Genomic Studies of Autoimmune Rheumatic Disease

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

Wildfire Smoke on Epigenome and Health: An additional collaboration with Genetic Epidemiology and Genomics Laboratory (GEGL) at the University of California Berkeley (UCB) involves an award from the Intramural Targeted Climate Change & Health (ITCCH) funding program at NIEHS. Data and sample collection for this project is underway to examine the effects of wildfire smoke on the epigenome and health. To date, approximately 10,000 samples have already been collected and extracted. DNA methylation profiling and SNP genotyping is in progress. Additional data collection of exposure assessments is also under way. Sjögren’s Disease Glands: From the SICCA cohort, labial salivary glands (LSG) from 1,453 participants were profiled on the Illumina Infinium MethylationEPIC BeadChip v1 array. Standard QC pipelines were applied and a total of 9 cell-type proportions were estimated that included epithelial cells, fibroblasts and 7 types of immune cells. Statistical methods applied include randomly separating 80% of the dataset into a discovery and 20% into a validation cohort. All analyses were conducted in the discovery cohort and then significant results were tested in the validation cohort for the same direction of effect and p<0.05. Differentially methylated positions (DMPs) were identified and validated. There were 2,659 differentially methylated positions associated with SjD in any cell type. For a subset of 89 individuals where single nuclei expression data was available, expression of genes annotated to cell-specific DMPs were tested for association with SjD. In cell-type specific analyses, fibroblasts had the largest number of DMPs, followed by epithelial cells and natural killer cells. These results demonstrate that there are differences in DNA methylation in the LSG across different cell types associated with SjD. This might improve our understanding of the underlying biology of SjD and help identify therapeutic targets. Smoking and Sjögren’s Disease: Cigarette smoking has been linked to the development of autoimmune diseases including SjD. Since DNA methylation is altered by cigarette smoking, it may provide some insight into the influence of smoking on SjD progression. Using the previously mentioned set of methylation data in 1,453 LSG, DNA methylation was compared with different self-reported smoking behaviors using linear models. In current smokers, approximately 106k CpG sites were differentially methylated compared to approximately 13 CpG sites in former smokers. Results suggest that cigarette smoking may perturb normal cellular processes in salivary glands. The limited number of statistically significant differentially methylated results in former smokers suggests that quitting smoking may allow individuals to acquire DNA methylation profiles comparable to individuals who have never smoked. WES in Sjögren’s Disease: Whole Exome Sequencing (WES) data has been generated at NISC on 768 samples of European and East Asian descent from the SICCA cohort. These samples consist of extreme phenotypes of 438 severe SjD cases with an ACR-EULAR score of 8 or 9 versus 330 symptomatic controls with a score of 0. Additional WES data has also been generated from 384 healthy controls recruited through the Kaiser Permanente Northern California Plan (KPNC) Research Program controls . Previous gene-based analysis was performed on an earlier subset of 384 to test for association using the rare variants in each gene region. The previous analysis on a smaller subset showed there were no genes significantly associated with severe SjD and rare or deleterious variants were more common in the non-cases. However, several variants have a higher frequency among SjD cases and symptomatic controls compared to the gnomAD database. Future analysis plans include determining whether there are rare genetic variants associated with severe SjD case status versus healthy controls. WES in SLE Disease: We have also generated WES data at NISC on 384 SLE samples of Asian SLE patients with the goal of identifying rare coding variants specific to clinical phenotypes among Asians. These 384 are from our Lupus Genetics Project and the CLUES cohort. SLE patients of Asian descent experience higher rates of lupus nephritis (LN) than patients of European descent. Analyses are currently underway for a SLE case only analysis examining LN as a dichotomous variable and age of disease onset as a continuous variable. Additional analysis plans include comparing these 384 SLE cases to approximately 12,000 healthy matched controls of Asian descent from the UK Biobank data. We expect that results from this study can help determine whether differences in the genetic architecture of SLE patients, specifically rare and low frequency variants, are associated with disease development and clinical phenotype. WES combined SLE and SjD: In the past year, the joint calling of genotypes from the SjD and SLE WES datasets have been completed. Analysis plans in the coming year include combining both SLE and SjD sets into a combined analysis to examine shared variants across the two diseases. Lupus DNA Methylation Studies: In the past year, multiple projects began on 1,882 SLE patients profiled on the Illumina EPICv2 BeadChip, which covers approximately 937,000 CpG sites across the genome. DNA methylation has been associated with the development of SLE as well as the heterogeneity among SLE patients. To identify biologically relevant molecular subtypes, clustering analyses were performed on the methylation data to investigate associations with SLE clinical characteristics. Four distinct clusters were identified using DNA methylation values. Clusters 1 and 2 suggested milder disease with an association with discoid rash. Cluster 3 exhibited older age of enrollment and Cluster 4 was associated with severe disease as reflected by younger age of diagnosis. These results support the potential of DNA methylation profiles to define molecular subtypes of SLE. An additional project with this dataset of 1,882 individuals with SLE includes the examination of methylation patterns among SLE patients with and without lupus nephritis. The dataset of 1,882 was randomly split into a discovery and validation cohort. Among the discovery cohort, differentially methylated positions associated with nephritis were identified using linear regression models. Statistically significant results from the discovery cohort were then tested in the validation cohort. 5,783 CpG sites were identified in the discovery cohort and approximately 3% of these CpG sites were statistically significant in the validation cohort. The 3 % in the validation cohort involved genes in innate immunity and inflammation. These results suggest that methylation patterns may provide opportunities to further subtype SLE patients for risk stratification in clinical settings. A third project investigates whether DNA methylation patterns differ between genetically similar SLE individuals living in Peru and the United States. DNA methylation data and genotype data were utilized from 1,671 SLE patients in the United States and 74 individuals from Peru. Principal components analysis (PCA) was performed to identify genetically similar individuals with SLE. This resulted in a genetically similar set of 86 individuals from the US and 64 individuals from Peru. Results indicated 13,736 CPG sites that were differentially methylated by country. The CpG sites with the largest differences in methylation were in genes involved in immune function. Understanding how these differences shape DNA methylation patterns may offer insights into SLE pathogenesis across populations. A fourth project examines the role of clonal hematopoiesis of indeterminate potential (CHIP) with epigenetic dysregulation in SLE. CHIP involves age-related somatic mutations in hematopoietic stem cells frequently affecting epigenetic regulators such as DNMT3A and TET2. An epigenome-wide association study was conducted to investigate how mutations in DNMT3A and overall CHIP mutation burden (TMLoad) influence DNA methylation. CHIP mutations were identified in 1,073 SLE patients using a hybrid capture-based targeted sequencing panel covering 22 canonical CHIP genes and profiled for methylation data on the Illumina Infinium MethylationEPIC v2 array. Linear regression models were run with CHIP mutation status as the predictor. Twelve differentially methylated CpG sites were associated with mutations in DNMT3A, 783 CpG sites in TET2 and 1,549 CpG sites with TMLoad. Pathway analysis indicated relevant processes such as chromatin remodeling, transcriptional regulation, and immune-related and hematopoietic differentiation. These results demonstrate that CHIP mutations may drive epigenetic changes in SLE. SLE FLARE studies: We profiled 59 SLE patients undergoing a disease flare on the Illumina EPIC chip v1, which consists of 850,000 CpGs genome-wide. These SLE flaring patients were undergoing routine clinical care at the time of recruitment. A flare was characterized by the treating rheumatologist and using the SLEDAI score, which is a validated measure of disease activity in lupus. Standard QC pipelines for the EPIC methylation chip were applied. Using linear regression models, we identified CpG sites where methylation changed between visits differentially depending on whether an SLE patient remitted at the follow-up visit. Models identifying the CpG sites accounted for the paired design and adjusted for blood cell proportions, time between visits, batch, age at flare, sex, medications and genetic principal components. Hierarchical clustering was performed on the most significant CpGs to identify patient subtypes. Nineteen patients fully remitted at follow-up visit. Methylation changed differentially at 1,953 CpG sites for remitters versus non-remitters and nearly half of those sites annotated to interferon-regulated genes. Three patient clusters were identified using methylation differences agnostic of clinical outcomes. All patients in the first cluster experienced complete flare remission. Longitudinal DNA methylation dynamics during active SLE were associated with remission status and identified subgroups of SLE patients with several distinct clinical and biological characteristics. Metabolomics in SLE: It is unknown whether metabolite levels are associated with SLE disease activity or are correlated with other biomarkers of SLE disease activity such as DNA methylation. Using a cohort of 40 multi-ethnic SLE patients recruited during a disease flare with a follow-up visit approximately three months later, the goal of this study was to identify whether changes in metabolites are associated with flare remission and if these changes are correlated with changes in DNA methylation. Metabolomic data was captured via LC-QTOF/MS, which captures small molecules produced endogenously by inflammation, oxidative stress, lipid peroxidation, and the gut flora. In a previous analysis, this patient set had whole blood DNA methylation profiles associated with SLE remission in 291 sites. Significant metabolite changes were tested for their association with methylation changes at these 291 sites using correlation coefficients. We also examined whether changes in metabolites were associated with three patient clusters previously identified from clustering methylation changes. Sixteen SLE patients fully remitted at follow-up. Nine metabolite changes were associated with remission status compared to non-remitting SLE patients. The nine metabolites included oleic acid and two isomers of adenosine. These metabolite changes were correlated with changes in DNA methylation at SLE-relevant loci. The integration of DNA methylation and metabolomics might help us better understand underlying biological pathways relevant for SLE flare remission and result in targeted therapies.