Twin-based Mendelian Randomization Analyses Highlight Smoking's Effects on Blood DNA Methylation, with Putative Reverse Causation.

Epigenome-wide association studies (EWAS) aim to identify differentially methylated loci associated with complex traits and disorders. EWAS of cigarette smoking shows some of the most widespread DNA methylation (DNAm) associations in blood. However, traditional EWAS cannot differentiate between causation and confounding, leading to ambiguity in etiological interpretations. Here, we apply an integrated approach combining Mendelian Randomization and twin-based Direction-of-Causation analyses (MR-DoC) to examine causality underlying smoking-associated blood DNAm changes in the Netherlands Twin Register (N=2577). Evidence across models suggests that current smoking's causal effects on DNAm likely drive many of the previous EWAS findings, implicating functional pathways relevant to several adverse health outcomes of smoking, including hemopoiesis, cell- and neuro-development, and immune regulation. Additionally, we find evidence of potential reverse causal influences at some DNAm sites, with 17 of these sites enriched for gene regulatory functional elements in the brain. The top three sites with evidence of DNAm's effects on smoking annotate to genes involved in G protein-coupled receptor signaling (GNG7, RGS3) and innate immune response (SLC15A4), elucidating potential biological risk factors for smoking. This study highlights the utility of integrating genotypic and DNAm measures in twin cohorts to clarify the causal relationships between health behaviors and blood DNAm.

Characterisation of ethnic differences in DNA methylation between UK-resident South Asians and Europeans.

Ethnic differences in non-communicable disease risk have been described between individuals of South Asian and European ethnicity that are only partially explained by genetics and other known risk factors. DNA methylation is one underexplored mechanism that may explain differences in disease risk. Currently, there is little knowledge of how DNA methylation varies between South Asian and European ethnicities. This study characterised differences in blood DNA methylation between individuals of self-reported European and South Asian ethnicity from two UK-based cohorts: Southall and Brent Revisited and Born in Bradford. DNA methylation differences between ethnicities were widespread throughout the genome (n = 16,433 CpG sites, 3.4% sites tested). Specifically, 76% of associations were attributable to ethnic differences in cell composition with fewer effects attributable to smoking and genetic variation. Ethnicity-associated CpG sites were enriched for EWAS Catalog phenotypes including metabolites. This work highlights the need to consider ethnic diversity in epigenetic research.

Complex trait methylation scores in the prediction of major depressive disorder.

BACKGROUND: DNA methylation (DNAm) is associated with time-varying environmental factors that contribute to major depressive disorder (MDD) risk. We sought to test whether DNAm signatures of lifestyle and biochemical factors were associated with MDD to reveal dynamic biomarkers of MDD risk that may be amenable to lifestyle interventions. METHODS: Here, we calculated methylation scores (MS) at multiple p-value thresholds for lifestyle (BMI, smoking, alcohol consumption, and educational attainment) and biochemical (high-density lipoprotein (HDL) and total cholesterol) factors in Generation Scotland (GS) (N=9,502) and in a replication cohort (ALSPACadults, N=565), using CpG sites reported in previous well-powered methylome-wide association studies. We also compared their predictive accuracy for MDD to a MDD MS in an independent GS sub-sample (N=4,432). FINDINGS: Each trait MS was significantly associated with its corresponding phenotype in GS (ßrange=0.089-1.457) and in ALSPAC (ßrange=0.078-2.533). Each MS was also significantly associated with MDD before and after adjustment for its corresponding phenotype in GS (ßrange=0.053-0.145). After accounting for relevant lifestyle factors, MS for educational attainment (ß=0.094) and alcohol consumption (MSp-value<0.01-0.5; ßrange=-0.069-0.083) remained significantly associated with MDD in GS. Smoking (AUC=0.569) and educational attainment (AUC=0.585) MSs could discriminate MDD from controls better than the MDD MS (AUC=0.553) in the independent GS sub-sample. Analyses implicating MDD did not replicate across ALSPAC, although the direction of effect was consistent for all traits when adjusting for the MS corresponding phenotypes. INTERPRETATION: We showed that lifestyle and biochemical MS were associated with MDD before and after adjustment for their corresponding phenotypes (pnominal<0.05), but not when smoking, alcohol consumption, and BMI were also included as covariates. MDD results did not replicate in the smaller, female-only independent ALSPAC cohort (NALSPAC=565; NGS=9,502), potentially due to demographic differences or low statistical power, but effect sizes were consistent with the direction reported in GS. DNAm scores for modifiable MDD risk factors may contribute to disease vulnerability and, in some cases, explain additional variance to their observed phenotypes. FUNDING: Wellcome Trust.

Genome-wide association meta-analysis identifies 29 new acne susceptibility loci.

Acne vulgaris is a highly heritable skin disorder that primarily impacts facial skin. Severely inflamed lesions may leave permanent scars that have been associated with long-term psychosocial consequences. Here, we perform a GWAS meta-analysis comprising 20,165 individuals with acne from nine independent European ancestry cohorts. We identify 29 novel genome-wide significant loci and replicate 14 of the 17 previously identified risk loci, bringing the total number of reported acne risk loci to 46. Using fine-mapping and eQTL colocalisation approaches, we identify putative causal genes at several acne susceptibility loci that have previously been implicated in Mendelian hair and skin disorders, including pustular psoriasis. We identify shared genetic aetiology between acne, hormone levels, hormone-sensitive cancers and psychiatric traits. Finally, we show that a polygenic risk score calculated from our results explains up to 5.6% of the variance in acne liability in an independent cohort.

Epigenetic Regulation of F2RL3 Associates With Myocardial Infarction and Platelet Function.

BACKGROUND: DNA hypomethylation at the F2RL3 (F2R like thrombin or trypsin receptor 3) locus has been associated with both smoking and atherosclerotic cardiovascular disease; whether these smoking-related associations form a pathway to disease is unknown. F2RL3 encodes protease-activated receptor 4, a potent thrombin receptor expressed on platelets. Given the role of thrombin in platelet activation and the role of thrombus formation in myocardial infarction, alterations to this biological pathway could be important for ischemic cardiovascular disease. METHODS: We conducted multiple independent experiments to assess whether DNA hypomethylation at F2RL3 in response to smoking is associated with risk of myocardial infarction via changes to platelet reactivity. Using cohort data (N=3205), we explored the relationship between smoking, DNA hypomethylation at F2RL3, and myocardial infarction. We compared platelet reactivity in individuals with low versus high DNA methylation at F2RL3 (N=41). We used an in vitro model to explore the biological response of F2RL3 to cigarette smoke extract. Finally, a series of reporter constructs were used to investigate how differential methylation could impact F2RL3 gene expression. RESULTS: Observationally, DNA methylation at F2RL3 mediated an estimated 34% of the smoking effect on increased risk of myocardial infarction. An association between methylation group (low/high) and platelet reactivity was observed in response to PAR4 (protease-activated receptor 4) stimulation. In cells, cigarette smoke extract exposure was associated with a 4.9% to 9.3% reduction in DNA methylation at F2RL3 and a corresponding 1.7-(95% CI, 1.2-2.4, P=0.04) fold increase in F2RL3 mRNA. Results from reporter assays suggest the exon 2 region of F2RL3 may help control gene expression. CONCLUSIONS: Smoking-induced epigenetic DNA hypomethylation at F2RL3 appears to increase PAR4 expression with potential downstream consequences for platelet reactivity. Combined evidence here not only identifies F2RL3 DNA methylation as a possible contributory pathway from smoking to cardiovascular disease risk but from any feature potentially influencing F2RL3 regulation in a similar manner.

Comparison of DNA methylation clocks in Black South African men.

Aims: DNA methylation clocks are widely used to estimate biological age, although limited data are available on non-European ethnicities. This manuscript characterizes the behavior of five DNA methylation clocks in 120 older Black South African men. Methods: The age estimation accuracy of the Horvath, Hannum and skin and blood clocks and the relative age-related mortality risk and predicted time to death portrayed by the PhenoAge and GrimAge biomarkers are investigated, respectively. Results: The results confirm the tendency of DNA methylation clocks to underestimate the biological age of older individuals. GrimAge more accurately characterizes biological decline in this African cohort compared with PhenoAge owing to the unique inclusion of smoking-related damage in the GrimAge estimate. Conclusions: Each clock provides a different fraction of information regarding the aging body. It is essential to continue studying under-represented population groups to ensure methylation-derived indicators are robust and useful in all populations.

Assessing the role of genome-wide DNA methylation between smoking and risk of lung cancer using repeated measurements: the HUNT study.

BACKGROUND: It is unclear if smoking-related DNA methylation represents a causal pathway between smoking and risk of lung cancer. We sought to identify novel smoking-related DNA methylation sites in blood, with repeated measurements, and to appraise the putative role of DNA methylation in the pathway between smoking and lung cancer development. METHODS: We derived a nested case-control study from the Trøndelag Health Study (HUNT), including 140 incident patients who developed lung cancer during 2009-13 and 140 controls. We profiled 850 K DNA methylation sites (Illumina Infinium EPIC array) in DNA extracted from blood that was collected in HUNT2 (1995-97) and HUNT3 (2006-08) for the same individuals. Epigenome-wide association studies (EWAS) were performed for a detailed smoking phenotype and for lung cancer. Two-step Mendelian randomization (MR) analyses were performed to assess the potential causal effect of smoking on DNA methylation as well as of DNA methylation (13 sites as putative mediators) on risk of lung cancer. RESULTS: The EWAS for smoking in HUNT2 identified associations at 76 DNA methylation sites (P < 5 × 10-8), including 16 novel sites. Smoking was associated with DNA hypomethylation in a dose-response relationship among 83% of the 76 sites, which was confirmed by analyses using repeated measurements from blood that was collected at 11 years apart for the same individuals. Two-step MR analyses showed evidence for a causal effect of smoking on DNA methylation but no evidence for a causal link between DNA methylation and the risk of lung cancer. CONCLUSIONS: DNA methylation modifications in blood did not seem to represent a causal pathway linking smoking and the lung cancer risk.

Investigating DNA methylation as a potential mediator between pigmentation genes, pigmentary traits and skin cancer.

Pigmentation characteristics are well-known risk factors for skin cancer. Polymorphisms in pigmentation genes have been associated with these traits and with the risk of malignancy. However, the functional relationship between genetic variation and disease is still unclear. This study aims to assess whether pigmentation SNPs are associated with pigmentary traits and skin cancer via DNA methylation (DNAm). Using a meta-GWAS of whole-blood DNAm from 36 European cohorts (N = 27,750; the Genetics of DNA Methylation Consortium, GoDMC), we found that 19 out of 27 SNPs in 10 pigmentation genes were associated with 391 DNAm sites across 30 genomic regions. We examined the effect of 25 selected DNAm sites on pigmentation traits, sun exposure phenotypes and skin cancer and on gene expression in whole blood. We uncovered an association of DNAm site cg07402062 with red hair in the Avon Longitudinal Study of Parents and Children (ALSPAC). We also found that the expression of ASIP and CDK10 was associated with hair colour, melanoma and basal cell carcinoma. Our results indicate that DNAm and expression of pigmentation genes may play a role as potential mediators of the relationship between genetic variants, pigmentation phenotypes and skin cancer and thus deserve further scrutiny.

Low-frequency variation in TP53 has large effects on head circumference and intracranial volume.

Cranial growth and development is a complex process which affects the closely related traits of head circumference (HC) and intracranial volume (ICV). The underlying genetic influences shaping these traits during the transition from childhood to adulthood are little understood, but might include both age-specific genetic factors and low-frequency genetic variation. Here, we model the developmental genetic architecture of HC, showing this is genetically stable and correlated with genetic determinants of ICV. Investigating up to 46,000 children and adults of European descent, we identify association with final HC and/or final ICV + HC at 9 novel common and low-frequency loci, illustrating that genetic variation from a wide allele frequency spectrum contributes to cranial growth. The largest effects are reported for low-frequency variants within TP53, with 0.5 cm wider heads in increaser-allele carriers versus non-carriers during mid-childhood, suggesting a previously unrecognized role of TP53 transcripts in human cranial development.

Variability of gene expression profiles in human blood and lymphoblastoid cell lines.

BACKGROUND: Readily accessible samples such as peripheral blood or cell lines are increasingly being used in large cohorts to characterise gene expression differences between a patient group and healthy controls. However, cell and RNA isolation procedures and the variety of cell types that make up whole blood can affect gene expression measurements. We therefore systematically investigated global gene expression profiles in peripheral blood from six individuals collected during two visits by comparing five of the following cell and RNA isolation methods: whole blood (PAXgene), peripheral blood mononuclear cells (PBMCs), lymphoblastoid cell lines (LCLs), CD19 and CD20 specific B-cell subsets. RESULTS: Gene expression measurements were clearly discriminated by isolation method although the reproducibility was high for all methods (range rho = 0.90-1.00). The PAXgene samples showed a decrease in the number of expressed genes (P < 1*10(-16)) with higher variability (P < 1*10(-16)) compared to the other methods. Differentially expressed probes between PAXgene and PBMCs were correlated with the number of monocytes, lymphocytes, neutrophils or erythrocytes. The correlations (rho = 0.83; rho = 0.79) of the expression levels of detected probes between LCLs and B-cell subsets were much lower compared to the two B-cell isolation methods (rho = 0.98). Gene ontology analysis of detected genes showed that genes involved in inflammatory responses are enriched in B-cells CD19 and CD20 whereas genes involved in alcohol metabolic process and the cell cycle were enriched in LCLs. CONCLUSION: Gene expression profiles in blood-based samples are strongly dependent on the predominant constituent cell type(s) and RNA isolation method. It is crucial to understand the differences and variability of gene expression measurements between cell and RNA isolation procedures, and their relevance to disease processes, before application in large clinical studies.

Mutation analysis of candidate genes within the 2q33.3 linkage area for familial early-onset generalised osteoarthritis.

In a genome-wide linkage scan of seven families with familial early-onset osteoarthritis (FOA), we mapped a FOA locus to a 5 cM region on chromosome 2q33.3-2q34 with a maximum LOD score of 6.05. To identify causal variants, 17 positional candidate genes and FRZB were sequenced for coding, splice sites, and 5' and 3' untranslated regions. The pathogenicity of possible disease-causing variants was evaluated using predicted effects on protein structure and function, splicing enhancers, degree of conservation and frequency in 790 unrelated subjects from the population-based Rotterdam study scored for the presence of radiographic signs of OA (ROA). Nine novel variants, identified in NRP2, XM_371590, ADAM23, IDH1, PIP5K3 and PTHR2, cosegregated with FOA, of which two were promising. IDH1 Y183C cosegregated in one family, involved a conserved amino-acid change and showed a damaging effect predicted by PolyPhen and SIFT. In the Rotterdam sample, carriers of IDH1 Y183C (0.02) had an increased but insignificant risk for generalised ROA. The second variant, NRP2 c.1938-21T>C cosegregated in three families. In the Rotterdam sample, carriers conferred an increased risk of 2.1 (95% confidence interval, 1.1-4.1, P=0.032) to have generalised ROA. Furthermore, two variants (NRP2 c.1938-21T>C and IDH1 c.933-28C>T) occurred together on the haplotypes segregating with FOA in two out seven families. This haplotype was rare in the Rotterdam sample (0.0013). Two promising variants in or near NRP2 and IDH1 may not be sufficient for the onset of FOA alone but might have a modulating role with FOA. Confirmation in other OA populations is required.

Strong linkage on 2q33.3 to familial early-onset generalized osteoarthritis and a consideration of two positional candidate genes.

A genomewide screen was performed in four extended families with early-onset generalized osteoarthritis (FOA) without dysplasia. The FOA phenotype within these families shows a dominant Mendelian inheritance pattern and may represent common osteoarthritis (OA) at later ages. An initial locus was confirmed by three additional families and refined by 14 markers to a two-point logarithm of odds score of 6.05 (theta=0.00) for marker D2S155 at chromosome 2q33.3. This locus coincided with the highest multipoint nonparametric linkage score of 4.70 (P-value=0.0013) at marker D2S2358. Haplotype analysis of family members delineated a narrow region with a number of possible positional candidates, of which we investigate here the two most likely ones: PTHR2, encoding parathyroid hormone receptor 2, and FZD5, encoding frizzled receptor 5. For FZD5, we did not observe a segregating variant, however, for PTHR2, a missense variant (A225S) cosegregated with FOA in one family. The frequency of the PTHR2 variant was rare in a population-based sample, aged 55-70 years (N=1228, 0.4%). Of the 11 carriers, 36% showed generalized radiographic OA as compared to 23% in the remaining population. None of the other families that contributed to the linkage revealed a segregating variant. Together, we have identified a locus on chromosome 2q33.3 for FOA. Candidate gene analysis suggested a possible association of a PTHR2 variant with generalized radiographic OA; it is, however, unlikely the major disease gene for the observed linkage to the FOA phenotype.

High microsatellite and SNP genotyping success rates established in a large number of genomic DNA samples extracted from mouth swabs and genotypes.

In this article, we present the genomic DNA yield and the microsatellite and single nucleotide polymorphism (SNP) genotyping success rates of genomic DNA extracted from a large number of mouth swab samples. In total, the median yield and quality was determined in 714 individuals and the success rates in 378,480 genotypings of 915 individuals. The median yield of genomic DNA per mouth swab was 4.1 microg (range 0.1-42.2 microg) and was not reduced when mouth swabs were stored for at least 21 months prior to extraction. A maximum of 20 mouth swabs is collected per participant. Mouth swab samples showed in, respectively, 89% for 390 microsatellites and 99% for 24 SNPs a genotyping success rate higher than 75%. A very low success rate of genotyping (0%-10%) was obtained for 3.2% of the 915 mouth swab samples using microsatellite markers. Only 0.005% of the mouth swab samples showed a genotyping success rate lower than 75% (range 58%-71%) using SNPs. Our results show that mouth swabs can be easily collected, stored by our conditions for months prior to DNA extraction and result in high yield and high-quality DNA appropriate for genotyping with high success rate including whole genome searches using microsatellites or SNPs.