dr. margarete fischer-bosch-institut für klinische pharmakologie … · 2018-08-02 · dr....
TRANSCRIPT
Dr. Margarete Fischer-Bosch-Institut für Klinische Pharmakologie
Robert Bosch StiftungStuttgart
Pharmacogenetics today:
What we don‘t know
STUTTGART
IKP
Pharmacogenomics and Drug Therapy
ME3015.ppt
Selection of appropriate drug and dose for the individual patient in order to
• achieve optimal therapeutic response
• avoid therapeutic failure
• minimize side effects and toxicity
ME3290.ppt
Pharmacogenetics and Drug Therapy
Candidate genes associated with response and side effects/toxicity are known
Value of pharmacogenetic test in predicting response and selection of appropriate and dose has been established in prospective clinical trials in two independent study cohorts
Ideal:
ME3289.ppt
Pharmacogenetics and Drug Therapy
Reality: Focus on one gene, limited number of mutations tested
Predictions based on case reports only
Retrospective studies with poor descriptionof patient characteristics, clinical outcome & confounders
ME2849.ppt
Prerequisites for Pharmacogenetic Testing
• Clear definition of phenotype (confounders)
• Genotype-phenotype relationship
• Sufficient sample size to identify all relevant
mutations associated with phenotype
• Association studies: Plausible biological hypothesis
ME3292.ppt
Pharmacogenetics and Drug Therapy
How is the phenotype drug concentration
related to drug response, side effects?
How predictive is the drug effect for efficacy,
clinical endpoint?
Pitfalls in Pharmacogenetic Testing
ME2838.ppt
• Of the known functionally important mutations only a limited number are tested
• Presence of unknown mutations
• Penetrance of gene
• Phenotype studied is only in part caused by candidate gene
• Other genetic and nongenetic factors contribute to phenotype
Schaeffeler et al. 2003, 2004, 2005
TPMT*11
TPMT*12
TPMT*13
TPMT*14
TPMT*15
84A>T
1A>G
-1G>A
TPMT*16488G>A
TPMT*17124C>G
374C>T
TPMT*18211G>A
TPMT*19365A>C
TPMT*20712A>G
TPMT*21205C>G
395G>A
I II III IV V VI VII VIII IX X
TPMT*2
TPMT*3A
TPMT*3B
TPMT*3C
238G>C
460G>A
460G>A
719A>G
719A>G
TPMT*3D
TPMT*4
TPMT*5
TPMT*6
TPMT*7
TPMT*8
460G>A 719A>G292G>T
146T>C
539A>T
681T>G
644G>A
A G
TPMT*9356A>C
TPMT*10430G>C
TPMT*22488A>C
ME3277
ME3291.ppt
Requirements for Pharmacogenetic Testing
How many mutations should be tested?
Restricted to most common mutations or as complete
as possible?
Example of TPMT: 22 loss of function mutations identified
Cohort of ~ 15000 patients: 10 new mutations identified.
1 patient with complete deficiency was classified
as heterozygous based on genotyping which was restricted
to 3 most common SNPs
Polygenic Nature of Drug Response: Antidepressants
ME3045
BBB BBBconc
.
conc
.
time [h] time [h]
MAO
HT Receptors
5-HT
TCA
Co
nce
ntr
atio
n
10
1
time [h]
time [h]
10
1
Co
nce
ntr
atio
n
10
1
time [h]
Co
nce
ntr
atio
n
CYP2D6
amplification
normal
mutations
MDR1
TPH2
SERT
Drug
Pharmacogenetics and Prediction of Response and Toxicity
ME3010.ppt
The example of TPMT
• Of the known functionally important mutations only a limited number are tested
• Presence of unknown mutations
• Penetrance of gene
• Phenotype studied is only in part caused by candidate gene
• Other genetic and nongenetic factors contribute to phenotype
ME3273
Dose dependent ADRs
Dose independent ADRs
• hematotoxicity (leukopenia, pancytopenia)• hepatotoxicity (most cases)
• pancreatitis• gastrointestinal disturbancies* (eg. nausea, vomiting, diarrhoea)• flu-like symptoms (such as fever*, headache)• rash*• arthralgia*, myalgia*
* commonly termed also as Azathioprine intolerance
Patients (n= 41) with Crohn‘s Disease and SevereMyelosuppression during Azathioprine Therapy
Colombel et al.Gastroenterology 2000
homozygousmutant
wildtype0
10
20
30
heterozygous
num
ber
n=4 n=8 n=299.8%
19.5%
70.7%TPMT genotype
29.3%
ME3278
13 (14%) comedicationallopurinol
391CMV Infection (36% received
ASA �12 mo )
53 wildtype (56%)
Schwab et al. unpublished data (Mai 2005)
12homozygous
28heterozygous
2het+allopurinol
42 mutant (44%)
onset within 3 - 8 weeks
95 patients with severe myelosuppression(pancytopenia)
Spontaneous reporting of cases (n=14 066); 92 % IBD
Patients on thiopurine therapy (n=4525)
Results of TPMT diagnostics in Stuttgart
ME3280
� DPD catalyzes 1st and rate limiting step
� commonly expressed Fe-S protein (predominately in human liver)
� cytosolic enzyme
� endogenous substrates known
� association to inborn error (familial pyrimidinemia) and
� severe 5-FU toxicity(Diasio et al., 1988)
�-alanine �-F-�-alanine �-amino-isobutyrate
Dihydropyrimidine Dehydrogenase (DPD)
dihydropyrimidinedehydrogenase (DPD)
ME2755.ppt
ME2843.ppt
798patients included
78 (9.7 %)concomitant chemotherapy
30 (3.8 %)incomplete documentation
5 other causes (0.6 %)
Reasons for exclusion
German Study-Groupon 5-FU Toxicity
GI tumors 96 %, breast cancer 2.4 %, CUP 1.6 %
685study patients
Pitfalls in Pharmacogenetic Testing:DPYD Mutations and 5-FU Toxicity
DPYD Exon 14 Skipping Mutation explains only ~ 15 % of 5-FU toxicity
patients 685
*2 / *2 0wt / *2 13
German Study-Groupon 5-FU Toxicity
2.2%
WHO grade IV III II I-0
number of pts 35 91 80 479
11.4%3.8% 0.8%
n=4 n=2 n=3 n=4Tox 4+3 vs 0-2:
95%CI 1.3-12, P = 0.019Odds Ratio 3.91.9%
8/2003
ME2844.ppt
Phenotype is only in part caused by candidate gene
Female Sex is a Risk Factor for 5FU Toxicity
0
10
20
30
40
50
60
70
WHO° 0-I476 patients
WHO° II81 patients
WHO° III92 patients
WHO° IV34 patients
p = 0.0015
mal
e
fem
ale
ME3030
Sex ratio [%] : 61:39 47:53 42:58 47:53
Total 56:44
Sex and DPYD*2 Allele associated 5-FU Toxicity
0
2
4
6
8
10
12
14
16
18
20
WHO° 0-I476 (4)
WHO° II81 (3)
WHO° III92 (2)
WHO° IV34 (4)
Dis
t rib
uti
on
of
DP
YD
*2 (
%)
male: p < 0.0001
female: n. s.
2
3
ME3035
patients (n)
4 1
31
0 0
no difference in allele *2 frequency betweenfemales (2.3%) and males (1.6%)
Nomogram for the Prediction of 5-FU Toxicity
DPD: Dihydropyrimidine dehydrogenase
Points
Age
Sex (DPD wt)
Sex (DPD*2)
Thymidylatesynthase
Folinic acid
Mode
Methylenetetrahydro-folate reductase
Total Points
0 10 20 30 40 50 60 70 80 90 100
0 20 40 60 80 100 120 140 160 180 200 220
90 80 70
25 60
26 100
Infusion
yes
no
wt
mt
mt
0.10.2
0.30.4
0.50.6
0.70.8
0.9Probability (WHO�3) [%]
ME3152
50
17
27
11
8
16
Bolus
wt
Predictive model for 5-FU WHO4-toxicity
Factor female malesex (DPD wt) 17 0sex (DPD*2) 9 (26) 100TS (wt) 16 16MTFR (wt) 8 8Folinic acid 27 27Bolus 11 11age 50 50total points 138 212Probability 0.48 0.95
Contribution of DPD*2 is negligible (4 %) in female but substantial (45 %) in male patients.
ME3151
What is the information content of a genetic test ?
ME2748.ppt
Sensitivity likelihood that a patient with a given phenotype (ADR)
will test positive
Specifity likelihood that a patient without the phenotype (ADR)
will test negative (i.e. not test false positive)
Positive likelihood that a patient with a positive test will have the
predictive value phenotype (ADR)
Negative likelihood that a patient with a negative test will not
predictive value have the phenotype (ADR)
ME3185.ppt
Genotype and Phenotype
Limitations of present approach: Usually focus on one gene
Like most disease phenotypes, drug phenotypes
(response, nonresponse, toxicity) are complex polygenic
traits with nongenetic factors contributing to the
manifestation of phenotypes.
The extent to which genetic factors contribute to phenotype
will depend whether the candidate gene is a gene of major,
moderate or minor effect.
ME3211
41 Candidate genes withpotential relevance for
statin response(Zineh 2005)
Contribution of polymorphism of
most genes in lowering cholesterol
will be moderate.
ME3225.ppt
LimitationsLimitations to to currentcurrent statinstatin pharmacopharmaco--geneticsgenetics studiesstudies ((ZinehZineh et al. 2005et al. 2005))
• Generally not multi-gene studies (or studiesconsidering combinations of several genes)
• Statistically significant results are not necessarilyclinically meaningful
• Many studies - few results replicated• Gene-environment, gene-disease and dietary
factors not contolled• Candidate polymorphisms often associated with
baseline cholesterol
ME3228.ppt
Pharmacogenomics of Statins: Outlook
• Genotype groups with diminished lipidresponse may still show clinically usefuleffects!
• How predictive is lipid lowering efficacy forclinical endpoints?
• Will genetic testing to predict response and toxicity be feasible and cost-effective?- Maybe, but expectations are probably too high
• Large studies with many genes are needed
ME3126Can Genotype-based Dose Selecetion Reduce Toxicityand Increase Response ?
Genotype Dose Concentration Response Dose Concentration Responseadjusted
0.1
0.2
1
4
PM
IM
UM
EM
Co
nce
ntr
atio
n 10
1
Time [h]
Time [h]
10
1
Co
nce
ntr
atio
n
10
1
Time [h]
Co
nce
ntr
atio
n
toxicity,
side effects
10
1
Time [h]
Co
nce
ntr
atio
n
side effects
drug response
no drug response
1
1
1
1
10
1
Time [h]
Time [h]
10
1
Time [h]
no toxicity,
no side effects
10
1
Time [h]
no side effects
drug response
drug response
Co
nce
ntr
atio
nC
on
cen
trat
ion
Co
nce
ntr
atio
n
10
1
Co
nce
ntr
atio
n
The Impact of CYP2D6 Genotype on Adverse Drug Reactionand Nonresponse During Treatment with Antidepressants
0
20
40
60
80
100
Adverse Effects Nonresponse
PM UM
expected observed expected observed expected observed
ME3019Kawanishi et al., 2004Rau et al., 2004
7 % 29 %
2 % 19 % 1 % 10 %
ME3287.ppt0.1 100101 MRS
Indi
vidu
als
10
20
40
30
IM: intermediatemetabolizer~ 10-15 %
PMpoor metabolizer
~5-10 %
UM:ultrarapid
metabolizer~ 5-10 %
EM: extensive metabolizer (~70-80 %)
N = 308
Sparteine
N
N
CYP2D6 Oxidation Phenotypes in Caucasians
ME3285.ppt
Limited Predictivity of CYP2D6 Gene Duplicationfor UM Phenotype
Indi
vidu
als
10
20
40
30
0.1 100101 MRS
N = 308
Carriers of 3 gene copies
6/16 UM (38%)
ME3286.ppt
CYP2D6 Protein and Propafenone EnzymeActivity in Human Liver
Correlation PPF-Protein
r=0.77
*4*4
*41*0
Dupl
0 25 50 75 100 125 1500
100
200
300
400
2D6 Protein pmol/mg
PP
F p
mo
l/mg
/min
N=125
Polygenic Nature of Drug Response: Antidepressants
ME3125
Co
nce
ntr
atio
n
10
1
time [h]
time [h]
10
1
Co
nce
ntr
atio
n
10
1
time [h]
Co
nce
ntr
atio
n
CYP2D6
amplification
normal
mutations
Drug
Drug Metabolism
Poor predictive value of CYP2D6 and CYP2C19 genotype for severe adverse drugreactions and non-response leading to discontinuation of treatment
• Comparable doses used; compliance
• Measurement of drug levels
• Coadministration of drugs: Phenocopying
• Coexisting diseases; age; gender
• Predictive value of genotype for phenotype:UM genotype predicts only 20 – 30 % of UM phenotype
Polygenic Nature of Drug Response: Antidepressants
Drug Transport
MDR1 at blood brain barrierComparable plasma concentration,yet different concentration at siteof action:
• Poor predictive value of drug concentration
• net concentration = influx (diffusion) – efflux (transport)
• Contribution of MDR1 polymorphism to
response only at comparable concentrations
BBB BBBconc
.
conc
.
time [h] time [h]
ME3124
Polygenic Nature of Drug Response: Antidepressants
Drug Target
1. Concentration of serotonin in synaptic cleft is influenced bybiosynthesis (TPH2), re-uptake(SERT) and catabolism (MAOA)
2. Inhibition of serotonin re-uptakedepends on drug concentration in synaptic cleft
3. Mutations of receptors and signalling pathways affectneurotransmitter and drug effects
MAOA
MDR1
HT Receptors
5-HT
TCA
TPH2
SERT
ME3122
Polygenic Nature of Drug Response: Antidepressants
ME3050
Drug Target
Contribution of receptor / signallingpathway to drug response shouldbe assessed at comparable plasmalevels
BBB: Stratification of patient groupsfor transporter polymorphisms
Comparable effect at receptor, but10 fold difference in concentrationrequired between genotypes
(Mason et al. 1999)
Gly389
Arg389
-11 -10 -9 -8 -7 -6 -5 -4 -3
log [Isoproterenol]A
den
ylyl
cycl
ase
acti
vity
[pm
ol/m
in/m
g]
0
15
30
45
60
75
ME3225.ppt
LimitationsLimitations to to currentcurrent statinstatin pharmacopharmaco--geneticsgenetics studiesstudies ((ZinehZineh et al. 2005et al. 2005))
• Generally not multi-gene studies (or studiesconsidering combinations of several genes)
• Statistically significant results are not necessarilyclinically meaningful
• Many studies - few results replicated• Gene-environment, gene-disease and dietary
factors not contolled• Candidate polymorphisms often associated with
baseline cholesterol
ME3228.ppt
Pharmacogenomics of Statins: Outlook
• Genotype groups with diminished lipidresponse may still show clinically usefuleffects!
• How predictive is lipid lowering efficacy forclinical endpoints?
• Will genetic testing to predict response and toxicity be feasible and cost-effective?- Maybe, but expectations are probably too high
• Large studies with many genes are needed