Influenza neuraminidase mutations and resistance to neuraminidase inhibitors

ABSTRACT

Mutations in influenza virus neuraminidase (NA) can lead to viral resistance to NA inhibitors (NAIs). To update global influenza NA mutations and resistance to NAIs, we investigated epidemic information from global regions for NAIs-resistant influenza strains and analyzed their NA mutations. Drug-resistant mutations in NA, especially new mutations occurred in 2016-2024, were updated. The H274Y mutation in N1, a major contributor to NAI resistance, peaked in 2008, significantly impacting public health in countries like Japan and the USA. Three main mechanisms of NAI resistance were identified: catalytic site mutations, structural hindrance, and monomer stability changes. Although global resistance rates of H1N1pdm09, H3N2, and influenza B have remained stable at around 1%, sporadic emergence of resistant strains highlights the need for continued vigilance. The evolution of drug-resistant, transmissible strains through compensatory mutations underscores the urgency of new antiviral strategies. Strengthening global surveillance and adjusting public health policies, such as improving vaccine coverage and prudent antiviral use, remain essential to mitigating future risks.

Introduction

Influenza viruses belong to the family Orthomyxoviridae and are categorized into types A, B, C, and D, with type A and B being the most significant in terms of human disease. Neuraminidase (NA) is a viral surface glycoprotein that helps to free influenza viruses from mucin-associated decoy receptors and to facilitate budding from infected cells. At present, NA has been considered to be one of the most promising targets for combating influenza. According to the different antigenicity of NA, influenza viruses can be divided into 11 subtypes (N1-N11), most of which have no cross-reactivity in serology [1,2]. NA is a tetramer formed by polymerization of four identical polypeptide monomers, each of which consists of a cytoplasmic tail, a transmembrane region, a stalk, and a catalytic head [3]. The partial N-terminal domain formed by the cytoplasmic tail and transmembrane region acts as apical targeting of NA to the apical plasma membrane, and its nearly 100% conserved sequence ensures the morphology and budding efficiency of viral particles [4]. The number and sequence variation of amino acid residues in stalk can be large and relate to the stability of tetramer [4]. The catalytic head of NA catalyzes the hydrolysis of sialic acid after virus maturation, releasing mature virus particles from infected cells; mature virus particles then infect other cells, initiating a new virus lifecycle and further expanding the scope of infection [5]. The resistance caused by NA mutations occurs only in the catalytic head, especially in the active site.

Since 2010, NA inhibitors (NAIs) have been the antiviral drugs most recommended by the World Health Organization (WHO) for the treatment of influenza A and B viruses. They bind to the active site of NA, competing with sialic acid binding and inhibiting its hydrolysis, reducing the transmission of mature virus particles to uninfected cells. Among NAIs, oseltamivir and zanamivir have been approved worldwide for the prevention and treatment of influenza [6], while peramivir has also been approved for influenza treatment in many countries such as China, the USA, and Japan; laninamivir requires more adequate clinical trials before it will be widely available in the market for clinical use [7].

Although there have been previous reports on the evolution of NA sequences and drug-resistant mutations [6,8,9], there are also many new developments worth updating in recent years. Therefore, this review firstly summarizes the prevalence of NA-resistant strains through the updated analysis of global clinical isolates. The H274Y mutation (N2 numbering, used throughout) is paid much more attention due to a high frequency of appearance in the world. The information on the drug resistance rate of seasonal influenza viruses and the new drug resistance mutation sites that have emerged since 2016 are updated. An interesting phenomenon in NA mutation is illustrated: the same point mutation produces different drug resistance effects, and multiple point mutations often produce synergistic effects, occasionally producing antagonistic effects. Lastly, three mechanisms of NA mutation leading to NAI resistance are summarized: disrupt hydrogen bonding interaction, cause steric hindrance, and change monomer stability.

Epidemiology of drug resistance in influenza virus

The Global Influenza Surveillance and Response System (GISRS) of the World Health Organization (WHO) counts influenza outbreaks worldwide every two weeks [10]. According to summary reviews of influenza seasons by hemisphere, only less than 1% of influenza A (H1N1) pdm09 virus in North America showed resistance to oseltamivir or peramivir from 2017 to 2018, but unfortunately, nearly 2% of influenza A (H1N1) pdm09 virus has emerged in Europe as being resistant to oseltamivir or zanamivir, with most of the drug-resistant viruses exhibiting H274Y mutations [11]. Less than 0.5% of influenza A(H1N1) pdm09 and B viruses in North America from 2018 to 2019 showed resistance to oseltamivir and zanamivir [12]; however, from 2018 to 2020, the NAI resistance rate of influenza A(H1N1) pdm09 was as high as 1.3% globally with over 75% exhibiting the H274Y mutation, which provides high resistance to oseltamivir and peramivir (IC₅₀ increases by over 100-fold), and in comparison, the resistance rate of influenza A (H3N2) was less than 0.1%; B/Victoria was 1.0%, and B/Yamagata was 1.1% [13]. The NAI resistance rate of influenza A(H1N1) pdm09 (the main epidemic virus strain) doubled in the following year (2019-2020), while the resistance rate was only 0.7% the previous year (2018–2019) [13]. From 2017 to 2018, the NAI resistance rate of A(H1N1) pdm09 was as high as 1.5%; H3N2 was nearly 0.4%, B/Victoria was over 1.1%, and B/Yamagata was over 0.6%, which compared with the previous year, the NAI resistance rates of all four types of influenza viruses showed a significant > 3-fold increase [14,15].

NAI resistance rates for major influenza viruses worldwide from 2014 to 2020 are presented in Table 1 [13–17]. The A(H1N1) pdm09 strain had the highest NAI resistance rate compared with other strains, which may be related to its higher incidence of H274Y mutation. The NAI resistance rate of H3N2 has remained below 0.5%, with the latest data showing < 0.1%. Although the NAI resistance rate of influenza B fluctuates greatly, the overall resistance rate is lower than that of influenza A [7]. The global monitoring data for NAI resistance from 2020 to 2022 has not yet been published or reported, possibly because but not limited to the following reasons: this period coincided with the outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and some flu symptoms were similar to those caused by this virus, making it difficult to distinguish the two infections based on symptoms alone; additionally, the global SARS-CoV-2 pandemic made it difficult to carry out the work of influenza monitoring, and as various countries focused mainly on that pandemic, there was a significant decrease in attention paid to influenza. Nevertheless, the global trend of NAI resistance in influenza virus outbreaks still requires close monitoring to ensure timely and effective prevention and treatment measures can be implemented.

Table 1.

Global monitoring of neuraminidase inhibitor-resistance rates in major influenza subtypes from 2014–2020.

Subtype
Year	A/(H1N1) pdm09	H3N2	B/Victoria	B/Yamagata
2018–2020	1.3%	<0.1%	1.0%	1.1%
2017–2018	1.5%	0.4%	1.1%	0.6%
2016–2017	0.5%	0.1%	0.4%	0.2%
2015–2016	1.8%	0.2%	0.5%	0.4%
2014–2015	0.5%	0.2%	0.7%	1.0%

NA mutations and influenza drug resistance

Most of the data reported so far on NA mutations leading to drug resistance are derived from reverse genetics and clinically isolated viruses. Most drug-resistant mutations occur in 19 highly conserved active sites, while a few non-active sites may also mutate. In this review, the newly discovered NA resistance mutation sites from 2016 to 2024 were assessed (Table 2). We searched the SciFinder, PubMed and Web of Science database for literature published since 2016 using the keywords: “influenza” “neuraminidase” and “resistance mutation.” A total of 1433 articles were retrieved. At the same time, the WHO annually report of “recommended composition of influenza virus vaccines for use” and “annual review of influenza season” were also reviewed. Previously unreported drug resistance mutation sites in NA of viruses with human hosts were summarized here. We found that most of the new resistance mutations occurred in the influenza B virus, while several new drug resistance mutations occur in H1N1pdm09 and H3N2 strains. Mutations H274Y plus G147R in A/N1 was found to cause extensive resistance to NAIs, which needs attention. In addition, drug resistance information for 19 active sites of NA and other frequently reported non-active sites in virus derived from clinical isolation and recombination were summarized (Table 3) to reveal the relationship between influenza virus NA sequence evolution and drug resistance. The NA active site consists of nineteen highly conserved residues, eight catalytic residues that directly interact with sialic acids (R118, D151, R152, R224, E276, R292, R371, and Y406) and 11 framework residues that maintain the structure of the active site (E119, R156, W178, S179, D198, I222, E227, H274, E277, N294, and E425). So far, the conserved catalytic residue R118, Y406 and the framework residue W178, S179, E277 and E425 in influenza A have not yet found drug-resistance mutation.

Table 2.

Newly emerged NAI-resistant mutations from 2016 to 2024.

Mutations conferring resistance	Subtype	Year isolated	Resistance level (fold change)e				Reference
			Oseltamivir	Zanamivir	Peramivir	Laninamivir
T43A	B	2017/2018	–	Y	–	–	[18]
P76S	B	2017/2018	Y	Y	–	–	[18]
H101L	B	2019	2	34	688	5	[19]
G104E	B	2016	87	1220	17724	701	[17,20]
G108Ea	B	2019	5	5	55	3	[19]
S110F	A/N1	2017	10.47	18.45	–	–	[21]
I115T	B	2017/2018	Y	Y	Y	Y	[18]
H134N	B	2016	4	158	72	41	[22]
T146I	B	2017/2018	–	–	Y	–	[23]
T146Kb	B	2016	6	3	5103	1	[19]
T146Pb	B	2018	2	4	128	3	[19]
R154Kd	B	2018/2019	–	–	–	–	[24]
A246Vc	A/N2	2022/2023	–	Y	–	–	[25,26]
S246G	A/N1	2022/2023	Y	N	N	N	[27]
S246P	B	2016	1.83	39.21	–	–	[21]
H273Q	B	2021	–	Y	–	–	[28]
N329K	A/N2	2016/2017	14	8–11	–	–	[14]
S334R	A/N2	2018/2019	Y	–	–	–	[24]
N340D	B	2016/2017	Y	–	–	–	[29]
I348T	B	2016/2017	Y	–	–	–	[20]
E358K	B	2016/2017	Y	–	–	–	[29]
G407S	B	2018	8.3	167.2	64.2	204	[30]
H431Y	B	2016/2017	Y	Y	Y	Y	[31]
D432N	B	2017/2018	–	–	Y	–	[18]
T436Pd	B	2018/2019	–	–	–	–	[24]
H439P	B	2019	<1	13	120	4	[19]
H439R	B	2019	<1	5	20	2	[19]
P124 T/V422I	B	2017	2.59	94.50	–	–	[21]
T146P/N169S	B	2019	31	573	10074	280	[19]
I222 V/S246N	A/N1	2023/2024	13	3	4	3	[32]
G247D/I361V	B	2019	2	6	46	3	[19]
H274Y/G147R	A/N1	2016	2600	5	1400	N	[33]
V303I/N342K	A/N2	2018/2019	Y	Y	–	–	[24]
H274Y/I222 V/S246Nd	A/N1	2023/2024	–	–	–	–	[34]

^aA mixed population of wild-type and mutant strains; ^b The mutation was detected and isolated in multiple years, and the table uses data from the earliest detected year; ^c One strain isolated from 1023 H3N2 strains did not specify the specific year and resistance level; ^d Resistance to NAIs but not specified which drug. ^e The fold change of IC₅₀ between mutant strain and wild strain, Y – “Yes” indicates resistance, N – “NO” indicates no drug resistance, and “ – ” indicates not date.

Table 3.

NA mutations resulting in NAI resistance.

Mutations conferring resistance	Subtype	Virus sourcea	Resistance level (fold change)b				Reference
			oseltamivir	zanamivir	peramivir	laninamivir
H101L	B	CI	–	34	688	–	[19]
E105K	B	CI	–	15	213	–	[35]
G108E	B	CI	–	–	55	–	[19]
S110F	A/N1	CI	10	18	–	–	[21]
E116G	B	RG	–	8000	–	–	[36]
E119A	A/N1	RG	–	58–255	67	–	[37,38]
E119A	A/N2	CI	–	25	–	–	[39]
E119A	A/N2	VP	–	>100	–	–	[40]
E119A	B	RG	3171	12538	13780	–	[41]
E119D	A/N1	RG	23–87	583–827	104–286	702	[37,42]
E119D	A/N2	CI	–	60	–	–	[39]
E119D	B	VP	–	1000	–	–	[40]
E119G	A/N1	RG	–	113–1485	51–164	108	[37,43–45]
E119G	A/N1	VP	–	>1000	–	>300	[46]
E119G	A/N2	CI	–	40	–	–	[39]
E119G	A/N9	VP	–	>100	–	–	[40]
E119G	A/N3	RG	24.5	523	13.6	–	[47]
E119V	A/N1	RG	60–1727	571–2144	25–5050	50	[43,44,48]
E119V	A/N2	CI	52–520	–	–	–	[49,50]
E119V	A/N2	RG	131–1028	–	–	–	[48,51–53]
E119V	A/N9	CI	84	–	–	–	[54]
E119V	A/N9	RG	91	–	–	–	[52]
H134N	B	CI	4	158	72	41	[22]
H134R	B	CI	–	308	358	–	[55]
H134Y	B	CI	–	–	75–92	–	[55]
H136N	B	RG	–	212	240	–	[56]
Q136K	A/N1	CI	–	36	80	–	[57]
Q136K	A/N1	RG	–	86–750	70	45	[43,58,59]
T146K	B	CI	–	3–192	3735–21893	–	[19]
T146K	B	RG	–	35–50	1997–2912	–	[19]
T146P	B	CI	–	–	128–3338	–	[19]
T146P	B	RG	–	–	746	–	[19]
D151E	A/N1	CI	–	10	–	–	[21]
R152K	A/N1	RG	18	–	–	–	[37]
R152K	A/N2	CI	188	29	–	–	[39]
R152K	B	CI	76–1000	>1000	–	–	[39,40,50,60]
R152K	B	RG	60	84	572	–	[56]
R156K	A/N1	RG	–	238	35	–	[38]
D198G	A/N1	RG	32	44	–	–	[45]
D198Y	B	RG	57	–	168	–	[41]
A200T	B	CI	–	–	77	–	[19]
I222L	A/N2	RG	9	–	–	–	[51]
I222K	A/N9	CI	32	–	–	–	[54]
I222M	A/N1	RG	36	–	–	–	[45]
I222R	A/N1	RG	53	–	10	–	[61]
R224K	A/N2	RG	>4000	>50	–	–	[62]
E227D	A/N1	RG	–	126	–	–	[63]
A246T	B	RG	–	11	–	–	[52]
S246P	B	CI	–	39	–	–	[21]
H274Y	A/N1	CI	200–2116	–	558	–	[21,50,64]
H274Y	A/N1	RG	177–2502	–	51–661	–	[37,44,45,48,61,63,65–67]
H274Y	A/N2	CI	400	–	–	–	[39]
H274Y	A/N3	RG	12229	–	–	–	[47]
H274Y	A/N9	RG	23	–	–	–	[52]
H274Y	B	RG	–	–	110	–	[41]
E276D	A/N2	RG	15	160	–	–	[62]
R292K	A/N1	RG	33	–	–	–	[37]
R292K	A/N2	CI	9375–110000	–	–	–	[39,40,49,50]
R292K	A/N2	RG	>10000	134	719	–	[48]
R292K	A/N2	VP	–	10–30	–	–	[40]
R292K	A/N3	RG	17960	22.5	41	–	[47]
R292K	A/N9	RG	3686–20612	22–54	563	35	[52,68,69]
R292K	A/N9	CI	>2000	16–66	345–1823	–	[69,70]
N294S	A/N1	RG	64–208	–	12	–	[37,44,48]
N294S	A/N2	CI	280	–	–	–	[49]
N294S	A/N2	RG	1879	–	–	–	[48]
N294S	B	RG	26	–	31	–	[41]
R371K	A/N2	RG	45	15	–	–	[62]
R371K	B	RG	101	145	352	–	[41]
H431R	B	CI	–	31	133	–	[55]
D432G	B	CI	–	–	148	–	[19]
I436N	A/N1	RG	30	71	20	20	[65]
W438R	B	CI	–	55	337	–	[55]
H439P	B	CI	–	13	120	–	[19]
H439R	B	CI	–	–	20	–	[19]
E119 V/P126S	A/N2	CI	413	–	–	–	[71]
E119 V/I222L	A/N2	RG	1571	–	–	–	[51]
E119 V/I222V	A/N2	CI	1006	–	–	–	[53]
E119 V/I222V	A/N2	RG	293	–	–	–	[53]
P124 T/V422I	B	CI	–	95	–	–	[21]
T146P/N169S	B	CI	31	573	10074	280	[19]
T146P/N169S	B	RG	44	225	7158	86	[19]
D151E/Q136K	A/N1	CI	–	25	18	–	[57]
G247D/I361V	B	CI	–	–	46	–	[19]
H274Y/I117M	A/N1	RG	546	–	179	–	[66]
H274Y/I117V	A/N1	RG	1896	–	523	–	[66]
H274Y/E119A	A/N1	RG	1173	56	14579	–	[37]
H274Y/E119D	A/N1	RG	790–3381	136–903	5958–33333	366	[37,42]
H274Y/E119G	A/N1	RG	225	224	＞33333	–	[37]
H274Y/Q136K	A/N1	CI	198	15	1805	–	[57]
H274Y/D151E	A/N1	CI	231	–	80	–	[57]
H274Y/D151G	A/N1	CI	1189	14	1161	–	[57]
H274Y/D151N	A/N1	CI	799	–	718	–	[57]
H274Y/D198N	A/N1	RG	843	–	–	–	[37]
H274Y/D198G	A/N1	VP	1818	–	5667	–	[72]
H274Y/I222M	A/N1	RG	1943–8024	–	400–3340	–	[45]
H274Y/I222R	A/N1	CI	9053	22	13092	–	[73]
H274Y/I222R	A/N1	RG	1647	16	17347	–	[61]
H274Y/I222V	A/N1	RG	971–1925	–	893–2707	–	[44,45,61]
H274Y/S334N	A/N1	RG	658	–	178	–	[44]
H274Y/I436N	A/N1	RG	4812	36	10459	22	[65]
E59G/E119 V/I222V	A/N2	RG	188	–	–	–	[53]
H274Y/Q136 K/D151N	A/N1	CI	356	–	2300	–	[57]

^aRG – reverse genetics, CI – clinically isolated, VP – in vitro passage; ^b The fold change of IC₅₀ between mutant strain and wild strain, “ – ” indicates not date.

Single-site and multi-site mutations

There are numerous reported NA mutations, but H274Y in the N1 subtype and E119 and R292 in the N2 subtype are the main mutations conferring drug resistance [74]. Because of the high conservation of residues at the active sites, slightly more single-site mutations occur there, and most active site mutations convey resistance in only one subtype. For example, H274Y only conveys a high level of oseltamivir resistance in N1, and E119 V and R292 K only convey a high level of oseltamivir resistance in N2 [8]. However, drug resistance caused by a small number of active site mutations can occur in multiple subtypes, such as the N294S mutation (N295S in N1 numbering) found in both N1 and N2 subtypes leading to reduced sensitivity to oseltamivir [8]. Other studies have reported that resistance from H274Y can also exist in group two recombinant viruses such as N3 (decreasing sensitivity to oseltamivir by > 10,00-fold) and N9 subtypes [47,52]. Despite this, the majority of NAI resistance induced by H274Y occurs in N1-type influenza viruses (Table 3). With continuous evolution, E119 V and R292 K, which once showed resistance in only one subtype, have also gradually conferred resistance to NAIs in other subtypes (Table 3).

Mutations at the same NA sites can lead to different drug-resistance phenotypes. There is controversy over whether the Y155H mutation causes effective drug resistance. Y155 is conserved in human-hosted N1 subtype viruses, while H155 is more common in N1 or N2 subtype viruses in pigs or poultry [75]. Monto et al. found Y155H in the human influenza virus A/Hokkaido/15/2002, which is highly resistant to oseltamivir (123-fold) and zanamivir (555-fold) that was attributed to outliers [75]. However, it was further demonstrated in a 2013 report involving some of the same authors that Y155H does cause a 30 to > 100-fold decrease in inhibitor sensitivity [76]. Perez-Sautu et al. isolated two strains of Y155H A(H1N1) pdm09 influenza virus from two patients, neither of which showed NAI resistance [77]. As a functional amino acid that can directly interact with substrates, when E276 is mutated into D276 in influenza virus N1 and N2 subtypes, NAI sensitivity changes are different. In 2006, Yen et al. constructed a recombinant virus using A/Wuhan/359/95 (H3N2) and introduced the E276D mutation. They found that the resulting sensitivity to zanamivir decreased by 160-fold and that of oseltamivir decreased by 15-fold [62]. The following year, Ho et al. showed that the E276D mutation in N1 did not lead to resistance to oseltamivir and zanamivir; they explained this as an individual amino acid influence between the two subtypes [63]. Similarly, Y347 is not the only mutation that can form additional hydrogen bonds with oseltamivir carboxylate groups in N1; Y252 can also form hydrogen bonds with residues 273 and 274 in 270 rings, which are not present in N2 (residue 347 wild-type is Gln) [63].

Most multi-site mutations produce synergistic effects and occasionally produce antagonistic effects. Okomo-Adhiamb et al. found that mutation of D151 near the NA 150 ring to E, G, or N did not cause resistance to oseltamivir, zanamivir, and peramivir in H1N1 seasonal influenza isolates, but when combined with H274Y mutation, it caused a large decrease in sensitivity to oseltamivir and peramivir [57]. Both H274Y and I222R single point mutations did not reduce susceptibility to zanamivir on their own, but H274Y and I222R combined mutations led to increased zanamivir resistance (up to 20-fold) [45,73]. The resistance conferred by these two combined mutations to oseltamivir and peramivir was much higher than that conveyed by single mutations [45,73].

The synergistic effect of different mutations around the same site reducing resistance has also been seen for H274Y + I222 V in Recombinant Pandemic A/Québec/144147/09(H1N1) viruses for drug sensitivity testing; although the sensitivity to oseltamivir conferred by the single mutations H274Y and I222 V decreased by 982-fold and 5.7-fold, respectively, and sensitivity to peramivir decreased by 263-fold and 1.9-fold, the sensitivity to oseltamivir and peramivir of the double-site mutant decreased by 1733-fold and 1331-fold, respectively [44]. Ferris et al. stably passed a constructed E119V + I222L mutant in MDCK cells and compared its drug sensitivity to wild-type virus (influenza A virus H3N2). They found that the E119V + I222L mutant had a 1571-fold higher resistance to oseltamivir than the wild-type, while the E119 V mutation only induced moderate resistance (197-fold) and the I222L mutation only induced mild resistance (9-fold) [51]. The above results suggest that closer attention must be paid to mutations of I222 and H274 because they can combine with other mutations to further reduce NAI sensitivity.

In addition to increasing drug resistance to certain NAIs, multi-site mutations can sometimes produce an antagonistic effect for other NAIs. When D151N co-mutates with three other sites, it unexpectedly antagonizes susceptibility to zanamivir; although Q136 K + H274Y double mutation results in severe resistance to oseltamivir (198-fold) and peramivir (1805-fold) and mild resistance to zanamivir (15-fold), when the D151N was synergistically mutated, the resistance to oseltamivir (356-fold) and peramivir (2300-fold) was further enhanced, but the resistance to zanamivir (4-fold) was reduced to near normal [57]. Similar antagonistic effects also exist for E59G. Baz et al. isolated and cultured multiple influenza virus strains with NA site mutations from an immunocompromised child (A/H3N2) over one year, and found that influenza virus mutants with the NA mutations E59G, E119 V, and I222 V in one sample showed 274-fold higher resistance to oseltamivir in NA inhibition tests compared with wild-type virus, and the oseltamivir resistance from E119 V and I222 V dual site mutations was 1006-fold higher than that of WT in another; the oseltamivir resistance conveyed by dual site mutation was stronger than that conveyed by triple site mutation [53].

Single-drug and cross-drug resistance

Because of the similar structure and mechanisms of action of the four NAIs, NA mutations leading to NAIs multi-drug resistance are more common than those leading to single drug resistance (Table 3). Fortunately, there are few cases in which infection with drug-resistant viruses conveys resistance to all four NAIs, allowing other NAIs to be used for treatment. For example, the N1 group mutation H274Y often leads to resistance to oseltamivir and peramivir, but has little impact on zanamivir (Table 3). The H7N9 strain isolated from Shanghai contains a rare R292 K mutation (R-23%, K-77%), conveying high resistance to oseltamivir; the recombinant virus (K-100%) constructed using this strain had moderate to severe cross-resistance to zanamivir, peramivir, and laninamivir [69]. The cross-resistance caused by NA mutation in influenza B is sometimes more serious than that in influenza A. Examining the B/Yamanashi/166/1998 influenza virus (Yamagata lineage), the introduction of E119A mutation greatly reduces the inhibitory effects of oseltamivir (3171-fold), zanamivir (12538-fold) and peramivir (13780-fold) [41]. If these cross-resistant viruses become prevalent, there will be no drugs available to treat them; because of this, attention should be paid to cross-resistant viruses and the pace of new drug research should be accelerated.

H274Y mutation in NA

The mutation rate of H274Y in the influenza A virus is far ahead of other active sites, and the two-site or three-site mutations carrying H274Y often cause multidrug resistance or aggravate drug resistance, causing greater pressure on influenza prevention and control (Table 3). To analyze the mutation H274Y in NA, a total of 17916 NA protein sequences were downloaded from NCBI's influenza virus database and subjected to multiple sequence alignment using MAFFT. The NA sequence data used in this review can be accessed through NCBI's influenza virus database (https://www.ncbi.nlm.nih.gov/genomes/FLU/Database/nph-select.cgi?go = 1), and was screened according to the following criteria: keyword: NA; search in sequence pattern: type A; human host: any country/region; NA protein: N1 subtype; collection date: from 1918 to 2024. In addition, we also searched and analyzed NA sequences from GISAID EpiFlu^TM database. The Search criteria is: Type A, N1, Host-Human, Location-All, Collection data-January 1, 1918 to September 5, 2024, Required Segments-NA. A total of 119564 N1 sequences were obtained and analyzed. Python was used to analyze the amino acids at 19 active sites (R118, D151, R152, R224, E276, R292, R371, Y406, and E119, R156, W178, S179, D198, I222, E227, H274, E277, N294, E425) [62] in the downloaded sequences. H274Y was found to have the highest mutation rate (exceeding 10%), far higher than the mutation rate of any other active site. However, in the GISAID EpiFlu™ database, the total mutation rate of H274Y is only 3% in the N1 sequences. This is primarily due to the fact that GISAID contains over 90,000 N1 sequences from 2015 to 2024, with the average mutation rate at this locus being just 1% over the past decade. This significantly lowers the overall mutation rate of H274Y. In contrast, the NCBI database contains fewer than 1,000 N1 sequences from the same period, leading to different mutation rate calculations. As a high-frequency mutation site in A(H1N1) pdm09, the frequency of the H274Y mutation is also very high in the literature; many NAI resistance studies have reported that the mutation of histidine to tyrosine at residue 274 often leads to high oseltamivir and peramivir resistance, posing severe challenges to the prevention and treatment of influenza.

The mutation rate at this site was analyzed globally over time (Figure 1(A), data from NCBI and GISAID), and statistical analysis was also conducted on the average mutation rate of H274Y in multiple countries with major epidemics from 1918 to 2024 (Figure 1(B), data from NCBI). The analysis results of sequence data from both databases reveal consistent mutation patterns. H274Y mutation was rarely reported before 2000. In 2001, the H274Y mutation rate reached 0.709% (1/141) in NCBI database and 0.592% (1/169) in GISAID database, but the mutant strain was not detected for the following three years. The strain gradually spread starting in 2005, reaching its peak prevalence in 2008. From 2005 to 2008, the prevalence of the H274Y mutation increased rapidly, rising from 1.575% (2/127) to 19.916% (427/2144) in the NCBI database and from 0.469% (1/213) to 37.692% (1078/2860) in the GISAID database. Sequencing data indicate that the prevalence of H274Y mutant strains increased by 12.6-fold in NCBI and by 80.4-fold in GISAID during this period. The mutation rate of H274Y was 15.123% (1264 / 8358) in the NCBI database and 11.540% (1334/11560) in GISAID database in 2009. From 2010 to 2014, the mutation rate of H274Y remained moderate in both the NCBI and GISAID databases, averaging around 3.5%. Since 2015, the N1 strain carrying the H274Y mutation has consistently maintained a low mutation rate in both databases.

Figure 1.

N1 H274Ymutation rate. (A). Worldwide prevalence of the N1 H274Y strain from 2000 to 2024. (B) Mutation rates of N1 H274Y strains in various countries from 1918 to 2024. A total of 17916 and 119564 NA protein sequences of influenza virus subtype N1 were downloaded from NCBI's influenza virus database and GISAID EpiFlu^TM database, respectively and then subjected to multiple sequence alignment using MAFFT. Python was used to analyze the mutation at the site of H274. The mutation rate was calculated as the frequency of H274Y mutation in the NA sequences of influenza virus isolated globally at a certain year. The sequences from NCBI were also used to calculate the mutation rates of N1 H274Y in various countries from 1918 to 2024. Only countries with more than 100 sequences data are counted.

Interestingly, the mutation rate of H274Y (A(H1N1) pdm09 high-frequency mutation site) in 2009 was lower than that in 2008, indicating that the ancestral strain of A(H1N1) pdm09 may have spread to humans in the months before this outbreak [78]. In April 2009, an epidemic broke out; the strain was named A(H1N1) pdm09. Since 2009, the A(H1N1) pdm09 replaced the previously seasonal H1N1 viruses.

According to the NA data obtained from NCBI, the frequency of H274Y mutation in influenza strains in various countries from 1918 to 2024 was compared (Figure 1(B)). The mutation rates were as follows: Argentina: 2.52%, Australia: 6.98%, Australia: 10.34%, Brazil: 2.87%, Canada: 4.89%, China: 10.10%, Denmark: 4.69%, Finland: 0.55%, France: 11.68%, Germany: 3.48%, Greece: 2.26%, India: 2.56%, Iran: 6.59%, Italy: 6.32%, Japan: 24.50%, Malaysia: 4.55%, Mexico: 6.85%, New Zealand: 4.71%, Nicaragua: 2.84%, Russia: 12.45%, Singapore: 1.03%, Spain: 4.75%, Thailand: 10.12%, UK: 6.14%, The USA: 11.96%, and Vietnam: 19.63%. The H274Y mutation frequency is over 10% in Australia, China, France, Japan, Russia, Thailand, the USA, and Vietnam, indicating that this strain may cause a burden on the healthcare systems of these countries; The high mutation rate (up to 20%) of H274Y may lead to severe global resistance to oseltamivir. Drug selective pressure leading to the large-scale spread of H274Y-resistant mutant strains may be one of the driving factors behind its spread [79].

Although the H274Y mutation has been circulating since 2001 or earlier, the reason why it has not caused a global epidemic is that its NA function is impaired; this decreased its adaptability and transmissibility. However, with NA antigenic drift, the H274Y mutant influenza virus has gradually become more virulent [80]. Fortunately, although the current N1 seasonal influenza is A(H1N1) pdm09, where H274Y mutations are common, the vast majority of circulating viruses do not carry the H274Y mutation and remain sensitive to oseltamivir.

Mechanisms of influenza virus resistance to NA inhibitors

NAIs have a strong affinity for NA; they compete with the influenza virus NA substrate N-acetylneuraminate (DANA) for the binding site of NA, inhibiting the hydrolysis of sialic acid and preventing virus spread to uninfected cells [6]. Zanamivir is an analog of DANA, formed by replacing the hydroxyl group at the C4 position with a positively charged guanidyl group [81]. This group, which is highly alkaline and has a wider range of positive charge distribution, occupies the small pocket on the side of the binding cavity, resulting in charge-based interactions such as hydrogen bonding and electrostatic binding with multiple amino acid residues (Glu119, Asp151, and Glu227) [81]. Comparing oseltamivir with the natural substrate DANA, the glycerol side chains of DANA C6 and C4-OH are replaced with – NH2 and pentyl ether side chains [82].

The mutation of NA leads to a variety of drug resistance mechanisms. Different subtypes of NA have different structures, and different site mutations lead to different resistance mechanisms to the same drug; even the same site mutation also has different resistance mechanisms for different drugs. Based on the complex characteristics of these mutations, three typical mechanisms are summarized for influenza drug resistance caused by NA mutations as follows.

Mutation disrupts hydrogen bonding interaction between NA and its inhibitor

Mutation in the NA active site directly leads to a decrease in the drug interaction capacity, resulting in drug resistance. Catalytic residues (R118, D151, R152, R224, E276, R292, R371, Y406) can directly interact with substrates and participate in catalytic reactions [62,83]. When these sites undergo mutations, a decline in drug-binding capacity may directly result. E276D is located in the active binding region S5, which belongs to the mixed polarity region and can accommodate hydrophobic units as well as polar units such as the glycerol side chains of sialic acid and zanamivir, forming hydrophobic interactions with the methylene group of E276 [84]. When glutamic acid at position 276 mutates into aspartic acid, the resistance to zanamivir increases, as the mutation prevents hydrogen bonding to glycerol groups O8 and O9 in zanamivir (Figure 2) [62].

Figure 2.

Catalytic site amino acid E276D mutation directly leads to a decrease or disappearance of hydrogen bonding between the binding pocket and the NAI. (A) E276 and zanamivir glycerol side chain form hydrogen bonding. (B) D276 and zanamivir space distance becomes larger, resulting in hydrogen bonding formation blocked.

Mutation causes steric hindrance between NA and its inhibitor

Framework amino acids (E119, R156, W178, S179, D/N198, I222, E227, H274, E277, N294, and E425) maintain the spatial conformation of enzyme active centers [83,85]. The NA sites interact with each other, and mutations to structural amino acids affect the structure of other sites around the hydrophobic pockets, leading to malformation or reduced local hydrophobicity which reduces the binding capacity of NAIs with hydrophobic side chains containing pentyl ether groups (oseltamivir or peramivir) to NA [8,86]. In the first phylogenetic group of influenza viruses (N1, N4, N5, N8), H274Y mutation affects the spatial orientation of E276 and causes the residue to get closer to the binding site; as a result, the hydrophobic pocket becomes smaller, and the affinity of oseltamivir with hydrophobic side chain to NA decreases. (Figure 3) [8,86,87]. In the second phylogenetic group (N2, N3, N6, N7, N9), when H274Y mutation replaces residue 274 with the large-volume group tyrosine (Y), the small threonine at residue 252 means that there is enough space to accommodate the tyrosine without affecting the spatial orientation of residue 276; because of this, this mutation does not affect oseltamivir binding [88].

Figure 3.

Mutation of framework amino acid H274Y causes spatial changes in E276, resulting in narrower hydrophobic pockets for NAI binding. (A) H274 wild-type makes oseltamivir hydrophobic backbone pentyl ether groups smoothly enter NA hydrophobic pocket. (B) Y274 mutant changes the orientation of E276 residue to narrow NA hydrophobic pocket.

One of the important reasons behind the resistance to zanamivir caused by guanidine is the mutation of the active site, which hinders the formation of hydrogen bonds between certain amino acid residues in the 150 ring; this results in a special configuration that interferes with the shape of the drug binding pocket [89]. In addition to directly affecting hydrogen bonding to drugs, the evolution of NA mutations can also indirectly affect the formation of one or more local hydrogen bonding networks between amino acid residues, limiting the free rotation of amino acid residues in the active site and spatially restricting the binding of NAIs [29,58].

Mutation leads to change in tetramer conformation of NA

The tetramer structure of NA requires each monomer to maintain mutual stability under certain energy maintenance. Mutations at certain sites can change interface bond energy between monomers, disrupting NA stability and reducing the sensitivity of the virus to NAIs. According to one report, after E105 K unit point mutation at the NA monomer interface, the free energy of the hydrogen bond formed between Lys and Gly at residue 141 decreased from – 493 kcal/mol to – 481 kcal/mol; this reduced the stability of the NA tetramer and resulted in a 15-fold and 213-fold decrease in sensitivity of virus to zanamivir and peramivir, respectively (Figure 4) [35].

Figure 4.

E105 K mutation causes changes in interfacial bond energy between monomers, disrupting the stability of NA and leading to a decrease in receptor affinity. (A) E105 maintains the stability among NA monomers, and NAIs could combine with NA smoothly. (B) K105 decreases the stability among monomers and reduces the affinity between NAIs and NA.

The impact of NA mutations on viral adaptability and transmission

Partial drug-resistant NA-mutant viruses may not be able to survive and spread well when they first appear, because long-term natural selection of the most suitable influenza virus is likely to cause some loss of function from accidental mutations, even if they can tolerate drug stress.

In the study of the evolution of NAs from 1999 to 2009, some researchers found that some site mutations produced antigen drift and gradually became common with the emergence of evolutionary pressure and epistatic interaction [80]. NA functionality is gradually reduced when H274Y is present, though the functional defects caused by the mutation are completely compensated after D354G mutation [80]. NA mutants that lack adaptability and transmissibility initially can also survive and spread at the host level because of epistatic effects. In 2007, the H274Y-mutated H1N1 influenza virus was found in Europe and widely spread to South Africa, Australia, New Caledonia, the Philippines, and other countries with low oseltamivir use, resulting in oseltamivir-resistant H1N1 being prevalent (85%) within 11 months [64]. This demonstrates that the H1N1 virus with H274Y mutation has significantly improved adaptability and transmissibility without drug pressure. I222 V can also compensate for the deficiencies caused by H274Y mutation; in addition, this mutation also has a synergistic effect on oseltamivir and peramivir resistance phenotypes of H274Y mutants [44]. The emergence of a double mutant virus with cross-resistance to multiple NAIs and high adaptive transmission is a concern requiring vigilance; once this type of virus spreads on a large scale, the consequences may be serious.

Although the adaptability and transmissibility of drug-resistant mutant viruses have been improved under selection pressure and epistasis, they are still surpassed by those of wild-type viruses. The replication and transmission of oseltamivir-resistant H274Y (seasonal H1N1), E119 V and R292 K (seasonal H3N2), and H274Y (A/H1N1 pdm09) in animal models are less than those of wild-type influenza [90]. If this were not the case, it would lead to a horrible drug-resistant virus pandemic, which could also be detrimental to the influenza virus itself in the long run. Fortunately, there have been no significant differences in clinical manifestations, complications, and hospitalization rates between the H274Y-mutated oseltamivir-resistant influenza virus H1N1 and the seasonal oseltamivir-sensitive influenza virus H1N1 [91].

Although some viral functions decrease in the early stages of mutation, negatively impacting survival, it is still possible that virus strains with higher resistance, adaptability, and transmission develop over time, compensating for mutations. Keeping ahead of this requires constant vigilance and accelerated development of new drugs against influenza viruses, especially those affecting drug-resistant strains.

NAI-resistance may render standard treatments ineffective, increasing the risk of severe illness and death, especially in vulnerable populations (e.g. the elderly, children, and immunocompromised individuals). Public health policies must adapt by regularly monitoring viral mutations and resistance patterns, updating treatment guidelines accordingly.

NA mutations can also affect the efficacy of flu vaccines. Although hemagglutinin (HA) has been the main target of influenza virus vaccine development, NA has gained increasing attention in the past few years [92]. Significant mutations in NA, like H1N1 virus with H274Y mutation, may necessitate adjustments to the vaccine strain selection. Public health agencies must ensure that vaccines are updated in response to circulating virus strains to maintain effectiveness.

Conclusion

The emergence of antiviral drug resistance is a major challenge to the treatment of influenza infection. In the review, the resistance rates of the four major surveillance influenza strains to existing NAIs generally showed an upward trend, which was a warning signal. In addition, as one of the main seasonal epidemic strains, the resistance of H1N1 pdm09 to NAIs is mostly related to H274Y mutation. Therefore, further research on the mutation of this site is one of the focuses of this review.

This review also updates the NA-resistant mutation sites discovered from 2016 to 2024, with the influenza B virus accounting for 80% of these (Table 2), which is very important for future influenza surveillance and prevention because there are two subtypes of influenza B virus (B/Victoria, B/Yamagata) in the four subtypes of influenza surveilled in various countries. Furthermore, this review summarizes the drug resistance caused by some NA site mutations, especially active site mutations. Many site mutations may cause significant or even severe resistance to the four existing NAIs, which greatly affects the prevention and treatment of influenza. Interestingly, the same mutation at some sites in NA in different strains may produce different or even opposite drug resistance phenotypes. Because of the influence of epistatic interaction, NA multi-site mutations often produce synergistic drug resistance effects, though antagonistic effects are occasionally seen. Both single and multiple point mutations in NA may lead to single-drug and cross-resistance to existing NAIs.

When investigating the global prevalence of the N1 H274Y strain from 2000 to 2024, it was found that the influenza virus carrying the H274Y mutation became widespread globally, reaching its peak prevalence in 2008. The H274Y mutation in H1N1 was of little concern until 2007. However, as the virus carrying this mutation continued to spread, the mutation rate in seasonal influenza H1N1 reached 20% in the NCBI Influenza Virus Database and even 37% in the GISAID EpiFlu™ database by 2008. Due to the undetected transmission of gene fragments from three distinct porcine genetic lineages (Eurasian swine, classical swine, and triple reassortant swine)，the natural recombination of these gene fragments led to the emergence of H1N1pdm09, characterized by high drug resistance and elevated transmissibility. This combination virus first emerged in April 2009 and eventually replaced the seasonal H1N1 strain [93].

From reports of influenza virus drug resistance mutations, we summarized that the mechanism of influenza virus NAI resistance mainly includes three mechanisms: first, catalytic amino acid mutation of the binding pocket causing a direct inhibitory effect, followed by framing or other amino acid site mutation causing an indirect structural hindrance, and finally amino acid site mutations at the monomer junction causing changes in monomer stability. In general, changes in NA amino acid residues lead to NAIs resistance, all by reducing the affinity of NA binding pockets to NAIs.

The mutation NA with poor adaptability and transmission continues to evolve, generating new mutations that can compensate for the functional defects caused by NA mutations, resulting in a qualitative leap in the viability and transmission ability of drug-resistant strains. The combination of terrible drug resistance and transmissibility will seriously threaten human public health security. Mutations that increase the transmissibility or virulence of influenza viruses have the potential to trigger new outbreaks or pandemics. Fortunately, according to global influenza surveillance data, including the WHO's Global Influenza Surveillance Report and biweekly influenza statistics, the resistance rate of H1N1pdm09, H3N2, and influenza B viruses to the four neuraminidase inhibitors (NAIs) has remained stable at approximately 1% over the past decade. Only a few isolated viruses have exhibited resistance to one or two specific NAIs, and these resistant strains have not become dominant in widespread outbreaks. Additionally, the mutation rate of H1N1pdm09 carrying the H274Y mutation has remained low. Based on this data, we predict that a global pandemic driven by drug-resistant H1N1pdm09 is unlikely in the coming years.

Although there are still few cases of high NAI resistance rates reported for seasonal influenza viruses, this must not be an excuse for researchers to stop the pace of new drug development; instead, research into new NAIs should be intensified to fully respond to the prevalence of new multi-drug resistant strains.

In addition, global influenza surveillance networks and national public health authorities must enhance real-time monitoring of viral mutations. Based on surveillance data, policies may need to be adjusted, including increasing vaccine coverage, developing broadly protective vaccines, promoting prudent antiviral use, and strengthening isolation or quarantine measures in affected regions.

Funding Statement

This work was supported by the Shenzhen Science and Technology Project (JCYJ20190808122605563 and JCYJ20220530153206014), the Natural Science Foundation of Top Talent of SZTU (GDRC202121), and the Fundamental Research Funds for Shenzhen Technology University (20211063010049).

Disclosure statement

No potential conflict of interest was reported by the author(s).