Characterizing the Genetics of Bronchiolitis by Viral Etiology: Is There a Shared Role in Asthma Development?

This is the authors' abstract. We don't add key points for this paper.

(See the Major Article by Pasanen et al on pages e193–202.) Bronchiolitis is a clinical syndrome of upper followed by lower respiratory tract symptoms in children <2 years of age [1, 2]. There are 2 major viral etiologies of bronchiolitis, respiratory syncytial virus (RSV) and human rhinovirus (RV) [3]. Viral etiology of bronchiolitis varies by age, with RSV causing most cases in infants <1 year of age (peak incidence: age 3–6 months [4]) and RV causing most disease after the first year of life [5]. In addition to age, the clinical syndrome, response to medications used to treat asthma, and impact on the airway epithelium differ [6, 7]. RSV bronchiolitis is additionally 1 of the most common reasons for infant hospitalization, respiratory failure, and death worldwide [1, 8]. Known risk factors for severe bronchiolitis include premature birth, chronic lung disease, and congenital heart disease, among others [1]. However, most bronchiolitis-associated hospitalizations are among infants without these predisposing conditions [9, 10], suggesting that host genetics, viral strain, and other unidentified environmental factors contribute to bronchiolitis severity. While the heritability of severe RSV bronchiolitis has been estimated to be approximately 22% [11], the genetic background, particularly of RV bronchiolitis, remains largely uncharacterized. Improved understanding of genetic factors underlying the distinct entities of RSV and RV bronchiolitis severity could aid in understanding the variable disease pathogenesis by viral etiology and identification of high-risk infants who may benefit from preventive interventions. In this issue of The Journal of Infectious Diseases, Pasanen et al address several of these unanswered questions by conducting the largest genome-wide association study (GWAS) of severe bronchiolitis requiring hospitalization (regardless of viral etiology) conducted to date using data from the FinnGen project [12]. While a few bronchiolitis GWASs have been previously published [13–15], most have focused on RSV bronchiolitis. By including all bronchiolitis hospitalizations, Pasanen et al aimed to assess genetic associations of the broad clinical entity of bronchiolitis, as well as by causative agent (RSV vs nonspecified, or what the authors refer to as non-RSV), which has important implications for understanding disease pathogenesis common to all-cause bronchiolitis and by viral etiology. The stratified analysis is important as RSV and RV bronchiolitis have distinct features suggesting they are 2 different diseases. While there may be genetic associations that are shared across viral etiologies, such as for immune response genes, the distinct genetic associations of RSV and RV bronchiolitis are likely to be diluted in a combined analysis. In the authors’ analysis of all-cause bronchiolitis, significant genome-wide associations were detected within cadherin-related family member 3 (CDHR3) and gasdermin B (GSDMB) loci despite the relatively small sample size (n = 1456 cases in FinnGen) for GWAS analysis [12]. Variants in these loci have been previously associated with RSV-negative bronchiolitis [16] and RV wheezing illness [17] in candidate gene studies. When stratified by the viral etiology of bronchiolitis, these results remained nominally significant in the group with nonspecified bronchiolitis (n = 484), likely a group representing predominantly RV bronchiolitis, but were not significant in the RSV bronchiolitis group (n = 263), suggesting that CDHR3 and GSDMB modulate susceptibility to bronchiolitis caused by viruses other than RSV, consistent with prior data [17]. The single-nucleotide polymorphism–based heritability estimate of all-cause bronchiolitis was 7.6%. Although the authors did not include estimates by viral etiology, heritability might be assumed to be higher for nonspecified bronchiolitis and lower for RSV bronchiolitis based on findings from the primary analysis. While inclusion of analyses stratified by viral etiology is a critically important and novel aspect of this study, non-RSV bronchiolitis was defined as all bronchiolitis in which RSV was not identified by International Classification of Diseases (ICD) codes. This group should instead technically be defined as nonspecified bronchiolitis rather than non-RSV bronchiolitis, as RSV was not excluded. While the approach taken to categorize bronchiolitis etiology is reasonable, including the assumption that the etiology of the majority of nonspecified bronchiolitis is due to RV, there is almost certainly misclassification bias. The viral etiology of bronchiolitis, even during hospitalization, is not always determined, the sensitivity of testing has improved over time but remains imperfect, and multiviral polymerase chain reaction (PCR) may not have been routinely used during the study years. The misclassification of a proportion of nonspecified bronchiolitis cases likely being attributable to RSV would be expected to decrease the average age of nonspecified bronchiolitis and drive the results for this subgroup toward the null. It is plausible to surmise that the authors’ findings of a significant association of the genes of interest with nonspecified bronchiolitis may have been significant if focused on RV bronchiolitis and not diluted by other viral etiologies, including RSV. This limitation, along with the small number of cases, reduced the power and precision of this important stratified analysis of distinct diseases based on bronchiolitis etiology. Although challenging, and recognizing that most GWASs of bronchiolitis conducted to date have been small, future studies of bronchiolitis genetics that accurately capture viral etiology may mitigate these issues. Next, the authors assessed the long-term impact of bronchiolitis on asthma. While the association between severe bronchiolitis in infancy due to RSV and RV and subsequent wheezing and asthma in childhood is well established, it has been unclear whether the association is causal or if severe bronchiolitis in infancy is merely a marker of shared genetic predisposition to bronchiolitis and asthma, or both [18]. The authors performed several analyses, including linkage disequilibrium score regression and a phenome-wide association study, to assess the genetic correlation between bronchiolitis and other complex traits. Findings from these combined analyses showed a high degree of shared genetic factors among severe bronchiolitis and asthma. Taken together, the shared genes are most strongly associated with nonspecified bronchiolitis. The findings are strongly suggestive that bronchiolitis—particularly nonspecified bronchiolitis, the majority of which is likely attributable to RV—is a marker of a predisposition to asthma, rather than causal. This has face validity, as RV bronchiolitis is a clinical phenotype of severe infection (not an exposure), occurs with advancing child age, and typically is not the first manifestation of RV infection. The association of RV bronchiolitis with asthma often occurs following the development of allergic sensitization, and the genes identified as associated with nonspecified bronchiolitis have been identified in asthma GWASs. However, whether this explains nearly all or just a portion of the association of RV bronchiolitis and asthma, whether it is specific only to RV, and the association with specific asthma phenotypes is less clear. This relationship is admittedly complex. The study by Pasanen et al provides new insights into the genetic background of severe bronchiolitis, including an analysis highlighting the important distinction between what should be considered 2 different acute diseases, RSV and RV bronchiolitis. In addition to the recognized clinical and pathologic differences in these diseases, the authors’ findings further demonstrate the important distinction in the genetics of these 2 diseases and their relationship with asthma. Studies of the association of bronchiolitis with asthma need to address whether bronchiolitis is an exposure versus a disease entity that has shared genetic architecture with asthma, as well as considering asthma phenotypes as distinct disease outcomes [19]. Financial support. This work was supported by the National Institutes of Health (K01 HL161257 which supports BMS). The content is solely the responsibility of the authors and does not necessarily represent the official views of this agency.