Increased virulence of neuraminidase inhibitor-resistant pandemic H1N1 virus in mice

Abstract

Pandemic H1N1 2009 (A[H1N1]pdm09) variants associated with oseltamivir resistance have emerged with a histidine-to-tyrosine substitution in the neuraminidase(NA) at position 274 (H274Y). To determine whether the H274Y variant has increased virulence potential, A(H1N1)pdm09 virus, with or without the H274Y mutation, was adapted by serial lung-to-lung passages in mice. The mouse-adapted H274Y (maCA04_H274Y) variants showed increased growth properties and virulence in vitro and in vivo while maintaining high NA inhibitor resistance. Interestingly, most maCA04_H274Y and maCA04 viruses acquired common mutations in HA (S183P and D222G) and NP (D101G), while only maCA04_H274Y viruses had consensus additional K153E mutation in the HA gene, suggesting a potential association with the H274Y substitution. Collectively, our findings highlight the potential emergence of A(H1N1)pdm09 drug-resistant variants with increased virulence and the need for rapid development of novel antiviral drugs.

Keywords: :

• pandemic influenza
• drug resistance
• mouse adaptation
• virulence
• neuraminidase

The pandemic H1N1 2009 (A[H1N1]pdm09) virus has begun to act as a seasonal influenza since it caused an influenza outbreak in 2009. Before the pandemic, especially during the 2007–2008 season, most seasonal H1N1 viruses had a neuraminidase (NA) gene with a substitution of tyrosine for histidine at residue 274 (H274Y, N2 numbering) which confers resistance to oseltamivir (OS),^1,2 the most widely used drug for clinical therapy of influenza-infected patients, albeit mostly with compromised viral fitness.^3,4 Fortunately, the A(H1N1)pdm09 virus remained sensitive to neuraminidase inhibitors (NAIs) such as OS. However, the widespread use of NAIs to treat patients that are believed to be infected with A(H1N1)pdm09 in many countries led the virus to evolve into drug-resistant variants.⁵ Furthermore, unlike previous seasonal H1N1 drug-resistant variants of A(H1N1)pdm09 retained comparable viral fitness in pathogenesis and transmissibility to those of their OS-sensitive counterparts.^6,7

A(H1N1)pdm09 virus has exhibited remarkable transmissibility among humans and even reverse-zoonotic transmission, such as to pigs and turkeys,^8,9 but the pathogenicity of the virus was milder than that of the pandemic virus in 1918. However, several studies have shown increased pathogenesis of A(H1N1)pdm09 in new host species through mouse adaptation and molecular-based substitutions related to enhanced virulence,^10-12 although the tested viruses still remained NAI-sensitive. Of note, in two of these studies, the mouse-adapted variants also showed enhanced fitness in ferrets,^11,12 suggesting mice are a relevant model for such studies.

The virulence potential of A(H1N1)pdm09 variants with drug-resistance mutations in the NA gene has been poorly understood. Increasing numbers of cases of oseltamivir (OS)-resistant variants, especially associated with the H274Y mutation, have been reported to the World Health Organization,¹³ implicating a potential threat of virulent and drug-resistant variants. Although several studies demonstrated that OS-resistant A(H1N1)pdm09 viruses are as virulent as their OS-sensitive virus counterparts,^6,7 the viruses showed basically mild virulence in vivo. Here, the virulence potential of drug-resistant A(H1N1)pdm09 variants conferred by the H274Y mutation in the NA gene (CA04_H274Y) was investigated, along with wild-type California/04/09 (CA04) virus in mice. Furthermore, to determine whether the CA04_H274Y can become more virulent than its CA04 virus counterpart, parallel mouse adaptation of these OS-resistant and OS-sensitive parental strains was performed by serial lung-to-lung passage for five times. Both viruses were generated by reverse genetics and the H274Y mutation was given by site directed mutagenesis in CA04 virus. The research protocol for the use of mice in this study was in strict accordance and adherence with relevant policies regarding animal handling as mandated under the Guidelines for Animal Use and Care of the Korea Center for Disease Control.

Initially, 10^4.5 TCID₅₀ in 30 µl were administered intranasally (i.n.) to groups of four BALB/c mice after light anesthesia with a mixture of Zoletil (VirBac) and Rompun (Bayer). Lungs from infected mice were collected 5 days (d) post-inoculation (p.i.), and the supernatant of the homogenate was administered to naïve mice for the subsequent four additional passages. Finally to compare pathogenicity, 10^4.0 TCID₅₀ of the mouse-adapted viruses and their parental strains (maCA04_A–D and maCA04_H274Y_A–D vs. CA04 and CA04_H274Y, respectively), were administered in groups of five mice and were monitored for morbidity and mortality for 12 d p.i. If the infected mice lost more than 30% of their body weight, they were euthanized. Three of four independently passaged, mouse-adapted CA04 viruses (maCA04) were 100% lethal, but the other one had a 60% survival rate starting at 7 d p.i. (Fig. 1A and B). All 4 independently passaged, mouse-adapted CA04_H274Y viruses (maCA04_H274Y) killed all infected mice within 7 d p.i. (Fig. 1C and D). These findings show that the OS-resistant variants can cause mortality and morbidity equivalent to that of the mouse-adapted wild-type A(H1N1)pdm09 virus after serial lung-to-lung passages.

Figure 1. Comparison and characterization of virulence between mouse-adapted and parental viruses in BALB/c mice. Five independent serial lung-to-lung passages for each virus (CA04 and CA04_H274Y) were performed. After adaptation, 10^4.0 TCID₅₀ of the passaged viruses and their parental strains were administered intranasally (i.n.). Body weight loss and survival rate of the parental CA04 and CA04_H274Y with their corresponding mouse-adapted variants (maCA04_A, maCA04_B, maCA04_C, maCA04_D, and maCA04_H274Y_A, maCA04_H274Y_B, maCA04_H274Y_C, and maCA04_H274Y_D, respectively) were monitored in groups of 5 mice for 12 d p.i. (A–D). The number of passages is represented by ‘P’ followed by the name of the individually passaged viruses on the figure. For further characterization of lethal mouse-adapted oseltamivir-resistant variants, mouse-adapted viruses were plaque-purified and selected. Next, 10^3.0 TCID₅₀ of plaque-purified mouse-adapted CA04, mouse-adapted CA04_H274Y, and their parental viruses were administered to mice i.n. Body weight loss (E) and survival rate (F) were monitored daily for 12 d p.i. Groups of 12 mice were also given 10^3.0 TCID₅₀, and lung samples were collected 1, 3, 5, and 7 d p.i. to determine lung viral titers (3 heads/day) (G). Standard hematoxylin and eosin staining of lung tissues infected by CA04 (H), CA04_H274Y (I), mouse-adapted CA04 (J), and mouse-adapted CA04_H274Y (K) was examined 5 d p.i. (magnification × 100). *P < 0.05 (unpaired t test, two-tailed) of viral lung titers between wild-type CA04 and CA04_H274Y viruses and mouse-adapted CA04 and CA04_H274Y viruses. Error bar shown in (C, E, and G) represents standard error mean (SEM).

To confirm the genetic stability of the conferred H274Y mutation in the NA gene and examine whether any additional mutations occurred in the viral genome during adaptation, the whole genome sequences of 8 mouse-adapted viruses were analyzed and compared with their parental viruses. We found 10 amino acid substitutions in 6 genes of the 8 mouse-adapted viruses during adaptation (Table 1); one synonymous silent mutation (G1068C) was additionally noted within the corresponding H1 HA2 protein region of the maCA04 viruses (data not shown). The conferred H274Y mutation in the NA gene of the CA04_H274Y virus was retained after mouse adaptation in all 4 independent parallel passages, showing that the OS resistance–inducing mutation did not create genetic instability or need other compensatory mutations in the NA gene to increase virulence. Three mutations (S183P and D222G in HA [H1 numbering] and D101G in NP) were almost synonymously found in our mouse-adapted CA04 (maCA04_A–C) and CA04_H274Y (maCA04_H274Y_A–D) viruses (Table 1), and were also correspondingly noted in previous mouse-adaptation studies.^10-12 The maCA04_D virus did not kill all mice and retained the wild-type sequence of 222D in the HA gene (Fig. 1B and Table 1). The D222G mutation in HA gene has been associated with severe-to-fatal cases of human A(H1N1)pdm09 infections and lethal swine H1N2 virus infection in ferrets.^14,15 Interestingly, all maCA04_H274Y viruses, but not maCA04 viruses, acquired a synonymous K153E mutation in the HA gene, suggesting a potential association with the H274Y mutation. On the other hand, maCA04_C virus had further mutations in PB1 (N105T and R721K) and HA (K119E) while maCA04_H724Y also incurred a S714R substitution in PB2. The E158G and T97I mutations found in PB2 and PA, respectively, of several mouse-adapted viruses in our study have been reported to be associated with increased polymerase activity or virulence.^11,16,17

Table 1. Amino acid substitutions identified after mouse adaptation of pandemic H1N1 2009 and oseltamivir-resistant variants

Virus strain	Amino acid residue at position no.
	PB2		PB1		PA	HA^a				NP	NA^b
	158	714	105	721	97	119	153	183	222	101	274
CA/04/09	E	S	N	R	T	K	K	S	D	D	H
maCA04_A	E	S	N	R	T	K	K	P	G	G ^c	H
maCA04_B	E	S	N	R	I	K	K	P	G	G	H
maCA04_C	E	S	T	K	T	E	K	P	G	G	H
maCA04_D	E	S	N	R	T	K	K	P	D	G	H
CA/04/09H274Y	E	S	N	R	T	K	K	S	D	D	Y
maCA04H274Y_A	G	S	N	R	T	K	E	P	G	G	Y
maCA04H274Y_B	E	R	N	R	T	K	E	P	G	G	Y
maCA04H274Y_C	E	S	N	R	T	K	E	P	G	G	Y
maCA04H274Y_D	G	S	N	R	T	K	E	P	G	G	Y

^a The number of the amino acid residue by H1 numbering. ^bThe number of the amino acid residue by N2 numbering. ^cThe boldface represents substitution of parental amino acid.

For further characterization of the virulent phenotype of OS-resistant variants, maCA04_H274Y_C and maCA04_A viruses, which only differ in HA₁₅₃ (E or K, respectively) were plaque-purified and selected as described elsewhere,¹⁶ and tested for morbidity, mortality, and viral lung titer. Groups of 5 mice were inoculated i.n. with 10^3.0 TCID₅₀ of the parental and plaque-purified mouse adapted viruses (CA04, CA04_H274Y, maCA04, and maCA04_H274Y) and monitored for survival and body weight loss daily for 12 d. Similar to the results of infection with the passaged viruses in lungs, mice infected with the plaque-purified mouse-adapted viruses (maCA04 and maCA04_H274Y) had dramatic weight loss and succumbed to infection within 9 d p.i. (Fig. 1E and F) although maCA04_H274Y killed all inoculated mice at 7 dpi. Viral replication kinetics in mouse lungs (n = 12) revealed that the mouse-adapted viruses yielded titers more than 10-fold higher than those of parental viruses at 1 and 3 d p.i., and no difference was observed between maCA04 and maCA04_H274Y at any time, which implies that the adaptation of the OS-resistant H274Y variant in mice increased growth properties to as high as that of the wild-type virus in vivo (Fig. 1G). The increased yields of mouse-adapted viruses in mouse lungs was also observed in MDCK cells and eggs, which yielded significantly higher titers (>10^1.3-fold, P < 0.05) than their parental strains (Table 2). To determine the 50% mouse lethal dose (MLD₅₀) of the viruses, we inoculated groups of 5 mice i.n. with 10-fold serial dilutions containing 10¹ to 10⁵ TCID₅₀ of the viruses. The maCA04 and maCA04_H274Y viruses showed more than 10^3.5-fold higher MLD₅₀ values,(2.0 and 1.5, respectively) than their parental viruses (>5.5 in both) (Table 2). Histopathologic analysis revealed that maCA04 and maCA04_H274Y viruses caused more severe lung tissue damage than their parental strains because intraepithelial infiltration of neutrophils and macrophages resulted in acute bronchointerstitial pneumonia at 5 d p.i. (Fig. 1H–K).

Table 2. Characteristics of growth efficiency and virulence, and neuraminidase-inhibitor susceptibility of wild-type and mouse-adapted pandemic H1N1 influenza viruses and their oseltamivir-resistant counterpart

Virus strain	Virus yield		MLD50	MST	Neuraminidase inhibitors IC50 ± SD, nM (Ratio)^a
	MDCK cells	Eggs
	(Log10TCID50/ml ± SD)	(Log10EID50/ml ± SD)	(Log10TCID50)	(dpi)	Oseltamivir	Zanamivir	Peramivir
CA/04/09	6.1 ± 0.3	6.7 ± 0.5	>5.5	>12	0.86 ± 0.03 (1)	0.6 ± 0.04 (1)	0.16 ± 0.07 (1)
maCA04	7.8 ± 0.2*	8.9 ± 0.4*	2	8	2.53 ± 0.04 (2.9)	1.99 ± 0.05 (3.3)	0.52 ± 0.04 (3.3)
CA04 H274Y	6.1 ± 0.3	7.1 ± 0.1	>5.5	>12	107.1 ± 0.07 (124.5)	0.33 ± 0.02 (0.55)	13.56 ± 0.03 (84.8)
maCA04 H274Y	7.7 ± 0.1*	8.4 ± 0.3*	1.5	6	128.1 ± 0.04 (149)	0.48 ± 0.05 (0.8)	27.11 ± 0.05 (169.4)

^a Compared with that of the wt CA/04/09 virus. *P< 0.05 compared with the value for the respective parental viruses (unpaired t test, two-tailed). SD, standard deviation; MLD, mouse lethal dose; MST, median survival time; wt, wild type; ma, mouse adapted.

To determine whether the CA04_H274Y variant and its adapted counterpart were resistant to NAIs, NA inhibition assays as described by Potieret al.¹⁸ were performed. Briefly, viruses were standardized to an NA activity 10-fold greater than that of the background and then incubated with serial 3-fold dilutions of drugs, including oseltamivir carboxylate (TRC Inc.), zanamivir (TRC Inc.), and peramivir (kindly provided by Green Cross Inc.). NA activity of viruses was determined using the NA-Star influenza NA inhibitor resistance detection kit (Applied Biosystems) according to the manufacturer’s instructions. Fifty percent inhibitory concentration (IC₅₀) values were calculated using nonlinear curve fitting with GraphPad Prism software (GraphPad Software). If a mutant virus showed <5-fold increase in IC₅₀ value over that of the wild-type virus, it was considered sensitive to NAIs. If a mutant virus showed >50-fold increase over the wild-type strain, it was considered highly resistant to NAIs. The IC₅₀ values of CA04 and maCA04 viruses were sensitive to all the NAIs tested (Table 2). CA04_H274Y, which showed high resistance to OS as well as peramivir, retained its low susceptibility to the two NAIs (149- and 169.4-fold in IC₅₀, respectively) (Table 2). Collectively, the results showed that OS-resistant variants could increase virulence without losing resistance to NAIs as much as their OS-sensitive counterparts during adaptation in a new host and share a molecular basis of pathogenesis similar to that in the wild-type A(H1N1)pdm09 virus.

Here, we show that parental CA04 and CA04_H274Y viruses had comparable replicative ability and virulence in mice which corresponded to previous studies.^6,7 However, we also show for the first time that mammalian adaptation of the CA04_H274Y virus through serial passage in mice resulted to increased growth properties and virulence in vitro and/or in vivo with no apparent alteration of NAI resistance implicating undiminished viral fitness conferred by the H274Y mutation. Increased in pathogenicity of CA04 and CA04_H274Y viruses appeared to be associated with previously described specific mutations, particularly HA-D222G, as a consequence of mouse adaptation of indicating parallel genetic evolution of A(H1N1)pdm09 viruses despite differences in OS-sensitivities. Notably, infection with a A(H1N1)pdm09 containing both NA-H274Y and HA-D222G mutations demonstrated persistent viral replication and induced severe illness that culminated to death,¹⁹ indicating that combination of this two mutations in the context of the A(H1N1)pdm09 virus could be lethal. It is also noteworthy that only maCA04_H274Y viruses additionally contained consensus K153E mutation in HA. The presence of HA-153E affects receptor binding specificity by decreasing affinity for both α-2-3- and α-2-6-sialic receptors.^20,21 Given that the H274Y mutation also confers decreased NA enzymatic activity,^22,23 it appears that the K153E mutation could be induced by the NA-H274Y to restore the balance between HA and NA activities. It is interesting to note that inoculation with maCA04_H274Y containing HA-153E induced rapid weight loss and killed all infected mice two days earlier than maCA04 giving it merit for more in-depth study.

Regardless of these results, the lethal OS-resistant variant remained sensitive to zanamivir, which is an alternative influenza antiviral agent for the variant (Table 2). However, extensive selective pressure by NAIs could result in other variants that could acquire multidrug resistance, such as the I223R mutation which also showed viral fitness comparable to that of the wild type.^24,25 Altogether, our findings emphasize the potential emergence of drug-resistant variants associated with high virulence as well as the need for the development of novel antiviral agents.

Acknowledgments

This work was supported by a grant from the Korea Healthcare Technology R&D Project (A103001) by the Ministry of Health and Welfare, Republic of Korea. We appreciate David Galloway for editorial assistance.

Submitted

05/15/2013

Revised

07/24/2013

Accepted

07/29/2013

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.