The NCI Approach To Drug Development

1
The NCI Approach To Drug Development

• Edward A. Sausville, M.D., Ph.D.
• Developmental Therapeutics Program
• National Cancer Institute

2
Goals Of Preclinical Drug Studies
Scientific framework

• Discovery of lead structures
• Refinement chemistry, pharmacology, efficacy
  Early development
• Late development formulation, dose form,
  toxicology

3
Goals Of Preclinical Drug Studies
Regulatory framework

• IND Investigational New Drug application
• approval by FDA to conduct human studies
• main criterion SAFETY AND LIKELY REVERSIBLE
  TOXICITY allows start of Phase I trials
  
• NDA New Drug Application
• basis for sale to public
• main criteria SAFETY AND SOME MEASURE OF
  EFFICACY result of Phase II/III trials

4
Cancer DrugsHow Do We Know We Have A Winner?
- PHASE III CLINICAL TRIAL WINNER
- PHASE II POTENTIAL WINNER
Time?
- PRECLINICAL MODEL (e.g., mouse or rat)
Cytotoxic
5
Cancer Drugs How To Pick A Winner?
VALIDATED CANCER TARGETS

• DNA
• Alkylators
• Antimetabolites
• Topo I / II
• Tubulin
• Receptors
• Nuclear
• Cell Surface (Immune?)
• Oncogene Proteins (2001 AT LAST!)

6
Empirical Drug Discovery
7
Problems With Empirical Models

• Lack of predictive power in vivo
• Poor correlation of non-human with human
  pharmacology
• Divorced from biology
• Inefficient many compounds screened
• developed, but have late clinical trials
  outcome
• at Phase III to define validation of compound
  action

8
KRN5500
Cell Membrane
9
Effect Of KRN5500 On Colo-205 Athymic Mouse
Xenografts
10
KRN5500 Plasma Concentrations On Effective
Schedule(20 mg/kg/d) In Mice
Plasma Concentration (?M)
Time (days)
11
Summary Of KRN5500 Phase I

• 26 patients as IV once per day over 5 days
• Dose limiting toxicity interstitial
  pneumonitis
• MTD 2.9 mg/M2/d x 5
• Achieve only 0.75 - 1 ?M at 3.7 mg/M2/d x 5
• 4/6 patients with gt25 incr Cmax have

grade 4 toxicity
Data of J. P. Eder, DFCI
12
In Vivo i.e., Intact Animal Tumor Models

• The information received depends on the
  question asked
• not all models are appropriate for all questions
• Drugs need different types of models at
  different times
• in their discovery / development life cycle
• Pharmacology models to qualify compound
• Efficacy models to define potential for
  biologic effect
• Target driven
• Target unselected
• Biology models to confirm ONLY the target
• or molecular events related to target are affected

13
In Vivo Activity vs Clinical Activity (39 Agents)
14
Rational Drug Discovery
PHARMACOLOGY (to affect target)
MOLECULAR TARGET SCREEN
Biochemical Engineered cell Animal
(yeast/worm/fish)
CHEMISTRY
15
How To Build A Cancer Cell
Cancer cells possess defined families of
lesions with common outcomes

• Misfiring or Abnormality of Cell Cycle
• Imbalance of Genes Regulating Cell Death
• Immortality / Telomerase
• Angiogenesis / Invasion Phenotype

- Oncogenes turned on - Suppressor genes
turned off - Mimic Growth Regulatory Signal
Transduction
G0
Tumor Size
(After Varmus, Bishop, Weinberg, Croce, Folkman,
Hanahan etc.)
16
Molecular Target Definition - How To?

• BIOLOGY
• RETROFIT ACTIVE MOLECULES
• CLASSICAL
• CHEMICAL GENETICS

Cytogenetics Breakpoints
Molecules (bcr-abl) Positive
selection from tumor DNA Active
oncogenes (signal transduction) Tumor
gene expression profiling (CGAP) Binding
partners (geldanamycin, rapamycin, fumagillin)
Computational algorithm (molecule
target) Cell metabolism / Cell cycle
effects Suggest single targets
Inefficient Libraries of molecules and
precisely defined organisms
- COMPARE - Cluster analysis
17
bcr-abl As Target Rationale

• Apparently pathogenetic in t9Q22 (Ph) CML/ALL
• Absence in normal tissues
• Modulate signal transduction events downstream

Maintenance of chronic phase Adjunct to bone
marrow transplantation
18
bcr-abl Fusion Protein
bcr
SH2
SH2
V
SH2/SH3
kinase
NT
DNA
Actin
bcr
autophosphorylation
Phosphorylation of other substances
McWhirter JR, EMBO 121533, 1993
19
Example Of Rational Approachbcr-abl directed
agents
Natural product empiric lead 1st
generation synthetic 2nd generation synthetic
in clinic
erbstatin
lavendustin
piceatannol
AG957
AG1112
CGP 57148B STI571
20
STI571An Oral In Vivo Bcr-abl Kinase Inhibitor
(days)
(hrs)
(days)
Tyr phosphorylation in vivo
Antitumor activity in vivo
le Coutre et al, JNCI 91163, 1999
21
Efficacy And Safety Of A Specific Inhibitor Of
The Bcr-abl Tyrosine Kinase In Chronic Myeloid
Leukemia
BRIAN J.DRUKER,M.D.,MOSHE TALPAZ,M.D.,DEBRA
J.RESTA,R.N.,BIN PENG,PH.D., ELISABETH
BUCHDUNGER,PH.D.,JOHN M.FORD,M.D.,NICHOLAS
B.LYDON,PH.D.,HAGOP KANTARJIAN,M.D., RENAUD
CAPDEVILLE,M.D.,SAYURI OHNO-JONES,B.S.,AND
CHARLES L.SAWYERS,M.D.
Ph Chromosome Cells
White Cell Count
in Metaphase
(cells x 10-3 / mm3)
Duration of Treatment with STI571 (Days)
NEJM 344 1031, 2001
22
NCI Drug Discovery
A Marriage Of Empirical And Rational
Opportunities

• Aids to find targets / link targets with drugs
• -Cancer Genome Anatomy Project (CGAP)
• -Developmental Therapeutics Program In Vitro Drug
  Screen
• Where the target is known build its assessment
  into the selection and development of a compound
  e.g., 17-allylamino 17-demethoxy geldanamycin
  (17-AAG)
• Where the target is unknown, define it or a
  proximal indicator of effect in parallel with
  conventional development path e.g., UCN-01
  protein kinase inhibitor

23
Cancer Genome Anatomy Project PROCESS
24
Gene Expression The Cells Fingerprint
Establishing for a cell the repertoire of genes
expressed, together with the amount of gene
products produced for each, yields a powerful
"fingerprint". Comparing the fingerprints of a
normal versus a cancer cell will highlight genes
that by their suspicious absence or presence
(such as Gene H ) deserve further scientific
scrutiny to determine whether such suspects play
a role in cancer, or can be exploited in a test
for early detection.
http//cgap.nci.nih.gov
25
NCI In Vitro Drug Screen
26
All Cell Lines
100
50
0
Percentage Growth
-50
-100
-7
-9
-8
-6
-5
-4
Log10 of Sample Concentration (Molar)
27
Pattern Recognition AlgorithmCOMPARE

• Goal COMPARE degree of similarity of a new
• compound to standard agents
• Calculate mean GI50, TGI or LC50
• Display behavior of particular cell line as
  deflection
• from mean
• Calculate Pearson correlation coefficient

1 identity 0 no correlation
28
Agents With Similar Mechanisms HaveSimilar Mean
Graphs
Leukemia NSCLC Small Cell Lung Colon CNS Me
lanoma Ovarian Renal
Taxol
Halichondrin B
Daunorubicin Topoisomerase II
Tubulin
29
Drug Target ClusteringsReveal Clues To Mechanism
Nature Genetics 24 236, 2000 http//dtp.nci.nih.
gov
30
Geldanamycin Structure
31
Benzoquinoid AnsamycinsInitial Cell Pharmacology

• Reduce levels or inhibit transformation by a
  large number of PTKs src, yes, fps, erbB1, lck
• e.g., 17AAG decrease erbB2 under conditions where
  overall transcription/translation little affected

150 100 50 0
MDA MB 453
p185 Protein p185 PY
Control
Hours
(Miller et al, Cancer Res 54 2724, 1994)
32
DTP, NCI In Vivo EvaluationOf Geldanamycin In
PC3 Prostate CA
Early Stage, Athymic Mouse Xenograft
Route of administration i.p.
Conclude Narrow therapeutic index on this
schedule Solubility of agent major problem for
other schedules
33
Geldanamycin Bead
34
Geldanamycin BeadsIdentify Hsp90 As Binding
Partner
1) Bead-Geld 2) Bead-Geld Geld
3) Bead-Geld Geldampicin 4) Bead
Neckers et al, PNAS 918324, 1994
35
Hsp 90
C.
A.
B.
36
Three Dimensional View Of Geldanamycin Binding
Pocket In Amino Terminus Of Hsp90
Stebbins et al, Cell 89239, 1997
37
17-AAG Binds To Hsp90 Shares Important Biologic
Activities With Geldanamycin
erbB2 ( of base line)
Raf-1 ( of base line)
dose (nM)
dose (nM)
Schulte Neckers, Cancer Chemother Pharmacol 42
273, 1998
38
UCN-01
IC50
Potent antiproliferative agent Cell cycle
arrest DNA-damage G2 checkpoint abrogation
37nM 300-600 nM 50nM
(DTP screen)
39
A DNA Damage G2 Checkpoint Is Mediated By CDKs
UCN-01 Action
40
UCN-01 Infusional Phase I TrialG2 Checkpoint
Abrogation
G2 checkpoint abrogation
G2 checkpoint abrogation
nM UCN-01
UCN-01 (mg/m2/day)
41
Challenges In Pursuing TheMolecular Therapeutics
Of Cancer

• Must change thinking from histologic to
  molecular
• diagnoses (CGAP, array technology)
• Develop new means (imaging, probes) to assess
• molecular pharmacodynamics
• Must move away from cytotoxicity as sole
  primary
• endpoint assess and evaluate cytostatic effect
• Promote patient participation in clinical
  trials
• Develop speed and efficiency in answering
  critical
• clinical questions

42
Goals For Cancer Drug ScreeningIn The New
Millennium

• Associate novel chemotypes with defined targets
• may utilize purified targets at the front end
• may define targets in pathway/organisms
• may retrofit molecules to targets or pathways
• by statistical approaches
• Allows facile tools for chemical/pharmacological
  
• optimization
• Define targets of relevance to and translatable
  in
• early clinical trials

43
SummaryDevelopmental Therapeutics Program, NCI

• Novel agents directed at molecular targets
• important to cancer pathogenesis
• and progression
• Interdisciplinary collaborators academia,
• industry, intramural NCI
• Contribute agents and regimens for use by
• intra / extramural investigators

44
Acknowledgements
NCI
V. Narayanan, R. Schultz J. Johnson, S. OBarr M.
Hollingshead, S. Stinson L. Rubinstein A.Monks,
N. Scudiero K. Paull, D. Zaharevitz, S. Bates S.
Holbeck, J. Weinstein A. Senderowicz A. Murgo, S.
Arbuck G. Kaur, P. Worland, Q. Wang P.
OConnor L. Neckers, L. Whitesell D. Newman
H. Piwnica-Worms Wash U V. Pollack
Pfizer M. Roberge U. Brit. Columbia
45
Our next speaker is Ms. Shannon Decker Office
of the Associate Director Developmental
Therapeutics Program