Background

Human adenovirus (HAdV) belongs to the Mastadenovirus genus of the Adenoviridae family, and is a linear double-stranded deoxyribonucleic acid (DNA) virus without envelope. The diameter is about 70–100 nm, and the total genome length is about 36kb [1]. The capsid of the HAdV is icosahedral symmetric and consists of the main capsid proteins (Hexon, Penton base, and Fiber) as well as secondary capsid proteins [2]. The Hexon protein contains two hypervariable domains, Loop1 (including the hypervariable region 1–6) and Loop2 (including the hypervariable region 7), which are the main target regions for inducing neutralizing antibodies and form ε antigen epitopes. The ε antigen epitopes are group-specific and type-specific, and are the basis for virus serum neutralization experiments and type identification [3]. The Penton base protein has two key sites, including the Arg-Gly-Asp (RGD) Loop and the hypervariable region (HVR) 1 region. The former is located in the hypervariable loop 2 region. The interplay between this region and the integrin expressed on the host cell surface is conducive to the virus internalization and thus promoting the virus infection [2]. The latter lies on the surface of the virus capsid, and recombination can occur at this site and its vicinity [4]. The Fiber protein contains N-terminal tail (Tail), stalk (Shaft), and C-terminal spherical domain nodules (Knob). Knob region can mediate the interaction between the virus and host cells [3]. So far, HAdV is separated into 7 species from A to G, including 116 types (http://hadvwg.gmu.edu/). HAdV is one of the important pathogens causing acute respiratory infection (ARI) in children, and the detection rate is about 5-10% [5], which can reach 20.1% in severe ARI children [6]. The treatment of HAdV infection is mainly symptomatic and supportive therapies, and there is no approved specific antiviral drug. In severe cases, HAdV infection can be fatal. Especially in infants and children with immunodeficiency or immunosuppression, HAdV infection can lead to serious clinical manifestations and long-term sequelae such as airway occlusion and bronchiectasis, posing a serious threat to children’s lives and health [7, 8]. The tissue tropism of different types of HAdV is different, which can lead to different illnesses. The main types related to respiratory infections are B species (HAdV-3, 7, 11, 14, 16, 21, 50, and 55), C species (HAdV-1, 2, 5, and 6), and E species (HAdV-4) [1].

HAdV-4 is one of the common types causing ARI, which is the only member of the E species. The prototype strain (RI-67 strain) was first isolated from an ARI case during the influenza pandemic in Fort Leonard Wood (Missouri, USA) in 1953 [9]. HAdV-4 is mostly sporadic, but it can also cause local small-scale epidemics or large outbreaks, with death cases reported [1]. In 1980, Faden et al. reported a case of a 5-and-a-half-year-old child with a previous history of epilepsy dying from respiratory distress caused by HAdV-4 infection in USA [10]. Ou et al. detected 16 positive cases of HAdV in the lung tissues of 175 children who died of pneumonia at Guangzhou Children’s Hospital from July 1988 to January 2005 by using nested polymerase chain reaction (PCR) and immunohistochemical analysis methods, among which 12 cases were of HAdV-4, accounting for 75%, indicating that it might be an important serotype for fatal pneumonia in children in Guangzhou [11]. In September 2013, an outbreak of ARI caused by HAdV-4 occurred in a kindergarten in Longgang District, Shenzhen, China. There were 17 cases of infection, including 4 hospitalized cases, all of which were 5-year-old children [12]. In recent years, the incidence rate related to HAdV-4 has gradually increased in many countries, including China, so it is considered to be one of the reasons for the major epidemic [13]. Coleman et al. performed genome sequencing on HAdV positive clinical samples collected from 396 pediatric patients in Singapore and found that compared with 2014–2015, the risk of HAdV-4 infection increased among pediatric patients admitted to Singapore General Hospital and KK Women’s and Children’s Hospital from 2016 to 2018 (OR, 13.6 [3.9–46.7]) [14]. Chen et al. found that among children hospitalized for ARI in Southern Taiwan, type 3 and 7 were the main types between November 1999 and March 2000, and from then on to 2001, type 4 became the main type, accounting for approximately 52% of the total number of HAdVs detected [15]. Ye et al. showed that the serum antibody positivity rate of HAdV-4 in healthy individuals in Southern China was 58.4%, and increased with age [16]. Recently, some studies have reported that HAdV-4 may have a new recombination phenomenon. Zhang et al. found that the HAdV-4 strain obtained in Hong Kong in 2014 might have undergone recombination of inverted terminal repeat sequences (ITRs), acquiring human transcription factor binding sites nuclear factor 1 (NF-1) from the internal ITRs of B species (3, 7, 14, 16, 21, 55), which was more conducive to replication and growth of HAdV-4 in human cells [17, 18]. Based on restriction enzyme analysis (REA), the HAdV-4 genotypes can be divided into two subtypes: HAdV-4p and HAdV-4a [19]. The Fiber protein of the HAdV-4 has the regions binding to Coxsackie-adenovirus receptor (CAR) that mediates the adhesion between HAdV-4 and target cells, which is a rate-limiting factor of HAdV-4 transfection into target cells [20].

As of now, there are few sequences of HAdV-4 strains from China. Only thirteen whole genome sequences of HAdV-4 published in GenBank are from China. The epidemic situation of HAdV-4 in China is little known [1, 21]. Although vaccines for HAdV-4 have been developed, they are only used in American military recruits. Currently, in China, there are no HAdV-4-related vaccines or drugs with safety and effectiveness applied [22]. In this study, the purpose was to clarify the prevalence and genetic characteristics of HAdV-4 in China, and provide data reference for the monitoring, prevention and control of HAdV-4, as well as the research and development of vaccines and drugs. ARI children hospitalized in six hospitals across six provinces in Northern and Southern China during 2017 and 2020 were selected as the research subjects. The clinical data of HAdV-4 positive patients were collected for clinical feature analysis, and bioinformatics analyses on the whole genome sequences, Hexon, Penton base, Fiber, and ITR gene sequences were conducted.

Methods

Patients and clinical data collection

A multicenter study was performed with children younger than 18 years hospitalized for ARI from 2017 to 2020. Patients were enrolled from six areas (Beijing, Guangzhou, Wenzhou, Yinchuan, Guiyang, and Shenyang) in Northern and Southern China. Children with ARI after 48 h of admission or recent history of HAdV infection were excluded. The clinical electronic medical record system was used to collect the clinical data.

Specimen collection and HAdV detection and typing

Respiratory samples including sputum (0.5 ml) and nasal/pharyngeal swabs were collected within 24 h after hospitalization. The sputum was stored in 3 ml virus transport solution, and the nasal/pharyngeal swabs were immersed in virus sampling tubes containing 3 ml virus transport solution (Yocon Hengye Biotechnology, Beijing, China). The samples were transported to laboratories of various hospitals for processing at low temperature (4 °C) and were divided into three parts and stored in the − 80 °C refrigerator. Ultimately, these samples were transported to virology laboratory of the Beijing Children’s Hospital affiliated to Capital Medical University through dry ice for centralized detection and analysis. The QIAamp MinElute Virus Spin Kit (QIAGEN, Hilden, Germany) was used to extract viral nucleic acids following the manufacturer’s instructions. The Multiple Respiratory Virus Panel (RVP) (Luminex Corporation, USA) was used to detect common respiratory viruses, including human enterovirus (HEV), human rhinovirus (HRV), respiratory syncytial virus (RSV), parainfluenza virus, HAdV, coronavirus, influenza virus, human bocavirus, and human metapneumovirus (HMPV), etc. For specimens tested HAdV positive, the amplification of the HVR 1–6 of the Hexon gene was completed by nested PCR (HotStar Taq Plus Master Mix Kit, QIAGEN, Germany) following the instructions [23]. The primers and amplified fragments are shown in Supplementary Table 1. The purification and sequencing of PCR products were performed by Sino Geno Max (Beijing, China). BLAST was used to align the sequencing results with the sequences in GenBank to determine the typing.

Virus isolation and sequencing of whole genome

To isolate the virus, the HEp-2 cells were used, onto which the HAdV-4 positive samples inoculated. The obtained nucleic acids of the strains were sent to BioGerm Medical Technology (Shanghai, China) to obtain the whole genome sequences using second-generation sequencing technology, which were annotated and uploaded to the GenBank database.

Amplifying and sequencing of major capsid proteins and ITR gene

HotStar Taq Plus Master Mix Kits (QIAGEN, Hilden, Germany) was used to perform PCR amplification to obtain full-length sequences of Hexon, Penton base, and Fiber genes following the instructions. The primers are shown in Supplementary Table 2. Based on Sanger sequencing, the sequencing of PCR products was performed by Sino Geno Max (Beijing, China). DNASTAR v7.1 (DNASTAR Inc., Madison, WI, United States) was used to assemble and edit the sequences. The obtained Hexon, Penton base, and Fiber genes were annotated and uploaded to the GenBank database. Using genomic DNA as a template, the ITR genes were sequenced with either a 5’-primer “HAdV-4-ITR-F” (5’-AACTCTTCTCGCTGGCACTCAA-3’) or a 3’-primer “HAdV-4-ITR-R” (5’-CCGCCCCTAACAGTCGCC-3’) [17].

Phylogenetic analysis

The GenBank database was used to select the reference sequences of the whole genome, Hexon, Penton base, and Fiber genes of HAdV-4 to perform phylogenetic analysis. The whole-genomic sequences that were not typed based on three major capsid protein genes and lacked isolation time and location and the Hexon, Penton base, and Fiber gene sequences with high consistency in region, time and sequence were excluded. The ITR sequences were extracted from the whole-genomic reference sequences of HAdV-4 obtained above, and the representative ITR reference sequences of the other types of HAdV were selected from GenBank. Multiple alignment of the whole-genomic sequences and Hexon, Penton base, Fiber, ITR genes was carried on by MAFFT (https://www.ebi.ac.uk/Tools/msa/mafft/) software. By MEGA v7.0.26 (Sudhir Kumar, Arizona State University, Tempe, AZ, United States) software, the phylogenetic trees were constructed using the neighbor-joining method and the Kimura two-parameter model. To evaluate the reliableness of the phylogenetic trees, the bootstrap method with 1,000 replicates was used.

Analysis of genetic variation

BioEdit v7.2.05 software (http://www.mbio.ncsu.edu/BioEdit/page2.html) was used to determine the genetic variations of the Hexon, Penton base, and Fiber genes of HAdV-4 obtained in this study.

Recombination analysis

Based on the whole-genomic sequences of the HAdV-4, recombination analysis was performed by Simplot v3.5.1 software. The window size, step size, distance model and tree model which were defaulted to 1,000, 200, “Kimura”, and “Neighbor-Joining”, respectively, were employed for the analysis.

Ethics

The Institutional Ethics and Review Committees of the Beijing Children’s Hospital affiliated to Capital Medical University approved this study (2017-k-15, 2019-k-357), which was conducted in accordance with the Declaration of Helsinki (revised in 2013). Parents/legal-guardian of each child signed an informed consent.

Results

Enrolled patients

After further excluding children with missing or duplicate clinical data or unqualified respiratory samples collected, 2847 ARI children were finally enrolled, of which 1407 (49.42%) patients were from Northern China and 1440 (50.58%) were from Southern China. The detailed geographical distribution of ARI children is shown in Table 1. The distribution of the gender and age is shown in Table 2. There were 1723 males (60.52%) and 1124 females (39.48%), with a proportion of 1.53: 1.

Table 1 Distribution of 2847 ARI patients in different areas of mainland China and the detection rate of HAdV
Full size table
Table 2 Demographic characteristics of 2847 ARI children
Full size table

Detection rate of HAdV and HAdV-4

The detection rate of HAdV is shown in Table 1. There were 156 (5.48%) HAdV positive samples detected, of which 67 were co-infected with other common respiratory viruses. forty-two cases were co-infected with one virus, eighteen cases were co-infected with two viruses, and 7 cases were co-infected with three viruses. The most common co-infection viruses were RSV (18 cases), HMPV (13 cases), and HEV and/or HRV (19 cases). Ninety-eight HAdV positive samples were successfully genotyped, and three species (B, C, and E) were detected, including seven genotypes (HAdV-1 ~ 7). The detection rates of seven genotypes of HAdV are shown in Table 1. Eleven HAdV-4 positive samples were detected, accounting for 0.39% of the total number of samples and 7.05% of the HAdV positive samples, respectively. The specific information of 11 HAdV-4 positive samples is shown in Table 1 and Supplementary Table 3. There were 2 cases (18.18%, 2/11) in 2017, 3 cases (27.27%, 3/11) in 2018, 5 cases (45.45%, 5/11) in 2019, and 1 case (9.10%, 1/11) in 2020.

Clinical characteristics of patients with HAdV-4 infection

The clinical characteristics of 11 patients with HAdV-4 infection are shown in Supplementary Table 4. Children aged < 6 months, 6 months-1 year, 1–3 years, 3–6 years, and 6–12 years made up 18.2% (2/11), 9% (1/11), 27.3% (3/11), 18.2% (2/11), and 27.3% (3/11), respectively. Six children were female and 5 children were male. All the 11 children had cough and 10 children (90.91%) developed fever, with a peak fluctuation of 38.2–40℃. Two children (18.18%) had conjunctivitis. Three children (27.27%) had shortness of breath and wheezing, among which 2 children had dyspnea and required nasal continuous positive airway pressure (NCPAP) respiratory support, as well as serious complications such as respiratory failure and mild elevation of liver enzymes. These two children were diagnosed with severe pneumonia, one of which developed hemophagocytic syndrome and checked in pediatric intensive care unit (PICU). The patient had a potential disease of ventricular septal defect and the longest hospital stay. Four patients (36.36%) experienced increased white blood cell. Six patients (54.55%) had increased procalcitonin and erythrocyte sedimentation rate, and 7 patients (63.64%) had increased C reactive protein. All patients had different degrees of chest X-ray abnormalities. Eight children (72.73%) were diagnosed with pneumonia, 1 child (9.09%) with bronchitis, and 2 children (18.18%) with upper respiratory tract infection (URTI). There was 1 child (9.09%) with mixed infection of HRV. Two children diagnosed with severe pneumonia did not have any other pathogenic infections. The total hospitalization time varied from 4 to 24 days. All the patients improved and were discharged without any deaths.

Phylogenetic analysis of the whole genome of HAdV-4

Two HAdV-4 strains were isolated from patients from Guangzhou, and the whole-genome sequences were obtained (Supplementary Table 3), of which the identities of the nucleotide and amino acid sequences were 99.9% and 100%, respectively. The genetic distance was 0.000. One hundred and thirty-four HAdV-4 whole-genomic reference sequences were obtained (Supplementary Table 5) from Genbank.

There were two phylogenetic branches of HAdV-4 according to the whole-genome sequence phylogenetic analysis: 4a and 4p. The two whole-genome sequences obtained in this study were located in the same branch (4a) as the thirteen HAdV-4 whole-genome reference sequences currently available in China, as well as most prevalent strains worldwide since the 1980s. The prototype strain (Genbank accession number: AY594253) and vaccine strains were in the same branch (4p) as most of the prevalent strains before the 1980s and several strains discovered in USA after that (Fig. 1). All the whole-genome sequences of HAdV-4 in China were clustered together, with the Bootstrap value of 97%, nucleotide and amino acid sequence identities of 96.6–100% and 96.5–100%, respectively, and the genetic distance of 0.000. The nucleotide and amino acid sequence identities on the whole-genome sequences of the HAdV-4 isolates obtained in this study and recent epidemic strains in China were 98.2–99.9% and 98-99.8%, respectively, with the genetic distance of 0.000. The mean genetic distances within group of 4a and 4p were 0.001, respectively, and the mean genetic distance between groups of 4a and 4p was 0.038. These findings suggested that there was high homology between the strains gained in this study and those previously obtained in China, including recent epidemic strains.

Fig. 1
figure 1

Phylogenetic analysis of the whole genome of HAdV-4. The phylogenetic tree was constructed by the neighbor-joining method and the Kimura two-parameter model using 1,000 replicates. The whole-genome sequences of HAdV-4 previously gained in China are indicated by cyan dots. Pink dots indicate the prototype strain and vaccine strains. Strains gained in this study are indicated by red dots

Full size image

Phylogenetic analysis of the Hexon, Penton base, and Fiber genes

Hexon, Penton base, and Fiber gene sequences were successfully obtained from all 11 HAdV-4 positive samples (Supplementary Table 3), among which the nucleotide and amino acid homologies were 99.9–100% for Hexon, 99.9–100% for Penton base, 99.7–100% for Fiber. The genetic distances were 0.000 for Hexon, 0.000 for Penton base, and 0.001 for Fiber. Based on the Genbank, there were 82, 79, and 82 reference sequences of Hexon, Penton base, and Fiber genes obtained, respectively (Supplementary Table 5).

According to the phylogenetic analysis of Hexon, Penton base, and Fiber, there were also two phylogenetic branches of HAdV-4: 4a and 4p (Fig. 2). Similar to the whole-genome phylogenetic analysis, the 11 isolated strains gained in this study, strains previously gained in China and most pandemic strains since the 1980s all over the word were located in the HAdV-4a branch. The prototype strain and vaccine strains were in the same branch (4p) as most of the prevalent strains before the 1980s and several strains discovered in USA after that. The phylogenetic tree based on the Hexon gene sequences (Fig. 2A) displayed that the nucleotide and amino acid sequence identities on the Hexon genes of the HAdV-4 isolates obtained in this study and recent epidemic strains in China were 99.8–100%, respectively, with the genetic distance of 0.000. The mean genetic distances within group of 4a and 4p were 0.000, respectively, and the mean genetic distance between groups of 4a and 4p was 0.029. The strains from different countries did not exhibit significant geographical clustering. The phylogenetic tree based on the Penton base gene sequences (Fig. 2B) displayed that the nucleotide and amino acid sequence identities on the Penton base genes of the HAdV-4 isolates obtained in this study and recent epidemic strains in China were 99.6–100% and 99.8–100%, respectively, with the genetic distance of 0.001. The mean genetic distances within group of 4a and 4p were 0.002 and 0.000, respectively, and the mean genetic distance between groups of 4a and 4p was 0.030. The 11 Penton base gene sequences were clustered with the strain from Beijing (BJ14) in China and strains discovered in USA from 2009 to 2020, with the Bootstrap value of 87%. The phylogenetic tree based on the Fiber gene sequences (Fig. 2C) displayed that the nucleotide and amino acid sequence identities on the Fiber genes of the HAdV-4 isolates obtained in this study and recent epidemic strains in China were 99.4–100%, respectively, with the genetic distance of 0.001. The mean genetic distances within group of 4a and 4p were 0.001 and 0.000, respectively, and the mean genetic distance between groups of 4a and 4p was 0.020. The 11 Fiber gene sequences, except for BCH20190735, all clustered together with sequences previously obtained in China, with the Bootstrap value of approximately 62%. BCH20190735 mainly clustered together with strains discovered in USA from 2010 to 2023, with the Bootstrap value of only 50%.

Fig. 2
figure 2

Phylogenetic analysis of three major capsid protein genes. The phylogenetic trees of (A) Hexon, (B) Penton base, and (C) Fiber genes were constructed by the neighbor-joining method and the Kimura two-parameter model using 1,000 replicates. The three major capsid protein gene sequences of HAdV-4 previously gained in China are indicated by cyan dots. Pink dots indicate the prototype strain and vaccine strains. Sequences of isolates gained in this study are indicated by red dots

Full size image

Amino acid variation analysis

In comparison with the HAdV-4 prototype strain, there were 9 amino acid substitutions (S138A, D139N, A147V, G199D, K214N, P240T, S256G, T258N, and T289K) of all 11 sequences obtained in this study in the Loop1 region of Hexon protein (Fig. 3A), which were observed in the HAdV-4a subtype, but were not found in the HAdV-4p subtype. One amino acid substitution (N432T) occurred in the Loop2 region of Hexon protein (Fig. 3A), which was found in the HAdV-4a subtype, but was not found in the HAdV-4p subtype except for 5 strains isolated from USA.

Compared with the prototype strain, all 11 sequences obtained in this study had 6 amino acid substitutions (S308A, A310T, G321D, A333V, E339K, and D353E) in the RGD Loop region and 1 amino acid substitution (D158N) in the HVR1 region of Penton base protein (Fig. 3B), which were observed in the HAdV-4a subtype, but were not found in the HAdV-4p subtype.

Compared with the prototype strain, the 11 sequences did not have any amino acid mutations in the key regions binding to CAR, but had 3 amino acid substitutions (K329R, H396I, and C425E) in the Knob region of Fiber protein (Fig. 3C), which were also observed in the HAdV-4a subtype, but were not found in the HAdV-4p subtype.

Fig. 3
figure 3

Amino acid variation analysis of three major capsid proteins. (A) Amino acid in the Loop1 and Loop2 of Hexon. (B) Amino acid in the RGD Loop and HVR1 of Penton base. (C) Amino acid in the key regions binding to CAR and Knob of Fiber. In the HAdV-4 prototype strain RI-67, the position of Loop1 was at 135–297 amino acid, and that of Loop2 was at 401–451 amino acid of Hexon. The position of RGD Loop was at 295–353 amino acid, and that of HVR1 was at 146–165 amino acid of Penton base. The regions binding to CAR were 253, 255, 256, and 274 amino acid location, and that of Knob was 318–425 amino acid location of Fiber

Full size image

Nucleotide variation analysis of ITRs

Through the nucleotide variation analysis of the ITR region (Fig. 4), it was found that the majority of HAdV-4 sequences in China, including all 11 isolates obtained in this study, and most prevalent strains from other countries since 1978 had NF-I binding sites. However, it was not found in the ITRs of the HAdV-4p subtype and the simian adenovirus (SAdV) either. The NF-I in HAdV-4 was almost identical to that in species B.

Fig. 4
figure 4

Nucleotide variation analysis of the ITR region. The 11 ITR gene sequences gained in this study, the prototype strain, vaccine strains, ITR gene sequences gained in China in the past, HAdV-4 strain that was first found NF-I binding sites adapting to host (Genbank accession number: KX384956, V0014/France/1978), and other HAdV types except HAdV-4 are all shown in the Figure. In the V0014 strain, the NF-I was 20–33 nucleotide location of the part of the ITR sequence shown in the figure

Full size image

Recombination analysis

One of the essential mechanisms of HAdV evolution is gene recombination. Previous studies have shown that HAdV-4 is an animal-derived virus from interspecific recombination. The genome skeleton of the virus is derived from SAdV-26, and the Loop1 region of the Hexon gene is embedded with HAdV-16 [18]. By Simplot v3.5.1 software, the recombination analysis was conducted on the two whole-genome sequences gained in this study and the prototype strain. There was no new recombination.

Discussion

This study is the largest study in China on the HAdV-4 acute respiratory tract infection in children with the widest geographic coverage and the largest number of included study centers to date. This prospective multicenter study showed that the positive detection rate of HAdV-4 in children hospitalized due to ARI was 0.39% in the total samples and 7.05% of the HAdV positive samples over the course of the study. Li et al. isolated HAdV-4 samples from ARI children hospitalized or treated in the outpatient and emergency department in four provinces of Northern and Southern China during 2013 to 2014. The detection rate of HAdV-4 in the total samples was 0.11%, while 2.99% in HAdV positive samples [1], which was lower than that in this study. On the one hand, this may be related to the short duration of the study and the fact that only nasopharyngeal swab samples were used. On the other hand, it suggests that the incidence rate of HAdV-4 in China has shown an increasing trend in recent years, which needs more attention. There is a relatively large difference in the positive detection rate of HAdV-4 from relevant foreign studies, ranging from 0.04 to 25.13% in the total samples and 2.07–26.9% in the HAdV positive samples [24,25,26]. The reasons may be that most of the studies abroad are single-center studies and the different research areas, testing methods, the number of included patients, the age of the patients, the source and collection time of the samples, the duration of the study, the prevalence of adenovirus at that time, and so on. Based on previous studies, Coleman et al. found that HAdV-4 mainly occurred in military personnel, accounting for 88.5% of all cases, and the incidence rate in the general population, including children, was only 11.5%, which was low [13].

In this study, the main manifestations of children with HAdV-4 infection were fever and cough. Two children had conjunctivitis. Two children were diagnosed with severe pneumonia and developed serious complications. Coleman et al. reported that HAdV-4 infection could cause common respiratory diseases such as URTI, pneumonia and ocular diseases such as conjunctivitis, corneal conjunctivitis, etc. The severity of the diseases varied from mild/self-limiting to serious, even death, and hospitalization frequently occurred [13]. Most domestic and foreign literatures reported that respiratory diseases were most common in HAdV-4 infection. Compared with other common respiratory viruses, HAdV infection including HAdV-4 is more likely to lead to fever (body temperature ≥ 38℃), and the frequency of shortness of breath and dyspnea is relatively low [27]. Among respiratory HAdVs, HAdV-3, HAdV-4 and HAdV-7 are mainly associated with severe infection, especially HAdV-7, which has a stronger replication ability and virulence, and can promote and exacerbate cytokine responses, leading to more severe respiratory inflammation [28]. However, in a few countries, such as Vietnam, Greece, Nigeria, Tunisia, Türkiye, Saudi Arabia and Iraq, ocular diseases were more common in the reported HAdV-4 cases [13]. The reason may be the bias because of less HAdV-4 cases reported in these countries. Japan experienced multiple HAdV-4 epidemics from 1979 to 2006 [29, 30], in which approximately 82.3% of HAdV-4 cases were only ocular infections, especially keratoconjunctivitis [13], which may be related to the more active detection of HAdV-4 ocular infection in Japan [31]. In a systematic review and meta-analysis of databases consisting of PubMed and Embase between January 2000 and January 2020, keratoconjunctivitis was most commonly caused by species D of HAdV, especially HAdV-8 [32]. Notably, HAdV-8 infection could cause more severe symptoms, such as pseudomembrane formation, conjunctival hyperemia, and bloody eye secretion [33]. In this study, All the patients improved and were discharged without death, which may be the bias because of the small sample size. Coleman et al. systematically reviewed 144 articles on HAdV-4 infection published between 1960 and 2020. There was a total of 14 related deaths reported in children, mainly in USA, Singapore, and China, all of which died of respiratory complications [13].

In this study, the whole-genome sequences of 2 strains and the Hexon, Penton base, Fiber sequences of 11 HAdV-4 clinical isolates were obtained to performed phylogenetic analyses. The phylogenetic trees of the whole-genome, Hexon, Penton base, and Fiber sequences based on the genetic distance all formed two branches (HAdV-4p and HAdV-4a), which were consistent with previous REA research results and related reports [1, 18]. We found that the prototype strain (RI-67 strain) from USA belonged to HAdV-4p, which was mainly isolated before the 1980s. Recently, only four HAdV-4p strains were detected, which were from 2000, 2001, 2004, and 2015 in USA, respectively. HAdV-4a first appeared in France in 1978 and has been monitored in Japan, USA, China and other countries since then. In recent years, the detection frequency of HAdV-4a has gradually increased in various countries, gradually replacing HAdV-4p as the main epidemic strain, which is basically consistent with previous research conclusions [1]. All the HAdV-4 strains in China, including the clinical strains gained in this study, have always been situated in the same branch (HAdV-4a) as most of the foreign strains since the 1980s, and the mean genetic distance within the group was very close (≤ 0.002), indicating that the strains obtained in this study had high homology with most strains from different countries and regions, and the HAdV-4 genome was relatively stable in time and geographical space, which would facilitate the immune protection effect of the vaccine, as found by Li et al. [1]. HAdV-4a was the mainly subtype of HAdV-4 in China. Based on the homology analysis, the nucleotide and amino acid sequence identities on the whole-genomes and three genes of the HAdV-4 isolates obtained in this study and recent epidemic strains in China were not less than 98%, respectively, and the mean genetic distance was very close (≤ 0.001), indicating that the genomes of HAdV-4 strains prevalent in China had high conservation and stability. Recently, Gonzalez et al. reported HAdV-4 could be divided into two phylogroups (PGs): PGs I and II, comprising the prototype (p)-like (4p) and a-like (4a) genomes, respectively [34]. Previous studies have shown that the vaccines of HAdV-4 were live oral formulations of the nonattenuated p-like strains exclusively licensed for military use [35,36,37], so it would be necessary to develop new vaccines.

In comparison with the prototype strain, the HAdV-4 isolates obtained in this study had amino acid mutations in the antigen neutralizing sites of the Hexon protein, the HVR1 and RGD Loop regions of Penton base protein, and the Knob region of Fiber protein. The HAdV-4 strains prevalent in China exhibited approximately identical amino acid mutations with those belonging to the HAdV-4a subtype abroad. Similar mutations were also reported in previous studies [1]. However, except for the amino acid substitution (N432T) occurring in the Loop2 region of Hexon protein, which was found in the 5 strains (Genbank accession numbers: KX384957, KX384946, AP014852, EF371058 and AP014853) isolated from USA in 1981, 1986, 2000, 2001 and 2004 of HAdV-4p, these mutations were almost not found in HAdV-4p, indicating that these mutations might be related to the differences in virulence, pathogenicity, and prevalence between HAdV-4a and 4p. Tian et al. reported that a monoclonal antibody, MN4b, had strong neutralizing activity against HAdV-4 and could recognize a conformational epitope (418AGSEK422) in HVR7 of the Hexon protein. A recombinant virus, rAd3-A4R7-1 (containing the related HVR7 neutralizing epitope of HAdV-4 in the Hexon region of HAdV-3), was successfully constructed in vitro and successfully induced an anti-serum that could inhibited HAdV-4 infection. Previous studies and this study all showed that the amino acids of the above epitope were extremely conserved and no amino acid mutations were found, which would provide a new idea for the prevention and treatment of HAdV-4 in the future [38].

One of the proverbial features in HAdV genetics is recombination, which is also an important factor driving HAdV evolution [39]. Previous studies found that HAdV-4 originated from zoonotic diseases, which were generated by recombination of SAdV-26 genome and HAdV-16 Hexon gene [18]. The same results were obtained from the recombination analysis of the two HAdV-4 strains gained in this study, indicating that the two strains were evolved from the prototype strain. In addition, the majority of the sequences in China (including all the 11 isolates obtained in this study) and most of the sequences abroad in recent years found NF-I in the ITR region. This motif existed in almost all the types of HAdV, but not in SAdVs, HAdV-4 prototype strain, and simultaneously circulating isolates, including the vaccine strains. Moreover, NF-I in HAdV-4 was almost identical to that in species B, which was the same as Dehghan’s finding, so it was speculated that it might be derived from the recombination of species B [18]. NF-I is necessary for effective virus replication. It may allow the virus to expand to the immune immature population, and allow HAdV-4 to adapt to new hosts, which was served as an example of the molecular evolutionary of zoonotic and host adaptation for both “new” and “urgent” human pathogens [17]. The recombination event first appeared in 1978 (Genbank accession number: KX384956, V0014/France/1978) [40]. This will further explain why the incidence rate of HAdV-4 has gradually increased in recent years. However, the NF-I motif is located at the forefront of the genome, making it difficult to sequence [41], which may be the reason why this motif was not found in the 5 strains (Genbank accession numbers: OK323253, OK323260, OK323261, OK323262 and OK323263) isolated from Beijing, China in 2019. In addition, no new recombination of HAdV-4 was found in this study, which was consistent with the results of other studies [41], indicating that the HAdV-4 genome circulating in recent years was relatively stable, and the vaccine could maintain long-term effectiveness.

There were also some limitations in this study. Firstly, although the samples came from six hospitals in the north and south of China, China had a vast territory and the findings gained in this study could not represent the HAdV-4 infection in ARI children in all regions of China. Secondly, this study only included hospitalized ARI children and the number of HAdV-4 positive samples was limited.

Conclusions

This multicenter study showed that the detection rate of HAdV-4 in children hospitalized with ARI in China between 2017 and 2020 was relatively low, and the detection rate of HAdV-4 was 0.39% in the total samples and 7.05% of all HAdV positive samples. ARI caused by HAdV-4 may be severe and can also be accompanied by eye diseases. HAdV-4 prevalent in China belonged to the HAdV-4a subtype and the genome was high conservative and stable. This study further clarified the prevalence and genetic characteristics of HAdV-4 epidemic in China and provided reference for the monitoring, prevention and control of HAdV-4, as well as the research and development of vaccines and drugs.