Pharmacogenomics in Drug Development

1
Pharmacogenomics in Drug Development
FDA Science Board April 9, 2003 Brian B. Spear,
Ph.D. Director, Pharmacogenomics Abbott
Laboratories
2
Presentation Overview

• Current application of pharmacogenomics in drug
  development
• Industry concerns regarding the conduct of
  pharmacogenomic studies and submission of
  pharmacogenomic data

3
Current Applications of Pharmacogenomics

• Primary uses relate to interpretation of clinical
  trial results, data quality, study design, and
  biomarkers.
• Targeting drugs at genetically-defined
  populations is not a primary focus in
  pharmaceutical development
• Three areas of greatest activity
• Clinical genotyping
• Pre-clinical gene expression
• Clinical gene expression

4
Clinical Pharmacogenetics

• Phase I studies
• Explain outliers or patient-to-patient
  variability in PK
• Exclude or include specific patients
• Normalize genotype frequencies
• Bridge to other populations

5
Example Desipramine PK Parameters
Genotyping can increase trial safety and explain
outlying data

• Drug interaction study
• CYP2D6 poor metabolizers (2 null alleles)
  excluded.
• One outlier with slow metabolism
• Outlier has 6 null allele and 9 allele with
  reduced enzymatic activity.
• Expected occurrence of null/9 genotype is 0.4

50
CYP2D6 6/9
40
t1/2, hr
30
20
10
Katz et al., Abbott Labs.
6
Clinical Pharmacogenetics

• Phase II/III studies
• Identify genetically-defined groups with more
  pronounced or rapidly progressing disease
• Exclude/include at-risk individuals
• Stratify studies based on genotypes
• Clinical response
• Risk of adverse events
• Where appropriate, develop drugs for specific
  groups
• Identify genetic markers associated with clinical
  outcomes

7
Example Gene Association using SNP Mapping
Association of anonymous SNP markers with
Alzheimers Disease
APOe
Courtesy Allen Roses, GSK
Similar methods can be used to identify genes
associated with drug effect or drug adverse
reactions
8
Pre-clinical Gene Expression

• Toxicogenomics
• Predict toxicity of candidate compounds
• Identify mechanisms of toxicity
• Identify potential biomarkers for toxicity or
  efficacy for future clinical studies

9
Example Toxicogenomics of Hepatotoxic Compounds
52 compounds rat liver mRNA analyzed by
microarray
Waring, et al., Abbott
10
Gene Expression Studies Can Produce Enormous
Amounts of Data
Page 2 of 28 Page Document
Statistical methods are being developed for
interpretation
Gene p-value
11
Clinical Gene Expression

• Biomarkers for drug response
• Biomarkers for drug-induced toxicity
• Comparison of human response to pre-clinical
  animal models
• Identify genes with variants that may define
  patient populations
• Identify proteins as potential biomarkers

12
Example Gene Expression in Clinical
TrialsResponse to Cyclosporin and rhIL11 in
Psoriasis

• Evaluation of gt7000 genes in microarray
• 159 found to associate with psoriasis
• 142 found to associate with improvement of
  psoriatic skin in response to therapeutic agents
• Gene expression reflects drug response

Non-Responder
Responder
Avg. PSI 9 8.7 5.8
4.1 5.6
Avg. PSI 9.5 9.5 8 9
9
1
10
Fold Change (lesion/treatment)
Fold Change (lesion/treatment)
1
0.1
0.1
0
1
4
8
12
0
1
4
8
12
Treatment Week
Treatment Week
Andrew J. Dorner Molecular Medicine,
Wyeth
Self-organizing map analysis of drug response for
psoriasis-related genes
13
The Challenge in High-Density Genomic Analyses

• Modern micro-array and whole genome analyses can
  generate tens of thousands of data points
• Analysis is dependent upon statistical methods
  which are themselves experimental
• No clear methods to determine validity of
  conclusions
• Results can be subject to multiple
  interpretations
• Genomics data and biological impact is
  incompletely understood

14
Submission of Data From Pharmacogenomic Studies

• Drug developers are hesitant to initiate
  high-density pharmacogenomic studies and
  reluctant to share data with regulators
• Analytical methods have not been developed to the
  point where valid conclusions can be drawn
• Data can be subjected to multiple statistical
  methods
• Reviewers might lack appropriate training or
  expertise
• Results may be mis- or over-interpreted
• Review may impact review timeline
• These may lead to unfavorable regulatory impact
  and jeopardize a drug development program

15
Commentary on Preliminary FDA Proposal on
Submission of Pharmacogenomic Data

• Favorable aspects
• Lowering of risk in conducting high-density
  pharmacogenomic studies
• Evaluation by qualified experts
• Consistent evaluation covering multiple compounds
• Pharmacogenomics review independent of medical
  review timeline
• Joint FDA/Industry effort to provide common basis
  for research exemption process

16
Commentary on Preliminary FDA Proposal on
Submission of Pharmacogenomic Data

• Uncertain aspects
• Definition of Pharmacogenomic Data
• Terms under which public health concerns would
  overrule research exemption
• Process for feedback from FDA to companies
• Unfavorable aspects
• Possible future rescinding of research exemption
• Potential requirement for additional studies for
  drug registration

17
Key Challenges in Pharmacogenomics
Data submission is only one of many issues facing
industry

• Complexity of biological responses genetics
  isnt everything
• Value of a pharmacogenomic study is often unknown
  until it has been completed
• No clear regulatory pathway for pharmaco-genomics
  (including assays)
• Financial constraints weigh against programs with
  uncertain outcomes

18
Unresolved Issues in Application of
Pharmacogenomics

• What are reasonable expectations of the role of
  genetics in drug responses?
• If a relationship is identified between a
  genotype and a response, will that lead to
  specific labeling requirements, even if the drug
  is safe and effective for the general population?
• Will collection of DNA in a clinical trial be a
  green light for the FDA to request
  pharmacogenetic studies?

19
Unresolved Issues in Application of
Pharmacogenomics

• Will the division of the patient population into
  multiple genetic subgroups lead to a request for
  larger studies to enable statistical power for
  each group?
• What will be the regulatory requirements for
  tests indicated on the drug label (IVD vs
  homebrew) and for tests used in genotyping for
  registrational studies?
• Under what conditions will it be possible to
  label a drug based on testing of only a
  pharmacogenetically defined patient group?

20
Conclusions

• Pharmacogenomics is becoming an integral part of
  drug discovery and development
• Excellent progress is being made in cooperative
  programs between industry and the FDA
• Clarity in the FDAs expectations of
  pharmacogenomics will encourage the use of these
  new technologies

21
(No Transcript)