Last month, we hosted AEpiA’s sixth flagship conference, Epigenetics 2015, in beautiful Hobart, Tasmania.
What an exciting and inspiring three days! Thank you Hobart for welcoming us and well done to Adele Holloway, Jo Dickinson and all the organising committee for their hard work over the past many months. AEpiA was delighted with the attendance at all the sessions and AEpiA president Sue Clark was thrilled to hear all the lively epigenetics discussion that ensued.
On the Friday evening we were welcomed to Government house by Her Excellency Professor the Honourable Kate Warner, AM, Governor of Tasmania, for a memorable evening of music, canapes and socialising. AEpiA congratulates Joshua Ho from Victor Chang Cardiac Research Institute, who was awarded the Illumina Early Career Research Award for his wonderful work on chromatin organisation.
Some of you have sent us photos and written highlights of the meeting – please scroll down to enjoy them all. Please email us if you would like to share any more stories or pictures. Thanks to all attendees and we hope to see you all at Epigenetics 2017 in Queensland!
Ksenia Skvortsova (PhD student) from the Garvan Institute writes about her highlights:
Highly conserved epigenome remodelling during the vertebrate phylotypic periodProf Ryan Lister (University of Western Australia) described a comprehensive vision on the role of epigenetic processes during early organism development in a very beautiful non-trivial model system.
He compared epigenome remodeling during the highly conserved vertebrate phylotypic period (VPP) of three different species: Mouse, Xenopus and Zebrafish. The phylotypic period is a period of early embryonic development characterized by morphological similarity of distinct vertebrate species.
Comparison of epigenetic landscape remodeling during multiple equivalent stages of VPP between organisms revealed the presence of conserved phyloDMRs (differentially methylated regions). Those phyloDMRs were predominantly undergoing hypomethylation and were associated with development-activated enhancers. Those phyloDMR-harbouring enhancers were associated with orthologous genes of three studied species. Those genes possess similar gene transcription dynamics and involved in the regulation of body plan and organ formation. Regarding the mechanism of such conserved hypomethylation of enhancers of genes crucial for development they observed the specific enrichment of DNA hydroxymethylation at the future phyloDMRs, as well as increase of TET3 enzyme expression prior to the phylotypic period. Levels of hydroxymethylation enrichment reflected the extent of hypomethylation. Triple knockout of all TET enzymes affected hypomethylation of phyloDMRs resulting in the conservation of highly methylated state confirming that TET-mediated demethylation is essential for phyloDMRs epigenetic reprogramming during VPP.
Low dose DNA-demethylating agents target colorectal cancer-initiating cells by activation of MDA5/MAVS/IRF7 pathwayDaniel de Carvalho (University of Toronto) presented an impressive non-traditional view on the mechanism of DNA demethylating agents’ (5-aza) action on tumour cells.
Along with the genes undergoing rapid expression changes upon 5-aza treatment due to their DNA methylation status changes, they have observed the group of late-response genes with a delayed increase of expression upon 5-aza treatment without changes in their DNA methylation status. Strikingly, that group is enriched for the genes involved in the anti-viral response. Knockout of genes encoding virus receptors (as well as other members of virus recognition pathway) results in the inability of 5-aza to induce the reactivation of IFN-stimulated genes. En masse, genome-wide demethylation of gene bodies caused by 5-aza, results in the increased production of dsRNA, which leaves the nucleus and get recognized by the members of the virus recognition pathway (i.e. MDA5, MAVS, IRF7). Hereby, DNA demethylating agents seem to work via induction of a “viral mimicry state”. Moreover, cancer-initiating cells seem to be more sensitive to the “viral mimicry state” induced by DNA demethylating agents. Innate dsRNA pattern recognition may be a general surveillance mechanism to detect global DNA demethylation. Some of the colorectal cancer patients have active IRF7, the virus recognition pathway member. And those patients possess a better response to 5-aza treatment. Generally, activity of IRF7 is associated with the diminished frequency of cancer stem cells.
Dr Ruth Pidsley (post-doc) from the Garvan Institute writes:
Intergenerational Risk for Spontaneous Preterm Birth: Insight from Epigenetic StudiesThe second day was opened with a talk by Dr Alicia Smith from Emory University, USA. She gave us an overview of her recent study investigating the epigenetic contribution to spontaneous preterm birth in African American mothers.
Using 450K arrays to measure DNA methylation of maternal leukocytes and fetal umbilical samples she found a strong correlation between fetal and maternal DNA at CpG sites associated with preterm birth. Many of these same sites could be attributed to a nearby genetic change. Together this raises the fascinating possibility that methylation may mediate the relationship between genetic risk factors and preterm birth.
Alicia then changed tack to finish her talk with new data extending Steve Horvath’s epigenetic clock calculator: showing exciting evidence that DNA methylation can be used as a predictor of neonatal age.
Roll over Weismann: extracellular vesicles in the transgenerational transmission of environmental effects
Dr Cath Suter was the first to talk in the Developmental Epigenetics session, with a bold title instructing the evolutionary biologist August Weismann to “roll over”! – challenging the dogma that somatic cells do not communicate with the germline.
Cath built a strong case giving evidence that, despite being transcriptionally inert, sperm carry more tRNAs and piwi proteins than other cell types, making it likely that these components are transported in from outside. One strong contender for this mechanism is exosomes, which are small microvesicles which carry complex molecular cargo between cells. She touched on their current work showing that Sertoli cells in the testis are able to secrete exosomes. This intriguing mechanism offers a plausible way for somatic cells to influence germ cells, thereby passing on the effects of environment exposures to the next generation. “Roll over Weismann” indeed!
Dr Jane Loke (post-doc) and A/Prof Jeff Craig (group leader) from Murdoch Childrens Research Institute write:
Comprehensive analysis of chromatin landscapeJoshua Ho (Victor Chang Cardiac Research Institute), winner of the ‘Early Career Research Award” at Epigenetics 2015, gave an elegant talk on understanding the epigenetic landscape.
He compared histone modification and other patterns of chromatin state across difference species. Bringing together a large number of datasets from the ENCODE and modENCODE consortia, and using a novel non-parametric machine learning method (hiHMM) developed by his team, Joshua concluded that patterns in promoters, gene bodies, and enhancers tend to be largely conserved across humans, flies and worms.
Notable differences across these three species were identified in repressive chromatin regions. These regions were noted as ‘repressive chromatin’ for their enrichment of H3K9me3. Details of this results can be read in this publication. Joshua even started an epigenome project in the filamentous fungus Aspergillus nidulans.
Epigenomics – from mapping to functional interpretation
Jörn Walter leads the German epigenome program called DEEP, which links various epigenomic mapping and functional analysis groups in Germany. DEEP is part of the International Human Epigenome Consortium (IHEC).
Jörn’s group uses a method, NOMe-seq, that uses a GpC methyltransferase (M.CviPI) and next generation sequencing to generate a high resolution footprint of nucleosome positioning genome-wide using less than 1 million cells while retaining endogenous DNA methylation information from the same DNA strand. In his talk Jörn compared methods of genome-wide chromatin accessibility; he saw many similarities in the results, but some method-specific differences. He also found massive regional loss of DNA methylation accompanying differentiation, that topology-associated chromatin domains (TADs) can change over time and that, counterintuitively, loss of methylation was associated with an increase in heterochromatic state. He concluded that local differences in the topology of chromatin are like ‘codes’ in each tissue. Focusing on cells from the brain, he showed that neurons and non-neuronal cells (glia) showed distinct methylomes and that non-CpG DNA methylation was greatly enhanced in neurons but not glia. He also found that different brain cells ‘age’ differently for DNA methylation. Finally, Prof Walter showed that cell heterogeneity made a huge difference to data interpretation, with proportions of neutrophils in being the largest confounder in whole blood samples.
Epigenetic transitions leading to heritable, RNA-mediated de novo silencing in Arabidopsis thalianaDonna Bond (University of Cambridge, United Kingdom) gave an overview of her research into the role of epigenetic transitions in RNA-mediated de novo silencing in Arabidopsis thaliana.
RNA-directed methylation (RdDM) is essential for maintaining the integrity of plant genomes. Using the virus-induced gene silencing (VIGS) of an active FWA epiallele, Donna’s team established that VIGS-mediated RdDM requires RNA Polymerase V (PolV) and Domains Rearranged Methyltransferase 2 (DRM2) and that this mechanism can be initiated by a process guided by 21-22nt sRNAs and reinforced by 24-nt sRNAs.
Regulation of intron retention by DNA methylationJustin Wong (University of Sydney, Australia) presented his team’s progress into determining the mechanisms regulating intron retention (IR) that is frequently observed in normal cells.
Using whole genome bisulfite sequencing and RNA-seq they addressed the question whether epigenetic mechanisms are associated with intron retention. They discovered that reduced levels of DNA methylation directly regulate IR by reducing the DNA-methylation mediated Mecp2 binding near exon-intron boundaries, which in turn may increase RNA Pol II elongation rate that leads to less efficient splicing.
Sophie Rousseaux directs the EpiMed (Medical Epigenetics) translational research facility at the at the French Institut National de la Santé et de la Recherche Médicale (INSERM). Sophie presented a few different areas of research her lab is focusing on.
Firstly they are investigating the proteins BRD4, a bromodomain chromatin-binding protein, and NUT, a testis specific protein of unknown function. NUT is known to bind to BRD4 and recruits p300 to enhance it’s activity in a feed-forward loop, creating hyper acetylation in cancer. iHDACs have been shown to help reverse this cancer phenotype and tumour regression. They are investigating this relationship further. Another part of her labs work is looking at how ectopic gene expression can be used as biomarkers for caner. She stated that male germ cell differentiation involves major global chromatin reorganization and that male germ cell specific genes drive male genome reprogramming. Further the activation of these genes is seen in many cancers creating a cell identity crisis coupled with a deregulation of the epigenome. To investigate this further they looked at 300 small cell lung cancers with no lymph node invasion (T1NO), and a 10 year follow up, with both transcriptomic and methylome data. They found that even at an early stage of cancer, DNA demethylation and ectopic expression of the germ cell genes was seen. They narrowed down the list to genes to 26 that were significantly associated with poor prognosis and high expression of 3 or more from the panel of 26 was linked to poor survival. This research has since been repeated in breast cancer, leukemias and lymphomas. Going forward Sophie’s lab is interested in looking for new histone modifications and understanding how they are correlated with cancer, in particular they are interested in crontylation.
Intergenerational Risk for Spontaneous Preterm Birth: Insight from Epigenetic Studies
Alicia Smith was an invited speaker from Emory University, USA. She is an Assistant Professor in the department of Psychiatry and Behavioral Science within the School of Medicine. Alicia spoke about her research on the relationship between DNA methylation and preterm birth. Preterm birth was defined as a gestational period less than 37 weeks and is already known to be associated with increased risk for chronic disease.
Her group is interested in the long term affect of preterm birth based on CpG signature and how this is linked with epigenetic aging, also known as the “epigenetic clock”. They performed 450K bead chip arrays on maternal blood and umbilical cord blood at delivery. They found a correlation between DNA methylation patterns and gestational age (GA). Interestingly, in a validation cohort they were able to predict GA based on methylation patterns. They also found there are more than 1000 CpG sites that are the same between the mother and neonate and these sites are enriched for immune, metabolism and cardiovascular pathways. They are hoping this research can lead to using the methylation pattern in the mother’s blood to predict disease risk for their babies. Alicia’s talk provided interesting insight to the epigenetics of preterm birth and how we might be able to use methylation data to infer disease risk.
We’ve had LADs (Lamina Associated Domains) and TADs (Topological Associating Domains), and now Philippe Collas has added a brand new layer of understanding to 3D chromatin conformation with the introduction of GADs (GlyNAcylated H2B). Collas’ results show that in progenitor cells, GADs serve as a ‘placeholder’ for the LADs that these cells will inherit upon differentiation. This is one of the first insights into how 3D chromatin conformation is first established in the cell.
Qian Du (PhD student) from the Garvan discusses Pilar Blancafort’s talk:
Combinatorial epigenetic approaches for genome engineering: form zinc fingers to novel epigenetic editing toolsRe-engineering the genome via gene therapy to treat disease/cancer has been difficult and dangerous thus far with limited success. But what if we can precisely alter genome function without altering the DNA?
A/Prof Blancafort’s research aims to precisely re-engineer the expression and epigenome of cancer-related genes using zinc-finger, TALEN and CRISPR dCas9 technology. This gives her the precision of targeting specific oncogenes or TSGs, without the dangers that come with re-engineering the DNA itself.
Her strategy is to recombine the DNA-binding domain of each of the 3 technologies with epigenetic enzymes to specifically alter the epigenome. These effectors include de novo DNA methyltransferase DNMT3a, DNA demethylase TET3 and histone acetyltransferase P300. She has tried this so far in mouse models, attempting to reactivate or repress the MASPIN tumour suppressor gene. What Blancafort finds is that single agent modifiers alone do not have a large effect on the target gene; however, combinations of these targeted modifiers are highly synergistic. This gives an additional level of control over effect size as well as improves target specificity. Non-specific binding by any of the modifiers alone will not be as detrimental given the low effect size of single agents. However, multiple targeting from combined modifiers will apply a large effect on the target gene. Furthermore, she has figured out an effective way of delivering these agents to the cell using nanoparticles called ‘dendrons’. So far has had success reactivating OCT4 and suppressing SOX2 in breast cancer cell lines and mouse models.
Reprogramming the epigenome in aging and cancerProfessor Jean-Pierre Issa (Temple University School of Medicine, Philadelphia) gave an inspiring talk on the process of epigenetic drift (ED), where the methylation profile of an individual changes over time, and in particular affects the stability of the gene expression profile of the tissue in question.
A strong believer in methylation state as the key driver of cell stability, he illustrated how ED is a central player in both the aging process and the onset of cancer. Gross methylation differences typify aging and cancerous tissue, but the nexus of CpG island promoter methylation and transcription in testis, liver and blood is elucidated by a more subtle non-linear relationship – in particular the point of inflection occurs in the methylation domain of 0-10%. Issa also detailed the translational elements of these findings, where exposure to different lifestyle factors and drug treatment can radically alter the methylation profile of many types of cells, giving hope for development of pathways to treatment and prevention of common malignancies.
Intron retention is an epigenetically regulated mechanism of gene expression control
Professor of Medicine at the University of Sydney and clinical haematologist at the Royal Prince Alfred Hospital, Sydney, Professor John Rasko and his research team at the adjoining Centenary Institute have worked tirelessly over the last few years characterising intron retention (IR) as a distinct and vital epigenetic mechanism. They were rewarded in 2013 with a publication in Cell describing granulopoiesis as being regulated by IR through nonsense-mediated decay (NMD).
Importantly, epigenetic marks such as methylation at exon-intron boundaries and H3.3 lysine6 trimethylation control the rate of IR in a number of cell lines. For example, a methylated splice site may recruit Mecp2 binding to DNA, slowing the rate of RNA polymerase II elongation and increasing the chance of a splice event. The obverse of this process is that RNA pol II skips this site and retains the intron. The resulting transcripts contain a premature termination codon and are hence targeted by the cell for NMD. Rasko opined that this post-transcriptional gene downregulation ought to be seen as part of the normal process of cell differentiation, rather than aberrance.
Dr Phuc Loi Luu (post-doc) from the Garvan Institute discusses two of his highlights:
Multiplex bisufite-PCR resequencing of prognostic and predictive biomarkers in breast cancer using clinical FFPE DNA
This talk by Darren Korbie (University of Queensland) caught my attention because this technique requires very low DNA volume as input for breast cancer prognostic analysis – approximately 50 nanograms, which can be extracted directly from FFPE. Darren demonstrated that is method of DNA methylation analysis is quite robust compared to the method with fluidigm platform in spite despite this low input. The method can be run in high-throughput with either each primer for one amplicon or multi primer for multi amplicons.
This can be achieved by designing multiplex amplicon primers. Darren’s team have developed a primer design tool with optimisation for bisulfite treament DNA, Primer Suite, which they have made available online. They have applied this method to breast cancer samples and discovered distinct regions of differential methylation in triple negative breast cancer compared to normal breast cells. This analysis included 200 amplicons on both FFPE and cfDNA. The results were robust compared to previous data. In short, this method can give high coverage of reads on amplicon regions, low input DNA, low cost and robustness. Importantly, it can be used for breast cancer prognosis. Read more in Darren’s recent publication or visit Primer Suite online.
Breast cancer subtypes as defined by methylome-wide analysis
Eric Joo spoke about exciting work ongoing at Melissa Southey’s lab at the University of Melbourne. They and their team are investigating DNA methylation patterns in subtypes of breast cancer.
While breast cancer is currently grouped into molecular subtypes which have distinct gene expression profiles, the factors driving differential gene expression remain unknown, and furthermore, there is clear heterogeneity within these subgroups. Using 427 breast tumours from the Melbourne Collaborative Cohort Study (MCCS) they applied the Infinium HumanMethylation 450K assay, which has provided methylation information on 469,554 CpG sites across breast cancer subtypes Luminal A, Luminal B, HER2 and triple-negative. Distinct methylation signatures were revealed between tumour subtypes as well as within subtypes. Regression analysis also revealed 4988 Differentially Methylated Regions (DMRs) associated with molecular subtypes, and with hormone receptor status. Overall, their work is demonstrating that DNA methylation signatures can be used to further subgroup breast cancers; this may provide novel molecular markers with important prognostic value and potentially guide the clinical application of emerging epigenetic therapies.