Polymorphism in the gene encoding toll-like receptor 10 may be associated with asthma after bronchiolitis

Toll-like receptors (TLRs) recognise microbes that contribute to the severity of bronchiolitis and the subsequent risk of asthma. We evaluated whether post-bronchiolitis asthma was associated with polymorphisms in the TLR3 rs3775291, TLR4 rs4986790, TLR7 rs179008, TLR8 rs2407992, TLR9 rs187084, and TLR10 rs4129009 genes. The gene polymorphisms were studied at the age of 6.4 years (mean) in 135 children hospitalised for bronchiolitis in infancy. The outcome measure was current or previous asthma. Current asthma was more common (30%) in children with the variant AG or GG genotype in the TLR10 rs4129009 gene versus those who were homozygous for the major allele A (11%) (p = 0.03). The adjusted odds ratio (aOR) was 4.30 (95% CI 1.30–14.29). Asthma ever was more common (34.6%) in girls with the TLR7 variant AT or TT genotype versus those who were homozygous for the major allele A (12.5%) (p = 0.03). The adjusted OR was 3.93 (95% CI 1.06–14.58). Corresponding associations were not seen in boys. There were no significant associations between TLR3, TLR4, TLR8, or TLR9 polymorphisms and post-bronchiolitis asthma. Polymorphism in the TLR10 gene increases and in the TLR7 gene may increase the risk of asthma in preschool-aged children after infant bronchiolitis.

and TLR8 genes are associated with susceptibility to asthma and related atopic disorders 15 and to susceptibility to respiratory viral infections 17 . Signalling via TLR7 and TLR9 affects the function of eosinophils, engendering a link between viral infection and allergic exacerbations 18 . Although TLR10 is a pattern-recognition receptor without known ligand specificity, it has shown to be a modulatory receptor with mainly inhibitory properties 19 .
We have prospectively followed 166 children who were hospitalised for bronchiolitis at less than 6 months of age 2 . At 5 to 7 years of age, 127 of the children attended a clinical control visit, and questionnaire data were available for another 39 children 2 . We have previously studied the TLR1 rs5743618, TLR2 rs5743708, and TLR6 rs5743810 polymorphisms and reported their associations with post-bronchiolitis asthma at preschool age 10 . The present study was carried out to complement this exploratory study series by evaluating whether the TLR3 rs3775291, TLR4 rs4986790, TLR7 rs179008, TLR8 rs2407992, TLR9 rs187084, and TLR10 rs4129009 polymorphisms are associated with post-bronchiolitis asthma. The aim of this study was to compare these polymorphisms between children with and without current asthma, current atopic dermatitis, or current allergic rhinitis at preschool age, or with and without asthma ever combining current and previous asthma in children hospitalised for bronchiolitis in infancy.

Results
The mean age of the 135 patients was 6.4 years at the control visit, and 51% were males. Asthma ever was present in 37 patients (27.4%), current asthma in 18 (13.3%), atopic dermatitis in 46 (34.1%), and allergic rhinitis in 39 (28.9%). The genotypes and minor allele frequencies (MAF) and population data on the MAFs are listed in Table 1. The MAFs of the cases and the Finnish population MAF data 20 did not differ substantially in terms of TLR3 rs3775291, TLR4 rs4986790, TLR7 rs179008, TLR8 rs2407992, TLR9 rs187084, or TLR10 rs4219009 genes.
The TLR3 genotype was wild (CC) in 45.9% and variant (TC or TT) in 54.1% of the cases. The TLR4 genotype was wild (AA) in 83.7% and variant (AG) in 16.3% of the cases. The TLR9 genotype was wild (TT) in 32.1% and variant (TC or CC) in 67.9% of the cases. There were no significant associations between the TLR3, TLR4, or TLR9 genotypes and asthma ever, current asthma, current atopic dermatitis, or current allergic rhinitis (Table 2).
In females, the TLR7 genotype was wild (AA) in 60.6% and variant (AT or TT) in 39.4% of the cases. In males, allele A was present in 79.4% and allele T in 20.6%. Asthma ever was present in 34.6% of the girls who had the variant AT or TT genotype compared to 12.5% of those who were homozygous for the major allele A (p = 0.03) ( Table 2). The odds ratio (OR) adjusted for age was 3.71 (95% confidence intervals [CI] 1.08-12.77). This association was significant in logistic regression adjusted first for early-life risk factors, and then separately for current confounders (data not shown). The association remained significant in logistic regression adjusted for age, early-life risk factors, and current confounders in the same model (OR 3.93, 95% CI 1.06-14.58). The corresponding figures in boys were 33.3% (allele A present) and 37.5% (allele T present) (p = 1.00). There were no significant associations between the TLR7 genotypes and current asthma, current atopic dermatitis, or current allergic rhinitis in either girls or boys (Table 3).
In females, the TLR8 genotype was wild (GG) in 34.8% and variant (GC or CC) in 65.2% of the cases. In males, allele G was present in 50.7% and allele C in 49.3%. There were no significant associations between the TLR8 genotypes and asthma ever, current asthma, current atopic dermatitis, or current allergic rhinitis in either girls or boys ( Table 3).
The TLR10 genotype was wild (AA) in 84.3% and variant (AG or GG) in 16.7% of the cases. Current asthma was present in 30.0% of the children who had the variant AG or GG genotype compared to 10.6% of those who were homozygous for the major allele A (p = 0.03) ( Table 2). The OR adjusted for age and gender was 3.74 (95% CI 1. 19-11.78). This association was significant in logistic regression adjusted first for early-life risk factors, and then separately for current confounders (data not shown). The association remained significant in logistic regression adjusted for age, gender, early-life risk factors, and current confounders in the same model (OR 4.30, 95% CI 1. 30-14.29). There were no statistically significant associations between TLR10 gene polymorphisms and asthma ever, current atopic dermatitis, or current allergic rhinitis ( Table 2).

Discussion
There were three main results in our study on the association of TLRs with asthma at 5 to 7 years of age after hospitalisation for bronchiolitis at less than 6 months of age. Firstly, current asthma was more common in children who had the variant TLR10 rs4129009 genotype. Secondly, asthma ever was more common in girls who had the variant TLR7 rs179008 genotype. Both findings were robust to adjustments with known early-life risk factors for asthma as well as with current confounders at the age of 5 to 7 years. However, TLR10 and TLR7 gene polymorphisms had no significant associations with current allergy. And thirdly, there were no significant associations between TLR3 rs3775291, TLR4 rs4986790, TLR8 rs2407992, or TLR9 rs187084 polymorphisms and earlier or current asthma or allergy.
TLRs play a pivotal role in promoting and controlling innate immune responses. Functional gene polymorphism alters the amino acid structure of the receptor, as has been shown in the cases of TLR3 rs3775291 (Leu > Phe), TLR4 rs4986790 (Asp > Gly), TLR7 rs179008 (Glu > Leu), TLR9 rs187084 (Arg > Trp), and TLR10 rs4219009 (Ile > Leu) gene polymorphisms 21,22 . The consequence of the mutation is dependent on its location. Mutation in the extracellular domain of the receptor may further lead to an altered binding affinity and subsequent immune response 23 , whereas mutation in the cytoplasmic TIR (toll/interleukine-1 receptor) domain, as in the case of TLR10 rs4219009, may result in an altered downstream signalling, despite normal binding 12,24 . Although polymorphism in the TLR8 rs2407992 (2040 C/G) does not change the amino acid (651Leu > Leu), it can potentially affect TLR8 splicing 15 .
TLR10 is a modulatory pattern-recognition receptor with mainly inhibitory properties, and it is able to reduce TLR2 responses by increasing the production of anti-inflammatory IL-1Ra 19 . Further, a recent meta-analysis revealed that polymorphisms of the IL-1Ra encoding genes were associated with asthma, especially in Caucasian populations 25 . Our finding that the TLR10 gene polymorphism was associated with current asthma is in accordance with these observations. The genetic variation in TLR10 rs4129009, which was also determined in the present study, was associated with asthma risk in two independent samples from the USA 26 . In addition, in a Canadian-Australian study, a weak association was observed between another TLR10 polymorphism (rs11096957) and atopic asthma 27 .
In a large German study, a protective effect of genetic variants on atopic asthma was identified in the TLR2-associated heterodimer network consisting of TLR1, TLR6, and TLR10 12 . Corresponding findings in the genes encoding TLR1, TLR2, and TLR6 were also seen in the present post-bronchiolitis cohort, but the direction of the effect was opposite 10 . The variant genotype in the TLR1 gene was associated with asthma during the first 6 years of life, and asthma was present in only two children with the wild genotype in all three polymorphisms 10 . In the most recent study from this cohort 28 , polymorphism of TLR6 was associated with bronchial hyper-reactivity, and if all of the four genes including TLR10 presented with the wild genotype, exercise-induced responses in resistance at 5HZ by impulse oscillometry were significantly smaller than in those with one or more variant genotypes. These findings are in accordance with our current observations stressing the role of the variant TLR10 genotypes in the emergence of post-bronchiolitis asthma. The differences between the German 12 and Finnish cohorts may be due to different asthma phenotypes, allergic asthma in the German study, and post-bronchiolitis asthma in the current study.
The German study reported that primary cells derived from carriers of protective TLR1, TLR6, and TLR10 variants showed augmented inflammatory responses, increased Th1 cytokine expression, and reduced Th2-associated IL-4 production after specific stimulation 12 . The suppressed secretion of allergy-related cytokines, like IL-4, IL-15, and IL-13, seems to be associated with asthma phenotypes not related to allergy 29 .
We found preliminary evidence that the role of the TLR7 gene in asthma may differ between girls and boys, which may be explained by the situation of the TLR7 rs179008 in the X chromosome 30 . A recent experimental study reported that a TLR7 gene defect and early pneumovirus infection in mice interacted with each other, first leading to a severe bronchiolitis-like disease, then to Th2-dominated immunity, and finally to an asthma-like pathology 31 . A Danish study revealed that TLR7 rs179008 polymorphism was associated with asthma 15 . An American study found that TLR7 was expressed in human airway nerves and contributed to relaxation of the airways via the production of nitric oxide 16 . Therefore, normal TLR7 function may be protective against airway  Table 2. Genotypes of TLR3 rs3775291, TLR4 rs4986790, TLR9 rs187084, and TLR10 rs4129009 encoding genes in relation to asthma and allergy at preschool age in 135 former bronchiolitis patients.
hyper-reactivity and asthma, whereas TLR7 polymorphism may predispose to asthma, and our observations suggest that the influence is greater in girls than in boys. Polymorphism in the TLR8 rs2407992 gene had no association with current or earlier asthma, atopic dermatitis, or allergic rhinitis. This finding contradicts the results of the Danish study 15 , in which the same TLR8 polymorphism was associated with asthma, atopic dermatitis, allergic rhinitis, and elevated allergen-specific immunoglobulin E. A Swedish case-control study found an association between TLR8 gene variation and allergic rhinitis in adults 9 .
Polymorphisms in the TLR3 rs3775291, TLR4 rs4986790, or TLR9 rs187084 genes were not associated with current or earlier asthma or allergies. A previous study from the present cohort offered preliminary evidence that the wild TLR3 rs3775291 genotype increased the risk for repeated post-bronchiolitis wheezing 32 . The negative result of the present study at the preschool age was due, at least partly, to the more beneficial outcome of subjects hospitalised for RSV bronchiolitis compared to subjects hospitalised for rhinovirus bronchiolitis. Herein, subsequent asthma was more common after rhinovirus bronchiolitis than after RSV bronchiolitis 2, 33 . Recent meta-analyses have failed to find any associations between TLR4 rs4986790 polymorphism and asthma 34 , and between TLR9 rs187084 polymorphisms and asthma 35 .
There were certain limitations in the present study. The number of patients was relatively small for genetic studies. In addition, blood samples for genetic studies were not available from all cases with sufficient follow-up data available, although we consider the 81% proportion as acceptable. The small number of patients means a risk of type-2 statistical error. On the other hand, we carried out multiple analyses for polymorphisms of six different TLR-encoding genes, which means a risk of type-1 statistical error. We considered our study as an exploratory study aimed at finding preliminary evidence for associations, if present, which needs to be confirmed or rejected in future confirmatory studies. Therefore, and because only two polymorphisms were associated with asthma risk, we did not regard any multiplicity adjustments as necessary 36,37 . Multivariate logistic regression was used, allowing adjustments with early-life risk factors and current confounding factors, but the power of the study was not sufficient for incorporation of all six polymorphisms in the same model.
The strengths of the present study are the prospective design, relatively long follow-up period of six years, extensive virological test panel available during hospitalisation for bronchiolitis, and careful data collection during bronchiolitis and subsequent control visits. The homogeneity of study populations, as in the present study, is a clear benefit for genetic studies. We consider the revealed significant association between TLR10 polymorphism and current asthma at preschool age a reliable finding, since the MAF figures were the same, 0.08, in both cases and population-based controls from the FIN data 20 . Since TLR7 and TLR8 genes are located in the X chromosome, the analyses were carried out separately for both sexes, and indeed, the findings seemed to be different in girls and boys. RSV caused over 70% of the bronchiolitis cases, but subsequent asthma was more common after rhinovirus bronchiolitis than after RSV bronchiolitis 2 . Therefore, the final analyses on the role of TLR10 rs4129009 were performed with multivariate logistic regression adjusted for age, sex, early-life risk factors, and current risk factors for asthma, including the RSV aetiology of early-life bronchiolitis, and the conclusions did not change.
In conclusion, polymorphism in the TLR10 gene seems to increase the risk of post-bronchiolitis asthma in preschool-aged children, and polymorphism in the TLR7 gene seems to increase the risk of post-bronchiolitis asthma in preschool-aged girls. This result was rather strong, since the associations were robust to adjustments for early-life and current risk factors for asthma.

Methods
The study was conducted at the Department of Paediatrics, Tampere University Hospital, Finland, and the design was previously described 2 . In brief, 166 previously healthy, full-term infants hospitalised for bronchiolitis at less  Table 3. Genotypes of TLR7 rs179008 and TLR8 rs2407992 encoding genes in relation to asthma and allergy at preschool age in 135 former bronchiolitis patients. TLR7 and TLR8 are presented separately for females and males. One test result of a TLR7 male presented AT heterozygote and was deleted from the analyses.
of age. In infancy, bronchiolitis was defined as an acute lower respiratory illness characterised by rhinorrhea, cough, and diffuse wheezes or crackles 38 . Early-life data were collected by interviewing the parents during hospitalisation using structured questionnaires 2 . This showed that 14.5% of the mothers and 6% of the fathers had asthma, and 29.5% of the children had atopic dermatitis or food allergy. Data on the viral aetiology of bronchiolitis were studied on admission by antigen detection and polymerase chain reaction (PCR), and a viral infection was identified in most cases: RSV in 70.5% and rhinovirus in 12.7% 2 . During the follow-up study, 127 children attended the clinical control visit, and an additional 39 children returned the study questionnaire. The structured questionnaire completed by the parents of both groups consisted of separate questions on wheezing episodes, asthma diagnoses, and use of bronchodilators and inhaled corticosteroids (ICS) for the preceding 12 months. The presence of atopic dermatitis and allergic rhinitis, night cough in the absence of infections, and prolonged cough for more than four weeks were also charted for the preceding 12 months. The subjects of the present study consisted of those 135 children for whom samples for genetic studies are available.
Definition of asthma. Current asthma was defined as the current use of continuous maintenance medication with ICS for asthma, or suffering from doctor-diagnosed episodes of wheezing, with a prolonged or night cough during the preceding 12 months and with a diagnostic finding in the exercise challenge test (ECT) 2 . The ECT consisted of free running outdoors for 8 minutes and measurements of pre-and post-exercise airway resistance by impulse oscillometry (Jaeger, Master Screen IOS, Höchberg, Germany). An increase in resistance of 35% or more at 5 Hz was considered to be pathological 2,39 . Previous asthma before the control visit was defined as the previous use of inhaled ICS as continuous or intermittent maintenance medication for asthma 2 . If the child had either previous or current asthma, the term asthma ever was used. Allergic rhinitis was defined as episodes of watery nasal discharge not accompanied by fever or by other symptoms of respiratory tract infection 2 . Atopic dermatitis was defined by doctor-diagnosed eczema and atopy 2 .
TLR9 rs187084 (1486 T/C) genotyping was performed using BspTI restriction enzyme (ThermoFhisher Scientific, Waltham, USA) for digestion of the PCR product 28 . High-resolution melting analysis (HMR) (Roche Diagnostics Light Cycler 480, Basel, Switzerland) was used for genotyping of TLR10 rs4129009 (2322 A/G), as described previously 28 . There were 135 samples available for genotyping of the TLR3, TLR4, TLR7, and TLR8. For the genotyping of the TLR9 and TLR10, 134 samples were available.