Translating Genomes | Personalizing Medicine
•Download as PPTX, PDF•
1 like•1,118 views
Gene-editing evolved - Combining CRISPR, rAAV and ZFNs for maximum versatility and minimal hassle
Recommended
More Related Content
What's hot
What's hot (20)
Viewers also liked
Similar to Translating Genomes | Personalizing Medicine
Similar to Translating Genomes | Personalizing Medicine (20)
More from Candy Smellie
More from Candy Smellie (14)
Recently uploaded
Recently uploaded (20)
Translating Genomes | Personalizing Medicine
- 1. Translating Genomes | Personalizing Medicine Gene-editing evolved - Combining CRISPR, rAAV and ZFNs for maximum versatility and minimal hassle Dr. Chris Lowe R&D Director, Cell Line Engineering
- 2. 2 Presenter Dr. Chris Lowe PhD R&D Director, Cell Line Engineering Chris obtained his PhD in the field of Medical Genetics from the University of Cambridge where he engaged in research into the genetic causes of Type 1 diabetes. He joined Horizon Discovery in 2011 and has been responsible for the gene editing group since 2013.
- 3. 3 Content of the Presentation Introduction of Horizon Discovery GENESIS™ - Horizon’s precision genome editing platform Systematic optimisation of the GENESIS™ platform Combining CRISPR and rAAV technologies to improve targeting efficiency Custom cell line development service Summary
- 4. Translating genetic information into personalized medicines Genomics Translational Genomics Personalised Medicine 4
- 5. 5 Horizon Discovery’s Mission “To translate the human genome and accelerate the discovery of personalized medicines” Tailoring the right drugs...to the right patients...at the right time
- 6. Genome Editing: Creating accurate genetic models Isakoff et al., Cancer Research, Jan 2006 Di Nicolantonio et al., PNAS, Dec. 2008 6 Large growth induction phenotype Transforming alone Milder growth induction phenotype Non-transforming alone
- 7. Genome Editing: The Right Tool For The Right Outcome
- 8. GENESIS™: The Right Tool For The Right Outcome 8 rAAV • High precision / low thru-put • Any locus, wide cell tropism • Well validated, KI focus • Exclusive to HD Zinc Fingers • Med precision / med thru-put • Good genome coverage • Well validated / KO Focus • Licensed from Sigma CRISPR • New but high potential • Capable of multi-gene targeting • Simple RNA-directed cleavage • Combinable with AAV • Extensive IP position
- 9. rAAV: Modify Any Genomic Loci, in Any Way, with Perfect Precision DNA-vectors that use a natural homologous recombination (HR) in cells to alter genomic sequences No DNA-breaks created or required (rAAV stimulates HR directly, 1000x better than plasmid vectors) 9 Efficient at performing all types of alterations Wide tropism Hard to generate multi-allelic KO’s quickly like nucleases Ideal for ‘deep-biology’ & disease model generation Homologous Recombination (HR) using single-stranded DNA recombinant Adeno Associated Viruses Nature Genetics 18, 325- 330 (1998)
- 10. rAAV: How Does It Work? AAV = Adeno Associated Virus (ssDNA)
- 11. rAAV: What You Can Do with rAAV Gene-Editing Point mutations/SNPs RNAi rescue Insertional gene disruption Gene deletions Long range deletions Translocations Amplifications > 40 different parental cell lines now targeted; 500 projects, covering 16 tissue types
- 12. 12 Nuclease Methods: ZFNs Double strand breaks are repaired by either NHEJ, or HDR in tandem with a donor Low off target risk High efficiencies of knockout Reliable gene knock-outs Double Strand Break Non-Homologous End Joining
- 13. 13 Nuclease Methods: CRISPR Analogous to ZFNs/TALENs, but much simpler: no protein engineering required Short ‘guide’ RNAs with homology to target loci direct a generic nuclease (Cas9) Guide RNA + Cas9 are delivered into the cell Cas9 cleavage is repaired by either NHEJ, or HDR in tandem with a donor High efficiencies of knockout or knock-in Multiplexing (multiple gene KOs in parallel) possible hCAS9 + Guide RNA ‘Nick’ or Break KO CRISPR components delivered into cell by transfection or electroporation Guide RNA CAS 9 PAM sequence Matching genomic sequence Genomic DNA OR Donor KI Jinek M, Science 2102. Mali P, Science 2013. Cong L, Science 2013
- 14. Nuclease Methods: Cas9 Wild-Type or Cas9 Nickase? Cas9 Wild type Cas9 Nickase (Cas9n) Induces double strand break Only “nicks” a single strand Only requires single gRNA Requires two guide RNAs for reasonable activity Concerns about off-target specificity Reduced likelihood of off-target events High efficiency of cleavage Especially good for random indels (= KO) Guide efficiency dictated by efficiency of the weakest gRNA Nishimasu et al Cell
- 15. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation
- 16. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Gene copy number Number and nature of modified alleles Effect of modification on growth Normal human karyotype HeLa cell karyotype
- 17. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Transfection/electroporation Single-cell dilution Optimal growth conditions
- 18. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Sequence source Off-target potential Guide proximity Wild-type Cas9 or mutant nickase
- 19. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Ran et al Cell (2013)
- 20. Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation NT Cas9 wt only gRNA uncut 1 2 3 4 5 600 500 400 300 200 100 +ve 700 700 600 500 400 300 200 100 Key Considerations for a Gene-Editing Experiment Number of gRNAs gRNA activity measurement
- 21. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Modification effects on expression or splicing Type of donor (AAV, oligo, plasmid) Cas9 Cut Site Donor sequence modifications Donor size Selection based strategies Genomic Sequence Donor Sequence containing mutation
- 22. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Number of cells to screen Screening strategy Modifications on different alleles Homozygous or heterozygous modifications versus mixed cultures % Cells Targeted
- 23. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation Confirmatory genotyping strategies Off-target site analysis Heterozygous knock-in Wild type Genetic drift/stability Modification expression Contamination
- 24. Key Considerations for a Gene-Editing Experiment Gene Target Specifics Cell Line gRNA Design gRNA Activity Donor Design Screening Validation How many copies? Is it suitable? What’s my goal? (Precision vs Efficiency) Does my guide cut? Have I minimised re-cutting? How many clones to find a positive? Is my engineering as expected?
- 25. Horizon’s Experience + Developments: rAAV + CRISPR Combinations Nucleases have historically been less efficient at performing user-defined KIs vs KOs Combining rAAV with a nuclease allows very high efficiency KIs and KOs 25 % Green cells (FACs)
- 26. Horizon’s Experience + Developments: Advances in AAV design Novel negative selection targeting strategies Gene-targeting frequencies at the CDK2 locus 26 • Reduce background of NHEJ integrations ‡ Targeting frequency is the number of correctly targeted colonies per 100 drug-resistant colonies screened. § The fold increase is the targeting frequency of the ShRNA vectors divided by the targeting frequency of the no ShRNA vector (set at 1).
- 27. Horizon’s Experience & Developments: Cas9-FOK1 dimers Fusion of the dimerization-dependent FokI nuclease to a catalytically inactive Cas9 DNA modification requires dimerization of the Fok1 pairs Dimerization can only occur using two closely spaced gRNAs Improved specificity relative to Cas9n Mutagenic frequencies at known off target sites
- 28. Case Study: Disruption of the MAPK3 gene in the A375 cell line (copy number = 3) Conserved exon 3 targeted 96 Clones Screened 28 Positive for cutting 7 Clones Sequenced 3 Clones with indels on all three alleles Using CRISPR to Generate Gene KOs and KIs ENSEMBL
- 29. Using CRISPR to Generate Gene KOs and KIs Case Study: Disruption of the MAPK3 gene in the A375 cell line (copy number = 3) 1 2 3 Parental Allele 1 Allele 2 Allele 3
- 30. Applications of Horizon’s Engineered Cell Lines 30 Case study: DLD-1 BRCA2 null cell line: DLD-1 BRCA2 null cells show selective sensitivity to the PARP inhibitor olaparib
- 31. Applications of Horizon’s Engineered Cell Lines 31 Case study: SW48 PI3Ka cell lines Resistance to a tyrosine kinase inhibitor is conferred by PTEN deletion or activating mutations of PIK3CA
- 32. Accessing GENESIS™: Custom Cell Line Development Service The only ‘one-stop genome editing shop’ (ZFNs, CRISPR & rAAV) Full custom services - modify any gene/loci to your requirements No project too tough; including inducible alterations (KI or KOs) Extensive know-how on editing in range parental cell-lines Continuum of price, speed and design to meet all needs Delivery of a validated custom cell line from as low as $30,000 Horizon’s scientists are experts at all forms of gene editing and so have the experience to help guide customers towards the approach that best suits their project 32 Point Mutations Gene Knockouts Deletions Insertions Translocations Amplifications
- 33. Horizon Discovery Products & Services 33
- 34. Learn more about sgRNA screening in our upcoming free webinar Webinar: RNAi screening in drug discovery – introducing sgRNA technologies Tue 9th Dec at 4 pm (GMT) 34 Shalem et al Science 2014
- 35. Useful Resources From Horizon Free gRNAs in Cas9 wild type vector – www.horizondiscovery.com/guidebook Technical manuals for working with CRISPR - http://www.horizondiscovery.com/talk-to-us/technical-manuals In the Literature Exploring the importance of offset and overhand for nickase - http://www.cell.com/cell/abstract/S0092- 8674(13)01015-5 sgRNA whole genome screening: • Shalem et al - http://www.sciencemag.org/content/343/6166/84.short • Wang et al - http://www.sciencemag.org/content/343/6166/80.abstract On the web Feng Zhang on Game Changing Therapeutic Technology (Link to Feng’s Video) Guide design - http://crispr.mit.edu/ CRISPR Google Group - https://groups.google.com/forum/#!forum/crispr
- 36. Your Horizon Contact: Dr. Chris Lowe R&D Director c.lowe@horizondiscovery.com +44 (0)1223 655580 Horizon Discovery Ltd, Building 7100, Cambridge Research Park, Waterbeach, Cambridge, CB25 9TL, United Kingdom Tel: +44 (0) 1223 655 580 (Reception / Front desk) Fax: +44 (0) 1223 862 240 Email: info@horizondiscovery.com Web: www.horizondiscovery.com
Editor's Notes
- Welcome and thank you for joining me today to talk about Gene-editing at Horizon Discovery. In the webinar today I would like to introduce you to Horizon Discovery and describe our gene editing platform. I’ll also discuss in general terms how we optimise each element of a gene engineering project to enhance the successful delivery – including some examples of engineering projects that have been performed. We’ll then move on to present some newer advances we have developed which improve gene engineering efficiencies before wrapping up with how Horizon can help you with your gene engineering needs.
- So we live now in what can be thought of as the post-genomic era – in which it is possible to screen individuals for genetic variations which lead to disease progression or resistance to certain therapeutics - information is no longer the bottle neck, the emphasis is shifting to the ‘what does it all mean’. And this is the shift to translational genomics, so understanding the impact of this multitude of mutations that are being identified, and ultimately this leads us into the realm of personalized medicine in which individuals obtain the therapies they require rather than the average population requires. Genome editing is enabling the promise of the genomic era to be realized in the form of novel therapeutics and diagnostics and is being driven by the capability to efficiently introduce targeted alterations into any specific gene in living cells to elucidate their effect.
- At Horizon, our mission is to translate the human genome and accelerate the discovery of personalised medicines by creating better models to improve the drug discovery pipeline. The ultimate aim of this is to ensure that right drugs are given to the right patients at the right time – be this by providing tools to identify new targets and therapeutics or by pharmacogenomics
- So lets start with the basic question of “Why would you be interested in gene engineering” Historically, to understand the effect of a specific mutation, classic overexpression studies have been performed, in which a mutated gene of interest is introduced randomly into the cell under the control of an exogenous promoter. This can result in unclear results due to the non-physiological levels of expression and regulation. For example, on the slides we can see two studies investigating the effect of the same mutations. The first overexpresses PI3Kinase mutations and resulted in an oncogenic phenotype. However, when these same mutations are introduced into the endogenous PI3K gene, where physiological levels of expression and endogenous mechanisms of regulation are retained, a much milder growth phenotype is observed which alone would not be considered as transforming. This more accurate modelling of the impact of specific mutations enables the multifactoral genotypes present if complex disease to be investigated with greater confidence. Similar limitiations can be seen with si and sh RNA studies which attempt to knock-down the expression of genes of interest. However these are rarely complete knockdowns and the residual expression cannot be discounted. Whereas permanent, stable disruption of a gene removes such confounding issues.
- So there is a clear need to generate improved models and the field of genome editing has advanced tremendously in recent years to a point at which many of the initial limitations, which put people of venturing down this road, have been overcome to the point at which genome engineering could now be considered the norm. There exist now a range of technologies that allow the modification of genomes - most commonly the introduction of point mutations or gene knockouts. But also the insertion or deletion of specific sequences or exons as well as more complex events such as translocation and amplifications. Horizon is the only source of rAAV expertise and is uniquely capable of exploiting multiple platforms: recombinant adeno associated virus commonly known as rAAV, Zinc Finger nucleases and most recently the CRISPR/Cas9 technology. This ability to access all these technologies allows our scientists to deploy the best technology, or combination of technologies to achieve the goals of your project. On the next slide we will summarise the key features of each of these technologies and touch on their strengths and weaknesses.
- As I said, Horizon has exclusive access to the use of AAV for in vitro gene engineering and our scientists have extensive experience of using this technology having used it for many years to provide a custom gene engineering service, as a result we have experience of a wide range of cell lines covering a number tissue types which I’ll touch on again later in the webinar. The major benefits of rAAV are its high precision which allows us to ensure only a single, predictable modification has occurred in the target genome i.e. no off-targets. There is a wide tropism meaning a wide range of cell types can be engineered, and this technology is particularly suited to the generation of precise knock-ins. The main limitation of the AAV technology is that in general, only a single allele can be targeted at one time, therefore sequential targeting events may be required for complex projects. In comparison, nucleases such as Zinc Finger and CRISPR are able to perform multi-allelic targeting, with their particular strength being knock-outs. ZFNs are a well validated an understood technology providing a medium throughput option. The latest addition to the gene engineering family is CRISPR. Like ZFNs this is a nuclease based approach, but in comparison to ZFNs is a much simpler technology, requiring only the generation of specific gRNA to target the nuclease activity and we will talk in more detail about the specifics of each technology in the following slides. And although Horizon was based around the AAV technology, we consider ourselves technology agnostic and deploy the right tool, or tools required to generate the specific cell line model.
- Starting with AAV, the technology on which Horizon’s gene editing platform has historically been based. It’s a single stranded DNA virus that has been demonstrated to be particularly efficient at driving homologous recombination in cells, with levels of homologous recombination 1000x greater compared to double-stranded plasmid techniques. rAAV exploits the ability of the simple single-stranded virus to deliver a sequence of homology to ANY target locus within human and mammalian genomes. What separates this approach from all the other genome editing approaches that involve introduction of a double strand break is that is uses ONLY homologous recombination and is therefore not subject to the “off-target” effects that can occur with nucleases. That makes this the most controlled and precise mode of genome engineering. AAV does, however, have some disadvantages to nucleases and that is one of the reasons that Horizon made the decision to broaden our portfolio to include nucleases. But we still like having the precision and complete control afforded by AAV at our disposal.
- Lets look now in a little more detail about how AAV is deployed as a genome engineering tool. As I said, AAV is a single-stranded, linear DNA virus with a 4.7 kb genome. For the purpose of genome editing, the viral genome is replaced almost in entirety with the targeting vector sequence, shown on the second line of the cartoon, with only the ITRs or inverted terminal repeats which are shown in green remaining. The rest of the genome is replaced with regions of homology to the target region, known as homology arms, and in the majority of cases these homology arms flank a selection cassette. The mutation or mutations of interest are incorporated into the homology arms. After entry of the vector into the cell, target-specific homologous DNA is believed to activate and recruit HR-dependent repair factors, but the mechanism by which it achieves this is still largely unknown. HR occurs along the homology arms resulting in the mutations present in the vector homology arms being introduced into the host genome. By including a selection cassette we can select for cells that have integrated the targeting vector, and then screen for clones which have undergone targeted insertion rather than random integration, which will generally be around 1%. The presence of this selection cassette can be used to our advantage and allows for high confidence exclusion of off-target integrations.
- AAV has been used to perform a wide range of genome editing events, ranging from standard KI and KO approaches, which it achieves with enhanced precision over any technique, to introduce SNPs, and indels. Generate RNAi rescue models, where silent mutations that abrogate si or shRNA binding can be introduced in an allele specific manner. Gene disruptions have been generated in a number of ways, including promoter traps, in which the endogenous promoter is hijaked by the introduction of a selection cassette preventing expression of the endogenous gene, together with the deletion of specific exons or even whole genes. More complex examples are translocations such as the EML4-ALK translocation (EML4-Alk / Crizotinib Dx Standard), with current projects modelling other amplifications and translocations. So AAV has been demonstrated to be a potent and versatile gene editing technique and has been deployed across over 40 different parental cell line backgrounds, covering 16 tissue types to generate over 500 models.
- Moving on to the nuclease technologies, Zinc Finger nucleases are one of the more established technologies and exploit the non-specific cleavage domain from Fok1. The beauty of this system is that FOK1 only functions as an active nuclease as a dimer. Thus a pair of ZFNs are required for nuclease activity to occur, each consisting of a FOK1 monomer fused to specific zinc finger domains on opposite strands of the DNA a defined distance from each other (5-7bp), these zinc finger domains provide the targeting specificity for the system. Once generated the ZFN pair are introduced into the target cell line by transfection or electroporation, the pair identify their target loci in the nucleus, dimerise and generate double strand breaks in the DNA. In the absence of a DNA repair template such as a synthetic donor, the double strand break is repaired by the mechanism of non-homologous end joining, which is an imprecise mechanism and can result is small deletions or insertions which disrupt the gene by knocking the coding sequence out of frame. If a donor sequence, carrying a desired mutation is introduced alongside the ZFN pairs then HR can be used to repair the double strand break and thus introduce the desired mutation into the genome. The benefit of this technology over, for example CRISPR, is that the risk of off-target nuclease activity is lower due to requirement that both monomers must be in close proximity. Whereas one of the limitations is the requirement for suitable targeting sites in the correct proximity to each other.
- The relative new kid on the block when it comes to genome engineering is the CRISPR Cas9 system, where CRISPR stands for clustered regularly interspaced short palindromic repeats)/cas(crispr-associated), Is very easy to design – which is a major advantage over other nucleases technologies. All that is required is a guide RNA of 20 bp which is complementary to the sequence flanking a protospacer adjacent motif, or PAM site which is a 3 bp sequence consisting of nGG, in the genomic target . In complex with a Tracr RNA, the gRNA directs CAS9 to the target DNA via Wastson-Crick base-pairing whereupon Cas9 mediates cleavage of target DNA to create a double-stranded breaks. As with ZFNs, CRISPR components are delivered by transfection or electroporation, and co-transfection with a donor sequence can result in the double stand break being repaired via homologous recombination thus introducing the desired point mutations, although currently the main strength of the CRISPR system is creating KOs by NHEJ in multiple alleles. Given the relatively new arrival of this technology there are still a number of unanswered questions, in particular is the issue of off-target nuclease activity caused by miss-match binding of the gRNA. The extent to which this is an issue is still under active investigation in many labs. A number of strategies are already in play to overcome this potential limitation of the technology, for example…… n alternative strategy has been proposed in which shorter gRNA sequences, of 17 bp, significantly reduced the number of off-target events, but had no impact on the targeting efficiency at the desired locus, and this work was published earlier this year in nature biotechnology. N17GG/N18GG is as effective as N20GG, but with fewer side effects too Yangfang Fu, Nature Biotechnology 2014.
- In the previous slide I was describing the wild-type Cas9, which introduces a double strand break at the target locus. However, other variants of Cas9 have been created by mutating the nuclease such that it will only nick one strand of the dsDNA, creating a nickase form of the protein. Nicks will in general be repaired by the base excision repair pathway which is significantly higher fidelity than NHEJ. Targeting strategies using the nickase are designed with two gRNAs, one to recruit the nickase to each strand of the DNA, only after which a DSB will be introduced. And this strategy as been shown to be around 1000X more precise than WT CAS9 with a single guide. A ‘Nickase’ version of Cas9 with tandem guides is ~1000x more precise Ann Ran F et al, Cell 2013. Mali P et al, Nature Biotechnology 2013. This increase in specificity is unfortunately at the expense of some efficiency at you’re at the mercy of your weakest guide in the pair
- So that was a brief overview of the different technologies we employ at Horizon. And as we are running gene editing projects every day at Horizon we’ve learnt from experience that there are various ways that things can go wrong if you don’t consider the following, and I want to briefly run through each of these one at a time. Given our key advantage of being technology agnostic, we use the final deliverables of a specific project, as well as the characteristics of the desired cell lines, to determine the technology or technologies to deploy. And this choice can been driven by a number of factors which need to be considered before embarking on an engineering project and we’ll run through some of these now assuming the goal is to introduce a point mutation into a specific cell line.
- The first group of considerations regard the quirks of your specific target gene. How many copies of your gene exist in your cell line? Many of us use transformed human cell lines and there are not many that are actually diploid. For example, many people have used HeLa cells over they years to create models and this cell line is on average quadroploid, as you can see in the bottom right hand side image. Thus they may carry multiple copies of an allele you are trying to modify. Clearly his would have an impact for both KO and KI experiments……KO of 4 alleles is a less frequent event than KO of 2 alleles. Similarly, is KI on one out of 4 alleles sufficient to recapitulate your desired phenotype. Do you need to modify all alleles present? Would KO of one allele and modification of the other be viable/acceptable? This is something that happens frequently with CRISPR. When you make the modification do you expect it affect the growth of the cells? If the effect is detrimental to cell growth or viability this will have an impact on the chances of success, so a thorough literature review to assess any si or shRNA data is prudent as evidence of a growth impact can lead may alter your editing strategy to more of a conditional or inducible approach.
- The second category, and this is probable the one that causes us here at Horizon the most trouble/has required a lot of hard-won expertise - the suitability of the cell line. For example: Can you get DNA into cells Can they be transfected, electoporated or infected This can have a major impact on the technology deployed. Can the cell line be single cell diluted (SCD)? All technologies require a single modified clone to be isolated and expanded to produce the final engineered cell line, and some cell line are unable to tolerate this treatment., even if the cell line grows well in normal culture. We assess a whole panel of media formulations, additives, diff seeding densities (see bottom right image) to identify the conditions most suitable for SCD. Similarly, if the cell line does tolerate SCD, how well does it SCD. Do the cells separate easily or do they tend to clump, as in the example shown in the upper right panel where there is an example of a cell line that can be nicely suspended into a single cell suspension whereas in the upper image is a cell line that you might consider to be sticky, in that they prefer to clump together and it can be difficult to obtain a single cell suspension meaning that possibly multiple rounds of SCD are required to obtain a clonal population. Horizon has now a panel of over 70 cell lines covering 16 different tissue types, including lung, pancreas, breast, brain and kidney that have been shown to be amenable to gene engineering based on their ability to SCD and nucleofect/transduce, which provides a great starting point for engineering projects and has the added bonus of reducing timelines and costs as this characterisation phase can be significantly reduced
- The design of reagents to be used in your editing event is crucial, and although this slide is focused on gRNA design for CRISPR, one consideration is common to all technologies and that is What source you are using for your genomic sequence? Even a single base discrepancy can be the difference between success and failure with CRISPR. It is important that you know what the target looks like IN YOUR cell line. For AAV engineering, there are very few instances in which a donor cannot be designed with confidence, and Sigma are able to custom design ZFNs reliably. However, design rules for CRISPR are still being generated and refined. Horizon together with Desktop Genetics has developed a CRISPR design tool, called gUIDEbook, available via our website, to design gRNA. The design algorithm takes into account all the sites that are obviously a perfect match and up to 1, 2, 3, or 4bp mismatched potential off-target effects elsewhere in the genome. However, sometimes these cannot be avoided; common to have some mismatches. The important consideration is if the potential off-target is in coding or non-coding sequence, and if non-coding is it in a regulatory region? When using a nuclease with a donor it is important to be aware of the distance from the gRNA to the site at which the desired mutation is to be introduced. The closer the double strand break is to the modification you want to make, the most effective the guide will be.
- There is also the question of which Cas9 you wish to use. If it is the nickase version of Cas9, then 2 gRNA are required to induce the double strand breaks. Data suggests that two nicks that result in a 5’ overhang are most efficient at being modified It has also been shown that the distance or “offset” between the two guides is important for efficiency, as can be seen in the graphs in the lower panel, which show that frequencies of indels caused by NHEJ drop quickly as the gRNA offset increases beyond 100 bp with the optimum range being-4bp to 20 bp and therefore this is an important consideration when designing your reagents.
- Once you have guide designs it can pay dividends to validate their activity before hand, in an robust cell line such as HEK293, especially if the chosen cell line has low transfection efficiency. In general we design up to 5 guides (or 10 if using pairs) per target and then assess their activity using the semi quantitiative Surveyor assay, which detects the mismatches created by NHEJ. This done,. In the example on the slide, the 3rd lane on the agarose gel contains a PCR product in he absence of the cel1 enzyme used in the surveyor assay. In the following 5 lanes, labelled 1-5, these are PCR products amplified from cells which have been transfected with Cas9 and a gRNA. The PCR product has been denatured and allowed to anneal creating heterodimers in which one strand has an alteration created by NHEJ and he other remains wildtype. The CEL1 enzyme detects these mismatches and cleaves the PCR product, resulting in the smaller fragments seen in lanes labelled 1 to 4, the gRNA in lane 5 does not seem to be as effective a gRNA. Once you have confirmed that your gRNA direct Cas9 to the correct loci, you can then take just one or two forward for gene targeting in your more challenging cell line of choice.
- Donor design is just as crucial to get right. Especially if the donor is being used in conjunction with a nuclease then the donor must be sufficiently different from the nuclease target to prevent re-cutting once the homologous recombination has occurred. This can be achieved by introducing silent mutations, but these mismatches can impact on the efficiency of HR so we want to introduce as few changes a possible in order to achieve the highest efficiencies, and they may also impact on expression and splicing so this must be considered. Other considerations are whether a selection marker can be employed and which donor type is employed, be it short oligos, plasmids or AAV, is a key decision, and I’ll touch on this again in a few slides
- Armed with all the above information on the nature of the cell line, your transfection efficiency and your guide activity, you next consideration will be how many cells do you need to screen to have a chance of finding a positive. If you look at the pie-chart in the top right hand corner, this gives an example breakdown of the potential frequencies you might expect for the allele modifications in nuclease with donor experiment. AS you can see, aiming to obtain a homozygous knock-in is the least likely event to occur and therefore you must adjust your experiment and screens to account for this if this is your desired outcome. The method you use to screen will depend largely on the modification you are introducing – if for example you’re inserting a tag, you can screen using PCR at that locus. If you’re introducing a frameshift that will disrupt a restriction site, this can be used. Finally, in many knock-outs the modification on each allele will be different, and so the surveyor assay, or similar, can be used followed by sequencing
- Finally, once you have identified a clone or clones that you believe contain your desired mutation the ultimate step is the validate these. Of most interest is certainly going to be the nature of the modifications introduced at your target site – whether for example insertion deletions are present on all of your alleles, and if so, whether they result in frameshifts. You may also wish to assess the off-target cutting in your clones, whether mutations have been introduced at your predicted off target sites, and you can do this using sequencing or Finally, there are various other factors that will be critical to the utility of your cell line. Does the modification express (if it’s a knock-in) or not (if it’s a knock-out). Does your cell line remain genomically stable over multiple passages. And finally, given the length of time cells must remain in culture, and also the degree of handling, we always test our engineered lines for contamination with mycoplasma and other microbes.
- So you can see, there’s a lot to consider and a lot of value in working with a company that has experience in assessing and overcoming these issues. What I would like to do now is present a few brief slides highlighting a few recent improvements Horizon have made to improve the efficiency of genome editing.
- As we have access to multiple technologies we are able to exploit the strengths of each in combination – evidence in the literature suggests that DSBs can stimulate HDR by rAAV. Here we demonstrate that , by combining a CRISPR induced double strand break with rAAV delivery of the donor we can improve efficiency of HDR by 50 fold. In the example shown we are using an in house testing system, which is a cell line containing a mutant (non-fluoresecent) GFP, and we can measure efficiency of targeting by rescue of fluoresence, due to repair by the donor sequence using FACS. This combination has now been employed in “real-life” engineering events and has shown similar levels of targeting improvement.
- Similarily, The majority of transduced cells undergo random integration of the rAAV vector within the genome by Non-homologous end joining (NHEJ) and correctly targeted cells are distinguished from these random integrations by PCR screening. To reduce these off-target events we have developed negative selection cassettes, incorporated into the rAAV vector, to select against random integrations and increase the population of correctly targeted cells. The negative selection marker, is placed outside the region of homology in the rAAV region of homology. Cell that have undergone homologous recombination will lose the marker, whereas cells in which the construct integrates randomly by NHEJ will retain the marker and be eliminated by its expression. The negative selection targets either an essential housekeeping gene required for cell survival; a gene that sensitises to a drug when knocked down (i.e. HPRT followed by HAT selection) or shRNA against a positive selection marker. We have found that use of shRNA across a range of targeting events (inc CDK2, EGFR, BRAF and Abl) have increased efficiency 10-fold, achieving targeting efficiencies of 20-30%. This increases the range of cell lines that can be efficiently targeted and also increases the complexity of targeting events that can be attempted.
- Outside of Horizon, work continues to improve the specificity of the CRISPR system. One approach has been to combine the dimerisation dependent Fok1 nuclease with a catalytically inactive Cas9. like the nickase, it requires a pair of gRNAs, but unlike the nickase, cuts will only occur once a pair of Fok1 subunits are in proximity and can dimerise. This exploits the benefits of ZFNs with the greater targeting versatility of the CRISPR system.
- So now I’dlike to run through a few real-life examples of how we have used the various technologies. In this project, the goal was to knock-out MAPK3 in a triploid cell line. gRNAs designed to cut within conserved exon 3 to introduce indels and induce frame shift mutations Cells were transfected with the gRNA and Cas9 plasmids and 96 clones screened by PCR to identify those that had modified the three copies of the MAPK3 gene Twenty eight clones showed evidence of cutting at the target locus Seven clones were sequenced and of these three showed out of frame indels on all three alleles The very fact that NHEJ is error prone means modifications that are introduced are random, and this can mean that each allele will have a different modification (making sequencing difficult). We therefore use Top cloning to deconvolute
- DNA from the clones was analysed by PCR and TOPO cloning followed by sequencing of the products 5 base deletion 4 base deletion 2 base deletion and 21 base insertion (combined = 19 base insertion) It is important that the deletion is not divisible by three in order to know the transcript out of frame. Ultimately the goal of genome editing is essentially to create “patients in a tube” which recapitulate patient phenotypes and enable the development of new drugs or targets. The following two slides show examples of such models…..
- In this example, the cell line was again created using AAV, with the difference being that the selection cassette was removed by cre-recombinase in order to disrupt the active sites in exon 11 of BRCA2, resulting in a null phenotype. Some cancers in patients carrying BRCA1/2 mutations carry a unique vulnerability as the cancer cells have increased reliance on PARP to repair their DNA and enable them to continue dividing. This means that drugs which selectively inhibit PARP may be of significant benefit in patients whose cancers are susceptible to this treatment. Olaparib is a PARP inhibitor, shown to act against cancers carrying BRCA1 and BRCA2 mutations, and, as can be seen in the lower graph, this cell line nicely recapitulates the drug response phenotype. These two examples demonstrate that genome engineering can create models that recapitulate patient phenotypes, and are therefore a valuable tool in the drug discovery pipeline. Experiments are cell growth experiments comparing growth of different cell lines when treated with the drug. Olaparib experiment: Olaparib (AZD-2281) is an inhibitor of poly ADP ribose polymerase (PARP), an enzyme involved in DNA repair. It acts against cancers in people with hereditary BRCA1 or BRCA2 mutations, which includes many ovarian, breast and prostate cancers. AstraZeneca is currently running a Phase III trial of Olaparib for patients with BRCA mutated ovarian cancer in April 2013. Patients with BRCA1/2 mutations may be genetically predisposed to developing some forms of cancer, and are often resistant to other forms of cancer treatment, but this also sometimes gives their cancers a unique vulnerability, as the cancer cells have increased reliance on PARP to repair their DNA and enable them to continue dividing. This means that drugs which selectively inhibit PARP may be of significant benefit in patients whose cancers are susceptible to this treatment (Wikipedia) Experiment shows that BRCA2 null cells are more sensitive to Olaparib than parental cells.
- In this second example, the panel of cell lines in this study were created using AAV. The top panel shows schematics of how they were created by the introduction of either frameshift mutations, in the case of PTEN to create a PTEN null cell line, or activating mutations in the case of PI3KCA to create constitutively active PIK3CA cell lines. As you can see in the proliferation assay data in the lower panel. The cell lines were used to demonstrate, very cleanly, that either deletion of PTEN, or activation of PIK3CA lead cells to become resistant to tyrosine kinase inhibitor being assessed. Tyrosine kinase experiment: Tyrosine Kinase experiment was a client project so we can not disclose the drug other than it was a tyrosine kinase inhibitor in the experiment. Experiment shows that when you either delete PTEN or active PI3K by mutations, cells become resistant to this tyrosine kinase inhibitor.
- I would like wrap up now by saying that Horizon is the only one stop genome editing shop, able to offer a fully customisable gene editing service. We have experience of complex projects ad have developed a wide range of expertise that can be deployed to create the models you require.
- Before we move onto questions I would like to highlight that in addition to our gene editing services, Horizon is able to offer a range of services to support this discovery pipeline, from basic research through to validated targets. At Horizon, cell line models can be accessed by multiple routes to suit your requirements. Previously created isogenic cell lines can be accessed via our range of of-the-shelf cell lines, or else we can provide reagents and design advice via the GenAssist service through to a full custom cell line development project, generating your specific mutation in the cell line of your choice, performed by Horizons experienced scientists. Target discovery and drug development services are provided by Horizon Discovery Services and our CombinatoRx platform, whilst Sage labs provide custom development of in vivo models. These services can be accessed individually or in combination to allow horizon to support projects of any size.
- I’d also like to take a moment to bring to your attention another webinar which may be of interest which will be discussing Synthetic lethality Target Identification via CRISPR sgRNA screening – an approach we believe holds great promise in uncovering new targets - if you would like to hear more about this exciting new area, please register for our sgRNA screening webinar on the 9th Dec. Cancer is a genetic disease Genome sequencing has generated 100’s of potential targets for cancer therapy However, most targets are rare and poorly characterised Most of them can’t be drugged directly e.g., tumour suppressors If tumour suppressor loss is the prime cancer initiating event, agents exploiting this loss may assist with overcoming tumour heterogeneity Systematic de-orphaning required to find key druggable downstream targets - exploiting co-dependence or synthetic lethality
- This final slide contains links to some useful resources and I’d be happy to take any questions.
- Additional slide options