Genetic and antigenic characterization of influenza A(H3N2) in Cameroon during the 2014-2016 influenza seasons

The first outbreak of influenza A(H3N2) occurred in 1968 and caused the third flu pandemic of the 20th century. It has affected multiple countries over time. The best strategy to reduce the burden of influenza is through vaccination whose efficacy varies with respect to the circulating strains. This study was performed to better understand the molecular evolution of influenza A(H3N2) and assess vaccine efficacy in Cameroon. Complete sequences of three gene segments were obtained from 2014 to 2016 influenza seasons in Cameroon. Hemagglutinin (HA), Neuraminidase (NA) and matrix (M) genes of 35 A(H3N2) virus strains were amplified and sequenced. Predicted vaccine efficacy was measured using the Pepitope model. Phylogenetic analysis of the HA gene showed that all Cameroonian strains had evolved away from the 3C.1-A/Texas/50/2012-like clade. Globally, 2014 virus strains clustered with the 2015–2016 vaccine strain, 3C.3a-A/Switzerland/9715293/2013, whereas 2015 and 2016 virus strains clustered with the 2016–2017 vaccine strain, 3C.2a-A/HongKong/4801/2014. In order to determine the genotypic drug susceptibility to neuraminidase inhibitors and amantadine, the NA and M2 protein coding sequences were analyzed. There was no strain with characteristic mutation for resistance to neuraminidase inhibitors, per contra; all strains possessed the substitution S31N, peculiar of resistance to adamantanes. There was drift in influenza A(H3N2) dominant epitopes B (2014 and 2015) to epitopes A (2016) with a theoretical efficiency in vaccine ranging from low to moderate. The presence of several antigenic site mutations among H3N2 virus strains between 2014–2016 influenza seasons in Cameroon confirms the progressing evolution of circulating H3N2 strains.


Introduction
Influenza A virus is a major cause of acute respiratory disease in humans and is responsible for approximately 250,000-500,000 deaths annually worldwide [1]. Pandemic influenza A virus infection resulted in significant morbidity and mortality in 1918 (H1N1), 1957 (H2N2), 1968 (H3N2), and 2009 (H1N1) [2,3]. The first outbreak of influenza A/H3N2 was reported in 1968 and has since affected multiple countries [2]. In several seasons, A(H3N2) strains are more prevalent than the other co-circulating viral sub-types [4,5] with high morbidity and mortality rates [6,7]. Influenza A nomenclature is based on the genotypic properties of two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA). Currently, 18 HA and 11 NA subtypes have been recognized [8]. The primary target of host neutralizing antibodies is the HA glycoprotein where it inhibits binding to the sialic acid receptors present in the respiratory tract [9]. There are five identified antigenic sites for A(H3N2), namely, A, B, C, D and E [10]. Non-recognition of some strains by neutralizing antibodies present in the host is the consequence of mutations at these sites [11]. Continuous build-up of mutations at these antigenic sites is the basis for the evolutionary dynamics observed in the influenza virus resulting in antigenic drift [12]. About seven A(H3N2) clades (designated clades 1 to 7) and many sub-clades have evolved over the past few years, providing a significant challenge for the annual design of effective influenza vaccines [13]. At present, viruses in clade 3 are the dominant group, which has diversified further into three major subclades designated 3A, 3B, and 3C, with subclade 3C containing several distinguishable genetic groups [13,14].
The best strategy to reduce the burden of influenza is through vaccination whose efficacy varies with respect to the circulating strains [15]. Although influenza vaccination is very limited in most of Africa, particularly Sub-Saharan Africa, there have been previous reports of introduction of viruses that have undergone drift due to antiviral and vaccine pressure [16,17] from other regions. There is therefore, a crucial need for monitoring the genetic and antigenic characteristics as well as drug susceptibility of influenza viruses as a component of pandemic preparedness [16]. To better understand the molecular evolution of influenza and assess vaccine efficacy in Cameroon, three gene segments of A(H3N2) from 2014 to 2016 were characterized.

Sample collection and preparation
Respiratory samples were collected between January 2014 and June 2016 from outpatients with clinical evidence of influenza-like illness (ILI), defined as person with sudden onset of fever >38˚C and cough or sore throat, at 12 sentinel sites involved in influenza surveillance in Cameroon. Naso-pharyngeal and/or oropharyngeal swabs were collected and put in 2 ml cryovials containing virus transport medium and stored at 4˚C before transportation to the Pasteur Centre of Cameroon (PCC) for analysis. All samples were stored anonymously. This study was specifically approved by the Cameroon National Ethics committee (N˚2016/08/798/CE/ CNERSH/SP). Viral RNA was extracted then from 140 μL of clinical samples with Qiagen RNA kit according to the manufacturer's instructions into a final volume of 60 μL elution buffer. Extracts were analyzed for the presence of influenza by a real-time reverse transcription-polymerase chain reaction (RT-PCR) in an ABI Prism 7300 or 7500 thermocycler (Applied Biosystems, Foster City, California, USA). Positive samples for influenza H3N2 with Ct below 30 after subtyping were eligible for sequencing. A convenient sample size of 35 was randomly selected from the flu database based on their geographic origin and distribution over time. Samples coinfected with other influenza subtypes were not included.

Amplification
The first PCR was performed using Superscript III One Step RT-PCR System (Invitrogen, Carlsbad, USA) and a set of primers (S1 Table). The following procedure was used for amplification of 1192 bp of HA gene, 1459bp of NA gene and 1027 bp of M gene of influenza A(H3N2). Overall a total reaction volume of 25 μL contained the following reagents: 12.5 μL of 2x PCR buffer, 0.5 μL of 10 μM forward and reverse primer, 0.12 μL RNAsin 40 U/μL, 0.5μL Superscript RT/Platinum Taq enzyme and 5μL RNA. This mixture was run using the following program for HA and M genes: 45˚C for 30min, 55˚C for 15min, 94˚C for 2min, 30 cycles at (94˚C for 45sec, 45˚C for 45sec, 72˚C for 5min) and 72˚C for 5min. For NA gene, the following program was used: 45˚C for 30min, 55˚C for 15min, 94˚C for 2min, 15 cycles at (94˚C for 30sec, 55˚C for 30sec, 72˚C for 2min), 15 cycles at (94˚C for 30sec, 45˚C for 30sec, 72˚C for 3min) and 72˚C for 5min. The first PCR product was then used as a template for semi-nested PCR. A total reaction volume of 50 μL contained 38.3 μL of water, 5 μL of 10x PCR buffer, 1 μL of 10 mM dNTP mix, 1.45 mM of 50 mM MgCl 2 , 1μL of 10μM forward and reverse primer, 0.25 μL of Taq DNA polymerase 5U/μL and 2μL of first PCR product. This mixture was run using the following program for all gene segments: 94˚C for 5min, 30 cycles at (94˚C for 45sec, 45˚C for 45sec, 72˚C for 90sec) and 72˚C for 5min. PCR products were run in a gel and bands corresponding to the appropriate size were sent for sequencing at GENEWIZ UK, LTD (Hope End, UK) with the corresponding primer sets.

Phylogenetic analysis and antigenic characterization
Sequences were edited and assembled with CLC Main Workbench version 5.5. The sequences of reference strains of known clades and northern hemisphere vaccine strains recommended by WHO [18] included in phylogenetic analysis were obtained from the GenBank and GISAID databases. The phylogenetic tree of the coding nucleotide sequences was generated by MEGA version 6.0 using a neighbor-joining method. The number of bootstrap replications was set to 1000 and bootstrap values above 50% were labeled on major tree branches for reference. The amino acid residues in epitopes A to E of influenza A(H3N2) was previously identified [19].

Nucleotide sequence accession numbers
Influenza A(H3N2) virus sequences included in the analysis were submitted to Genbank under accession numbers KY653817 to KY653919. S2 Table provides detailed information for the Cameroon influenza gene sequences.

Prediction of glycosylation sites
The prediction of potential N-liked glycosylation sites was performed with an online server: NetNGlyc 1.0. [20]. This server considers the amino acid alignment Asn-X-Ser/Thr, where X can be any amino acid except Asp or Pro. A threshold value of >0.5 suggests glycosylation.

Prediction of vaccine efficacy using P epitope model
The predicted vaccine efficacy of the influenza A(H3N2) seasonal influenza virus was estimated using the P epitope method [21][22][23]. P epitope measures the antigenic distance between the vaccine strain and the dominant circulating strains, and it is calculated as follows: (number of mutations in the dominant epitope)/(number of amino acids in the dominant epitope). The mathematical formula linking vaccine efficacy for influenza A(H3N2) and P epitope is given by E = −2.47 × P epitope + 0.47 for [23]. When the P epitope between the vaccine and circulating strains is null, the influenza A(H3N2) vaccine efficacy is 47% [21]. This model correlates best with vaccine efficacy as compared to phylogenetic analyses or antisera hemagglutination inhibition assay [21].

Results
A total of 103 gene segments from the 35 virus strains were analyzed (35 HA, 35 NA and 33M). Nucleotide and amino acid sequences for the 3 selected genes were compared within the Cameroon virus strains and with the WHO vaccine strains, WHO reference strains and available sequences of other African virus strains.
Amino acids at positions 26, 27, 30 and 31 of the transmembrane region of the M2 protein are the residues frequently associated with amantadine-and rimantadine-resistant strains [33]. Amino acid sites were checked as follows for mutations associated with reduced susceptibility or resistance to adamantanes: L26F, V27A/T, A30T/V, S31N/R, G34E, A30V, S193F and R45H [25,26]. All strains possessed the S31N amino acid substitution.

Estimation of vaccine efficacy for A(H3N2)
To assess the effect of the accumulated mutations in the HA1 domain on predicted vaccine efficacy in a given year, the P epitope method was used to evaluate how closely the vaccine strain resemble the circulating strain (Table 2). Theoretically, when P epitope in the dominant epitope is higher than 0.19, the vaccine efficacy becomes negative [21,22]. For 2014, the HA1 sequences mostly had a dominant mutation in epitope B (128,159,186,198) and the P epitope of 0.191 with respect to A/Texas/50/2012 vaccine strain, suggesting that the latter poorly matched the multiple circulating strains present that year consequently leading to a negative vaccine efficacy against these strains of -0.11% (E = -0.05% of 47%, P epitope = 0).

Discussion
This is the first report of genome characterization of influenza H3N2 in Cameroon. We analyzed the coding sequences of 3 gene segments of influenza A/H3N2 virus strains from 2014 to 2016 influenza seasons in Cameroon and compared them with those of the WHO reference strains, WHO recommended vaccine strains [18]and with other epidemiologically relevant African strains from both GenBank and GISAID databases. We observed different phylogenetic clusters generated for different gene segments. The HA gene sequences indicated that all the Cameroon strains had evolved away from the 3C.1-A/Texas/50/2012-like clade by acquiring the genetic markers N145S, Y186G, P198S and F219S which has been identified as being characteristic of clades 3C.2 and 3C.3 [13,34] [22]. Several amino acid mutations on epitope A of the HA gene were responsible for the drift between the circulating strains and the 2014-2015 vaccine strains, 3C.1-A/Texas/50/2012-like strain. This corroborates with the P epitope > 0.19 observed, and eventually resulting in a negative vaccine efficacy. The vaccine change to the A/Switzerland/9715293/2013-like strain for the following season resulted in more than 80% of the circulating H3N2 strains not covered by the selected vaccine. Notwithstanding the poor results with the previous influenza seasons, the 2016-2017 selected vaccine strain showed moderate efficacy when compared to the circulating virus strains. It is noteworthy that vaccination against influenza virus, is not a common phenomenon in the Cameroon population, however a better selection of annual vaccine with the use of P epitope as suggested by other authors, [21] will improve vaccine efficacy and encourage uptake of flu vaccine. This may result in reduced burden caused by influenza due to herd immunity [35]. The acquisition of a mutation in influenza A viruses conferring resistance to an antiviral agent may occur as a result of drug selection, spontaneous mutation or through genetic reassortment with another drug resistant influenza A virus [26]. Neuraminidase inhibitors target the NA glycoprotein which plays a key role in the production of new viral progeny. There have been rare reports of resistance to neuraminidase inhibitors in influenza H3N2 virus strains [27], and this was confirmed Cameroonian virus strains by the absence of previous mutations characteristic of this resistance. There was however a novel amino an amino acid mutation, Q136H, that had not been previously reported. Further investigation on the clinical relevance of this mutation will be essential for a better understanding of the molecular evolution of influenza A(H3N2).
Amantadine resistance reported in most H3N2 viruses worldwide was confirmed in the Cameroonian virus strains by an S31N substitution in the M2 protein [25,27]. High prevalence of amantadine resistance in influenza A viruses was observed in other countries, irrespective of its use [25,26]. Resistance to M2 inhibitors first appeared following extensive drug use in Asia and the US after the SARS (Severe Acute Respiratory Syndrome) epidemic in 2004. Since there is no history of exposure to anti-viral drugs against influenza and eventually to adamantanes, Cameroonian virus strains with S31N mutation indicated that strains already harbouring drug resistant mutation were most likely introduced in the community. These resistant strains probably have a selective advantage over the sensitive strains that favour their evolution and fitness [36].
As reported by Mostafa et al., glycosylation and deglycosylation are important viral mechanisms to (i) mask antigenic epitopes and thus inhibit binding to neutralizing antibodies, (ii) adjust receptor-binding affinity, or (iii) regulate virulence of influenza [14]. The Cameroonian H3N2 virus strains shared at least nine potential N-glycosylation sites with the 2014-2017 vaccine strains, and had similar glycosylation patterns. Group 3C.2a-like strains had acquired 2 additional glycosylation sites: 126NWT128 and 158NYT160. The predicted HA glycosylation patterns may contribute to a potential antigenic escape of influenza H3N2 viruses from antibodies raised against vaccine strain and previous circulating strains or increase viral fitness by yet-to-be-characterized mechanisms.

Conclusion
The presence of several antigenic site mutations with high frequency among H3N2 virus strains between 2014-2016 influenza seasons confirms the progressing evolution of circulating H3N2 strains. These results lay more emphasis on the relevance of routine influenza surveillance as well as the genetic and antigenic characterization in facilitating the prompt and efficient selection of the influenza strains which will be included in the flu vaccine. This data will be very valuable in tracking the phylogenetic evolution of influenza viruses within sub-Saharan Africa.
Supporting information S1