Changes in the Length of the Neuraminidase Stalk Region Impact H7N9 Virulence in Mice

Abstract

The neuraminidase stalk of the newly emerged H7N9 influenza virus possesses a 5-amino-acid deletion. This study focuses on characterizing the biological functions of H7N9 with varied neuraminidase stalk lengths. Results indicate that the 5-amino-acid deletion had no impact on virus infectivity or replication in vitro or in vivo compared to that of a virus with a full-length stalk, but enhanced virulence in mice was observed for H7N9 encoding a 19- to 20-amino-acid deletion, suggesting that N9 stalk length impacts virulence in mammals, as N1 stalk length does.

TEXT

In February and March 2013, a novel influenza A (H7N9) virus emerged and spread quickly across China, resulting in a total of 672 laboratory-confirmed cases in humans, with at least 271 deaths as of 23 June 2015 (http://www.who.int/influenza/human_animal_interface/HAI_Risk_Assessment/en/) occurring during the subsequent years. Influenza A virus is an enveloped, negative-sense, single-stranded RNA virus, with a segmented genome comprising eight gene segments: hemagglutinin (HA), neuraminidase (NA), basic polymerase 2 (PB2), basic polymerase 1 (PB1), acidic polymerase (PA), nucleoprotein (NP), matrix protein (M), and nonstructural protein (NS). Influenza A viruses may encode up to 16 proteins: HA, NA, PB2, PB1, PB1-F2, PB1-N40, PA, PA-X, PA-N155, PA-N182, NP, M1, M2, M42, NS1, and NS2 (1, 2). Cleavage of the HA precursor polypeptide (HA0) into subunits HA1 and HA2 is required for virus infection, and the amino acid residues at the HA0 cleavage site are a molecular hallmark for virulence in chickens. Almost all of the highly pathogenic avian influenza viruses (HPAIVs) possess a multibasic amino acid motif in the HA cleavage site, unlike with low-pathogenicity avian influenza viruses (LPAIVs) (3–5). Based on the molecular characteristics of the HA0 cleavage site and asymptomatic infections in chickens, the new H7N9 virus was confirmed as an LPAIV (6, 7). Nevertheless, this H7N9 virus caused severe disease in humans, with a high mortality rate (∼40%). The underlying mechanisms behind the virulence of H7N9 influenza virus in humans are yet not fully understood.

Apart from the human-strain-like mutations in HA and PB2, a 5-amino-acid deletion in the NA stalk region (69th to 73rd amino acids) was detected in this novel H7N9 virus (6). Notably, this deletion appeared for the first time in N9, based on NA sequences deposited in GenBank. Apart from that of H7N9, NAs with shortened stalks have previously been reported with different influenza subtypes, such as H5N1, H6N1, H7N1, H7N3, H9N2, and H2N2 (8–14). The shortened NA stalk was considered a relic of molecular evolution which resulted from the early adaptation of influenza virus from wild aquatic birds to terrestrial poultry (8, 10, 11). Viruses containing short-stalk NA were shown to have altered rates of virus growth, virulence, and transmission in chickens and mammals (9, 12, 13, 15–17), in which the functional balance between HA and NA in the influenza virion was found to be a critical determinant (18–23). Hence, NA stalk length, levels of glycosylation on HA, and mutations in the HA and NA antigens represent potential evolutionary strategies for the control of HA-NA functional balance to optimize virus fitness (12, 18, 20–22).

The NA segment from the novel H7N9 virus was found to possess a 5-amino-acid deletion, but it is not known whether this deletion confers any particular biological advantage with regard to human infections. To investigate the functions of this shortened NA stalk in virus replication and pathogenesis, wild-type A/Anhui/1/2013(H7N9) (AH-H7N9, short-stalk NA) and a modified virus containing a 5-amino-acid (69-QISNT-73) supplement in the NA stalk (AH-NA/su-H7N9, long-stalk NA) were rescued by reverse genetics as described previously (24, 25). The motif for the 5 amino acids was based on previous N9 sequences deposited in GISAID and GenBank. The rescued stock viruses have comparable virus titers, as determined in egg embryos, as well as in the MDCK and A549 cell lines (Table 1), suggesting that AH-H7N9 and AH-NA/su-H7N9 possess comparable growth capacities.

TABLE 1.

Infectivities and NA activities of the AH-H7N9 (short-stalk NA) and HA-NA/su-H7N9 (long-stalk NA) viruses

Virus	Infectivity					NA activitya
	HA titer	EID50/0.1 ml	TCID50/0.1 ml		MLD50 (EID50)	Km (μM)	Vmax (fluo AU/s)	Vmax ratio
			MDCK cells	A549 cells
AH-H7N9	27	108.75	106.81	102.5	107.5	194 ± 31	20 ± 2	1.0
AH-NA/su-H7N9	26	108.5	107.0	102.5	106.88	185 ± 45	36 ± 8	1.8

^afluo AU, fluorogenic arbitrary units. The V_max ratio is the ratio of the AH-NA/su-H7N9 to the AH-H7N9 V_max values.

To examine the effect of the shortened stalk on NA activity, the NA enzymes of β-propiolactone-inactivated AH-H7N9 and AH-NA/su-H7N9 were first diluted to equal titers (64 HA units and 10^8.5 50% egg infective doses [EID₅₀]/0.1 ml) and then quantified by the fluorogenic substrate 4-methylumbelliferyl-N-acetylneuraminic acid as described previously (25). As shown in Table 1, the Michaelis-Menten constant (K_m) values between the inactivated AH-H7N9 and AH-NA/su-H7N9 viruses were similar, whereas the maximum velocity of the substrate conversion (V_max) of AH-NA/su-H7N9 was greater than that of AH-H7N9 (Table 1), which is consistent with the results of the virus's elution from chicken red blood cells (CRBC) and human red blood cells (HuRBC). The elution of AH-NA/su-H7N9 was completed after an incubation period of 2 or 3 h from CRBC or HuRBC, respectively, while AH-H7N9 took 3 or 4 h under the same experimental conditions (Fig. 1). These results indicate that the 5-amino-acid deletion in the N9 stalk region has no effect on NA activity in a monovalent substrate assay but may affect the release of viruses from erythrocytes. These results are consistent with those of previous studies with N1 (9, 15, 26).

FIG 1.

Virus elution from chicken or human erythrocytes. Viruses with HA titers of 1:64 (2⁶) were incubated with equal volumes of 1% chicken (A) or human (B) RBC at 4°C for 30 min or 60 min, respectively. Samples were then transferred to 37°C, and reductions in HA titer were recorded periodically. The HA titer following incubation at 37°C is expressed as a percentage of the HA titer at time zero at 4°C (y axis). The data are presented as the means ± standard deviations from four independent experiments and were compared to each other by two-way analysis of variance (ANOVA). **, P < 0.01; ***, P < 0.001.

To explore the impact of N9 stalk length on virus replication, multicycle growth curves of AH-H7N9 and AH-NA/su-H7N9 were analyzed in MDCK and A549 cells (Fig. 2). The replication ability of AH-H7N9 was shown to be comparable to that of AH-NA/su-H7N9. Substantial deviations in virus titers between the two viruses over the measured time points were not detected.

FIG 2.

Multiple-cycle growth curves of the H7N9 viruses with a long or short NA stalk. The growth kinetics of the AH-H7N9 and AH-NA/su-H7N9 viruses were measured by multiple-cycle growth curve analysis in MDCK (A) and A549 (B) cells. The cells were inoculated at a multiplicity of infection (MOI) of 0.001. Supernatants were collected every 12 h until 96 hpi, and the virus titers were determined by measuring 50% tissue culture infective doses (TCID₅₀) in MDCK cells. The data are presented as the means ± standard deviations of results from three independent experiments and were compared to each other by two-way ANOVA.

To further evaluate the contributions of N9 stalk deletion to virus replication and pathogenicity in vivo, BALB/c mice were challenged with an intranasal (i.n.) inoculation of AH-H7N9 or AH-NA/su-H7N9 at a dose of 10⁶ EID₅₀. Changes in body weight were similar between the two challenge groups throughout the experiment. In both groups, the body weight gradually decreased after 2 days postinfection (dpi), reaching the lowest values around 6 dpi. All mice eventually recovered, and their body weights returned to levels similar to those of the PBS control group by 14 dpi (Fig. 3A). In addition, the AH-H7N9 and AH-NA/su-H7N9 viruses exhibited comparable lung virus titers (LVTs) throughout the infection (Fig. 3B). Histopathological analysis revealed that both the AH-H7N9 and the AH-NA/su-H7N9 virus caused severe bronchopneumonia at 7 dpi (Fig. 3C and D). Levels of interferon gamma (IFN-γ), monocyte chemoattractant protein 1 (MCP-1), tumor necrosis factor alpha (TNF-α), interleukin 6 (IL-6), IL-10, and IL-12p70 in mouse lungs were significantly increased at 3, 5, and 7 dpi after challenge with AH-H7N9 and AH-NA/su-H7N9, compared to levels in control mice. However, substantial differences were not observed between the two viruses in terms of cytokine release (Fig. 3E to J).

FIG 3.

Pathogenicities of the H7N9 viruses with long and short NA stalks in BALB/c mice. AH-H7N9 and AH-NA/su-H7N9 were inoculated into the mice of each group by the intranasal route at a dose of 10⁶ EID₅₀. Mice in the control group were mock infected with phosphate-buffered saline (PBS). (A) Body weights were monitored daily for a 14-day observation period, and weight changes are expressed as a percentage of the initial value (measured on the day of challenge). (B) Three mice from each group were euthanized at 3, 5, and 7 dpi, respectively. The lung virus titers (LVTs) were measured in special pathogen-free (SPF) embryonated eggs and are expressed as log₁₀ EID₅₀/ml. (C and D) Hematoxylin- and eosin-stained lung sections (magnification, ×600) at 7 dpi of mice challenged with AH-H7N9 or AH-NA/su-H7N9. Inflammatory cell infiltration, deciduous epithelium mucosa and inflammatory cells in the bronchial lumen, and hemorrhage are indicated by thick solid arrows, thick white arrows, and triangles, respectively. Scale bar = 50 μm. (E to J) Cytokine released into the lungs after challenge with AH-H7N9 or AH-NA/su-H7N9. The data are presented as means ± standard deviations and were analyzed by two-way ANOVA and a paired-sample t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001). The black asterisks represent the P value for the AH-H7N9- or AH-NA/su-H7N9-infected mice in a comparison with PBS (mock)-infected mice.

To further evaluate the impact of NA stalk length on the virulence of the different recombinant H7N9 viruses in mice, the 50% median lethal dose (MLD₅₀) was determined by challenge of BALB/c mice with a dose of 10⁶, 10⁷, or 10⁸ EID₅₀ of AH-H7N9 or AH-NA/su-H7N9. The MLD₅₀ values of the two viruses revealed that the 5-amino-acid deletion in the N9 stalk did not lead to an increase of virulence in mice (Table 1). However, the continual evolution in the NA stalk of H7N9 viruses may create stalks of various lengths, including a severely shortened version similar to that of N1 from H5N1 virus. Furthermore, larger deletions of 19 to 20 amino acids in the H5N1 NA stalk were previously found to impact virulence (9, 13). Therefore, we further investigated the pathogenicity of H7N9 viruses with a similar long (19- or 20-amino-acid) deletion in the NA stalk region (Fig. 4A to D), based on knowledge gained from N1 stalk deletions (deletions of the 49th to 68th and 54th to 72nd amino acids) in the H5N1 virus. We also designed a deletion model (54th to 73rd amino acids) that contains both the deletion of the 54th to 72nd amino acids of the N1 stalk and the deletion of the natural 69th to 73rd amino acids of the N9 stalk (Fig. 4D). Notably, longer deletions in the NA stalk enhanced the virulence of recombinant AH-NA-del-1 (deletion of the 49th to 68th amino acids), AH-NA-del-2 (deletion of the 54th to 72nd amino acids), and AH-NA-del-3 (deletion of the 54th to 73rd amino acids) viruses in mice (Fig. 4A and B), with MLD₅₀ values of 10^6.5, 10^6.0, and 10^6.0 EID₅₀, respectively. In addition, the LVTs of the AH-NA-del-1, AH-NA-del-2, and AH-NA-del-3 groups were a litter higher than those of the AH-H7N9 group, although there was no statistically significant difference (Fig. 4C). Consistently with observed clinical symptoms, histopathological changes (Fig. 4E to H) and cytokine secretion into the lung (Fig. 4I to L) caused by AH-NA-del-1, AH-NA-del-2, and AH-NA-del-3 infections were significantly stronger than those caused by AH-H7N9 virus. Interestingly, AH-NA-del-1, AH-NA-del-2, and AH-NA-del-3 displayed lower HA titers (2⁴ to 2⁵, 2⁴, and 2³ to 2⁴, respectively) and EID₅₀ titers (10^−8.5, 10^−7.5, and 10^−7.75 per 0.1 ml, respectively) than AH-H7N9. The balance between HA and NA may have been altered by the longer deletions in the NA stalk of H7N9 virus, which is currently a topic of investigation in our laboratory.

FIG 4.

Pathogenicities to mice of the H7N9 virus with different deletions in the NA stalk. (A and B) Levels of infection of 10⁶ EID₅₀ (A) and 10⁷ EID₅₀ (B) of wild-type H7N9 (AH-H7N9) or laboratory-generated mutant H7N9 containing increasingly larger deletions in the NA stalk (AH-NA-del-1, AH-NA-del-2, and AH-NA-del-3) were inoculated into BALB/c mice (n = 5 per group) by the intranasal route. Body weights were monitored daily over 14 days, and weight changes are expressed as a percentage of the initial value, as measured at the time of challenge. After the infection of the H7N9 viruses at a dose of 10⁶ EID₅₀, three mice from each group were euthanized at 3, 5, and 7 dpi, respectively. (C) LVTs were measured in SPF embryonated eggs and are expressed as log₁₀ EID₅₀/ml. The data were analyzed among the AH-NA-del-1 to -3 groups and the AH-H7N9 group by two-way ANOVA (*, P < 0.05). (D) The different deletions generated in the NA stalk of the mutant H7N9 virus are shown. The dots in the frame represent deleted amino acids. The 5 amino acids (69-QISNT-73) that were deleted from the wild-type H7N9 virus are in blue. (E to H) Hematoxylin- and eosin-stained lung sections (magnification, ×400) of mice infected at 5 dpi with AH-H7N9 or AH-NA-del-1 to -3. Inflammatory cell infiltration, deciduous epithelium mucosa and inflammatory cells in the bronchial lumen, and hemorrhage are indicated by thick solid arrows, thick white arrows, and triangles, respectively. Scale bar = 100 μm. (I to L) Cytokines released into the lung after challenge with AH-H7N9 or AH-NA-del-1 to -3. The data are presented as means ± standard deviations and are analyzed by two-way ANOVA and a paired-sample t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001). The black and red asterisks represent the P value for the virus infection groups in a comparison with the PBS (mock)-infected group and the AH-H7N9 to AH-NA-del-1 to -3 groups, respectively.

In summary, our data indicate that the naturally occurring 5-amino-acid deletion in the NA stalk of H7N9 virus did not have a significant impact on NA activity, viral replication in vitro and in vivo, or pathogenesis in mice, even though this deletion reduces the virus release from CRBC and HuRBC. The 5-amino-acid deletion in the NA stalk most likely did not substantially impact the balance between HA and NA, which play critical roles in viral entry and release. However, a longer deletion in the N9 stalk region of about 19 or 20 amino acids leads to enhanced H7N9 virus pathogenesis in mice, which has previously been observed with H5N1 and H9N2 viruses possessing shortened NA stalk regions (9, 13, 27). Longer deletions in the NA stalk may have disrupted the balance between HA and NA, impacting H7N9 virulence in mice as well as the virus HA titers. Understanding the effect of NA stalk length and its relationship to the virulence of H7N9 in mammals will be important for future surveillance studies evaluating the potential threat to humans of circulating H7N9 viruses.