Identification of genes influencing a phenotype of interest is frequently achieved through genetic screening by RNA interference (RNAi) or knockouts. However, RNAi may only achieve partial depletion of gene activity, and knockout-based screens are difficult in diploid mammalian cells. Here we took advantage of the efficiency and high throughput of genome editing based on type II, clustered, regularly interspaced, short palindromic repeats (CRISPR)-CRISPR-associated (Cas) systems to introduce genome-wide targeted mutations in mouse embryonic stem cells (ESCs). We designed 87,897 guide RNAs (gRNAs) targeting 19,150 mouse protein-coding genes and used a lentiviral vector to express these gRNAs in ESCs that constitutively express Cas9. Screening the resulting ESC mutant libraries for resistance to either Clostridium septicum alpha-toxin or 6-thioguanine identified 27 known and 4 previously unknown genes implicated in these phenotypes. Our results demonstrate the potential for efficient loss-of-function screening using the CRISPR-Cas9 system.

SummaryAcute myeloid leukemia (AML) is an aggressive cancer with a poor prognosis, for which mainstream treatments have not changed for decades. To identify additional therapeutic targets in AML, we optimize a genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screening platform and use it to identify genetic vulnerabilities in AML cells. We identify 492 AML-specific cell-essential genes, including several established therapeutic targets such as DOT1L, BCL2, and MEN1, and many other genes including clinically actionable candidates. We validate selected genes using genetic and pharmacological inhibition, and chose KAT2A as a candidate for downstream study. KAT2A inhibition demonstrated anti-AML activity by inducing myeloid differentiation and apoptosis, and suppressed the growth of primary human AMLs of diverse genotypes while sparing normal hemopoietic stem-progenitor cells. Our results propose that KAT2A inhibition should be investigated as a therapeutic strategy in AML and provide a large number of genetic vulnerabilities of this leukemia that can be pursued in downstream studies.

The accuracy of replicating the genetic code is fundamental. DNA repair mechanisms protect the fidelity of the genome ensuring a low error rate between generations. This sustains the similarity of individuals whilst providing a repertoire of variants for evolution. The mutation rate in the human genome has recently been measured to be 50–70 de novo single nucleotide variants (SNVs) between generations. During development mutations accumulate in somatic cells so that an organism is a mosaic. However, variation within a tissue and between tissues has not been analysed. By reprogramming somatic cells into induced pluripotent stem cells (iPSCs), their genomes and the associated mutational history are captured. By sequencing the genomes of polyclonal and monoclonal somatic cells and derived iPSCs we have determined the mutation rates and show how the patterns change from a somatic lineage in vivo through to iPSCs. Somatic cells have a mutation rate of 14 SNVs per cell per generation while iPSCs exhibited a ten-fold lower rate. Analyses of mutational signatures suggested that deamination of methylated cytosine may be the major mutagenic source in vivo, whilst oxidative DNA damage becomes dominant in vitro. Our results provide insights for better understanding of mutational processes and lineage relationships between human somatic cells. Furthermore it provides a foundation for interpretation of elevated mutation rates and patterns in cancer.

SummaryThe genetic basis of naive pluripotency maintenance and loss is a central question in embryonic stem cell biology. Here, we deploy CRISPR-knockout-based screens in mouse embryonic stem cells to interrogate this question through a genome-wide, non-biased approach using the Rex1GFP reporter as a phenotypic readout. This highly sensitive and efficient method identified genes in diverse biological processes and pathways. We uncovered a key role for negative regulators of mTORC1 in maintenance and exit from naive pluripotency and provided an integrated account of how mTORC1 activity influences naive pluripotency through Gsk3. Our study therefore reinforces Gsk3 as the central node and provides a comprehensive, data-rich resource that will improve our understanding of mechanisms regulating pluripotency and stimulate avenues for further mechanistic studies.

The 21-23 nucleotide single-stranded RNAs classified as microRNAs (miRNA) perform fundamental roles in a wide range of cellular and developmental processes. miRNAs regulate protein expression through sequence-specific base pairing with target messenger RNAs (mRNA) reducing both their stability and the process of protein translation1, 2. At least 30% of protein coding genes appear to be conserved targets for miRNAs1. In contrast to the protein coding genes3, 4, no public resource of miRNA mouse mutant alleles exists. We have generated a library of highly germ-line transmissible C57BL/6N mouse mutant embryonic stem (ES) cells with targeted deletions for the majority of miRNA genes currently annotated within the miRBase registry5. These alleles have been designed to be highly adaptable research tools that can be efficiently altered to create reporter, conditional and other allelic variants. This ES cell resource can be searched electronically and is available from ES cell repositories for distribution to the scientific community6.

Genome-wide CRISPR-based knockout (CRISPR-KO) screening is an emerging technique which enables systematic genetic analysis of a cellular or molecular phenotype in question. Continuous improvements, such as modifications to the guide RNA (gRNA) scaffold and the development of gRNA on-target prediction algorithms, have since been made to increase their screening performance. We compared the performance of three available second-generation human genome-wide CRISPR-KO libraries that included at least one of the improvements, and examined the effect of gRNA scaffold, number of gRNAs per gene and number of replicates on screen performance. We identified duplicated screens using a library with 6 gRNAs per gene as providing the best trade-off. Despite the improvements, we found that each improved library still has library-specific false negatives and, for the first time, estimated the false negative rates of CRISPR-KO screens, which are between 10% and 20%. Our newly-defined optimal screening parameters would be helpful in designing screens and constructing bespoke gRNA libraries.

Despite progress in understanding its genomics and molecular pathogenesis, the therapeutic landscape of acute myeloid leukaemia (AML) has changed little in the last 40 years. Whilst our improved molecular understanding of AML permits some optimism that progress may be forthcoming, an alternative approach for the identification of therapeutic targets is the agnostic interrogation of AML genomes for genetic vulnerabilities. In this study we apply a new and technically robust CRISPR-Cas9 platform to perform genome-wide screens for genetic vulnerabilities in human cancers. To do this, we develop and validate a CRISPR-based functional genomics toolkit composed of: i) lentiviral gRNA expression vectors harbouring an improved sgRNA scaffold, ii) Cas9 activity reporters for choosing cell line clones with high Cas9 nuclease activity and iii) an improved human genome-wide CRISPR library composed of 90,709 gRNAs targeting 18,010 genes. We first describe the timescale over which cells lacking individual essential genes are depleted from a pool of isogenic cells, thus providing the first such genome-wide framework for mammalian cells. As well as being of fundamental interest, such a temporal framework can be used to decide the length of time required for performing genetic screens and to select therapeutic targets. We then proceeded to perform drop-out screens with 30-day latencies in 5 AML cell lines (MV4-11, MOLM-13, OCI-AML2, OCI-AML3 and HL-60) and also in the non-AML lines HT-29 (colorectal adenocarcinoma) and HT-1080 (fibrosarcoma). Drop-out genes were identified using the MAGeCK algorithm as those showing significant depletion across their ≥5 cognate sgRNAs. From each cell line, more than 1,000 genes dropped out at FDR <20%, with the exception of MV4-11 where the number was slightly lower. Using these data we identified 881 "pan-essential genes" defined as those displaying significant depletion across ≥5 cell lines including HT-29 and HT-1080. These 881 genes can be used as a standard set of quality-control genes for future screens. Of these, 335 genes were depleted in all 7 cell lines, showing remarkable consistency across different cellular contexts. Next, we looked for genes that are specifically essential to AML cells by extracting genes depleted in at least 1 of the 5 AML cell lines, but not in HT-29 or HT-1080. This analysis identified approximately 150-200 essential genes for each cell line yielding a total of 510 AML-specific genes. Of these, 59 genes including RUNX1, CEBPA, CEBPB, MEN1, DOT1L and SMARCB1 were essential to 3 or more AML cell lines. GO analysis of these 59 genes showed particular enrichment in processes pertaining to chromatin modification and organisation, transcriptional regulation and nucleotide metabolism. We proceed to validate a number of novel drop-out genes using CRISPR-Cas9 with new sgRNAs and where possible with existing clinical/pre-clinical inhibitors. Furthermore, we identify oncogene-specific cell vulnerabilities, even for leukaemias driven by closely related oncogenes such as the MLL-AF4 (MV4-11) and MLL-AF9 (MOLM-13) fusion genes, which differed in their dependency on several genes including KAT2A and SRPK1. To validate these findings in primary cells, we generate a novel Rosa26-Ef1a-Cas9 mouse model and cross this with mice carrying Flt3-ITD. We then transformed Lin- haemopoietic cells from RosaCas9/+/Flt3ITD/+ mice using MLL-AF4- or MLL-AF9 -expressing retroviruses and validate the findings of our screens using sgRNAs against murine Kat2a and Srpk1. Taken together, these data dissecting the individual vulnerabilities of highly similar initiating mutations demonstrate the power of our screen to identify specific vulnerabilities for individual oncogenes and suggest that similar screens may also help to guide programmes of personalised medicine for patients based on the complement of somatic mutations within their cancer, which in some cases could be achieved through re-purposing of existing therapeutics. Disclosures McDermott: 14M Genomics: Other: co-founder, stock-holder and consultant.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.