Building Better Biologics

Claes Gustafsson, PhD

Co-Founder, CCO

3

ATUMMachine Learning

+Synthetic Biology

• Founded in 2003

• Headquarter in SF Bay Area

• ~85 employees

• Genes/proteins/cell lines

• Acquired Migs April 2016

• 100% employee owned

• Rebranded to Atum Dec 2016

Ecosystem of Integrated Tools

Genes Proteins Cell Lines

Engineer

Ma

ch

ine

Le

arn

ing

Gene Synthesis

Electra Cloning

Phosphoramidite chemistry

DNA ATLAS

Design of Experiment

Leap-In transposase

Sequence-Activity Relationship

Codon Optimization

Data Integration

Robotics

Protein EngineeringDNA Visualization

Markers

Assembly tools

Genetic elements

Assay tools

Engineered Cells

De

sig

n a

xis

Manufacturing axis

PatentsExpertise/Trade secrets

Vectors

DoE/QbD

Vectorology

cGMP cell banking

High Throughput Biologics ProductionGene Design & Synth Protein Expr & Eng Cell Line Dev

Consistent Automated Process

Protein Analytics and CharacterizationProtein XX

XX

Design

BuildLearn

60

60

120

120

Test

Design of Experiment

Synthetic

BiologyMachine

Learning

Biology

Drug

Candidate

Genomic

Data

Codon Optimization

0

0.5

1

1.5

2

2.5

3

-0.5 0 0.5 1 1.5 2 2.5 3

Me

as

ure

d E

xp

res

sio

n

Model-Predicted Expression

FP

scFv

Phi29Pol

R2 = 0.67R2(CV) = 0.59

R²(scFv) = 0.0013

R²(Pol) = 0.0029

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Re

lati

ve

Ex

pre

ss

ion

CAI

~200µg/ml~80µg/ml

• J Biotechnol 2014 10:187. Mellitzer, et al.

• Prot Exp Purif 2012 83:37. Gustafsson, et al.

• Methods Enzymol 2011 498:43. Welch, et al.

• PLoS ONE 2009 4(9):e7002. Welch, et al.

Gene Variants1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

ALA 4 7 1 7 8 3 6 6 5 4 2 2 3 8 2 8 7 4 5 5 6 3 1 5 8 3 5 3 4 7 2 2 1 8 6 6 7 4 6 4 3 8 1 1 2 7 5 Hi

ARG 2 4 7 1 7 8 3 6 6 5 4 2 1 3 8 2 8 7 4 5 5 6 3 5 6 8 3 5 3 4 7 2 2 1 8 6 5 7 4 6 4 3 8 1 1 2 7 Hi

ASN 4 2 4 8 2 8 8 4 6 6 6 2 4 2 4 8 2 8 8 4 6 6 6 6 8 6 8 4 6 4 4 8 2 2 2 6 8 6 8 4 6 4 4 8 2 2 2 Hi

ASP 6 4 2 4 8 2 8 8 4 6 6 2 6 4 2 4 8 2 8 8 4 6 6 6 2 8 6 8 4 6 4 4 8 2 2 6 2 8 6 8 4 6 4 4 8 2 2 Hi

CYS 6 6 4 2 4 8 2 8 8 4 6 2 6 6 4 2 4 8 2 8 8 4 6 6 2 2 8 6 8 4 6 4 4 8 2 6 2 2 8 6 8 4 6 4 4 8 2 Hi

GLN 5 5 5 3 1 3 7 1 7 7 3 1 5 5 5 3 1 3 7 1 7 7 3 5 1 1 1 7 5 7 3 5 3 3 7 5 1 1 1 7 5 7 3 5 3 3 7 Hi

GLU 3 5 5 5 3 1 3 7 1 7 7 1 3 5 5 5 3 1 3 7 1 7 7 5 7 1 1 1 7 5 7 3 5 3 3 5 7 1 1 1 7 5 7 3 5 3 3 Hi

GLY 8 4 6 5 5 3 2 3 8 2 7 2 7 3 5 6 6 4 1 4 7 1 8 5 4 8 2 1 1 7 6 7 4 6 3 6 3 7 1 2 2 8 5 8 3 5 4 Hi

HIS 4 4 4 2 2 2 4 2 4 4 2 6 8 8 8 6 6 6 8 6 8 8 6 6 8 8 8 6 6 6 8 6 8 8 6 2 4 4 4 2 2 2 4 2 4 4 2 Hi

ILE 5 8 4 8 5 5 1 4 1 8 4 6 2 3 7 3 2 2 6 7 6 3 7 5 1 4 8 4 1 1 5 8 5 4 8 2 6 7 3 7 6 6 2 3 2 7 3 Hi

LEU 4 5 8 4 8 5 5 1 4 1 8 6 7 2 3 7 3 2 2 6 7 6 3 5 8 1 4 8 4 1 1 5 8 5 4 2 3 6 7 3 7 6 6 2 3 2 7 Hi

LYS 5 1 5 5 1 5 5 5 1 1 1 7 3 7 3 3 7 3 3 3 7 7 7 5 1 5 1 1 5 1 1 1 5 5 5 3 7 3 7 7 3 7 7 7 3 3 3 Hi

PHE 4 8 2 8 8 4 6 6 6 4 2 8 6 2 8 2 2 6 4 4 4 6 8 6 8 4 6 4 4 8 2 2 2 8 6 4 2 6 4 6 6 2 8 8 8 2 4 Hi

PRO 1 3 7 1 7 7 3 5 5 5 3 7 7 5 1 7 1 1 5 3 3 3 5 6 6 8 4 6 4 4 8 2 2 2 8 4 4 2 6 4 6 6 2 8 8 8 2 Hi

SER 4 2 3 8 2 8 7 3 5 6 5 7 6 8 5 2 8 2 1 5 3 4 3 6 7 5 8 3 5 3 4 8 2 1 2 4 1 3 2 5 3 5 6 2 8 7 8 Hi

THR 5 4 2 3 8 2 8 7 3 5 6 7 3 6 8 5 2 8 2 1 5 3 4 6 2 7 5 8 3 5 3 4 8 2 1 4 8 1 3 2 5 3 5 6 2 8 7 Hi

TYR 6 6 4 2 4 8 2 8 8 4 6 8 4 4 6 8 6 2 8 2 2 6 4 6 2 2 8 6 8 4 6 4 4 8 2 4 8 8 2 4 2 6 4 6 6 2 8 Hi

VAL 6 5 6 4 1 4 8 2 7 7 3 7 4 3 4 6 7 6 2 8 1 1 5 6 1 2 1 7 6 7 3 5 4 4 8 4 7 8 7 1 4 1 5 3 6 6 2 Hi

5'AT Up y n n n y y y n y n n n n y y y n n n y n y y n n y y y n n n y n y y y y n n n y y y n y n n y

RCO n y n n n y y y n y n n y n y y y n n n y n y n y n y y y n n n y n y y n y n n n y y y n y n y

Fluorescence Intensity

Neg

Gene1

Gene2

Gene3

DoE design matrix

Iterative Codon Optimization

And many many more…….

Vectorology for mAb Expression

Enhancer Promoter Intron IRES/2A Localization polyA

CMV Chick Actin Chick Actin Picorna IRES hnRNP A1 Beta Globin

EF1α CMV CMV A SLC7A1 TUS BGH

SV40 EF1α CMV C FGF-2 IRES REV_EIAVY HSV-TK

Synthetic GAPDH EF1α FMDV 2A Jarid2 SV40 early

12

Rituxan

ATUM vectors

IgG

titer

from

tra

nsie

nt H

EK

050

100

150

200

250

300

pC

DN

A3.4 v4

v12 v5 v9 v3

v13.1

v22 v2 v6

v10

v14

v23

v24

v11

v13.2

v16

v25

v17

v18 v7

v19

v20

v15 v8 v1

Expression Vector Optimization

Commercial antibody 1

13

Rituxan

Herceptin (ATUM)

Vectibix

MM121V

23LN

Herceptin

ATUM vectors

IgG

titer

from

tra

nsie

nt H

EK

050

100

150

200

250

300

pC

DN

A3.4 v4

v12 v5 v9 v3

v13.1

v22 v2 v6

v10

v14

v23

v24

v11

v13.2

v16

v25

v17

v18 v7

v19

v20

v15 v8 v1

Expression Vector Optimization

Six different

commercial

antibodies

v13dv5

v4

v13F

v11

v9

v3

v14

v12

v7

v10

v18

v16 v19

v2

v6

v1

v17

v15

v8

v20

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

-0.4 -0.2 0.0 0.2 0.4

Mapping Vector to Antibody

FW1 FW4FW3FW2CDR1 CDR2 CDR3VH

Total explored diversity: ~2x 1019.

Total number of genes made: 96

Measured output:

• Titer post purification

• Tm

• Kinetic off-rate

2 options, 3 options, 4 options, 5 options on 2 leads: FW1, FW2

mAb Framework Protein Engineering

Modeling of Titer

observed titer (mg/l)

pre

dic

ted

tite

r (m

g/l

)

0

0 60

60

120

120Kd scale

0

10-10

parent IgG

Impact of Sequence Variables

impact on titer

imp

ac

t o

n K

d

positivenegative

positive

negative

1

10

100

1000

10000

0.01 0.1 1 10

Cell Based Assay (log log)

IC5

0 c

ell

line

A

IC50 cell line B

ParentRound 1Round 2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

affinity | thermostability | solubility

Multidimensional Optimization

Binding vs Stability vs Humanization vs Titer vs Etc Etc

Winner✓ Cell binding

✓ Antigen binding

✓ SEC HPLC

✓ Tm stable

✓ Humanized

Making Stable Cell Line is a Pain

Gene of

interest

Traditional technology • Integration rate: <<1%• Amplification required? • Concatemer formation• Transgene rearrangement

Two New TransposasesActivity tested in adherent CHO K1 cells

136 constructs

300,000 bp

Synthesize 95 Variants Containing 60 Mutations

Substitutions 1-60

Va

ria

nts

1-9

5

3 substitutions per variant Each substitution

occurs 4-5 times

Five Rounds of Protein GPS®

Hyperactive transposase with >300-fold increased activity

0

1

2

3

4

5

6

7

0 10 20 30 40 50 60 70 80 90 100

inte

grat

ion

fre

qu

ency

, x1

0-3

Variants

Round 1Round 2Round 3Round 4Round 5

500 Constructs1,000,000 bp

Transposase Advantages

Gene of

interest

Traditional technology • Integration rate: <<1%• Amplification required? • Concatemer formation• Transgene rearrangement• Looong timelines

Transposase technology• Integration rate: >90%• ~10 single copy integrations• Structural integrity• Unlimited payload• Unique locations• Footprint-free excision• 10 days from transfection to

stable pool • Grams/Liter

26

Merry Xmas from the elves and Santa at ATUM

Thank You

Technology presented is protected by

US patents 9580697, 9574209, 9534234,

9493521, 9428767, 9290552, 9206433,

9102944, 8975042, 8825411, 8635029,

8412461, 8401798, 8323930, 8158391,

8126653, 8005620, 7805252, 7561973,

7561972 and pending applications

Claes Gustafsson, PhD

cgustafsson@atum.bio