May 15, 2020 J.A. Zuris, Et Al. Highly Efficient Multi-Gene Knockout and Transgene Knock-In

Home , Gene knockout

Highly Efficient Multi-Gene Knockout and Transgene Knock-in using CRISPR-Cas12a in Induced Pluripotent Stem Cells for the Generation of Engineered Cell Immunotherapies

JA Zuris, CM Margulies, R Viswanathan, JN Edelstein, R Pattali, KW Gareau, SN Scott, S Lele, AC Wilson, JI Moon, N Dilmac, AE Friedland, RA Morgan

John Zuris, Ph.D.

May 15th, 2020

▪ John Zuris is an employee and shareholder at Editas Medicine

CRISPR Gene Editing to Develop Differentiated, Transformational Medicines for High Unmet Need

In Vivo CRISPR Medicines Engineered Cell Medicines

Develop best-in-class medicines for Leverage AAV-mediated editing with SaCas9 hemoglobinopathies using Cas12a and solid into additional therapeutic areas tumors using iPSC-derived cells

Maintain Best-in-class Platform & Intellectual Property, and Advance Organizational Excellence

CRISPR: clustered regularly interspaced short palindromic repeat; SaCas9: Staphylococcus aureus CRISPR-associated protein 9; Cas12a: CRISPR-associated protein 12a; iPSC: induced pluripotent stem cell © 2020 Editas Medicine 3 First in human CRISPR-Cas Gene Editing: EDIT-101 for LCA10

Chromosome 12 position q21.32 Mutated CEP290 Gene

EDIT-101 PAM Mutation

Genome Editing Deletion

Cut and Remove

Repaired CEP290 Gene Transcription and normal mRNA splicing Protein translation Restored CEP290 Protein

Inversion 323 64 IVS26 Exon 26 Exon 27

Deletion

First patient dosed at Oregon Health & Science University (OHSU) Casey Eye Institute, in February 2020

Kozak-ATG pA

SV40 U6 323 U6 64 hGRK1 Sa Cas9 SD/SA

NLS NLS

CRISPR Gene Editing to Develop Differentiated, Transformational Medicines for High Unmet Need

In Vivo CRISPR Medicines Engineered Cell Medicines

Develop best-in-class medicines for Leverage AAV-mediated editing with SaCas9 hemoglobinopathies using Cas12a and solid into additional therapeutic areas tumors using iPSC-derived cells

Maintain Best-in-class Platform & Intellectual Property, and Advance Organizational Excellence

CRISPR: clustered regularly interspaced short palindromic repeat; SaCas9: Staphylococcus aureus CRISPR-associated protein 9; Cas12a: CRISPR-associated protein 12a; iPSC: induced pluripotent stem cell © 2020 Editas Medicine 6 There Are Many Nucleases to Choose from but for Ex Vivo Gene Editing Using AsCas12a Offers Several Potential Advantages to SpCas9

Editas general suite of nucleases AsCas12a is a more specific nuclease compared to SpCas9 Matched Target Site (20-Ns): Variant PAM Frequency (bp) TTTVNNNNNNNNNNNNNNNNNNNNNGGRRT

SpCas9 NGG 1 in 8

SaCas9 NNGRRT 1 in 32

SaCas9 KKH NNNRRT 1 in 8

AsCas12a TTTV 1 in 43

AsCas12a RR TYCV/CCCC 1 in 18

AsCas12a RVR TATV 1 in 43

LbCas12a TTTV 1 in 43

PAM sites are not limiting for general KO and large transgene KI applications

Small ~40 nt Cas12a guide is advantageous for manufacturing and sequence fidelity Matched Target Matched Target Matched Target References on AsCas12a target specificity: Sites 1→25 Sites 1→25 Sites 1→25 Kim et al. Nat Biotech 2016 Strohkendl et al. Mol Cell 2018 From - Gotta et al. Genome Engineering: Frontiers of CRISPR/Cas Cold Spring Harbor, NY October 10, 2019 Swarts et al. Biochem Soc Trans 2019 © 2020 Editas Medicine 7 Engineered Cas12a (EngCas12a) Robustly Edits at Higher Efficiency than WT Cas12a and is Capable of Efficient Knock-in in T Cells

EngCas12a has superior activity High knock-out across loci High AAV6-mediated knock-in

Editing at Multiple Sites in B2M Exon Efficient Editing at Key Targets Knock-in of a Single Transgene Cargo

100 ** 60 ***

S 80

k y 40

60 c – + b

TCR GFP

o g

n 30 n

i – +

K t B2M mCherry

i 40 d

% 20

% 20 10

0 0 TRAC B2M CIITA PD1 6 ly 6 ly V n V n A o A o A A T cells P ls + N + l P R P e N N C R R lt A Based on AsCas12a Ultra nuclease from IDT Up to 40% double knock-in

Robust HbF in human CD34+ cells in vivo High specificity of WT nuclease retained in EngCas12a 10100%0

) Have since verified this attribute in many other cell types

A 80

F 60

(

F 40 >90% editing in HSCs with EngCas12a

b H 52% 20 4% >50% HbF levels that persist after engraftment 0 Unedited EHDBITG-13/021 Edited From – De Dreuzy et al. Annual Meeting of the American Society of Hematology, Orlando. FL, December 9, 2019 © 2020 Editas Medicine 9 The Case for Developing an Allogeneic Medicine for Oncology

Autologous Start Patient Centralized Production Patient Hospital Apheresis Hospital COGs Scope Journey Manufacturing Time - Weeks Treatment $$$

Allogeneic Multiple Patient Healthy Apheresis/gene edit Expand Cells Hospital Treatments Donor $$

Pluripotent Stem Cell

CD34+ Cells

ab T cells gd T cells NK cells

Immune System ADAPTIVEAdaptive INNATEInnate

Tumor Recognition • ab T cell receptor • gd T cell receptor • Innate receptors • CAR • Innate receptors • Antibody-directed • CAR • CAR

Graft-vs-Host Risk Higher Lower

Multiple tumor-recognition mechanisms; potential to NK cells induce adaptive immune combine with therapeutic antibodies to enhance ADCC responses against tumors

• Loss of MHC • NKG2D • NCRs • CD16, etc…

Rezvani et al. Mol Therapy 2017 Souza-Fonseca-Guimaraes, et al., Trends in Immunology, 2019, Vol. 40, 142-158

From - Morgan et al. Keystone Emerging Cellular Therapies: Cancer and Beyond/Engineering the Genome February 10 Banff, Alberta 2020

To date, no evidence of GvHD in patients treated with allogeneic NK cells, unlike off-the-shelf T cells

TGF-b TisG OverexpressedFβ is an immuno sinu pManypress iSolidve fac Tumorstor

abundant in the solid tumor microenvironment

)

B A

F NK cell

m (

Matched TCGA normal and GTEx database Knockout TGFBR2

Stomach cancer HNSCC n=408 (pt) n=529 (pt) n=211 (normal) n=44 (normal)

Souza-Fonseca-Guimaraes, et al., Trends in Immunol, 2019, Vol. 40, 142-158 Modified from Saetersmoen et al. Seminars Immunopathology 2019

From - Morgan et al. Keystone Emerging Cellular Therapies: Cancer and Beyond/Engineering the Genome February 10 Banff, Alberta 2020 © 2020 Editas Medicine 13 Efficient KO of TGFBR2 in Healthy Donor NK Cells with EngCas12a Leads to Expected Loss of TGF-b Signaling

Rationale Editing Biology

KO NK cells, minimal pSMAD2/3 NK cell Editing at TGFBR2 in Three Different Donors D a ta 1in presenceD a ta 1 of TGF-β

100 6 pSMAD2/3 6 N o E P

3 N o E P /

3 3 technical replicates

/ CNoon Et P

/ D

S T G F B R 2 K O 2

80 D T G F B R 2 K O

G A

D KT OG F B R 2 K O N

M 4

y S

60 p

b M

Donor 1 p

z i

Donor 2

t z

40 l

i i

d 2

l a

d 2 e

P a E

Donor 3 m

r o

% 20 2

N m

r 0 0 0 o 0 2 0 6 0 1 8 0 0 2 0 6 0 1 8 0 0.001 0.01 0.1 1 N 10 T im e (m in u te s ) RNP concentration (µM) 0 T im e (m in u te s ) 0 2 0 Stimu6la0 tion wit1h8 010ng/mL TGF-β

T im e (m in u te s ) Cell cycle Anti-tumor progression function

From - Morgan et al. Keystone Emerging Cellular Therapies: Cancer and Beyond/Engineering the Genome February 10 Banff, Alberta 2020 © 2020 Editas Medicine 14 Efficient Transgene Knock-in with EngCas12a in Healthy Donor NK Cells Enables Testing of New Biology

50% knock-in with EngCas12a Tumor-targeting CAR knock-in enhances killing

2:1 E:T ratio

50 1.0

40 0.9

GFP Knock-in

i -

30 S 0.8

20 h 0.7 %

p CAR Knock-in S

10 TRAC site 1 0.6 TRAC site 2 0 0.5 1 10 0 25 50 75 100 125 AAV6 MOI x 1e5 (vg/cell) Time (h)

Enhanced tumor killing with CAR

Summary of what we achieved with EngCas12a platform for potentially making medicines for oncology:

- EngCas12a retains intrinsic AsCas12a specificity advantage over SpCas9

- EngCas12a can robustly edit targets across cell types, including functional KO of TGFBR2 in NK cells

- EngCas12a can efficiently knock-in transgene cargos into T cells and NK cells

Potential Knock-ins

NK cell

Knockout TGFBR2

Off-the-shelf Multiple gene Scalable Off-the-shelf iPSC Hospital edited MCB Patient Treatments Manufacturing $ Advancing Universal Allogeneic Cell Medicines to Treat Cancer

Differentiated iPSC Edited clonal iPSC line iNK cells somOuratic process:cell

Editing + Dedifferentiate Differentiate clonal identification 0 0 Technology from High editing efficiency Infinitely renewable Developing a Highly engineered line BlueRock Therapeutics with engineered Cas12a proprietary, scalable Defined genome differentiation method Low COGs Genome characterization builds Off-the-shelf on in vivo editing capabilities

Potential Knock-ins

First key step in our iPSC editing platform:

Demonstrate highly efficient dsDNA break occurrence (i.e. knock-out) in order to maximize opportunity knock-in efficiency

Potential Knock-out: TGFBR2

Potential target sites for iNK Potent RNP targeting AAVS1 Electroporation pulse code and buffer screen

100 100 100 S

S Recommended Lonza conditions for SpCas9

S 80 80 80

y y y 60 60

60 b

EC < 100 nM

50 g

t t

t 40 i

i 40 40

% 20

20 % 20

0 0 0

1 M 3 4 5 6 0.01 0.1 1 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 S 2 t t t t V B e e e e A rg rg rg rg A a a a a RNP Concentration (µM) Condition T T T T

Additional advances to our iPSC editing platform:

- Made up to four edits at once and developed assays to assess which edit combinations show most promise

- Developed method to edit cells in a consistent manner so that editing and differentiation results are consistent

- Developed method to detect the possible infrequent occurrence of a translocation between two on-target edits

Factors to consider while making a Optimization is needed for plasmid-based highly-edited iPSC-derived NK cell transgene knock-in in iPSCs

Several parameters that could be tested: Donor type

Attributes Electroporation Additives AAV Plasmid ssODN Pulse Code Rock/CloneR Buffer BCL-XL Efficient transgene insertion High Low Moderate NHEJ Inhibitors p53 Inhibitors Toxicity when delivered to cells Moderate High Moderate

Off-target integration events Low Moderate Low RNP Dose Donor Composition Donor Nature Homology arm length Delivery of large cargos Moderate HIGH Low Plasmid Linear plasmid AAV Manufacturing challenge High LOW High Long ssDNA

Efficient Editing at Potential Targets Efficient Knock-in of Transgene Cargo

Editing with EngCas12a RNP + mCherry plasmid mCherry plasmid only

100 7.7 1.9 95.8 0.07

S 80

- -

y 60

MHC MHC I MHC I

i t

i 40

% 20 72.1 18.4 4.11 0.03 0 AAVS1 B2M iPSCs mCherry+ mCherry+

This knock-in efficiency should greatly reduce the number of iPSC clones that need to be analyzed

Process of going from an edited iPSC to a differentiated iNK cell iNK killing of SKOV3 cells Advancing Universal Allogeneic Cell Medicines to Treat Cancer

Differentiated 100 iPSC Edited clonal iPSC line iNK cells somatic cell 80 Editing + Dedifferentiate Differentiate

clonal identification s i

s 60

y l

0 o

t WT y

0 C

Technology from High editing efficiency Infinitely renewable Developing a Highly engineered line 40 KO % BlueRock Therapeutics with engineered Cas12a proprietary, scalable Defined genome differentiation method Low COGs Genome 20 characterization builds Off-the-shelf on in vivo editing capabilities 0 1:4 1:2 1:1 2:1 4:1 8:1 E:T Ratio

▪ EngCas12a is more potent than Wild Type Cas12a but retains its strongly superior specificity compared to SpCas9 ENGCAS12A AS A POWERFUL GENE EDITING PLATFORM ▪ EngCas12a is capable of highly efficient KO and transgene KI across cell types

▪ Allogeneic NK cells are an effective cancer cell therapy without evidence of Graft vs Host Disease WHY NK CELLS FOR CANCER? ▪ Using EngCas12a we obtain robust editing to reduce TGF-β impact on NK cells

▪ iPSCs offer the potential to create Off-the-Shelf therapies at low costs BUILDING A PLATFORM TO ▪ Using the best nuclease possible can increase chances of finding clones with the MAKE HIGHLY ENGINEERED desired edit and inserted construct IPSC-DERIVED NK CELLS ▪ Edited iPSCs can differentiate into iNKs with enhanced tumor killing ability

Editas Colleagues Partners iNK Program Team Allergan Sickle Program Team Bristol-Myers Squibb Healthy Donor NK Program Team BlueRock Therapeutics Discovery Biology Team Sandhill Therapeutics Lead Discovery Team Integrated DNA Technologies Biological Development Team GenEdit Boulder Chemistry Team