What is Amphora

1
(No Transcript)
2
What is Amphora?

• A small pharma with a world-class discovery
  pipeline of NCEs for cancer, inflammation, and
  CNS disorders advancing toward the clinic.
• Proprietary microfluidics-based, systems
  biology-focused discovery engine utilizing HTS
  industrialization and QC techniques enabling SAR
  from the primary screen.
• Mineable 30 million datapoint (and growing)
  database of high-content chemical biology
  information.

3
Who is Amphora?
Employees

• US based
• Research Triangle Park, NC
• Operations Commenced in 2002
• Privately Held
• Venture Capital Funded
• Strategic Partnerships
• Sanofi-Aventis
• JohnsonJohnson

• 60 Employees with
• Over 40 different Biotech Pharmaceutical
  companies represented
  
• Six Sigma Culture
• gt 80 are Scientists
• Biochemists
• Biologists
• Cellular Pharmacologists
• Chemists
• Engineers
• Enzymologists
• Pharmacologists

4
Amphoras Pipeline
1H06

1H06
1H06
1H06

5
The Engine .. Amphoras Technology
6
Amphoras Industrialized Lab
7
State-of-the-Art Technology Applied to Multiple
Target Families
8
(No Transcript)
9
Enables High Quality HTS Data
Precision radius 95 Confidence Interval
727 gt 6s
6s
3s lt 1327 lt 6s
3s
86,580 lt 3s
The precision radius represents the half-width of
a 95 confidence interval around the percent
inhibition. As the precision radius decreases,
our power to distinguish between compounds
exhibiting similar inhibition increases SAR from
HTS
10
HTS Results are Reproducible
11
Correlate Well with IC50 Data
Calculation of IC50s from primary inhibition data
conc
Inh
conc IC50
Irwin H. Siegel, Enzyme Kinetics, John Wiley
Sons, 1975
12
Enables SAR from HTS Data
120,000 chemical compounds x 65 Targets
Chemical diversity

• Drug-like properties
• Pure
• Characterized
• Quantitated

Biological diversity

• Validated Targets
• Pathways
• Gene Families
• Liabilities/Counter Screens

SAR and selectivity information allow early
decision-making
13
Amphoras Drug-Like Library
What do successful drugs look like?
Iterative optimization of 1000 parameters

• 130,000 compounds
• 95 purity
• Fully characterized

14
Aqueous Quantitation is used
Compounds
Analysis for a random sample of 2000 compounds
from the Amphora library (Popa-Burke et al.,
Anal. Chem. 2004, 76, 7278 7287)
15
to Determine Actual IC50 Values
Irwin H. Siegel, Enzyme Kinetics, John Wiley
Sons, 1975
16
.. Target Families .. ADME Surrogates.....
Cellular Pharmacology
The Assays
17
Kinase Family

• 75 targets developed
• Represents 15 of the kinase genome
• 58 targets screened against the full library
• gt14 million mineable data points
• 28 therapeutic targets
• Coverage of all subfamilies
• 66 kinase panel running monthly
• Full enzymology support through Lead Optimization

18
The Amphora Developed Kinases (75)
TK

Abl Abl (mt) Axl BTK BMX c-Kit CSK EGFR EphB2 EphB
4 ErbB4
PDGFRa PDGFRb SRC SYK TRKA TRKB TIE2 VEGFR3 Yes ZA
P70
Flt1 Flt3 Fyn HCK INSR IGF1R KDR LCK LynA LynB M
et

Other
(32)
(1)
(6)
CMGC
(1)
CDK1/cycB CDK1/cycA CDK2/cycE CDK5/p35 DYRK2 GSK3a
GSK3b MAPK1 p38a p38b p38g p38d
AGC
AKT1 AKT2 AKT3 MSK1 MSK2 p70SGK1 PDK1
PKA PKCbII RSK2 ROCK1 ROCK2 SGK1
(12)
(12)
(13)
(10)
19
Peptide Screens Among Kinase Groups Reveal
Potential Homologous Domains
S/T
S/T, Y
Y
20
Protease Family

• 36 protease assays developed
• Represents 10 of the protease genome
• 24 proteases vs. full library (in process)
• 6,250,000 mineable data points
• All 24 are therapeutic targets
• Coverage of 4 of 5 sub-families
• 24 protease panel running monthly (Currently)
• All enzymology for Hit-to-Lead in place

21
Protease Panel
Currently in Panel
3Q05
4Q05
TBD
ADAM 10 Furin ACE 2 Caspase 8 Caspase 9 Caspase
10 Factor VIIa Factor IXa Trypsin Pepsin
uPa Cathepsin H
Plasmin Seprase(FAP) MMP10 MMP11 BACE2 Granzyme
B RCE1 ADAM12 ADAM23 DPPII/QPP/DPP7 IDE
(insulysin) USPs (DUBs)
ADAMTS1 ADAMTS4 ADAMTS5 ADAM9 ADAM15 Caspase
4 Cathepsin A Cathepsin K Cathepsin S Cathepsin
L Cathepsin G Cathepsin C/DPP1
22
Epigenetic Family

• Full platform project began in Q1 2005
• Substrates have been identified and/or assays
  developed for gt10
• By the end of 2005 Amphora will have
• 12-18 assays developed
• Coverage of all major families as well as all
  relevant epigenetic families (methylases,
  acetylases)
• 3 million mineable datapoints
• HDAC selectivity capability will be a competitive
  advantage

Assays under development
23
Amphoras Microfluidic Ion Channel Assay Format
Agonist Antagonist Chip Formats
24
Microfluidic Ca Detection Assay

• Cells are loaded with two calcium indicator dyes
• fluo4 (emission 530)
• fura red (emission 685)
• Ratio of emissions used to calculate response
• 17 second compound sip 3 second buffer
  wash-out
• 125-150 cells/data point

M1 Muscarinic Receptor Carbachol Concentration
Response
25
Physiological Signal Modifiers Can Be Monitored
With Microfluidic Platform
TRP Family
26
TRPV1 Antagonists Lead Series Overview

• Human TRPV1 transiently expressed at high levels
  in HEK293 cells via BacMam infection
• IC50 values calculated using antagonist
  concentrations measured in aqueous buffer
• Channel activity monitored via calcium influx
• Screening actives vs. agonist (capsaicin)-activate
  d TRPV1

27
The Amphora Engine
2002 2003 2004 2005 2006
Infrastructure Creation
Proof of principle experiments
Kinase Platform
Ion Channel Platform
Protease Platform
HDAC Platform
Amphora Database 28 Million Searchable Data
Points
28
AMPHORA PIPELINE
October 6, 2005
September 2, 2005
29
Selectivity Profiling vs 96 Enzymes Full panel
runs first week of every month Capacity of 4224
compounds in duplicate
24 Proteases
66 Kinases
30
ADME Surrogates Predict Clinical Parameters

• Microsomal Stability and inhibition of cytochrome
  P450 are predictive of clearance and drug-drug
  interactions

Inh C B A
Target A
Liability Target B
Liability Target C
Compounds
31
Cellular Pharmacology Capabilities

• Cellular Proliferation Assays
• Alamar Blue reduction
• 12 point dose response curves in duplicate
• Apoptosis Assays
• Caspase 3 Induction
• Single point or 8 point dose response curves in
  duplicate
• Cytotoxicity Assays
• LDH release
• 4 point dose response in duplicate
• Mechanistic Assays
• Multiplex bead-based Technology for Cellular
  Signaling Proteins
• Luminex Technology
• All assays are 96-well plate based

32
Hallmarks Of Cancer
Six general routes that cancer cells use to
escape internal and external death
stimuli Adapted from The Hallmarks of
Cancer, Douglas Hanahan and Robert A. Weinberg.
Cell, 2000 10057-70.
33
Amphora and the Hallmarks of Cancer
Kinase, protease and epigenetic targets in HTS
complete, panel, or queue
FGFR1 FGFR4 AXL IGF1R ERK1 CK2 MMP 1, 2, 7,
9 EZHZ HDAC III HMT
AKT1 AKT2 AKT3 BMX PIM1 BTK PDK1 IGF1R XIAP
VEGFR1 / FLT1 VEGFR2 / KDR / FLK1 VEGFR3 /
FLT4 TIE2 EPHB4 / HTK PDGFR-b MMP 2, 9,
14 ADAM10, 17
SRC MET FYN YES ROCK1 ROCK2 TRKB / NTRK2 LCK
Aurora-A Aurora-B PLK1 CHEK1 CHEK2
Active Program
34
AKT1 Isozyme Selective Program
Kinase, protease and epigenetic targets in HTS
complete, panel, or queue
AKT1 AKT2 AKT3 BMX PIM1 BTK PDK1 IGF1R XIAP
Active Program
35
Targeting AKT1 for Oncology
The PI3K/AKT pathway regulates fundamental
cellular functions such as cell growth, survival
and proliferation
SURVIVAL
PROLIFERATION
36
AKT and Cancer

• PI3K/AKT pathway
• Frequently disregulated in human cancer resulting
  in activation of AKT
• Thought to have a role in chemoresistance e.g.
  Iressa
• Gene amplification of AKT is found in ovarian,
  breast and colon cancer
• Levels of activated AKT have been correlated with
  progression in a number of cancer types
• Pancreatic, melanoma, thyroid, colorectal and
  breast
• PTEN alterations
• Loss-of-function extremely common in
  glioblastoma, melanoma, prostate, endometrial
  cancers
• Significant of breast, lung and lymphomas have
  mutations in PTEN
• PI3K gene amplification in colorectal, gastric,
  breast and brain cancers
• In colorectal cancer with mutated p110 gene
  AKT1 was predominant isoform

37
AKT1 HTS Analysis
38
AKT1 Lead Series SAR Overview
Extensive SAR substituent effects provide
enhanced potency
Strict SAR
Extensive SAR functionalization
provides improved potency, physical properties
39
Panel Data Helps Direct Chemistry
40
Amphora AKT Inhibitors Are Highly Selective
Relative activity determined by ratio of IC50.
41
AKT Inhibition is PH Domain-Dependent
inh. _at_ 30 mM

• Inhibition mechanism is similar to recent Merck
  publication

42
AKT1 Lead Compound Summary
1 Values are for a single compound. 2 Ki and
IC50 values calculated using inhibitor
concentrations measured in aqueous buffer. 3
Number of screens against which the active
analogs in the series have been tested. 4
Mechanism appears to involve active site binding
and requires the PH domain.
43
AKT1 Lead Rapidly Optimized to in vivo Efficacy
Studies

• Series with SAR from screen
• Approximately 165 compounds to identify potential
  candidate
• 3 Chemistry expansions of 35, 110 and 20 each
• Pre-clinical evaluation of lead compound underway
  
• Clinical advisors in place
• Expect first dose in man during 1H06
• Backup series have been identified

44
Targeted Tyrosine Kinase Program (TTK)
Kinase, protease and epigenetic targets in HTS
complete, panel, or queue
FGFR1 FGFR4 AXL IGF1R ERK1 CK2 MMP 1, 2, 7,
9 EZHZ HDAC III HMT
EGFR ERBB2 / HER2 ERBB4 / HER4 PDGFR-a FMS /
CSFR1 KIT FLT3 / STK1 / FLK2 ABL / BCR-ABL
Active Program
45
Selection of a Targeted Tyrosine Kinase (TTK)
Inhibitor

• Targets involved with successful clinical kinase
  inhibitors
• e.g. Gleevec, Iressa, Tarceva, SU11248, others
• Combinations of these targets desirable
• e.g. ABL, EGFR, FLT-3, c-KIT, PDGFRs
• Selectivity against other target classes
• e.g. phosphatases, proteases, ion channels
• Balance selectivity to reduce side effects
• Maintain multi-target activity to reduce
  resistance due to mutations

46
Profiles of Kinase Inhibitors, Marketed Drugs and
Competitor Compounds

• Targets sorted alphabetically
• Panel shown above includes kinases and proteases

47
IC50 Profile of Clinical Kinase Inhibitors
CML and GIST
NSCLC
NSCLC
GIST and RCC
48
Potency Profiles of Clinical Inhibitors at
Steady-state PK
Imatinib (Gleevec) 400 mg/d (28 d) s.s. 4.2 mM
Tarceva (OSI-774) 150 mg/d (24 d) s.s. 4.9 mM
Gefitnib (Iressa) 225 mg/d (14 d) s.s. 0.6 mM
SUTENT (SU11248) 50 mg/d (14 d) s.s 0.23 mM
49
Target Similarity Can Be Measured By Compounds

• These three targets look similar to the
  compounds
• The degree of similarity between them can be
  explicitly measured
• A chemical genomic tree can be derived from
  pairwise target similarities across the entire
  database

50
Compound/Target Relationships
FYN
ZAP70
MET
CASPASE-7
P38-ALPHA
VHR
CD45
P38-BETA
CASPASE-2
SHP-1
TCPTP
LYNA
LCK
PDK1
CASPASE-3
MAPKAPK-3
BMX
P70S6K1
TRKB
HCK
DAPK1
PDGFR-ALPH
MSK2
AKT1
PLK1-002
KIT
NILL
FLT-3
SRC
ABL-T315I
PLK1
CSK
CDK2-CYCLI
ABL1
CK1
CHEK2
BACE-1
CDK1
CDK5
CDK2
AURORA-A
PRAK
GSK-3-ALPH
CHEK1
GSK-3-BETA
CK2
DYRK2
PKA
Diversity screen reveals Chemogenomic
relationships
INSR
MAPK1
C-TAK1
P38-GAMMA
AKT2-002
AKT2
P38-DELTA
SYK
PP2AC
AKT3-002
SGK1
MSK1
MAPKAPK-2
PTP-BETA
AKT3
LAR
PAK2
NEK2
51
Poly Tyrosine Kinase Inhibitor
52
Starting Point for a Selective Multi-Kinase
Series

• Group FLT-3, c-KIT, and PDGFR as primary target
• Sequence similarity at the active site
• c-KIT, FLT-3, PDGFRs commonly mis-regulated in
  cancers
• Mining and evaluation of multiple scaffolds from
  the existing profiling database
• Use SAR from 1o data and start new chemistry

53
Database Mining for TTK Inhibitors
54
Series Selection TTK Inhibitors
18 distinct, selective chemotypes identified
3 Scaffolds selected for chemical expansion
55
Industrialization of Enzymology and Lead
Identification
New compound synthesis SAR from 1o screen
1
2
3
Multiple scaffolds enter queue
56
Selectivity is Retained for Series 3
Profiling at 10 mM, ranked by Ki at 30 mM

• Good agreement between profile and potency
• Excellent selectivity and potency seen for
  c-KIT, PDGFRa, PDGFRb
• Weaker activity against FLT-3, AURORA-A

57
Rapid Progression from HTS Hits to Lead
Identification

• 3 scaffolds (213 compounds) selected for
  expansion based on SAR, chemical properties, and
  IP
• All compounds profiled in monthly enzyme panel
• Enzymology in parallel against c-KIT, FLT-3,
  PDGFR
• Series prioritized by SAR and mechanism
• 2 Selective series with low nM potency against
  multiple oncology targets in less than 5 weeks

58
TTK Lead Series Overview
Series 3
Series 2
Data is from one cycle of chemistry optimization
59
Aurora Kinase Program
Kinase, protease and epigenetic targets in HTS
complete, panel, or queue
Aurora-A Aurora-B PLK1 CHEK1 CHEK2
Active Program
60
Aurora Kinase Lead Series Overview
Series 3
Series 1
Series 5
Series 4
61
Shortening the Time to the Clinic
The bottom line 1 year from the Amphora
database to candidate selection
62
The Drug Discovery Process
63
Acknowledgements
Amphora
Perkins, Erin Popa-Burke, Ioana Pressley,
David Rayburn, Chris Read, Alysha Riley,
Jeff Savas, Peter Schweiner, Mike Sparkman,
Darren Steed, Paul Suto, Carla Teilmann,
Jon Thompson, Tracy Turnbough, Julie Van De Carr,
Gretchen Veasey, Lee Walker, Clint Williams,
Kevin Zink, Jennifer
Aquesta, Larry Allen, Dwayne Arroway,
Jamie Birkos, Steve Blackwell, Len Blake,
Sean Bobasheva, Anna Brock, Dean Cheatham,
Lynn Chen, Ke Chen, Min Churchill, Carl Clark,
Jennifer Coudurier, Louis Dickson,
John Dougherty, Bob Evans-Storms,
Rose Galasinski, Scott Greene, Nathalie Hallam,
Rhonda
Hamsioglu, Berk Hardy, Brian Hodge, Nick Hunt,
Ken Johnson, Daniel Johnston, Rob Kidd,
Alison Lewis, Rod Licause, Joe Luft
Chris Mendoza, Jose Miller, Jennifer Mitchell,
Lou Ann Mohney, Rob Murray, Jeffrey Norris,
Jacqueline Nye, Beth Orcutt, Matt Patterson, Tom
64
Acknowledgements
Duke - Comprehensive Cancer Center Kim Lyerly and
Gay Devi Herb Hurwitz and Andy Nixon Duke
Brain Tumor Center Henry Friedman and Jeremy
Rich Duke Jeff Rathmell Dana Farber Cancer
Institute Tom Roberts and Jean Zhao Memorial
Sloan Kettering Neal Rosen and David
Solit Vanderbilt Mace Rothenberg Carlos Arteaga
Harvard Medical School Lew Cantley and Jeff
Engelman Gary Gilliland UNC Kim
Rathmell NCI Phillip Dennis Emory Harry Findley
Scientific and Clinical Advisors Greg Verdine
Harvard Stephen Carter Dan Hoth
65
(No Transcript)