Building a deep immuno-oncology portfolio in less than a year. David Johnson, PhD PEGS Spring 2018

Transcription

1 Building a deep immuno-oncology portfolio in less than a year David Johnson, PhD PEGS Spring 2018

2 Improving early development years, >$billion for an innovative new drug Attrition rate Target discovery Lead discovery CMC Phase I Phase II Phase III 5 yrs 3 yrs 2 yrs 2 yrs 2 yrs 3 yrs

3 Improving early development Speed + Attrition rate = Cost + Products Attrition rate Functional genomic screens Massively parallel discovery Fully natural, fully human Personalized clinical studies Target discovery Lead discovery CMC Phase I Phase II Phase III 2 yrs 2 yrs 2 yrs 2 yrs 2 yrs 3 yrs

4 Immune checkpoint targets Lots of groups working on a few immune checkpoint targets GigaGen generated antibodies against 17 known immune checkpoint targets We are using these resources to investigate how immune checkpoint interact Eventual goal is to discover novel targets and novel combinations Nature Biotechnology 32, (2014)

5 Checkpoints in the tumor microenvironment RNA-seq from 38 cholangiocarcinoma tumors correlation coefficient

6 Large portfolios are critical for combo screens Owning a vast diversity of antibodies against immune checkpoint targets helps identify candidate combinations. PD1 associates with LAG3 in T cells mabs block PD1-LAG3 association Oncotarget 6(29), (2015)

7 GigaGen s founders + = David Johnson Genomics PhD Stanford COO & Co-Founder of Natera Everett Meyer Immunology MD/PhD Stanford Stanford Professor World-leading immune genomics technology for drug discovery MAbs 2017 Nov/Dec; 9(8): MAbs 2017 Nov/Dec; 9(8): MAbs 2018 Apr; 10(3):

8 Ultra-fast single cell capture Throughput of >3 million cells per hour. carrier oil B cells Droplet reactors lysis mix carrier oil

9 Fast & comprehensive antibody discovery Humanized mouse immunization Generate scfv library scfv screening antigen binding scfv expression Immunize humanized mouse with soluble or cell-surface expressed antigen Isolate millions of B cells into droplets and amplify antibody coding DNA Express millions-diverse scfv libraries on microbial cell surfaces Flow sort for binders against antigens Deep sequence before and after sorting to identify antibody clades of interest

10 Yeast display discovers high affinity binders Yeast display libraries cover the full diversity of each mouse (millions of antibodies) Native heavy-light chain pairing is retained Deep sequencing (Illumina) quantifies all sequences in preand post-sort libraries Three rounds of sorting captures the highest affinity antibodies Fc (-) Control LAG-3-His 1 st sort LAG-3-His 2 nd sort LAG-3-His 3 rd sort Deep sequencing Antigen (PE) c-myc (AF488)

11 PD-1: different Abs in different tissues & mice Mouse A Mouse B Mouse C LN spleen LN spleen LN pre post pre post pre post pre post pre post 99.0% 262 high affinity binders 30.6% 9.0% 2.2% 0.0% Enriched scfv >0.1% post-sort abundance are shown

12 PD-1: a diversity of antibody sequences 262 unique high affinity scfv binders to PD-1 Enriched scfv >0.1% post-sort abundance Each color represents a different mouse Distance = Sequence Similarity Circle/Square = Spleen/Lymph Node Lines indicate less than 9 amino acid difference across heavy & light chain

13 LAG-3: different Abs in different tissues & mice Mouse A LN spleen pre post pre post Mouse B LN spleen pre post pre post Mouse C LN spleen pre post pre post 5.3% 467 high affinity binders 4.0% 3.0% 2.2% 1.5% 1.0% 0.6% 0.2% 0.0% Enriched scfv >0.1% post-sort abundance are shown

14 LAG-3: a diversity of antibody sequences 467 unique high affinity scfv binders to LAG-3 Enriched scfv >0.1% post-sort abundance Each color represents a different mouse Distance = Sequence Similarity Circle/Square = Spleen/Lymph Node Lines indicate less than 9 amino acid difference across heavy & light chain

15 Prioritizing antibodies for development Manufacturability: Pilot expression as full-length mab in CHO cells Affinity kinetics: array surface plasmon resonance (SPR) using Carterra technology Epitope binning: Array SPR using Carterra technology Cell surface binding & polyspecificity: FACS of antibody vs. CHO cells expressing target or irrelevant target In vitro efficacy: reporter assays (Promega) and CMV recall assays Mouse efficacy CMC scale up, toxicology, PK IND

16 PD-1: a diversity of antibody kinetics 65% of scfv bound antigen when reformatted as mabs K D nM (pembro benchmark was 16nM in our hands)

17 LAG-3: a diversity of antibody kinetics 88% of scfv bound antigen when reformatted as mabs K D nM (benchmark was 0.5nM in our hands)

18 FACS cell surface binding assay FACS binding assay: Bind antibody to cells that express target Bind antibody to cells that express irrelevant target Absolute intensity and spread between curves indicates on-target and off-target binding Specific binder Non-binder Target cells Irrelevant cells Target cells Irrelevant cells

19 PD-1: FACS cell surface binding data All antibodies that bound soluble antigen also bound cell surface antigen No antibodies bound cells expressing irrelevant target representative plots shown

20 LAG-3: FACS cell surface binding data 58% of antibodies that bound soluble antigen also bound cell surface antigen No antibodies bound cells expressing irrelevant target representative plots shown

21 PD-1: a diversity of blocking functions No anti-pd-1 antibodies beat pembro so far Poor response to immunogen means fewer antibodies, less likely to block well PD-1 blockade pembro K D =24nM K D =96nM K D =24nM K D =33nM K D =66nM K D =10nM no binding Fc control Antibody Concentration (nm)

22 LAG-3: a diversity of blocking functions Two antibodies beat benchmark Higher affinities more likely to beat benchmark LAG-3 blockade K D =1.8nM K D =0.8nM BMS benchmark K D =3.0nM K D =8.3nM K D =3.5nM K D =11nM Fc control Antibody concentration (nm)

23 PD-1: clear delineation of functional epitopes Blocking epitope bin Non-blocking epitope bin

24 LAG-3: functional epitopes less clear not blocking blocking & not blocking blocking not blocking not blocking blocking (beats BMS benchmark) blocking (benchmark) not blocking not blocking not blocking blocking blocking

25 Current directions Immunization campaigns with DNA and cell-expressed targets Yeast FACS with cell lysates to capture membraneembedded targets Exploring in vitro assays to identify synergistic antibody combinations Expression profiling of drug-exposed immune cells and terminally exhausted immune cells to discover new targets and further non-clinical and clinical development of promising candidates

26 Thank You