Molecular characterisation of human adenoviruses associated with respiratory infections in Uganda

Globally, human adenoviruses (HAdV) have been implicated in a broad spectrum of both sporadic and epidemic disease outbreaks [1, 2]. Since their first isolation in 1953 [3], human adenoviruses have evolved into a diverse group of viruses causing a broad range of infections with significant human morbidity, mortality and world-wide spread [4, 5]. They cause infections of the lower and upper respiratory tract illnesses, keratoconjunctivitis, urinogenital infections and acute gastroenteritis in both children and adults [2, 6]. Although adenovirus infections are generally self-limiting [7], human adenoviruses types HAdV-B3, HAdV-B7 and HAdV-E4 have been implicated in both upper and lower respiratory tract infection sometimes with fatal outcomes in both immunocompromised and immune-competent individuals [7, 8].

There are a number of adenovirus pathogens circulating within populations. Up to 113 HAdV genotypes have been recently characterized using whole genome analysis and these include the 51 serotypes originally recorded [9]. These have been classified into seven species under the genus Mastadenoviruses [10]. The seven species; HAdV-A to HAdV-G have been classified based on their biological properties and DNA homology [11]. Characterisation was initially done using serologyand sequencing of the hexonand fiber protein genes but has evolved to include the penton base protein gene and recently; whole genome sequencing (WGS) [9, 12]. Whole genome sequencing provides a clear discrimination of genotypes including recombinants [13]. Species HAdV-B (types 3, 4, 7, 11, 14, 16, 21, 34, 35 and 55), HAdV-C (types 1, 2, 5 and 6) and HAdV-E (type 4) are not only associated with respiratory tract infections but may also be involved in other infections in humans [2, 14].

As the threat for novel and re-emerging infections with pandemic capability increases, there is a justifiable need to carry out baseline surveillance and genetic analysis of endemic pathogens circulating within populations with WGS provifing for the availability of high-resolution genomic data facilitating a deeper insight into the molecular evolution of human adenoviruses [15].

The emergence of novel adenovirus strains has been attributed to recombination of genomes of different viral strains and mutations serving as a major pathway in the molecular evolution of new types and emergent pathogens [7, 16]. Recombination may lead to altered pathogenicity and altered tissue specificities [2] such as HAdV-B55, a highly contagious HAdV type which is a recombinant of HAdV-B11 hexon gene and HAdV-B14 penton base and fiber protein genes with altered tissue tropism and antibody specificity [9, 13].

Human adenovirus is one of the many respiratory disease pathogens that present with influenza like symptoms. In Uganda, influenza like illnesses (ILI) are common with 46.6% of samples from those ILI cases yielding one or more respiratory viruses of which 8.7% are adenoviruses [17]. However, adenovirus outbreaks are not usually reported as they most likely go undetected due to lack of awareness of adenoviruses as causative agents and scarcity of diagnostic tests. Identification and characterization of new and re-emerging adenoviruses is important in the prevention and control of disease outbreaks as it would aid in predicting and preparing for future disease occurrences. As more and more adenovirus-vectored vaccines get approval for use against pandemic diseases, knowledge of circulating adenoviruses is critical to inform deployment of such vaccines in those locations following increased risk in adenovirus seropositive populations.

A retrospective analysis was carried out on samples collected between the period of 2008 and 2016 from patients with influenza-like illness at selected health care facilities. These samples were collected from established sentinel sites for Makerere University Walter Reed Project Influenza surveillance programme: Mulago national neferral hospital in Kampala city, Kayunga district hospital, Bugiri district hospital, Gulu regional referral hospital and Jinja Regional Referral Hospital. A total of 2,298 nasopharyngeal swab samples were collected from individuals aged 6 months and above presenting at the hospitals’ outpatient department with influenza-like symptoms; a fever (above 38⁰C) plus either cough or sore throat within 72 h of presentation. Samples were collected using a dacron swab in a 2-ml cryovial containing virus transport medium enriched with antibiotics and stored at a central laboratory facility at -80 °C until processing.

This study was carried out to determine the prevalence and genetic relationships of human adenoviruses associated with respiratory infections. Epidemiological investigations have revealed that multiple genetic variants may exist within a single particular genotype of HAdV [19] thus the need for routine surveillance and in-depth molecular characterization to identify the prevalence of the different adenovirus genetic variants. Adenovirus infection spread is usually dependant on age of an individual, geographic and seasonal difference dynamics and other demographic factors that affect disease occurrence such as close living conditions and the immune status of an individual [11, 20]. Studies have reported a varied prevalence of up to 49.6% in a military population in Korea [21], 8.7% in Uganda [17], 2.0– 6.1% in hospitalized patients in Guangzhou City, China [22] and among organ transplant patients, 64% were reported to have developed adenovirus disease [23].

A prevalence of 9.8% (n = 225) was observed in this study. This is in agreement with the 8.7% HAdV prevalence that was previously reported for Kampala and Entebbe by Balinandi et al. [17] and 9.4% in Uganda by Mimbe et al.[24]. The observed HAdV prevalence highlights the importance of HAdV detection in clinical settings and surveillance in communities. Indeed studies elsewhere in Africa have reported higher prevalence of up to 30.8% among 6381 patients in Senegal [25] and 19.6% in Tunisia [11].

It was observed that children below 5 years of age were more likely to be infected with HAdV than adults older than 15 years (p = 0.028). Generally, children aged 5 years and below are more susceptible to respiratory viral infections as documented in various studies in sub-Saharan Africa [26]. A higher infection prevalence of 10.3% was observed among children less than 6 years. This is consistent with a study in China that reported HAdV detection rate of 8.55% and 9.4% in Chongqing and Hunan Province respectively in hospitalized children with Respiratory Tract Infections (RTIs) [3]. High infection rates of up to 19.5% have been recorded in Tunisia [26] among children with community acquired lower acute respiratory infections (LARIs) conforming to the trend of infection in our study. This could be attributed to the health-seeking behaviors of parents of infants which is more frequent than in adults and older children [27] as well as previous exposure during early childhood; most of the human population (90%) have been exposed to at least one serotype of HAdV [28].

Throughout the world, HAdV has been shown to cause between 5 and 15% of upper acute respiratory infections (UARIs) and about 5% of LARIs in children below 5 years [11]. This observation is routine for adenoviruses as seen in most of the published data that indicates children to be more susceptible to respiratory adenovirus infections [8, 29]. This is attributed to the immunity not being well developed in children thus making them more susceptible to various infections [2]. Furthermore, transmission among children easily occurs since they usually spend time playing together in their social groups and their personal hygiene habits such as washing hands, proper sneezing and coughing etiquette are not well adhered to.

There was no significant difference in the prevalence of respiratory HAdV infections between males (10.1%) and females (9.5%) and a similar trend is reported in Korea and Egypt [22]. However, in Croatia and Jordan males were reported to be more likely to acquire HAdV infection than females (for Croatia; male-to-female ratio = 3.2:1, for Jordan; male to-female ratio 1.9:1) [30, 31].

Phylogenetic analysis showed HAdV-B and HAdV-C as the only species detected within the tested population. Of the15 isolate sequences, 10 clustered with HAdV‐B and 5 with HAdV‐C. HAdV-B and HAdV-C infections have been reported to be significantly associated with younger children[32] as we observed in this study. Reports from elsewhere have shown varying prevalence between the different species. Cameroon found that HAdV‐B infections were more common and severe than HAdV‐C [33] while Egypt and Senegal reported HAdV‐C to be more predominant [5, 25]. This difference in the HAdV species dominance can be attributed to factors such as variations in geographic location, time period and sampling methodologies.

Human adenoviruses have a worldwide spread thus most individuals have been exposed and will possess anti-HAdV antibodies [28]. In the recent years, the study of adenoviruses has gained interest due to theirpotential use in gene therapy and as delivery vectors in cancer treatments and vaccines [34]. These vectors have mainly been based on replication defective HAdV species C types 2 ad 5 in preclinical and clinical models of gene therapy [34]. However, in most populations many individuals are immune to HAdV-C2 and HAdV-C5 based vectors due to previous infection exposure which has presented a major limitation in their use [35].

The most prevalent genotype in the current study was HAdV-B3which has been reported to be associated with sporadic infection in community and institutional out breaks [36]. HAdV-B3 respiratory infections have also been reported in other parts of the world and have become a global concern [12, 37]. In the United States, HAdV-B3 respiratory infection among the civilians and military population were 34.6% and 2.6% respectively [19]. Of the HAdV-C detected, HAdV-C5 was the commonest which differs from reports in Egypt by Demian et al.[5] who reported HAdV-C1 as the commonest HAdV-C genotype identified. Despite the fact that HAdV C infections are not associated with life-threatening disease in the presence of a functional immune system, latent and persistent infections have been reported [38]. Other genotypes detected in this study were HAdV-C1, HAdV-C2 and HAdV-B7out of which HAdV-B7 and HAdV-B3have been implicated in inflammation of the upper respiratory tract and pneumonia with possible severe disease consequences [39]. Human adenovirus type B3 has been indicated to have a world-wide distribution owing to continuous mutations that has led to a high degree of heterogeneity of the hypervariable regions (HVRs) of HAdV-B3 strains circulating globally [32, 36]. Analysis of the genetic distances showed a close similarity between HAdV-B3and HAdV-B7sequences. In some studies, recombination events between HAdV-B3 and HAdV-B7 have been reported [40] which could explain the close similarity observed in this study. More discriminatory genotyping methods such as whole genome sequencing or PCR targeting different genes encoding the fiber protein, the penton base protein and the hexon protein genes could have been able to discriminate these genotypes more. In this study, we used PCR with primers that targeted the conserved loop 4 of the Hexon protein gene which has been used in other HAdV genotyping studies [18]. This could thus explain the close similarity between some of these genotypes. All the detected genotypes are commonly seen in acute respiratory infections and have been isolated from lungs in children with pneumonia. HAdV-E; a species that has commonly been associated with high rates of febrile respiratory illness in US military recruits [41] was not detected in this study and indeed has been indicated to be rare even in other studies [5].

HAdV stays in circulation throughout the year without a specific seasonal pattern [4]. The peaks of HAdV infections in the current study were observed in the months of February and March, then September, October and December. In 2014, a different trend of infection was observed where by the highest peaks appeared to be in July, August, October and December. A study in Egypt by Demian et al. [5] reported most of the isolates in his study to have been obtained during a four-month period from March through June suggesting that the peak of circulation occurred during that time frame. Some studies reported peaks to occur from late part of the wet season (March to May and September to December). Studies in Kenya reported HAdVs to be associated with higher land surface temperature (≥ 40 °C) more than cooler land surface temperature and this could explain the higher peaks in these years [4]. The presence of specific genotypes did not appear to correlate with any time period except for HAdV-B7 that first appeared in 2014 and continued appearing years afterwards. We were unable to give a conclusive seasonal, age, sex and region circulation trend of the different genotypes as not all the positives were sequenced.

We confirmed the presence of different species of HAdV (HAdV-B3, HAdV-B7, HAdV-C1, HAdV-C2, HAdV-C5, HAdV-C6) associated with respiratory infections especially in children in Uganda. More studies are needed to describe the molecular dynamics of HAdV in Ugandan population through whole genome sequencing for more in-depth characterisation especially in lieu of the common adenovirus-vectored vaccines and therapuetics.