zealand/2047/2009 (pdm/H1N1/2047), A/New Zealand/1212a/2009 (A/H1N1/1212a), A/New Zealand/1212b/2009 (A/H1N1/1212b), A/New Zealand/1212c/2009 (A/H1N1/1212c), A/New Zealand/1212d/2009 (A/H1N1/1212d), A/New Zealand/1212e/2009 (A/H1N1/1212e), and A/New Zealand/1212f/2009 (A/H1N1/1212f).

Genotyping of viruses.

Virus samples were genotyped initially as described in Ducatez et al. 2010 followed by more specific genotyping using the longer primer pairs (Table 1) under the same RT-PCR conditions (22).

TABLE 1.

Primers used for new genotyping assay in NZ reassortment analysis

Segment and primer	Sequence (5′–3′)	Length (nt)a	Tm (°C)	Amplicon (nt)
M
MseasFOR 301-328	CGAAAGCTTAAGAGGGAGATAACATTCC	28	64.6	350
MseasREV 629-654	GGCTCTCATTGCCTGCACCATCTGCC	26	70.9
MpdmFOR 267-291	CCCGAACAACATGGATAGAGCAGTT	25	64.6	530
MpdmREV 766-794	AGATCCCAATGATATTTGCTGCAATGACG	29	64.6
NS
NSseasFOR 71-93	CAAGATCTAGGCGATGCCCCC	21	66.5	590
NSseasREV 637-660	CGTTTCTGTGTTGTAGTGAATGG	23	61.0
NSpdmFOR 60-81	GCGATTTGCAGACAATGGATTG	22	61.0	240
NSpdmREV 282-300	TCGTGACATCTCCTCGAGG	19	62.3
NP
NPseasFOR 135-159	CGAGCTTAAGCTCAATGATTATGAG	25	61.3	930
NPseasREV 1049-1068	CCCCCTTGGAAGTACTCTTG	20	62.5
NPpdmFOR 423-447	CAACCTGAATGATGCCACATATCAG	25	62.9	500
NPpdmREV 921-942	GCTGACCACTTGGCTGTTTTGG	22	64.5
PB2
PB2seasFOR 18-44	GCTAAGGAATTTGATGTCACAATCTCG	27	63.1	300
PB2seasREV 296-316	CCACTGGTCCATTTCTGTTCC	21	62.6
PB2pdmFOR 546-571	GCTGGCAATAACAAAAGAGAAGAAAG	26	61.4	520
PB2pdmREV 1146-1164	CCCGCTTACTATCAACTGG	19	60.2
PB1
PB1seasFOR 513-531	GGAGTCGATGGACAGAGGC	19	64.5	220
PB1seasREV 712-735	TGGGGTTGCAATTGCTCTGCG	21	64.5
PB1pdmFOR 297-322	CCCAGGAATATTTGAGAATTCATGCC	26	63.0	730
PB1pdmREV 1005-1023	GGGTGCCATGCTCAGGATG	19	64.5
PA
PAseasFOR 960-978	CACTGTTGTCAAGCCACAC	19	60.2	525
PAseasREV 1485-1506	GCCGTACAAATTGGTCTTTCTC	22	60.8
PApdmFOR 153-171	CCATTTCATCGACGAACGG	19	60.2	220
PApdmREV 345-370	GTGTTACTCCAATTTCAATGAACCGG	26	63.0
NA
NA-seas-226-43F	GCTGGAGAGGACAAAACG	18	59.9	650
NAseasREV 856-876	GCATACACACATCACTATGCC	21	60.6
NA-seas-226-43F	GCTGGAGAGGACAAAACG	18	59.9	540
NApdmREV 745-765	GATCTTGTATGAGGCCTGTCC	21	62.3

^ant, nucleotides.

Sequence analysis.

Sequences were analyzed using MEGA 5.1 (26) and the Influenza Research Database (27) online toolkit: for the comparison of amino acid sequences with consensus sequences of seasonal and pandemic H1N1 viruses, sequences of pdm/H1N1 (human isolates, April to June 2009, retrieved 16 October 2013) and sequences of A/H1N1 (human isolates, June 2008 to June 2009, retrieved 16 October 2013) were aligned by segment using the multiple sequence comparison by log expectation algorithm and consensus nucleotide sequences generated for each segment and each lineage (single nucleotide polymorphism analysis). Overall, 309 to 310 sequences were used to create a consensus for each segment of A/H1N1, and 617 to 658 sequences were used to create a consensus for each segment of pdm/H1N1. Consensus sequences of Oceania 2007-2009 isolates were generated in the same manner. These consensus sequences were utilized to identify point mutations in the viruses of interest using MEGA 5.1 and CLUSTALW alignment of nucleotides, followed by conversion into protein sequence.

Growth kinetics in NHBE cells.

Fully differentiated NHBE cells were washed extensively with PBS to remove mucin and then incubated on the apical side with virus at a multiplicity of infection (MOI) of 0.1 in 500 μl of infection medium for 1 h at 37°C and 5% CO₂. The inoculum was aspirated, and the cells were washed with PBS before their return to the air-liquid interface. Progeny virus was recovered at 1, 12, 24, 36, 48, and 72 h postinoculation by washing the apical side of the cells with 500 μl of infection medium for 30 min before removal of the infection media and return to the air-liquid interface. Virus titers were determined on MDCK cells as 50% tissue culture infectious doses (TCID₅₀)/ml by serial dilution of nasal washes or of supernatants from NHBE cells, as previously described (28). Briefly, 96-well tissue culture plates were seeded to 100% confluence with MDCK cells and washed once with warm PBS. The cells were inoculated with 100-μl serial dilutions of sample in 1 μg/ml TPCK-trypsin for 72 h. Hemagglutinin (HA) titers were determined with 0.5% turkey red blood cells (29, 30). Data visualization and statistical analysis was carried out in Prism 6.0b for Mac OS X (GraphPad Software, San Diego, CA) using one-way analysis of variance (ANOVA) of means and Bonferroni's multiple-comparison test.

Inoculation of ferrets.

Nine- to fifteen-week-old ferrets (Triple F Farms, Sayre, PA) determined to be seronegative for A/H1N1, A/H3N2, influenza B, and pdm/H1N1 by HI assay were housed in the St. Jude Children's Research Hospital Animal Resource Center. Food and water were made available to the ferrets ad libitum. Experiments were carried out according to the rules of the Institutional Animal Control and Use Committee, the policies of the National Institutes of Health, and the regulations of the Animal Welfare Act. For inoculations, ferrets were anesthetized with isoflurane (Baxter, Deerfield, IL) and inoculated intranasally with 10⁶ TCID₅₀ of the indicated viruses in a total volume of 1 ml. The ferrets were monitored at least once a day, and nasal washes, weights, and temperatures (subcutaneous transponder; BioMedic Data Systems, Inc., Seaford, DE) were taken every 2 days. For the nasal washes, the ferrets were anesthetized with ketamine (50 mg/kg), and the washes were collected using 1 ml of PBS. Prior to storage at −80°C, bovine serum albumin (7.5%; Sigma, St. Louis, MO) was added to the nasal washes at a dilution of 1/20 (vol/vol). The donor ferrets were inoculated and housed separately from naive ferrets for 24 h. Naive ferrets were then brought either in direct contact (same cage) or in aerosol contact (cage divided by two, offset, perforated metal sheets) with the donor ferrets. Aerosol-contact ferrets were handled prior to direct contact and donor ferrets when taking nasal washes, temperature, and weight. Ferrets were euthanized using ketamine-xylazine (50 and 5 mg/kg, respectively; AnaSed), followed by Euthasol (1 ml/ferret).

HA inhibition assays.

Ferret sera were treated with receptor-destroying enzyme (Denka Seiken, Japan) according to the manufacturer's instructions, and HA inhibition (HI) assays carried out according to standard methods using homologous antigens and 0.5% turkey red blood cells (http://www.who.int/csr/resources/publications/influenza/whocdscsrncs20025rev.pdf).

Accession number(s).

The full genome sequences for six viruses (A/H1N1/3958, A/H1N1/1212a, A/H1N1/1212b, A/H1N1/1212c, A/H1N1/1212d, A/H1N1/1212e, and A/H1N1/1212f) have been deposited in GenBank under accession numbers KJ130145 to KJ130208.

RESULTS

Reassortant viruses were recovered from human clinical material.

During the first wave of pdm/H1N1 in the Southern hemisphere country New Zealand, 11 patients were identified as naturally coinfected with already circulating oseltamivir-resistant A/H1N1 and newly introduced pdm/H1N1 (8). In the present study, we plaque purified the clinical material from one coinfected patient using MDCK cells. The patient was a 29-year-old female of European ethnicity who did not receive the seasonal influenza vaccine in 2009 (8). The patient presented to her general practitioner on 14 June 2009 reporting rapid onset of influenza-like illness (on 14 June 2009) with lethargy, malaise, cough, coryza, and some difficulty breathing. There were no reported preexisting conditions for this patient. She presented with a fever of 38.7°C, a blood pressure of 124/70, and a respiratory rate of 24/min. A diagnostic sample was taken, and oseltamivir (Tamiflu) was prescribed (8), which was started within 24 h of symptom onset. The patient was also prescribed paracetamol, ibuprofen, and salbutamol. The general practitioner noted that the patient was the most severe case seen by her that year and that hospitalization was considered but not carried out. The patient improved quickly after the start of treatment and recovered fully without complications.

Clinical material from this patient (patient 1212; case 1 in Peacey et al. [8]) was serially diluted and plaqued on MDCK cells. Approximately 72 h later, 96 plaques were picked and grown in MDCK-SIAT1 cells. These 96 samples were then tested by specific rRT-PCR (M gene) for the presence of influenza A. Of 96 samples, 85 (89%) were found to contain detectable levels of influenza A virus (M-gene positive; cutoff at threshold cycle [C_T] = 40). These 85 influenza virus samples were then tested by two rRT-PCR runs, one specific for seasonal H1 and one specific for pandemic H1 (WHO reference). Of the 85 influenza virus-positive plaques, 74 contained only seasonal H1 (87%), and 1 contained only pandemic H1 (1.2%). Nine samples contained detectable levels of both seasonal H1 and pandemic H1 (11%) (Table 2).

TABLE 2.

Summary of genotyping results for 96 picked plaques for patient 1212

Category	No. of plaques in group	% Total
		Expt 1	Expt 2	Expt 3
Plaques picked	96	100
M gene positive	85	88.5	100
Seasonal plus pandemic HA	9		10.6
Only pandemic HA	1		1.2
Only seasonal HA	74		87.1
Samples with single HA	75			100
Samples with several internal genotypes	51			68
Samples with one genotype	21			28
Pandemic H1N1	1			1.3
Seasonal H1N1	5			6.7
Reassortants with one genotype	15			20
Any samples with reassortants	69			92
Unclear samples	3			4

The 75 plaque isolates of one HA type were partially genotyped by a PCR-based method as described previously (22) to determine whether any reassortants had been generated. Partial genotyping indicated that >75% of the isolates contained a reassortant genotype or possibly a mixed virus population (Table 2). Six partially genotyped samples were chosen due their already apparent reassortant genotypes that were different from each other. We further purified these six samples by two rounds of plaque purification on MDCK cells. These six reassortant viruses used in subsequent experiments and animal challenges had the following passage history (where “Pl” indicates plaquing and “P” indicates passage): Pl_(MDCK) P_(MDCK-SIAT) Pl_(MDCK) P_(MDCK) Pl_(MDCK) P_(MDCK). The six viruses were then sequenced, and the reassortant genotypes were confirmed.

We also genotyped the remaining viruses containing only one HA type by two genotyping assays: the first assay (22) was found to give multiple double bands, and therefore a more stringent second assay was developed with longer primers and higher primer melting temperatures (Table 1). The genotypes were determined for the remaining plaques. We found among the 75 samples with a single HA type (by rRT-PCR) 51 mixed virus populations and 21 samples with single detectable genotypes. Of these 21 samples, 1 virus was of fully pandemic genotype, 5 were of fully seasonal genotype, and 15 were of reassortant genotype. Three samples remained unclear in the genotype due to limited initial RNA-extraction material from the original plaques (Table 2; see Table S1 in the supplemental material).

Reassortants contain expected molecular markers and several additional point mutations.

Six reassortants of different genotypes were chosen for further analysis (Table 3). Sequence analysis revealed the presence of known molecular markers of antiviral resistance (reviewed in reference 31). All reassortants possessed the seasonal neuraminidase gene segment and contained the marker H275Y for oseltamivir resistance. The adamantane resistance marker in the M2 protein, 31N, was present in the reassortants containing the pandemic M gene (A/H1N1/1212d, A/H1N1/1212e, and A/H1N1/1212f).

TABLE 3.

Genotypes of six clinical reassortants recovered from a patient^a

^aFor virus abbreviations, see Materials and Methods. Gray boxes, pandemic gene segments; white boxes, seasonal gene segments.

Virulence and transmission markers were identified in the isolates as previously described for A/H1N1 or pdm/H1N1 gene segments: the PB2 E627K marker (32) was present in the seasonal segments (A/H1N1/1212c, A/H1N1/1212e, and A/H1N1/1212f), whereas Q591R (33) was detected in the pandemic segments (A/H1N1/1212a, A/H1N1/1212b, and A/H1N1/1212d).

Five of six reassortants contained hybrid ribonucleoprotein (RNP) complexes containing different combinations of pandemic and seasonal PB2, PB1, PA, and NP, while one of the reassortants (A/H1N1/1212a) contained a pandemic RNP complex and seasonal HA, NA, M, and NS gene segments (Table 3). Hybrid RNP complexes were apparent in many of the 75 genotyped viruses (see Table S1 in the supplemental material).

In order to investigate the potential presence of compensatory mutations in the six reassortants, consensus sequences of all eight pdm/H1N1 and A/H1N1 gene segments were generated using “early” pandemic (April-June 2009) or “late” seasonal (June 2008 to June 2009) H1N1 isolates (see Materials and Methods). Ten protein sequences were used for amino acid comparisons: PB2, PB1, PA, HA, NA, NP, M1, M2, NS1, and NS2. The A/H1N1 isolate A/New Zealand/3958/2009 (A/H1N1/3958) and the pdm/H1N1 isolate A/New Zealand/2047/2009 (A/H1N1/2047), along with reference sequences from Oceania, were also compared to the reassortant viruses (Table 4). Overall, nine positions of interest were identified that were present in the reassortants but absent from the worldwide consensus sequences, the New Zealand A/H1N1 and pdm/H1N1 viruses, or other available sequences from Oceania during this time period (Table 4).

TABLE 4.

Residue differences between reassortants and appropriate parental strains and consensus sequences for A/H1N1 and pdm/H1N1^a

^aResidues shaded in dark gray or light gray in column 1 indicate different segments, including the switch from seasonal to pandemic for ease of viewing. Red lettering in the table body indicates residue changes in reassortant gene segments that are different from the consensus sequence, or reference viruses from Oceania. Letters shaded in blue or yellow indicate a seasonal gene segment or pandemic gene segment, respectively. *, consensus sequences for seasonal H1N1 gene segments (A/H1N1) or pandemic H1N1 gene segments (pdm/H1N1) were generated as described in Materials and Methods.

Reassortants replicate in differentiated primary human bronchial epithelial cells comparable to parental strains.

Fully differentiated NHBE cells have been previously utilized to assess the relative ability of influenza A viruses to replicate in the human airway (34–36). The NHBE cells used in the present study were harvested from the bronchi of one human donor followed by culturing and differentiation in an air-liquid interface, as described in Materials and Methods.

To compare the fitness of the six selected reassortant viruses to A/H1N1 and pdm/H1N1 viruses from 2009, multistep growth curves were carried out in NHBE cells using an MOI = 0.1 inoculation dose. All six reassortants, as well as the New Zealand isolates A/H1N1/3958 and pdm/H1N1/2047, replicated to comparable levels with viral titers up to 7.8 log₁₀ TCID₅₀/ml (Fig. 1). No statistically significant differences in virus growth were detected between any of the viruses by one-way ANOVA of means and an unpaired t test. Overall, the six reassortants examined did not exhibit overt fitness deficiency or enhancement in NHBE cells compared to contemporary A/H1N1 and pdm/H1N1 isolates.

FIG 1.

Clinical reassortants replicate efficiently in differentiated NHBE cells. Cells were inoculated apically at an MOI of 0.1 for 1 h and then returned to an air-liquid interface. Virus was recovered by washing the apical cell surface with PBS at the indicated time points. TCID₅₀ values were determined on MDCK cells. Displayed are the means of three replicate wells with the standard deviations from one experiment. The virus abbreviations are as defined in Materials and Methods.

Reassortant viruses are capable of aerosol transmission between ferrets.

Ferrets are widely used to evaluate the potential transmissibility of influenza viruses in humans (reviewed in references 37 and 38). Two clinical reassortants, A/H1N1/1212e and A/H1N1/1212f, which differed in genotype were selected for further study in ferrets (Table 3).

Two ferrets per group were inoculated intranasally with one of four viruses: A/H1N1/3958, pdm/H1N1/2047, A/H1N1/1212e, and A/H1N1/1212f. All donor ferrets became productively infected as determined by virus shedding (Fig. 2 and Table 5). On day 1 postinoculation of the donor ferrets, one direct contact and one aerosol contact ferret were introduced to each donor.

FIG 2.

Clinical reassortants replicate and transmit in ferrets. (A to D) Donor ferrets were inoculated intranasally with 10⁶ TCID₅₀ of the indicated viruses and placed with direct and aerosol contact ferrets 1 day postinoculation. Shown are infectious titers from nasal washes expressed as TCID₅₀/ml.

TABLE 5.

Clinical signs, virus shedding, and seroconversion of donor and contact ferrets^a

^aFor virus abbreviations, see Materials and Methods. Weight loss is expressed as the number of ferrets/ferrets per group (maximum weight loss as a percentage of initial body weight), virus shedding is expressed as the number of ferrets/ferrets per group (days of shedding postinoculation of donor ferrets), the peak virus titer in nasal washes is expressed as the log₁₀ TCID₅₀/ml (day of peak titer), and the homologous hemagglutinin inhibition (HI) titer was determined using 0.5% turkey red blood cells.

In our experimental setting the seasonal A/H1N1/3958 virus transmitted to direct contact ferrets, with virus titers in nasal washes of contact animals detectable on day 2 postcontact (Fig. 2A and Table 5). Aerosol transmission was detectable in one contact ferret on day 2 postcontact and in the second aerosol contact ferret on day 6 postcontact. The aerosol contact animals had no direct contact with each other. Virus titers in nasal washes and duration of shedding were comparable in the donor and contact ferrets, although the second aerosol contact was still shedding on the last sampled time point (Fig. 2A and Table 5). Ferrets inoculated with pandemic pdm/H1N1/2047 displayed a similar pattern of virus titers and direct contact transmission to seasonal A/H1N1/3958; however, it did not transmit to either aerosol contacts (Fig. 2B and Table 5). Reassortants A/H1N1/1212e and A/H1N1/1212f displayed virus titers similar to those of A/H1N1/3958 and pdm/H1N1/2047 in inoculated animals and direct contacts (Fig. 2C and D and Table 5). Reassortant A/H1N1/1212e transmitted to one of two aerosol contacts with titers detectable on day 4 postcontact, whereas A/H1N1/1212f transmitted to both aerosol contact ferrets and with virus titers detectable from days 2 and 4 postcontact, respectively (Fig. 2 and Table 5).

In regard to body weight changes, one-way ANOVA of means indicated a significant difference in weight changes within the donor ferret group. Following up with unpaired t tests of means assuming equal standard deviations, ferrets inoculated with pdm/H1N1/2047 displayed significantly greater weight loss compared to ferrets inoculated with A/H1N1/3958, A/H1N1/1212e, or A/H1N1/1212f (Table 5; P = 0.0013, 0.0407, and 0.0151, respectively). Direct contact ferrets infected with A/H1N1/3958 displayed significant weight gain compared to direct contact ferrets in the other three groups pdm/H1N1/2047 (P = 0.0084), A/H1N1/1212e (P = 0.0293), and A/H1N1/1212f (P = 0.0081) (Table 5). Among the eight aerosol contact ferrets, two ferrets experienced mild weight loss: one infected with A/H1N1/3958 and one infected with reassortant A/H1N1/1212e (Table 5). Even though both aerosol contact ferrets in the reassortant A/H1N1/1212f group shed virus and seroconverted, neither displayed weight loss. The only significant difference in weight change was between the pdm/H1N1/2047 aerosol contact ferrets that never shed virus and did not seroconvert, and the aerosol contact ferrets to A/H1N1/1212e (P = 0.0345).

Overall, body temperature changes were minimal, although >1°C temperature increases were detected in several ferrets infected with A/H1N1/1212f, one ferret infected with A/H1N1/1212e, and one ferret infected with pdm/H1N1 (data not shown). Consistent with the ferrets' virus shedding, 21 of 24 ferrets seroconverted (Table 5): the two aerosol contact ferrets of the pdm/H1N1 group and one aerosol contact ferret of the reassortant A/H1N1/1212e group that had no detectable virus titers in their nasal washes did not seroconvert. In conclusion, although few animals were used in this assessment, both reassortant viruses did not display overt growth fitness differences compared to seasonal A/H1N1/3958 and pdm/H1N1/2047 and transmitted to both direct and aerosol contacts.

DISCUSSION

During the first wave of the H1N1 pandemic in New Zealand from May to September 2009, several patients became naturally coinfected with seasonal H1N1 and pandemic H1N1 viruses (8). The possibility of reassortment between pdm/H1N1 and other circulating human or animal influenza A viruses, as well as the generation of pathogenic or oseltamivir-resistant reassortants, has been raised previously (39–47). In addition, a human coinfection with pdm/H1N1 and A/H3N2 leading to reassortment has recently been reported (14).

In this study, clinical material from a patient naturally coinfected with A/H1N1 and pdm/H1N1 was plaqued on MDCK cells in an attempt to isolate reassortants from the clinical material and not through coculture in cells. Eighty-five virus-containing samples were recovered and, of these, 75 contained only one HA type (seasonal or pandemic) as determined by the highly sensitive and specific WHO-recommended diagnostic rRT-PCR method (21). Since these 75 samples no longer contain both parental HA types, the reassortants identified (internal genes) would have either (i) had to have arisen in the patient or (ii), in the case of the identified mixed virus populations, have been the result of reassortment in cell culture of reassortants already present in the clinical specimen. We cannot exclude the possibility of further reassortment in the cell culture during plaque purification and virus amplification since we did identify by PCR-based genotyping (22; this study) a number of samples containing >1 genotype in several genes. However, since only one HA type was detected by rRT-PCR, it is highly likely that some reassortment events had occurred in the patient even for those samples. A different approach to the isolation of individual viruses from the original clinical material would have been to use limiting dilutions, although complete confidence in the exclusion of any virus mixture in one round of purification is difficult to ensure and would require deep sequencing of samples to identify minor variants.

Marshall et al. previously reported that, in the absence of a segment mismatch, reassortment occurs frequently (48). Interestingly, we recovered mostly viruses containing the seasonal HA segment. It is possible that viruses with the pdmH1 segment were less fit than those with seasonal H1 in either the patient or in the plaque recovery in MDCK cells.

The relative amounts of pdmH1 and seasonal H1 may have been different in the patient sample: the Centers for Disease Control and Prevention has previously reported that the seasonal H1 and pdmH1 primer/probe panels (the rRT-PCR Flu Panel [IVD, K080570] and the rRT-PCR Swine Flu Panel, respectively) displayed a similar analytical performance with equivalent limits of detection (49, 50). Peacey et al. reported threshold cycle (C_T) values for seasonal (18.13) and pandemic HA (22.74) in the clinical sample of patient 1212 used in this study, indicating a possibly 10-fold-higher RNA copy number for seasonal H1 than pdmH1 (8). Our recovery of 87% seasonal HA in the MDCK plaquing would be consistent with this observation. However, pdmH1N1 internal gene segments were readily detected in the virus material recovered, indicating either equitable amounts of pdmH1N1 were present in the clinical sample or that reassortants had already arisen in the patient. Lastly, it has been previously shown that pdmH1N1 forms smaller plaques on MDCK cells than seasonal H1N1, indicating a possible disadvantage in our specific MDCK-cell-based plaque recovery method (43, 51).

To assess the relative fitness of six of the reassortants, two well-described model systems were utilized: differentiated NHBE cells and ferrets. NHBE cells have been shown to support growth of both A/H1N1 and of pdm/H1N1 (52). We found that all reassortants were capable of replication in differentiated NHBE cells similar to that of A/H1N1 and pdm/H1N1 strains used as benchmarks, indicating the reassortants could grow in the human respiratory system.

Two reassortants were assessed for their ability to transmit between ferrets as a surrogate in vivo model for transmission between humans. The two reassortants were of different genotypes and transmitted to direct contact ferrets and aerosol contact ferrets indicating these viruses could potentially be transmitted between humans. All of the viruses caused little to no pathogenic effect in the infected ferrets, with the pandemic strain causing the most severe weight loss in donor ferrets. Our results are consistent with previous studies of reassortants between pdm/H1N1 and A/H1N1 generated in vitro or in vivo that have reported various degrees of pathogenicity ranging from apathogenic to low pathogenic (40, 45, 52).

Overall, our study, based on clinical material from an influenza patient, clearly confirms the potential of A/H1N1 and pdm/H1N1 viruses to reassort in humans. However, the phenotypes of recovered reassortants were more consistent with the seasonal parental strain (A/H1N1), and at most intermediate in pathogenicity between A/H1N1 and pdm/H1N1 in ferrets.

None of the reassortants we describe have ever been reported from patient samples, either in New Zealand or elsewhere. The apparent lack of success of such reassortant viruses in the human population may be the result of significant levels of herd immunity against the seasonal HA protein, which appeared to be dominant in the patient sample (8) and was overrepresented in MDCK-cell-plaqued viruses. The seasonal A/H1N1 lineage circulating in New Zealand during early 2009 was replaced by the new pdm/H1N1 viruses in New Zealand from April or May onward, with no A/H1N1 detected between 2010 and 2014 (53–56). Therefore, in the context of the newly introduced pandemic virus and the New Zealand population, the seasonal-HA-containing reassortants generated in the patients may have been unable to spread past the index case. It may therefore be a testament to the high fitness of the pandemic H1N1 genotype in the human population (possibly due to immune escape) that reassortants with seasonal H1 did not become established. Overall, reassortants of pdm/H1N1 have very rarely been reported in humans (57) and most often were the result of zoonotic infection with reassortant swine viruses which already contained the pandemic H1N1 M-gene (58–60).

Several reports have described reassortants of pdm/H1N1 viruses in swine. This may indicate different influenza virus fitness dynamics in pigs, possibly due to variable herd immunity in different swine populations (age and production type) (41, 61–63). Many of the described swine influenza reassortants with pdm/H1N1 harbored swine-endemic HA and NA genes and the pdm/H1N1 matrix (M) gene, including the variant swine H3N2 viruses that have infected humans in the United States (59, 60). The pandemic H1N1 M gene was shown to increase virus transmissibility and to undergo positive selection in a pool of reassortants (59, 64–66). In the present study, all characterized reassortants contained the seasonal H1 and N1 gene segments, whereas the pdm/H1N1 M gene was present in three of six of the characterized reassortants. Overall, the pandemic M gene was identified in 16 of 46 samples that contained a single M gene by PCR-based genotyping assays (see Table S1 in the supplemental material).

Even though in our model systems the clinical reassortants exhibited a high level of fitness, incompatibilities between A/H1N1 and pdm/H1N1 reassortants generated in vitro have been reported, particularly with respect to hybrid RNP complexes (39, 47, 67–69). We therefore examined the reassortants' amino acid signatures in detail and detected seven positions with unique point mutations. Among these, the PB2 D256G mutation (present in A/H1N1/1212c) has been shown to enhance the polymerase activity of a reassortant H5N1 virus (67) and is located in a region of PB2 interacting with the NP segment (70). The point mutation PA N350K of reassortant A/H1N1/1212a is in a region of human host adaptation sites identified through bioinformatics analysis, although the function is unknown (71). Reassortant A/H1N1/1212b contains a point mutation in seasonal H1 (H137Y) that falls within the antigenic site (130-loop) of the HA and has been detected previously in pdm/H1N1 in Taiwan but did not become fixed in that virus population (72). A D272G mutation was present in the seasonal H1 of all reassortants and has been previously reported in a few A/H1N1 clade 2B viruses (73). As described by Peacey et al. (8), the 29-year-old female patient was not vaccinated, and the illness resolved without hospitalization, although an antiviral agent (oseltamivir) was administered. The patient sample was obtained on the day of symptom onset and before oseltamivir therapy was started.

The point mutation I123V in NS1 was present in both pdm/H1N1/2047 and all examined reassortants with a pandemic NS segment but not in the consensus of “early” pdm/H1N1 viruses. Nelson et al. reported this mutation as unique to clade 7 pdm/H1N1 viruses and to have occurred early during the evolution of pdm/H1N1 (April 2009) (74). Overall, these reassortants seem to have undergone only limited drift, which serves as additional evidence of their recent emergence. Further studies on the point mutations identified in our reassortants could reveal novel means for A/H1N1 and pdm/H1N1 to overcome incompatibilities during reassortment and could further add to the understanding of the basis of influenza A and pdm/H1N1 reassortment in humans.

Conclusions.

The recovery of multiple reassortant influenza A viruses from a naturally coinfected patient is described. Pandemic and seasonal H1N1 viruses cocirculating in New Zealand in 2009 were compatible and produced a variety of reassortant genotypes. Several of the reassortants were assessed and showed high fitness in models for human transmission comparable to parental A/H1N1 and pdm/H1N1.

Funding Statement

This study was supported by contract HHSN266200700005C (Centers of Excellence for Influenza Research and Surveillance [CEIRS]) from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, by American Lebanese Syrian Associated Charities (ALSAC), and by the New Zealand Ministry of Health.