This year’s Canadian Epigenetics, Environment, and Health Research Consortium (CEEHRC) meeting was held in Banff, Alberta, as a joint meeting with the International Human Epigenome Consortium (IHEC), making it the biggest CEEHRC meeting yet with over 260 attendees! The background of the beautiful Canadian Rockies created the perfect “epigenetic landscape” for discussing all the exciting advancements in epigenetics, including whole genome bisulfite sequencing, single-cell analyses, and multi-omics. And, as usual, the two poster sessions were full of lively discussions that carried their momentum late into the night.
Epigenomics of Single Cells
Bing Ren (University of California, San Diego)
The meeting kicked off on Sunday night with the first of three keynote lectures by Bing Ren, who is best known for his pioneering work describing the three-dimensional structure of the genome with the 4D Nucleosome project. The overarching goal of his lab is to connect the ~90% of GWAS hits that show up in non-coding regions to transcriptional outcomes. He discussed three current research programs in his lab:
- Epigenomic analysis of human brain cell types uncover role microglia specific enhancers in Alzheimer’s disease (AD). The group mapped cell type-specific chromatin landscapes using ATAC-seq and ChIP-seq for H3K27ac and H3K4me3 in 4 FACS sorted cell types: microglia, neurons, oligodendrocytes, and astrocytes. They found that AD risk variants are strongly enriched in microglia enhancers and not other cell types, a pattern not seen in other disorders that primarily show enrichment in neurons. They utilized proximity-ligation assisted ChIP-seq (PLAC-seq aka HiChIP) to characterize cell-type specific enhancer-promoter interactome, allowing them to assign non-coding risk variants to specific genes, which highlights that the variant doesn’t always regulate the nearest gene. An example for this principle was a microglia-specific enhancer containing an AD risk variant that drives microglia-specific expression of the BIN1 gene, which they confirmed by using CRISPR to knockout the gene in stem cells that were differentiated, where only in microglia does deletion of enhancer alter protein levels. Overall, they assigned 41 genes with active promoters to 261 credible set variants.
- Construction of mouse brain cell atlas by single nuclear analysis of chromatin accessibility. The Ren lab has been utilizing the combinatorial cellular indexing single-cell ATAC-seq method from Shendure lab, which uses tn5 tagmentation, alongside with their Single Nucleus Analysis Pipeline for ATAC-seq (SnapATAC) approach to dissect heterogenous mouse forebrain tissue and profile chromatin accessibility in rare cell types .They found that sequence variants associated neurological traits are enriched at cis elements in a cell-type specific manner. In the end, they identified 93 major and rare cell types to identify half a million-candidate cis-regulatory elements. Their ultimate goal is to produce a single-cell chromatin accessibility map of the entire mouse brain.
- Joint analysis of single-cell open chromatin and gene expression from same cells. Single-cell multi-omics was one of the hot topics of this meeting, and the Ren lab is blazing the trail. Their main goal was to increase throughput in order to profile more cells and they developed Paired-seq: an ultra-high throughput method, inspired by split-seq, that utilizes ligation-based barcoding (which means no complex equipment required). They utilized paired-seq to analyze mouse adult cerebral cortex and fetal forebrain to detect cell type abundance dynamics during development in heterogenous tissue. Their method allowed them to assay the correlation between enhancer accessibility and promoter transcription activity to identify cell-specific promoter-enhancer pairs. They were also able to define the trajectory of differentiation in the developing mouse forebrain using their data.
Reading and Programming the Epigenome
John Stamatoyannopoulos (University of Washington)
The second keynote was given by John Stamatoyannopoulos on Monday morning. He discussed two research programs taking place in his lab: Read and Write. The goal of read is to advance regulatory genomics by transitioning from discovery to detection. While previous annotations of regulatory DNA have focused on cell context, he is interested in a paradigm shift based on defining a “regulatory vocabulary” for “archetypal regulatory elements”, where transcription factor binding sites explain why they behave differently in different cell types. He also let us know that while we’ve been obsessing over ATAC-seq, old-fashioned DNAaseI hypersensitivity assays have advanced to the point of nucleotide-level resolution, and (in his opinion) are strictly superior to ATAC-seq in defining footprints of DNA bound proteins due to the Tn5 insertion bias. Along the same paradigm-shifting lines, his “write” research program has demonstrated that everyone’s favorite genetic multi-tool, CRISPR/dCas9, is not ideal to study the positioning of regulatory proteins because it has to “blow up” the DNA double helix to bind and can only target the ~11% of the genome with a PAM sequence. Instead, his group has been re-mastering the older TALEN technology to create modular, synthetic transcription factors which they’ve used to identify single nucleotide “keyhole” sites that trigger consistent, potent regulatory activity.
Resistance and Transformation in T-cell Development: Gene Networks Against Epigenetic Constraint
Ellen Rothenberg (Caltech)
The third keynote was given on Tuesday morning by Ellen Rothenburg who described her work on the epigenetic mechanisms of T-cell development. Working in this well-characterized model system allows her group to get at questions of causation that aren’t possible in other studies. She’s found that after a precursor cell commits to becoming a T-cell, there’s a ~4 day delay before it actually happens. It turns out that the DNA binding protein Bcl11b is the key to reprogramming, and that there’s a complex choreography of transcription factors and chromatin modifiers that dance along the genome before it can be fully expressed. A combinatorial input of TFs determine Bcl11b activity at certain stage; however, these TFs are also present before Bcl11b is active, so what else triggers it expression? Dr. Rotherberg’s group wanted to determine if a “major peak” enhancer was the reason. They used a fluorescent protein system that allowed visualization of the timing of activation of the two Bcl11b alleles in the same cell. They found that the activation of the gene is stochastic (i.e. no imprinted).
The Uterine and Placental Epigenome During Pregnancy as a Function of Maternal Age
Myriam Hemberger (University of Calgary)
Myriam Hemberger discussed her lab’s research into placental epigenomics. In their research they are investigating declining reproductive performance in older female mice, where older mothers produce abnormal embryos with defective placentas. They have noted that the developmental defects caused by having an older mother are rescued by a young environment (embryo transfer), which means maternal age-related defects don’t all come down to the egg. They’ve seen that decidualization is delayed in aged females, where gene expression profiling on embryonic day (E) 11.5 shows that key markers are deregulated, and longitudinal analysis of other time points shows the developmental gene expression pattern is delayed. There is blunted hormone responsiveness in uterine cells of aged females (hormone production per se remains seemingly unchanged) and they wanted to investigate whether epigenomic alterations associated with age are a contributing factor. They uncovered dramatic (global and locus specific) epigenetic (H3K4me3) changes in uterine stromal cells of aged females, where the ChIP-seq peak width is reduced. They then characterized epigenetic modifier expression during decidualization (E3.5, E9.5, E10.5, E11.5, E12.5) and noted that DNA methylation is widely gained. They characterized the CGi methylation dynamics in E3.5 uteri of young and aged females and noted that hypermethylation of a subset of promoter-associated CGIs interferes with adequate gene up-regulation (some Hox and Wnt genes), which appears to be due to a large reduction of Tet1 and 5hmC in aged E3.5 uteri/glandular epithelial cells. Ultimately, these findings lead them to conclude that the uterine environment is the dominant factor over the oocyte for increased developmental failure in aged mice.
Regulatory Principles Governing Enhancer Function in Development and Disease
Emma Farley (University of California, San Diego)
Emma Farley discussed her lab’s uses of an unorthodox model organism, the sea squirt (Ciona intestinalis), which looks like a clear tube as an adult but has a remarkably similar embryo to vertebrates, to study developmental enhancers. They can transfect millions of sea squirt embryos at a time, with unique enhancer arrangements driving GFP expression, and have found that there’s a “regulatory syntax” behind developmental tissue-specificity. It turns out that suboptimal binding specificity, spacing and orientation are what keeps enhancers from driving expression in one tissue and not others; creating the most optimal combination of these factors leads to GFP expression throughout the whole embryo.
Speaking of big picture questions, this year EpiGenie partnered with CEEHRC/IHEC to produce a video where we asked attendees to answer the question “What is epigenetics?” as part of knowledge translation effort, and we also conducted some exciting interviews with thought leaders at the meeting. Check out our video playlist of the six interviews below:
EpiGenie would like to thank our team members Ben Laufer (University of California, Davis), Kathryn Vaillancourt (McGill University, Canada), and Eric Chater-Diehl (SickKids, Canada) for contributing this conference coverage.