Epidemiology and genotypic diversity of human metapneumovirus in paediatric patients with acute respiratory infection in Beijing, China

Background Acute respiratory tract infections (ARTIs) causes high amounts of morbidity and mortality worldwide every year. Human metapneumovirus (HMPV) is a major pathogen of ARTIs in children. In this study, we aimed to investigate the epidemiology and genotypic diversity of HMPV in children hospitalized with ARTIs in Beijing, China. Methods Hospitalized children aged < 14 years with ARTIs were enrolled from April 2017 to March 2018; nasopharyngeal aspirates were collected and subjected to real-time polymerase chain reaction tests for HMPV. HMPV-positive samples were genotyped based on a partial N gene. Whole genome sequences were determined for samples with high viral loads. Results 4.08% (52/1276) enrolled paediatric patients were identified as having HMPV infection. The epidemic season is winter and early spring, children aged ≤ 4 years were more susceptible to HMPV infection (47/52, 90.38%). The co-infection rate were 36.54% (19/52), the most common co-infected virus were influenza and respiratory syncytial virus. The main diagnoses of HMPV infection were pneumonia (29/52, 55.77%) and bronchitis (23/52, 44.23%), while the main clinical manifestations were cough, fever, rhinorrhoea, and sneeze. Among 48 HMPV-positive specimens, A2b (19/48, 39.58%) and B1 (26/48, 54.17%) were the main epidemic subtypes. Patients with HMPV genotype A infection had a higher viral load compared to genotype B patients (6.07 vs. 5.37 log10 RNA copies/ml). Five complete sequences of HMPV were obtained. This is the first report of a whole genome sequence of HMPV-B1 isolated in China. Conclusions HMPV is an important respiratory pathogen in paediatric patients. Cases of HMPV infection could burden hospitals in the epidemic season. HMPV viral loads and genotypes have no correlation with co-infection or clinical characteristics.

In this study, we describe the prevalence rate, co-infection status, and clinical characteristics of HMPV in paediatric patients with ARTIs. Additionally, we explore the association between viral loads and symptoms of HMPVinfected patients and analyse the distribution of HMPV genotypes in Beijing, China. We also obtained five whole genome sequences of HMPV.

Study population and specimen collection
From April 2017 to March 2018, nasopharyngeal aspirates (NPAs) were collected from children (aged < 14 years) with ARTIs who were hospitalized in the Beijing Friendship hospital. ARTIs were defined by the presence of at least two of the following symptoms: fever, cough, sneeze, wheezing, nasal obstruction, sore throat, and dyspnoea. Children who had been enrolled but discharged were considered as a new case if they were readmitted for a new episode of ARTIs. Demographic data and clinical characteristics were recorded on a form from each patient's record of medical history and examination. All NPAs were collected in tubes with viral transport medium and kept at − 80 °C until use.

Detection of HMPV
The viral nucleic acids were extracted from 200 µl of each specimen with a QIAamp MinElute Kit (Qiagen, Germany) in accordance with the manufacture's protocols. HMPV was identified by using a One-step RT-PCR Kit (Ambion, USA) in accordance with the manufacture's protocols. HMPV forward primer (5′-CAT ATA AGC ATG  CTA TAT TAA AAG AGT CTC -3′), reverse primer (5′-CCT ATT TCT GCA GCA TAT TTG TAA TCAG-3′),  and  probe (FAM-TGY AAT GAT GAG GGT GTC ACT GCG GTTG-TAMRA) were used as previously described [10] to amplify a 163-bp fragment from the N gene. PCRpositive products were confirmed by sequencing. Viral loads were detected by real-time reverse transcription PCR (qRT-PCR), and the standard curve was generated as described previously [11].

HMPV genotyping
HMPV cDNA was synthesized using SuperScript III First-Strand Synthesis System for Reverse Transcription-PCR (RT-PCR) (Invitrogen, USA) in accordance with the manufacture's protocols. The partial sequence of N gene (1200 bp) was amplified using a nested PCR under the following thermocycling conditions: an initial denaturing at 94 °C for 5 min, followed by 35 cycles of 94 °C for 40 s, 46 °C for 40 s, and 72 °C for 1 min, and a final extension at 72 °C for 10 min. The following newly designed primers were used: outer forward primer, 5′-TTA ART TAC AAA AAA ACA TGG GAC -3′; outer reverse primer, 5′-AAA GAA TAT CTT TTC CTT CAGGG-3′; internal forward primer, 5′-ATG GGA CAA GTG AAA ATG TCTC-3′; and internal reverse primer, 5′-AAT TAC TCA TAA TCA TTT TGA CTG -3′. Specimens that failed to be amplified were considered as untyped. The PCR products were sequenced by Sanger sequencing (TSINGKE, Beijing, China). A neighbour-joining (NJ) tree was constructed by the Tamura-Nei model in MEGA 7.0 using 1000 bootstrap replicates. The web server of FindModel (https ://www.hiv.lanl.gov/conte nt/seque nce/findm odel/ findm odel.html) was used to find best available nucleotide substitution model. Reference strains of HMPV were acquired from GenBank, including AF371337, AY297749, DQ843659, AY525843, and FJ168778. Avian metapneumovirus C (AMPV C, AY590688) was used as an outgroup to root the tree.

Sequencing
HMPV-positive samples with high copy numbers (> 10 5 copies/ml) were further used for whole genome sequencing. Fourteen pairs of primers with overlap were designed according to the reference strain CAN97-83 (GenBank accession number: AY297749) ( Table 1). Ex-Taq (TaKaRa, China) was used to perform PCR. The PCR products were sequenced by Sanger sequencing, and the resulting sequences were assembled using Sequencher 5.0. Comparisons were made using published HMPV sequences selected from GenBank. A NJ-tree was constructed using the Tajima-Nei model (MEGA 7.0) with 1000 bootstrap replicates. Identity within the analysed sequences was analysed by BioEdit. Recombination events were detected by Simplot.

Statistical analysis
Data analysis was performed using SPSS 13.0, and categorical data was compared by χ 2 and Fisher's exact tests. Two-tailed p values of < 0.05 were considered to indicate statistical significance.

Epidemiology of HMPV
A total of 1276 hospitalized children with ARTIs were enrolled between April 2017 and March 2018; 4.1% (52/1276) of their samples were positive for HMPV. As shown in Table 2, the male/female ratio was 1.26 (29:23) (p = 0.251), indicating the sex difference was not statistically significant. The median age of HMPV-infected children was 36 months (IQR: 1-168 months), and the detection rates were significantly different among different age groups (p = 0.019). The HMPV prevalence of patients aged ≤ 4 years (47/898, 5.2%) was significantly higher than that of patients aged 4-14 years (5/378, 1.3%) (p = 0.001). The seasonal distribution of HMPV infection from April 2017 to March 2018 is shown in Fig. 1. 59.6% (31/52) of positives were found in spring, none (0/52) in summer, 3.8% (2/52) in autumn and 36.5% (19/52) in Table 1 Primer sets for amplification of the whole genome of HMPV a The reference of location is based on AY297749 strain

Co-infection with other respiratory viruses
Among the 52 HMPV-infected children, 36.5% (19/52) were co-infected with other respiratory viruses (Table 3). Most co-infections involved RSV or IFV-A, (both account for 11.5%, 6/19). The linear range of the standard curve of the N-gene is 10 3 -10 12 copies/ml. The viral loads of HMPV-positive patients ranged from 7 × 10 3 to 9.61 × 10 7 copies/ml. At the single time point when the viral load was measured, there was no statistical difference in the viral loads between HMPV mono-infections and co-infections (p = 0.398) (Fig. 2), and the viral loads had no relationship with the tested demographic or clinical features (i.e. gender, age, temperature, and hospitalization; data not shown).

Clinical characteristics of HMPV infections
The clinical characteristics, diagnoses, and hospital stay lengths of HMPV-positive patients are listed in There was no significant difference in these symptoms between the HMPV-positive patients with co-infections and those without co-infections. Additionally, hospital stay, fever and cough have no significant difference between patients aged < 2 years and 2-14 years (P = 0.05, 0.187, 1.00, respectively).

Phylogenetic analysis of HMPV
A portion of the N gene (813 bp) was amplified in 48 HMPV-positive specimens (the sequences are available in Additional file 1). The phylogenetic analysis conducted on these sequences revealed that 39.6% (19/48)  of strains were grouped into the A2b lineage, 54.2% (26/48) of strains were grouped into the B1 lineage, and 6.3% (3/48) of strains were grouped into the B2 lineage; subtypes A1 and A2a were not detected (Fig. 3a). The positives rates of genotype A2b and B1 were higher in the spring (23.1%, 12/52 and 30.8%, 16/52, respectively). Additionally, HMPV genotype A patients had a significantly higher viral load (6.07 ± 1.21 log 10 RNA copies/ml, n = 22) than genotype B patients (5.37 ± 0.92 log 10 RNA copies/ml, n = 26) (p = 0.029). There was no statistical difference in the main clinical features (i.e. gender, age, temperature, hospitalization, and virus loads) between patients infected with HMPV genotype A and those infected with HMPV genotype B (data not shown).

Whole genome analysis of HMPV
To characterize the whole genome sequences of HMPV epidemic strains, genomes were obtained for five of the HMPV strains by using Sanger sequencing; the GenBank accession numbers for these sequences are MK820375 and MN745084 to MN745087. Thirty HMPV sequences were downloaded from GenBank as reference strains.
The Simplot analyses showed no evidence of a recombination event (data not shown). A phylogenetic tree based on the whole-genome sequences of HMPV was constructed using the NJ method. The resulting phylogenetic tree shows that four strains from this study belong to the A2b lineage, and are closest to the gz-01 (China) strain in genetic distance. Additionally, the first complete genome (bj0123) of an HMPV-B1 genotype in China was obtained. The results was consistent with the phylogenetic tree constructed based on the partial N gene sequences (Fig. 3b). Fusion protein (F) is an envelope protein of HMPV that has two functional sites. One is a cleavage site, Arg-Gln-Ser-Arg↓ (RQSR↓), which is important for fusion progress between HMPV and the host cell, and the other one is an Arg-Gly-Asp motif (RGD), which mediates the binding of virus and cellular α v β 1 integrins [12,13]. Based on an alignment of the obtained HMPV sequences with reference sequences, there were no nucleotide mutations detected in these functional sites (data not shown). The sequence identity were 80.0-100.0% among all the analysed sequences, G gene and SH gene have high genetic variability (48.3-100.0% and 66.4-100.0%, respectively). The identity among five obtained HMPV whole genomes were 80.1%-99.6%, while the identity of the four A2b genomes were 99.1-99.6%, and the identity between bj0123 and the four A2b genome is about 80.1%.

Discussion
HMPV has been identified a leading cause of ARTIs since it was discovered in 2001. According to a serological study, HMPV infection has existed for at least 70 years [1]. Almost all children eventually become infected by HMPV, and adults can be re-infected by HMPV throughout their lives due to incomplete immunity. The prevalence of HMPV in children is approximately 2.0-18.2% [4,[14][15][16] [4,17], indicating that the prevalence of HMPV could  be different in across regions and years. Therefore, it is necessary to establish continuous epidemiological surveillance in a wider area. The detection rates of HMPV infection between male and female patients in our study were not significantly different (p = 0.251), but there was a significance difference in the detection rates among age groups, with patients aged ≤ 4 years (4.8%) having a higher rate of detection compared with the patients aged 5-14 years (1.7%) (p = 0.001). Notably, 90.4% of HMPV infections detected in this study occurred in patients aged ≤ 4 years, which is consistent with findings from prior studies [4,15]. The peak age range for HMPV infections was 1-2 years (6.6%). HMPV infections have a clear tendency toward seasonal distribution, supported by the result of the present study (p = 0.000), but they vary based on different climate and geography factors. Here, HMPV was detected most frequently in the winter and spring. The HMPV detection rate was highest in March 2018, which is similar to findings from studies in Lanzhou (China) and Korea [18,19].
Co-infection with respiratory viruses and bacteria is commonly seen in cases of ARTIs. Coinfections of HMPV with other respiratory viruses have been reported many times, and patients with such co-infections are more likely to have a fever, lead to severe pneumonia, and cause higher hospitalization rates [4,18,20]. The proportion of co-infection in this study was 36.5%. IFV and RSV were the most frequently co-infected viruses, which may be related to the overlap between the epidemic season of HMPV with those of IFV and RSV [21]. There was no significance difference observed between the co-infection group and mono-infection group in any of the assessed clinical manifestations, diagnoses, or hospital stay length. Peng et al. indicated that a high HMPV copy number is correlated with disease severity and hospital stay duration [22]; however, we did not find any relationship between the HMPV viral load and clinical features. There are also studies reporting that the viral load of asymptomatic children infected with HMPV is significantly lower than that of symptomatic children [23], which needs to be verified in future studies.
HMPV infection can cause both URTI and LRTI [3]. The risk factors associated with severe HMPV infection include preterm birth, early age, low immune function, nosocomial infection, and chronic pulmonary, cardiopathy, or neurological disease [24]. In the present study, all HMPV-infected patients had an URTI, 55.8% had pneumonia, and 44.2% had bronchitis. Additionally, 94.2% of HMPV-positive children had an abnormal chest radiograph, and the main clinical characteristics of HMPV infection were cough, fever, and rhinorrhoea. No severe cases were found in our cohort, and 78.9% of our patients were discharged within 7 days.
HMPV can be divided into genotypes A and B. Vicente et al. reported that type A HMPV is more virulent than type B, whereas Papenburg et al. concluded that type B HMPV was associated with severe infection [25,26], and some studies found no evidence for differential severity among different HMPV genotypes [18]. In the present study, we found patients with HMPV genotype A infection had a higher viral load compared to genotype B patients (p = 0.029), which is consistent with Oong's report [27]. However, there was no significant difference between HMPV genotypes A and B in terms of their epidemiological characteristics, hospital stay, or viral loads. We analysed the distribution of HMPV subtypes in Beijing based on a partial sequence of the N gene. Subtypes A2b, B1, and B2 were detected, but subtypes A1 and A2a were not found. The most prevalent subtype was B1 (54.5%), followed by A2b (40.9%), which is consistent with previous work conducted in Beijing, Changsha, Kuala Lumpur, and Jordanian [7,14,28,29]. Finally, to study the genome variation of HMPV epidemic strains in Beijing, we obtained five complete HMPV genome sequences with 80.1-99.6% identity, and four of them have an identity of 99.1-99.6%. Phylogenetic tree results show that four strains belong to sublineage A2b and have the closest genetic distance to the gz-01 (China) strain, one strain belongs to sublineage B1, which is the first report of whole genome sequence of HMPV-B1 isolated in China. No recombination events or mutations in important function sites were found.
The study is limited by the lack of a control group without ARTIs and its short research duration. Followup studies with long-term continuous monitoring that include a control group are needed to analyse the clinical features more accurately.

Conclusion
HMPV is an important virus in paediatric patients. The findings provide characteristics about the epidemiological and genotypic diversity of HMPV infections in Beijing, which will help to provide a theoretical basis for the prevention and control of diseases and to enrich the HMPV genome database of China.