Methods for Gene Editing Measurement and Off-Target Discovery

Transcription

1 Methods for Gene Editing Measurement and Off-Target Discovery NIST-FDA Genome Editing Workshop April 23, 2018 Gaithersburg, MD Christopher Wilson, h.d. Senior Director, Lead Discovery editasmedicine.com 1

2 Disclosures for Chris Wilson Employee and stock option holder of Editas Medicine 2

3 Agenda LCA10 and CE290 background UDiTaS to measure large deletions and inversions at the CE290 editing site to measure translocations development of accuracy standards for structural changes Specificity approaches Digenome Statistical framework for describing off-target verification 3

4 Leber Congenital Amaurosis Type 10 Infantile-onset of poor vision, nystagmus, and a flat electroretinogram 1 Caused by autosomal recessive mutations in the CE290 gene at chromosome12q ~85% of LCA10 patients from northwest Europe have a IVS26 mutation in intron 26, c a>g 1-7 CE290 is present in the connecting cilium and is important for ciliogenesis, ciliary trafficking, and outer segment function and structure 8 WT hotoreceptor Outer segment Inner segment Discs Connecting cilium CE290 LCA10 hotoreceptor Outer segment Inner segment Degenerated discs low / no CE290 1 Weleber RG, 2013 LCA Gene Reviews; 2 den Hollander AI, Koenekoop RK, Am J Hum Genet 2006;79:556; 3 Stone EM, Am J Ophthalmol 2007;144:791; 4 CE290_database 2017; 5 errault I, Hum Mutat 2007;28:416; 6 Vallespin E, IOVS 2007;48:5653; 7 Simonelli F, IOVS 2008;48:4284; 8 Rachel RA, Cilia 2012;1:22 4

5 Gene Editing to Repair CE290 Splicing Defect DNA CE290 with IVS26 mutation IVS26 Exon 26 * Exon 27 CE290 corrected with EDIT-101 Inversion IVS26 Exon 26 Exon 27 * Deletion Transcription mrna Exon 26 X Exon 27 Exon 26 Exon 27 Translation rotein p.cys998x rematurely truncated and non-functional Full length, functional CE290 NIST-FDA NIST-FDA Genome Editing Workshop April 23, 2018 Gaithersburg, MD Christopher Wilson 5

6 Challenges with CR-NGS assays when making multiple edits roductive deletion 1.1 kb 3 CR assays needed to measure editing at CE290 intron 26 locus Even with rigorous standards it is difficult to cross compare assays roductive Inversion 1.1 kb Another set need for inversions 1.1 kb ddcr sufficient to measure the deletion but unable to distinguish inversions from wild type locus roductive deletion NIST-FDA NIST-FDA Genome Editing Workshop April 23, 2018 Gaithersburg, MD Christopher Wilson 6

7 Uni-Directional Targeted Sequencing UDiTaS* An NGS method for measuring junctions (and indels!) a b c Custom Transposon Genomic DNA from edited Editing event cells CE290 locus guide 64 guide Tn5* Tn5* 3 ddc ß UMI Barcode ß ooling Barcode ß i5 Custom Transposon i5 primer i5 BC UMI CR Round 1 CR Round 2 + Tagmentation ~ 2 kb fragments Sequencing Library Sequence specific primer BC Barcode i7 2 nd round primer i7 1,176 bp Detection of editing UMI UDiTaS primer Detection of large deletions UMI New junction UDiTaS primer Detection of inversions New junction UMI UDiTaS primer * Giannoukos, et al., UDiTaS, a genome editing detection method for indels and genome rearrangements, BMC Genomics, :212 7

8 UDiTaS finds novel structural changes and is robust U-2 OS cells nucleofected with plasmids expressing SaCas9 + grna64 + grna323 gdna from stable HEK293 line with deletion and inversion mixed with HEK293 gdna at various ratios 8

9 UDiTaS in action: hundreds of samples from mouse pharmacology experiments Measured by UDiTaS Total Editing Rates (%) Cas9 mrna (Molecule/ug RNA) K/D relationship: On-target Editing Correlates with Transgene Expression by EDIT-101 in HuCE290 KI Mice 9

10 UDiTaS can measure translocations CD4+ human primary T cell nucleofected with 2 RNs: TRAC chr14 1 Schematic Type 1 (%) 2 (%) No edits n/a Indels n/a Dicentric n/a n/a B2M chr 15 2 Acentric n/a No edits n/a 3.89 Indels n/a possible outcomes 7 measurable with primers 1 and 2 All 7 events detected Dicentric n/a 0.42 Acentric n/a n/a Balanced n/a n/a Balanced Acentric 0.66 n/a Dicentric n/a

11 Accuracy of translocation measurements Created series of 6 plasmid standards at the predicted TRAC-B2M translocation adding SNs every ~10bp TRAC primer (#1)! = 4.49 & '.(1 * * * * * * * * TRAC B2M * * * * * * * * TRAC B2M! = & '.)3 B2M primer (#2) Individually dilute each plasmid ~3,000 molecules to 3e6 molecules (3 logs) Spike into mouse genomic DNA Run UDiTaS and AM-Seq Demonstrates high accuracy and linearity of UDiTaS! = 6.06 & '.()! = & '.)/ NIST-FDA NIST-FDA Genome Editing Workshop April 23, 2018 Gaithersburg, MD Christopher Wilson 11

12 Editas approach to editing specificity Gene editing agent(s) In silico closest matches Biochemical Digenome-Seq Cellular GUIDE-Seq Discovery hase Targeted Sequencing Assay anels in cells as relevant as possible Verification hase Off-target sites Risk assessment and mitigation 12

13 Digenome-Seq with Lead Guides (64 and 323) CAS9 + Guide + genomic DNA 16 Align & hours WGS analyze Identify overabundant start sites CE290 Guide 323 on-target example EMX1 (SpCAS9) CE290 g64 (SaCAS9) CE290 g323 (SaCAS9) No CAS9 85 Count of sites Includes on-target site nm 100 nm 1,000 nm 10 nm 100 nm 1,000 nm 10 nm 100 nm 1,000 nm - 13

14 Statistical framework for sensitivity calculations Similar to small molecule safety profiling Off-target measurements needs to include the limit of detection Express no detected offtargets as <LLoD; eg: Editing at chr1: is <0.1% Main determinants are: Read count Input DNA amount # of UMIs for >= n obs. 1,000, ,000 10,000 1, Binomial sampling distribution in NGS as part of an LLoD determination 95% confidence % editing n

15 Summary LCA10 and CE290 background UDiTaS to measure large deletions and inversions at the CE290 editing site to measure translocations development of accuracy standards for structural changes Specificity approaches Digenome Statistical framework for describing off-target verification (and lack of off-target measurement) 15

16 Thank you. editasmedicine.com 16