Molecular characterization of influenza viruses collected from young children in Uberlandia, Brazil - from 2001 to 2010

Background Influenza remains a major health problem due to the seasonal epidemics that occur every year caused by the emergence of new influenza virus strains. Hemagglutinin (HA) and neuraminidase (NA) glycoproteins are under selective pressure and subjected to frequent changes by antigenic drift. Therefore, our main objective was to investigate the influenza cases in Uberlândia city, Midwestern Brazil, in order to monitor the appearance of new viral strains, despite the availability of a prophylactic vaccine. Methods Nasopharyngeal samples were collected from 605 children less than five years of age presenting with acute respiratory disease and tested by immunofluorescence assay (IFA) for detection of adenovirus, respiratory syncytial virus, parainfluenza virus types 1, 2, and 3 and influenza virus types A and B. A reverse transcription-PCR (RT-PCR) for influenza viruses A and B was carried out to amplify partial segments of the HA and NA genes. The nucleotide sequences were analyzed and compared with sequences of the virus strains of the vaccine available in the same year of sample collection. Results Forty samples (6.6%) were tested positive for influenza virus by IFA and RT-PCR, with 39 samples containing virus of type A and one of type B. By RT-PCR, the type A viruses were further characterized in subtypes H3N2, H1N2 and H1N1 (41.0%, 17.9%, and 2.6%, respectively). Deduced amino acid sequence analysis of the partial hemagglutinin sequence compared to sequences from vaccine strains, revealed that all strains found in Uberlândia had variations in the antigenic sites. The sequences of the receptor binding sites were preserved, although substitutions with similar amino acids were observed in few cases. The neuraminidase sequences did not show significant changes. All the H3 isolates detected in the 2001-2003 period had drifted from vaccine strain, unlike the isolates of the 2004-2007 period. Conclusions These results suggest that the seasonal influenza vaccine effectiveness could be reduced because of A H3N2 variants that circulated in 2001-2003 years. Thus, an early monitoring of variants circulating in the country or in a region may provide important information about the probable efficacy of the vaccine that will be administered in an influenza season. Electronic supplementary material The online version of this article (doi:10.1186/s12879-015-0817-z) contains supplementary material, which is available to authorized users.


Background
Influenza viruses are pathogens responsible for causing respiratory disease worldwide and represent a major threat to public health due to annual epidemics and the potential of pandemics. These agents are members of Orthomyxoviridae family and are classified into types A, B and C. Influenza viruses of type A can be further divided into subtypes based on the antigenic properties of the surface glycoproteins hemagglutinin (HA) and neuraminidase (NA) [1]. Currently, 18 HA and 11 NA variants are known [2] and they are used to subtype the influenza virus isolates, with subtypes H1N1 and H3N2 characterized by the sustained transmission in the human population [3]. HA and NA are under selective pressure and undergo frequent changes through antigenic drift, resulting in the appearance of new strains that evade the host immune system [4]. Distinct antigenic sites located in the H1 sequence (Sa, Sb, Ca and Cb) and five major variable sites within domain 1 of H3 (A through E) are targets of neutralizing antibodies and virus isolates carrying substitutions at these sites have been related with epidemics in humans in limited time periods [5,6]. Vaccination is the main strategy to reduce the public health burden caused by influenza. However, its effectiveness depends on how close the HA sequences of the virus strains used in the vaccine are to those circulating in the same influenza season [7]. Sequence analysis of the glycoprotein genes allows the identification of new strains and may also reveal geographical regions where the detected strains have circulated. Therefore, knowledge of the circulating strains and genetic analysis of HA and NA are invaluable to guide effective measures for influenza prevention and control. In Brazil, influenza virus sequences and strain information are currently available from only a few locations and surveillance efforts have been increased after the emergence of the pandemic influenza A/H1N1 virus in humans, in 2009 [8]. In Uberlândia, in the southeastern region of the country, however, information about circulating strains is not known. Thus, the purposes of this study were to detect the influenza viruses in young children presenting with acute respiratory disease, to characterize the strains from the identified cases and to compare the HA and NA sequences of these strains with the sequences of the virus variants used in the vaccines of the same flu season.

Specimens
Nasopharyngeal aspirates (NPA) were collected from children under five years old presenting with acute respiratory disease (ARD), within five days of onset of clinical symptoms, as described by Costa et al. [9]. The collection of clinical specimens was performed in public health service units in Uberlândia city, in the state of Minas Gerais, southeastern Brazil, from 2001 to 2010 and aliquots from the specimens were stored at −70°C. Initially, the specimens were screened by immunofluorescence assay (IFA) with the Respiratory Panel I Viral Screening and Identification Kit (Millipore/Chemicon International, Inc., Temecula, CA, USA), following the manufacturer's instructions, to detect the presence of influenza viruses A and B, respiratory syncytial viruses A and B, parainfluenza virus types 1, 2 and 3 and adenovirus. RT-PCR for detection of influenza virus RNA was used with all samples that tested positive for influenza virus by IFA and, depending on the availability, with IFA-negative or -inconclusive samples. The study was approved by the Ethics Committee of Universidade Federal de Uberlândia (protocol number 326/08) and a written consent was obtained by one of the patient's parents.

RNA extraction and RT-PCR
Total RNA was extracted from 140 μL NPA by using Viral RNA mini kit (Qiagen, Hilden, Germany) according to the manufacturer's recommendations.
For reverse transcription, the reaction was carried out in a total volume of 20 μL, which contained 5 μL total RNA, 0.5 mM each deoxynucleoside triphosphate, 1 μM primer for RT (Table 1) [10], 0.1 M DTT, 100 U SuperScript™ III Reverse Transcriptase (Life Technologies Corp, Grand Island, NY, USA), and buffer supplied by the enzyme manufacturer. The first three components were mixed, incubated at 95°C for 3 min and cooled on ice for 1 min. Then, the other components were added to the tube, the volume was completed with diethyl pyrocarbonate (DEPC)-treated water, the mixture was incubated at 50°C for 60 min, and the enzyme was inactivated at 95°C for 3 min. Four μL cDNA was submitted to first-round PCR, in a total reaction volume of 20 μL, which contained 1 mM MgCl 2 , 0.2 mM of each deoxyribonucleoside triphosphate, 1 or 0.25 μM each primer (for HA or NA sequence amplification, respectively) (each sequence was amplified separately), 1 U of Platinum Taq DNA polymerase (Life Technologies) and buffer supplied by the enzyme manufacturer. The reaction was carried out with the following conditions: 94°C for 1 min; 35 cycles of 94°C for 1 min, 50°C or 55°C (respectively, for HA or NA sequence amplification) for 1 min, and 72°C for 1.5 min. The first-round amplification product was diluted 1:100 in water and 0.8 μL was submitted to a second-round amplification, with similar conditions used in the first-round PCR, except for the following: primer concentrations were changed to 0.125 or 0.25 μM (HA or NA, respectively), MgCl 2 to 1.5 mM, and annealing temperature to 52 or 57°C (HA or NA, respectively) for 1 min. PCR products were separated in a 1.0% agarose gel in tris-borate-EDTA buffer [11], stained with GelRed TM (Biotium, Hayward, CA, USA) and visualized under ultraviolet light.

Nucleotide sequencing and analysis
The amplified products were purified by using GenElute PCR Clean-Up Kit (Sigma-Aldrich, Inc., St. Louis, MO, USA) and sequencing was performed by using either DYEnamic™ ET Dye Terminator Kit and MegaBACE 1000 sequencer (GE Healthcare, Buckinghamshine, UK) or BigDye Terminator v.3.1 and ABI-Prism 3100 Genetic Analyzer (Applied Biosystems/Life Technologies, Foster City, CA, USA), following the manufacturer's instructions. Sequences were edited by using SeqMan Pro (Lasergene version 10, DNASTAR, Inc. Madison, WI, USA) and search of similar sequences was carried out using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Sequence alignment was performed by using the Clustal W method in MegAlign (DNASTAR). The nucleotide sequences of the characterized isolates in this study were submitted to GenBank under accession numbers KF918346 to KF918392.

Results
A total of 605 clinical samples were collected from patients with ARD between 2001 and 2010 and tested by IFA. Thirty seven (6.1%) samples showed to be reactive for influenza viruses: 36 of type A and one of type B; and from remaining 568 samples, 254 tested negative, 122 were inconclusive, and 192 were positive for other respiratory viruses (HRV, RSV, hMPV, PIV and Adv), which were the subject of other studies [9,[12][13][14][15]. Among the negative and inconclusive samples, 218 were tested by RT-PCR for the amplification of HA and NA gene segments and three other positive samples for influenza virus type A were detected. The remaining 158 samples were not tested due to insufficient amount. Therefore, the total prevalence found for these viruses in the study was 6.6% (40/605).
Most children were outpatients (33/40; 82.5%), males (26/40; 65%; p = 0.013), and the median age was 14.5 months, ranging from one to 60 months. Influenza cases peaked in July, though they were observed from February to September in the time period of the study.
Out of the 40 positive samples, the HA sequence amplification by subtype-specific PCR was obtained from 25 of them: 16 were subtyped as H3, eight as H1, and one was characterized as type B. Regarding NA sequence, four specimens were subtyped as N1 and 24 as N2. The subtyping of influenza A viruses revealed that 41.0% (16/39), 17.9% (7/39) and 2.6% (1/39) of infections were attributed to H3N2, H1N2 and to H1N1 viruses, respectively [see Additional file 1].   Table 2.
Regarding the catalytic and framework sites within the NA sequence, the analysis showed that they were conserved in all H1N1 and H3N2 isolates of this study (composition of the catalytic site is R118, D151, R152, R224, E276, R292, R371, and Y406 and composition of the framework sites is E119, R156, W178, S179, D/N198, I222, E227, H274, E277, N294 and E425). Nevertheless, amino acid variations were observed in other positions (data not shown). None of the isolates carried a known oseltamivir resistance mutation in the NA sequence (positions 151, 152, 222, 224, 274, 276, 292 and 294).
In this study, we detected infections caused by influenza viruses from the end of summer to the end of winter, and this observation was consistent with the findings reported by Carraro et al. [28] during 2001-2003 influenza seasons in São Paulo state. Besides, the number of influenza cases observed in the 2001-2003 period in Uberlândia was higher than in other flu seasons, which may be the cause of the higher rate of influenza-related morbidity reported in other parts of the world [29].
H1N1 and H3N2 are usually the most common influenza A subtypes found infecting humans [30], and H3N2 viruses have been the dominant strain in most years since they first emerged in 1968. Still, this subtype had caused one of the most serious respiratory infections [31][32][33] until 2009, when the novel swine-origin influenza H1N1 virus emerged and was responsible for the new pandemic [34]. In this study, H3N2 also predominated, with 41.0% of isolates that were subtyped, and was responsible for the majority of hospitalizations (in five out of six hospitalizations).
The co-circulation of influenza virus subtypes is found commonly and in the 2002 season in Uberlândia, H3N2 co-circulated with subtype H1N2 and influenza B virus, which was detected in only one clinical specimen. Chieochansin et al. [35] also reported only one case involving influenza B virus in 302 samples collected from  [36][37][38] and were also found in our study. In Brazil, in the north region, H1N2 cases were detected between January and April 2003 (epidemiological weeks 2 to 14) [Personal Communications, Wyller Mello and Mirleide Santos], whereas in Uberlândia, which is located in the southeastern Brazil, these strains circulated as early as July 2002.
Sequence analyses of subtypes H1N1, H1N2 and H3N2 and influenza B virus found in this study showed high identities with variants found in other continents in previous years, such as in Asia, Africa and North America, suggesting a global origin of new strains, which are then spread to other countries through human mobility [39,40].
Every year, WHO recommends specific vaccine strains for influenza vaccine production. This strategy needs to be followed since hemagglutinin and neuraminidase are under selective pressure and undergo frequent antigenic changes in order to evade the host's immune system [4]. Comparison of the deduced amino acid sequences of Uberlandia's isolates and those of the vaccine strains showed that they were related closely in the majority.  [41], who reported that after 2002, this was one of the regions in which the highest number of substitutions in H3 subtype virus was detected. Also, divergences from nine to 11 amino acid residues in three or four antigenic sites, by comparing the HA sequence of the vaccine strain and those of the viruses detected between 2001-2003, may have caused a certain decrease in protection efficiency of the vaccine that was available in the same time period in Brazilpossibly configuring a mechanism of immune evasion by these variants [42].
For H1 subtype, important amino acid changes in the HA sequence are clustered in the variable antigenic sites, which have been proposed based on sequence homology to the H3 sites B + D, B, C + D + A and E, respectively [43]. In the present study, we detected small numbers of variations in the Ca and Sb sites in the HA of viruses detected in 2002 and 2006 in comparison to the vaccine strain. In H1 isolates, we found Gln and Gly at positions 226 and 228, respectively. These determinants, however, have been reported in H1 viruses from both human and avian hosts [44]. In addition, H1 D190 and D225 determinants of receptor-binding specificity of H1N1 from humans [45] were also observed.
The antigenic regions A through E of the HA1 domain are potential targets for neutralizing antibodies, and amino acid changes in these regions have been associated with annual epidemics in humans. Still, it has been proposed that epidemiologically important drift variants generally exhibit four or more amino acid substitutions located in two or more antigenic sites in this domain [46]. These substitutions may generate a new strain against which immunity acquired previously, by either infection or vaccination, is no longer effective [47]. On the other hand, the receptor binding sites are likely to resist changes [48]. Our results showed that all of the 2001-2003 H3 sequences presented at least eight substitutions located at no less than two antigenic sites. In regard to HA protein binding to sialic acid, Leu and Ser at positions 226 and 228, respectively, have been reported to be important in H3 subtypes and are determinants of influenza viruses that infect humans [44]. In this study, at position 226, we found Val in the H3N2 viruses found in the 2001-2004 period and Ile in the 2004-2007 period ( Figure 2). Nevertheless, Val and Ile, like Leu, are neutral, nonpolar amino acids and thus the presence of any of them may have preserved the properties of the sialic acid binding pocket of HA, as suggested by others [49]. Still, mutations at this position may represent a selective advantage to the virus [50].
Increasing number of N-linked glycosylation sites have been shown to attenuate H3N2 influenza viruses in a mouse model [51]. On the other hand, addition or removal of a glycosylation site may generate virus variants that evade the host immune system, including the response generated by vaccination, and maintain a sustained circulation of the virus within the human population [52]. We found one isolate (323, Figure 2) in a sample collected in 2003 that contained an additional glycosylation site located in an antigenic site, when compared to the vaccine strain.
Concerning the NA, variations that promote modifications in the antigenic properties usually occur in the catalytic or framework sites of the protein [53]. The amino acid substitutions found in the NA sequences of the Uberlândia isolates, in comparison with the sequences of the vaccine strains of each corresponding year, were not located in these sites and, therefore, may not have led to significant antigenic changes.
Circulation of naturally resistant H1N1 viruses to oseltamivir has been reported elsewhere, even in the absence of selective drug pressure [54]. The extensive use of this antiviral, especially after the emergence of the pandemic H1N1 virus, in 2009, may select and increase the frequency of resistant variants in flu cases. In this study, oseltamivir-resistant viruses were not detected. It should be noted, however, that the