In the present study, rare variants (minor allele frequency (MAF) ≤ 1%) in 121 asthma-associated loci were investigated and analysed for their cumulative burden of association with asthma onset in a cohort of 576 cases with physician-diagnosed asthma.
The demographic and clinical characteristics of the Norwegian cases with asthma and the controls are presented in Supplementary Table 2. Briefly, twice as many females as males were present and approximately two-thirds of the cases developed asthma before the age of 16 (n = 361/554, 64.4%). Thirty-three percent of the cases reported poor asthma control, with an ACT (11) score of ≤ 19 (n = 111/337, 32.9%). About half of the cases with an ACT ≤ 19 developed asthma before age 16. Furthermore, spirometric measurements showed noticeable airway obstruction (reduced FEV1) among the cases. Several clinical variables were significantly different between the cases and controls (p < 0.05, Supplementary Table 2). Approximately 1% of the controls had other lung diseases (10).
The mean target coverage was 381 X and 97% of all the targeted bases had coverage of > 20 X. After quality control filtering, 2764 of the variants previously not reported or with a MAF ≤ 1% in ExAC and an in-house database distributed in 92 genes were available for gene-level SKAT analyses (13). The SKAT combines the effects of multiple gene variants. Each of the 92 genes selected for the SKAT analysis contained ≥ 10 variants across the analysed cases. The results of the cumulative burden tests of the 92 genes with the age of onset in the 576 cases are shown in Table 1. Four genes, phospholipase A2, group VII (PLA2G7), elongator acetyltransferase complex subunit 1 (ELP1), hedgehog interacting protein (HHIP) and spermatogenesis-associated serine-rich 2 like (SPATS2L), were associated with the age of onset (p = 0.015, p = 0.028, p = 0.033, and p = 0.034, respectively, uncorrected for multiple testing) in our Norwegian cohort.
Table 1
Genes associated with asthma onset in the SKAT.
Gene
|
N cases in total
|
N rare variants
|
N cases with rare variants
|
Q stats
|
p-value
|
PLA2G7
|
572
|
16
|
16
|
15.2
|
0.015
|
ELP1
|
572
|
28
|
29
|
25.3
|
0.028
|
HHIP
|
572
|
16
|
21
|
18.8
|
0.033
|
SPATS2L
|
511
|
24
|
28
|
23.4
|
0.034
|
The rare variant association analysis included variants located in the exons, splice sites and UTRs with a MAF ≤ 1% in the Exome Aggregation Consortium (ExAC, broadinstitute.org) database and a MAF ≤ 1% in the in-house controls. Frequencies of the included variants were timely checked against the gnomAD v2.1.1 database (broadinstitute.org). The SKAT analyses were adjusted for sex.
|
Table 2: Demographic and baseline clinical characteristics of the carriers and non-carriers of the rare variants in PLA2G7.
Eight of the 16 variants included in the SKAT for PLA2G7 have not previously been reported in either ExAC or gnomAD. Three non-synonymous variants and one splice variant were detected (Supplementary Table 3). The pathogenicity of the three nonsynonymous variants Phe110Leu, Ser261Phe, and Val279Phe and the splice variant c.539 + 1G > A is unknown. Other variants in PLA2G7 are associated with atopic dermatitis and asthma (15). It is hypothesised that these variants might affect the catalytic function of PLA2G7 to hydrolyse PAF; the prolonged presence of PAF might influence IgE levels by increasing the recruitment of inflammatory cells such as B-cells.
Carriers of the rare PLA2G7 variants had significantly higher FeNO levels than the non-carriers (18 parts per billion (ppb) vs. 13 ppb, p = 0.013) (Table 2). The carriers also had a significantly lower FEV1 (carriers vs. non-carriers: 84.1% of predicted vs. 91.8% of predicted, p = 0.044) and an increased effect of broncho-dilation with the β-2-agonist (carriers vs. non-carriers; 87.6% of predicted vs. 96.0% of predicted, p = 0.018). Although not significant, a higher proportion of the PLA2G7 rare variant carriers had poor asthma control according to an ACT score ≤ 19 (carriers vs. non-carriers; 41.7% vs. 32.6%) and experienced allergies along with their asthma (carriers vs. non-carriers; 87.5% vs. 73%). Similarly, the level of total IgE was more than twice as high in the PLA2G7 rare variant carriers than in the non-carriers (carriers vs. non-carriers: 124 (16.8–317) vs. 50 (17–132) kU/L).
We also found an association of cumulative burden of rare variants in ELP1, HHIP and SPATS2L with the onset of asthma (p = 0.028, p = 0.033, and p = 0.035, respectively; Table 1). Few studies have investigated the involvement of these genes in asthma. ELP1 regulates NF-κB signalling, which influences immunological pathways, and has been linked to childhood asthma outside of Europe (16). HHIP is involved in lung development and has been associated with lung function (17), and a homozygous variant of SPATS2L was shown to affect the bronchodilator response by increasing the levels of the β2-adrenergic receptor (18). Further statistical analyses did not reveal significant differences between the carriers and non-carriers of the associated genes or variations in the clinical variables in our study (Supplementary Table 4). Supplementary Table 3 shows the distribution of the variants in ELP1, HHIP and SPATS2L included in the SKAT analyses.
The role of rare variants in complex diseases has been widely discussed owing to variations in results (19). In asthma, studies have shown the effects of rare variants in both the coding and non-coding regions of genes (20). Our study would have benefitted from a larger sample size and replication cohort; however, this was limited by the high cost of sequencing at the time of the study initiation. However, clinical variation among the subgroups of patients is known to be associated with specific variants, and our correlation with the age of onset might strengthen this study.
We found an association between the rare variants in PLA2G7, ELP1, HHIP, and SPAST2L and asthma onset (p < 0.05, uncorrected for multiple testing). Furthermore, a higher disease burden, with increased FeNO and reduced FEV1 was observed among the PLA2G7 rare variant carriers.