Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Epigenome-wide association study identifies DNA methylation markers for asthma remission in whole blood and nasal epithelium

Epigenome-wide association study identifies DNA methylation markers for asthma remission in whole... Background: Asthma is a chronic respiratory disease which is not curable, yet some patients experience spontane‑ ous remission. We hypothesized that epigenetic mechanisms may be involved in asthma remission. Methods: Clinical remission (ClinR) was defined as the absence of asthma symptoms and medication for at least 12 months, and complete remission (ComR) was defined as ClinR with normal lung function and absence of airway hyperresponsiveness. We analyzed differential DNA methylation of ClinR and ComR comparing to persistent asthma (PersA) in whole blood samples (n = 72) and nasal brushing samples (n = 97) in a longitudinal cohort of well char‑ acterized asthma patients. Significant findings of whole blood DNA methylation were tested for replication in two independent cohorts, Lifelines and Epidemiological study on the Genetics and Environment of Asthma (EGEA). Results: We identified differentially methylated CpG sites associated with ClinR (7 CpG sites) and ComR (129 CpG sites) in whole blood. One CpG (cg13378519, Chr1) associated with ClinR was replicated and annotated to PEX11 (Per‑ oxisomal Biogenesis Factor 11 Beta). The whole blood DNA methylation levels of this CpG were also different between ClinR and healthy subjects. One ComR‑associated CpG (cg24788483, Chr10) that annotated to TCF7L2 ( Transcription Factor 7 Like 2) was replicated and associated with expression of TCF7L2 gene. One out of seven ClinR‑associated CpG sites and 8 out of 129 ComR‑associated CpG sites identified from whole blood samples showed nominal significance (P < 0.05) and the same direction of effect in nasal brushes. Conclusion: We identified DNA methylation markers possibly associated with clinical and complete asthma remis‑ sion in nasal brushes and whole blood, and two CpG sites identified from whole blood can be replicated in independ‑ ent cohorts and may play a role in peroxisome proliferation and Wnt signaling pathway. Keywords: Asthma remission, DNA methylation, Whole blood, Nasal brushes Introduction to now. However, some asthma patients may grow out of Asthma is a chronic airway disorder that affects more this disease, which is called asthma remission. than 300 million people around the world. Asthma can- Asthma remission is more common in patients with not be completely cured with the available treatment up childhood onset asthma, and the proportions of remis- sion differ in different age groups at follow-up (33–53% in adolescence, 6–33% in young adulthood and 11–52% in adulthood) [1, 2]. There are two types of asthma remis - *Correspondence: g.h.koppelman@umcg.nl Department of Pediatric Pulmonology and Pediatric Allergy, Beatrix sion, one is defined by absence of asthma symptoms and Children’s Hospital, University Medical Center Groningen, University medication for at least one year, which is called “clinical of Groningen, PO Box 30.001, 9700 RB Groningen, the Netherlands remission” (ClinR). Some ClinR subjects may still have Full list of author information is available at the end of the article © The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Qi et al. Clin Transl Allergy (2020) 10:60 Page 2 of 13 airway hyperresponsiveness (AHR) or a low lung func- in two independent cohorts, and also verified the top tion. Therefore, “complete remission” (ComR) was put results from whole blood DNA in cells obtained by nasal forward [3]. In addition to criteria of ClinR, ComR has brushing. to meet additional criteria of normal lung function and absence of AHR. This is a more rare phenomenon, that Methods takes place in 5–22% of asthmatics [2]. A full description of methods is provided in the online Asthma remission is associated with both genetic and supplement. environmental factors. Environmental factors includ- ing breast feeding and having pets in childhood were Study populations reported to be positively associated with asthma remis- Subjects included in this study were from long-term sion [4]. A recent genome-wide association study follow-up studies in the University Medical Center Gro- (GWAS) identified one SNP (single nucleotide polymor - ningen (UMCG). This study consists of two sources of phism) that was associated with ClinR and three SNPs subjects, (1) subjects from the third visit (2013–2014) of with ComR. One of these SNPs was also associated with a longitudinal study described previously by Carpaij et al expression of known asthma genes including ILRL1 and [4], and (2) additional asthma remission subjects that IL13 in lung tissue [5]. were followed up from previous genetic studies [11] and Epigenetic mechanisms such as DNA methylation may were re-invited in 2013–2014 to take nasal brushes and help to build a link between genetic factors and the envi- blood samples, during which the remission status was ronment. DNA methylation can be regulated by SNPs, assessed again. The participants included in this cohort and can reflect environmental exposures, ageing, cell type were different from those of the cohort described by Ver - constitution and activation [6]. DNA methylation refers meulen et  al. [10] from our research institute. All par- to the addition of a methyl-group to cytosine, which usu- ticipants were diagnosed with asthma at baseline with ally happens when a cytosine is located next to a guanine a doctor diagnosis of asthma and AHR. Then, subjects in the 5′ to 3′ direction (CpG site), and this may relate to had at least one follow-up medical examination dur- the regulation of gene expression [7]. Epigenome-wide ing adulthood in which their asthma status was evalu- association studies (EWAS) have provided insights into ated by questionnaires and in most subjects additionally the development of asthma and its remission. CpG sites with spirometry and an AHR test [5]. The medical ethical associated with asthma in both blood and nasal epithe- board of the UMCG approved the studies and all partici- lial cells have been identified, such as CpG sites located in pants gave written informed consent. DICER1, STX3, and LIPIN1 in blood cells, and CDHR3, We replicated findings in whole blood in two independ - FBXL7 and NTRK1 in nasal epithelial cells [8, 9]. Ver- ent cohorts: the Lifelines population-based cohort in The meulen et  al. [10] identified 4 CpG sites and 42 regions Netherlands (where we could replicate our results on that were differentially methylated between remission ClinR using Illumina 450 K array) and the Epidemiologi- and persistent asthma (PersA) in bronchial biopsies, and cal study on the Genetics and Environment of Asthma top CpG sites were annotated to genes including ACKR2 (EGEA) cohort in France, a case control and family and DGKQ by gene expression. study on asthma, where we could replicate our results on Although this latter paper provided a proof of con- ClinR and ComR with methylC-capture sequencing [12]. cept of the relation between asthma remission and DNA No cohort was available to replicate our nasal findings. methylation, bronchial biopsies are not easy to obtain Extensive information on these cohorts is described in for further studies. DNA methylation in whole blood the Additional file 1. and nasal epithelium is a good proxy for bronchial epi- thelium and can therefore also help to understand the Phenotype definition mechanism of asthma remission. Here, we hypothesize The presence of PersA, ClinR and ComR was determined that epigenetic mechanisms may be involved in asthma at the most recent visit. ClinR was defined according to remission, reflected in different DNA methylation pat - the following criteria: (1) No use of any asthma medi- terns in whole blood and nasal epithelium. To test this cation, and (2) no symptoms (asthma attacks and/ or hypothesis, we performed an EWAS of whole blood and wheezing) in the past year. ComR was defined as ClinR nasal DNA in subjects with PersA, ClinR and ComR. combined with (3) no AHR (PC (provocative concen- We investigated a longitudinal cohort in which asthma tration causing a 20% fall in FEV (forced expiratory −1 was initially carefully defined and the remission status volume in 1 s)) methacholine ⩽ 39.3 mg mL )), and (4) was assessed during follow-up (median 39  years). We FEV % predicted pre-bronchodilator > 80%. PersA was subsequently replicated the whole blood DNA results defined as the presence of asthma symptom and/or the Qi  et al. Clin Transl Allergy (2020) 10:60 Page 3 of 13 use of asthma medication. Detailed phenotype defini - of ± 250  kb) by cis expression quantitative trait DNA tions of the replication cohorts are described in the Addi- methylation (cis-eQTM) analysis. Blood eQTM was tional file 1. assessed in the BIOS consortium dataset [19]. Matched nasal RNA sequencing data and DNA methylation data DNA methylation measurements and statistical analyses was used to perform eQTM analysis for significant nasal DNA was extracted from 72 whole blood and 103 nasal CpG sites and these results were additionally replicated brushing samples taken at the most recent visit. Genome- in the larger collection of nasal brushes obtained in the wide DNA methylation was determined using Illumina PIAMA study [20]. The genes that were identified in Infinium HumanMethylation450 BeadChips. After qual - eQTM analysis were used for pathway analysis by Con- ity control, 72 whole blood samples and 97 nasal epi- sensusPathDB [21]. To check if the results of replicated thelium samples (of in total 103 unique subjects) with CpG sites were affected by inhaled corticosteroids (ICS), 436,824 CpG sites remained for following steps. we stratified asthma subjects by ICS usage, and compared We used robust linear regression to determine the the DNA-methylation in the remission group to asthma differential methylation between persistent asthma with ICS and asthma without ICS group, respectively, by and asthma remission: (1) PersA versus ClinR, and (2) t-test. To evaluate the impact of BMI and allergic rhinitis PersA versus ComR, in whole blood and nasal brush- on the results, we performed sensitivity analyses on the ing samples, with adjustment for covariates that are replicated blood CpG sites by additionally adjusting for known to affect DNA methylation (age, sex, smoking allergic rhinitis and BMI respectively. We also performed status, pack years, and batch). For whole blood samples, sensitivity analysis for all significant nasal CpG sites. we performed adjustment on the percentage of mono- cytes, B cells, NK cells, CD4 + T cells, CD8 + T cells, Results and granulocytes, which were predicted by the House- Subject characteristics man [13] algorithm using minfi [14] package. For nasal Subject characteristics of the discovery and replication brushing samples, we applied the R package sva [15] to cohorts are shown in Table  1. Regarding the discovery estimate significant surrogate variables (SVs), represent - cohort, the median [range] duration of follow-up of the ing unknown latent factors that capture heterogeneity entire population was 39 [4–49] years. Most of the sub- in data, such as cell type composition. One SV for each jects (group 1, n = 76) were from the study described by analysis (ClinR and ComR) was generated and added to Carpaij et  al. [4] at the third visit; the remaining (group the model respectively. Differentially methylated regions 2, n = 27) were remission subjects added to this study (DMR) were identified using comb-p v0.48 [16] and to further enrich remission cases. Characteristics of DMRcate [17]. subjects from the two groups are shown in Additional file  2: Table  S1. By definition, group 2 had a higher pro - Replication and meta‑analyses portion of remission subjects, and showed better lung Genome-wide significant CpG sites that passed Bonfer - function. Of the 103 subjects included in this study, 54 –7 roni correction (P < 1.14 × 10 , which is 0.05 / 436,824) (52.4% of total subjects) had ClinR, and 20 of the ClinR in whole blood DNA were selected for replication in two subjects had ComR (19.4% of total subjects). At baseline, independent cohorts, Lifelines and EGEA. Lifelines did subjects in remission later had higher FEV and FVC not include an assessment of AHR, so only included a level compared to PersA, but no difference regarding ClinR phenotype available for replication. The weighted FEV1%predicted and FVC %predicted values was iden- Z-score method was used to meta-analyze the results tified among the groups (Additional file  3: Figure S1). of discovery and replication cohorts, considering EGEA Within the whole dataset, 72 subjects (44 PersA, 28 ClinR used methylC-capture sequencing method, which was and 10 of ClinR subjects are ComR) had whole blood different from discovery and Lifelines (450  K array), to methylation data and 97 (44 PersA, 53 ClinR and 19 of investigate DNA methylation. CpG sites that passed the ClinR subjects are ComR) had nasal methylation data, –7 epigenome-wide significance threshold of P < 1.14 × 10 with 66 subjects providing both (Fig. 1). White blood cell (Bonferroni correction) in the meta-analysis of all studies composition was estimated, and we did not observe sig- were considered to be replicated. Finally, top sites identi- nificant cell composition differences among PersA, ClinR fied in whole blood were verified in nasal brushes. and ComR group in six cell types. (Additional file  3: Fig- ure S2). Annotation and functional relevance Significant CpG sites were annotated by GREAT v3.0.0 Differential methylation in the discovery cohort of ClinR [18]. We correlated the significant CpG sites we identi - Whole blood DNA methylation levels at seven individual fied to the expression level of gene nearby (with a region CpG sites and 19 differentially methylated regions (DMR) Qi et al. Clin Transl Allergy (2020) 10:60 Page 4 of 13 Table 1 Characteristics of study participants in the discovery and replication cohorts Discovery cohort Lifelines EGEA PersA ClinR ComR PersA ClinR Healthy PersA ClinR ComR (n = 3) (n = 49) (n = 54) (n = 20) (n = 99) (n = 25) (n = 636) (n = 106) (n = 15) Characteristics at last visit Age at 49.8 (48.5; 50.5 (48.3; 50.5 (48.4; 46.9 43.2 46.3 52.6 (36.7; 49.2 (36.7; 30.5 (28.2; remission 51.3) 57.6) 59.8) (40.7;50.7) (35.7;50.3) (39.2;52.9) 61.4) 57.6) 59.1) status, years Male, n (%) 31 (63.3%) 37 (68.5%) 15 (75.0%) 58 (58.6%) 10 (40.0%) 384 (60.4%) 49 (46.7%) 7 (46.7%) 1 (33.3%) Duration of 39 (38;40.5) 39 (37; 42) 39 (37;41) 11.3 (10.7; 11.6 (10.7; 11.6 (10.0; follow‑up, 12.1) 12.2) 12.7) years Current‑ 4 (8.2%) 8 (14.8%) 2 (10.0%) 27 (27.3%) 14 (56.0%) 207 (32.5%) 10 (9.5%) 2 (13.3%) 1 (33.3%) smoking, n (%) Ex‑smoking, 13 (26.5%) 18 (33.3%) 7 (35.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 35 (33.3%) 5 (33.3%) 0 n (%) Never‑ 32 (65.3%) 28 (51.9%) 11 (55.0%) 72 (72.7%) 11 (44.0%) 429 (67.5%) 60 (57.1%) 8 (53.3%) 2 (66.7%) smoking, n (%) ICS, n (%) 31 (63.3%) 0 0 48 (48.5%) 0 (0.0%) 0 (0.0%) 72 (70.6%) 0 0 Allergic 33 (67.3%) 30 (55.6%) 9 (45.0%) 59 (59.6%) 13 (52.0%) 123 (19.3%) 70 (67.3%) 11 (73.3%) 2 (66.7%) rhinitis, n (%) BMI 25.7 (23.7; 26.8 (23.7; 26.0 (22.8; 25.7 26.0 25.8 23.9 (22.4; 22.5 (19.9; 22.0 (19.9; 28.5) 28.3) 28.6) (22.9;28.0) (23.7;29.6) (23.6;28.1) 27.5) 28.0) 22.4) FEV1% pred 80.6 (71.1; 90.2 (82.9; 101.8 (85.9; 81.2 87.1 101.5 87.0 (71.7; 95.2 (87.5; 97.5 (90.7; 91.9) 104.7) 110.0) (70.9;91.6) (74.0;93.4) (96.1;108.7) 98.4) 107.2) 121.5) FVC% pred 94.2 (88.4; 101.7 (95.2; 106.6 (101.2; 94.1 100.3 104.0 98.5 (84.8; 95.2 (86.9; 96.3 (86.9; 103.7) 111.5) 115.9) (85.8;102.2) (90.4;109.2) (97.5;111.4) 108.9) 106.1) 119.2) FEV1/FVC 0.67 (0.60; 0.72 (0.68; 0.72 (0.68; 0.67 0.67 0.79 0.69 (0.61; 0.81 (0.78; 0.80 (0.80; 0.72) 0.75) 0.76) (0.62;0.76) (0.62;0.76) (0.75;0.83) 0.77) 0.82) 0.91) Characteristics at baseline Age, years 10.0 (9.0; 11.0 (9.0; 11.5 (9.8; 40.6 (25.7; 37.0 (26.0; 18.9 (18.8; 11.0) 17.5) 15.8) 48.7) 46.0) 46.4) FEV1% pred 76.6 (67.4; 77.1 (67.5; 78.1 (73.5; 81.7 (66.2; 90.6 (82.0; 82.0 (80.1; 84.0) 83.8) 83.7) 95.9) 103.0) 122.9) FVC% pred 88.8 (84.0; 88,3 (80.9; 88.9 (79.8; 95.5 (83.7; 86.4 (81.9; 84.5 (70.1; 96.1) 95.3) 96.5) 104.3) 100.6) 118.8) FEV1/FVC 0.73 (0.64; 0.73 (0.66; 0.77(0.66; 0.70 (0.63; 0.86 (0.82; 0.86 (0.83; 0.78) 0.79) 0.80) 0.80) 0.89) 0.96) Start 48 (98.0%) 50 (92.6%) 19 (95.0%) 59 (62.1%) 18 (81.8%) 52 (49.5%) 8 (53.3%) 2 (66.7%) asthma before age 16, n (%) Continuous data are presented as median (25th percentile; 75th percentile), category data are presented as number (percentage) Italic values represent significant differences compared with persistent asthma with P < 0.05 in the discovery cohort were significantly associated with ClinR (Fig.  2a, Table 2, file  2: Table S3). These four eQTM genes were enriched in Additional file  2: Table  S2). The Q-Q plot (Additional two pathways: G Protein Signaling Pathways and Hemo- file  3: Figure S3a) and the inflation factor λ of 1.064 indi - stasis (P < 0.01, Additional file 2: Table S4). cated no obvious inflation of the results. Three out of the We also identified four epigenome-wide significant seven CpG sites were significantly correlated with gene- CpG sites and 24 DMRs that associated with ClinR expression level in cis (FDR < 0.05), resulting in four CpG- in nasal brushes (Table  S5-S6). The Q-Q plot (Addi - gene pairs that showed negative correlation (Additional tional file  3: Figure S4a) and the λ of 1.125 indicated Qi  et al. Clin Transl Allergy (2020) 10:60 Page 5 of 13 Nasal brush Whole blood samples samples N=31 N= 66 N=6 DNA methylation in nasal brush samples (N=97) DNA methylation in whole blood samples (N=72) PersA ComR ClinR only PersA ComR ClinRonly N=44 N=19 N=34 N=44 N=10 N=18 PersAvs ComR PersAvs ClinR PersAvs ComR PersAvs ClinR 21 candidate CpG sites 4 candidate CpG sites 129 candidate CpG sites 7 candidate CpG sites Replication in EGEA Annotation and functional relevance ReplicationinEGEA and Lifelines Meta-analysis Meta-analysis 1 CpG replicated 1 CpG replicated Annotation and functional relevance Fig. 1 Study design. In the discovery panel,103 samples were assessed in this study, 72 subjects had whole blood DNA methylation data, 97 subjects had nasal DNA methylation data and 66 subjects have both. Epigenome‑ wide association studies were performed on clinical remission (ClinR) and complete remission (ComR) respectively in both whole blood and nasal samples. Significant CpG sites identified from whole blood were further replicated in two independent cohorts. Results of ComR were replicated in EGEA study (93 out of 129 probes available in EGEA); results of ClinR were replicated in EGEA study (4 out 7 probes available) and Lifelines (all 7 probes available) Fig. 2 Manhattan plot of association between clinical remission (a)/ complete remission (b) and DNA methylation in whole blood in discovery –7 cohort. In total, 436,824 CpGs were tested. The red line represents the genome‑ wide significance threshold (Bonferroni correction, P < 1.14 × 10 ). Highlighted sites represent replicate CpG sites no obvious inflation of the results. We verified the were correlated to expression level of four genes in cis CpG sites identified from whole blood samples in (nominal significant P < 0.05) in nasal brushes, with one nasal brushes, and found that one CpG (cg15404785) pair (cg07673230-RPL30 (Ribosomal Protein L30)) also associated with ClinR showed nominal significance being nominal significant and in the same direction in the PIAMA dataset (Additional file 2 : Table S8). (P < 0.05) and the same direction of effect in nasal brushes (Table  S7). Four ClinR-associated CpG sites Qi et al. Clin Transl Allergy (2020) 10:60 Page 6 of 13 –7 Table 2 Seven genome-wide significant CpG sites (P value < 1.14 × 10 ) of ClinR in relation to methylation in whole blood CpG CHR Great annotation (distance Discovery ReplicationEGEA ReplicationLifelines Metaanalysis to TSS) Coef SE P N Coef SE P N Coef SE P N P.meta Direction cg10125195 12 LACRT (+ 133) 1.32E−02 2.30E−03 9.03E−09 72 1.098 0.748 0.142 80 (69/11) 2.73E−04 3.56E−03 0.939 124 1.58E−04 + + + cg12713893 18 SNRPD1 (+ 44) 4.64E−03 8.63E−04 7.52E−08 72 7.127 5.359 0.184 120 (106/14) 4.19E−04 1.11E−03 0.705 124 2.90E−04 + + + cg15404785 14 TRMT5 (− 279,717); TMEM30B − 3.09E−02 5.17E−03 2.43E−09 72 0.423 0.150 0.005 105 (93/12) − 4.93E−03 6.11E−03 0.420 124 7.58E−02 − + − (+ 20,765) cg27252019 17 MKS1 (+ 550) − 3.86E−02 7.24E−03 1.01E−07 72 − 0.033 0.269 0.902 91 (80/11) 5.71E−03 1.01E−02 0.572 124 1.80E−02 – + cg06185924 10 ACBD5 (− 42) 1.10E−02 2.06E−03 9.36E−08 72 NA NA NA 106 (92/14) − 4.72E−03 2.33E−03 0.043 124 1.05E−01 + NA− cg13378519 1 PEX11B (+ 18) − 5.99E−03 1.07E−03 2.43E−08 72 NA NA NA 119 (105/15) − 1.08E−02 2.61E−03 3.57E05 124 2.58E−11 −NA− cg17900884 1 ICMT (− 91) 3.58E−03 6.28E−04 1.15E−08 72 NA NA NA 114 (101/13) − 1.78E−04 9.86E−04 0.857 124 9.16E−04 + NA− P.meta: the P value of meta-analysis of results from discovery, EGEA and Lifelines by using weighted Z-score method Direction: the direction of regression coefficient in discovery, EGEA and Lifelines NA: the result for this probe was not available in EGEA a –7 The CpG site was replicated with meta P value < 1.14 × 10 N (persA/ClinR), N varies from one CpG to another because of the methylC-capture sequencing method used and QC criteria in EGEA Qi  et al. Clin Transl Allergy (2020) 10:60 Page 7 of 13 Table 3 Top ten CpG sites in meta-analysis of ComR in relation to methylation in whole blood showing the same direction of effect in discovery and replication cohort CpG CHR Great annotation (distance to TSS) Discovery ReplicationEGEA Metaanalysis Coef SE P N Coef SE P N P.meta Direction cg24788483 10 HABP2 (− 401132), TCF7L2 (+ 201,644) − 1.31E−02 2.11E−03 4.62E−10 54 − 1.348 0.544 0.013 56 (53/3) 8.49E−10 – cg15223066 10 ITGB1 (− 23,896), NRP1 (+ 354472) − 8.56E−03 1.39E−03 6.65E−10 54 − 0.658 0.470 0.161 92 (89/3) 1.13E−06 – cg26909813 17 RPRML (605) − 2.68E−02 4.15E−03 1.06E−10 54 − 0.621 0.539 0.249 95 (92/3) 1.52E−06 – cg02341571 15 PATL2 (+ 150) − 1.69E−02 2.71E−03 4.39E−10 54 − 0.253 0.388 0.515 64 (62/2) 2.59E−06 – cg00269245 16 SLC12A3 (− 41116), NUP93 (+ 93986) 1.13E−02 1.86E−03 1.14E−09 54 0.832 1.109 0.453 70 (68/2) 4.62E−06 + + cg24833566 4 MTNR1A (− 152680), FAT1 (+ 15608) 1.02E−02 1.71E−03 2.38E−09 54 0.658 0.758 0.386 77 (74/3) 6.87E−06 + + cg25881850 11 KCNK4 (− 1426) − 1.23E−02 2.23E−03 3.43E−08 54 − 0.838 0.816 0.304 68 (67/1) 9.09E−06 – cg24730612 14 FOS (− 95503), TMED10 (− 6640) − 1.47E−02 2.50E−03 4.15E−09 54 − 0.260 0.489 0.595 62 (61/1) 1.09E−05 – cg23250019 11 SLC29A2 (+ 818) 2.15E−02 4.00E−03 7.62E−08 54 0.645 0.541 0.234 82 (79/3) 1.61E−05 + + P.meta: the P value of meta-analysis of results from discovery and EGEA by using weighted Z-scores method Direction: the direction of regression coefficient in discovery and EGEA A list of all 129 significant CpG sites identified in discovery cohort and the information of replication in EGEA is shown in Additional file 2: Table S9 a –7 The CpG site was replicated with meta P value < 1.14 × 10 N (persA/ComR), N varies from one CpG to another because of the methylC-capture sequencing method used and different QC criteria in EGEA Qi et al. Clin Transl Allergy (2020) 10:60 Page 8 of 13 Differential methylation in the discovery cohort of ComR compared to PersA in whole blood, but showed no obvi- We identified 129 individual CpG sites and 53 regions ous difference in nasal brushes (Additional file  3: Figure that were significantly associated with ComR in whole S5). In whole blood cis-eQTM analysis, this CpG was sig- blood DNA (Fig.  2b, Table  3, Additional file  2: Table  S9, nificantly negatively correlated to the expression level of –7 10). Q-Q plot is shown in Additional file  3: Figure S3b TCF7L2 gene (P = 9.12 × 10 ). and λ was 1.178. The eQTM analysis identified 99 CpG- To evaluate potential SNP effects within the probe for gene pairs for 45 out of the 129 CpG sites (FDR < 0.05, the two replicated CpG sites, the β value distributions Additional file  2: Table  S3). These eQTM genes were were visually assessed in the discovery cohorts (Addi- enriched in eight pathways (P < 0.01, Additional file  2: tional file  3: Figure S6), and no bimodal distribution was Table S4), and the top pathways were Activation of SMO detected indicating no underlying SNP effect within the (Smoothen), Platelet Adhesion to exposed collagen, and probe. The regional co-methylation plots of genes anno - Cilium Assembly. tated to replicated CpG sites are shown in the Additional In nasal brushes, we identified 21 CpG sites and 62 file  3: Figure S7, which showed that cg13378519 is located DMRs that were associated with ComR (Additional file  2: in the promoter region of PEX11B gene, and cg24788483 Table S11, 12). Q-Q plot is shown in Additional file  3: Fig- is located in heterochromatin region of TCF7L2 gene. ure S4b and λ was 1.318. Eight of the ComR-associated We did not identify any SNP that was associated with the CpG sites identified from whole blood also showed nomi - two CpG sites using an online database [19]. nal significance in nasal brushes with the same direction We verified if the results of the two replicated CpG of effect (Additional file  2: Table  S13). Seven ComR- sites were affected by inhaled corticosteroids (ICS), associated CpG sites were correlated to expression level known to affect DNA methylation at multiple sites in of nine genes in cis (nominal significant P < 0.05) in nasal the genome (Additional file  3: Figure S8). The boxplots brushes, and four pairs also showed nominal significant showed that the methylation levels in the remission and in the same direction in PIAMA dataset (Additional groups (both ClinR and ComR) were significantly lower file 2: Table S8). than both persistent asthma with ICS usage group and persistent asthma without ICS usage group, which indi- Replication study cated that the results were not affected by the use of ICS. We replicated the top findings in whole blood in two We evaluated the impact of rhinitis on our results, as independent cohorts. In replication cohort EGEA, four rhinitis shows comorbidity with asthma. The replicated ClinR-associated CpG sites and 93 ComR-associated ClinR-associated CpG site showed similar effect size and CpG sites were available, with one CpG site for ClinR and P value before and after adjusting for rhinitis, while the one for ComR showing nominal significance (P < 0.05) ComR-associated CpG showed similar effect size yet the and the same direction as the discovery cohort. In the P value changed from 4.62E-10 to 4.07E-05 after correc- replication cohort Lifelines, which only had the ClinR tion for rhinitis (Additional file  2: Table S14). Additional phenotype, seven ClinR-associated CpG sites were all adjustment for BMI did not change the results of the available with one CpG with P < 0.05 and the same direc- two replicated CpG sites, indicating that BMI was not a tion. After meta-analysis, one CpG associated with ClinR confounder (Additional file  2: Table S14). The significant and another one CpG associated with ComR were rep- CpG sites identified from nasal samples showed similar –7 licated (meta P < 1.14 × 10 ). The ClinR-associated CpG effect size and P value before and after adjusting for BMI (cg13378519) was located on Chr 1 and close to gene and for rhinitis (Additional file 2: Table S15). PEX11B (Peroxisomal Biogenesis Factor 11 Beta). The We also compared our results from the discovery CpG site was not available in the EGEA study and was cohort with the published DNA methylation study of only replicated in Lifelines. This CpG was lower meth - asthma remission in bronchial biopsies (Vermeulen et al. ylated in ClinR subjects compared to PersA (Fig.  3a). A [10]). One CpG (cg13525448), located close to LBX1 similar trend was also found in nasal brushes although (Ladybird Homeobox  1) gene and TLX1 (T Cell Leuke- the association was not statistically significant (Fig.  3b). mia Homeobox  1) gene, in their top list of remission vs The methylation levels at this CpG also differed between persistent asthma reached nominal significance (P < 0.05) ClinR and healthy subjects in Lifelines cohort (Fig.  3c), and showed the same direction of effect in our results of indicating that DNA methylation status of asthma remis- ComR in whole blood (Additional file  2: Table S16). Two sion subjects is not equal to that of healthy subjects. genes of their DMRs associated with ClinR, at the gene The ComR-associated CpG (cg24788483) was located PTCHD3 (patched domain-containing protein 3) and on chr10 and close to gene HABP2 (Hyaluronan Bind- LOC100507389, were also present in our results of DMRs ing Protein 2) and TCF7L2 (Transcription Factor 7 Like associated with ClinR in nasal brushes (Additional file  2: 2). This CpG was lower methylated in ComR subjects Table S6). Qi  et al. Clin Transl Allergy (2020) 10:60 Page 9 of 13 (a) cg13378519 in blood (discovery) (b) cg13378519 in nasal brush (discovery) (c) cg13378519 in blood (Lifelines) 0.16 0.16 0.16 0.12 0.12 0.12 0.08 0.08 0.08 0.04 0.04 0.04 PerAsthma ClinR ComR PerAsthmaClinR ComR Asthma ClinR Healthy Fig. 3 Boxplot illustrating DNA methylation levels of cg13378519 in persistent asthma, clinical remission and complete remission subjects. a DNA methylation levels in whole blood in discovery cohort; b DNA methylation levels in nasal brushes in discovery cohort; c DNA methylation levels in whole blood in replication cohort (Lifelines) the pancreatic insulin secretory response to incretins Discussion [25, 26]. Previously, the risk of asthma was reported to In this study, we identified one CpG site associated with be higher in type 2 diabetes patients [27] and it has been ClinR and a different CpG site associated with ComR suggested that treatment targeting insulin resistance may from whole blood, that were replicated in independent have a positive effect on asthma patients [28]. However, cohorts. We also revealed that 25 CpG sites were associ- it has never been studied whether this treatment could ated with asthma remission phenotypes in nasal brushes. contribute to asthma remission. In addition, we could In the nasal dataset, two differentially methylated regions not infer from the current results of association analysis were previously observed in another DNA methylation whether the DNA methylation signal we identified is the study of asthma remission using bronchial biopsies. This cause or consequence of asthma remission. is the first epigenome-wide association study of asthma Although remission subjects do not have asthma remission in both whole blood and nasal epithelial cells. symptoms any more, they still may be different from This study is important since understanding asthma healthy people. Airway abnormalities such as basement remission may provide new leads for future asthma treat- membrane thickening still exist in clinical and complete ments, and epigenetic studies may help to reveal these remission subjects [29]. Vermeulen et  al. [10] reported cellular mechanisms. that the DNA methylation profile in remission subjects is The two replicated CpG sites were annotated to genes different from that in healthy subjects. Our data in whole which were not known to be involved in asthma (remis- blood for our significantly replicated CpG confirmed this sion) before. ClinR-associated CpG cg13378519 was observation. The methylation levels of the ClinR associ - located in the promoter region of PEX11B gene. PEX11B ated CpG (cg13378519) were different between ClinR encodes a protein of the PEX11 family [22]. Overex- and healthy subjects, which indicated that the methyla- pression of the PEX11B gene in human cells induces tion status of subjects with asthma remission may not peroxisome proliferation [23, 24]. Our results could simply return to that of healthy people. indicate a possible role of peroxisomal proliferation as DNA methylation is cell-type specific. Both of the stud - being involved in asthma remission. ComR-associated ied cells or tissues are highly heterogeneous and this may CpG cg24788483 was located in the TCF7L2 gene and confound the association between DNA methylation also correlated with the expression level of this gene. and disease. Thus, we applied the widely used House - The TCF7L2 gene product is a transcription factor man’s cell type correction method in blood samples, and that plays a role in the Wnt signaling pathway. SNPs in for nasal brushes, we used SVA which showed good per- TCF7L2 were previously reported to be strongly asso- formance in cell type correction when reference data is ciated with type-2 diabetes and this gene plays a role in Beta value Beta value Beta value Qi et al. Clin Transl Allergy (2020) 10:60 Page 10 of 13 lacking [30]. DNA methylation is also tissue specific. during the airway remodeling of persistent asthma [37]. It is yet unknown if mechanisms of remission include Blood is considered as an easily accessible surrogate tis- local, tissue specific airway (epithelial) effects, immune sue for asthma study, and previous studies showed some cell effects, or both. Matched data from whole blood and shared gene expression pattern and pathways between nasal brush samples enable us to compare DNA meth- blood and airway/ lung [38, 39]. Our results suggest that ylation profiles associated with asthma remission in two genes related to asthma remission associated CpG sites different tissues, as proxies of both immunological and may reflect a role for airway epithelial barrier and airway airway specific mechanisms. Although we identified sev - remodeling in asthma remission. eral interesting CpG sites related to remission in nasal Corticosteroids treatment is usually used to control epithelium, our results of nasal brushes still need further inflammation and improve control of asthma. One pre - replication. In previous studies, several DNA methylation vious study identified differences in DNA methylation sites were significantly associated with asthma in both of CpG sites in blood associated with systemic corticos- whole blood and nasal cells [9, 31]. In our study, although teroid treatment in patients with COPD, suggesting the no significant enriched cross-tissue effect of DNA meth - potential effect of corticosteroids on the DNA meth - ylation in asthma remission was shown, we did identify ylation profile [40]. In our study, we therefore assessed nine CpG sites in blood DNA that could be replicated in the effect of ICS on the methylation levels of the two DNA of nasal brushed cells. This indicated that there may replicated CpG sites and found that the results were be shared DNA methylation signals in blood and nasal not affected by ICS usage. However, the effect of corti - cells. We interpret these findings as either indicative of costeroid treatment on DNA methylation in nasal and the presence of cross-tissue epigenetic mechanisms in blood samples of patients with asthma and remission is blood and epithelial cells, or the possibility that the DNA largely unknown and should be further investigated in methylation signal in nasal brushed cells is partly driven the future. by immune cells in the nose [20], that are also present in Rhinitis and asthma often co-exist and show shared blood. genetic origins [41]. A previous study of our group also DNA methylation might be related to the regulation of showed a shared DNA methylation signal of asthma and gene expression, and eQTM analysis may help to under- rhinitis, suggesting the importance of considering the stand the function of CpG sites. Among the eQTM genes presence of rhinitis when doing EWAS of asthma in nasal that correlated with ClinR-associated CpG sites in whole epithelium [20]. In this study, we showed that additional blood, the protein encoded by PRKCH (Protein Kinase adjustment for rhinitis changed the results on one of the C Eta) plays a key role in epithelial tight junction regu- replicated CpG sites (cg24788483) identified from blood. lation which is important in maintaining the integrity This remission associated CpG site was associated with and function of the airway epithelial barrier [32]; PDE1B active rhinitis in the absence of asthma, suggesting that (Phosphodiesterase 1B) encoded a protein belonging blood DNA methylation profiles of persistent asthma and to phosphodiesterases (PDEs) family, and various PDE rhinitis overlap. This indicates that we should control for inhibitors showed anti-inflammatory, anti-remodelling allergic rhinitis in future studies of asthma remission. and bronchodilator effect and are potential treatment of When comparing our results with those of Vermeulen asthma [33]. In whole blood, 99 eQTM genes that were et  al. [10] in bronchial biopsies, one CpG (cg13525448) associated with ComR were enriched in eight path- in their top list of remission vs persistent asthma reached ways, among which three pathways were related to air- nominal significance (P < 0.05) in our data and showed way epithelial function: activation of SMO (Smoothen), the same direction in our results of ComR in whole blood cilium assembly and focal adhesion. The activation of (Additional file  2: Table  S16). This CpG site is anno - SMO activity in bronchial epithelia enhanced the aller- tated to LBX1 gene and TLX1 gene by position. Besides, gen-induced goblet cell metaplasia, which is defined among their DMRs associated with ClinR, regions at the as a reversible transformation of airway epithelial cells gene PTCHD3 and LOC100507389 also showed up in our to mucous cells such as goblet cells and may occur in results of ClinR in nasal brushes. SNPs in PTCHD3 gene asthma [34]. Airway cilia are important for clearance were previously associated with asthma in African Amer- of inhaled particles and pathogens. One study showed ican children [42]. asthma patients had less ciliated cells in airway samples There are strengths and limitations in this study. than healthy people [35], and changes in ciliary function To the best of our knowledge, this is the only cohort may be relevant for the development of asthma in chil- worldwide to study asthma remission in nasal brushes, dren [36]. Focal adhesion between the cell membrane which enable us to compare the DNA methylation pro- and matrix are essential elements of airway smooth mus- file in relation to asthma remission in whole blood and cle cells migration which may play an important role nasal epithelium. However, because of the uniqueness Qi  et al. Clin Transl Allergy (2020) 10:60 Page 11 of 13 of this data, we could not replicate our results in nasal and after adjusting for BMI and rhinitis. Table S15. Summary statistics of brushes in another independent cohort. Notably, we had significant CpGs (identified from nasal samples) before and after adjusting for BMI and rhinitis. Table S16. Look up of CpGs associated with ClinR matched DNA methylation and gene expression data in identified by Vermeulen et al. in this study. nasal brushes, which help us to get a better understand- Additional file 3: Figure S1. Lung function of different groups at baseline ing of the function of CpG sites associated with asthma and last visit. In this figure, t ‑test was used in comparing means of any two remission in nasal brushes. Regarding the limitations groups (ns: P>0.05, *: P<0.05, **: P<0.01, ***: P<0.001, ****: P<0.0001). Fig‑ ure S2. Estimated cell proportions among different groups. In this figure, of this study, firstly, this study was performed in a rela - t‑test was used in comparing means of any two groups (ns: P>0.05). Fig ‑ tively small sample with limited number of ComR cases, ure S3. Quantile–quantile plot for epigenome‑ wide meta‑analysis of the and the results were replicated in EGEA which is also association between asthma remission and blood DNA methylation (n = 72). (a) ClinR, (b) ComR. Figure S4. Quantile–quantile plot for epigenome‑ with few ComR cases, which lead to low power of rep- wide meta‑analysis of the association between asthma remission and lication of ComR results. Secondly, although we studied nasal DNA methylation (n = 97). (a) ClinR, (b) ComR. Figure S5. Boxplot over 450,000 DNA methylation sites, this only repre- illustrating DNA methylation levels of cg24788483 in persistent asthma, clinical remission and complete remission subjects in discovery cohort. (a) sents 1.6% of the human DNA methylome. Thirdly, we DNA methylation levels in blood, (b) DNA methylation in nasal brushes. studied mixed contributions of cells in whole blood and Figure S6. Density distributions of DNA methylation levels of two nasal brushes; and although we corrected for cell types replicated CpGs in 72 blood samples from discovery cohorts. Figure S7. Regional association plot of two replicated CpGs. For each plot from top using established methods, potentially stronger results to bottom the tracks included are: 1) Log10(P values) from the discovery may be anticipated when studying pure cell types, as was 4 years model with CpGs indicated by dots. 2) Annotation tracks for the previously shown for DNA methylation sites associated plotted genomic region taken from UCSC Genome Browser. 3) Pairwise correlation matrix across the displayed CpGs. Figure S8. DNA methylation with asthma in purified blood eosinophils [8]. Finally, levels of two replicated CpGs in remission subjects and asthma patients we used a cross sectional analysis at follow up to inves- stratified by ICS usage. In this figure, t ‑test was used in comparing means tigate asthma remission. Future studies, with prospective of any two groups (*: P<0.05, ***: P<0.001, ****: P<0.0001). designs, should be performed to investigate if these DNA methylation sites can predict future asthma remission. Acknowledgements In conclusion, we identified replicable DNA methyla - We thank the participants in this study and all staff involved in the study through the years. tion signals associated with clinical and complete asthma remission, which may play a role in peroxisome prolifera- Authors’ contributions tion and Wnt signaling pathway. This could help in iden - CQ, JMV, HMB, CJX and GHK contributed to the design of the study. CQ, CJX, DA van der P and DA contributed to data analysis. JMV, AMEN, FND and tifying the underlying mechanisms of asthma remission VS contributed to study procedures, participant recruitment or laboratory and also of the chronicity of the disease. procedures. All authors contributed to results interpretation. CQ, CJX and GHK contributed to the initial manuscript draft. All authors revised the manuscript draft for important intellectual content. All authors read and approved the Supplementary information final manuscript. Supplementary information accompanies this paper at https ://doi. org/10.1186/s1360 1‑020‑00365 ‑4. Funding The Biobank‑Based Integrative Omics Studies (BIOS) Consortium is funded by BBMRI‑NL, a research infrastructure financed by the Dutch government (NWO Additional file 1. Supplementary materials. 184.021.007). Additional file 2: Table S1. Characteristics of subjects from two groups. Table S2. Differentially methylated regions associated with ClinR in Availability of data whole blood identified by both comb ‑p (Sidak p ‑ value) and DMRcate DNA methylation data has been deposited at the European Genome‑phe ‑ (FDR) methods. Table S3 Correlation analysis of DNA methylation and nome Archive (EGA), which is hosted by the EBI and the CRG, under accession gene expression levels (in cis, +/‑ 250 kb) for asthma remission in blood. number EGAS00001004766. Table S4. Pathway analysis of eQTM genes correlated with ClinR and ComR in blood. Table S5. Genomewide significant (pvalue < 1.14E‑07) Ethics approval and consent to participate CpGs associated with ClinR in nasal brushes. Table S6. Differentially This study was approved by the medical ethical board of University Medical methylated regions associated with ClinR in nasal brushes identified by Center Groningen, and all participants gave written informed consent. both comb‑p (Sidak p ‑ value) and DMRcate (FDR) methods. Table S7. Look up of ClinR‑assocaited CpGs (indentified from blood) in the results of Competing interests nasal brushes. Table S8. Correlation analysis of DNA methylation and gene GHK reports grants from Lung Foundation of the Netherlands, TEVA the expression levels (in cis, +/‑ 250 kb) for asthma remission in nasal brushes. Netherlands, Vertex, GSK, Ubbo Emmius Foundation, TETRI foundation, outside Table S9. 129 genome‑ wide significant (pvalue < 1.14E‑07) ComR‑asso ‑ the submitted work. The rest of the authors declare that they have no relevant ciated CpGs in whole blood in discovery cohort. Table S10. Differentially conflict of interests. methylated regions associated with ComR in whole blood identified by both comb‑p (Sidak p ‑ value) and DMRcate (FDR) methods. Table S11. Author details Genome‑ wide significant (p value < 1.14E‑07) CpGs associated with ComR Department of Pediatric Pulmonology and Pediatric Allergy, Beatrix Children’s in nasal brushes. Table S12. Differentially methylated regions associated Hospital, University Medical Center Groningen, University of Groningen, PO with ComR in nasal brushes identified by both comb ‑p (Sidak p ‑ value) and Box 30.001, 9700 RB Groningen, the Netherlands. GRIAC Research Institute, DMRcate (FDR) methods. Table S13. Look up of ComR‑associated CpG University Medical Center Groningen, University of Groningen, Groningen, sites (indentified from blood) in the results of nasal brushes. Table S14. The Netherlands. Department of Epidemiology, University Medical Center Summary statistics of replicated CpGs (identified from blood) before Groningen, University of Groningen, Groningen, The Netherlands. University Qi et al. Clin Transl Allergy (2020) 10:60 Page 12 of 13 Grenoble Alpes, Inserm, CNRS, Team of environmental epidemiology Bioinformatics. 2012;28(22):2986–8. https ://doi.org/10.1093/bioin forma applied to Reproduction and Respiratory health, IAB (Institute for Advanced tics/bts54 5. Biosciences), 38000 Grenoble, France. Research Group of Bioinformatics 17. Peters TJ, Buckley MJ, Statham AL, et al. De novo identification of differen‑ and Computational Genomics, CiiM, Centre for individualized infection medi‑ tially methylated regions in the human genome. Epigenetics Chromatin. cine, A Joint Venture Between Hannover Medical School and the Helmholtz 2015;8:6. https ://doi.org/10.1186/1756‑8935‑8‑6. Centre for Infection Research, Hannover, Germany. Department of Gastroen‑ 18. McLean CY, Bristor D, Hiller M, et al. GREAT improves functional interpreta‑ terology, Hepatology and Endocrinology, TWINCORE, Centre for Experimental tion of cis‑regulatory regions. Nat Biotechnol. 2010;28(5):495–501. https :// and Clinical Infection Research, A Joint Venture Between the Hannover doi.org/10.1038/nbt.1630. Medical School and the Helmholtz Centre for Infection Research, Hannover, 19. the BIOS Consortium, Bonder MJ, Luijk R, et al. Disease variants alter tran‑ Germany. Department of Internal Medicine, Radboud University Medical scription factor levels and methylation of their binding sites. Nat Genet. Center, Nijmegen, the Netherlands. 2017;49(1):131–8. https ://doi.org/10.1038/ng.3721. 20. Qi C, Jiang Y, Yang IV, et al. Nasal DNA methylation profiling of asthma Received: 4 August 2020 Accepted: 26 November 2020 and rhinitis. J Allergy Clin Immunol. 2020. https ://doi.org/10.1016/j. jaci.2019.12.911. 21. Herwig R, Hardt C, Lienhard M, Kamburov A. Analyzing and interpreting genome data at the network level with ConsensusPathDB. Nat Protoc. 2016;11(10):1889–907. https ://doi.org/10.1038/nprot .2016.117. 22. Koch J, Pranjic K, Huber A, et al. PEX11 family members are membrane References elongation factors that coordinate peroxisome proliferation and 1. Upham JW, James AL. Remission of asthma: the next therapeutic frontier? maintenance. J Cell Sci. 2010;123(19):3389–400. https ://doi.org/10.1242/ Pharmacol Ther. 2011;130(1):38–45. https ://doi.org/10.1016/j.pharm thera jcs.06490 7. .2011.01.002. 23. Schrader M, Reuber BE, Morrell JC, et al. Expression of PEX11beta medi‑ 2. Carpaij OA, Burgess JK, Kerstjens HAM, Nawijn MC, van den Berge M. A ates peroxisome proliferation in the absence of extracellular stimuli. J Biol review on the pathophysiology of asthma remission. Pharmacol Ther. Chem. 1998;273(45):29607–14. https ://doi.org/10.1074/jbc.273.45.29607 . 2019;201:8–24. https ://doi.org/10.1016/j.pharm thera .2019.05.002. 24. Ebberink MS, Koster J, Visser G, et al. A novel defect of peroxisome divi‑ 3. Vonk JM, Postma DS, Boezen HM, et al. Childhood factors associated with sion due to a homozygous non‑sense mutation in the PEX11β gene. J asthma remission after 30 year follow up. Thorax. 2004;59(11):925–9. https Med Genet. 2012;49(5):307–13. https ://doi.org/10.1136/jmedg enet‑2012‑ ://doi.org/10.1136/thx.2003.01624 6. 10077 8. 4. Carpaij OA, Nieuwenhuis MAE, Koppelman GH, van den Berge M, Postma 25. Grant SFA, Thorleifsson G, Reynisdottir I, et al. Variant of transcription DS, Vonk JM. Childhood factors associated with complete and clinical factor 7‑like 2 ( TCF7L2) gene confers risk of type 2 diabetes. Nat Genet. asthma remission at 25 and 49 years. Eur Respir J. 2017;49(6):1601974. 2006;38(3):320–3. https ://doi.org/10.1038/ng173 2. https ://doi.org/10.1183/13993 003.01974 ‑2016. 26. Pearson ER. Translating TCF7L2: from gene to function. Diabetologia. 5. Vonk JM, Nieuwenhuis MAE, Dijk FN, et al. Novel genes and insights in 2009;52(7):1227–30. https ://doi.org/10.1007/s0012 5‑009‑1356‑1. complete asthma remission: a genome‑ wide association study on clinical 27. Thomsen SF, Duffy DL, Kyvik KO, Skytthe A, Backer V. Risk of asthma in and complete asthma remission. Clin Exp Allergy. 2018;48(10):1286–96. adult twins with type 2 diabetes and increased body mass index: type 2 https ://doi.org/10.1111/cea.13181 . diabetes, obesity and asthma in twins. Allergy. 2011;66(4):562–8. https :// 6. Qi C, Xu C‑ J, Koppelman GH. The role of epigenetics in the development doi.org/10.1111/j.1398‑9995.2010.02504 .x. of childhood asthma. Expert Rev Clin Immunol. 2019;15(12):1287–302. 28. Carpaij OA, van den Berge M. The asthma–obesity relationship: underly‑ https ://doi.org/10.1080/17446 66X.2020.16869 77. ing mechanisms and treatment implications. Curr Opin Pulmon Med. 7. Gibney ER, Nolan CM. Epigenetics and gene expression. Heredity. 2018;24(1):42–9. https ://doi.org/10.1097/MCP.00000 00000 00044 6. 2010;105(1):4–13. https ://doi.org/10.1038/hdy.2010.54. 29. Broekema M, Timens W, Vonk JM, et al. Persisting remodeling and less 8. Xu C‑ J, Söderhäll C, Bustamante M, et al. DNA methylation in child‑ airway wall eosinophil activation in complete remission of asthma. hood asthma: an epigenome‑ wide meta‑analysis. Lancet Respir Med. Am J Respir Crit Care Med. 2011;183(3):310–6. https ://doi.org/10.1164/ 2018;6(5):379–88. https ://doi.org/10.1016/S2213 ‑2600(18)30052 ‑3. rccm.20100 3‑0494O C. 9. Forno E, Wang T, Qi C, et al. DNA methylation in nasal epithelium, atopy, 30. Kaushal A, Zhang H, Karmaus WJJ, et al. Comparison of different cell type and atopic asthma in children: a genome‑ wide study. Lancet Respir Med. correction methods for genome‑scale epigenetics studies. BMC Bioinfor ‑ 2019;7(4):336–46. https ://doi.org/10.1016/S2213 ‑2600(18)30466 ‑1. mat. 2017;18(1):216. https ://doi.org/10.1186/s1285 9‑017‑1611‑2. 10. Vermeulen CJ, Xu C‑ J, Vonk JM, et al. Differential DNA methylation in 31. Cardenas A, Sordillo JE, Rifas‑Shiman SL, et al. The nasal methylome as a bronchial biopsies between persistent asthma and asthma in remission. biomarker of asthma and airway inflammation in children. Nat Commun. Eur Respir J. 2019. https ://doi.org/10.1183/13993 003.01280 ‑2019. 2019;10(1):3095. https ://doi.org/10.1038/s4146 7‑019‑11058 ‑3. 11. Nieuwenhuis MA, Siedlinski M, van den Berge M, et al. Combining 32. Suzuki T, Elias BC, Seth A, et al. PKC eta regulates occludin phospho‑ genomewide association study and lung eQTL analysis provides evi‑ rylation and epithelial tight junction integrity. Proc Natl Acad Sci USA. dence for novel genes associated with asthma. Allergy. 2016;71(12):1712– 2009;106(1):61–6. https ://doi.org/10.1073/pnas.08027 41106 . 20. https ://doi.org/10.1111/all.12990 . 33. Zuo H, Cattani‑ Cavalieri I, Musheshe N, Nikolaev VO, Schmidt M. Phos‑ 12. Allum F, Shao X, Guénard F, et al. Characterization of functional methyl‑ phodiesterases as therapeutic targets for respiratory diseases. Pharmacol omes by next‑ generation capture sequencing identifies novel disease ‑ Ther. 2019;197:225–42. https ://doi.org/10.1016/j.pharm thera .2019.02.002. associated variants. Nat Commun. 2015;6:7211. https ://doi.org/10.1038/ 34. Xu C, Zou C, Hussain M, et al. High expression of Sonic hedgehog in aller‑ ncomm s8211 . gic airway epithelia contributes to goblet cell metaplasia. Mucosal Immu‑ 13. Houseman E, Accomando WP, Koestler DC, et al. DNA methylation arrays nol. 2018;11(5):1306–15. https ://doi.org/10.1038/s4138 5‑018‑0033‑4. as surrogate measures of cell mixture distribution. BMC Bioinformat. 35. Vieira Braga FA, Kar G, Berg M, et al. A cellular census of human lungs 2012;13(1):86. https ://doi.org/10.1186/1471‑2105‑13‑86. identifies novel cell states in health and in asthma. Nat Med. 2019. https 14. Aryee MJ, Jaffe AE, Corrada‑Bravo H, et al. Minfi: a flexible and com‑ ://doi.org/10.1038/s4159 1‑019‑0468‑5. prehensive Bioconductor package for the analysis of Infinium DNA 36. Dizier M‑H, Nadif R, Margaritte ‑ Jeannin P, et al. Interaction between the methylation microarrays. Bioinformatics. 2014;30(10):1363–9. https ://doi. DNAH9 gene and early smoke exposure in bronchial hyperresponsive‑ org/10.1093/bioin forma tics/btu04 9. ness. Eur Respir J. 2016;47(4):1072–81. https ://doi.org/10.1183/13993 15. Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The sva package 003.00849 ‑2015. for removing batch effects and other unwanted variation in high‑ 37. Gerthoffer WT. Migration of Airway Smooth Muscle Cells. Proceed Am throughput experiments. Bioinformatics. 2012;28(6):882–3. https ://doi. Thoracic Soc. 2008;5(1):97–105. https ://doi.org/10.1513/pats.20070 org/10.1093/bioin forma tics/bts03 4. 4‑051VS . 16. Pedersen BS, Schwartz DA, Yang IV, Kechris KJ. Comb‑p: software for com‑ bining, analyzing, grouping and correcting spatially correlated P‑ values. Qi  et al. Clin Transl Allergy (2020) 10:60 Page 13 of 13 38. Morrow JD, Chase RP, Parker MM, et al. RNA‑sequencing across three 42. White MJ, Risse‑Adams O, Goddard P, et al. Novel genetic risk factors for matched tissues reveals shared and tissue‑specific gene expression asthma in African American children: precision medicine and the SAGE II and pathway signatures of COPD. Respir Res. 2019;20(1):65. https ://doi. Study. Immunogenetics. 2016;68(6–7):391–400. https ://doi.org/10.1007/ org/10.1186/s1293 1‑019‑1032‑z.s0025 1‑016‑0914‑1. 39. Halloran JW, Zhu D, Qian DC, et al. Prediction of the gene expression in normal lung tissue by the gene expression in blood. BMC Med Genomics. Publisher’s Note 2015;8:77. https ://doi.org/10.1186/s1292 0‑015‑0152‑7. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 40. Wan ES, Qiu W, Baccarelli A, et al. Systemic steroid exposure is associ‑ lished maps and institutional affiliations. ated with differential methylation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2012;186(12):1248–55. https ://doi. org/10.1164/rccm.20120 7‑1280O C. 41. Ferreira MA, Vonk JM, Baurecht H, et al. Shared genetic origin of asthma, hay fever and eczema elucidates allergic disease biology. Nat Genet. 2017;49(12):1752–7. https ://doi.org/10.1038/ng.3985. Ready to submit your research ? Choose BMC and benefit from: fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clinical and Translational Allergy Springer Journals

Epigenome-wide association study identifies DNA methylation markers for asthma remission in whole blood and nasal epithelium

Loading next page...
 
/lp/springer-journals/epigenome-wide-association-study-identifies-dna-methylation-markers-MmBLoBiM0k

References (45)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
eISSN
2045-7022
DOI
10.1186/s13601-020-00365-4
Publisher site
See Article on Publisher Site

Abstract

Background: Asthma is a chronic respiratory disease which is not curable, yet some patients experience spontane‑ ous remission. We hypothesized that epigenetic mechanisms may be involved in asthma remission. Methods: Clinical remission (ClinR) was defined as the absence of asthma symptoms and medication for at least 12 months, and complete remission (ComR) was defined as ClinR with normal lung function and absence of airway hyperresponsiveness. We analyzed differential DNA methylation of ClinR and ComR comparing to persistent asthma (PersA) in whole blood samples (n = 72) and nasal brushing samples (n = 97) in a longitudinal cohort of well char‑ acterized asthma patients. Significant findings of whole blood DNA methylation were tested for replication in two independent cohorts, Lifelines and Epidemiological study on the Genetics and Environment of Asthma (EGEA). Results: We identified differentially methylated CpG sites associated with ClinR (7 CpG sites) and ComR (129 CpG sites) in whole blood. One CpG (cg13378519, Chr1) associated with ClinR was replicated and annotated to PEX11 (Per‑ oxisomal Biogenesis Factor 11 Beta). The whole blood DNA methylation levels of this CpG were also different between ClinR and healthy subjects. One ComR‑associated CpG (cg24788483, Chr10) that annotated to TCF7L2 ( Transcription Factor 7 Like 2) was replicated and associated with expression of TCF7L2 gene. One out of seven ClinR‑associated CpG sites and 8 out of 129 ComR‑associated CpG sites identified from whole blood samples showed nominal significance (P < 0.05) and the same direction of effect in nasal brushes. Conclusion: We identified DNA methylation markers possibly associated with clinical and complete asthma remis‑ sion in nasal brushes and whole blood, and two CpG sites identified from whole blood can be replicated in independ‑ ent cohorts and may play a role in peroxisome proliferation and Wnt signaling pathway. Keywords: Asthma remission, DNA methylation, Whole blood, Nasal brushes Introduction to now. However, some asthma patients may grow out of Asthma is a chronic airway disorder that affects more this disease, which is called asthma remission. than 300 million people around the world. Asthma can- Asthma remission is more common in patients with not be completely cured with the available treatment up childhood onset asthma, and the proportions of remis- sion differ in different age groups at follow-up (33–53% in adolescence, 6–33% in young adulthood and 11–52% in adulthood) [1, 2]. There are two types of asthma remis - *Correspondence: g.h.koppelman@umcg.nl Department of Pediatric Pulmonology and Pediatric Allergy, Beatrix sion, one is defined by absence of asthma symptoms and Children’s Hospital, University Medical Center Groningen, University medication for at least one year, which is called “clinical of Groningen, PO Box 30.001, 9700 RB Groningen, the Netherlands remission” (ClinR). Some ClinR subjects may still have Full list of author information is available at the end of the article © The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/ zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Qi et al. Clin Transl Allergy (2020) 10:60 Page 2 of 13 airway hyperresponsiveness (AHR) or a low lung func- in two independent cohorts, and also verified the top tion. Therefore, “complete remission” (ComR) was put results from whole blood DNA in cells obtained by nasal forward [3]. In addition to criteria of ClinR, ComR has brushing. to meet additional criteria of normal lung function and absence of AHR. This is a more rare phenomenon, that Methods takes place in 5–22% of asthmatics [2]. A full description of methods is provided in the online Asthma remission is associated with both genetic and supplement. environmental factors. Environmental factors includ- ing breast feeding and having pets in childhood were Study populations reported to be positively associated with asthma remis- Subjects included in this study were from long-term sion [4]. A recent genome-wide association study follow-up studies in the University Medical Center Gro- (GWAS) identified one SNP (single nucleotide polymor - ningen (UMCG). This study consists of two sources of phism) that was associated with ClinR and three SNPs subjects, (1) subjects from the third visit (2013–2014) of with ComR. One of these SNPs was also associated with a longitudinal study described previously by Carpaij et al expression of known asthma genes including ILRL1 and [4], and (2) additional asthma remission subjects that IL13 in lung tissue [5]. were followed up from previous genetic studies [11] and Epigenetic mechanisms such as DNA methylation may were re-invited in 2013–2014 to take nasal brushes and help to build a link between genetic factors and the envi- blood samples, during which the remission status was ronment. DNA methylation can be regulated by SNPs, assessed again. The participants included in this cohort and can reflect environmental exposures, ageing, cell type were different from those of the cohort described by Ver - constitution and activation [6]. DNA methylation refers meulen et  al. [10] from our research institute. All par- to the addition of a methyl-group to cytosine, which usu- ticipants were diagnosed with asthma at baseline with ally happens when a cytosine is located next to a guanine a doctor diagnosis of asthma and AHR. Then, subjects in the 5′ to 3′ direction (CpG site), and this may relate to had at least one follow-up medical examination dur- the regulation of gene expression [7]. Epigenome-wide ing adulthood in which their asthma status was evalu- association studies (EWAS) have provided insights into ated by questionnaires and in most subjects additionally the development of asthma and its remission. CpG sites with spirometry and an AHR test [5]. The medical ethical associated with asthma in both blood and nasal epithe- board of the UMCG approved the studies and all partici- lial cells have been identified, such as CpG sites located in pants gave written informed consent. DICER1, STX3, and LIPIN1 in blood cells, and CDHR3, We replicated findings in whole blood in two independ - FBXL7 and NTRK1 in nasal epithelial cells [8, 9]. Ver- ent cohorts: the Lifelines population-based cohort in The meulen et  al. [10] identified 4 CpG sites and 42 regions Netherlands (where we could replicate our results on that were differentially methylated between remission ClinR using Illumina 450 K array) and the Epidemiologi- and persistent asthma (PersA) in bronchial biopsies, and cal study on the Genetics and Environment of Asthma top CpG sites were annotated to genes including ACKR2 (EGEA) cohort in France, a case control and family and DGKQ by gene expression. study on asthma, where we could replicate our results on Although this latter paper provided a proof of con- ClinR and ComR with methylC-capture sequencing [12]. cept of the relation between asthma remission and DNA No cohort was available to replicate our nasal findings. methylation, bronchial biopsies are not easy to obtain Extensive information on these cohorts is described in for further studies. DNA methylation in whole blood the Additional file 1. and nasal epithelium is a good proxy for bronchial epi- thelium and can therefore also help to understand the Phenotype definition mechanism of asthma remission. Here, we hypothesize The presence of PersA, ClinR and ComR was determined that epigenetic mechanisms may be involved in asthma at the most recent visit. ClinR was defined according to remission, reflected in different DNA methylation pat - the following criteria: (1) No use of any asthma medi- terns in whole blood and nasal epithelium. To test this cation, and (2) no symptoms (asthma attacks and/ or hypothesis, we performed an EWAS of whole blood and wheezing) in the past year. ComR was defined as ClinR nasal DNA in subjects with PersA, ClinR and ComR. combined with (3) no AHR (PC (provocative concen- We investigated a longitudinal cohort in which asthma tration causing a 20% fall in FEV (forced expiratory −1 was initially carefully defined and the remission status volume in 1 s)) methacholine ⩽ 39.3 mg mL )), and (4) was assessed during follow-up (median 39  years). We FEV % predicted pre-bronchodilator > 80%. PersA was subsequently replicated the whole blood DNA results defined as the presence of asthma symptom and/or the Qi  et al. Clin Transl Allergy (2020) 10:60 Page 3 of 13 use of asthma medication. Detailed phenotype defini - of ± 250  kb) by cis expression quantitative trait DNA tions of the replication cohorts are described in the Addi- methylation (cis-eQTM) analysis. Blood eQTM was tional file 1. assessed in the BIOS consortium dataset [19]. Matched nasal RNA sequencing data and DNA methylation data DNA methylation measurements and statistical analyses was used to perform eQTM analysis for significant nasal DNA was extracted from 72 whole blood and 103 nasal CpG sites and these results were additionally replicated brushing samples taken at the most recent visit. Genome- in the larger collection of nasal brushes obtained in the wide DNA methylation was determined using Illumina PIAMA study [20]. The genes that were identified in Infinium HumanMethylation450 BeadChips. After qual - eQTM analysis were used for pathway analysis by Con- ity control, 72 whole blood samples and 97 nasal epi- sensusPathDB [21]. To check if the results of replicated thelium samples (of in total 103 unique subjects) with CpG sites were affected by inhaled corticosteroids (ICS), 436,824 CpG sites remained for following steps. we stratified asthma subjects by ICS usage, and compared We used robust linear regression to determine the the DNA-methylation in the remission group to asthma differential methylation between persistent asthma with ICS and asthma without ICS group, respectively, by and asthma remission: (1) PersA versus ClinR, and (2) t-test. To evaluate the impact of BMI and allergic rhinitis PersA versus ComR, in whole blood and nasal brush- on the results, we performed sensitivity analyses on the ing samples, with adjustment for covariates that are replicated blood CpG sites by additionally adjusting for known to affect DNA methylation (age, sex, smoking allergic rhinitis and BMI respectively. We also performed status, pack years, and batch). For whole blood samples, sensitivity analysis for all significant nasal CpG sites. we performed adjustment on the percentage of mono- cytes, B cells, NK cells, CD4 + T cells, CD8 + T cells, Results and granulocytes, which were predicted by the House- Subject characteristics man [13] algorithm using minfi [14] package. For nasal Subject characteristics of the discovery and replication brushing samples, we applied the R package sva [15] to cohorts are shown in Table  1. Regarding the discovery estimate significant surrogate variables (SVs), represent - cohort, the median [range] duration of follow-up of the ing unknown latent factors that capture heterogeneity entire population was 39 [4–49] years. Most of the sub- in data, such as cell type composition. One SV for each jects (group 1, n = 76) were from the study described by analysis (ClinR and ComR) was generated and added to Carpaij et  al. [4] at the third visit; the remaining (group the model respectively. Differentially methylated regions 2, n = 27) were remission subjects added to this study (DMR) were identified using comb-p v0.48 [16] and to further enrich remission cases. Characteristics of DMRcate [17]. subjects from the two groups are shown in Additional file  2: Table  S1. By definition, group 2 had a higher pro - Replication and meta‑analyses portion of remission subjects, and showed better lung Genome-wide significant CpG sites that passed Bonfer - function. Of the 103 subjects included in this study, 54 –7 roni correction (P < 1.14 × 10 , which is 0.05 / 436,824) (52.4% of total subjects) had ClinR, and 20 of the ClinR in whole blood DNA were selected for replication in two subjects had ComR (19.4% of total subjects). At baseline, independent cohorts, Lifelines and EGEA. Lifelines did subjects in remission later had higher FEV and FVC not include an assessment of AHR, so only included a level compared to PersA, but no difference regarding ClinR phenotype available for replication. The weighted FEV1%predicted and FVC %predicted values was iden- Z-score method was used to meta-analyze the results tified among the groups (Additional file  3: Figure S1). of discovery and replication cohorts, considering EGEA Within the whole dataset, 72 subjects (44 PersA, 28 ClinR used methylC-capture sequencing method, which was and 10 of ClinR subjects are ComR) had whole blood different from discovery and Lifelines (450  K array), to methylation data and 97 (44 PersA, 53 ClinR and 19 of investigate DNA methylation. CpG sites that passed the ClinR subjects are ComR) had nasal methylation data, –7 epigenome-wide significance threshold of P < 1.14 × 10 with 66 subjects providing both (Fig. 1). White blood cell (Bonferroni correction) in the meta-analysis of all studies composition was estimated, and we did not observe sig- were considered to be replicated. Finally, top sites identi- nificant cell composition differences among PersA, ClinR fied in whole blood were verified in nasal brushes. and ComR group in six cell types. (Additional file  3: Fig- ure S2). Annotation and functional relevance Significant CpG sites were annotated by GREAT v3.0.0 Differential methylation in the discovery cohort of ClinR [18]. We correlated the significant CpG sites we identi - Whole blood DNA methylation levels at seven individual fied to the expression level of gene nearby (with a region CpG sites and 19 differentially methylated regions (DMR) Qi et al. Clin Transl Allergy (2020) 10:60 Page 4 of 13 Table 1 Characteristics of study participants in the discovery and replication cohorts Discovery cohort Lifelines EGEA PersA ClinR ComR PersA ClinR Healthy PersA ClinR ComR (n = 3) (n = 49) (n = 54) (n = 20) (n = 99) (n = 25) (n = 636) (n = 106) (n = 15) Characteristics at last visit Age at 49.8 (48.5; 50.5 (48.3; 50.5 (48.4; 46.9 43.2 46.3 52.6 (36.7; 49.2 (36.7; 30.5 (28.2; remission 51.3) 57.6) 59.8) (40.7;50.7) (35.7;50.3) (39.2;52.9) 61.4) 57.6) 59.1) status, years Male, n (%) 31 (63.3%) 37 (68.5%) 15 (75.0%) 58 (58.6%) 10 (40.0%) 384 (60.4%) 49 (46.7%) 7 (46.7%) 1 (33.3%) Duration of 39 (38;40.5) 39 (37; 42) 39 (37;41) 11.3 (10.7; 11.6 (10.7; 11.6 (10.0; follow‑up, 12.1) 12.2) 12.7) years Current‑ 4 (8.2%) 8 (14.8%) 2 (10.0%) 27 (27.3%) 14 (56.0%) 207 (32.5%) 10 (9.5%) 2 (13.3%) 1 (33.3%) smoking, n (%) Ex‑smoking, 13 (26.5%) 18 (33.3%) 7 (35.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 35 (33.3%) 5 (33.3%) 0 n (%) Never‑ 32 (65.3%) 28 (51.9%) 11 (55.0%) 72 (72.7%) 11 (44.0%) 429 (67.5%) 60 (57.1%) 8 (53.3%) 2 (66.7%) smoking, n (%) ICS, n (%) 31 (63.3%) 0 0 48 (48.5%) 0 (0.0%) 0 (0.0%) 72 (70.6%) 0 0 Allergic 33 (67.3%) 30 (55.6%) 9 (45.0%) 59 (59.6%) 13 (52.0%) 123 (19.3%) 70 (67.3%) 11 (73.3%) 2 (66.7%) rhinitis, n (%) BMI 25.7 (23.7; 26.8 (23.7; 26.0 (22.8; 25.7 26.0 25.8 23.9 (22.4; 22.5 (19.9; 22.0 (19.9; 28.5) 28.3) 28.6) (22.9;28.0) (23.7;29.6) (23.6;28.1) 27.5) 28.0) 22.4) FEV1% pred 80.6 (71.1; 90.2 (82.9; 101.8 (85.9; 81.2 87.1 101.5 87.0 (71.7; 95.2 (87.5; 97.5 (90.7; 91.9) 104.7) 110.0) (70.9;91.6) (74.0;93.4) (96.1;108.7) 98.4) 107.2) 121.5) FVC% pred 94.2 (88.4; 101.7 (95.2; 106.6 (101.2; 94.1 100.3 104.0 98.5 (84.8; 95.2 (86.9; 96.3 (86.9; 103.7) 111.5) 115.9) (85.8;102.2) (90.4;109.2) (97.5;111.4) 108.9) 106.1) 119.2) FEV1/FVC 0.67 (0.60; 0.72 (0.68; 0.72 (0.68; 0.67 0.67 0.79 0.69 (0.61; 0.81 (0.78; 0.80 (0.80; 0.72) 0.75) 0.76) (0.62;0.76) (0.62;0.76) (0.75;0.83) 0.77) 0.82) 0.91) Characteristics at baseline Age, years 10.0 (9.0; 11.0 (9.0; 11.5 (9.8; 40.6 (25.7; 37.0 (26.0; 18.9 (18.8; 11.0) 17.5) 15.8) 48.7) 46.0) 46.4) FEV1% pred 76.6 (67.4; 77.1 (67.5; 78.1 (73.5; 81.7 (66.2; 90.6 (82.0; 82.0 (80.1; 84.0) 83.8) 83.7) 95.9) 103.0) 122.9) FVC% pred 88.8 (84.0; 88,3 (80.9; 88.9 (79.8; 95.5 (83.7; 86.4 (81.9; 84.5 (70.1; 96.1) 95.3) 96.5) 104.3) 100.6) 118.8) FEV1/FVC 0.73 (0.64; 0.73 (0.66; 0.77(0.66; 0.70 (0.63; 0.86 (0.82; 0.86 (0.83; 0.78) 0.79) 0.80) 0.80) 0.89) 0.96) Start 48 (98.0%) 50 (92.6%) 19 (95.0%) 59 (62.1%) 18 (81.8%) 52 (49.5%) 8 (53.3%) 2 (66.7%) asthma before age 16, n (%) Continuous data are presented as median (25th percentile; 75th percentile), category data are presented as number (percentage) Italic values represent significant differences compared with persistent asthma with P < 0.05 in the discovery cohort were significantly associated with ClinR (Fig.  2a, Table 2, file  2: Table S3). These four eQTM genes were enriched in Additional file  2: Table  S2). The Q-Q plot (Additional two pathways: G Protein Signaling Pathways and Hemo- file  3: Figure S3a) and the inflation factor λ of 1.064 indi - stasis (P < 0.01, Additional file 2: Table S4). cated no obvious inflation of the results. Three out of the We also identified four epigenome-wide significant seven CpG sites were significantly correlated with gene- CpG sites and 24 DMRs that associated with ClinR expression level in cis (FDR < 0.05), resulting in four CpG- in nasal brushes (Table  S5-S6). The Q-Q plot (Addi - gene pairs that showed negative correlation (Additional tional file  3: Figure S4a) and the λ of 1.125 indicated Qi  et al. Clin Transl Allergy (2020) 10:60 Page 5 of 13 Nasal brush Whole blood samples samples N=31 N= 66 N=6 DNA methylation in nasal brush samples (N=97) DNA methylation in whole blood samples (N=72) PersA ComR ClinR only PersA ComR ClinRonly N=44 N=19 N=34 N=44 N=10 N=18 PersAvs ComR PersAvs ClinR PersAvs ComR PersAvs ClinR 21 candidate CpG sites 4 candidate CpG sites 129 candidate CpG sites 7 candidate CpG sites Replication in EGEA Annotation and functional relevance ReplicationinEGEA and Lifelines Meta-analysis Meta-analysis 1 CpG replicated 1 CpG replicated Annotation and functional relevance Fig. 1 Study design. In the discovery panel,103 samples were assessed in this study, 72 subjects had whole blood DNA methylation data, 97 subjects had nasal DNA methylation data and 66 subjects have both. Epigenome‑ wide association studies were performed on clinical remission (ClinR) and complete remission (ComR) respectively in both whole blood and nasal samples. Significant CpG sites identified from whole blood were further replicated in two independent cohorts. Results of ComR were replicated in EGEA study (93 out of 129 probes available in EGEA); results of ClinR were replicated in EGEA study (4 out 7 probes available) and Lifelines (all 7 probes available) Fig. 2 Manhattan plot of association between clinical remission (a)/ complete remission (b) and DNA methylation in whole blood in discovery –7 cohort. In total, 436,824 CpGs were tested. The red line represents the genome‑ wide significance threshold (Bonferroni correction, P < 1.14 × 10 ). Highlighted sites represent replicate CpG sites no obvious inflation of the results. We verified the were correlated to expression level of four genes in cis CpG sites identified from whole blood samples in (nominal significant P < 0.05) in nasal brushes, with one nasal brushes, and found that one CpG (cg15404785) pair (cg07673230-RPL30 (Ribosomal Protein L30)) also associated with ClinR showed nominal significance being nominal significant and in the same direction in the PIAMA dataset (Additional file 2 : Table S8). (P < 0.05) and the same direction of effect in nasal brushes (Table  S7). Four ClinR-associated CpG sites Qi et al. Clin Transl Allergy (2020) 10:60 Page 6 of 13 –7 Table 2 Seven genome-wide significant CpG sites (P value < 1.14 × 10 ) of ClinR in relation to methylation in whole blood CpG CHR Great annotation (distance Discovery ReplicationEGEA ReplicationLifelines Metaanalysis to TSS) Coef SE P N Coef SE P N Coef SE P N P.meta Direction cg10125195 12 LACRT (+ 133) 1.32E−02 2.30E−03 9.03E−09 72 1.098 0.748 0.142 80 (69/11) 2.73E−04 3.56E−03 0.939 124 1.58E−04 + + + cg12713893 18 SNRPD1 (+ 44) 4.64E−03 8.63E−04 7.52E−08 72 7.127 5.359 0.184 120 (106/14) 4.19E−04 1.11E−03 0.705 124 2.90E−04 + + + cg15404785 14 TRMT5 (− 279,717); TMEM30B − 3.09E−02 5.17E−03 2.43E−09 72 0.423 0.150 0.005 105 (93/12) − 4.93E−03 6.11E−03 0.420 124 7.58E−02 − + − (+ 20,765) cg27252019 17 MKS1 (+ 550) − 3.86E−02 7.24E−03 1.01E−07 72 − 0.033 0.269 0.902 91 (80/11) 5.71E−03 1.01E−02 0.572 124 1.80E−02 – + cg06185924 10 ACBD5 (− 42) 1.10E−02 2.06E−03 9.36E−08 72 NA NA NA 106 (92/14) − 4.72E−03 2.33E−03 0.043 124 1.05E−01 + NA− cg13378519 1 PEX11B (+ 18) − 5.99E−03 1.07E−03 2.43E−08 72 NA NA NA 119 (105/15) − 1.08E−02 2.61E−03 3.57E05 124 2.58E−11 −NA− cg17900884 1 ICMT (− 91) 3.58E−03 6.28E−04 1.15E−08 72 NA NA NA 114 (101/13) − 1.78E−04 9.86E−04 0.857 124 9.16E−04 + NA− P.meta: the P value of meta-analysis of results from discovery, EGEA and Lifelines by using weighted Z-score method Direction: the direction of regression coefficient in discovery, EGEA and Lifelines NA: the result for this probe was not available in EGEA a –7 The CpG site was replicated with meta P value < 1.14 × 10 N (persA/ClinR), N varies from one CpG to another because of the methylC-capture sequencing method used and QC criteria in EGEA Qi  et al. Clin Transl Allergy (2020) 10:60 Page 7 of 13 Table 3 Top ten CpG sites in meta-analysis of ComR in relation to methylation in whole blood showing the same direction of effect in discovery and replication cohort CpG CHR Great annotation (distance to TSS) Discovery ReplicationEGEA Metaanalysis Coef SE P N Coef SE P N P.meta Direction cg24788483 10 HABP2 (− 401132), TCF7L2 (+ 201,644) − 1.31E−02 2.11E−03 4.62E−10 54 − 1.348 0.544 0.013 56 (53/3) 8.49E−10 – cg15223066 10 ITGB1 (− 23,896), NRP1 (+ 354472) − 8.56E−03 1.39E−03 6.65E−10 54 − 0.658 0.470 0.161 92 (89/3) 1.13E−06 – cg26909813 17 RPRML (605) − 2.68E−02 4.15E−03 1.06E−10 54 − 0.621 0.539 0.249 95 (92/3) 1.52E−06 – cg02341571 15 PATL2 (+ 150) − 1.69E−02 2.71E−03 4.39E−10 54 − 0.253 0.388 0.515 64 (62/2) 2.59E−06 – cg00269245 16 SLC12A3 (− 41116), NUP93 (+ 93986) 1.13E−02 1.86E−03 1.14E−09 54 0.832 1.109 0.453 70 (68/2) 4.62E−06 + + cg24833566 4 MTNR1A (− 152680), FAT1 (+ 15608) 1.02E−02 1.71E−03 2.38E−09 54 0.658 0.758 0.386 77 (74/3) 6.87E−06 + + cg25881850 11 KCNK4 (− 1426) − 1.23E−02 2.23E−03 3.43E−08 54 − 0.838 0.816 0.304 68 (67/1) 9.09E−06 – cg24730612 14 FOS (− 95503), TMED10 (− 6640) − 1.47E−02 2.50E−03 4.15E−09 54 − 0.260 0.489 0.595 62 (61/1) 1.09E−05 – cg23250019 11 SLC29A2 (+ 818) 2.15E−02 4.00E−03 7.62E−08 54 0.645 0.541 0.234 82 (79/3) 1.61E−05 + + P.meta: the P value of meta-analysis of results from discovery and EGEA by using weighted Z-scores method Direction: the direction of regression coefficient in discovery and EGEA A list of all 129 significant CpG sites identified in discovery cohort and the information of replication in EGEA is shown in Additional file 2: Table S9 a –7 The CpG site was replicated with meta P value < 1.14 × 10 N (persA/ComR), N varies from one CpG to another because of the methylC-capture sequencing method used and different QC criteria in EGEA Qi et al. Clin Transl Allergy (2020) 10:60 Page 8 of 13 Differential methylation in the discovery cohort of ComR compared to PersA in whole blood, but showed no obvi- We identified 129 individual CpG sites and 53 regions ous difference in nasal brushes (Additional file  3: Figure that were significantly associated with ComR in whole S5). In whole blood cis-eQTM analysis, this CpG was sig- blood DNA (Fig.  2b, Table  3, Additional file  2: Table  S9, nificantly negatively correlated to the expression level of –7 10). Q-Q plot is shown in Additional file  3: Figure S3b TCF7L2 gene (P = 9.12 × 10 ). and λ was 1.178. The eQTM analysis identified 99 CpG- To evaluate potential SNP effects within the probe for gene pairs for 45 out of the 129 CpG sites (FDR < 0.05, the two replicated CpG sites, the β value distributions Additional file  2: Table  S3). These eQTM genes were were visually assessed in the discovery cohorts (Addi- enriched in eight pathways (P < 0.01, Additional file  2: tional file  3: Figure S6), and no bimodal distribution was Table S4), and the top pathways were Activation of SMO detected indicating no underlying SNP effect within the (Smoothen), Platelet Adhesion to exposed collagen, and probe. The regional co-methylation plots of genes anno - Cilium Assembly. tated to replicated CpG sites are shown in the Additional In nasal brushes, we identified 21 CpG sites and 62 file  3: Figure S7, which showed that cg13378519 is located DMRs that were associated with ComR (Additional file  2: in the promoter region of PEX11B gene, and cg24788483 Table S11, 12). Q-Q plot is shown in Additional file  3: Fig- is located in heterochromatin region of TCF7L2 gene. ure S4b and λ was 1.318. Eight of the ComR-associated We did not identify any SNP that was associated with the CpG sites identified from whole blood also showed nomi - two CpG sites using an online database [19]. nal significance in nasal brushes with the same direction We verified if the results of the two replicated CpG of effect (Additional file  2: Table  S13). Seven ComR- sites were affected by inhaled corticosteroids (ICS), associated CpG sites were correlated to expression level known to affect DNA methylation at multiple sites in of nine genes in cis (nominal significant P < 0.05) in nasal the genome (Additional file  3: Figure S8). The boxplots brushes, and four pairs also showed nominal significant showed that the methylation levels in the remission and in the same direction in PIAMA dataset (Additional groups (both ClinR and ComR) were significantly lower file 2: Table S8). than both persistent asthma with ICS usage group and persistent asthma without ICS usage group, which indi- Replication study cated that the results were not affected by the use of ICS. We replicated the top findings in whole blood in two We evaluated the impact of rhinitis on our results, as independent cohorts. In replication cohort EGEA, four rhinitis shows comorbidity with asthma. The replicated ClinR-associated CpG sites and 93 ComR-associated ClinR-associated CpG site showed similar effect size and CpG sites were available, with one CpG site for ClinR and P value before and after adjusting for rhinitis, while the one for ComR showing nominal significance (P < 0.05) ComR-associated CpG showed similar effect size yet the and the same direction as the discovery cohort. In the P value changed from 4.62E-10 to 4.07E-05 after correc- replication cohort Lifelines, which only had the ClinR tion for rhinitis (Additional file  2: Table S14). Additional phenotype, seven ClinR-associated CpG sites were all adjustment for BMI did not change the results of the available with one CpG with P < 0.05 and the same direc- two replicated CpG sites, indicating that BMI was not a tion. After meta-analysis, one CpG associated with ClinR confounder (Additional file  2: Table S14). The significant and another one CpG associated with ComR were rep- CpG sites identified from nasal samples showed similar –7 licated (meta P < 1.14 × 10 ). The ClinR-associated CpG effect size and P value before and after adjusting for BMI (cg13378519) was located on Chr 1 and close to gene and for rhinitis (Additional file 2: Table S15). PEX11B (Peroxisomal Biogenesis Factor 11 Beta). The We also compared our results from the discovery CpG site was not available in the EGEA study and was cohort with the published DNA methylation study of only replicated in Lifelines. This CpG was lower meth - asthma remission in bronchial biopsies (Vermeulen et al. ylated in ClinR subjects compared to PersA (Fig.  3a). A [10]). One CpG (cg13525448), located close to LBX1 similar trend was also found in nasal brushes although (Ladybird Homeobox  1) gene and TLX1 (T Cell Leuke- the association was not statistically significant (Fig.  3b). mia Homeobox  1) gene, in their top list of remission vs The methylation levels at this CpG also differed between persistent asthma reached nominal significance (P < 0.05) ClinR and healthy subjects in Lifelines cohort (Fig.  3c), and showed the same direction of effect in our results of indicating that DNA methylation status of asthma remis- ComR in whole blood (Additional file  2: Table S16). Two sion subjects is not equal to that of healthy subjects. genes of their DMRs associated with ClinR, at the gene The ComR-associated CpG (cg24788483) was located PTCHD3 (patched domain-containing protein 3) and on chr10 and close to gene HABP2 (Hyaluronan Bind- LOC100507389, were also present in our results of DMRs ing Protein 2) and TCF7L2 (Transcription Factor 7 Like associated with ClinR in nasal brushes (Additional file  2: 2). This CpG was lower methylated in ComR subjects Table S6). Qi  et al. Clin Transl Allergy (2020) 10:60 Page 9 of 13 (a) cg13378519 in blood (discovery) (b) cg13378519 in nasal brush (discovery) (c) cg13378519 in blood (Lifelines) 0.16 0.16 0.16 0.12 0.12 0.12 0.08 0.08 0.08 0.04 0.04 0.04 PerAsthma ClinR ComR PerAsthmaClinR ComR Asthma ClinR Healthy Fig. 3 Boxplot illustrating DNA methylation levels of cg13378519 in persistent asthma, clinical remission and complete remission subjects. a DNA methylation levels in whole blood in discovery cohort; b DNA methylation levels in nasal brushes in discovery cohort; c DNA methylation levels in whole blood in replication cohort (Lifelines) the pancreatic insulin secretory response to incretins Discussion [25, 26]. Previously, the risk of asthma was reported to In this study, we identified one CpG site associated with be higher in type 2 diabetes patients [27] and it has been ClinR and a different CpG site associated with ComR suggested that treatment targeting insulin resistance may from whole blood, that were replicated in independent have a positive effect on asthma patients [28]. However, cohorts. We also revealed that 25 CpG sites were associ- it has never been studied whether this treatment could ated with asthma remission phenotypes in nasal brushes. contribute to asthma remission. In addition, we could In the nasal dataset, two differentially methylated regions not infer from the current results of association analysis were previously observed in another DNA methylation whether the DNA methylation signal we identified is the study of asthma remission using bronchial biopsies. This cause or consequence of asthma remission. is the first epigenome-wide association study of asthma Although remission subjects do not have asthma remission in both whole blood and nasal epithelial cells. symptoms any more, they still may be different from This study is important since understanding asthma healthy people. Airway abnormalities such as basement remission may provide new leads for future asthma treat- membrane thickening still exist in clinical and complete ments, and epigenetic studies may help to reveal these remission subjects [29]. Vermeulen et  al. [10] reported cellular mechanisms. that the DNA methylation profile in remission subjects is The two replicated CpG sites were annotated to genes different from that in healthy subjects. Our data in whole which were not known to be involved in asthma (remis- blood for our significantly replicated CpG confirmed this sion) before. ClinR-associated CpG cg13378519 was observation. The methylation levels of the ClinR associ - located in the promoter region of PEX11B gene. PEX11B ated CpG (cg13378519) were different between ClinR encodes a protein of the PEX11 family [22]. Overex- and healthy subjects, which indicated that the methyla- pression of the PEX11B gene in human cells induces tion status of subjects with asthma remission may not peroxisome proliferation [23, 24]. Our results could simply return to that of healthy people. indicate a possible role of peroxisomal proliferation as DNA methylation is cell-type specific. Both of the stud - being involved in asthma remission. ComR-associated ied cells or tissues are highly heterogeneous and this may CpG cg24788483 was located in the TCF7L2 gene and confound the association between DNA methylation also correlated with the expression level of this gene. and disease. Thus, we applied the widely used House - The TCF7L2 gene product is a transcription factor man’s cell type correction method in blood samples, and that plays a role in the Wnt signaling pathway. SNPs in for nasal brushes, we used SVA which showed good per- TCF7L2 were previously reported to be strongly asso- formance in cell type correction when reference data is ciated with type-2 diabetes and this gene plays a role in Beta value Beta value Beta value Qi et al. Clin Transl Allergy (2020) 10:60 Page 10 of 13 lacking [30]. DNA methylation is also tissue specific. during the airway remodeling of persistent asthma [37]. It is yet unknown if mechanisms of remission include Blood is considered as an easily accessible surrogate tis- local, tissue specific airway (epithelial) effects, immune sue for asthma study, and previous studies showed some cell effects, or both. Matched data from whole blood and shared gene expression pattern and pathways between nasal brush samples enable us to compare DNA meth- blood and airway/ lung [38, 39]. Our results suggest that ylation profiles associated with asthma remission in two genes related to asthma remission associated CpG sites different tissues, as proxies of both immunological and may reflect a role for airway epithelial barrier and airway airway specific mechanisms. Although we identified sev - remodeling in asthma remission. eral interesting CpG sites related to remission in nasal Corticosteroids treatment is usually used to control epithelium, our results of nasal brushes still need further inflammation and improve control of asthma. One pre - replication. In previous studies, several DNA methylation vious study identified differences in DNA methylation sites were significantly associated with asthma in both of CpG sites in blood associated with systemic corticos- whole blood and nasal cells [9, 31]. In our study, although teroid treatment in patients with COPD, suggesting the no significant enriched cross-tissue effect of DNA meth - potential effect of corticosteroids on the DNA meth - ylation in asthma remission was shown, we did identify ylation profile [40]. In our study, we therefore assessed nine CpG sites in blood DNA that could be replicated in the effect of ICS on the methylation levels of the two DNA of nasal brushed cells. This indicated that there may replicated CpG sites and found that the results were be shared DNA methylation signals in blood and nasal not affected by ICS usage. However, the effect of corti - cells. We interpret these findings as either indicative of costeroid treatment on DNA methylation in nasal and the presence of cross-tissue epigenetic mechanisms in blood samples of patients with asthma and remission is blood and epithelial cells, or the possibility that the DNA largely unknown and should be further investigated in methylation signal in nasal brushed cells is partly driven the future. by immune cells in the nose [20], that are also present in Rhinitis and asthma often co-exist and show shared blood. genetic origins [41]. A previous study of our group also DNA methylation might be related to the regulation of showed a shared DNA methylation signal of asthma and gene expression, and eQTM analysis may help to under- rhinitis, suggesting the importance of considering the stand the function of CpG sites. Among the eQTM genes presence of rhinitis when doing EWAS of asthma in nasal that correlated with ClinR-associated CpG sites in whole epithelium [20]. In this study, we showed that additional blood, the protein encoded by PRKCH (Protein Kinase adjustment for rhinitis changed the results on one of the C Eta) plays a key role in epithelial tight junction regu- replicated CpG sites (cg24788483) identified from blood. lation which is important in maintaining the integrity This remission associated CpG site was associated with and function of the airway epithelial barrier [32]; PDE1B active rhinitis in the absence of asthma, suggesting that (Phosphodiesterase 1B) encoded a protein belonging blood DNA methylation profiles of persistent asthma and to phosphodiesterases (PDEs) family, and various PDE rhinitis overlap. This indicates that we should control for inhibitors showed anti-inflammatory, anti-remodelling allergic rhinitis in future studies of asthma remission. and bronchodilator effect and are potential treatment of When comparing our results with those of Vermeulen asthma [33]. In whole blood, 99 eQTM genes that were et  al. [10] in bronchial biopsies, one CpG (cg13525448) associated with ComR were enriched in eight path- in their top list of remission vs persistent asthma reached ways, among which three pathways were related to air- nominal significance (P < 0.05) in our data and showed way epithelial function: activation of SMO (Smoothen), the same direction in our results of ComR in whole blood cilium assembly and focal adhesion. The activation of (Additional file  2: Table  S16). This CpG site is anno - SMO activity in bronchial epithelia enhanced the aller- tated to LBX1 gene and TLX1 gene by position. Besides, gen-induced goblet cell metaplasia, which is defined among their DMRs associated with ClinR, regions at the as a reversible transformation of airway epithelial cells gene PTCHD3 and LOC100507389 also showed up in our to mucous cells such as goblet cells and may occur in results of ClinR in nasal brushes. SNPs in PTCHD3 gene asthma [34]. Airway cilia are important for clearance were previously associated with asthma in African Amer- of inhaled particles and pathogens. One study showed ican children [42]. asthma patients had less ciliated cells in airway samples There are strengths and limitations in this study. than healthy people [35], and changes in ciliary function To the best of our knowledge, this is the only cohort may be relevant for the development of asthma in chil- worldwide to study asthma remission in nasal brushes, dren [36]. Focal adhesion between the cell membrane which enable us to compare the DNA methylation pro- and matrix are essential elements of airway smooth mus- file in relation to asthma remission in whole blood and cle cells migration which may play an important role nasal epithelium. However, because of the uniqueness Qi  et al. Clin Transl Allergy (2020) 10:60 Page 11 of 13 of this data, we could not replicate our results in nasal and after adjusting for BMI and rhinitis. Table S15. Summary statistics of brushes in another independent cohort. Notably, we had significant CpGs (identified from nasal samples) before and after adjusting for BMI and rhinitis. Table S16. Look up of CpGs associated with ClinR matched DNA methylation and gene expression data in identified by Vermeulen et al. in this study. nasal brushes, which help us to get a better understand- Additional file 3: Figure S1. Lung function of different groups at baseline ing of the function of CpG sites associated with asthma and last visit. In this figure, t ‑test was used in comparing means of any two remission in nasal brushes. Regarding the limitations groups (ns: P>0.05, *: P<0.05, **: P<0.01, ***: P<0.001, ****: P<0.0001). Fig‑ ure S2. Estimated cell proportions among different groups. In this figure, of this study, firstly, this study was performed in a rela - t‑test was used in comparing means of any two groups (ns: P>0.05). Fig ‑ tively small sample with limited number of ComR cases, ure S3. Quantile–quantile plot for epigenome‑ wide meta‑analysis of the and the results were replicated in EGEA which is also association between asthma remission and blood DNA methylation (n = 72). (a) ClinR, (b) ComR. Figure S4. Quantile–quantile plot for epigenome‑ with few ComR cases, which lead to low power of rep- wide meta‑analysis of the association between asthma remission and lication of ComR results. Secondly, although we studied nasal DNA methylation (n = 97). (a) ClinR, (b) ComR. Figure S5. Boxplot over 450,000 DNA methylation sites, this only repre- illustrating DNA methylation levels of cg24788483 in persistent asthma, clinical remission and complete remission subjects in discovery cohort. (a) sents 1.6% of the human DNA methylome. Thirdly, we DNA methylation levels in blood, (b) DNA methylation in nasal brushes. studied mixed contributions of cells in whole blood and Figure S6. Density distributions of DNA methylation levels of two nasal brushes; and although we corrected for cell types replicated CpGs in 72 blood samples from discovery cohorts. Figure S7. Regional association plot of two replicated CpGs. For each plot from top using established methods, potentially stronger results to bottom the tracks included are: 1) Log10(P values) from the discovery may be anticipated when studying pure cell types, as was 4 years model with CpGs indicated by dots. 2) Annotation tracks for the previously shown for DNA methylation sites associated plotted genomic region taken from UCSC Genome Browser. 3) Pairwise correlation matrix across the displayed CpGs. Figure S8. DNA methylation with asthma in purified blood eosinophils [8]. Finally, levels of two replicated CpGs in remission subjects and asthma patients we used a cross sectional analysis at follow up to inves- stratified by ICS usage. In this figure, t ‑test was used in comparing means tigate asthma remission. Future studies, with prospective of any two groups (*: P<0.05, ***: P<0.001, ****: P<0.0001). designs, should be performed to investigate if these DNA methylation sites can predict future asthma remission. Acknowledgements In conclusion, we identified replicable DNA methyla - We thank the participants in this study and all staff involved in the study through the years. tion signals associated with clinical and complete asthma remission, which may play a role in peroxisome prolifera- Authors’ contributions tion and Wnt signaling pathway. This could help in iden - CQ, JMV, HMB, CJX and GHK contributed to the design of the study. CQ, CJX, DA van der P and DA contributed to data analysis. JMV, AMEN, FND and tifying the underlying mechanisms of asthma remission VS contributed to study procedures, participant recruitment or laboratory and also of the chronicity of the disease. procedures. All authors contributed to results interpretation. CQ, CJX and GHK contributed to the initial manuscript draft. All authors revised the manuscript draft for important intellectual content. All authors read and approved the Supplementary information final manuscript. Supplementary information accompanies this paper at https ://doi. org/10.1186/s1360 1‑020‑00365 ‑4. Funding The Biobank‑Based Integrative Omics Studies (BIOS) Consortium is funded by BBMRI‑NL, a research infrastructure financed by the Dutch government (NWO Additional file 1. Supplementary materials. 184.021.007). Additional file 2: Table S1. Characteristics of subjects from two groups. Table S2. Differentially methylated regions associated with ClinR in Availability of data whole blood identified by both comb ‑p (Sidak p ‑ value) and DMRcate DNA methylation data has been deposited at the European Genome‑phe ‑ (FDR) methods. Table S3 Correlation analysis of DNA methylation and nome Archive (EGA), which is hosted by the EBI and the CRG, under accession gene expression levels (in cis, +/‑ 250 kb) for asthma remission in blood. number EGAS00001004766. Table S4. Pathway analysis of eQTM genes correlated with ClinR and ComR in blood. Table S5. Genomewide significant (pvalue < 1.14E‑07) Ethics approval and consent to participate CpGs associated with ClinR in nasal brushes. Table S6. Differentially This study was approved by the medical ethical board of University Medical methylated regions associated with ClinR in nasal brushes identified by Center Groningen, and all participants gave written informed consent. both comb‑p (Sidak p ‑ value) and DMRcate (FDR) methods. Table S7. Look up of ClinR‑assocaited CpGs (indentified from blood) in the results of Competing interests nasal brushes. Table S8. Correlation analysis of DNA methylation and gene GHK reports grants from Lung Foundation of the Netherlands, TEVA the expression levels (in cis, +/‑ 250 kb) for asthma remission in nasal brushes. Netherlands, Vertex, GSK, Ubbo Emmius Foundation, TETRI foundation, outside Table S9. 129 genome‑ wide significant (pvalue < 1.14E‑07) ComR‑asso ‑ the submitted work. The rest of the authors declare that they have no relevant ciated CpGs in whole blood in discovery cohort. Table S10. Differentially conflict of interests. methylated regions associated with ComR in whole blood identified by both comb‑p (Sidak p ‑ value) and DMRcate (FDR) methods. Table S11. Author details Genome‑ wide significant (p value < 1.14E‑07) CpGs associated with ComR Department of Pediatric Pulmonology and Pediatric Allergy, Beatrix Children’s in nasal brushes. Table S12. Differentially methylated regions associated Hospital, University Medical Center Groningen, University of Groningen, PO with ComR in nasal brushes identified by both comb ‑p (Sidak p ‑ value) and Box 30.001, 9700 RB Groningen, the Netherlands. GRIAC Research Institute, DMRcate (FDR) methods. Table S13. Look up of ComR‑associated CpG University Medical Center Groningen, University of Groningen, Groningen, sites (indentified from blood) in the results of nasal brushes. Table S14. The Netherlands. Department of Epidemiology, University Medical Center Summary statistics of replicated CpGs (identified from blood) before Groningen, University of Groningen, Groningen, The Netherlands. University Qi et al. Clin Transl Allergy (2020) 10:60 Page 12 of 13 Grenoble Alpes, Inserm, CNRS, Team of environmental epidemiology Bioinformatics. 2012;28(22):2986–8. https ://doi.org/10.1093/bioin forma applied to Reproduction and Respiratory health, IAB (Institute for Advanced tics/bts54 5. Biosciences), 38000 Grenoble, France. Research Group of Bioinformatics 17. Peters TJ, Buckley MJ, Statham AL, et al. De novo identification of differen‑ and Computational Genomics, CiiM, Centre for individualized infection medi‑ tially methylated regions in the human genome. Epigenetics Chromatin. cine, A Joint Venture Between Hannover Medical School and the Helmholtz 2015;8:6. https ://doi.org/10.1186/1756‑8935‑8‑6. Centre for Infection Research, Hannover, Germany. Department of Gastroen‑ 18. McLean CY, Bristor D, Hiller M, et al. GREAT improves functional interpreta‑ terology, Hepatology and Endocrinology, TWINCORE, Centre for Experimental tion of cis‑regulatory regions. Nat Biotechnol. 2010;28(5):495–501. https :// and Clinical Infection Research, A Joint Venture Between the Hannover doi.org/10.1038/nbt.1630. Medical School and the Helmholtz Centre for Infection Research, Hannover, 19. the BIOS Consortium, Bonder MJ, Luijk R, et al. Disease variants alter tran‑ Germany. Department of Internal Medicine, Radboud University Medical scription factor levels and methylation of their binding sites. Nat Genet. Center, Nijmegen, the Netherlands. 2017;49(1):131–8. https ://doi.org/10.1038/ng.3721. 20. Qi C, Jiang Y, Yang IV, et al. Nasal DNA methylation profiling of asthma Received: 4 August 2020 Accepted: 26 November 2020 and rhinitis. J Allergy Clin Immunol. 2020. https ://doi.org/10.1016/j. jaci.2019.12.911. 21. Herwig R, Hardt C, Lienhard M, Kamburov A. Analyzing and interpreting genome data at the network level with ConsensusPathDB. Nat Protoc. 2016;11(10):1889–907. https ://doi.org/10.1038/nprot .2016.117. 22. Koch J, Pranjic K, Huber A, et al. PEX11 family members are membrane References elongation factors that coordinate peroxisome proliferation and 1. Upham JW, James AL. Remission of asthma: the next therapeutic frontier? maintenance. J Cell Sci. 2010;123(19):3389–400. https ://doi.org/10.1242/ Pharmacol Ther. 2011;130(1):38–45. https ://doi.org/10.1016/j.pharm thera jcs.06490 7. .2011.01.002. 23. Schrader M, Reuber BE, Morrell JC, et al. Expression of PEX11beta medi‑ 2. Carpaij OA, Burgess JK, Kerstjens HAM, Nawijn MC, van den Berge M. A ates peroxisome proliferation in the absence of extracellular stimuli. J Biol review on the pathophysiology of asthma remission. Pharmacol Ther. Chem. 1998;273(45):29607–14. https ://doi.org/10.1074/jbc.273.45.29607 . 2019;201:8–24. https ://doi.org/10.1016/j.pharm thera .2019.05.002. 24. Ebberink MS, Koster J, Visser G, et al. A novel defect of peroxisome divi‑ 3. Vonk JM, Postma DS, Boezen HM, et al. Childhood factors associated with sion due to a homozygous non‑sense mutation in the PEX11β gene. J asthma remission after 30 year follow up. Thorax. 2004;59(11):925–9. https Med Genet. 2012;49(5):307–13. https ://doi.org/10.1136/jmedg enet‑2012‑ ://doi.org/10.1136/thx.2003.01624 6. 10077 8. 4. Carpaij OA, Nieuwenhuis MAE, Koppelman GH, van den Berge M, Postma 25. Grant SFA, Thorleifsson G, Reynisdottir I, et al. Variant of transcription DS, Vonk JM. Childhood factors associated with complete and clinical factor 7‑like 2 ( TCF7L2) gene confers risk of type 2 diabetes. Nat Genet. asthma remission at 25 and 49 years. Eur Respir J. 2017;49(6):1601974. 2006;38(3):320–3. https ://doi.org/10.1038/ng173 2. https ://doi.org/10.1183/13993 003.01974 ‑2016. 26. Pearson ER. Translating TCF7L2: from gene to function. Diabetologia. 5. Vonk JM, Nieuwenhuis MAE, Dijk FN, et al. Novel genes and insights in 2009;52(7):1227–30. https ://doi.org/10.1007/s0012 5‑009‑1356‑1. complete asthma remission: a genome‑ wide association study on clinical 27. Thomsen SF, Duffy DL, Kyvik KO, Skytthe A, Backer V. Risk of asthma in and complete asthma remission. Clin Exp Allergy. 2018;48(10):1286–96. adult twins with type 2 diabetes and increased body mass index: type 2 https ://doi.org/10.1111/cea.13181 . diabetes, obesity and asthma in twins. Allergy. 2011;66(4):562–8. https :// 6. Qi C, Xu C‑ J, Koppelman GH. The role of epigenetics in the development doi.org/10.1111/j.1398‑9995.2010.02504 .x. of childhood asthma. Expert Rev Clin Immunol. 2019;15(12):1287–302. 28. Carpaij OA, van den Berge M. The asthma–obesity relationship: underly‑ https ://doi.org/10.1080/17446 66X.2020.16869 77. ing mechanisms and treatment implications. Curr Opin Pulmon Med. 7. Gibney ER, Nolan CM. Epigenetics and gene expression. Heredity. 2018;24(1):42–9. https ://doi.org/10.1097/MCP.00000 00000 00044 6. 2010;105(1):4–13. https ://doi.org/10.1038/hdy.2010.54. 29. Broekema M, Timens W, Vonk JM, et al. Persisting remodeling and less 8. Xu C‑ J, Söderhäll C, Bustamante M, et al. DNA methylation in child‑ airway wall eosinophil activation in complete remission of asthma. hood asthma: an epigenome‑ wide meta‑analysis. Lancet Respir Med. Am J Respir Crit Care Med. 2011;183(3):310–6. https ://doi.org/10.1164/ 2018;6(5):379–88. https ://doi.org/10.1016/S2213 ‑2600(18)30052 ‑3. rccm.20100 3‑0494O C. 9. Forno E, Wang T, Qi C, et al. DNA methylation in nasal epithelium, atopy, 30. Kaushal A, Zhang H, Karmaus WJJ, et al. Comparison of different cell type and atopic asthma in children: a genome‑ wide study. Lancet Respir Med. correction methods for genome‑scale epigenetics studies. BMC Bioinfor ‑ 2019;7(4):336–46. https ://doi.org/10.1016/S2213 ‑2600(18)30466 ‑1. mat. 2017;18(1):216. https ://doi.org/10.1186/s1285 9‑017‑1611‑2. 10. Vermeulen CJ, Xu C‑ J, Vonk JM, et al. Differential DNA methylation in 31. Cardenas A, Sordillo JE, Rifas‑Shiman SL, et al. The nasal methylome as a bronchial biopsies between persistent asthma and asthma in remission. biomarker of asthma and airway inflammation in children. Nat Commun. Eur Respir J. 2019. https ://doi.org/10.1183/13993 003.01280 ‑2019. 2019;10(1):3095. https ://doi.org/10.1038/s4146 7‑019‑11058 ‑3. 11. Nieuwenhuis MA, Siedlinski M, van den Berge M, et al. Combining 32. Suzuki T, Elias BC, Seth A, et al. PKC eta regulates occludin phospho‑ genomewide association study and lung eQTL analysis provides evi‑ rylation and epithelial tight junction integrity. Proc Natl Acad Sci USA. dence for novel genes associated with asthma. Allergy. 2016;71(12):1712– 2009;106(1):61–6. https ://doi.org/10.1073/pnas.08027 41106 . 20. https ://doi.org/10.1111/all.12990 . 33. Zuo H, Cattani‑ Cavalieri I, Musheshe N, Nikolaev VO, Schmidt M. Phos‑ 12. Allum F, Shao X, Guénard F, et al. Characterization of functional methyl‑ phodiesterases as therapeutic targets for respiratory diseases. Pharmacol omes by next‑ generation capture sequencing identifies novel disease ‑ Ther. 2019;197:225–42. https ://doi.org/10.1016/j.pharm thera .2019.02.002. associated variants. Nat Commun. 2015;6:7211. https ://doi.org/10.1038/ 34. Xu C, Zou C, Hussain M, et al. High expression of Sonic hedgehog in aller‑ ncomm s8211 . gic airway epithelia contributes to goblet cell metaplasia. Mucosal Immu‑ 13. Houseman E, Accomando WP, Koestler DC, et al. DNA methylation arrays nol. 2018;11(5):1306–15. https ://doi.org/10.1038/s4138 5‑018‑0033‑4. as surrogate measures of cell mixture distribution. BMC Bioinformat. 35. Vieira Braga FA, Kar G, Berg M, et al. A cellular census of human lungs 2012;13(1):86. https ://doi.org/10.1186/1471‑2105‑13‑86. identifies novel cell states in health and in asthma. Nat Med. 2019. https 14. Aryee MJ, Jaffe AE, Corrada‑Bravo H, et al. Minfi: a flexible and com‑ ://doi.org/10.1038/s4159 1‑019‑0468‑5. prehensive Bioconductor package for the analysis of Infinium DNA 36. Dizier M‑H, Nadif R, Margaritte ‑ Jeannin P, et al. Interaction between the methylation microarrays. Bioinformatics. 2014;30(10):1363–9. https ://doi. DNAH9 gene and early smoke exposure in bronchial hyperresponsive‑ org/10.1093/bioin forma tics/btu04 9. ness. Eur Respir J. 2016;47(4):1072–81. https ://doi.org/10.1183/13993 15. Leek JT, Johnson WE, Parker HS, Jaffe AE, Storey JD. The sva package 003.00849 ‑2015. for removing batch effects and other unwanted variation in high‑ 37. Gerthoffer WT. Migration of Airway Smooth Muscle Cells. Proceed Am throughput experiments. Bioinformatics. 2012;28(6):882–3. https ://doi. Thoracic Soc. 2008;5(1):97–105. https ://doi.org/10.1513/pats.20070 org/10.1093/bioin forma tics/bts03 4. 4‑051VS . 16. Pedersen BS, Schwartz DA, Yang IV, Kechris KJ. Comb‑p: software for com‑ bining, analyzing, grouping and correcting spatially correlated P‑ values. Qi  et al. Clin Transl Allergy (2020) 10:60 Page 13 of 13 38. Morrow JD, Chase RP, Parker MM, et al. RNA‑sequencing across three 42. White MJ, Risse‑Adams O, Goddard P, et al. Novel genetic risk factors for matched tissues reveals shared and tissue‑specific gene expression asthma in African American children: precision medicine and the SAGE II and pathway signatures of COPD. Respir Res. 2019;20(1):65. https ://doi. Study. Immunogenetics. 2016;68(6–7):391–400. https ://doi.org/10.1007/ org/10.1186/s1293 1‑019‑1032‑z.s0025 1‑016‑0914‑1. 39. Halloran JW, Zhu D, Qian DC, et al. Prediction of the gene expression in normal lung tissue by the gene expression in blood. BMC Med Genomics. Publisher’s Note 2015;8:77. https ://doi.org/10.1186/s1292 0‑015‑0152‑7. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 40. Wan ES, Qiu W, Baccarelli A, et al. Systemic steroid exposure is associ‑ lished maps and institutional affiliations. ated with differential methylation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2012;186(12):1248–55. https ://doi. org/10.1164/rccm.20120 7‑1280O C. 41. Ferreira MA, Vonk JM, Baurecht H, et al. Shared genetic origin of asthma, hay fever and eczema elucidates allergic disease biology. Nat Genet. 2017;49(12):1752–7. https ://doi.org/10.1038/ng.3985. Ready to submit your research ? Choose BMC and benefit from: fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Clinical and Translational AllergySpringer Journals

Published: Dec 11, 2020

There are no references for this article.