Designer Drugs in thePost-Genomic Era

Thomas Ferrin, Ph.D.

Resource for Biocomputing, Visualization,and Informatics

University of California, San Francisco

Pittsburgh Supercomputer Center May 13, 2003

Resource for Biocomputing,Visualization, and Informatics

The RBVI is a NIH/NCRR BiomedicalTechnology Research Center

We create innovative computational andvisualization-based data analysis methods andalgorithms, turns these into easy-to-usesoftware tools, and apply these tools forsolving a wide range of genomic and molecularrecognition problems within the complexsequence structure function triad

Application areas

Gene characterization and interpretation

Drug design

Variation in drug response due to genetic factors

Protein engineering

Biomaterials design

Bioremediation

Prediction of protein function from sequence andstructure

Sample RVBI projects

•New methods for large-scale data collection,storage, analysis, and presentation forpolymorphism (SNP) genotyping project

•Extensible visualization tools for comparativestudies of protein sequence, structure, andfunction

Fortune - Oct 30, 2000

Case Report #1: Michael Adams-Conroy

Young child born to abusive mother, adoptedat age 3, with signs of fetal alcoholsyndrome, obsessive-compulsive disorder,Tourette’s syndrome, and attention-deficithyperactivity disorder. Prescribed Prozac tohelp control emotional outbursts.

Child dies suddenly; toxicology tests showmassive overdose of Prozac. Adoptiveparents investigated for homicide and theirother two children put into protectivecustody.

Michael Adams-Conroy (continued)

Sharp-eyed psychiatrist notices unusuallyhigh levels of other metabolites in toxicologyreport, indicating child may have had anenzyme deficiency inhibiting Prozac frombeing metabolized normally.

Subsequent genetic testing showed child haddefect in 2D6 gene which resulted inabnormal liver enzyme that metabolizesantidepressants.

Adoptive parents exonerated.

Case Report #2

Patient: 3-year old boy

Diagnosis: Acute LymphoblasticLeukemia (ALL)

Standard therapy: 6-mercaptopurine(6-MP)

Result: Adverse Drug Reactionleading to acute bone marrowsuppression

Normal Mechanism of Action

6-MP

6-METHYLMERCAPTOPURINE

CH3S

S

TPMT

6-THIOGUANINE NTs

TPMT

TPMTX

THIOPURINE METHYLTRANSFERASE (TPMT) GENESARE DEFECTIVE IN 1:300 PEOPLE

This leads to elevated levels ofThioguanine Nucleotides

6-MP

6-METHYLMERCAPTOPURINE

CH3S

S

TPMT

6-THIOGUANINE NTsX ↑ ↑ ↑

ADVERSE DRUGREACTION (ADR)

THERAPEUTICRESPONSE

NO RESPONSE

PEOPLE DIFFER IN THEIR RESPONSE TO DRUGS

TESTING FOR TPMT GENES IS NOW AVAILABLE

CHILDREN WITH DEFECTIVE TPMT GENESSHOULD RECEIVE A LOWER DOSE OF 6-MP

Drug (Brand Name) Perscribed For... Adverse Reaction Gene at Cause

Imipramine (Tofrannil) Depression, ATD Heartbeat irregularity CYP2D6

Isoniazid (Laniazid) Tuberculosis Liver toxicity NAT2

Warfarin (Coumadin) Preventation of blood clots Internal bleeding CYP2C9

5-fluorouracil (Adrucil) Cancer Severe immune suppression DPD

Clarithromycin (Biaxin) Antibiotic Heartbeat irregularity KCNE2

Azathioprine (Imuran) Rheumatoid arthritis Severe immune suppression TPMT

Adverse Drug Reactions

ADRs may kill 30,000 - 40,000 Americans eachyear and cause 2,200,000 serious nonfatalreactions. JAMA 1998 June 3;279(21):1684

Drugs with known genetically-linked potential forfatal adverse reactions (partial list):

Pharmacogenetics and Pharmacogenomics:Impact on Drug Discovery and Development

Time 157(2) Jan.15, 2001

Genetics vs. Genomics

PHARMACOGENETICS: Genetic basis fordifferences in drug response

PHARMACOGENOMICS: The use of geneticsand genomics in drug discovery and development

Age

Diet

Gender

Disease

ORGANISM

Exercise

Alcohol andDrug Use

OccupationalExposures

Smoking

Genetics

RESPONSE

Environmental vs. GeneticInfluence on Drug Response

Toxic

Ineffective

TherapeuticDrugBloodLevel

Time

Absorption Distribution

ExcretionMetabolism

Target Cell

ReceptorsSignaling Proteins

Determinants of Human Drug Response

Human Genetic Variation

Sequence differences between an individual’smaternal and paternal genomes occur on averageabout every 500 to 2000 bp.

Most of these are SNPs. Deletions and insertionsare less common.

Estimated 240,000-400,000 common cSNPs.

Typical person estimated to be heterozygous for24,000-40,000 non-synonymous SNPs.

Human Genetic Variation

106 genes relevant to cardiovascular disease,endocrinology and neuropsychiatry

560 single nucleotide polymorphisms (SNPs) in114 chromosomes• 392 cSNPs• 47% of cSNPs lead to amino acid changes

Cargill et al. Nature Genet. 22:231, 1999

Halushka et al. Nature Genet. 22:239, 1999

Human Genetic Variation

75 candidate genes for blood pressurehomeostasis and hypertension

874 SNPs in 148 chromosomes• 387 cSNPs• 54% of cSNPs lead to amino acid changes

Pharmacogenetics of Membrane Transporters

$12-million, 4-year NIH grant• Kathleen Giacomini and Ira Herskowitz, co-PIs, plus ~20other UCSF researchers

Major Project Goal:• Understand the genetic basis for variation in response todrugs which interact with membrane transporters

- This class of proteins is of great pharmacologicalimportance, as it provides the target for about 30% ofthe most commonly used prescription drugs and is amajor determinant of the absorption, distribution andelimination of many others.

PMT project goals - continued

•Determine the amount of genetic variation (single-nucleotidepolymorphisms) in at least 40 transporter genes by examiningthe DNA from an ethnically diverse sample of 250 people.

•Test the performance of these transporter variants in cellcultures and determine, through clinical phenotype studies, ifpeople with those variants respond differently to drugs in aclinically significant way.

•Provide access to the data from these studies to the generalscientific community through the World Wide Web tofacilitate collaborative research and to speed development ofnew drug treatments.

The Corriel Cell Collection

African American (AA) - 100

Caucasian (CA) - 100

Asian American (AS) - 30

Mexican American (ME) - 10

Pacific Islander (PA) - 7

TOTAL - 247

PMT Intranet Website

Used by ~100 researchers at UCSF

Effective data analysis and display driven byiterative design/refinement cycle, successfulbecause the bioinformatics team works closelywith the molecular biologists• Jill Mesirov, Whitehead: “Bioinformatics needs to betightly integrated with the scientific research, not aservice function”

Flexibility key!• Multiple ways to display same data• Simple download mechanism for scientists who want toload raw data into Excel spreadsheets

PMT Scientist-Users Are aDemanding Bunch…

Web Demo

You’ve now seen DNAand AA sequences

What about structure?

Why is Structure Important?

Sequence Structure Function

• Current research areas:- Prediction of structure from sequence- Prediction of function from sequence and structure- Understanding evolutionary changes- Engineering proteins for specialized function

• Applications in pharmacogenomics …- Improvements in drug discovery and development process- Prediction of drug response- Avoidance of toxic side effects

Growth in Protein Structures

The Structural Genomics Initiatives

“The next step beyond the human genome project”

$150 million in NIH grants to establish 9 U.S. centers• Goals:

- Speed the determination of three-dimensional atomic-scalemaps of proteins

- 35,000 structures by 2005- Identify all proteins expressed in an organism - “proteomics”

See http://www.nigms.nih.gov/funding/psi.html for additional information

Center Lead Institution TargetNY Struct. Genomics Res. Consortium Rockefeller Univ. Bacteria/yeast/humanNortheast Struct. Gen. Consort. Rutgers Univ. Roundworm/fly/humanSoutheast Collab. for Struct. Gen. Univ. of Georgia Bacteria/roundworm/humanBerkeley Struct. Genomic Center Lawrence Berkeley Lab. BacteriaJoint Ctr. for Struc. Genomics Scripps Research Inst. Roundworm/humanTB Struct. Genomics Consortium Los Alamos Nat. Lab. TuberculosisMidwest Ctr. for Struct. Genomics Argonne National Lab. Archaea/bacteria/eukaryaCtr. for Eukaryotic Struct. Genomics Univ. of Wisconsin Arabidopsys thalianaStruct. Gen. of Pathogenic Protozoa Univ. of Washington Protozoans

The X-ray Crystallography Pipeline

Target Selection

“Crystallomics”• Isolation• Expression• Purification• Crystallization

Data Collection

Structure Solution

Structure Refinement

Functional Annotation

Publish

Anticipated BioinformaticsDevelopments:• Determination of distanthomologs and domain recognition

• Automation of proceduresthrough empirical rules

• Software integration fordecision support and dataanalysis

• Automated annotation throughsequence and structuralalignments, prediction ofprotein-protein and protein-ligand interactions, and motifrecognition

The Same Protein?

Visualizing 3D Structure:The Chimera Molecular Modeling System

Chimera is an extensible interactive 3-Dmodeling system designed to allow developersto quickly incorporate novel algorithms andanalysis tools• ~30 extensions written to date• Extensions are written in the Python programminglanguage- Easy to learn, even for novice programmers- Offers object-oriented language features

• Extensions can control standard user interfacefeatures (e.g. camera, help, menus, toolbar) as wellas their own custom interfaces

Sample Chimera Extension

Volume Viewer• an extension for visualizing three-dimensional (3D)numerical data sets

Electrostatic potential(surfaces)

Electron density(mesh)

Electrostatic potential(solids)

Sample Chimera Extension

ViewDock• rapid screening of promising drug candidates foundwith the UCSF DOCK program

Sample Chimera Extension

Multalign Viewer• simultaneous display of protein sequence andstructure

Sample Chimera Extension

Collaboratory• supports collaborative studies of molecular structureamong scientists at multiple remote locations

Summary

We are in the midst of a profound and exciting new era inbioinformatics and computational biology

The data made available by the various genome andstructural genomics projects will occupy researchers fordecades to come

High performance computing and the internet play a criticalrole in the navigation, analysis, and dissemination of thisdata and the resulting scientific knowledge

The potential impact on drug development and treatment ofhuman disease is enormous

Acknowledgements

Collaborators & Staff• Dr. Conrad Huang, Dr. ElaineMeng, Prof. Patricia Babbitt,Prof. Kathy Giacomini, GregCouch, Eric Pettersen, AlConde, Tom Goddard, SusanJohns, Doug Stryke, MichikoKawamoto

NIH National Center forResearch Resources• P41-RR01081

National Institute ofGeneral Medical Sciences• GM61390

Additional information

RBVI:www.rbvi.ucsf.edu

PMT project:www.pharmacogenetics.ucsf.edu

Chimera:www.cgl.ucsf.edu/chimera