Genome wide analyses have brought new insights to the genetic basis of chronic inflammatory diseases, which have seen a significant rise in numbers during the 21st century. This increase in prevalence is considered too rapid to be accounted for by changes in the frequency of genetic risk variants. However the environment, which is a key factor in the development of common inflammatory diseases, may impact the underlying genetics of these disorders through alterations to the epigenome. Christopher Bell from King’s College London, UK, and colleagues probe the human DNA methylome of peripheral blood to uncover epigenetic variations that may have a key role to play in inflammatory disease, as published in their recent study in Genome Medicine. Here Bell discusses their comparative epigenomic approach, using chimpanzee and rhesus macaque genomes, and the potential clinical impact of their findings.
What was the main goal of this research?
The overarching goal of our work is to use epigenomic analysis to understand human disease. For this particular paper the focused aim was to identify strong human-specific epigenetic variation, in whole blood, by a triangulation comparison between human, chimpanzee and rhesus macaque. Thus we hoped to identify potential regulatory change in this immunologically important tissue type. These changes will be driven by facilitative genetic differences between the three primates, but may additionally also represent environmentally influenced variation. We were fortunate to be able to perform this comparative epigenomic analysis due to a fantastic collaboration with Lutz Walter and Christian Roos at the German Primate Centre.
Since the advent of high throughput SNP (single nucleotide polymorphism) array GWAS (genome-wide association studies), there has been considerable success in identifying common genetic variants associated with common diseases. However, we also know environmental factors have a strong impact on these diseases. This is particularly the case when we consider the inflammatory and metabolic conditions that have had a dramatic rise in prevalence in the last 50 years. This is too short a timeframe to be due to change in risk allele frequency, and is hypothesised to be due to the modern human environment.
These environmental factors, through the interface of epigenetic changes, may be translated into biological effectors on the genome. Due to both the replication stability of DNA methylation, but also its potential plasticity, it is proposed as a biomarker of quantitative lifetime environmental exposures or accrued pathogenic alterations. Therefore by defining this human-specific epigenome we may identify human species-specific physiological differences and vulnerabilities, as well as start to look within this variation for a potential subtle imprint of the modern environment.
What inspired you to take a comparative epigenomic approach to probing the human DNA methylome for disease assocations?
We have been and continue to be involved in common disease Epigenome-Wide Association Studies (EWAS). Whilst exciting strong signals with aging and smoking have been discovered, indicating the promise of the field, the non-cancer disease associations identified so far have only been very small DNA methylation changes. Therefore we took the approach that to increase our understanding this human population variation, which we were trying to associate with disease state, we needed to take a step back and learn more about fixed species-specific variation. This was partly inspired by an excellent paper from De and Babu (Proc Natl Acad Sci U S A 2010, 107, 13004-13009) discussing the ‘time-invariant’ principle of genome evolution, in that similar processes occur across all three time frames of species-, population-, and cancer-evolution. We have used this concept previously to investigate epigenetic change, firstly by focusing on direct facilitative genetic effects, identifying by inter-primate comparison the set of human-specific CpGs that we termed CpG ‘beacons’.
Furthermore regulatory change has been proposed as a major driver in the delineation of primate species phenotypic variation since the classic paper by King and Wilson in 1975 (Science 1975, 188, 107-116). Therefore to be able to assay this epigenetic regulatory level in a genome-wide study is another step in fulfilling their pioneering theory and was further inspiration for this work.
Finally the potential of environmentally-driven epigenetic effectors needs thorough exploration. Starting to define human-specific Differentially Methylated Regions (s-DMRs) comprising the human-specific DNA methylome is part of this process. Also we saw that a comparative blood cell methylome analysis could clearly be interesting with respect to the recent rise of common inflammatory diseases.
What surprised you when you started looking at the data?
At that point in time when we first looked at the data, there were still only a few genome-wide human methylomes published and no chimpanzee or rhesus macaque datasets. So it was obviously exciting to look at data no one has ever seen before. What surprised us initially was just how similar, on the kilobase plus scale, these methylomes looked. The bigWig MeDIP-seq tracks files nearly mirrored each other (see image below and the comparative trimethylome). This neatly brought home, due to the sequence similarity between these homologous primate species, just how genetically driven (predominately due to CpG density) the DNA methylome is.
Secondly, although not a direct surprise, as we had hypothesised this, but a pleasant one nonetheless, was that an immunological gene was our most significantly different locus and furthermore that its entire pathway was also involved. The Leukotriene B4 receptor, (LTB4R/BLT1), is implicated across the gamut of human inflammatory disease. That we have also been able to corroborate this finding with other datasets and studies was great, including across other primates as well as with both wild and captive born.
Who is going to be interested in this research?
We expect that this research would be of interest to all those researchers working in the genomics and epigenomics of common human disease, as a further avenue to think of in disentangling these disorders. Excellent work by Ziller and colleagues has started to define the dynamic human DNA methylome from multiple tissues (Nature 2013, 500, 477-481), and defining the human-specific methylome is an important additional step in the application of epigenomics to genomic medicine.
Of further particular interest for epigeneticists, currently unpicking the genetic effects on methylation, we found, through comparative motif analysis within CpG dense regions, trans-species evidence for a role of the transcriptional repressor, CTCF, in reducing methylation. Also we observed additional support for the interesting association of methylation variability within enhancer regions, due to an enrichment of s-DMRs in these loci.
With respect to immunology, we were able to show functional effects of our epigenetic findings in LTB4R, through another excellent collaboration with Grzegorz Woszczek and his graduate student Holly Foster at the MRC/Asthma UK Centre in Allergic Mechanisms of Asthma, UK. Exploring inter-human variability of this locus, as well as other s-DMRs, in association with disease is an area we, and others, may also wish to pursue. Our results also further emphasise the uniqueness of the human immune system, and the potential caution that is required in extrapolating human diseases from model organisms.
On the back of this work we have further proposed the ‘s-DMR hypothesis’, whereby regions of high sequence similarity between the primates, where human-specific epigenetic variation can be identified, may be enriched for human-specific environmentally-driven DMRs.
What kind of impact will your findings and similar studies have on the clinic?
Epigenomic analysis is poised to revolutionise pathological interpretation, due to the powerful ability of these cell-type specific signatures to deconvolute heterogeneous biopsy tissue samples.
Our work is at the beginning of the next stage of implementing the capabilities of epigenomics, whereby we will hopefully be able to precisely quantitate environmental exposure. We can see the dawning of this potential already with tobacco smoking, where instead of a physician relying on unreliable and imprecise self-reported measures, long term blood-derived DNA methylation changes, at the AHRR locus and others, could be measured. This can also be used to measure passive or in utero exposure.
Precisely delineating these environmental effects may also lead to new causative associations, as well as a molecular understanding of the pathophysiological processes involved in factors already associated with disease. That is, in how they modify the epigenome, and therefore influence genome function. Our work in defining the human-specific methylome, facilitates the beginning of outlining variability and potential impacts of modern human environments on the epigenome. The recently published reconstruction of an archaic human methylome by Gokhman and colleagues (Science 2014, Apr 17), through modeling of the expected degradation of DNA, shows that with this and accumulation of further datasets, we will be able to start to plot the chronicity of when these methylation changes arose. Moreover, once we can fully pick apart the subtle genetic effects on the methylome, due to particular transcription factor binding motif mutations for instance, as well as separate minute true change from technical variation, even more elusive signatures may be able to be identified.
What are you up to next?
Having dissected the human epigenome from an evolutionary perspective, as well as being involved in a number of EWAS, this has led us to the clear conclusion that the most powerful model to identify human disease-associated epigenetic variation is discordant Monozygotic Twins. Fortuitously the chance to pursue this avenue of research has become available due to the opportunity to work with Prof. Tim Spector and the TwinsUK cohort at King’s College London, where I am now based. We will also continue to use comparative evolutionary tools, as well as integrated omics approaches, to further our understanding of common human disease.