In Vivo and In Vitro Alterations in Influenza A/H3N2 Virus M2 and Hemagglutinin Genes: Effect of Passage in MDCK-SIAT1 Cells and Conventional MDCK Cells

Abstract

No mutations were detected in the hemagglutinin gene of influenza A/H3N2 virus isolates from patients undergoing short-term amantadine treatment. However, genetic changes occurred after serial passage in either MDCK or MDCK-SIAT1 cells. Our results showed that only a few mutations were observed in MDCK-SIAT1-passaged isolates in the presence of amantadine.

Viral resistance to amantadine emerged quickly in vivo when patients received amantadine for the treatment of influenza A virus infection (3, 5). The genetic basis of resistance to amantadine is associated with amino acid substitutions in the transmembrane region of the M2 gene (2). We documented a high frequency of amantadine-resistant influenza A/H3N2 virus isolates with a Ser-31-Asn mutation in the M2 gene and dual changes in the hemagglutinin (HA) gene at residues 193 and 225 (clade N lineage) (10). However, little is known about whether the HA changes (i) were synergistic with those occurring in the M2 gene in response to drug selection pressure or (ii) occurred separately and were randomly associated with fitness-improving mutations. To address this question, we analyzed changes in the M2 and HA genes of influenza A viruses from clinical samples and from the same isolates that were serially passaged in Madin-Darby canine kidney (MDCK) cells or SIAT1-transfected MDCK cells in the presence or absence of amantadine. The MDCK cell line is commonly used for influenza virus isolation, while the MDCK-SIAT1 cell line is a recently introduced MDCK variant characterized by an overexpression of sialyl-α2,6-galactose that improves the binding of human influenza A virus to the cell receptor (6).

Nasopharyngeal swabs were collected from patients with an influenza-like illness who visited a pediatric outpatient clinic in Niigata City, Japan, from 2000 to 2002. Samples were collected at the first visit and at the second visit after 3 to 5 days of amantadine treatment. One hundred microliters of each sample was inoculated into MDCK cells for virus isolation. Antigenic characterization was performed by hemagglutination inhibition test (12). Influenza A viruses were screened for amantadine susceptibility by the 50% tissue culture infective dose/0.2-ml method (5) and verified by M2 gene sequencing of the transmembrane region to a detect mutation at position 26, 27, 30, 31, or 34 that confers resistance (12). After screening, influenza A/H3N2 viruses that were originally amantadine sensitive and became amantadine resistant after drug treatment were selected and analyzed for this study (in vivo).

MDCK-SIAT1cells (kindly donated by Mikhail Matrosovich, Institute of Virology, Philipps University, Marburg, Germany) were passaged as described elsewhere (6). The selected parental amantadine-sensitive strains were inoculated into MDCK cells or MDCK-SIAT1 cells and sequentially passaged 10 times in the presence or absence of amantadine at a final concentration of 2.0 μg/ml. The viruses were analyzed after the 3rd and 10th passages in the absence or presence of amantadine in MDCK or MDCK-SIAT1 cells (in vitro). Viral RNA extraction and cDNA synthesis were performed as described elsewhere (1). After amplification of the M2 and HA genes, direct sequencing of PCR products was performed with an ABI 3100 DNA sequencer (11). The transmembrane region of the M2 channel protein and the coding regions of the HA1 domain (amino acid residues 1 to 329) and the HA2 domain (amino acid residues 1 to 208) were analyzed. No virus plaque purification and cloning of PCR products was performed prior to sequencing.

Our in vivo study showed that all of the virus isolates (n = 7) obtained from patients treated with amantadine possessed M2 gene mutations, whereas HA changes were not observed after amantadine treatment (Table 1).

TABLE 1.

Sequence analysis of H3N2 variants selected in vivo and in vitro

Patient	In vivo				In vitrod
					MDCK cells								SIAT1-MDCK cells
	Virus from patient before treatment		Virus from patient after amantadine treatment		Without amantadine				With amantadine				Without amantadine				With amantadine
					3rd-passage virus		10th-passage virus		3rd-passage virus		10th-passage virus		3rd-passage virus		10th-passage virus		3rd-passage virus		10th-passage virus
	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA	M2	HA
1	NDa	ND	S31N	—b	ND	R229G	ND	R229G	ND	P221L	S31N	S137Y, P221L	ND	—	ND	—	ND	—	S31N	—
2	ND	ND	S31N	—	ND	R229I	ND	R229I	ND	R229I	S31N	R229I	ND	—	ND	—	ND	—	S31N	—
3	ND	ND	A30T	—	ND	R220G	ND	R220G	ND	R229K, I236V	S31N	R229K, I236V	ND	—	ND	—	ND	—	L26F	—
4	ND	ND	V27A	—	ND	R229K	ND	R229K	ND	R229K	V27A	R50G, V106A, D144N, G186S, S199P, R229I, S247C, P273S,	ND	—	ND	—	ND	—	A30V	—
5	ND	ND	S31N	—	ND	V226I	ND	V226I	ND	V226I	S31N	V226I	ND	—	ND	—	ND	—	V27G	—
6	ND	ND	S31N	—	ND	—	ND	R83K, H183L, A147Tc	ND	—	V27A	R83K, H183L, A147Tc	ND	—	ND	R83K, A147Tc	ND	—	S31N	R83K, A147Tc
7	ND	ND	S31N	—	ND	H183L	A30T	H156N, T147A,c G150Ec	ND	H183L	S31N	H156N, T147A,c G150Ec	ND	—	ND	—	ND	—	A30V	H156N, T147Ac G150Ec

^aND, mutation not detected.

^b—, No mutation was detected compared to the virus collected from the patient at the first clinic visit.

^cSubstitutions which are located within the HA2 subunit.

^dViruses used in the in vitro study were those obtained from patients before amantadine treatment.

Our in vitro study showed that all of the viruses developed M2 gene mutations after 10 passages in both MDCK and MDCK-SIAT1 cells in the presence of amantadine, but the M2 mutation sites differed between the two cell lines for five of the isolates (Table 1). On the other hand, no mutations were observed in amantadine-free cultures after 10 passages, except for one isolate that showed an A30T substitution in M2 (conferring amantadine resistance) when grown in MDCK cells.

Analysis of the HA gene showed that six out of seven isolates developed mutations after three passages in MDCK cells, and eventually all of the isolates showed mutations in the HA1 and HA2 domains after 10 passages in the presence or absence of amantadine (Table 1). Most of the HA mutation sites and the type of amino acid substitutions were similar between the isolates passaged with or without amantadine and between 3rd- and 10th-passage isolates. However, the number of mutation sites increased after the 10th passage when cells were cultivated in the presence of amantadine.

HA changes were not observed in viruses after the 3rd passage in MDCK-SIAT1 cells and occurred in only one or two viruses after the 10th passage without or with amantadine, respectively (Table 1). The HA mutation sites and types of amino acid substitutions found were similar to those found in MDCK-passaged viruses but were far less numerous.

Our clinical surveillance results suggested that the short-term amantadine treatment did not drive HA mutations, and thus, the appearance of clade N was due to the combined events of reassortment (7, 13), amantadine-driven point mutations in the M2 gene, and fitness-improving mutations in the HA gene (11).

The HA gene was more variable in MDCK-passaged viruses than that in their MDCK-SIAT1-passaged counterparts. Mutations in amantadine-free culture accumulated mostly at the receptor-binding (RB) pocket in the HA1 subunit (residues 156, 220, and 229) (14, 15), whereas changes in the HA2 subunit did not have any functional significance. We assume that frequent mutations near RB sites can be attributed to adaptation of the viruses to MDCK cells, which express larger amounts of NeuAcα2,3Gal and smaller amounts of NeuAcα2,6Gal than do human airway epithelial cells (9). Egg-adapted influenza viruses also showed high specificities of binding to NeuAcα2,3Gal, presumably resulting from key amino acid changes at the RB site (at position 226) (4). On the other hand, MDCK-SIAT1 cells, engineered to overexpress NeuAcα2,6Gal, led to fewer HA mutations and could thus be more reliable for multipassage analysis of human influenza virus. Similar results for MDCK-SIAT1-passaged viruses were recently demonstrated by Oh et al., but different amino acid substitutions in the HA1 subunit were observed (8).

In our study, amantadine drove more HA1 mutations in both cell lines (more stable in MDCK-SIAT1 cells) than under amantadine-free conditions, and most of them were key amino acids at or near RB sites (positions 144, 156, 183, 186, 199, 220, 221, 226, 229, and 236) (13-15). Thus, amantadine may contribute to the appearance of a mutated HA gene with an RB site property altered by an unknown mechanism. While our clinical study demonstrated that amantadine did not affect the HA gene in the short term, the in vitro results suggested the possibility that longer exposure to the drug affects the HA gene in vivo, since MDCK-SIAT1 cell-passaged viruses also underwent changes in the HA gene in the presence of amantadine.

Given the small number of samples tested, our results elucidated the differences in the HA gene changes seen in vivo and in vitro in a comparison of two cell lines during the development of amantadine resistance. These results suggested that careful interpretation is needed after consecutive passages in MDCK cells, but MDCK-SIAT1 cells are more favorable for analysis of the RB domain and the molecular epidemiologic phylogeny of the HA gene.