Epigenetic scores of blood-based proteins as biomarkers of general cognitive function and brain health.

BACKGROUND: Epigenetic Scores (EpiScores) for blood protein levels have been associated with disease outcomes and measures of brain health, highlighting their potential usefulness as clinical biomarkers. They are typically derived via penalised regression, whereby a linear weighted sum of DNA methylation (DNAm) levels at CpG sites are predictive of protein levels. Here, we examine 84 previously published protein EpiScores as possible biomarkers of cross-sectional and longitudinal measures of general cognitive function and brain health, and incident dementia across three independent cohorts. RESULTS: Using 84 protein EpiScores as candidate biomarkers, associations with general cognitive function (both cross-sectionally and longitudinally) were tested in three independent cohorts: Generation Scotland (GS), and the Lothian Birth Cohorts of 1921 and 1936 (LBC1921 and LBC1936, respectively). A meta-analysis of general cognitive functioning results in all three cohorts identified 18 EpiScore associations (absolute meta-analytic standardised estimates ranged from 0.03 to 0.14, median of 0.04, PFDR < 0.05). Several associations were also observed between EpiScores and global brain volumetric measures in the LBC1936. An EpiScore for the S100A9 protein (a known Alzheimer disease biomarker) was associated with general cognitive functioning (meta-analytic standardised beta: - 0.06, P = 1.3 × 10-9), and with time-to-dementia in GS (Hazard ratio 1.24, 95% confidence interval 1.08-1.44, P = 0.003), but not in LBC1936 (Hazard ratio 1.11, P = 0.32). CONCLUSIONS: EpiScores might make a contribution to the risk profile of poor general cognitive function and global brain health, and risk of dementia, however these scores require replication in further studies.

Genetic control of DNA methylation is largely shared across European and East Asian populations.

DNA methylation is an ideal trait to study the extent of the shared genetic control across ancestries, effectively providing hundreds of thousands of model molecular traits with large QTL effect sizes. We investigate cis DNAm QTLs in three European (n = 3701) and two East Asian (n = 2099) cohorts to quantify the similarities and differences in the genetic architecture across populations. We observe 80,394 associated mQTLs (62.2% of DNAm probes with significant mQTL) to be significant in both ancestries, while 28,925 mQTLs (22.4%) are identified in only a single ancestry. mQTL effect sizes are highly conserved across populations, with differences in mQTL discovery likely due to differences in allele frequency of associated variants and differing linkage disequilibrium between causal variants and assayed SNPs. This study highlights the overall similarity of genetic control across ancestries and the value of ancestral diversity in increasing the power to detect associations and enhancing fine mapping resolution.

General and specific patterns of cortical gene expression as spatial correlates of complex cognitive functioning.

Gene expression varies across the brain. This spatial patterning denotes specialised support for particular brain functions. However, the way that a given gene's expression fluctuates across the brain may be governed by general rules. Quantifying patterns of spatial covariation across genes would offer insights into the molecular characteristics of brain areas supporting, for example, complex cognitive functions. Here, we use principal component analysis to separate general and unique gene regulatory associations with cortical substrates of cognition. We find that the region-to-region variation in cortical expression profiles of 8235 genes covaries across two major principal components: gene ontology analysis suggests these dimensions are characterised by downregulation and upregulation of cell-signalling/modification and transcription factors. We validate these patterns out-of-sample and across different data processing choices. Brain regions more strongly implicated in general cognitive functioning (g; 3 cohorts, total meta-analytic N = 39,519) tend to be more balanced between downregulation and upregulation of both major components (indicated by regional component scores). We then identify a further 29 genes as candidate cortical spatial correlates of g, beyond the patterning of the two major components (|ß| range = 0.18 to 0.53). Many of these genes have been previously associated with clinical neurodegenerative and psychiatric disorders, or with other health-related phenotypes. The results provide insights into the cortical organisation of gene expression and its association with individual differences in cognitive functioning.

A Large-Scale Genome-Wide Study of Gene-Sleep Duration Interactions for Blood Pressure in 811,405 Individuals from Diverse Populations.

Although both short and long sleep duration are associated with elevated hypertension risk, our understanding of their interplay with biological pathways governing blood pressure remains limited. To address this, we carried out genome-wide cross-population gene-by-short-sleep and long-sleep duration interaction analyses for three blood pressure traits (systolic, diastolic, and pulse pressure) in 811,405 individuals from diverse population groups. We discover 22 novel gene-sleep duration interaction loci for blood pressure, mapped to genes involved in neurological, thyroidal, bone metabolism, and hematopoietic pathways. Non-overlap between short sleep (12) and long sleep (10) interactions underscores the plausibility of distinct influences of both sleep duration extremes in cardiovascular health. With several of our loci reflecting specificity towards population background or sex, our discovery sheds light on the importance of embracing granularity when addressing heterogeneity entangled in gene-environment interactions, and in therapeutic design approaches for blood pressure management.

Understanding species-specific and conserved RNA-protein interactions in vivo and in vitro.

While evolution is often considered from a DNA- and protein-centric view, RNA-based regulation can also impact gene expression and protein sequences. Here we examined interspecies differences in RNA-protein interactions using the conserved neuronal RNA binding protein, Unkempt (UNK) as model. We find that roughly half of mRNAs bound in human are also bound in mouse. Unexpectedly, even when transcript-level binding was conserved across species differential motif usage was prevalent. To understand the biochemical basis of UNK-RNA interactions, we reconstituted the human and mouse UNK-RNA interactomes using a high-throughput biochemical assay. We uncover detailed features driving binding, show that in vivo patterns are captured in vitro, find that highly conserved sites are the strongest bound, and associate binding strength with downstream regulation. Furthermore, subtle sequence differences surrounding motifs are key determinants of species-specific binding. We highlight the complex features driving protein-RNA interactions and how these evolve to confer species-specific regulation.

Multi-trait analysis characterizes the genetics of thyroid function and identifies causal associations with clinical implications.

To date only a fraction of the genetic footprint of thyroid function has been clarified. We report a genome-wide association study meta-analysis of thyroid function in up to 271,040 individuals of European ancestry, including reference range thyrotropin (TSH), free thyroxine (FT4), free and total triiodothyronine (T3), proxies for metabolism (T3/FT4 ratio) as well as dichotomized high and low TSH levels. We revealed 259 independent significant associations for TSH (61% novel), 85 for FT4 (67% novel), and 62 novel signals for the T3 related traits. The loci explained 14.1%, 6.0%, 9.5% and 1.1% of the total variation in TSH, FT4, total T3 and free T3 concentrations, respectively. Genetic correlations indicate that TSH associated loci reflect the thyroid function determined by free T3, whereas the FT4 associations represent the thyroid hormone metabolism. Polygenic risk score and Mendelian randomization analyses showed the effects of genetically determined variation in thyroid function on various clinical outcomes, including cardiovascular risk factors and diseases, autoimmune diseases, and cancer. In conclusion, our results improve the understanding of thyroid hormone physiology and highlight the pleiotropic effects of thyroid function on various diseases.

Genome-wide association study of preserved ratio impaired spirometry (PRISm).

BACKGROUND: Preserved ratio impaired spirometry (PRISm) is defined as a forced expiratory volume in 1 s (FEV1) <80% predicted and FEV1/forced vital capacity ≥0.70. PRISm is associated with respiratory symptoms and comorbidities. Our objective was to discover novel genetic signals for PRISm and see if they provide insight into the pathogenesis of PRISm and associated comorbidities. METHODS: We undertook a genome-wide association study (GWAS) of PRISm in UK Biobank participants (Stage 1), and selected single nucleotide polymorphisms (SNPs) reaching genome-wide significance for replication in 13 cohorts (Stage 2). A combined meta-analysis of Stage 1 and Stage 2 was done to determine top SNPs. We used cross-trait linkage disequilibrium score regression to estimate genome-wide genetic correlation between PRISm and pulmonary and extrapulmonary traits. Phenome-wide association studies of top SNPs were performed. RESULTS: 22 signals reached significance in the joint meta-analysis, including four signals novel for lung function. A strong genome-wide genetic correlation (rg) between PRISm and spirometric COPD (rg=0.62, p<0.001) was observed, and genetic correlation with type 2 diabetes (rg=0.12, p=0.007). Phenome-wide association studies showed that 18 of 22 signals were associated with diabetic traits and seven with blood pressure traits. CONCLUSION: This is the first GWAS to successfully identify SNPs associated with PRISm. Four of the signals, rs7652391 (nearest gene MECOM), rs9431040 (HLX), rs62018863 (TMEM114) and rs185937162 (HLA-B), have not been described in association with lung function before, demonstrating the utility of using different lung function phenotypes in GWAS. Genetic factors associated with PRISm are strongly correlated with risk of both other lung diseases and extrapulmonary comorbidity.

Integration of datasets for individual prediction of DNA methylation-based biomarkers.

BACKGROUND: Epigenetic scores (EpiScores) can provide biomarkers of lifestyle and disease risk. Projecting new datasets onto a reference panel is challenging due to separation of technical and biological variation with array data. Normalisation can standardise data distributions but may also remove population-level biological variation. RESULTS: We compare two birth cohorts (Lothian Birth Cohorts of 1921 and 1936 - nLBC1921 = 387 and nLBC1936 = 498) with blood-based DNA methylation assessed at the same chronological age (79 years) and processed in the same lab but in different years and experimental batches. We examine the effect of 16 normalisation methods on a novel BMI EpiScore (trained in an external cohort, n = 18,413), and Horvath's pan-tissue DNA methylation age, when the cohorts are normalised separately and together. The BMI EpiScore explains a maximum variance of R2=24.5% in BMI in LBC1936 (SWAN normalisation). Although there are cross-cohort R2 differences, the normalisation method makes a minimal difference to within-cohort estimates. Conversely, a range of absolute differences are seen for individual-level EpiScore estimates for BMI and age when cohorts are normalised separately versus together. While within-array methods result in identical EpiScores whether a cohort is normalised on its own or together with the second dataset, a range of differences is observed for between-array methods. CONCLUSIONS: Normalisation methods returning similar EpiScores, whether cohorts are analysed separately or together, will minimise technical variation when projecting new data onto a reference panel. These methods are important for cases where raw data is unavailable and joint normalisation of cohorts is computationally expensive.

Copy-number variants and polygenic risk for intelligence confer risk for autism spectrum disorder irrespective of their effects on cognitive ability.

Rare copy number variants (CNVs) and polygenic risk for intelligence (PRS-IQ) both confer risk for autism spectrum disorder (ASD) but have opposing effects on cognitive ability. The field has struggled to disentangle the effects of these two classes of genomic variants on cognitive ability from their effects on ASD risk, in part because previous studies did not include controls with cognitive measures. We aim to investigate the impact of these genomic variants on ASD risk while adjusting for their known effects on cognitive ability. In a cohort of 8,426 subjects with ASD and 169,804 controls with cognitive assessments, we found that rare coding CNVs and PRS-IQ increased ASD risk, even after adjusting for their effects on cognitive ability. Bottom decile PRS-IQ and CNVs both decreased cognitive ability but had opposing effects on ASD risk. Models combining both classes of variants showed that the effects of rare CNVs and PRS-IQ on ASD risk and cognitive ability were largely additive, further suggesting that risk for ASD is conferred independently from its effects on cognitive ability. Despite imparting mostly additive effects on ASD risk, rare CNVs and PRS-IQ showed opposing effects on core and associated features and developmental history among subjects with ASD. Our findings suggest that cognitive ability itself may not be the factor driving the underlying risk for ASD conferred by these two classes of genomic variants. In other words, ASD risk and cognitive ability may be two distinct manifestations of CNVs and PRS-IQ. This study also highlights the challenge of understanding how genetic risk for ASD maps onto its dimensional traits.

Gene-educational attainment interactions in a multi-population genome-wide meta-analysis identify novel lipid loci.

Introduction: Educational attainment, widely used in epidemiologic studies as a surrogate for socioeconomic status, is a predictor of cardiovascular health outcomes. Methods: A two-stage genome-wide meta-analysis of low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), and triglyceride (TG) levels was performed while accounting for gene-educational attainment interactions in up to 226,315 individuals from five population groups. We considered two educational attainment variables: "Some College" (yes/no, for any education beyond high school) and "Graduated College" (yes/no, for completing a 4-year college degree). Genome-wide significant (p < 5 × 10-8) and suggestive (p < 1 × 10-6) variants were identified in Stage 1 (in up to 108,784 individuals) through genome-wide analysis, and those variants were followed up in Stage 2 studies (in up to 117,531 individuals). Results: In combined analysis of Stages 1 and 2, we identified 18 novel lipid loci (nine for LDL, seven for HDL, and two for TG) by two degree-of-freedom (2 DF) joint tests of main and interaction effects. Four loci showed significant interaction with educational attainment. Two loci were significant only in cross-population analyses. Several loci include genes with known or suggested roles in adipose (FOXP1, MBOAT4, SKP2, STIM1, STX4), brain (BRI3, FILIP1, FOXP1, LINC00290, LMTK2, MBOAT4, MYO6, SENP6, SRGAP3, STIM1, TMEM167A, TMEM30A), and liver (BRI3, FOXP1) biology, highlighting the potential importance of brain-adipose-liver communication in the regulation of lipid metabolism. An investigation of the potential druggability of genes in identified loci resulted in five gene targets shown to interact with drugs approved by the Food and Drug Administration, including genes with roles in adipose and brain tissue. Discussion: Genome-wide interaction analysis of educational attainment identified novel lipid loci not previously detected by analyses limited to main genetic effects.

Identification of circulating proteins associated with general cognitive function among middle-aged and older adults.

Identifying circulating proteins associated with cognitive function may point to biomarkers and molecular process of cognitive impairment. Few studies have investigated the association between circulating proteins and cognitive function. We identify 246 protein measures quantified by the SomaScan assay as associated with cognitive function (p < 4.9E-5, n up to 7289). Of these, 45 were replicated using SomaScan data, and three were replicated using Olink data at Bonferroni-corrected significance. Enrichment analysis linked the proteins associated with general cognitive function to cell signaling pathways and synapse architecture. Mendelian randomization analysis implicated higher levels of NECTIN2, a protein mediating viral entry into neuronal cells, with higher Alzheimer's disease (AD) risk (p = 2.5E-26). Levels of 14 other protein measures were implicated as consequences of AD susceptibility (p < 2.0E-4). Proteins implicated as causes or consequences of AD susceptibility may provide new insight into the potential relationship between immunity and AD susceptibility as well as potential therapeutic targets.

Uridylation of the histone mRNA stem-loop weakens binding interactions with SLBP while maintaining interactions with 3'hExo.

Histone mRNA degradation is controlled by the unique 3' stem-loop of histone mRNA and the stem-loop binding protein (SLBP). As part of this process, the 3' stem-loop is trimmed by the histone-specific 3' exonuclease (3'hExo) and uridylated by the terminal uridylyl transferase 7 (TUT7), creating partially degraded intermediates with short uridylations. The role of these uridylations in degradation is not fully understood. Our work examines changes in the stability of the ternary complex created by trimming and uridylation of the stem-loop to better understand the role of this process in the histone mRNA life cycle. In this study, we used fluorescence polarization and electrophoretic mobility shift assays to demonstrate that both SLBP and 3'hExo can bind to uridylated and partially degraded stem-loop intermediates, although with lower affinity. We further characterized this complex by performing 1-µs molecular dynamics simulations using the AMBER force field and Nanoscale Molecular Dynamics (NAMD). These simulations show that while uridylation helps maintain the overall shape of the stem-loop, the combination of uridylation and dephosphorylation of the TPNK motif in SLBP disrupts key RNA-protein interactions. They also demonstrate that uridylation allows 3'hExo to maintain contact with the stem-loop after partial degradation and plays a role in disrupting key base pairs in partially degraded histone mRNA intermediates. Together, these experiments and simulations suggest that trimming by 3'hExo, uridylation, and SLBP dephosphorylation weakens both RNA-protein interactions and the stem-loop itself. Our results further elucidate the role of uridylation and SLBP dephosphorylation in the early stages of histone mRNA degradation.

Predictive blood biomarkers and brain changes associated with age-related cognitive decline.

Growing evidence supports the use of plasma levels of tau phosphorylated at threonine 181, amyloid-ß, neurofilament light and glial fibrillary acidic protein as promising biomarkers for Alzheimer's disease. While these blood biomarkers are promising for distinguishing people with Alzheimer's disease from healthy controls, their predictive validity for age-related cognitive decline without dementia remains unclear. Further, while tau phosphorylated at threonine 181 is a promising biomarker, the distribution of this phospho-epitope of tau in the brain is unknown. Here, we tested whether plasma levels of tau phosphorylated at threonine 181, amyloid-ß, neurofilament light and fibrillary acidic protein predict cognitive decline between ages 72 and 82 in 195 participants in the Lothian birth cohorts 1936 study of cognitive ageing. We further examined post-mortem brain samples from temporal cortex to determine the distribution of tau phosphorylated at threonine 181 in the brain. Several forms of tau phosphorylated at threonine 181 have been shown to contribute to synapse degeneration in Alzheimer's disease, which correlates closely with cognitive decline in this form of dementia, but to date, there have not been investigations of whether tau phosphorylated at threonine 181 is found in synapses in Alzheimer's disease or healthy ageing brain. It was also previously unclear whether tau phosphorylated at threonine 181 accumulated in dystrophic neurites around plaques, which could contribute to tau leakage to the periphery due to impaired membrane integrity in dystrophies. Brain homogenate and biochemically enriched synaptic fractions were examined with western blot to examine tau phosphorylated at threonine 181 levels between groups (n = 10-12 per group), and synaptic and astrocytic localization of tau phosphorylated at threonine 181 were examined using array tomography (n = 6-15 per group), and localization of tau phosphorylated at threonine 181 in plaque-associated dystrophic neurites with associated gliosis were examined with standard immunofluorescence (n = 8-9 per group). Elevated baseline plasma tau phosphorylated at threonine 181, neurofilament light and fibrillary acidic protein predicted steeper general cognitive decline during ageing. Further, increasing tau phosphorylated at threonine 181 over time predicted general cognitive decline in females only. Change in plasma tau phosphorylated at threonine 181 remained a significant predictor of g factor decline when taking into account Alzheimer's disease polygenic risk score, indicating that the increase of blood tau phosphorylated at threonine 181 in this cohort was not only due to incipient Alzheimer's disease. Tau phosphorylated at threonine 181 was observed in synapses and astrocytes in both healthy ageing and Alzheimer's disease brain. We observed that a significantly higher proportion of synapses contain tau phosphorylated at threonine 181 in Alzheimer's disease relative to aged controls. Aged controls with pre-morbid lifetime cognitive resilience had significantly more tau phosphorylated at threonine 181 in fibrillary acidic protein-positive astrocytes than those with pre-morbid lifetime cognitive decline. Further, tau phosphorylated at threonine 181 was found in dystrophic neurites around plaques and in some neurofibrillary tangles. The presence of tau phosphorylated at threonine 181 in plaque-associated dystrophies may be a source of leakage of tau out of neurons that eventually enters the blood. Together, these data indicate that plasma tau phosphorylated at threonine 181, neurofilament light and fibrillary acidic protein may be useful biomarkers of age-related cognitive decline, and that efficient clearance of tau phosphorylated at threonine 181 by astrocytes may promote cognitive resilience.

Unraveling Neuro-Proteogenomic Landscape and Therapeutic Implications for Human Behaviors and Psychiatric Disorders.

Understanding the genetic basis of neuro-related proteins is essential for dissecting the molecular basis of human behavioral traits and the disease etiology of neuropsychiatric disorders. Here, the SCALLOP Consortium conducted a genome-wide association meta-analysis of over 12,500 individuals for 184 neuro-related proteins in human plasma. The analysis identified 117 cis-regulatory protein quantitative trait loci (cis-pQTL) and 166 trans-pQTL. The mapped pQTL capture on average 50% of each protein's heritability. Mendelian randomization analyses revealed multiple proteins showing potential causal effects on neuro-related traits such as sleeping, smoking, feelings, alcohol intake, mental health, and psychiatric disorders. Integrating with established drug information, we validated 13 out of 13 matched combinations of protein targets and diseases or side effects with available drugs, while suggesting hundreds of re-purposing and new therapeutic targets. This consortium effort provides a large-scale proteogenomic resource for biomedical research on human behaviors and other neuro-related phenotypes.

General and specific patterns of cortical gene expression as spatial correlates of complex cognitive functioning.

Gene expression varies across the brain. This spatial patterning denotes specialised support for particular brain functions. However, the way that a given gene's expression fluctuates across the brain may be governed by general rules. Quantifying patterns of spatial covariation across genes would offer insights into the molecular characteristics of brain areas supporting, for example, complex cognitive functions. Here, we use principal component analysis to separate general and unique gene regulatory associations with cortical substrates of cognition. We find that the region-to-region variation in cortical expression profiles of 8235 genes covaries across two major principal components : gene ontology analysis suggests these dimensions are characterised by downregulation and upregulation of cell-signalling/modification and transcription factors. We validate these patterns out-of-sample and across different data processing choices. Brain regions more strongly implicated in general cognitive functioning (g; 3 cohorts, total meta-analytic N = 39,519) tend to be more balanced between downregulation and upregulation of both major components (indicated by regional component scores). We then identify a further 41 genes as candidate cortical spatial correlates of g, beyond the patterning of the two major components (|ß| range = 0.15 to 0.53). Many of these genes have been previously associated with clinical neurodegenerative and psychiatric disorders, or with other health-related phenotypes. The results provide insights into the cortical organisation of gene expression and its association with individual differences in cognitive functioning.

Refining epigenetic prediction of chronological and biological age.

BACKGROUND: Epigenetic clocks can track both chronological age (cAge) and biological age (bAge). The latter is typically defined by physiological biomarkers and risk of adverse health outcomes, including all-cause mortality. As cohort sample sizes increase, estimates of cAge and bAge become more precise. Here, we aim to develop accurate epigenetic predictors of cAge and bAge, whilst improving our understanding of their epigenomic architecture. METHODS: First, we perform large-scale (N = 18,413) epigenome-wide association studies (EWAS) of chronological age and all-cause mortality. Next, to create a cAge predictor, we use methylation data from 24,674 participants from the Generation Scotland study, the Lothian Birth Cohorts (LBC) of 1921 and 1936, and 8 other cohorts with publicly available data. In addition, we train a predictor of time to all-cause mortality as a proxy for bAge using the Generation Scotland cohort (1214 observed deaths). For this purpose, we use epigenetic surrogates (EpiScores) for 109 plasma proteins and the 8 component parts of GrimAge, one of the current best epigenetic predictors of survival. We test this bAge predictor in four external cohorts (LBC1921, LBC1936, the Framingham Heart Study and the Women's Health Initiative study). RESULTS: Through the inclusion of linear and non-linear age-CpG associations from the EWAS, feature pre-selection in advance of elastic net regression, and a leave-one-cohort-out (LOCO) cross-validation framework, we obtain cAge prediction with a median absolute error equal to 2.3 years. Our bAge predictor was found to slightly outperform GrimAge in terms of the strength of its association to survival (HRGrimAge = 1.47 [1.40, 1.54] with p = 1.08 × 10-52, and HRbAge = 1.52 [1.44, 1.59] with p = 2.20 × 10-60). Finally, we introduce MethylBrowsR, an online tool to visualise epigenome-wide CpG-age associations. CONCLUSIONS: The integration of multiple large datasets, EpiScores, non-linear DNAm effects, and new approaches to feature selection has facilitated improvements to the blood-based epigenetic prediction of biological and chronological age.

Genetic mechanisms of 184 neuro-related proteins in human plasma.

Understanding the genetic basis of neuro-related proteins is essential for dissecting the disease etiology of neuropsychiatric disorders and other complex traits and diseases. Here, the SCALLOP Consortium conducted a genome-wide association meta-analysis of over 12,500 individuals for 184 neuro-reiated proteins in human plasma. The analysis identified 117 cis-regulatory protein quantitative trait loci (cis-pQTL) and 166 trans-pQTL. The mapped pQTL capture on average 50% of each protein's heritability. Mendelian randomization analyses revealed multiple proteins showing potential causal effects on neuro-reiated traits as well as complex diseases such as hypertension, high cholesterol, immune-related disorders, and psychiatric disorders. Integrating with established drug information, we validated 13 combinations of protein targets and diseases or side effects with available drugs, while suggesting hundreds of re-purposing and new therapeutic targets for diseases and comorbidities. This consortium effort provides a large-scale proteogenomic resource for biomedical research.

Epigenome-Wide Association Study Reveals CpG Sites Associated with Thyroid Function and Regulatory Effects on KLF9.

Background: Thyroid hormones play a key role in differentiation and metabolism and are known regulators of gene expression through both genomic and epigenetic processes including DNA methylation. The aim of this study was to examine associations between thyroid hormones and DNA methylation. Methods: We carried out a fixed-effect meta-analysis of epigenome-wide association study (EWAS) of blood DNA methylation sites from 8 cohorts from the ThyroidOmics Consortium, incorporating up to 7073 participants of both European and African ancestry, implementing a discovery and replication stage. Statistical analyses were conducted using normalized beta CpG values as dependent and log-transformed thyrotropin (TSH), free thyroxine, and free triiodothyronine levels, respectively, as independent variable in a linear model. The replicated findings were correlated with gene expression levels in whole blood and tested for causal influence of TSH and free thyroxine by two-sample Mendelian randomization (MR). Results: Epigenome-wide significant associations (p-value <1.1E-7) of three CpGs for free thyroxine, five for free triiodothyronine, and two for TSH concentrations were discovered and replicated (combined p-values = 1.5E-9 to 4.3E-28). The associations included CpG sites annotated to KLF9 (cg00049440) and DOT1L (cg04173586) that overlap with all three traits, consistent with hypothalamic-pituitary-thyroid axis physiology. Significant associations were also found for CpGs in FKBP5 for free thyroxine, and at CSNK1D/LINCO1970 and LRRC8D for free triiodothyronine. MR analyses supported a causal effect of thyroid status on DNA methylation of KLF9. DNA methylation of cg00049440 in KLF9 was inversely correlated with KLF9 gene expression in blood. The CpG at CSNK1D/LINC01970 overlapped with thyroid hormone receptor alpha binding peaks in liver cells. The total additive heritability of the methylation levels of the six significant CpG sites was between 25% and 57%. Significant methylation QTLs were identified for CpGs at KLF9, FKBP5, LRRC8D, and CSNK1D/LINC01970. Conclusions: We report novel associations between TSH, thyroid hormones, and blood-based DNA methylation. This study advances our understanding of thyroid hormone action particularly related to KLF9 and serves as a proof-of-concept that integrations of EWAS with other -omics data can provide a valuable tool for unraveling thyroid hormone signaling in humans by complementing and feeding classical in vitro and animal studies.