Background
Human cytomegalovirus (HCMV), the representative member of the beta-herpesvirus, is a widespread viral agent that infects the majority of the world’s population and then establishes lifelong latency. Although infection of healthy individuals is usually asymptomatic, HCMV can be responsible for serious diseases with multi-organ involvement and frequent fatal consequences in ‘at-risk’ categories of individuals, such as those with a deficient immune system due to natural, iatrogenic (e.g. bone marrow or organ transplant patients) or acquired (e.g. HIV-infected subjects) causes [
1‐
3]. In addition, HCMV is still considered to be the main viral cause of birth defects and long-term neurological and sensory
sequelae following congenital infection [
4‐
7].
The consequences of HCMV congenital disease have been reportedly considered as exceeding that caused by other childhood diseases [
8] so that the virus has been assigned the highest priority for vaccine development [
9] even though, to date, there is no licensed vaccine. On that basis, many studies are still addressing the characterization of HCMV strains and the mechanisms being responsible for infection in utero
, with the goal of finding reliable markers to distinguish congenital from post-natal infections.
Many gaps remain in our knowledge about the mechanisms that determine infection outcome and the duration and severity of clinical manifestations, which may involve immunological factors of the host as well as purely viral determinants [
10]. Although little data is available about the impact of HCMV virulence factors on infection outcome, several Authors sustain a key role of the HCMV envelope glycoproteins, such as gB [
11‐
13]. Indeed, in addition to being a target of neutralising antibodies and crucial for the virus interaction with cell receptors, gB is encoded by the UL55 gene presenting a number of polymorphic regions which account for its genotypic and phenotypic variability, giving rise to four principal subtypes (gB1-gB4) of HCMV circulating strains [
14,
15].
More recently, other envelope glycoproteins have been indicated as putative HCMV virulence factors, such as the glycoproteins N (gN) and O (gO) [
16‐
18]. Similarly to gB, the genes (UL73 and UL74) coding for these glycoproteins possess hypervariable regions, resulting in a number of gN and gO subtypes. The gN variants are as follows: gN1, gN2, gN3a, gN3b, gN4a, gN4b, gN4c; in relation to gO, four main clades have been described, gO1-gO4, which can be further divided into seven genetic variants (gO1a, gO1b, gO1c, gO2a, gO2b, gO3, gO4) [
19]. Glycoprotein N is involved in virus attachment to the host cell and viral spread, while gO participates in the fusion of the viral envelope to the host cell membrane, promoting HCMV penetration, envelope acquisition and release [
16,
17,
20‐
22].
Considering that the genetic polymorphisms underlying the specific differences between gB, gN and gO subtypes can influence the ability of HCMV to preferentially target specific host cells, it is very likely that they play an important role in defining HCMV infection outcome [
12,
16,
23]. It is also of note that genes encoding the above-mentioned glycoproteins generally act in a coordinated and synergistic way [
17,
19,
24]. Thus, in the quest to identify predictive biomarkers of infection outcome, studies addressing the combined polymorphic patterns of HCMV genes encoding envelope glycoproteins are much more representative than those focussed on single polymorphisms.
Based on the aforementioned notions, the present study focussed on HCMV gB, gN and gO gene polymorphisms in viral strains present in urine samples of paediatric patients with congenital or post-natal HCMV infection, to investigate whether the prevalence of combined genetic variants may be associated with congenital infection.
Discussion
HCMV congenital infection is one of the major causes of birth defects and late neurological and sensory
sequelae in children. Around 0.2–2 % of newborns acquire HCMV infection in utero [
25,
26], but only 10–15 % of children with congenital infection show relevant clinical signs at birth [
5,
27‐
29].
HCMV envelope glycoproteins showing genetic polymorphisms among circulating strains have been considered as potential virulence markers, and may be responsible for differential HCMV tropism to specific cell types and differential ability to disseminate and interfere with normal tissue development [
13,
16,
21,
23]. A number of studies have focussed on these viral components, although most of them have only addressed singular glycoprotein subtypes. In particular, UL55 (coding for gB), the less polymorphic gene among those coding the main envelope glycoproteins, has been analysed in different categories of subjects at risk of developing diseases upon HCMV infection [
30‐
32]. Some Authors found that all the four dominant gB genotypes can be transmitted to the
foetus in the case of congenital infections [
33‐
35], while others found gB1 to be the prevalent genotype in congenital infections [
11,
36]; a prevalence of certain genotypes in patient cohorts characterised by distinct geographical or ethnic contexts has been also reported [
37,
38]. Collectively, these data do not favour the hypothesis that specific gB genotypes alone could provide reliable markers of congenital infection, partially due to the low genetic variability of the
locus (9.5 %) and also because most of the gB subtypes demonstrate a similar probability of being associated with congenital infection [
33]. Conversely, the UL73 gene, coding for the envelope glycoprotein N, possesses highly hypervariable regions (50 % variability). The UL74 gene coding for gO also shows a considerable (30 %) variability, but has not been largely analysed in this category of at-risk subjects [
38]. In summary, although no conclusive results concerning the infection outcome and tissue tropism of HCMV variants carrying specific glycoprotein polymorphisms have been produced to date, the genetic characterisation of the infecting HCMV strains appears useful for the infection prognosis and, thus, remains the focus of many current studies [
13]. As already mentioned, it is unlikely that just a single polymorphic gene product is responsible for a specific infection outcome and/or tissue tropism, especially in the case of a viral agent as complex as HCMV. Instead, it is probable that the viral phenotype is determined by a combination of polymorphic genetic
loci that operate synergistically [
19,
24]; this is likely the case for UL73 and UL74 (coding for gN and gO, respectively).
In this study we have addressed the polymorphisms of HCMV genes encoding gB, gN and gO envelope glycoproteins in a cohort of paediatric patients with congenital or post-natal HCMV infection, in order to assess whether the presence of specific genetic combinations of viral glycoprotein subtypes may constitute reliable markers of infection outcome. By first analysing the distribution of single genotypes, we found a significant predominance of gN4 in congenital infections, in accordance with other studies [
17,
18], implicating this genotype as a potential prognostic marker of HCMV congenital infection. Furthermore, our results support a significant association of congenital infections with the gN4c genetic variant; a similar trend was found for gO3. The RFLP results were further confirmed by applying DNA sequencing and phylogenetic analysis to randomly selected samples.
With regard to combined genotypes, we detected a statistically significant association of gN4 and gO3 genetic variants in congenitally infected children; it is noteworthy that, among those presenting clinical signs at birth (7/19), all were gN4 and nearly all (6 out of 7) showed the gN4-gO3 combination. On the other hand, most of the post-natally infected infants did not show any clinical signs at birth; a minority of them presented clinical signs not compatible with HCMV infection (see Fig.
4).
It is also of interest to consider that although congenitally infected newborns with clinical signs are at higher risk of developing long-term neurological and sensory
sequelae (about 49 % on average) [
6,
39‐
41], a considerable portion of asymptomatic babies (about 13.5 %) eventually manifest such problems [
6,
26]. Moreover, the above mentioned percentages may be under-estimated in the literature because data on late disabilities are often incomplete and follow-up periods too short to identify late
sequelae.
Thus, the development of prognostic markers of HCMV congenital infection is of great importance for the identification and characterisation of predominant genetic variants of HCMV in at-risk patients, as well as for the development of protective vaccines. With regard to the latter, the major problem concerns the high variability of circulating strains that are differentially involved in clinical manifestations with different degrees of severity, making it highly difficult to choose a viral product that would present the highest level of safety whilst effectively eliciting humoral as well as cellular immunity. In this context, it is noteworthy that the gN viral glycoprotein is one of the viral products eliciting the highest neutralising antibody response and that anti-gN4 immunity also seems to protect against frequent re-infection [
42]; it is also likely that specific genetic variants of the gN4 protein (such as gN4c as shown here) could enhance this relevant feature even further.
Another interesting aspect of HCMV congenital infections is the possible presence of multiple viral strains. Indeed, although a number of studies have outlined the notion that mixed infections are more frequently found in adult immunocompromised hosts [
15,
43], more recently published studies have also described the presence of multiple HCMV strains in congenital infections at a relatively high rate (15–46 %) [
44,
45].
Surprisingly, we didn’t find mixed HCMV infections in our study population. This might be attributed to the fact that our sample size, although sufficient to obtain statistically significant results, was quite small, thus associated with a lower probability of detecting mixed infections; for instance, similar results were also obtained recently by other Authors [
7]. Moreover, it has been reported that a mixed infection has a higher chance of being detected if the less represented genotype constituted at least 25 % of the total viral population [
7]; thus, we can speculate that additional genotypes may have been present in a smaller proportion in our study population. Another interesting notion, highlighted by Ross and collaborators [
44], is that in some cases only one type of HCMV genotype could be found in a specific biological sample, while other genotypes could prevail in samples obtained from different compartments of the same child. This observation could also account for the apparent absence of mixed genotypes in our study population as only urine samples were tested.
Work is in progress in our laboratory to collect and analyse more than one type of biological sample from HCMV infected babies.
In summary, identification of the predominant molecular combinations of glycoprotein subtypes in congenital HCMV infections, like gN4-gO3 as shown in the present study, is crucial for unravelling the complicated interplay between glycoprotein genetic variation and HCMV pathogenicity. Identification of such combinations will also permit a more accurate prenatal diagnosis and focussed, long-term follow-up care for infected babies.
Conclusions
This study is one of few to address not only the polymorphisms of individual genes encoding HCMV glycoproteins, but also their combination in one of the most important categories of at-risk subjects, i.e. congenitally infected children.
Although the sample size considered in the present study is limited, it has sufficient power to detect significant differences in the two paediatric groups considered. Indeed, we reveal that the gN4-gO3 genotype association is statistically significant in the congenitally infected cohort.
These results indicate that the gN4 (mostly the gN4c variant) and gO3 glycoprotein genotypes could provide reliable markers of congenital infection. The identified variants could also constitute ideal candidates for prophylactic measures that aim to counteract the onset of late disabilities in congenitally infected children.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MCA conceived and designed the experiments; RVS and IR performed the experiments; MCA, FDC, MCM, DM and AC analyzed the data and performed statistical analysis; MCA and AC contributed reagents/materials/analysis tools; RVS and IR helped to draft the manuscript. MCA and CC wrote the paper; AC and CC revised critically the manuscript; All authors read and approved the final manuscript.