Genome engineering using CRISPR/Cas9: what could go wrong? Genome - PowerPoint PPT Presentation

Genome engineering using CRISPR/Cas9: what could go wrong? Genome Editing 2020 St Edmund Hall, Oxford, UK 12 th March 2020 Katharina Boroviak

Introduction to CRISPR/Cas9 Doudna, Charpentier and colleagues apply Cas9 in gene targeting.

Introduction to CRISPR/Cas9 G A A U A G A C G A U G C U A G C C G U A guide RNA G C A G C A A G G G G U U G A U G U U A U C G A A U G U A UUUUUU UGAA A A U A A NNNNNNNNNNNNNNNNNNNN G UAAGGCUAGUCCGUUAUCAACUU G A Cas9 3’ 5 ’ NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNCCNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGGNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 3’ 5 ’ genomic target PAM

Introduction to CRISPR/Cas9 G A A U A G A C G A U G C A U G C G C A U guide RNA G C A G C A A G G G G U U G A U G U U A C G U A A U G U A UUUUUU UGAA A A U A A NNNNNNNNNNNNNNNNNNNN G UAAGGCUAGUCCGUUAUCAACUU G A 3’ 5 ’ NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNCCNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGGNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 3’ 5 ’ Cas9

Introduction to CRISPR/Cas9 Insertion of small ssDNA oligos  loxP, point Indels, mutations, … Seamless insertion of Small or large deletions, targeting constructs  HDR inversions and GFP, lacZ , … duplications HR NHEJ DSB is repaired 3’ 5 ’ NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN NNNCCNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN NNNGGNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN 3’ 5 ’

ESC vs CRISPR/Cas9 ~3 – 6 weeks ~8 – 10 weeks 10 - 11 weeks 4 - ? weeks Mashiko D, Young SAM, Muto M, et al (2013) Feasibility for a large scale mouse mutagenesis by injecting CRISPR/Cas plasmid into zygotes. Dev Growth Differ. doi: 10.1111/dgd.12113

ESC vs CRISPR/Cas9 ~3 – 6 weeks ~2 – 6 weeks ~8 – 10 weeks 10 - 11 weeks 4 - ? weeks ~6 weeks Is it really this simple? Mashiko D, Young SAM, Muto M, et al (2013) Feasibility for a large scale mouse mutagenesis by injecting CRISPR/Cas plasmid into zygotes. Dev Growth Differ. doi: 10.1111/dgd.12113

Delivery of CRISPR/Cas9 via cytoplasmic injection Delivery of Cas9 (mRNA or protein) and gRNA via cytoplasmic injection into one cell mouse embryos (= E0.5) followed by embryo transfer 1-2 hours after injection or culture to blastocyst (= E3.5) for further analysis. Brendan Doe

Mosaicism For initial studies (e.g. test different versions of Cas9 mRNA/protein) we targeted tyrosinase ( Tyr ) - re-arrangements/indels in this region should be non-lethal - ability to score mutation events phenotypically (C57BL6/N zygotes: heterozygous mutation = black coat colour; homozygous mutation = albino patches or complete albinism).

Mosaicism  Up to 5 different alleles observed in mice and at blastocyst stage

Mouse genome engineering Routine: • Critical exon (CE) deletions CRISPR/Cas9 in zygotes • Single nucleotide polymorphism (SNPs) • Conditional Knock-outs • Knock-ins (reporters, humanisation, conditional point mutations, ESC Ectopic/over-expressors) • EUCOMM/KOMP ‘knockout first’ models R&D using CRISPR/Cas9 in zygotes: • Optimization of conditions • Conditional alleles • Homologous recombination • Genomic rearrangements

Mouse genome engineering Routine: • Critical exon (CE) deletions CRISPR/Cas9 in zygotes • Single nucleotide polymorphism (SNPs) • Conditional Knock-outs • Knock-ins (reporters, humanisation, conditional point mutations, ESC Ectopic/over-expressors) • EUCOMM/KOMP ‘knockout first’ models R&D using CRISPR/Cas9 in zygotes: • Optimization of conditions • Conditional alleles • Homologous recombination • Genomic rearrangements

Mouse genome engineering Routine: • Critical exon (CE) deletions CRISPR/Cas9 in zygotes • Single nucleotide polymorphism (SNPs) • Conditional Knock-outs • Knock-ins (reporters, humanisation, conditional point mutations, ESC Ectopic/over-expressors) • EUCOMM/KOMP ‘knockout first’ models R&D using CRISPR/Cas9 in zygotes: • Optimization of conditions • Conditional alleles • Homologous recombination • Genomic rearrangements

Mouse genome engineering Routine: • Critical exon (CE) deletions CRISPR/Cas9 in zygotes • Single nucleotide polymorphism (SNPs) • Conditional Knock-outs • Knock-ins (reporters, humanisation, conditional point mutations, ESC Ectopic/over-expressors) • EUCOMM/KOMP ‘knockout first models R&D using CRISPR/Cas9 in zygotes: • Optimization of conditions • Conditional alleles • Homologous recombination • Genomic rearrangements

Genomic Rearrangements Genomic rearrangements often present in inherited disease and cancer and often involve gross alterations of chromosomes: • Deletions • Duplications • Insertions • Inversions • Translocations Modelling these structural variants in mice required multistep processes in ES cells  Time consuming and limited their availability.

Genomic Rearrangements

Genomic Rearrangements What are the limits of using CRISPR/Cas9 to generate rearrangements directly in mouse zygotes?

Genomic Rearrangements

Genomic Rearrangements

Genomic Rearrangements

Genomic Rearrangements

Genomic Rearrangements Injections and births G0 mice with Embryos Pups born Deletion size Deletion (%) Inversion (%) Duplication (%) transferred (%) 155kb 105 46 (44%) 11 (24%) 14 (30%) 1 (2%) 545kb 114 68 (60%) 12 (18%) 12 (18%) 1 (1%) 1.15Mb 103 48 (47%) 14 (29%) 10 (21%) 0 (0%)

Genomic Rearrangements

Genomic Rearrangements

Genomic Rearrangements Example #1 Founder endpoint PCR: WT, DEL, INV RV Offspring endpoint PCR: WT, DEL, INV RV X

Genomic Rearrangements Example #1 Founder endpoint PCR: WT, DEL, INV RV Offspring endpoint PCR: WT, DEL, INV RV X

Genomic Rearrangements Example #2 Founder endpoint PCR: WT, DUP1 Offspring endpoint PCR: WT, DUP1

Genomic Rearrangements Example #2 Founder endpoint PCR: WT, DUP1 Offspring endpoint PCR: WT, DUP1

Genomic Rearrangements Other examples 1.15Mb region endpoint PCR: INV RV Duplication of yellow and red probes

Genomic Rearrangements Other examples Chr1 region spanning 664kb from Cfhr1 to Cfh * endpoint PCR: WT * Unpublished data: Daniel Gitterman, Matthew Pickering

Genomic Rearrangements Although CRISPR/Cas9 is very efficient in generating genomic rearrangements, be careful on how you characterize them! • Loss of primer binding sites can result in incomplete genotype. • Endpoint PCR can identify correct deletion event but the excised fragment can re-integrate locally and potentially restore activity of the excised gene. • WT animals used as littermate controls might actually harbour a rearrangement. • Even if all the PCRs match up, internal fragments can be lost. Observed similar events in other genomic regions as well as in some cases of our routine critical exon deletions.  Thorough analysis of alleles generated by CRISPR/Cas9 is crucial.  Do not breed F0s together and assume they generate the desired line.

Off Targets? • Used two gRNAs within exon2 of Tyr locus Number of nucleotide mismatches CRISPR Sequence Off-target totals 0 1 2 3 4 Target only 1 0 0 4 51 55 Tyr2F Target +1 (DNA bulge) 0 3 5 75 - 83 Target -1 (RNA bulge) 2 2 8 245 - 257 395 Target only 1 0 2 15 161 178 Tyr2R Target +1 (DNA bulge) 1 1 12 324 - 338 Target -1 (RNA bulge) 0 4 71 911 - 986 1,502 • Analyse the genomes of both parents and offspring using WGS • Separate de novo mutations from variations within the population/strain • If off-targets are an issue in our conditions, we would expect to see an increase of de novo mutations in treated animals compared to the controls.

Off Targets? targeting efficiency: experiment #embryos number of origin experiment # pups analysed (%) # with mosaic genotype (%) average variants per embryo transferred genotype confirmed pups (% of pups analysed) 474 142 (30%) 76 (54%) 35 (25%) 1 historical Cas9 protein (all) Tyr2R RNP (equiv. 30 13 (43%) 11 (85%) 7 (54%) 2 historical conditions) 8 8 (100%) 0 (0%) 0 (0%) 0 current parent 15 11 (73%) 0 (0%) 0 (0%) 0 current No Injection 15 11 (73%) 0 (0%) 0 (0%) 0 current Sham Injection 15 6 (40%) 0 (0%) 0 (0%) 0 current Cas9 only 14 8 (57%) 7 (88%) 6 (75%) 2 current Tyr2F 13 11 (85%) 10 (91%) 7 (64%) 2 current Tyr2R • 18/20 alleles detected by MiSeq were also detected by WGS • The two missing alleles were at low frequency (10%) and hence filtered out during WGS analysis due to low read number

Off Targets? • In our conditions, using cytoplasmic injection, we see efficient rate of on- target mutation but no introduction of off-targets • Pick best available gRNA and compare design using multiple design tools

Off Targets? How the controls can skew the results:

Off Targets? How the controls can skew the results: