Genomic Studies of Autoimmune Rheumatic Disease

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

During the current reporting period, the longitudinal study examining 101 SLE pairs from the CLUES cohort has been published online and is currently undergoing final editing. One study in progress is examining environmental exposures in 300 SLE patients from the CLUES cohort and 100 controls matched by age, sex and race. Serum samples from SLE patients were profiled using liquid chromatography coupled with Q-TOF mass spectrometry (LC/Q-TOF MS) methods for measuring environmental chemicals. The goal of this study is to develop a novel method to characterize the serum levels of circulating environmental chemicals and metabolites in this cohort of deeply phenotyped SLE patients. An additional study to characterize the metabolome is being planned on flaring SLE patients. In addition to capturing hazardous exposures, LC-QTOF/MS can also capture small molecules produced endogenously by inflammation, oxidative stress, lipid peroxidation, and the gut flora, referred to as the metabolome. We aim to characterize the metabolome at different stages of disease activity in SLE patients as well as comparing the metabolome in SLE patients to healthy controls. During this current reporting period for our studies on Sjogren's syndrome, we are in the process of generating data on glands for approximately 1,000 Sjogren's syndrome cases and symptomatic non-cases from the SICCA cohort. The glands are currently being prepared for profiling on the Illumina Infinium MethylationEPIC BeadChip array. In collaboration with the NISC, we have also generated Whole Exome Sequencing data on approximately 400 samples. The following sections describe our progress on the following studies in the current reporting period for the SLE FLARE study, Epigenetic Aging in Lupus, and Sjogren's syndrome subtypes. SLE FLARE study: We profiled 53 SLE patients undergoing a disease flare on the Illumina EPIC chip 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 with the SLEDAI score, which is a validated measure of disease activity in lupus. Quality control on the methylation data included quantile normalization and background subtraction with dye-bias normalization. In addition, CpG sites with high detection p-values, cross-reactive probes, and those with SNPs in or near probes were filtered out. Using the R package limma, differentially methylated positions (DMPs) associated with SLEDAI score were identified. Models identifying the DMPs accounted for the paired design and adjusted for blood cell proportions, batch, age, sex, medications and genetic principal components. Hierarchical clustering was performed on the 5,000 most significant DMPs to identify patient subtypes. Gene ontology enrichment analysis was performed on the set of DMPs associated with each subtype. Hierarchical clustering revealed three clusters of patients. For Cluster 1, the most significant gene ontology pathways were predominantly phosphorylation processes and regulation of defense response by viruses. For Cluster 2, import into the nucleus and adrenergic receptor signaling were the most significant gene ontology pathways. Lastly, synaptic transmission and several neuronal membrane pathways were the top gene ontology pathways for Cluster 3. Epigenetic Aging in lupus: Epigenetic clocks based on DNA methylation have been shown to correlate with biological aging and predict adverse health outcomes. Chronological age does not capture the inter-individual variability of underlying biological processes in individuals. Accelerated aging based on the epigenetic clock is defined as having a higher predicted biological age than chronological age. DNA methylation data were generated using the Illumina Infinium Human MethylationEPIC BeadChip from whole blood samples in a multi-ethnic cohort of 323 SLE cases and 99 healthy controls. Background correction with dye-bias normalization using the preprocessNoob function in the minfi R package was performed on the methylation data. The methylClock R package was used to estimate the biological age using the DNAm Pheno Age clock. The DNAm Pheno Age uses 513 CpGs to estimate the biological age. DNAm Pheno Age was then regressed on chronological age, SLE, female, batch and race/ethnicity to test for association of accelerated aging in SLE patients compared to healthy controls. Separate models were also run for association testing of income and education. DNAm Pheno Age was regressed on chronological age, income or education, female, batch and race/ethnicity. In a case only analysis of SLE patients, association testing was run in separate models for anti-DsDNA status, presence of lupus nephritis and the SLE severity index by adjusting for chronological age, female, batch and race/ethnicity. Accelerated aging was associated with SLE status while adjusting for chronological age, sex, race/ethnicity and batch. Among SLE patients, accelerated aging was associated with positive anti-dsDNA status, presence of lupus nephritis, and SLE severity. Sjogren's Syndrome subtypes: Sjogren's Syndrome (SS) is a heterogenous disorder which poses challenges in the diagnosis, management and treatment of the disease. In order to characterize disease subtypes, we performed a cluster analysis of genome-wide methylation data. The genome-wide methylation data was profiled on DNA extracted from gland tissue using the Illumina 450k Infinium Methylation BeadChip in 28 patients and the Infinium MethylationEPIC (EPIC) platform for 103 patients. All 131 patients exhibited at least one symptom related to Sjogren's. The 131 patients include 64 patients who were classified as SS cases based on the ACR-EULAR 2016 classification criteria and 67 symptomatic non-cases. For the analysis, the intersection of CpG sites from the 450K chip and the EPIC ship resulted in an initial set of 452,832 CpG sites. After quality control processing, 336,040 CpG sites remained. Hierarchical clustering analysis identified four clusters which distinguished clinically severe and mild subgroups of SS. Clusters 1 and 2 showed the largest distance from clusters 3 and 4. Clusters 1 and 2 consisted of 93% of the SS cases while clusters 3 and 4 consisted of 73% of the non-cases. An analysis of clinical variables show that patients in clusters 1 and 2 had more severe clinical features compared to clusters 3 and 4. Differential methylation analysis showed that hypomethylation at the MHC and hypermethylation at other genome regions characterize the epigenetic differences between these SS subgroups. Differences in allele frequencies at established SS risk loci showed that severe SS cases had higher risk allele frequencies at SNPs which tag HLA-DRB1, HLA-DQA1, HLA-DQB1 and HLA-DQA2 within the MHC region. These results show that epigenetic profiling provides additional information about the biological subtypes of SS that could contribute to improved SS classification. A manuscript for the Sjogren's syndrome subtypes has been submitted for publication and is currently under revision.