Transcript Methotrexate study - UCSF Helen Diller Family

Personalizing Medicines
Tumor Genomics
E.g., Trastuzumab (Herceptin)
Germline DNA
Tamoxifen: Used to treat estrogen-positive
breast cancer
Tamoxifen:is Is
a Pro-drug
Tamoxifen
a Pro-Drug
Jin Y et al: J Natl Cancer Inst 97:30, 2005
CP1229323-3
CYP2D6 Has Reduced Function Alleles
CYP2D6
Polymorphisms Endoxifen
*4 Allele
Tamoxifen
Tamoxifen
CYP3A4
Tamoxifen
Tamoxifen
Inactive Prodrug
CYP2D6
Endoxifen
Endoxifen
CYP2D6
Polymorphisms
*10 Allele
Endoxifen
In Japanese
Active Drug
Individuals with CYP2D6*4 Have Lower Levels
of Endoxifen
Plasma Endoxifen
(nM)
P<0.001, r2=0.24
CYP2D6 *4/*4
Polymorphisms
CYP2D6*4 (most common genetic variant associated with
the CYP2D6 poor metabolizer state)
Jin Y et al: J Natl Cancer Inst 97:30, 2005
Individuals with CYP2D6*4 Have Have Poorer
Relapse Free Survival
%
With Relapse
Free Survival
CYP2D6 WT/WT
CYP2D6 *4/WT
CYP2D6 *4/*4
P=0.020
CYP2D6 *4/*4
Polymorphisms
Years after randomization
Goetz et al J Clin Oncol. 2005;23(36):9312-8.
Women with CYP2D6 Polymorphisms Have
Poorer Response to Tamoxifen
N = 1325
CYP2D6
Polymorphisms
*4, and other
Schroth, W. et al. JAMA 2009;302:1429-1436.
Japanese Women with CYP2D6 Polymorphism,
Have Poorer Response to Tamoxifen
Cancer Science, May 2008, page 995-99
Plasma
Endoxifen
(nM)
Japanese Women with CYP2D6 Polymorphism,
Have Poorer Response to Tamoxifen
Wt/Wt,
Venlafaxine Sertraline
no
inhibitor
Jin Y et al: J Natl Cancer Inst 97:30, 2005
Paroxetine
*4/*4, no
inhibitor
NIH Pharmacogenomics Research Network
EUROPE
Research
Groups
CANADA
Collaborating
Sites
NWAP
Network
Resources
UW-EXOME
PHAT
PPII
PHONT
PARC
PHRAT
PAAR
XGEN
PAPI
PMT
WU-NGS
PAT
PAAR4KIDS
PGBD
PNAT
P-STAR
PG-POP
GAP-J
BCM-HGSC
TAIWAN
JAPAN
PEAR
PGRN-CGM
ALASKA
Hawaii
Cancer Pharmacogenomics
Riken- Center for Genomic Medicine:
NIH Pharmacogenomics Research Network:
To identify genetic predictors of
therapeutic and adverse drug response
using genomewide approaches.
Personalized Medicines
Picture from CGM Website
Strategies in Genomic Studies
Disease Genetic Studies
Patients
Identify Mutant/
Polymorphic Genes
Biological/
Pharmacologic
Mechanism
Genomewide
Markers
Diabetics
Controls
Genomewide Association of Seven Common Diseases
Bipolar
Coronary Artery
Disease
Crohn’s
Hypertension
P value
-log10
Rheumatoid
Arthritis
P=10-15
Wellcome Trust,
Nature, 2007
Chromosome Number
Type 1
Diabetes
Type 2
Diabetes
January of 2009, > 200 Genomewide Association
Studies Identifying Disease Risk Alleles
Number of Studies
60
Fewer than 20 Pharmacogenomic Studies
40
Drug Response/
Toxicity
20
0
Cardiovascular/
Aging/
Drug
Inflammatory
Physiologic/
Diabetes/
Psychiatric/
Response/
Cancer
Disease
Biochem Trait
Obesity
Other
Toxicity
http://www.genome.gov/gwastudies/
Strategies in Genomic Studies
Pharmacogenomic Studies
Patients
Identify Mutant/
Polymorphic Genes
Biological/
Pharmacologic
Mechanism
Genomewide
Markers
Receive Drug
Adverse Reaction
Receive Drug
No
Adverse Reaction
CALGB 40101 – 2 X 2 Factorial Design
Adjuvant Therapy for Women with Breast
Cancer with 0-3 Positive Nodes
Stratification
AC q 2 wk
60 mg/m2
600 mg/m2
Pre-Post
Menopausal
ER/PgR
4 cycles – 8 wk
6 cycles – 12 wk
4 cycles – 8 wk
Paclitaxel q 2 wk
175 mg/m2
6 cycles – 12 wk
AC = doxorubicin/cyclophosphamide
Target Accrual 4646 pts
Pharmacogenetic analysis of predictors of
paclitaxel and AC toxicity
Paclitaxel Candidate Gene Studies are
Inconclusive
Genes
ABCB1
ABCC1
ABCC2
Endpoints
Findings
ABCG2
Pharmacokinetics
Small Populations
CYP1B1
Toxicities
CYP2C8
Outcomes (DFS, OS)
Inconsistent
results
CYP3A4
CYP3A5
MAPT
TP53
SLCO1B3
Riken Center for Genomic Medicine and PGRN
Collaboration for Genome Wide Association
Study of CALGB 40101
• Genome wide analysis of CALGB 40101 samples
using Illumina 610-Quad platform
• Doxorubicin/Cyclophosphamide Arm (n=919)
• Paclitaxel Arm (n=1040)
Study Design
1040 Subjects
Illumina 610-Quad
(≈592K SNPs)
QC
(Subject CR >99%, SNP
CR >95%; Chr 1-23,
MAF > 0.5%, IBD)
1029 Subjects/554,450 SNPs
‘Genetic Caucasian’
PC1,PC2,PC3 all within 2
SDs of mean values
from all patients
self-declaring as ‘White’
Principal Components Analysis
859 Genetically Western Europeans
Time to Event
Analysis
Ordinal
Regression
Replication Study
Logistic
Regression
170 Remaining Subjects
Regression,
Admix Mapping
Principal Component Analysis Identifies 859
Genetically Similar Subjects of Western European
Descent
Neuropathy Grading Criteria
Activities of Daily Living (ADL)
Instrumental ADL: Preparing meals, shopping for groceries or clothes,
using the telephone, managing money, etc.
Self care ADL: Bathing, dressing and undressing, feeding self, using the toilet,
taking medications, and not bedridden.
Paclitaxel Sensory Peripheral Neuropathy
in ‘Genetic Europeans’
Grade 3
59
Grade 2
147
Omitted
4
Grade 0
280
Grade 1
369
Maximum grade sensory peripheral neuropathy during
treatment or follow-up: 859 subjects
Analysis Strategy
 Ordinal Logistic Regression
Highest grade sensory peripheral neuropathy
Additive genetic model
Covariate: Log cumulative dose/BSA at first instance of
maximum grade toxicity. If no toxicity, equal to
cumulative dose at end of treatment.
 Time to Event Analysis
“Event” - The first occurrence of a grade 2 or higher
sensory peripheral neuropathy.
“Time” - Cumulative dose/BSA at event. If no toxicity,
equal to cumulative dose at end of treatment.
Additive genetic model
Time to Event Analysis
EPHA5
FGD4
PITPNA
GRIP1
Ordinal Regression
FZD3
Probability of NOT having an Event
Probability of NOT having an Event
Time to Event Analysis
EPHA5
GRIP1
Ephrin Signaling is Important in Axon Guidance
EphrinB2 Levels Increase in the Dorsal
Root Ganglion in Response to Injury
Kobayashi et al. Spine 32:1592-1598, 2007
Probability of NOT having an Event
Probability of NOT having an Event
Time to Event Analysis
NDRG1
FGD4
FGD4 is Implicated in Congenital
Neuropathic Disease
Am J Hum Genet 81:158-164, 2007
Demyelination in Patients with FGD4
Mutations
Sural nerve
biopsy
M298R
E543fs
Stendel et al. Am J Human Genet 81:158-164, 2007
Ordinal Regression
FZD3
The Wnt Receptor FZD3 is Critical for
Neurite Outgrowth
Endo et al. Mol Cell Biol 28:2368-2379, 2008
Gene Ontology Analysis is Being
Applied to GWA Studies
AJHG 2007
AJHG 2009
Exploratory Gene Association
Networks (EGAN)

Developed by Jesse Paquette at the Biostatistics and
Computational Biology Core (BCB) at the Helen Diller Cancer
Center

Tool for visualizing gene-gene networks and pathway
enrichment using several publicly available databases

http://akt.ucsf.edu/EGAN/

Paquette and Tokuuasu Bioinformatics 26:285-6, 2010
EGAN: CALGB 40101 Neuropathy GWA Data
Used top 2500 candidate SNPs (based on P-value)
Illumina’s annotation used for gene assignments
SNP with smallest P-value used for each gene
Taking into account genes with multiple hits and SNPs
without gene annotation:
 1154 SNPs were analyzed
“Background” Gene Space
 Genes represented on the Illumina 610quad considered
as all “potential” hits (i.e. the denominator in
calculations of pathway “enrichment”)
SNP Hits are Enriched in Neuronal
Development and Cell Adhesion Pathways
GO Process
All Genes
40101 “Hits”
Enrichment
Nervous System
Development
787
137
1.5 x 10-12
Cell Adhesion
715
126
5.1 x 10-12
Central Nervous
System Development
299
64
4.1 x 10-10
Cell Development
562
95
1.9 x 10-8
Hemophilic Cell
Adhesion
124
33
3.3 x 10-8
Cell-Cell Adhesion
259
53
6.3 x 10-8
Neurogenesis
372
68
1.0 x 10-7
Neuron Projection
Development
176
40
1.4 x 10-7
Generation of Neurons
347
64
1.7 x 10-7
Cell part Morphogenesis
168
38
3.3 x 10-7
Synaptic Transmission
264
51
7.1 x 10-7
Neuron Differentiation
309
57
7.9 x 10-7
Visualization of Protein-Protein Interactions
Replication Strategies

BioBank Japan – no dosing data

Remaining 40101 samples

Scottish Ovarian Cancer Trial: 454 women
receiving 175 mg/m2 paclitaxel plus
carboplatin

SWOG 0221: Phase III trial of AC + G vs. Q 2
wks AC followed by paclitaxel weekly or Q 2
wks
SNPs for Replication
Analyses
Time to Event
Ordinal
Total
Number SNPs
Cutoff: <1 x 10-6
Number SNPs
Cutoff: <1 x 10-5
8
48
4 (*0)
28 (*4)
12
72
(*Number of Shared SNPs)
Additional Phenotypes

Ovarian function

Breast cancer outcomes
Kathy Giacomini
Pharmacogenomics:
Focus on breast
cancer
Deanna Kroetz
Joan Venticinque