Epstein–Barr virus (EBV) and cytomegalovirus (CMV) are herpesviruses commonly contracted during infancy, and they are prevalent in the human population worldwide. They belong to what is referred to as the human virome or virobiota – a collection of both eukaryotic and prokaryotic viruses are found in humans (1–3). The majority of adults (over 90%) are considered to be seropositive for EBV (4, 5), and the transmission commonly occurs via saliva (6). In early childhood, primary infection with EBV is commonly mild or even asymptomatic, while during adolescence or in adults, the EBV infection more often manifests as acute infectious mononucleosis (5, 7). In relation to CMV, between 60 and 90% of adults are considered to be seropositive, however not without substantial geographic variation that is associated with socioeconomic status (8). CMV transmission occurs through various body fluids including maternal genital secretions at birth and breast-milk, as well as via urine from children and adults (9, 10). EBV can infect B-lymphocytes and epithelial cells (11), while myeloid cells are the main CMV reservoirs (12). After primary infection, EBV and CMV establish latency and persist for life. Both viruses are then chronically or intermittently shed for the lifetime of the host, and their re-activation is under tight immunological control. Still, herpesviruses can be a serious threat to the immunocompromised host (2, 4).

There has been a marked change in life-style and environmental exposures in many parts of the world during the last decades, which has been accompanied by a steep increase of immune-mediated disorders, such as allergy and asthma. Also, the higher sanitation standards of westernized countries and prompt antibiotic availability are believed to cause alterations in our exposure to beneficial microbes, which may be disadvantageous to immune development and maturation (2, 13–15). Using a well-characterized prospective longitudinal birth cohort, we have previously reported that EBV and CMV seropositivity relates both to allergic sensitization (16, 17) and modulated immune function in children (18–20). The presence of early-life microbiota was also investigated in this cohort, where a group of lactobacilli [Lactobacillus (L.) casei, L. paracasei, and L. rhamnosus] and Staphylococcus (S.) aureus DNA were related to opposite allergic profiles at 5 years of age (21). As the order and timing of colonization with these microbial agents has been suggested to affect functional capacity of the infant immune system (15–21), herein we investigated the seropositivity for EBV and CMV at four time points up to 10 years of age in the same children and assessed the possible association between environmental exposures, including gut microbiota colonization following birth and the acquisition of herpesviruses over a 10-year period.

In our study, we took advantage of a longitudinal prospective birth cohort of children followed until 10 years of age, thereby providing a unique opportunity to investigate early environmental exposures and their possible influence on EBV and CMV infections over time. Our results show that these two viruses are acquired with different kinetics and that the EBV and CMV seroprevalence at age of 10 years was 47 and 57%, respectively. When relating these findings to maternal age as well as to early exposures such as duration of breastfeeding, siblings, and day-care attendance, we observed that a delayed EBV and CMV infection was associated with older maternal age. Further, we found associations between the early-life gut colonization and modulation of odds to contract CMV infection as S. aureus colonization early in life reduced the time to CMV acquisition.

Herpesvirus infections are delayed in affluent compared to developing countries (30). In line with these observations, we found relatively low herpesvirus seroprevalence in our study cohort of Swedish children, with around 57% CMV and 47.1% EBV seropositive children at 10 years of age. The prevalence of congenital CMV, which is observed when the virus is transmitted to the fetus upon primary infection, reactivation, or re-infection, has been shown to occur in 0.2–2% of pregnant women in westernized countries [reviewed in Ref. (31)], arguing against a major influence on our results. Also, to avoid misinterpretation of the seropositivity and/or infection kinetics due to interference with maternal herpesvirus-specific antibodies in the circulation of newborn children, we choose to exclude data concerning seropositivity at 6 months of age. In addition, similar P-values we generated when analyses were performed with or without EBV and CMV serostatus data from 1-year olds. When comparing the kinetics between EBV and CMV infections, our results showed that CMV was often acquired earlier in life than EBV, and CMV was also more common in our study population at 10 years of age than EBV. In agreement with our observations of the kinetics of the time of contracting EBV and CMV, which was most rapid during the first 2 years of life for both viruses, children are more likely to be exposed to these herpesviruses through their contact with family and day-care members early in life (9). The reason for CMV being the most prevalent of these two viruses at 2–10 years of age is not clear but may be in part connected to the maturation and expansion of virus-specific memory cells. In agreement with this, we have shown that CMV seropositive infants who are EBV naive have a large population of IFN-γ-producing T cells (19, 20), the presence of which associates with the inhibition of B-cell transformation upon in vitro EBV infection of B cells derived from children in this cohort (20). However, other factors such as differences in the prevalence of herpesviruses among the parents, transmission kinetics, the infectivity, and/or the reactivation potential of each virus may also play a role.

We also found that a delay in both EBV and CMV infections was related to older maternal age at the time of birth. Such associations with maternal age have been previously described for CMV, where the infection was more frequent in children whose mothers were 25 years or younger – which was suggested to be linked to higher CMV viral shedding in younger mothers (32). Alternatively, we can speculate that social and behavioral changes and/or altered hygiene practices with increased maternal age may also play a role in delayed infection pattern with herpesviruses. Whether other coupling factors other than those the results were adjusted for could explain this finding may be possible. Yet, socioeconomic inequality, which is a central risk factor for microbial infections, does not explain our findings since children that became seropositive for either herpesviruses at 2 years of age had similar socioeconomic status to those remaining seronegative at 10 years of age.

The microbiota plays an integral role in supplying important signals for immune priming, immune development, and affecting the susceptibility to pathogens (33, 34). The interactions between our commensal microbiota and bacterial pathogens, especially in early life, have been extensively pursued during the past decades (14, 35, 36), while the interplay between viruses and commensals has only recently begun to be appreciated (2, 3, 33, 37). Indeed, the influence of microbiota on viral infections can have a dual role: it can culminate in protection of the host against the virus and in promotion of the viral establishment (34, 37, 38). In our study, we investigated the presence of lactobacilli or S. aureus DNA in fecal samples as proxies for early-life microbiota, since we have earlier found in our cohort that these are linked to immunomodulatory properties during early childhood (21, 25). Our results indicate that an early S. aureus colonization decreased the time to CMV acquisition. This finding is intriguing in relation to earlier observations on increased risk of allergy in infants who were CMV infected but EBV naive (16). How microbes in the gut condition the host resistance against herpesviruses in humans remains to be addressed. Bleotu et al. found that infection of HeLa cells with herpes simplex virus (HSV) in the presence of inactivated S. aureus induced TNF, IL-6, and IL-8 cytokines, and suggested that bacterial antigens could contribute to altered herpesvirus cell binding and multiplication rate in vitro (39).

Some limitations should be acknowledged. As different species and/or strains of commensals can play a different role, a full microbiota analysis could bring more information about their composition and possible association with CMV, EBV, and other viruses, especially in a larger cohort of children. Also, it could be argued that the statistical model used here may lose stability due to the size of data sets. Nonetheless, using TR as our read-out enabled us to include additional subjects with occasional missing data. Further, interactions with viruses and bacteria could shape immune responses and have previously been described for better (40) or for worse (41, 42) but mainly in animal models. Therefore, despite the limited number of children and commensal species investigated here, our study is unique with its longitudinal approach with data sets following birth to 10 years of age and is among the first to describe an association between early infant gut colonization and timing of CMV contraction. Mechanisms behind the contribution of cross talk between herpesviruses and the gut microbiota for the development of an allergic phenotype remain yet unclear and should be elucidated in future studies.

In conclusion, we report that both EBV and CMV seropositivity occurs relatively late in Swedish children and that maternal age seems to be an important factor for the odds of acquiring EBV or CMV infection. Further, our data suggest that a connection between harboring certain gut species and tuning of host reactivity against herpesviruses may exist, which could ultimately influence maturation and functional competence of the immune system and potentially also allergic sensitization during childhood.

CC-Q performed data collection and experimental analyses and wrote the manuscript. MJ performed data collection and experimental analyses. J-OP performed statistical analyses. EJ performed experimental analyses. TK performed experimental analyses and contributed reagents/materials/analysis tools. CN conceptualized and designed the study and contributed reagents/materials/analysis tools. SS-H conceptualized and designed the study, performed experimental analyses, and wrote the manuscript. ES-E conceptualized and designed the study, contributed reagents/materials/analysis tools, performed experimental analyses, and wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer PK and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.