Abstract
| Originalsprache | Englisch |
|---|---|
| Zeitschrift | The Lancet |
| Jahrgang | 383 |
| Ausgabenummer | 9933 |
| Seiten (von - bis) | 1990-1998 |
| Seitenumfang | 9 |
| ISSN | 0140-6736 |
| DOIs | |
| Publikationsstatus | Veröffentlicht - 01.01.2014 |
Fördermittel
We have identified and replicated a specific association between BMI and methylation of ). HIF3A in whole blood DNA. We recorded the same association in DNA from adipose tissue, which is of high relevance to bodyweight and obesity, implying that it is biologically relevant. Although some preliminary reports are available of whole-blood methylation profiles in relation to indices of body composition and obesity, 24–27 we are the first to have undertaken a large-scale analysis with replication of the principal finding ( panel HIF3A is a component of the hypoxia inducible transcription factor (HIF), which regulates a wide variety of cellular and physiological responses to reduced oxygen concentrations by controlling expression of many target genes. 30 It is a heterodimer that is composed of a β subunit (ARNT) and one of three α subunits (HIF1A, EPAS1, and HIF3A). The binding of each α subunit to ARNT targets different sets of downstream genes in a cell-specific manner. 30 In the case of HIF3A, a further layer of complexity is added by the fact that the HIF3A locus is subject to much alternate splicing, leading to at least seven variants with differing targets. 31 The induction of target genes by HIF3A binding to ARNT is generally weaker than is that evoked by HIF1A and EPAS1 binding to ARNT. 30,31 Furthermore, especially in situations in which the amount of ARNT could be limiting, at least some isoforms of HIF3A seem to hinder the response to hypoxia by sequestering ARNT and restricting its binding to HIF1A and EPAS1. 32,33 Although the main focus on HIF has been its role in cellular and vascular response to changes in oxygen tension during normal development or pathological processes (eg, cardiovascular disease and cancer 30 ), compelling and increasing experimental data suggest that the HIF system also plays a key part in metabolism, energy expenditure, and obesity. 34–37 Specifically, targeted disruption of either HIF1A or ARNT in adipocytes in transgenic mice is associated with reduced fat formation and protection from obesity and insulin resistance induced by high-fat diets. 34 Similarly, systemic use of an antisense oligonucleotide to HIF1A for 8 weeks in mice with diet-induced obesity substantially suppresses HIF1A expression in liver and adipose tissue and is associated with increased energy expenditure and weight loss. 35 In the hypothalamus, HIF signalling (primarily via EPAS1) has a role in glucose sensing and regulation of energy balance and weight by affecting expression of pro-opiomelanocortin. 36 Although HIF3A has not been investigated as thoroughly as the other α subunits in this context, it has been shown to have a role in the cellular response to glucose and insulin, and functions as an accelerator of adipocyte differentiation. 38,39 Furthermore, siRNA inhibition of HIF3A in Hep3B cells significantly downregulates mRNA expression of ANGPTL4 , 31 which could have a role in acquired obesity. 40 The cross-sectional nature of our analysis means that we cannot assign a cause–effect association directly from the association we observed between HIF3A methylation and BMI. Previous studies 41,42 have shown that DNA sequence variation can affect levels of methylation at individual sites (methylation quantitative trait loci). To investigate directionality of the association between HIF3A methylation and BMI, we searched for genetic variants that associate with HIF3A methylation to establish whether these variants also associate with BMI in turn. We identified significant independent associations between genotypes at two SNPs—rs8102595 and rs3826795, upstream of HIF3A —and methylation at one of our identified HIF3A probes, cg22891070. However, we identified no association between these variants and BMI in our cohorts or in the large GIANT genome-wide association meta-analysis of BMI which included more than 123 000 individuals. Our analysis of GIANT data had more than 95% power to detect an association for both SNPs if one existed ( appendix p 8 ). These findings suggest that the association between increased methylation and higher BMI is not causal. Furthermore, the finding that methylation in HIF3A in skin was not associated with BMI, despite a strong methylation quantitative trait locus for cg22890170 in this tissue, also indicates the absence of causal directionality. Therefore, our findings suggest that increased methylation at the HIF3A locus is a result of increased BMI. An alternative possibility is that the association between methylation at HIF3A and BMI is due to a confounding factor which affects both variables. However, we did not observe the association between HIF3A methylation and BMI in skin. Furthermore, we did not observe any association with other characteristics associated with BMI, such as physical activity or diabetes. The mechanism by which increased BMI could lead to rises in HIF3A methylation is unknown. Obesity predisposes individuals to obstructive sleep apnoea, 43 which is associated with intermittent hypoxia. In turn, hypoxia activates HIF signalling. Therefore, chronic upregulation of HIFs in response to obstructive sleep apnoea could result in secondary changes in methylation of the HIF genes. However, the association of methylation level at the HIF3A locus showed a linear correlation across the range of BMI levels, and increased methylation was not confined to obese individuals ( appendix p 13 ). Furthermore, the association of BMI with variation in methylation was specific to HIF3A and was not noted for HIF1A and EPAS1 . We identified a significant inverse association between HIF3A methylation and HIF3A expression in adipose tissue. The association was only recorded with one of five HIF3A expression probes on the genome-wide expression array ( appendix p 6 ), suggesting that the effect of methylation could be transcript-specific. 31 In this context, we note that all three CpG sites at the HIF3A locus that were associated with BMI are situated within regions of open chromatin as identified by formaldehyde-assisted isolation of regulatory elements (FAIRE) in H1-hESC cells and K562 cells, suggesting that these sites lie in a regulatory region. 44 However, two of the expression probes analysed (ILMN_1663015 and ILMN_1687481) are reported to tag the same HIF3A transcript ( appendix p 6 ), and the reason for the discrepant findings for these two probes is unclear. Therefore, further work needs to be done to confirm the effect of methylation on expression and any transcript specificity. However, our finding supports the possibility that even if the association between increased methylation of HIF3A and BMI is secondary, an alteration in HIF signalling as a result of obesity-induced HIF3A methylation could still have an important role in some of the deleterious downstream effects of the disorder. Although we recorded significant associations between increased HIF3A methylation in blood DNA and increased BMI in three different cohorts, the strength of the association varied substantially across the different cohorts. The gradient of the relation between methylation at HIF3A and BMI was four-times steeper in the discovery cohort than in the second population-based replication cohort (KORA), despite a similar distribution of methylation values. Whether this difference represents an element of winner's curse 45 or reflects other variation in the characteristics of the cohorts (including the presence of disease in some) is unclear. Even in the discovery cohort, we noted a difference in the level of association between the individuals who had had myocardial infarction and the healthy blood donors. The strength of the association in the blood donors was similar to that in the MARTHA cohort, which comprised patients with deep vein thrombosis, suggesting that the variation is not entirely related to disease status. Therefore, further studies are needed to identify factors that affect HIF3A methylation and modulate the association between BMI and HIF3A methylation in whole-blood DNA. Further work is also necessary to deduce the timing of the variation in methylation at the HIF3A locus in relation to BMI and whether it is dynamic or not. Blood is readily accessible for DNA analyses. By contrast with genetic analyses, a challenge of epigenetic analyses is that circulating leucocytes—the source of DNA in blood—are composed of several different cell subtypes that could each show cell-type specific variation in DNA methylation patterns. To an extent, as we have shown, this variation can be assessed and statistical adjustment done. Perhaps a more fundamental issue for the epigenetics community is whether analysis of blood DNA methylation is worthwhile and can reflect changes in relevant tissues for a phenotype. In this regard, our finding of an association between BMI and specific HIF3A methylations sites in both blood and adipose tissue DNA supports the use of whole-blood DNA methylation profiling for identification of relevant epigenetic changes and provides a rationale for other studies of this type. We used a strict sequential replication design to avoid the penalty of multiple testing for confirmation of the association of probes identified in the discovery cohort. We also started with a fairly small discovery cohort. Therefore, we recognise that we have probably missed associations between methylation of other genes and BMI. Meta-analyses of the datasets used in our study together with other datasets could yield additional insights into epigenetic changes associated with BMI. In summary, we have reported a novel association of increased BMI in adults of European origin with increased methylation at the HIF3A locus in blood cells and in adipose tissue. The finding extends reports linking HIF and obesity in experimental models and provides direct evidence in people that perturbation of HIF signalling could have an important role in mediation of some of the downstream adverse responses to increased BMI. Contributors KJD, CPN, PD, and NJS conceived the study. JE, CH, FC, AHG, WHO, HS, and NJS were responsible for recruitment and phenotyping of the discovery (Cardiogenics) cohort. LT and EM generated methylation array data for the discovery cohort. KJD and CPN analysed data for the discovery cohort, supervised by JRT. DA, P-EM, FG, and D-AT provided data from the primary replication cohort (MARTHA) and did analyses. SW, HG, MW, AP, and CG provided data from the second replication cohort (KORA) and did analyses. JKS, TDS, and PD provided data from the MuTHER cohort and did analyses. KJD, CPN, and NJS wrote the report. All authors reviewed the report and provided comments. Declaration of interests We declare that we have no competing interests. Acknowledgments This work was done as part of the Cardiogenics Project, which is funded by the European Union ( LSHM-CT 2006–037593 ). The MARTHA project was supported by a grant from the Program Hospitalier de Recherche Clinique, and the methylation array typing was funded by the Canadian Institutes of Health Research ( grant MOP 86466 ) and the Heart and Stroke Foundation of Canada ( grant T6484 ). Statistical analyses of the MARTHA datasets were done in the C2BIG computing centre (UPMC, Paris, France), which is funded by the Fondation pour la Recherche Médicale and Région Ile de France. The KORA study was initiated and financed by the Helmholtz Zentrum München—German Research Center for Environmental Health, which is funded by the German Federal Ministry of Education and Research and by the State of Bavaria. The MuTHER study was funded by a programme grant from the Wellcome Trust ( 081917/Z/07/Z ), and receives support from the National Institute for Health Research BioResource Clinical Research Facility and Biomedical Research Centre based at Guy's and St Thomas' NHS Foundation Trust and King's College London. CPN is funded by the National Institute for Health Research Leicester Cardiovascular Biomedical Research Unit, and this work comes under the portfolio of translational research supported by this unit. DA was supported by a PhD grant from the Région Ile de France (CORDDIM). FG holds a Canada Research Chair. JE, FC, HS, D-AT, and NJS collaborate under a Fondation Leducq Grant ( 12CVD02 ). TDS is a European Research Council Senior Investigator and is holder of an ERC Advanced Principal Investigator award. PD is supported by the Wellcome Trust core grant to the Wellcome Trust Sanger Institute ( 098051 ), which funded DNA methylation analysis for MuTHER. NJS holds a chair funded by the British Heart Foundation and is a National Institute for Health Research Senior Investigator. We thank the staff from the genotyping facilities at the Wellcome Trust Sanger Institute for sample preparation, quality control, and typing for the Cardiogenics and MuTHER cohorts.
UN SDGs
Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung
