Discussion
For the first time, antibodies against a vast panel of different EVs were analyzed to shed light on the possible role of EVs as risk-modifying factors in the development of allergic sensitization and/or disease. The cumulative number of infections by different EV serotypes was not associated with IgE sensitization. However, the cumulative number of species A EV infections showed a trend for protective association with the development of allergic disease. This inverse association was significant for IgE-sensitized cases with atopic dermatitis, whereas for hay fever and asthma only a statistically insignificant inverse tendency was observed. Especially asthma is a heterogeneous disease and many patients lack atopic background25, which could explain the lack of association. Another explanation could be limited statistical power, since the number of cases with asthma or hay fever were lower than those with atopic dermatitis.
Certain individual EV serotypes showed an association with IgE sensitization and/or allergic disease, including Echo3 and 14 which both showed an inverse association and EV74 which showed a risk association. Interestingly, Echo3 has been reported to associate with a reduced risk of allergic diseases also in a previous study 11. Additionally, we wish to highlight CAV6, CAV16, Echo5, and Echo14 as intriguing protective players in the development of allergic diseases. Especially CAV6 exposure, a common virus causing hand, foot and mouth disease, was significantly less frequent in cases with allergic disease than in their matched controls and CAV6 was also less frequent in cases with stronger (> 0.7kU/L) IgE sensitization. The inverse CAV6 association was observed for boys with IgE sensitization and boys with atopic disease, but not for girls. This sex difference is parallel with our previous report of rhinoviruses’ association with reduced risk of atopy within boys but not in girls21, and indicates that viruses’ influence on atopy risk may be more pronounced in males. The fact that only few of the 41 EV serotypes showed association with atopy may relate to immunofenotype differences between virus serotypes and strains, as suggested by our previous report showing very different immune system activation by individual CBV1 strains27. It is also known that EVs differ in their ability to infect various white blood cell populations26. Thus, it is possible that certain EV types, such as CAV6, have a unique “anti-allergic” immunophenotype due to their specific interaction with the immune system. This hypothesis requires further studies.
Previous studies have not identified any specific microbes driving the inverse association between microbial exposures and the risk of allergic disease. Instead, it seems that a vast variety of microbes, perhaps even the total microbiome, is involved4. Our finding of an inverse association between the cumulative exposure to different EV types and allergic disease fits with the hypothesis of more than a single culprit microbe involved. Accordingly, even though we found some significant associations when analyzing the individual serotypes, we want to encourage opting for as large virus panel as possible in future studies on EVs and atopy.
Strengths of this study are the comprehensive panel of the most common circulating EV serotypes and the utilization of a highly serotype specific neutralization assay, which can be considered the gold standard for determining serotype specific EV antibodies. The endpoints of our study were also solid, as the IgE sensitization was determined with a wide allergen panel including both inhalant and food allergens, and information on allergic diseases was based on either a validated atopy questionnaire (ISAAC) or a diagnosis made by a physician.
The limitations of the study include the lack of correction of Pvalues for the great number of comparisons performed when analyzing the individual EV serotypes. In our study using a Bonferroni coefficient of 41 (number of EV serotypes studied) leads to a statistical significance cutoff value P < 0.0012 and wipes out the significant associations for individual serotypes. Bonferroni correction has been criticized for being overly conservative, and therefore we held to the generally accepted P value <0.05. One must keep in mind that without correction for multiple comparisons the results for individual EV serotypes must be kept preliminary and they need confirmation in further studies.
A further limitation is that the study subjects did not represent the general population but carried specific type 1 diabetes-associated HLA haplotypes. However, no consistent relationship has been shown between allergic diseases and type 1 diabetes 27,28, and neither islet autoimmunity nor HLA was associated with atopic sensitization in the current cohort. Lastly, we were unable to ascertain the chronological order of the EV infections and IgE sensitization, as the latter was measured only at the age of 5 years.
In conclusion, we show that there is an inverse association between cumulative EV infections and allergic disease but not with allergic sensitization alone. Whether this relationship is causal and how EVs could obstruct the disease progression from sensitization to allergic disease needs to be further studied. We also want to highlight serotypes Echo3, Echo5, Echo14, CAV6 and CAV16 as possible protective serotypes and EV74 as a possible risk factor in allergic sensitization and disease.
Acknowledgements: The authors would like to thank the families who have participated in the DIPP study as well as the study nurses and other personnel running the study.
Impact statement: Hygiene hypothesis has long been proposed to explain the increasing incidence of allergic sensitization and disease. Enteroviruses, the most common viruses infecting human beings, have been associated with the risk of allergy and asthma, but it is not known whether this association links to certain enterovirus types. The present study utilized the most comprehensive serotype panel thus far applied and was able to strengthen previous indications of a possible role of enteroviruses in the regulation of the risk of allergic diseases.