Pharmacogenomics and Pharmacogenetics:Are We Ready for Personalized Medicine?David GurwitzSackler Faculty of Medicine, Tel-Aviv University GRIB, Barcelona, July 12, 2005
Hooked on Drugs: Mean lifetime drug consumption (UK) Prescription ~ 14,000 OTC ~ 30,000 Annual USA expenditures on prescription drugs: ~US$ 70 Billions “Cradle to Grave”at The British Museum
Drug safety is questioned.. Vioxx.. Celebrex.. Bextra.. What next??
1999 IOM Report (USA) Estimated cost of adverse drug reactions: US $15 billion/year in the USA alone • Estimated number of drug-related hospitalizations: • ~1,000,000 per year in the USA • ~50,000 - 100,000 toxicity-related death per year Kohn et al (eds): To Err is Human. Building a Safer Health System. Committee on Quality of Health Care in America. (1999) Institute of Medicine. National Academy Press: Washington, DC.
UK ADRs Study • 6.5% of new hospital admissions to internal medicine wards are directly related to ADRs!(1,225 admissions out of 18,820 during six months) • 4% of bed occupancy - directly due to ADRs • Annual UK cost: EUR 706 million (direct hospitalization costs - actual costs much higher!) • Women: 59% of ADRs (while only 52% of admissions) Pirmohamed et al (July 2004) Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ. 329:15-19.
All patients with same diagnosis Remove: (1) non-responders (2) toxic responders Treat: Responders and Patients Not Predisposed to Toxicity Which is more urgent? What should our priorities be?
“Primum non nocere” (Galen, 131-201 AD)The priority - reducing drug toxicity! The priorities for fitting personalized medicine should be like fitting a belt for your trousers: The first priority is ensuring that it does not harm you – Ensuring that it holds your trousers is second priority..
Improving drug efficacy Difficulties with ensuring better Drug Efficacy: • Tailoring the most effective drug and dosage for the individual patient is extremely complex. (exception: in oncology – some success) • Hundreds of polymorphic gene alleles are involved • Many non-genetic factors participate in drug efficacy:Gender; age; environment; diet; physical activity; smoking; family support etc.
Are we Ready for Personalized Medicine? Better Drug Efficacy? – Not Yet! We are not yet ready, with few exceptions: - Oncology (Herceptin, anti-EGFR) - Depression (5-HTT genotyping) - High blood pressure – Bidil (FDA approved June 2005)(ethnic considerations in drug prescription as an interim phase – ACE inhibitors ineffective in African-Americans, preference for NO donors & calcium channel inhibitors) • In most medical disciplines, we are not likely to be ready for Personalized Medicine in the clinic before 2025; maybe only ~2050? – but there will be exceptions..
Are we Ready for Personalized Medicine? Better Drug Safety? – YES, almost! • We are (almost) ready: it is possible with existing tools,resources, and knowledge! • Case in point: CYP2D6, CYP2C19, soon CYP2C9.. • Requires relatively less expenditures on new research and on new diagnostic tools
(Almost) Ready for Personalized Medicine –If we focus first on better drug safety Consideration #1: Society – almost ready.. - Growing public awareness on costs of drug toxicity (both economic costs and human suffering) - Growing awareness of genetic factors affecting health for complex diseases (not just for “genetic disorders”) - Worries about the “graying of society” and the associated soaring health care costs in the EU and USA
(Almost) Ready for Personalized Medicine Consideration #2: Government – almost ready Drug regulatory agencies are (almost) ready: USA: In March 2005, the FDA has issued the “Guidance for Industry: Pharmacogenomic Data Submissions” http://www.fda.gov/cber/gdlns/pharmdtasub.htm Europe: In February 2005, WHO/CIOMS Working Group on Pharmacogenetics has issued the report, “Pharmacogenetics Towards improving treatment with medicines” (Council for International Organizations of Medical Sciences)
(Almost) Ready for Personalized Medicine –If we focus first on better drug safety Consideration #3: Pharmaceutical Industry • Pharma, which traditionally objected to PGx as a threat to market size and profits, begin to see the advantages: - Lower costs for clinical trials (less participants) - New “orphaned markets” - Can charge more for safer drugs - Marketing a diagnostic test along with the drug can bring added revenues
Drug Development:Typically ~10 years, ~$1.7 billion CompoundsTime (years) 10,000 _____ Discovery (2–6 y) 250 ______ Preclinical (2-4 y) 5 ___ Phase I (~1 y) (~100 healthy volunteers) _____ Phase II (~2 y) (~300 patients) ______ Phase III (~3 y) (~3000 patients – save here – less patients !) FDA review and approval _____ (~2 y) Note: patents are granted for 25 y
(Almost) Ready for Personalized Medicine –If we focus first on better drug safety Consideration #4: Diagnostic Tools • Diagnostic tools focusing on better drug safety have been approved by the FDA and the EC and are entering the market: • Roche Diagnostics “AmpliChip P450”: - approved by the EC in September 2004, and by the FDA in December 2004
The Next Diagnostic Chips? Additional diagnostics are needed: General: CYP2C9; CYP3A5; CYP2B6; MDR-1; UDP Glucuronosyltransferases (UGTs); N-acetyltransferases (NATs) Oncology: thiopurine methyltransferase (TMPT); Thymidilate synthase; dihydropyrimidine dehydrogenase (for 5-FU dosing)
The Human Genome Variation Project • Identify all major polymorphic gene alleles • Characterize polymorphic alleles frequencies in different ethnic groups • Identify phenotypic effects of polymorphic alleles • Proposed in 1991 (by Luca Cavalli-Sforza et al) • Never approved • HapMap Project: started on October 2002
HapMap Project: Caucasians; Han Chinese; Japanese; Yoruba
Single Nucleotide Polymorphisms • >10 million SNPs are known (number depends on how we define a SNP) • About 120,000 estimated SNPs are within exons • Mean of 4 estimated exonic SNPs per gene • SNPs freely accessible on http://snp.cshl.org/
A Need for Data Sharing • Data about relationships between drug target haplotypes and drug response phenotype would be needed from many thousands of patients • This is beyond the capacity of academic research teams • The only likely solution: • - consortia for collection and sharing of PGx data • - give incentives for data sharing by Pharma (tax breaks? accelerated drug approval? ) - build upon the model of the PharmGKB
Current Practice: “Trial and Error”- What about Safety? Toxicity No Effect Oops! Too Much Too Little Decrease dose Increase dose Favorable Effect No effect Toxicity Increase dose or change drug Decrease dose or change drug
CYP450 Content in Human Liver Low levels of P4502D6 & P4502C19
CYP2D6 • Minor P450 enzyme in human liver • Metabolizes ~30% of drugs • Low enzyme activity “poor metabolizers” higher risk of drug toxicity and drug interactions
CYP2D6 Substrates Debrisoquine Amphetamine Dexfenfluramine Ouanoxan Ondansetron Antiarrhythmics Encainide Flecainide S-mexillitene Lidocaine Antipsychotics Perphenazine Thioridazine Haloperidol Risperidone Minaprine Venlafaxine Beta Blockers Propafenone S-metoprolol Propranolol Timolol Alprenolol Bufuralol Carvedilol Antidepressants Fluoxetine Fluvoxamine Paroxetine Amitriptylline Desipramine Clomipramine Imipramine Analgesics Dextromethorphan Codeine Tramadol
CYP2D6 Poor Metabolizers Caucasians 5% -10% African-Americans 6% Africans 2% – 19% Japanese 0.5% Chinese 0.7% In poor metabolizers: - Higher risks for toxic reactions - Higher risks for drug interactions (e.g., codeine and propranolol)
CYP2C19 Substrates Omperazole Lansoprazole Pantoprazole S-mephenytoin Hexobarbital R-mephobarbital Phenytoin Diazepam Citalopram Warfarin Proguanil Teniposide Nilutamide Indomethacin Moclobemide Propranolol (in part) Imipramine (in part) Clomipramine (in part) Amitryptylline (in part)
CYP2C19 “Poor Metabolizers” Caucasians 2% - 3% Africans 3% - 5% Chinese ~14% Koreans ~14% Japanese ~20%
Acetaminophen Metabolism: Identifying “UGTs poor metabolizers” could save lives Glucuronidation (60%) Sulfation (35%) CYP2E1(5%) Toxic!
Personalized Psychiatry? Psychiatry is unique among medical disciplines: - Poor understanding of disease biology - Lack of objective diagnostic and follow-up tools - Drugs work very slowly (weeks – months): • Huge morbidity and economic cost while trying to optimize pharmacotherapy • Large potential for PGx Gurwitz & Weizman (2004) Pharmacogenomics 5:1-5
Personalized Psychiatry Notable example: - The SSRIs antidepressants:Selective Serotonin Reuptake Inhibitors (fluoxetine, fluvoxamine, paroxetine, etc.) Only ~60% of patients respond favorably; this is unrelated to blood SSRI concentrations
SSRIs: Next Best Money Makers for Pharma.. Prozac: US annual sales = $1,577M Zoloft Luvox: $126M All: metabolized by CYP2D6 Paxil, Seroxat: US annual sales = $2,468M
Depression The most prevalent affective disorder: 10%-20% of the population (lifetime) • Decreased raphe serotonin • Excessive activity of the serotonin transporter??
Drug Efficacy and 5-HTT Polymorphism L = 44 bp insertion Gurwitz D (2000). Clinical Genetics 57, 247-249
SSRI Efficacy HAM-D ratingL/LL/SS/S N 16 26 16 Week 0 25.1 28.4 27.7 Week 1 21.5 24.4 24.8 Week 2 16.6 20.0 23.4 Week 3 11.0 15.3 22.2 Week 4 7.4 10.3 20.6 (p = 0.030) (paroxetine = 40 mg/day) Zanardi et al (2000) J Clin Psychopharmacol. 20, 105-107
Human Lymphocytes:Saturable Binding of[3H]Imipramine Binding isotherm Scatchard analysis
Human Lymphocytes:Functional Uptake of[3H]5-HT 5-HT uptake velocity Lineweaver-Burk plot
Human Lymphocytes:Expression of 5-HTT 1,2 = human lymphocytes 3 = human lymphoblasts 4,5 = rat striatum • What is the role of 5-HT in immune regulation? • What is the role of the 5-HTT in human lymphcoytes? • Is the 5-HTT plymorhism implicated in response to drugs affecting the immune system?Example: TCA as therapy for neuropathic pain?
Human Choline Transporter I89V • hCT-1 with I39V shows: • - Similar gene expression levels • - Similar level of membrane insertion - 40% less [3H]choline uptake
Human Choline Transporter I89V(expressed in transfected COS-7 cells) Similar [3H]HC-3 binding 40% less [3H]choline uptake
Human Choline Transporter I89V (expressed in transfected COS-7 cells) Could there be an effect on drug efficacy in AD? PD? Other CNS disorders? Similar cellular distribution Similar Na+ dependence
NLGIP study: CYP2D6 & CYP2C19 Luo et al (2004) American Journal of Pharmacogenomics 4: 395-401