Skip to main content
PMC home page
PLOS Genetics logoLink to PLOS Genetics
. 2022 May 16;18(5):e1009973. doi: 10.1371/journal.pgen.1009973

Epigenetic and transcriptional dysregulation in CD4+ T cells in patients with atopic dermatitis

Amy A Eapen 1,2,3,#, Sreeja Parameswaran 3,#, Carmy Forney 3, Lee E Edsall 3, Daniel Miller 3, Omer Donmez 3, Katelyn Dunn 3, Xiaoming Lu 3, Marissa Granitto 3, Hope Rowden 3, Adam Z Magier 1, Mario Pujato 3, Xiaoting Chen 3, Kenneth Kaufman 3,4,5,6, David I Bernstein 7, Ashley L Devonshire 1,5, Marc E Rothenberg 1,5, Matthew T Weirauch 3,4,5,*, Leah C Kottyan 1,3,5,*
Editor: Devin M Absher8
PMCID: PMC9135339  PMID: 35576187

Abstract

Atopic dermatitis (AD) is one of the most common skin disorders among children. Disease etiology involves genetic and environmental factors, with 29 independent AD risk loci enriched for risk allele-dependent gene expression in the skin and CD4+ T cell compartments. We investigated the potential epigenetic mechanisms responsible for the genetic susceptibility of CD4+ T cells. To understand the differences in gene regulatory activity in peripheral blood T cells in AD, we measured chromatin accessibility (an assay based on transposase-accessible chromatin sequencing, ATAC-seq), nuclear factor kappa B subunit 1 (NFKB1) binding (chromatin immunoprecipitation with sequencing, ChIP-seq), and gene expression levels (RNA-seq) in stimulated CD4+ T cells from subjects with active moderate-to-severe AD, as well as in age-matched non-allergic controls. Open chromatin regions in stimulated CD4+ T cells were highly enriched for AD genetic risk variants, with almost half of the AD risk loci overlapping AD-dependent ATAC-seq peaks. AD-specific open chromatin regions were strongly enriched for NF-κB DNA-binding motifs. ChIP-seq identified hundreds of NFKB1-occupied genomic loci that were AD- or control-specific. As expected, the AD-specific ChIP-seq peaks were strongly enriched for NF-κB DNA-binding motifs. Surprisingly, control-specific NFKB1 ChIP-seq peaks were not enriched for NFKB1 motifs, but instead contained motifs for other classes of human transcription factors, suggesting a mechanism involving altered indirect NFKB1 binding. Using DNA sequencing data, we identified 63 instances of altered genotype-dependent chromatin accessibility at 36 AD risk variant loci (30% of AD risk loci) that might lead to genotype-dependent gene expression. Based on these findings, we propose that CD4+ T cells respond to stimulation in an AD-specific manner, resulting in disease- and genotype-dependent chromatin accessibility alterations involving NFKB1 binding.

Author summary

Gene expression is regulated in stimulated CD4+ T cells in a disease-dependent manner in patients with atopic dermatitis (AD). AD-specific regions of chromatin accessibility and binding of the NFKB1 transcription factor are enriched for AD genetic risk variants. Clinically, CD4+ T cells in the peripheral blood of patients with AD respond to stimulation in a disease- and genotype-dependent manner.

Introduction

Atopic dermatitis (AD) is one of the most common skin disorders in children, affecting nearly 20% of children worldwide and constituting a significant social and financial burden for affected children and their families [1]. Although AD onset often occurs during childhood, up to 80% of patients with AD have persistent disease continuing into adulthood [2,3]. Currently, patients with moderate-to-severe AD are treated using a “one-size-fits-all” approach, but recent investigations have revealed several different AD endotypes [4]. Both genetic and environmental factors have been implicated in AD pathogenesis [5], and 29 independent AD risk loci have been identified using genome-wide association studies (GWAS) [6,7].

Immunologically, AD involves skin barrier defects and CD4+ T cells that localize to the skin, producing inflammatory cytokines and amplifying epidermal dysfunctions [8]. This can result in allergic sensitization via disruption of the skin barrier, contributing to subsequent development of other allergic diseases along the atopic march, including allergic rhinitis, food allergies, and asthma [9]. Recent studies suggest that early and aggressive management of AD may prevent allergic sensitization and further progression of the atopic march [10–12].

AD genetic risk variants are enriched for genes with genotype-dependent expression (expression quantitative trait loci, eQTLs) in the skin and in CD4+ T cells. This study focused on CD4+ T cells because of their critical role in shaping the immune response in AD and other allergic diseases. Notably, in transcriptional studies on food allergies, the most robust disease-specific expression in CD4+ T cells was detected after stimulation with CD3/CD28 or antigen-loaded dendritic cells [13–15], the two stimulatory pathways that activate the nuclear factor kappa B (NF-κB) protein [16,17]. NF-κB signaling controls the transcription of inflammatory cytokines, such as interleukin-6, and adhesion molecules such as the intercellular adhesion molecule-1, which is responsible for skin inflammation and disruption of the skin barrier [18,19]. An important role for ΝF-κΒ in CD4+ T cells in AD was recently established using a mouse model of AD, in which mice injected with CD4+ T cells with inhibited ΝF-κΒ signaling showed marked improvement in AD-like skin lesions compared to those injected with CD4+ T cells with a control vector [20].

In this study, we hypothesized that that the chromatin accessibility and binding of transcription factors may regulate genetic AD risk loci. To test this hypothesis, we measured chromatin accessibility, NFKB1 binding, and gene expression levels in stimulated CD4+ T cells in subjects with active moderate-to-severe AD as well as in age- and ancestry-matched healthy and non-allergic controls. We identified 34,216 chromatin regions across the genome that were accessible in an AD-dependent manner. These regions were highly enriched for DNA sequence motifs recognized by NF-κB transcription factors (TFs). We performed ChIP-seq for NFKB1 in patients with AD and in control individuals, and identified 20,322 genomic loci with AD-dependent NFKB1 occupancy. Whole genome sequencing of these individuals and the application of our Measurement of Allelic Ratios Informatics Operator (MARIO) method [21] revealed 63 instances of genotype-dependent chromatin accessibility at 36 AD risk variant loci that might lead to genotype-dependent gene expression. Collectively, our findings demonstrate that the pathoetiology of AD involves epigenetic changes in CD4+ T cells via NF-κB-mediated gene expression regulation.

Results

We created a set of 3,143 AD-associated genetic risk variants at 29 independent risk loci (See METHODS and S1 Table). Application of our Regulatory Element Locus Intersection (RELI) method [21] to these variants using expression quantitative trait loci (eQTL) data obtained from the Genotype-Tissue Expression (GTeX) database [22] and the Database of Immune Cell Expression, eQTL, and Epigenomics (DICE) [23] revealed the strongest enrichment for CD4+ T cells and in the skin (sun-exposed and sun-unexposed) (Table 1). This analysis indicates that alteration of gene regulatory mechanisms in CD4+ T cells is likely an important factor underlying AD-associated genetic risk.

Table 1. Expression quantitative trait loci (eQTL) enrichment at atopic dermatitis (AD) risk loci.

Application of our regulatory element locus intersection (RELI) method [21] to 3,143 AD variants across 29 independent risk loci using eQTL data obtained from the Genotype-Tissue Expression (GTeX) database [22] and the Database of Immune Cell Expression, eQTL and Epigenomics (DICE) [23].

Cell line/track Overlap Corrected
P-value		 Enrichment
CD4+ T cell: DICE 6/29 3.85 x 10−5 6.84
Sun-unexposed skin: GTEx 14/29 4.91 x 10−8 4.67
Sun-exposed skin: GTEx 10/29 2.9 x 10−4 4.19
Esophageal mucosa: GTEx 8/29 2.34 x 10−6 4.09
Open in a new tab

We recruited six patients with moderate-to-severe AD (average eczema area and severity index (EASI) score = 30) and six age- and ancestry-matched controls. Participant demographics are presented in Table 2. Adults with persistent AD had experienced disease onset in childhood. As expected, the mean total IgE levels in AD subjects (180.8 kU/L) were higher than those in the controls (61.7 kU/L). Peripheral blood was obtained from each subject and CD4+ T cells were isolated and stimulated for 45 h with anti-CD3/CD28 beads (Fig 1).

Table 2. Demographics of six age-matched AD case and control subjects.

Age
(years)		 Gender
(% male)		 EASI
(0–72)		 IGA
(1–4)		 Age of onset
(years)		 Asthma Allergic rhinitis Food allergy Total IgE Number of sensitizations
Children with active AD
(n = 2)		 10
(8–12)		 1/2
(50%)		 31.9
(31.3–32.4)		 3.5
(3–4)		 4.5
(2–7)		 0 1/2
(50%)		 0 16.5
(6–27)		 3.5
(1–6)
Children without AD
(n = 2)		 13.5
(10–17)		 1/3 (33%) n/a n/a n/a n/a n/a n/a 150.5
(23–278)		 n/a
Adults with active AD
(n = 4)		 44.2
(25–67)		 1/4
(25%)		 29.2
(17.1–50.9)		 3.3
(3–4)		 4
(0.33–10)		 3/4
(75%)		 5/5
(100%)		 3/4
(75%)		 290.3
(18–648)		 5.3
(4–8)
Adults without AD
(n = 4)		 35.3
(18–53)		 0/4
(0%)		 n/a n/a n/a n/a n/a n/a 17.2
(8–36)		 n/a
Open in a new tab

Fig 1. Study Design.

Fig 1

Open in a new tab

Figure created in BioRender.

Global mapping of the chromatin accessibility landscape in AD CD4+ T cells

We performed assay for Transposase-Accessible Chromatin followed by sequencing (ATAC-seq) to map genome-wide chromatin accessibility. The data obtained were of high quality, with an average of almost 70,000 peaks per dataset, an average fraction of reads inside peaks (FRiP) score of 0.32, and an average transcription start site (TSS) enrichment score of 20.5 (S2 Table). Pairwise comparisons of each dataset identified strong agreement between subjects within case and control cohorts (S1A and S1B Fig).

We assessed the overlap of chromatin accessibility data with AD genetic risk variants using the RELI method [21]. Seven of the twelve ATAC-seq datasets were significantly enriched for AD risk loci with 7–14 overlapped risk loci for each subject (RELI pcorrected = 0.01–1.6 x 10−3) (S3 Table).

In pairwise assessments performed using MAnorm [24], most ATAC-seq peaks were shared between AD patients and demographically matched controls (85.9–96.0%). The remaining peaks were either stronger in AD (AD-specific) or control (control-specific) groups (representative analysis in Fig 2A; full results in S2 Fig). We identified 34,216 regions of chromatin across the genome that were accessible in an AD-dependent manner, yielding 357 AD-specific and 343 control-specific peaks that were present in three or more pairs (S3 Fig). We defined these ATAC-seq peaks that were AD-specific or control-specific in three or more subject pairs as “consistently AD-specific” and “consistently control-specific” peaks, respectively. Consistently AD-specific ATAC-seq peaks overlapped AD-associated genetic risk variants at 13 of the 29 AD risk loci (2.0-fold enrichment, pcorrected = 0.015) (S3 Table). As expected, there was substantial overlap between AD-consistent and control-specific ATAC-seq peaks since most chromatin states do not change from individual to individual within a cell type. We identified 5,631 AD-specific consistent ATAC-seq peaks (i.e., consistent peaks that were not consistent in controls). Likewise, we identified 10,058 control-specific consistent ATAC-seq peaks. In total, we identified 357 consistently AD-specific ATAC-seq peaks. Consistently AD-specific ATAC peaks have no overlap consistently control-specific ATAC peaks. These results indicate that chromatin is accessible in a disease-specific manner in CD4+ T cells at almost half of the AD risk loci.

Fig 2. Differential chromatin accessibility and transcription factor (TF) motif enrichment in atopic dermatitis (AD) subjects vs. matched controls.

Fig 2

Open in a new tab

Assay for transposase-accessible chromatin sequencing (ATAC-seq) peaks were identified for all cases and controls and compared for all subject pairs. (A) Differential chromatin accessibility analysis. For each matched pair of subjects, we identified shared, control-specific, and AD-specific peaks (see METHODS). A representative subject pair is shown in (A). Each row represents a single genomic locus where an ATAC-seq peak was identified in either the AD or control subject. The center of each row corresponds to the center of the ATAC-seq peak. Heatmap colors indicate the normalized ATAC-seq read count within the AD1 (right) or CTL1 (control) (left) subject–see key on the right. (B) Differential transcriptional factor (TF) motif enrichment analysis. Comparison of TF motif enrichment results within a representative AD-specific and control-specific matched subject pair. Each dot represents the enrichment of a particular motif (corrected negative log10 p-value). Select motif families are color-coded (see key on the upper left side). (C and D) Nuclear factor kappa B (NFKB) motif enrichment comparison between consistently AD-specific and consistently control-specific ATAC-seq peaks. “Consistently specific” peaks were defined as those peaks that were AD- or control-specific in at least three AD or control subjects, respectively. Results are shown for representative Cis-BP NFKB motif, M05887_2.00. Full results are provided in S5 Table.

To identify potential TFs whose binding might be affected by differential chromatin accessibility, we performed TF binding site motif enrichment analysis on AD- and control-specific ATAC-seq peaks. These analyses revealed that NF-κB DNA-binding motifs were more strongly enriched in the AD-specific ATAC-seq peaks than in the control-specific ATAC-seq peaks in five of the six matched pairs, with the remaining pair showing equivalent enrichment for NF-κB (Figs 2B, S2, and S4 Table). Motif enrichment analysis in consistently AD- and control-specific peaks confirmed that NF-κB binding motifs were highly enriched in an AD-specific manner, with ~17% of consistently AD-specific peaks containing predicted NFKB binding sites (P < 10−13) compared to ~4% of consistently control-specific peaks (P = 1) (Fig 2C and 2D and S5 Table). These data highlight the potential for more robust direct binding of NFKB to the genome in activated CD4+ T cells in the AD patients than in the matched controls.

NFKB1 binds in an AD-specific manner at hundreds of genomic loci in CD4+ T cells

We performed NFKB1 (p50) chromatin immunoprecipitation with sequencing (ChIP-seq) experiments to measure NF-κB binding to the genome within stimulated CD4+ T cells, and obtained an average of approximately 11,000 peaks per subject, with an average FRiP score of 0.012 (S2 Table). All ChIP-seq peak datasets displayed a highly significant overlap with a previously published CD4+ T cell NFKB1 ChIP-seq dataset (GSE126505) (S6 Table). As expected, the NF-κB DNA-binding motif was highly enriched in each of our NF-κB ChIP-seq datasets (Cis-BP NF-κB motif M05887_2.00 enrichment: 10−4158 < P < 10−123 (S7 Table)). We identified 20,322 genomic loci with AD-dependent NFKB1 occupancy. AD-specific NFKB1 ChIP-seq peaks were enriched for overlap with AD-specific ATAC-seq peaks in all six pairs (between 5.9 and 38.1-fold enrichment, 3.5 × 10−25 < P < 3.20 x 10−203) (S8 Table, reporting AD-specific ATAC and AD-specific ChIP). A global variability analysis of ATAC and ChIP peaks revealed minimal differences in variability within cases versus within controls. Substantially more variability was observed between subject matched pairs in the NFKB1 ChIP-seq experiments than in the ATAC-seq experiments, with a median of 51.5% shared NFKB1 peak as compared to a median of 91.9% shared ATAC-seq peaks (Fig 3). These results indicate substantially more differential NFKB1 binding than chromatin accessibility in AD subjects compared to that observed in matched controls.

Fig 3. Shared peaks in ATAC-seq and chromatin immunoprecipitation with sequencing (ChIP-seq) experiments between paired subjects.

Fig 3

Open in a new tab

Percentage of shared peaks between paired samples in ATAC-seq compared to NFKB1 ChIP-seq experiments.

We next sought to identify AD- and control-specific NFKB1 binding events by performing a pairwise assessment of NFKB1 peaks in the case vs. control subjects (see METHODS). This procedure identified shared, control-specific, and AD-specific NFKB1 peaks. A representative pair is shown in Fig 4A, and all pairs are shown in S4 Fig. Strikingly, NF-κB binding sites were more strongly enriched in the AD-specific NFKB1 ChIP-seq peaks than in the matched controls in five pairs (S4 Fig and S7 Table (a representative pair is shown in Fig 4B). We defined NFKB1 peaks that were AD-specific and/or control-specific in three or more subject pairs as “consistently AD-specific” and “consistently control-specific” peaks, respectively (S3C and S3D Fig). In total, we identified 131 AD-specific and 72 control-specific NFKB1 ChIP-seq peaks. Motif enrichment analysis revealed that NFKB binding motifs were also the most highly enriched motif class within the consistently AD-specific NFKB1 peaks (Fig 4C, 4D and S9 Table). In contrast, consistently control-specific peaks were not enriched for NFKB motifs, but instead were enriched for a wide range of other motif classes (S5A and S5B Fig and S9 Table). Collectively, these results indicate that AD-specific NFKB ChIP-seq peaks are strongly enriched for NFKB1 motifs, whereas control-specific NFKB peaks surprisingly are not.

Fig 4. Differential NFKB1 binding and TF motif enrichment in AD subjects vs. matched controls.

Fig 4

Open in a new tab

NFKB1 ChIP-seq peaks were identified for all case and control subjects and compared between matched subject pairs. (A) Differential NFKB1 binding analysis. For each matched pair of subjects, we identified the shared peaks, control-specific peaks, and AD-specific peaks (see METHODS). A representative subject pair is shown in (A). Each row represents a single genomic locus, where an NFKB1 ChIP-seq peak was identified in either the AD or control subject. The center of each row corresponds to the center of the ChIP-seq peak. Heatmap colors indicate the normalized ChIP-seq read counts within the AD2 (right) or CTL2 (left) subject–see key on the right. (B) Differential TF motif enrichment analysis. Comparison of TF motif enrichment results within a representative AD-specific and control-specific matched subject pair. Each dot represents the enrichment of a particular motif (corrected negative log10 p-value). Select motif families are color-coded (see key on the upper right side). (C and D) NFKB motif enrichment comparison between consistently AD-specific and consistently control-specific NFKB1 ChIP-seq peaks. “Consistently specific” peaks were defined as those peaks that were AD- or control-specific in at least three case or control subjects, respectively. Results are shown for representative Cis-BP NF-κB motif, M05887_2.00. All results are listed in S7 and S9 Tables.

Comparisons across different data types confirm the strong concordance among RNA-, ATAC-, and NFKB1 ChIP-seq results

Next, we measured gene expression levels in CD3/CD28-stimulated CD4+ T cells in each subject, with the goal of integrating these data with the chromatin accessibility and NFKB1 binding data. In case-control pairwise analysis, 15 genes were expressed at least 1.5-fold higher in the stimulated CD4+ T cells from patients with AD compared to the matched controls, while 16 genes were expressed 1.5-fold lower in the case group compared to the matched control group (S10 Table). These 31 genes were enriched for AD-related processes, such as the “regulation of immune system processes”, "lymphocyte activation”, and “cytokine-mediated signaling pathway” gene ontology (GO) biological processes as well as the “cytokine receptor binding”, “nitric oxide synthase binding”, and “RNA-polymerase II-specific DNA-binding transcription factor binding” GO molecular functions (S6 Fig).

The 100 kB region of DNA around AD-specific gene sets widely overlapped (94.7–100%) the ATAC-seq peaks in the six subjects with AD (S8 Table). There was substantial overlap (26.3–68.4%) between the 100 kb region of DNA around AD-specific gene sets with the AD-specific ATAC-peaks (Fig 5), indicating that possible enhancers proximal to the AD-specific genes were accessible for transcription in an AD-specific manner. Similarly, the 100 kb region of DNA around AD-specific NFKB1 ChIP-seq peaks overlapped the transcriptional start site of 47–95% of the AD-specific genes (S8 Table and Fig 5). In five of the six pairs, AD-specific NFKB1 ChIP-seq peaks overlapped a large proportion of the AD-specific genes (42.1–73.4%) (S8 Table). The 100 kb region around NFKB1 ChIP-seq peaks specific in AD overlapped 44.3% of the genes with increased expression in AD compared to the 15.4% of genes with increased expression in AD that were overlapped by 100 kb region around NFKB1 ChIP-seq peaks specific for controls (Fig 5). Importantly, there is AD specificity to the overlap of NFKB1 ChIP-seq peaks with AD-specific genes with AD-specific NFKB1 peaks overlapping more AD-specific genes (~80%) while control specific NFKB1 peaks overlapped only ~35% of AD-specific genes (Fig 5). These results are consistent with NFKB acting as both an activator and repressor depending on the context [25,26]. Collectively, these data indicate strong agreement between AD- and control-specific gene expression, chromatin accessibility, and NFKB1 binding.

Fig 5. Overlap of genes expressed in an AD-dependent manner with Control-specific and AD-specific ATAC-seq and ChIP-seq peaks.

Fig 5

Open in a new tab

AD-dependent genes can be upregulated in AD (Up in AD) or downregulated in AD (Down in AD). Overlap with of these gene sets with Control-specfic and AD-specific ATAC and ChIP-seq peaks is shown.

Allele-dependent chromatin accessibility at AD risk loci

Accumulating evidence reveals the important roles of allele-dependent gene regulatory mechanisms in various diseases [21,27–29]. To identify such mechanisms, we performed whole-genome sequencing of all subjects to identify the alleles present in AD genetic risk variants (S11 Table). We integrated these data with the functional genomics data produced in this study using the MARIO method, which measures the allele dependence of sequencing reads at genetic variants that are heterozygous [21]. Collectively, there were an average of 2.3 (range 0–5) heterozygous loci in AD cases and 2.6 (range 0–6) heterozygous loci among the controls (S11 Table), providing 124 opportunities to discover allele-dependent ATAC-seq or NFKB1 peaks at AD genetic risk variants.

Sixty AD-associated variants were located within an ATAC-seq peak in at least one subject and were found to be heterozygous in that subject. Notably, 36 of these 60 (60%) variants produced allele-dependent ATAC-seq peaks (S12 Table and Fig 6). 24 AD risk variants in 17 haplotypes overlap ATAC-peaks consistently found in subjects with AD. Of these, five overlap ATAC-seq peaks consistently found in subjects with AD but not consistently found in controls. Collectively, AD risk variants with allelic chromatin accessibility were found at nine independent risk loci (31% of AD risk loci). In 16 of the AD risk variants that were heterozygous and overlapped with ATAC-seq peaks, we observed allelic imbalance in multiple subjects. For example, at rs10791824 near the ovo-like transcriptional repressor 1 (OVOL1) gene, we measured a strong preference for the A allele across six individuals, with 85–100% of reads for all subjects having the A allele (total of 112 vs. 10, A vs. T reads). Twenty-eight AD risk variants with allelic ATAC-seq peaks were found to be eQTLs in stimulated CD4+ T cells based on DICE, as curated by the eQTL catalogue [30] (S13 Table); however, these associations of allelic expression were not robust enough to remain after multiple testing correction after accounting for all of the eQTL measurements in that study (i.e., across many cell types with and without stimulation).

Fig 6. Allele-dependent chromatin accessibility at AD risk loci.

Fig 6

Open in a new tab

(A) AD-associated genetic risk variants with allele-dependent ATAC-seq peaks in CD4+ T cells. Each variant is heterozygous and located within an ATAC-seq peak in the indicated individual, facilitating measurement of allelic ratios informatics operator (MARIO) analysis to identify allele-dependent behavior. Red tick marks indicate allelic accessibility in a subject who with AD. Complete data are presented in S12 Table. All results shown have MARIO allelic reproducibility score (ARS) values ≥ 0.4, and hence, are allele-dependent. In the cutout, the participant identifier and reads under the ATAC-seq peak overlapping rs10791824 mapping to the strong and weak bases are provided.

In contrast to the large number of observed allelic ATAC-seq peaks, only six unique AD-associated variants were located within at least one NFKB1 ChIP-seq peak and also heterozygous in any of the 12 subjects; moreover, none of them demonstrated genotype-dependent activity. Hence, future studies examining larger cohorts are required to identify allelic NFKB1 binding activity.

Discussion

Taken together, our data support a model in which stimulated peripheral blood CD4+ T cells from patients with active AD exhibit extensive differential chromatin accessibility and NFKB1 binding compared to their matched controls. We identified genotype-dependent chromatin accessibility for 36 AD genetic variants. Collectively, this study highlights plausible genetic risk mechanisms for AD via disease-specific epigenetic factors enriched at AD genetic risk loci. Current databases of eQTLs are not specific to patients with AD or other allergic diseases, and do not contain a sufficient number of participants to have the statistical power to allow the identification of moderate-sized eQTLs. Therefore, it is important to continue collecting control and AD-specific eQTL datasets. It is also important to perform molecular studies assessing genotype-dependent regulatory activity of AD risk variants in the context of CD4+ T cells to definitively identify allelic transcriptional mechanisms at these loci.

This study had limitations. We used strong stimulation (CD3/CD28 cross-linking) to model immune activation in patients. Use of different levels of stimulation may reveal additional disease-specific differences in future studies. We used CD4+ T cells in our functional genomic assays to quantitatively assess disease-specific and genotype-dependent mechanisms. The disease- and genotype-dependent effects observed in this study can be refined and validated in specific immune cell subsets as the technology and analytical framework for quantitative comparisons at the single-cell level are refined. This study used a paired analysis of cases with demographically matched controls. The findings of this manuscript represent epigenetic differences that are consistent across these matched pairs. Future studies will be needed in which ancestry and age are carefully matched for all cases and controls to support a pooled analytical strategy. This study used a paired analysis of cases with demographically matched controls. The findings of this manuscript represent epigenetic differences that are consistent across these matched pairs. Future studies will be needed in which age is carefully matched for all cases and controls to support a pooled analytical strategy. The paucity of AD-associated NFKB1 binding variants could in large part be due to the lower number of NFKB1 ChIP-seq peaks.

Differences in the subtypes of CD4+ T cells in the peripheral blood of patients with moderate-to-severe AD might explain some of the differential functional genomic effects identified in this study. AD patients exhibit an expansion of the T helper (Th)-9 subset of CD4+ T cells and an increased frequency of circulating CD25hiFoxp3+ T cells compared to controls [31–33]. Skin-homing Th22 T cell levels are also increased in patients with AD across all ages [32]. Additionally, the T cell profile naturally changes with age, and patients with AD display a unique profile of changes [34]. Natural aging leads to both quantitative and qualitative changes in the CD4+ T-cell compartment [35]. Although our study did not differentiate between the different CD4+ T cell subtypes, our case-control comparisons were performed between demographically matched controls. Future studies should investigate the role of these different subtypes based on the age and prevalence of genotype-dependent transcriptional dysregulation in AD patients compared to controls.

This study is important for the understanding of atopic dermatitis because a mechanistic understanding of the etiology of the disease could enable efficacious predictive tools and preventative therapeutics. Because atopic dermatitis is the first diagnosis of most patients who find themselves on the atopic march from AD to food allergy to asthma to allergic rhinitis, strategies to prevent and halt AD have the potential to prevent other atopic diseases. Current treatments for AD include the use of topical steroids and anti-inflammatory non-steroidal agents, such as anti-type 2 immunity biological agents and Janus kinase inhibitors, as well as aggressive moisturization and systemic immunosuppression in severe cases. There is an urgent need for novel therapeutics to address the complexity of this disease and to contribute to the development of personalized medicine against various subtypes/phenotypes that differ by age, disease chronicity, and underlying molecular mechanisms [4,36]. The results of this study support a continued focus on the development of therapeutics aimed at inhibiting the NF-κB signaling pathway. Numerous studies in mice and humans provide further impetus to elucidate this mechanism in AD [37,38].

In conclusion, the results of this study support a model in which stimulated CD4+ T cells from patients with AD display disease- and allele-dependent differences in chromatin accessibility and disease-dependent differences in NFKB1 binding. Based on the broad genotype-dependent chromatin accessibility at AD risk variants in stimulated CD4+ T cells, our data support allelic transcriptional regulation as an important epigenetic mechanism mediating disease risk in AD.

Methods

Ethics statement

This study was approved by the Cincinnati Children’s Hospital Institutional Regulatory Board (IRB #2018–4254). Prior to participation, all adults, children over 13 years of age, and parents of children below 13 years of age provided written informed consent; children 13 years and older provided assent for their participation.

Collection of AD-associated genetic risk variants

One hundred and twenty-two genetic variants that reached genome-wide significance in a GWAS of AD were identified from the Genome Wide Association Catalogue (https://www.ebi.ac.uk/gwas/) [39] and a genetic association study using the Illumina ImmunoChiP [40]. Twenty-nine independent genetic risk loci were identified via linkage disequilibrium pruning (r2 < 0.2). We identified 3,143 AD risk variants across these 29 loci by accounting for linkage disequilibrium (r2 > 0.8) based on 1000 Genomes Data [41] in the ancestry(ies) of the initial genetic association using PLINK (v.1.90b) [42]. S1 Table presents the twenty-nine AD risk loci that reached genome-wide significance at the beginning of this study; this supplemental table includes information regarding the major, minor, reference, non-reference, and risk alleles as well as allele frequencies, disease odds ratios, and additional associated diseases and phenotypes.

Patient recruitment

Patients with moderate-to-severe AD were recruited for this study from the Cincinnati Children’s Hospital Allergy Clinics, the Bernstein Allergy Group, and the Bernstein Clinical Research Center.

Matched healthy controls were recruited via advertisements. To reduce heterogeneity, the study inclusion criteria for subjects with AD were: 1) presence of atopy established by positive aeroallergen skin prick test and/or elevated serum total IgE levels; and 2) moderate-to-severe AD defined as an EASI score ≥ 17 and investigator global assessment score ≥ 3 [43]. These AD severity tools have previously been validated for AD severity scoring [43,44]. Control subjects were included if they had no history of atopic disease and displayed negative results from aeroallergen skin prick testing performed at their enrollment visit. Exclusion criteria included a history of any biologic therapy, oral steroids, or immunosuppressive medications in the past 6 months due to their effect on T cells and the transcriptome. S13 Table provides the age, sex, self-reported race and ethnicity of each subject in this study.

CD4+ T cell stimulation

We isolated the peripheral blood mononuclear cells (PBMCs) from AD and control individuals using Ficoll-Paque (GE Healthcare) density gradient separation. Then, 51.5 ± 20.1 million PBMCs were isolated from each participant. Whole genome sequencing of the DNA extracted from PBMCs was performed to identify the genetic variants. CD4+ T cells were isolated using magnetic column separation (Miltenyi Biotec; CD4+ T cell isolation kit, human); 17.7 ± 7.2 million CD4+ T cells were isolated from each subject. To activate NF-κB, we stimulated CD4+ T cells with CD3/CD28 crosslinking for 45 h (Gibco; Dynabeads Human T-Activator CD3/CD28). We performed ChIP-seq (2 million stimulated CD4+ T cells), ATAC-seq (50,000 stimulated CD4+ T cells), and RNA-seq assays (2.5 million stimulated CD4+ T cells). The detailed methods for each genetic and genomic assay are provided below.

ATAC-seq

Transposase Tn5 with adapter sequences was used to cut the accessible DNA [45]. These accessible DNA sequences with adaptor sequences were isolated, and libraries were prepared from 50,000 stimulated CD4+ T cells using the OMNI ATAC protocol [46]. The libraries were sequenced at 150 bases per end on an Illumina NovaSeq 6000 at the Cincinnati Children’s Hospital Medical Center (CCHMC) DNA Sequencing and Genotyping Core Facility. The quality of the sequencing reads was verified using FastQC (v. 0.11.2) (http://www.bioinformatics.babraham.ac.uk/projects/fastqc), and adapter sequences were removed using Cutadapt (Trimgalore v. 0.4.2). ATAC-seq reads were aligned to the human genome (hg19) using Bowtie2 [47]. Aligned reads were then sorted using samtools (v.1.8) [48] and duplicate reads were removed using Picard (v. 1.89) (https://broadinstitute.github.io/picard/). Peaks were called using MACS2 (macs2 callpeak -g hs -q 0.01) [49]. ENCODE blacklist regions (https://github.com/Boyle-Lab/Blacklist/tree/master/lists/hg19-blacklist.v2.bed.gz) [50] were removed. Differential chromatin accessibility was calculated using the MAnorm program [24] with thresholds of fold change greater than 1.5 and p-value less than 0.05.

ChIP-seq

CD4+ T cells from subjects were crosslinked, and nuclei were sonicated. Cells were incubated in a crosslinking solution (1% formaldehyde, 5 mM 4-(2-hydroxyethyl)-1-piperazineëthanesulfonic acid (HEPES) [pH 8.0], 10 mM sodium chloride, 0.1 mM ethylenediaminetetraacetic acid (EDTA), and 0.05 mM ethylene glycol tetraacetic acid (EGTA) in Roswell Park Memorial Institute (RPMI) culture medium with 10% fetal bovine serum (FBS)) and placed on a tube rotator at room temperature for 10 min. To stop crosslinking, glycine was added to a final concentration of 0.125 M, and the tubes were rotated at room temperature for 5 min. Cells were washed twice with ice-cold phosphate-buffered saline (PBS), resuspended in lysis buffer 1 (50 mM HEPES [pH 8.0], 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.25% Triton X-100, and 0.5% NP-40), and incubated for 10 min on ice. Nuclei were harvested after centrifugation at 5,000 rpm for 10 min, resuspended in lysis buffer 2 (10 mM Tris-HCl [pH 8.0], 1 mM EDTA, 200 mM NaCl, and 0.5 mM EGTA), and incubated at room temperature for 10 min. Protease and phosphatase inhibitors were added to both lysis buffers. Nuclei were resuspended in sonication buffer (10 mM Tris [pH 8.0], 1 mM EDTA, and 0.1% sodium dodecyl sulfate (SDS)). An S220 focused ultrasonicator (COVARIS) was used to shear chromatin (150–500-bp fragments) with 10% duty cycle, 175 peak power, and 200 bursts per cycle for 7 min. A portion of the sonicated chromatin was run on an agarose gel to verify fragment sizes. Sheared chromatin was pre-cleared with 10 μL of Protein G Dynabeads (Life Technologies) at 4°C for 1 h.

Immunoprecipitation of NFKB1-chromatin complexes was performed using an SX-8X IP-STAR compact automated system (Diagenode). Beads conjugated to antibodies against NFKB1 (Cell Signaling (D7H5M) Rabbit mAb #12540) were incubated with pre-cleared chromatin at 4°C for 8 h. The beads were then washed sequentially with buffer 1 (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 0.1% SDS, 0.1% NaDOC, and 1% Triton X-100), buffer 2 (50 mM Tris-HCl [pH 7.5], 250 mM NaCl, 1 mM EDTA, 0.1% SDS, 0.1% NaDOC, and 1% Triton X-100), buffer 3 (2 mM EDTA, 50 mM Tris-HCl [pH 7.5] and 0.2% sarkosyl sodium salt), and buffer 4 (10 mM Tris-HCl [pH 7.5], 1 mM EDTA, and 0.2% Triton X-100). Finally, the beads were resuspended in 10 mM Tris-HCl (pH 7.5) and used to prepare the libraries via ChIPmentation [51].

The ChIP-seq libraries were sequenced as single-end 100-base reads on an Illumina NovaSeq 6000 at the CCHMC DNA Sequencing and Genotyping Core Facility. The reads were processed and analyzed as described above for ATAC-seq. We also used publicly available NFKB1 ChIP-seq datasets (GSE126505) that were processed using the same analytical pipeline.

RNA-seq

Total RNA was extracted using the mirVANA Isolation Kit (Ambion) from stimulated CD4+ T cells of the control and AD subjects 45 h post-stimulation. RNA-seq libraries were sequenced as paired ends with 150 bases. FastQC and cutadapt were used to verify the read quality and remove adapters, respectively, as described above. RNA-seq reads were aligned to the hg19 (GrCh37) genome build (National Center for Biotechnology Information, NCBI) using Spliced Transcript Alignment with a Reference (STAR, v. 2.5.2a) [52]. The program featureCounts (subread/1.6.2) was used to count the reads mapped to each gene [53]. The fragments per kilobase per million mapped fragments (FPKM) values for the relative expression of each gene were used to calculate the pairwise AD case/control fold-change. Differential expression for pairwise subject comparisons was established as a fold-change greater than 1.5.

Whole genome sequencing and variant calling

DNA was isolated using a PureLink Genomic DNA Kit (Thermo Fisher). Whole-genome sequencing was performed using DNBseq next-generation sequencing technology. Libraries were sequenced on an Illumina NovaSeq to generate 100-base paired-end reads. Sequencing reads were aligned, and variants were identified using the Genome Analysis Toolkit (GATK) Unified Genotyper following GATK Best Practices 3.3 [54–56].

Enrichment analysis for functional genomic datasets

We used the RELI algorithm to estimate the significance of the overlap between genomic features generated in this study (TF binding events, histone marks, ATAC-seq peaks, etc.) [21,57,58]. As input, RELI takes the genomic coordinates of peaks from two datasets. RELI then systematically intersects these coordinates with one another and the number of input regions overlapping the peaks is counted. Next, a p-value describing the significance of this overlap is estimated using a simulation-based procedure in which the peaks from the first dataset are randomly distributed within the union coordinates of open chromatin from human cells. A distribution of expected overlap values is then created from 2,000 iterations of randomly sampling from the negative set, each time choosing a set of negative examples that match the input set in terms of the total number of genomic loci. The distribution of the expected overlap values from the randomized data resembles a normal distribution and can thus be used to generate a Z-score and corresponding p-value estimating the significance of the observed number of input regions that overlap each dataset.

Identification of allelic ATAC-seq and ChIP-seq reads using MARIO

To identify possible allele-dependent mechanisms in our functional genomics datasets, we applied our MARIO method [21]. In cases where a given variant is heterozygous in the cell assayed, both alleles are available for the TF to bind or for the chromatin to be accessible or not, offering a natural control for one another since the only variable that has changed is the allele. In brief, MARIO identifies common genetic variants that are (1) heterozygous in the assayed cell line (using NGS DNA sequencing data) and (2) located within a peak in a given ChIP-seq or ATAC-seq dataset. MARIO then examines the sequencing reads that map to each heterozygote within each peak for imbalance between the two alleles. To estimate the significance of the degree of allelic imbalance of a given dataset at a given heterozygote, we developed the MARIO Allelic Reproducibility Score (ARS), which is based on a combination of two predictive variables: the total number of reads overlapping the variant and the imbalance between the number of reads for each allele. We report allelic accessibility and NFKB1 binding at AD genetic risk variants in our ATAC-seq data with an ARS greater than or equal to 0.4 which is considered significantly allelic [21].

Supporting information

S1 Fig. ATAC-seq and NFKB1 ChIP-seq peak pairwise normalized read count comparisons.

Each dot represents a peak that is present in either of the two samples. The X- and Y-axis indicate the number of normalized read counts for that peak in the two samples. The red line indicates peaks with similar normalized read counts between the two samples. A. ATAC- seq scatter plot analysis for controls. B. ATAC-seq scatter plot analysis for AD cases. C. ChIP-seq scatter plot analysis for controls. D. ChIP-seq scatter plot analysis for AD cases. E. Global Variability analysis of the ATAC and ChIP peaks across AD and controls.

(PDF)

Click here for additional data file. (663.1KB, pdf)
S2 Fig. Case/control pairwise differential ATAC-seq analysis and transcription factor motif enrichment.

(Left plots). For each of the six matched sample pairs, differential ATAC- seq peak analysis was performed. ATAC-seq peaks that are equally strong between case and controls (“shared”), significantly stronger in the control (“control-specific”), or significantly stronger in the case (“AD-specific”) were identified using MAnorm (see Methods). The percent of peaks falling within each category is indicated to the left of each heatmap. Heatmap color indicates the normalized strength of the ATAC-seq signal. (Right plots). Transcription factor motif enrichment analysis was performed for each sample pair (see Methods). Each dot represents a transcription factor motif. The X-axis and Y-axis indicate the p-value of enrichment within control-specific and AD-specific ATAC-seq peaks, respectively. Motif family color key is provided at the top of each plot.

(PDF)

Click here for additional data file. (722.2KB, pdf)
S3 Fig. Number of consistently differential ATAC-seq and ChIP-seq peaks across sample pairs.

ATAC-seq and ChIP-seq peaks were called for all cases and controls. For each matched pair of samples, a MAnorm analysis identified shared peaks, control-specific peaks and AD-specific peaks (see Methods). Bar plots indicate the distributions of the number of control- specific (left) or case-specific peaks (right) for ATAC-seq (A,B) and ChIP-seq (C,D). For example, there are approximately 10 ATAC-seq peaks in the genome that are control-specific in 4 of the 6 matched pairs (see panel A).

(PDF)

Click here for additional data file. (331.5KB, pdf)
S4 Fig. Case/control pairwise differential NKFB1 ChIP-seq analysis and transcription factor motif enrichment.

(Left plots). For each of the six matched sample pairs, differential NKFB1 ChIP-seq peak analysis was performed. ChIP-seq peaks that are equally strong between case and controls (“shared”), significantly stronger in the control (“control-specific”), or significantly stronger in the case (“AD-specific”) were identified using MAnorm (see Methods). The percent of peaks falling within each category is indicated to the left of each heatmap. Heatmap color indicates the normalized strength of the ChIP-seq signal. (Right plots). Transcription factor motif enrichment analysis was performed for each sample pair (see Methods). Each dot represents a transcription factor motif. The X-axis and Y-axis indicate the p-value of enrichment within control- specific and AD-specific ChIP-seq peaks, respectively. Motif family color key is provided at the top of each plot.

(PDF)

Click here for additional data file. (607.7KB, pdf)
S5 Fig. Transcription factor motif enrichment comparison between consistently control-specific and consistently AD-specific NFKB1 ChIP-seq peaks.

The top five enriched TF motif families are shown. A. Percent of peaks containing predicted binding sites for the indicated motif. B. Motif enrichment p-value within those peaks (see Methods).

(PDF)

Click here for additional data file. (192.2KB, pdf)
S6 Fig. RNA-seq GO enrichment results.

Pairwise differential gene expression analysis was performed by comparing RNA-seq data for each of the cases with their matched controls (see Methods). Genes with 1.5-fold changes or greater in three or more pairs were identified. Pathway enrichment analysis was performed on these gene sets. The top five enriched GO biological pathways and GO molecular function categories are indicated within the top and bottom tables, respectively.

(PDF)

Click here for additional data file. (20.6KB, pdf)
S1 Table. Atopic Dermatitis risk variants and loci.

(XLSX)

Click here for additional data file. (68.6KB, xlsx)
S2 Table. Quality control summary.

(XLSX)

Click here for additional data file. (16.7KB, xlsx)
S3 Table. Regulatory Element Locus Intersection (RELI) analysis of ATAC-seq and ChIP datasets.

(XLSX)

Click here for additional data file. (16.7KB, xlsx)
S4 Table. Transcription factor motif analysis of ATAC-seq datasets.

(XLSX)

Click here for additional data file. (6.3MB, xlsx)
S5 Table. Transcription factor motif analysis of ATAC-seq peaks consistently found in samples from patients with atopic dermatitis and controls.

(XLSX)

Click here for additional data file. (855.7KB, xlsx)
S6 Table. Regulatory Element Locus Intersection (RELI) analysis of NFKB1 ChIP-seq datasets.

(XLSX)

Click here for additional data file. (11.8KB, xlsx)
S7 Table. Transcription factor motif analysis of NFKB1 ChIP-seq datasets.

(XLSX)

Click here for additional data file. (6MB, xlsx)
S8 Table. Regulatory Element Locus Intersection (RELI) integration of RNA-seq, ATAC-seq, and NFKB1 ChIP-seq datasets.

(XLSX)

Click here for additional data file. (21.4KB, xlsx)
S9 Table. Transcription factor motif analysis of NFKB1 ChIP-seq peaks consistently found in samples from patients with atopic dermatitis and controls.

(XLSX)

Click here for additional data file. (813.9KB, xlsx)
S10 Table. Analysis of RNA-seq datasets.

(XLSX)

Click here for additional data file. (178.5KB, xlsx)
S11 Table. Identification of AD risk variants in DNA-seq datasets.

(XLSX)

Click here for additional data file. (20.6KB, xlsx)
S12 Table. Identification of genotype-dependent.

(XLSX)

Click here for additional data file. (18.5KB, xlsx)
S13 Table. Demographics of participants in this study.

(DOCX)

Click here for additional data file. (13.7KB, docx)

Acknowledgments

We would like to thank the following physicians for their support in establishing our research clinic: Juan Pablo Abonia, MD; Jonathan Bernstein, MD; Sheharyar Durrani, MD; Stephanie Ward, MD; Justin Greiwe, MD; Michelle Lierl, MD; and Kimberly Risma, MD, PhD. We appreciate the support Bahram Namjou, MD and Andrew VonHandorf, PhD. Some figures were created in BioRender.

Data Availability

All raw and processed sequencing data generated in this study have been submitted to the NCBI Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) database under accession number GSE184238. A UCSC Genome Browser session is available at http://genome.ucsc.edu/s/ledsall/AtopicDermatitis.

Funding Statement

This study was funded in part by National Institutes of Health (NIH) through R01 DK107502, R01 AI148276, U19 AI070235, R01 AI130830, U01 HG011172, and P30 AR070549 to LCK. NIH funding awarded to MW that contributed to this study includes R01 HG010730, R01 NS099068, R01 GM055479, and U01 AI130830. NIH awards R01 AR073228 and R01 AI024717 to MW and LCK also contributed to this study. The Cincinnati Children’s Hospital Medical Center ARC Award 53632 to MW and LCK further contributed to this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.Paller AS, Spergel JM, Mina-Osorio P, Irvine AD. The atopic march and atopic multimorbidity: Many trajectories, many pathways. J Allergy Clin Immunol. 2019;143(1):46–55. Epub 2018/11/21. doi: 10.1016/j.jaci.2018.11.006 . [DOI] [PubMed] [Google Scholar]
  • 2.Silverberg JI, Hanifin JM. Adult eczema prevalence and associations with asthma and other health and demographic factors: a US population-based study. J Allergy Clin Immunol. 2013;132(5):1132–8. Epub 2013/10/08. doi: 10.1016/j.jaci.2013.08.031 . [DOI] [PubMed] [Google Scholar]
  • 3.Margolis JS, Abuabara K, Bilker W, Hoffstad O, Margolis DJ. Persistence of mild to moderate atopic dermatitis. JAMA Dermatol. 2014;150(6):593–600. Epub 2014/04/04. doi: 10.1001/jamadermatol.2013.10271 ; PubMed Central PMCID: PMC4352328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Czarnowicki T, He H, Krueger JG, Guttman-Yassky E. Atopic dermatitis endotypes and implications for targeted therapeutics. J Allergy Clin Immunol. 2019;143(1):1–11. Epub 2019/01/08. doi: 10.1016/j.jaci.2018.10.032 . [DOI] [PubMed] [Google Scholar]
  • 5.Otsuka A, Nomura T, Rerknimitr P, Seidel JA, Honda T, Kabashima K. The interplay between genetic and environmental factors in the pathogenesis of atopic dermatitis. Immunol Rev. 2017;278(1):246–62. Epub 2017/06/29. doi: 10.1111/imr.12545 . [DOI] [PubMed] [Google Scholar]
  • 6.Liang Y, Chang C, Lu Q. The Genetics and Epigenetics of Atopic Dermatitis-Filaggrin and Other Polymorphisms. Clin Rev Allergy Immunol. 2016;51(3):315–28. Epub 2015/09/20. doi: 10.1007/s12016-015-8508-5 . [DOI] [PubMed] [Google Scholar]
  • 7.Brown SJ. What Have We Learned from GWAS for Atopic Dermatitis? J Invest Dermatol. 2021;141(1):19–22. Epub 2020/06/12. doi: 10.1016/j.jid.2020.05.100 . [DOI] [PubMed] [Google Scholar]
  • 8.Werfel T, Allam JP, Biedermann T, Eyerich K, Gilles S, Guttman-Yassky E, et al. Cellular and molecular immunologic mechanisms in patients with atopic dermatitis. J Allergy Clin Immunol. 2016;138(2):336–49. Epub 2016/08/09. doi: 10.1016/j.jaci.2016.06.010 . [DOI] [PubMed] [Google Scholar]
  • 9.Davidson WF, Leung DYM, Beck LA, Berin CM, Boguniewicz M, Busse WW, et al. Report from the National Institute of Allergy and Infectious Diseases workshop on "Atopic dermatitis and the atopic march: Mechanisms and interventions". J Allergy Clin Immunol. 2019;143(3):894–913. Epub 2019/01/15. doi: 10.1016/j.jaci.2019.01.003 ; PubMed Central PMCID: PMC6905466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Horimukai K, Morita K, Narita M, Kondo M, Kitazawa H, Nozaki M, et al. Application of moisturizer to neonates prevents development of atopic dermatitis. J Allergy Clin Immunol. 2014;134(4):824–30 e6. Epub 2014/10/06. doi: 10.1016/j.jaci.2014.07.060 . [DOI] [PubMed] [Google Scholar]
  • 11.Simpson EL, Chalmers JR, Hanifin JM, Thomas KS, Cork MJ, McLean WH, et al. Emollient enhancement of the skin barrier from birth offers effective atopic dermatitis prevention. J Allergy Clin Immunol. 2014;134(4):818–23. Epub 2014/10/06. doi: 10.1016/j.jaci.2014.08.005 ; PubMed Central PMCID: PMC4180007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Skjerven HO, Rehbinder EM, Vettukattil R, LeBlanc M, Granum B, Haugen G, et al. Skin emollient and early complementary feeding to prevent infant atopic dermatitis (PreventADALL): a factorial, multicentre, cluster-randomised trial. Lancet. 2020;395(10228):951–61. Epub 2020/02/23. doi: 10.1016/S0140-6736(19)32983-6 . [DOI] [PubMed] [Google Scholar]
  • 13.Martino D, Neeland M, Dang T, Cobb J, Ellis J, Barnett A, et al. Epigenetic dysregulation of naive CD4+ T-cell activation genes in childhood food allergy. Nat Commun. 2018;9(1):3308. Epub 2018/08/19. doi: 10.1038/s41467-018-05608-4 ; PubMed Central PMCID: PMC6098117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Martino DJ, Bosco A, McKenna KL, Hollams E, Mok D, Holt PG, et al. T-cell activation genes differentially expressed at birth in CD4+ T-cells from children who develop IgE food allergy. Allergy. 2012;67(2):191–200. Epub 2011/11/15. doi: 10.1111/j.1398-9995.2011.02737.x . [DOI] [PubMed] [Google Scholar]
  • 15.Lindstedt M, Schiott A, Bengtsson A, Larsson K, Korsgren M, Greiff L, et al. Genomic and functional delineation of dendritic cells and memory T cells derived from grass pollen-allergic patients and healthy individuals. Int Immunol. 2005;17(4):401–9. Epub 2005/03/05. doi: 10.1093/intimm/dxh220 . [DOI] [PubMed] [Google Scholar]
  • 16.Cheng J, Montecalvo A, Kane LP. Regulation of NF-kappaB induction by TCR/CD28. Immunol Res. 2011;50(2–3):113–7. Epub 2011/07/01. doi: 10.1007/s12026-011-8216-z ; PubMed Central PMCID: PMC4066383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hayden MS, West AP, Ghosh S. NF-kappaB and the immune response. Oncogene. 2006;25(51):6758–80. Epub 2006/10/31. doi: 10.1038/sj.onc.1209943 . [DOI] [PubMed] [Google Scholar]
  • 18.Liu T, Zhang L, Joo D, Sun SC. NF-kappaB signaling in inflammation. Signal Transduct Target Ther. 2017;2. Epub 2017/11/22. doi: 10.1038/sigtrans.2017.23 ; PubMed Central PMCID: PMC5661633. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Barnes PJ, Karin M. Nuclear factor-kappaB: a pivotal transcription factor in chronic inflammatory diseases. N Engl J Med. 1997;336(15):1066–71. Epub 1997/04/10. doi: 10.1056/NEJM199704103361506 . [DOI] [PubMed] [Google Scholar]
  • 20.Tenda Y, Yamashita M, Kimura MY, Hasegawa A, Shimizu C, Kitajima M, et al. Hyperresponsive TH2 cells with enhanced nuclear factor-kappa B activation induce atopic dermatitis-like skin lesions in Nishiki-nezumi Cinnamon/Nagoya mice. J Allergy Clin Immunol. 2006;118(3):725–33. Epub 2006/09/05. doi: 10.1016/j.jaci.2006.05.024 . [DOI] [PubMed] [Google Scholar]
  • 21.Harley JB, Chen X, Pujato M, Miller D, Maddox A, Forney C, et al. Transcription factors operate across disease loci, with EBNA2 implicated in autoimmunity. Nat Genet. 2018;50(5):699–707. Epub 2018/04/18. doi: 10.1038/s41588-018-0102-3 ; PubMed Central PMCID: PMC6022759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Consortium GT. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580–5. Epub 2013/05/30. doi: 10.1038/ng.2653 ; PubMed Central PMCID: PMC4010069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Chandra V, Bhattacharyya S, Schmiedel BJ, Madrigal A, Gonzalez-Colin C, Fotsing S, et al. Promoter-interacting expression quantitative trait loci are enriched for functional genetic variants. Nat Genet. 2021;53(1):110–9. Epub 2020/12/23. doi: 10.1038/s41588-020-00745-3 ; PubMed Central PMCID: PMC8053422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Shao Z, Zhang Y, Yuan GC, Orkin SH, Waxman DJ. MAnorm: a robust model for quantitative comparison of ChIP-Seq data sets. Genome Biol. 2012;13(3):R16. Epub 2012/03/20. doi: 10.1186/gb-2012-13-3-r16 ; PubMed Central PMCID: PMC3439967. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Ashburner BP, Westerheide SD, Baldwin AS Jr. The p65 (RelA) subunit of NF-kappaB interacts with the histone deacetylase (HDAC) corepressors HDAC1 and HDAC2 to negatively regulate gene expression. Mol Cell Biol. 2001;21(20):7065–77. Epub 2001/09/21. doi: 10.1128/MCB.21.20.7065-7077.2001 ; PubMed Central PMCID: PMC99882. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Shaw J, Zhang T, Rzeszutek M, Yurkova N, Baetz D, Davie JR, et al. Transcriptional silencing of the death gene BNIP3 by cooperative action of NF-kappaB and histone deacetylase 1 in ventricular myocytes. Circ Res. 2006;99(12):1347–54. Epub 2006/11/04. doi: 10.1161/01.RES.0000251744.06138.50 . [DOI] [PubMed] [Google Scholar]
  • 27.Atak ZK, Taskiran II, Demeulemeester J, Flerin C, Mauduit D, Minnoye L, et al. Interpretation of allele-specific chromatin accessibility using cell state-aware deep learning. Genome Res. 2021;31(6):1082–96. Epub 2021/04/10. doi: 10.1101/gr.260851.120 ; PubMed Central PMCID: PMC8168584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Zhang S, Zhang H, Zhou Y, Qiao M, Zhao S, Kozlova A, et al. Allele-specific open chromatin in human iPSC neurons elucidates functional disease variants. Science. 2020;369(6503):561–5. Epub 2020/08/01. doi: 10.1126/science.aay3983 ; PubMed Central PMCID: PMC7773145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Abramov S, Boytsov A, Bykova D, Penzar DD, Yevshin I, Kolmykov SK, et al. Landscape of allele-specific transcription factor binding in the human genome. Nat Commun. 2021;12(1):2751. Epub 2021/05/14. doi: 10.1038/s41467-021-23007-0 ; PubMed Central PMCID: PMC8115691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.N. K, D. HJ, R. MJ, P. W, L. K, K. P, et al. eQTL Catalogue: a compendium of uniformly processed human gene expression and splicing QTLs. BioRxiv. 2020. [Google Scholar]
  • 31.Ma L, Xue HB, Guan XH, Shu CM, Zhang JH, Yu J. Possible pathogenic role of T helper type 9 cells and interleukin (IL)-9 in atopic dermatitis. Clin Exp Immunol. 2014;175(1):25–31. Epub 2013/09/17. doi: 10.1111/cei.12198 ; PubMed Central PMCID: PMC3898551. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Acevedo N, Benfeitas R, Katayama S, Bruhn S, Andersson A, Wikberg G, et al. Epigenetic alterations in skin homing CD4(+)CLA(+) T cells of atopic dermatitis patients. Sci Rep. 2020;10(1):18020. Epub 2020/10/24. doi: 10.1038/s41598-020-74798-z ; PubMed Central PMCID: PMC7582180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Biedermann T, Skabytska Y, Kaesler S, Volz T. Regulation of T Cell Immunity in Atopic Dermatitis by Microbes: The Yin and Yang of Cutaneous Inflammation. Front Immunol. 2015;6:353. Epub 2015/07/29. doi: 10.3389/fimmu.2015.00353 ; PubMed Central PMCID: PMC4500098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Czarnowicki T, He H, Canter T, Han J, Lefferdink R, Erickson T, et al. Evolution of pathologic T-cell subsets in patients with atopic dermatitis from infancy to adulthood. J Allergy Clin Immunol. 2020;145(1):215–28. Epub 2019/10/19. doi: 10.1016/j.jaci.2019.09.031 ; PubMed Central PMCID: PMC6957229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Lefebvre JS, Maue AC, Eaton SM, Lanthier PA, Tighe M, Haynes L. The aged microenvironment contributes to the age-related functional defects of CD4 T cells in mice. Aging Cell. 2012;11(5):732–40. Epub 2012/05/23. doi: 10.1111/j.1474-9726.2012.00836.x ; PubMed Central PMCID: PMC3444657. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Cabanillas B, Brehler AC, Novak N. Atopic dermatitis phenotypes and the need for personalized medicine. Curr Opin Allergy Clin Immunol. 2017;17(4):309–15. Epub 2017/06/06. doi: 10.1097/ACI.0000000000000376 ; PubMed Central PMCID: PMC5515628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Dajee M, Muchamuel T, Schryver B, Oo A, Alleman-Sposeto J, De Vry CG, et al. Blockade of experimental atopic dermatitis via topical NF-kappaB decoy oligonucleotide. J Invest Dermatol. 2006;126(8):1792–803. Epub 2006/04/22. doi: 10.1038/sj.jid.5700307 . [DOI] [PubMed] [Google Scholar]
  • 38.Nakamura H, Aoki M, Tamai K, Oishi M, Ogihara T, Kaneda Y, et al. Prevention and regression of atopic dermatitis by ointment containing NF-kB decoy oligodeoxynucleotides in NC/Nga atopic mouse model. Gene Ther. 2002;9(18):1221–9. Epub 2002/09/07. doi: 10.1038/sj.gt.3301724 . [DOI] [PubMed] [Google Scholar]
  • 39.Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005–D12. Epub 2018/11/18. doi: 10.1093/nar/gky1120 ; PubMed Central PMCID: PMC6323933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Ellinghaus D, Baurecht H, Esparza-Gordillo J, Rodriguez E, Matanovic A, Marenholz I, et al. High-density genotyping study identifies four new susceptibility loci for atopic dermatitis. Nat Genet. 2013;45(7):808–12. Epub 2013/06/04. doi: 10.1038/ng.2642 ; PubMed Central PMCID: PMC3797441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Genomes Project C, Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68–74. Epub 2015/10/04. doi: 10.1038/nature15393 ; PubMed Central PMCID: PMC4750478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. Epub 2015/02/28. doi: 10.1186/s13742-015-0047-8 ; PubMed Central PMCID: PMC4342193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Griffiths C, de Bruin-Weller M, Deleuran M, Fargnoli MC, Staumont-Salle D, Hong CH, et al. Dupilumab in Adults with Moderate-to-Severe Atopic Dermatitis and Prior Use of Systemic Non-Steroidal Immunosuppressants: Analysis of Four Phase 3 Trials. Dermatol Ther (Heidelb). 2021;11(4):1357–72. Epub 2021/06/19. doi: 10.1007/s13555-021-00558-0 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Schneider L, Tilles S, Lio P, Boguniewicz M, Beck L, LeBovidge J, et al. Atopic dermatitis: a practice parameter update 2012. J Allergy Clin Immunol. 2013;131(2):295–9 e1-27. Epub 2013/02/05. doi: 10.1016/j.jaci.2012.12.672 . [DOI] [PubMed] [Google Scholar]
  • 45.Buenrostro JD, Wu B, Chang HY, Greenleaf WJ. ATAC-seq: A Method for Assaying Chromatin Accessibility Genome-Wide. Curr Protoc Mol Biol. 2015;109:21 9 1–9 9. Epub 2015/01/07. doi: 10.1002/0471142727.mb2129s109 ; PubMed Central PMCID: PMC4374986. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Corces MR, Trevino AE, Hamilton EG, Greenside PG, Sinnott-Armstrong NA, Vesuna S, et al. An improved ATAC-seq protocol reduces background and enables interrogation of frozen tissues. Nat Methods. 2017;14(10):959–62. Epub 2017/08/29. doi: 10.1038/nmeth.4396 ; PubMed Central PMCID: PMC5623106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9(4):357–9. Epub 2012/03/06. doi: 10.1038/nmeth.1923 ; PubMed Central PMCID: PMC3322381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25(16):2078–9. Epub 2009/06/10. doi: 10.1093/bioinformatics/btp352 ; PubMed Central PMCID: PMC2723002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Feng J, Liu T, Qin B, Zhang Y, Liu XS. Identifying ChIP-seq enrichment using MACS. Nat Protoc. 2012;7(9):1728–40. Epub 2012/09/01. doi: 10.1038/nprot.2012.101 ; PubMed Central PMCID: PMC3868217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Amemiya HM, Kundaje A, Boyle AP. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. Sci Rep. 2019;9(1):9354. Epub 2019/06/30. doi: 10.1038/s41598-019-45839-z ; PubMed Central PMCID: PMC6597582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Schmidl C, Rendeiro AF, Sheffield NC, Bock C. ChIPmentation: fast, robust, low-input ChIP-seq for histones and transcription factors. Nat Methods. 2015;12(10):963–5. Epub 2015/08/19. doi: 10.1038/nmeth.3542 ; PubMed Central PMCID: PMC4589892. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15–21. Epub 2012/10/30. doi: 10.1093/bioinformatics/bts635 ; PubMed Central PMCID: PMC3530905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. 2014;30(7):923–30. Epub 2013/11/15. doi: 10.1093/bioinformatics/btt656 . [DOI] [PubMed] [Google Scholar]
  • 54.McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297–303. Epub 2010/07/21. doi: 10.1101/gr.107524.110 ; PubMed Central PMCID: PMC2928508. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43(5):491–8. Epub 2011/04/12. doi: 10.1038/ng.806 ; PubMed Central PMCID: PMC3083463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Van der Auwera GA, Carneiro MO, Hartl C, Poplin R, Del Angel G, Levy-Moonshine A, et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Curr Protoc Bioinformatics. 2013;43:11 0 1–0 33. Epub 2014/11/29. doi: 10.1002/0471250953.bi1110s43 ; PubMed Central PMCID: PMC4243306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Lu X, Chen X, Forney C, Donmez O, Miller D, Parameswaran S, et al. Global discovery of lupus genetic risk variant allelic enhancer activity. Nat Commun. 2021;12(1):1611. Epub 2021/03/14. doi: 10.1038/s41467-021-21854-5 ; PubMed Central PMCID: PMC7955039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Hass MR, Brissette D, Parameswaran S, Pujato M, Donmez O, Kottyan LC, et al. Runx1 shapes the chromatin landscape via a cascade of direct and indirect targets. PLoS Genet. 2021;17(6):e1009574. Epub 2021/06/11. doi: 10.1371/journal.pgen.1009574 . [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision Letter 0

John M Greally, Devin M Absher

20 Dec 2021

Dear Dr Kottyan,

Thank you very much for submitting your Research Article entitled 'Epigenetic and Transcriptional Dysregulation in CD4+ T cells of Patients with Atopic Dermatitis' to PLOS Genetics.

The manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important problem, but raised some substantial concerns about the current manuscript. Based on the reviews, we will not be able to accept this version of the manuscript, but we would be willing to review a much-revised version. We cannot, of course, promise publication at that time.

Should you decide to revise the manuscript for further consideration here, your revisions should address the specific points made by each reviewer. We will also require a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

If you decide to revise the manuscript for further consideration at PLOS Genetics, please aim to resubmit within the next 60 days, unless it will take extra time to address the concerns of the reviewers, in which case we would appreciate an expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments are included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool.  PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, use the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

We are sorry that we cannot be more positive about your manuscript at this stage. Please do not hesitate to contact us if you have any concerns or questions.

Yours sincerely,

Devin M Absher

Associate Editor

PLOS Genetics

John Greally

Section Editor: Epigenetics

PLOS Genetics

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: Thanks to the authors for this paper. I have only two criticisms:

1. The paper largely focuses on overlaps in genomic regions without applying any statistical evidence that those overlaps are more extreme than would be expected by chance alone. In the resubmission I would advise including hypergeometric testing or something similar to asses the strength of evidence.

2. The statement below is copied from the manuscript:

The 100 kB region of DNA around AD-specific gene sets widely overlapped (94.7-100%) the ATAC-seq peaks in the six subjects with AD (Supplemental Table 8). There was substantial overlap (26.3-68.4%) between the 100 kb region of DNA around ADspecific gene sets with the AD-specific ATAC-peaks, indicating that possible enhancers proximal to the AD-specific genes were accessible for transcription in an AD-specific manner.

The above gives no indication of the directionality of effect. Can you produce a graph or something similar to show the relationship between chromatin accessibilty and gene expression i.e logFC case v control (GE) v logFC case v control (ATAC)

3. Likewise is it possible to include a figure demonstrating the allele-specific chromatin accessibility i.e ATAC reads stratified by genotype in cases and controls or something similar?

Thanks

Reviewer #2: This is an interesting report from Eapen et al., that characterizes the epigenetic and transcriptional dysregulation of CD4 T cell in patients with atopic dermatitis (AD). AD affects approximately 20% of children and high rate of persistence into adulthood. There are 29 independent risk haplotypes identified. It is well know that CD4 T cells are the major effector cell type for AD and that NFKB signaling mediates the pathogenic inflammation. This paper seeks to determine if there are upstream effects in CD4 T cells at the level of epigenome and transcriptome changes that facilitate pathogenic T cell responses. The authors use a combination of genomic techniques to determine if AD risk haplotypes demonstrate altered chromatin accessibility, NFKB binding and gene expression changes consistent with AD using a case/control study design in CD4 T cells in patients with AD. They demonstrate that in stimulated CD4 T cells taken from patients with AD that open chromatin regions were enriched for AD risk variants and that there was strong enrichment for NFKB binding motifs in these peaks in AD patients but not in controls. They also demonstrated over 60 instances of genotype-dependent chromatin accessibility for 36 AD risk variants. Together they conclude that allele specific epigenetic and transcriptional regulation is an important feature of CD4 T cell responses to stimulation in patients with AD. While these results are interesting, there are several significant issues that need to be addressed to improve the overall quality of this manuscript. These are listed below in no specific order of importance:

1. The text is missing some basic proofreading

2. Testing the hypothesis that “AD loci may be epigenetically regulated” is vague (pp. 2, 128).

3. The FRiP scores are quite low with average of 32% indicating that the ATAC-seq data may have QC issues (typically >50% is expected), please address.

4. Analysis of shared peaks was performed by using pairwise assessments. It is not clear how case/control pairs were selected and why a more exhaustive case/control analysis was not performed (i.e., pairwise comparisons using all possible pairs), or why consensus peaks were not called from pooled data.

5. More generally, how were subjects matched?

6. After the pairing approach used to identify AD- or control-specific ATAC peaks, peaks that were AD- or control-specific in 3 or more of the pairings were classified as “consistently AD/control” (pp. 5, 176-178). It is not stated how much these peak sets overlap each other. Based on the data in the supplement (Supplemental Table 3), they share 80% of LD SNPs so it is likely they have some overlap.

7. Based on shared LD and Tag SNPs (Supplemental Table 3) it is NOT clear that chromatin is accessible in a “disease specific manner” (pp. 5, 181-182). Twelve of the 13 AD risk loci overlapped by “consistently AD” peaks are also overlapped by “consistently control” peaks.

8. On pp. 5, 170-171, authors state 75-88.4% of ATAC peaks were shared between AD and demographically matched controls, but on pp. 8, 231 authors state a median of 91.9% of ATAC peaks were shared. It is unclear how this inconsistency arose or if the comparison being made is different in the second statement in a way that is not clear from the text.

9. On pp. 22, 471-472 in the METHODS section, authors state that reads for ChIP-seq were processed the same way as ATAC-seq reads, but the MACS2 parameters needed to properly call peaks in these two data types should be different to avoid errors in the genomic coordinates of the called peaks, especially excluding parameters ‘-nomodel --shift --extsize’ for ATAC-seq peak calling.

10. On pp. 8, 228-231, it states that there is more variability in overlap between subject matched pairs for ChIP vs. ATAC peaks, but this may be entirely due to the fact that there is substantial variability in the number of ChIP peaks across samples (Supplemental Table 2). It would be useful to compare variability of overlap within controls and within AD to see if the variability between control vs. AD is any greater than the large variability already present in the samples.

11. The biological processes associated with the differentially expressed genes (15 are expressed 1.5x higher in AD, 16 are expressed 1.5x lower in AD; shown in a table in Supplemental Figure 6) are not particularly compelling. These seem like very high-level and general processes and many of the enriched pathways include hundreds or even over 1000 genes. More generally, only finding 31 differentially expressed genes seems surprisingly low.

12. The case/control RNA-seq comparisons were only performed pairwise (pp. 11, 280), which seems like a missed opportunity for comparing pooled case vs. pooled control data, which would presumably have more power to detect differentially expressed variants (this applies to the pairwise comparisons used more generally)

13. On pp. 12, 289-299, authors describe how a large % of AD-specific ATAC peaks overlap the 100kb region around AD-specific gene sets. They then state that NFKB1 ChIP-seq peaks overlap a large % of the TSS of AD-specific genes (they don’t state whether it is AD-specific NFKB1 peaks or ALL peaks). They then state that a lower % of AD-specific NFKB1 peaks overlap the AD-specific genes (does this mean that the peaks overlap the transcripts?). It would be helpful to show additional comparisons here to make the argument that control- and AD-ChIP/ATAC peaks are truly specific for control- and AD-genes. For example, what % of AD-specific genes show control-specific ATAC peaks in the 100kb vicinity? Is it similar to the % of AD-specific ATAC peaks, or is the % of AD-specific ATAC peaks significantly higher in the vicinity of AD-specific genes, etc.

14. Authors used MARIO method to integrate information across ATAC/ChIP/RNA-seq data with WGS for each of their subjects and find allele dependence of sequencing reads (read depth?) at het variants (het across the sample, or within individuals?) (pp. 13, 304-307)

15. Pp. 13, 311-313: The authors state that the MARIO analysis discovers AD-associated variants in ATAC peaks that are het in some subjects and produce allele-dependent ATAC peaks. This seems like a circular argument: the MARIO analysis uses differences in sequencing reads combined with genotyping data to discover alleles that affect sequencing reads, so why is it striking that it finds alleles that produce allele dependent ATAC peaks, since that is the very signal it is based on?

16. Fewer AD-associated NFKB1 binding variants were found, but this may be largely driven by the fact that there were ~7x more ATAC peaks than ChIP peaks (as the authors acknowledge pp. 14, 328-329).

17. It would be interesting search the AD-associated alleles in large public datasets and look analyze their frequencies, whether they have been associated with other diseases, etc.

18. The author should make an effort in the discussion, to connect how their data will impact the “atopic march” mentioned in the introduction.

19. There is an error in the readme file tab in Table 12 that should be changed. See line 15 of the table. S_read mislabeled as weak reads.

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #2: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: David Martino

Reviewer #2: No

Decision Letter 1

John M Greally, Devin M Absher

20 Apr 2022

Dear Dr Kottyan,

We are pleased to inform you that your manuscript entitled "Epigenetic and Transcriptional Dysregulation in CD4+ T cells in Patients with Atopic Dermatitis" has been editorially accepted for publication in PLOS Genetics. Congratulations!

Before your submission can be formally accepted and sent to production you will need to complete our formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Please note: the accept date on your published article will reflect the date of this provisional acceptance, but your manuscript will not be scheduled for publication until the required changes have been made.

Once your paper is formally accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosgenetics@plos.org.

In the meantime, please log into Editorial Manager at https://www.editorialmanager.com/pgenetics/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production and billing process. Note that PLOS requires an ORCID iD for all corresponding authors. Therefore, please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field.  This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

If you have a press-related query, or would like to know about making your underlying data available (as you will be aware, this is required for publication), please see the end of this email. If your institution or institutions have a press office, please notify them about your upcoming article at this point, to enable them to help maximise its impact. Inform journal staff as soon as possible if you are preparing a press release for your article and need a publication date.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Genetics!

Yours sincerely,

Devin M Absher

Associate Editor

PLOS Genetics

John Greally

Section Editor: Epigenetics

PLOS Genetics

www.plosgenetics.org

Twitter: @PLOSGenetics

----------------------------------------------------

Comments from the reviewers (if applicable):

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: My comments were addressed, thanks to the authors.

Reviewer #2: The authors have adequately addressed the concerns raised in the original version of the manuscript

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #2: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

----------------------------------------------------

Data Deposition

If you have submitted a Research Article or Front Matter that has associated data that are not suitable for deposition in a subject-specific public repository (such as GenBank or ArrayExpress), one way to make that data available is to deposit it in the Dryad Digital Repository. As you may recall, we ask all authors to agree to make data available; this is one way to achieve that. A full list of recommended repositories can be found on our website.

The following link will take you to the Dryad record for your article, so you won't have to re‐enter its bibliographic information, and can upload your files directly: 

http://datadryad.org/submit?journalID=pgenetics&manu=PGENETICS-D-21-01587R1

More information about depositing data in Dryad is available at http://www.datadryad.org/depositing. If you experience any difficulties in submitting your data, please contact help@datadryad.org for support.

Additionally, please be aware that our data availability policy requires that all numerical data underlying display items are included with the submission, and you will need to provide this before we can formally accept your manuscript, if not already present.

----------------------------------------------------

Press Queries

If you or your institution will be preparing press materials for this manuscript, or if you need to know your paper's publication date for media purposes, please inform the journal staff as soon as possible so that your submission can be scheduled accordingly. Your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version. PLOS Genetics may also choose to issue a press release for your article. If there's anything the journal should know or you'd like more information, please get in touch via plosgenetics@plos.org.

Acceptance letter

John M Greally, Devin M Absher

11 May 2022

PGENETICS-D-21-01587R1

Epigenetic and Transcriptional Dysregulation in CD4+ T cells in Patients with Atopic Dermatitis

Dear Dr Kottyan,

We are pleased to inform you that your manuscript entitled "Epigenetic and Transcriptional Dysregulation in CD4+ T cells in Patients with Atopic Dermatitis" has been formally accepted for publication in PLOS Genetics! Your manuscript is now with our production department and you will be notified of the publication date in due course.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

Soon after your final files are uploaded, unless you have opted out or your manuscript is a front-matter piece, the early version of your manuscript will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting PLOS Genetics and open-access publishing. We are looking forward to publishing your work!

With kind regards,

Agnes Pap

PLOS Genetics

On behalf of:

The PLOS Genetics Team

Carlyle House, Carlyle Road, Cambridge CB4 3DN | United Kingdom

plosgenetics@plos.org | +44 (0) 1223-442823

plosgenetics.org | Twitter: @PLOSGenetics

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Fig. ATAC-seq and NFKB1 ChIP-seq peak pairwise normalized read count comparisons.

    Each dot represents a peak that is present in either of the two samples. The X- and Y-axis indicate the number of normalized read counts for that peak in the two samples. The red line indicates peaks with similar normalized read counts between the two samples. A. ATAC- seq scatter plot analysis for controls. B. ATAC-seq scatter plot analysis for AD cases. C. ChIP-seq scatter plot analysis for controls. D. ChIP-seq scatter plot analysis for AD cases. E. Global Variability analysis of the ATAC and ChIP peaks across AD and controls.

    (PDF)

    Click here for additional data file. (663.1KB, pdf)
    S2 Fig. Case/control pairwise differential ATAC-seq analysis and transcription factor motif enrichment.

    (Left plots). For each of the six matched sample pairs, differential ATAC- seq peak analysis was performed. ATAC-seq peaks that are equally strong between case and controls (“shared”), significantly stronger in the control (“control-specific”), or significantly stronger in the case (“AD-specific”) were identified using MAnorm (see Methods). The percent of peaks falling within each category is indicated to the left of each heatmap. Heatmap color indicates the normalized strength of the ATAC-seq signal. (Right plots). Transcription factor motif enrichment analysis was performed for each sample pair (see Methods). Each dot represents a transcription factor motif. The X-axis and Y-axis indicate the p-value of enrichment within control-specific and AD-specific ATAC-seq peaks, respectively. Motif family color key is provided at the top of each plot.

    (PDF)

    Click here for additional data file. (722.2KB, pdf)
    S3 Fig. Number of consistently differential ATAC-seq and ChIP-seq peaks across sample pairs.

    ATAC-seq and ChIP-seq peaks were called for all cases and controls. For each matched pair of samples, a MAnorm analysis identified shared peaks, control-specific peaks and AD-specific peaks (see Methods). Bar plots indicate the distributions of the number of control- specific (left) or case-specific peaks (right) for ATAC-seq (A,B) and ChIP-seq (C,D). For example, there are approximately 10 ATAC-seq peaks in the genome that are control-specific in 4 of the 6 matched pairs (see panel A).

    (PDF)

    Click here for additional data file. (331.5KB, pdf)
    S4 Fig. Case/control pairwise differential NKFB1 ChIP-seq analysis and transcription factor motif enrichment.

    (Left plots). For each of the six matched sample pairs, differential NKFB1 ChIP-seq peak analysis was performed. ChIP-seq peaks that are equally strong between case and controls (“shared”), significantly stronger in the control (“control-specific”), or significantly stronger in the case (“AD-specific”) were identified using MAnorm (see Methods). The percent of peaks falling within each category is indicated to the left of each heatmap. Heatmap color indicates the normalized strength of the ChIP-seq signal. (Right plots). Transcription factor motif enrichment analysis was performed for each sample pair (see Methods). Each dot represents a transcription factor motif. The X-axis and Y-axis indicate the p-value of enrichment within control- specific and AD-specific ChIP-seq peaks, respectively. Motif family color key is provided at the top of each plot.

    (PDF)

    Click here for additional data file. (607.7KB, pdf)
    S5 Fig. Transcription factor motif enrichment comparison between consistently control-specific and consistently AD-specific NFKB1 ChIP-seq peaks.

    The top five enriched TF motif families are shown. A. Percent of peaks containing predicted binding sites for the indicated motif. B. Motif enrichment p-value within those peaks (see Methods).

    (PDF)

    Click here for additional data file. (192.2KB, pdf)
    S6 Fig. RNA-seq GO enrichment results.

    Pairwise differential gene expression analysis was performed by comparing RNA-seq data for each of the cases with their matched controls (see Methods). Genes with 1.5-fold changes or greater in three or more pairs were identified. Pathway enrichment analysis was performed on these gene sets. The top five enriched GO biological pathways and GO molecular function categories are indicated within the top and bottom tables, respectively.

    (PDF)

    Click here for additional data file. (20.6KB, pdf)
    S1 Table. Atopic Dermatitis risk variants and loci.

    (XLSX)

    Click here for additional data file. (68.6KB, xlsx)
    S2 Table. Quality control summary.

    (XLSX)

    Click here for additional data file. (16.7KB, xlsx)
    S3 Table. Regulatory Element Locus Intersection (RELI) analysis of ATAC-seq and ChIP datasets.

    (XLSX)

    Click here for additional data file. (16.7KB, xlsx)
    S4 Table. Transcription factor motif analysis of ATAC-seq datasets.

    (XLSX)

    Click here for additional data file. (6.3MB, xlsx)
    S5 Table. Transcription factor motif analysis of ATAC-seq peaks consistently found in samples from patients with atopic dermatitis and controls.

    (XLSX)

    Click here for additional data file. (855.7KB, xlsx)
    S6 Table. Regulatory Element Locus Intersection (RELI) analysis of NFKB1 ChIP-seq datasets.

    (XLSX)

    Click here for additional data file. (11.8KB, xlsx)
    S7 Table. Transcription factor motif analysis of NFKB1 ChIP-seq datasets.

    (XLSX)

    Click here for additional data file. (6MB, xlsx)
    S8 Table. Regulatory Element Locus Intersection (RELI) integration of RNA-seq, ATAC-seq, and NFKB1 ChIP-seq datasets.

    (XLSX)

    Click here for additional data file. (21.4KB, xlsx)
    S9 Table. Transcription factor motif analysis of NFKB1 ChIP-seq peaks consistently found in samples from patients with atopic dermatitis and controls.

    (XLSX)

    Click here for additional data file. (813.9KB, xlsx)
    S10 Table. Analysis of RNA-seq datasets.

    (XLSX)

    Click here for additional data file. (178.5KB, xlsx)
    S11 Table. Identification of AD risk variants in DNA-seq datasets.

    (XLSX)

    Click here for additional data file. (20.6KB, xlsx)
    S12 Table. Identification of genotype-dependent.

    (XLSX)

    Click here for additional data file. (18.5KB, xlsx)
    S13 Table. Demographics of participants in this study.

    (DOCX)

    Click here for additional data file. (13.7KB, docx)
    Attachment

    Submitted filename: AD_response_to_reviewer_comments_v8_FINAL_2.pdf

    Click here for additional data file. (2.6MB, pdf)

    Data Availability Statement

    All raw and processed sequencing data generated in this study have been submitted to the NCBI Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) database under accession number GSE184238. A UCSC Genome Browser session is available at http://genome.ucsc.edu/s/ledsall/AtopicDermatitis.

    Articles from PLoS Genetics are provided here courtesy of PLOS

    RESOURCES