Activation of the imprinted Prader-Willi Syndrome locus by CRISPR-based epigenome editing [ATAC-seq]

Epigenome editing with DNA-targeting technologies such as CRISPR-dCas9 can be used to dissect gene regulatory mechanisms and potentially treat associated disorders. For example, Prader-Willi Syndrome (PWS) is caused by loss of paternally expressed imprinted genes on chromosome 15q11.2-q13.3, although the maternal allele is intact but epigenetically silenced. Using CRISPR repression and activation screens in human induced pluripotent stem cells (iPSCs), we identified genomic elements that control expression of the PWS gene SNRPN from the paternal and maternal chromosomes. We showed that either targeted transcriptional activation or DNA demethylation can activate the silenced maternal SNRPN and downstream PWS transcripts. However, these two approaches function at unique regions, preferentially activating different transcript variants and involving distinct epigenetic reprogramming mechanisms. Remarkably, transient expression of the targeted demethylase leads to stable, long-term maternal SNRPN expression in PWS iPSCs. This work uncovers targeted epigenetic manipulations to reprogram a disease-associated imprinted locus and suggests possible therapeutic interventions. Overall design: iPSCs were maintained in StemCell mTeSR or mTeSR Plus, with ROCK inhibitor (Y-27632) after seeding or passaging. Stable polyclonal Tet1-dCas9 or VP64-dCas9-VP64 cell lines were established by transducing with lentivirus and selecting for transgene-expressing cassette with 1.5 ug/mL blasticidin for 5 days. Then, cells were transduced with gRNA lentivirus and selected for gRNA-expressing cassette with puromycin (1ug/mL) for 3 days. 14 days after gRNA virus transduction, cells were harvested with Accutase, pelleted, and the ATAC-seq protocol from Corces et al. 2017 was followed with 45,000 cells per sample with the following alterations: KAPA polymerase (Roche, KK2101) was used in the PCR step, following reaction condition recommendations for KAPA polymerase. Libraries were sequenced with 2x25 paired-end reads with a coverage of 50 million reads per sample. ATAC libraries were prepared according to Corces et al. 2017 Omni-ATAC protocol, with KAPA polymerase as the polymerase used for library PCR. Raw reads were trimmed using Trimmomatic to remove adapters. Trimmed reads were aligned to hg19 genome using Bowtie2. Reads mapping to ENCODE blacklisted regions were removed, and duplicated reads were removed with Picard MarkDuplicates. Peaks were called using MACS2. bamCoverage was used to generate rpkm-normalized bigwig files from deduplicated bam files for visualization. Union peak set was generated from narrowPeak files of all samples for each effector (VP64 or Tet1). Counts files were generated using the featureCounts function. Differential peak analysis was determined with DESeq2 on the feature counts files, where gene counts are fitted into negative binomial generalized linear models (GLMs) and Wald statistics to determine significantly differentially expressed genes, with an adjusted p-value threshold of p<0.01.