A Deoxyribonuclease 1 Like 3 genetic variant associates with asthma exacerbations

CAPSULE SUMMARY

The analysis of genetic variants from six genes whose expression is predictive of asthma exacerbations revealed a novel association of a regulatory polymorphism from DNASE1L3 in African-Americans and Latinos.

To the Editor,

Recently, Fricker et al.¹ reported that an expression signature in the sputum predicted efficiently exacerbations among asthma patients, showing the greatest performance for frequent exacerbators. This signature was based on six genes (ALPL, CLC, CPA3, CXCR2, DNASE1L3, and IL1B) that can differentiate airway asthmatic inflammatory phenotypes. Thus, we hypothesized that genetic variation within those genes may contribute to susceptibility to exacerbations. In the United States, African-descent minority populations show the highest rates of asthma exacerbations. In fact, African ancestry is a risk factor not only for asthma in Puerto Ricans and African-Americans but also for exacerbations among African-Americans^2,3. Hence, we performed a candidate-gene association study of asthma exacerbations in African-Americans and Latinos.

Details on the study populations and methods are described in the Online Repository. Genotype dosages with an imputation quality (R²)≥0.3, minor allele frequency ≥5%, and located within the gene limits ±100 kilobases were examined for association with asthma exacerbations. In the discovery, 1,002 African-American asthma patients from the Study of African Americans, Asthma, Genes & Environments (SAGE), 470 with exacerbations, were analyzed (Table 1). Asthma exacerbations were defined as a dichotomous variable (presence/absence) based on oral corticosteroids use, hospitalizations, or acute care due to asthma in the 12 months prior to study enrollment. Association testing was performed by logistic regression models with correction for age, sex, and two principal components (PCs). To account for multiple comparisons, we established a Bonferroni-corrected threshold for significance based on the number of independent single nucleotide polymorphisms (SNPs) within each genomic region by linkage disequilibrium (LD) pruning (r²<0.2). Replication was evaluated in 2,181 Latinos with asthma from the Genes, Environments & Admixture in Latino Americans (GALA II) study, including 1,283 exacerbators (Table 1). This two-stage design provided >80% statistical power to detect the association of common variants for different scenarios (Table E1).

Table 1.

Characteristics of the subjects included in the analyses.

	SAGE (n=1,002)		GALA II (n=2,181)
	Exacerbators (n=470)	Non-exacerbators (n=532)	Exacerbators (n=1,283)	Non-exacerbators (n=898)
Gender (% female)	229 (48.7)	286 (50.4)	582 (45.4)	403 (44.9)
Mean age in years ± SD	14.3±5.6a	14.7±4.1	12.2±3.1a	13.4±3.5
African ancestry (%)	79.4±10.7	78.4±12.1	18.9±13.2a	12.6±11.9
Native American ancestry (%)	NA	NA	22.9±22.6b	37.4±25.5
Asthma control, n (%)
Well controlled	60 (15.7)	182 (36.8)	210 (16.4)a	431 (48.0)
Partially controlled	112 (29.4)	150 (30.4)	533 (41.5)a	282 (31.4)
Poorly controlled	209 (54.9)	162 (32.8)	540 (42.1)a	185 (20.6)
Treatment category, n (%)
Step 1	111 (23.6)	249 (47.1)	210 (16.4)a	431 (48.0)
Step 2	183 (39.0)	210 (39.7)	533 (41.5)a	282 (31.4)
Step 3	176 (37.4)	70 (13.2)	540 (42.1)a	185 (20.6)
Baseline lung function
FEV1 (%) predicted	91.2±11.9	92.4±11.6	97.9±15.7a	103.6±14.4
FVC (%) predicted	96.5±11.8	97.0±10.9	102.0±16.0a	107.3±15.3
FEV1/FVC (%) predicted	94.3±8.3	95.3±7.6	96.4±7.6a	97.2±7.2
Total serum IgE (IU/ml)	168.3±4.9a	139.7±4.6	254.0±4.5a	147.4±4.7

Abbreviations: FEV₁, forced expiratory volume in one second; FVC, forced vital capacity; NA: Not available/applicable.

For continuous variables, the mean and standard deviation are displayed. For total serum IgE levels, the geometrical mean and standard deviation are reported. For categorical variables, the number and percentage of subjects in each category are shown.

^ap< 0.01 for the comparison between asthma cases with exacerbations and controls.

The severity medication regime was categorized into three levels based on the use of short beta agonists (step 1), one inhaled corticosteroid, leukotriene inhibitor, or theophylline tablet (step 2), more than one or a combination of an inhaled corticosteroid, leukotriene inhibitor, or theophylline tablet or a combination of inhaled corticosteroids and long-beta agonists (step 3).

Asthma control and treatment category were available for 875 and 999 individuals from SAGE, respectively. Asthma control and severity were available for 2,181 and 2,170 individuals from GALA II, respectively. FEV₁ was available for 807 and 2,068 SAGE and GALA II subjects. FVC was available for 821 participants from SAGE and 2,077 subjects from GALA II. FEV1/FVC was available for 804 and 2,068 SAGE and GALA II subjects. Total serum IgE levels were available for 784 and 1,856 SAGE and GALA II subjects.

A total of 4,089 SNPs was evaluated (Table E2). Among African Americans, an intronic variant in the deoxyribonuclease 1 Like 3 (DNASE1L3) region (Figure 1), rs67622929, exceeded the significance threshold corrected by multiple comparisons (0.05/57 independent variants, p=8.77×10⁻⁴) [odds ratio (OR) (95% confidence interval [CI]) for the C allele: 1.48 (1.18–1.87), p=7.9×10⁻⁴, imputation R²=0.99]. The association of rs67622929 was replicated in Latinos [OR (95% CI): 1.18 (1.01–1.37), p=0.03, imputation R²=0.97]. An inverse-variance weighted fixed-effect meta-analysis of effects was performed since no heterogeneity was detected (Cochran’s Q p=0.103), revealing that the C allele was significantly associated with increased odds of exacerbations [OR (95% CI) 1.26 (1.11–1.43), p=2.6×10⁻⁴]. Population stratification adjustment by African ancestry in African-Americans and African and Native ancestry in Latinos, instead of PCs, revealed similar results (Table E3). Furthermore, the C allele of rs67622929 was also associated with increased risk of exacerbations when adjusted by disease severity in a subset of African-Americans with data available [OR (95% CI): 1.40 (1.10–1.78), p=6.6 × 10⁻³] and Latinos [OR (95% CI): 1.18 (1.01–1.38), p=0.034], as well as in the meta-analysis [OR (95% CI): 1.24 (1.09–1.41), p=1.2×10⁻³, Cochran’s Q p=0.250]. To explore whether this association was population-specific, the SNP rs67622929 was evaluated for validation in European-descent individuals through online databases, and no evidence of association was found (p>0.05) (Table E4). However, the definition of the control group differed from our study, given that individuals without asthma were analyzed instead of non-exacerbators asthma patients. Therefore, the association of this SNP with severe asthma exacerbations among European-descent populations cannot be fully ruled out.

Figure 1.

Regional plot for DNASE1L3 in the discovery sample. The top hit (rs67622929) is shown as a purple diamond, and the rest of the genetic variants are colored based on their linkage disequilibrium (LD) with it using pairwise r² values from African populations from the 1000 Genomes Project. Chromosomal positions, the −log₁₀ transformed p-values and the recombination rate are plotted in the x-axis, right and left y-axis, respectively. The Bonferroni-corrected threshold for significance is indicated by the dashed line.

Next, we assessed the association of rs67622929 with asthma-related traits, including IgE levels and normalized lung function values at baseline. The linear regression models for normalized lung function values were adjusted by the 2 PCs. For log₁₀ IgE levels, age and sex were also included as covariates. The SNP rs67622929 was not associated (p>0.05) with FEV₁, FVC, FEV₁/FVC, or IgE (Table E3).

According to the GTEx database⁴, the C allele of rs67622929 acts as an expression quantitative trait locus (eQTL) that increases the expression of DNASE1L3 in several tissues, including cultured fibroblasts (p_GTEX=7.1×10⁻¹⁶) (Figure E1A, Table E5). This SNP is also a cis-splicing quantitative trait locus (sQTL) for DNASE3L1 in the lung (p_GTEX=9.4 × 10⁻⁹) (Figure E1B) and other tissues (Table E6). Furthermore, rs67622929 has been linked to FLNB in naive CD8 cells trough promoter capture Hi-C, according to OpenTargets⁵. FLNB is located close to DNASE1L3 (Figure 1) and it is transcribed in the opposite direction. FLNB is involved in actin, cadherin, and RNA binding, as well as many biological processes including antiviral mechanisms, cellular response to interferon-gamma, epithelial cell morphogenesis, and actin cytoskeleton organization. Interestingly, Flnb deficient mice show skeletal abnormalities and decreased cellular adhesion⁵.

DNASE1L3 encodes an endonuclease that participates in apoptosis, clearing DNA from circulating apoptotic bodies, and has an important role in autoimmunity protection⁶. Additionally, DNASE1L3 participates in different catabolic and metabolic processes, neutrophil activation, cell development, regulation of inflammasome activation, and cytokine secretion^5,6. Genetic variants from DNASE1L3 have been previously associated with neutrophil counts and systemic rheumatic diseases^5,6, and this is the first report of an association with asthma exacerbations. Interestingly, the allele associated with risk for asthma exacerbations in our study is associated with higher DNASE1L3 expression, an effect that is in the same direction as the one described for this gene in the expression signature associated with asthma exacerbations in the previous study¹. Additionally, DNASE1L3 is overexpressed in induced sputum from patients with eosinophilic asthma^1,7, and in endobronchial samples of asthma patients compared to healthy controls⁸. Despite all this evidence, the mechanistic implication of DNASE1L3 in asthma remains unclear, and functional studies are needed to disentangle it.

We acknowledge some limitations of our study. First, the association of rs67622929 would not be considered significant after applying a more restrictive correction for multiple comparisons accounting for all polymorphisms tested across the six genomic regions (0.05/384 independent variants=1.30×10⁻⁴). Second, the sample size of this study is modest compared to genetic studies of asthma. However, to our knowledge, this is one of the largest candidate-gene association studies of severe asthma exacerbations published to date. Third, heterogeneity in genetic admixture exists in the populations analyzed. While African-Americans are descendants of African and European ancestors, Latinos also have Native American admixture, with varying proportions of each component depending on the Latino subgroup⁹. In this regard, ethnicity-stratified analyses revealed that the effect in Latinos was driven mainly by Mexicans and other Latinos (Table E7).

In conclusion, the SNP rs67622929 was associated with severe asthma exacerbations in African-Americans and Latinos and showed regulatory effects over DNASE1L3 expression.

The authors thank the patients, families, recruiters, health care providers, community clinics, and the study coordinator (Sandra Salazar) participating in GALA II and SAGE.

ABBREVIATIONS

ALPL

alkaline phosphatase, liver/bone/kidney

CI

Confidence interval

CLC

Charcot-Leyden crystal galectin

CPA3

carboxypeptidase 3

CXCR2

C-X-C Motif chemokine receptor 2

DNASE1L3

Deoxyribonuclease 1-like 3

eQTL

Expression quantitative trait locus

FEV₁

Forced expiratory volume in one second

FLNB

Filamin B

FVC

Forced vital capacity

GALA II

Genes, Environments & Admixture in Latino Americans II

GTEx

Genotype-tissue expression

IgE

Immunoglobulin E

IL1B

Interleukin 1 Beta

LD

Linkage disequilibrium

OR

Odds ratio

NLRC4

NLR family CARD domain containing 4

NLRP3

NLR family, pyrin domain containing 3

RNA

Ribonucleic acid

SAGE

Study of African Americans, Asthma, Genes & Environments

SNP

Single nucleotide polymorphism

sQTL

splicing quantitative trait locus