Title: Expanding the Research Domains of Rheumatoid Arthritis Clinical Databases: The Promise of Pharmacogenetics Jeffrey Greenberg, MD, MPH NYU Hospital for Joint Diseases New York University School of Medicine December 2006
1 Expanding the Research Domains of Rheumatoid
Arthritis Clinical Databases The Promise of
PharmacogeneticsJeffrey Greenberg, MD,
MPHNYU Hospital for Joint DiseasesNew York
University School of MedicineDecember 2006
2Outline
- Why Pharmacogenetics
- Case Study TNF Antagonists
- Issues of Study Design and Analysis
- Future Directions
3How Variable is the Human Genome?
Size 3 billion base pairs of DNA
Content 39,114 genes Variation 1 of
base pairs are polymorphic
(i.e. 30 million base pairs)
4Pharmacogenetics and Personalized Medicine
Marsh, S. and McLeod HL. Hum. Mol. Genet. 2006
15R89-93.
5Why Is Pharmacogenetics Important?
Let us imagine the despair of the mouse
experimentalist when we suggest that he or she
randomly allocate treatments to animals of
different genetic backgrounds and immunized with
different disease triggers. .This is
essentially how we treat RA in 2006
Klareskog, L. Nature Clinical Practice
Rheumatology (2006) 2, 517.
6Arthritis Clinical Databases and Pharmacogenetics
- 6 biologic agents have been FDA approved for RA.
- Only 1 RCT (Early RA Trial) has published
pharmacogenetic studies. - Arthritis clinical databases may represent the
ONLY practical approach to advancing the field of
pharmacogenetics. - As more biologic agents are approved, biomarkers
that can predict response will be increasingly
important.
7GENETIC POLYMORPHISMS
Pharmacokinetic
Pharmacodynamic
- Receptors
- Ion channels
- Enzymes
- Immune molecules
- Transporters
- Plasma protein binding
- Metabolism
8Potential Consequences of Different Drug
Metabolism and Drug Receptor Genotypes
Evans WE and Relling MV. Science 286487-491,
1999.
9Polymorphic Drug Metabolizing Enzymes
10Proof of Principle Warfarin and Cytochrome P450
Cyp2C9 Alleles
Dervieux T et al, Mutat Res 2005180-194.
11 Methylene Tetrahydrofolate Reductase (MTHFR)
Dervieux T et al, Mutat Res 2005180-194.
12Effect of the MTHFR C677T Polymorphism on
Methotrexate Toxicity in RA Patients
Urano W et al. Pharmacogenetics 2002 12(3)
183-190
13Pharmacogenetics of TNF Antagonists for the
Treatment of RACase Study
14How Can We Predict Response (or Failure)
ofMethotrexate or TNF Antagonists in RA Patients?
Responders
Source Klareskog L et al. Lancet 2004 363
675-81.
15Comparison of RA StudiesMTX TNF Antagonist in
Established Disease
Achieving ACR Response
Infliximab 10 mg/kg/q4w
Adalimumab 40 q2w
Etanercept 25 biw
Lipsky Weinblatt
Weinblatt NEJM 2000 A R 2003
NEJM1999
16Can We Predict Response to Abatacept, but not
anti-TNF? Abatacept in TNF Inadequate
Responders (ATTAIN)
Genovese et al. N Engl J Med. 20053531114.
17Can We Predict Risk of Serious Infection for RA
Patients Treated with TNF Antagonists?
Meta-analysis of Risk of Serious Infection from
Adalimumab and Infliximab RCTs
Bongartz, T. et al. JAMA 20062952275-2285.
18Effect of HLA-DRB1 on ACR 50 Response
Etanercept vs Methotrexate in the Early RA
Etanercept Trial
N151 in the Etanercept 25 mg arm OR (95 CI)
for effect of SE 4.3 (1.8 10.3)
Criswell LA et al. Arthritis Rheum. 2004 50(9)
2750-2756.
19HLA-DRB1 and Response to TNF Antagonists Summary
of Published Literature
20Effect of the TNF 308 G/A Polymorphism on
Clinical Response to Infliximab (n53)EULAR Good
Response (Decrease of DAS-28 by 1.2)
P0.0086
Mugnier B. Arthritis Rheum. 2003 48(7) 1849-1852
21TNF Polymorphisms and Response to TNF Antagonists
Summary of Published Literature
Combination of TNF and IL-10 SNPs. Extended
Haplotype that included TNF/LTA and
HLA-DRB1region
22Genetic Risk Factors are Stronger Predictors of
Developing Infections (UTIs) than Clinical Risk
Factors Etanercept vs Methotrexate in the Early
RA Etanercept Trial
Odds Ratio 95 CI
Age gt65 1.49 0.63 3.52
Elevated ESR 1.26 0.62 2.56
RF positive 1.32 0.43 4.06
Steroid use 1.15 0.58 2.27
ETA vs MTX 1.05 0.58 1.91
TNF 238 A 2.45 0.99 6.13
LTA 365 C 1.70 1.04 2.77
FcGR3a F 1.76 1.01 3.10
Hughes LB et al. Genes and Immunity 2004 5
641-647.
23Study Design and Analysis Issues in
Pharmacogenetics
24SNP Selection for Pharmacogenetics
- Functional SNP of candidate gene
- High density genotyping of coding and non-coding
SNPs of a specific candidate gene - SNP(s) of multiple genes in a metabolic pathway
- Whole genome scan
25High Density Genotyping Project across the Whole
Genome Insights into the Correlation Structure
of Alleles
26Insights from the HapMap Project and Related
Studies
27Allele Frequencies of Drug Metabolizing Enzymes
and Other Genes Vary across Different Population
Groups
CYP1A1
A. Bantu, Ethiopian, Afro-Caribbean B.
Norwegians, Ashkenazi Jews, Armenians C.
Chinese, New Guineans From Wilson, et al.
Nature Genet 29265-269, 2001
GSTM1
CYP2C19
DIA4
NAT2
CYP2D6
28(No Transcript)
29Can We Tag Candidate Inflammatory Gene SNPs
Hypothesized to Modulate TNF Antagonist Response?
30Future Directions
31Example of a diagnostic DNA microarray of
polymorphisms of candidate genes to predict
response and risk of toxicity (e.g.
chemotherapeutic agent)
Evans WE, Relling MV. Pharmacogenomics
Translating functional genomics into rational
therapeutics. Science 286487-491, 1999.
32Pharmacogenomic Studies May Be More Relevant to
Predicting Response for Pleiotropic Drugs such as
TNF Antagonists
Roden, D. M. et. al. Ann Intern Med
2006145749-757
33Gene Profiling in White Blood Cells Predicts
Infliximab Responsiveness in Rheumatoid Arthritis
- 33 RA patients treated with Infliximab
- Responders vs nonresponders (decrease of DAS-28 ?
1.2) - PBMC isolated from venous blood and total RNA
extracted - mRNA collected at baseline hybridized to a
microarray of 10,000 non-redundant cDNAs. - Real-time, quantitative reverse transcription PCR
of selected mRNA also performed. - Statistical analysis included t- test and SAM
(Significance Analysis of Microarrays) with a
false discovery rate of lt1.
Lequerre T et al. Arthritis Research and Therapy
2006 8 R105
34Gene Profiling in White Blood Cells Predicts
Infliximab Responsiveness in Rheumatoid
ArthritisResults
- Overall 16/33 (48) were EULAR responders.
- The 33 patients randomly divided intoa)
Training cohort (n13) b) Validation cohort
(n20) - 41 mRNAs were differentially expressed in
responders versus nonresponders s a function of
the response to treatment - Differentially expressed genes were confirmed by
qRT-PCR - 20 transcript set
- 8 transcript set
Lequerre T et al. Arthritis Research and Therapy
2006 8 R105
35Gene Profiling in White Blood Cells Predicts
Infliximab Responsiveness in Rheumatoid Arthritis
20 transcript set 8 transcript set
Sensitivity 90 80
Specificity 70 100
PPV 75 100
NPV 88 83
36The Future Incorporate Biomarkers into Clinical
Trials
Evans and Johnson, Ann Rev Genom Hum Genet 29-39,
2001
37Pharmacogenetics and Personalized Medicine
Marsh, S. and McLeod HL. Hum. Mol. Genet. 2006
15R89-93.