Transcript Predictive factors of response to chemotherapy in colorectal cancer

Pharmcogenetics in
oncology
Pierre Laurent-Puig
INSERM, U775 Molecular basis of xenobiotic
response
AP-HP Hôpital Européen Georges Pompidou
Molecular Oncology and Pharmacogenetic laboratory
QuickTime™ et un
décompress eur TIFF (non c ompres sé)
sont requis pour v is ionner cette image.
Introduction
• Multiple active regimens for cancer but:
– Variation in response rate to the different
regimen
– Unpredictable toxicity for the different
regimens
• With choice came the time to decision
– Need to the development of tools which
help clinicians
Pharmacogenetics
Anticancer
drugs
Genetic
factors
Environmental
factors
Variations
in drug
response
or toxicity
Transport
Metabolism
Drug
target
Drug
interactions
Lee et al. Oncologist. 2005;10:104-111.
Cell
Introduction
• Pharmacogenetics : the science of incorporating
information of inherited genetic variability into
predicting treatment response.
• Polymorphisms in both individual’s genome as
well as tumor genome will affect drug toxicity
and drug response.
• Drug-related toxicity will be predicted mainly by
genotyping non tumour tissues.
• Drug response will be predicted both
genotyping non tumour and tumour tissues.
by
5-FU Pathway
Capecitabine
FBAL
Carboxyl esterase
-ureidopropionase
5’dFCR
Cytidine deaminase
FUPA
5’dFUR
DNA (Uracil misincorporation)
Thymidine phosphorylase
Tegafur
5F-uridine
Uridine
phosphorylase
Dihydropyrimidinase
80%-90%
5-FU
DHFU
Thymidine
phosphorylase
5F-deoxyuridine
Orotate
phosphoribosyl
transferase
Uridine
kinase
dUTP
Dihydropyrimidine
dehydrogenase
Thymidine
kinase
5-FUMP
Uridine
monophosphate
kinase
5-FUDP
Uridine
diphosphate
kinase
5-FUTP
dUTPase
dTMP
FdUDP
Ribonucleotide
reductase
dUMP
FdUTP dTDP
Uridine
diphosphate
kinase
dTTP
DNA
Methylene
tetrahydrofolate
reductase
dUTPase
5, 10-MTHF
Serine hydroxymethyl
transferase
Thymidylate
synthase
FdUMP
Uridine
monophosphate
kinase
dUDP
5-MTHF
THF
DHF
Folinic acid
(leucovorin)
Dihydrofolate
reductase
DPD and 5-FU toxicity
• IVS14+1G>A Polymorphism leads to the skipping of exon 14.
• As a result, the mature DPD mRNA lacks a 165 nucleotide segment
encoding amino acids 581–635
• Prevalence of IVS14+1G>A 0.75% to 2.2%
van Kuilenburg et al. Eur J Cancer. 2004;40:939-950.
5-FU Toxicity and the Prevalence
of IVS14 + 1G>A Mutations
Total
group
DPD
≤ 70%
DPD
> 70%
Patients (n = 60)
Heterozygotes
Homozygotes
16 (27%)
1 (2%)
15 (42%)
1 (3%)
1 (4%)
0
Patients (n = 25)
Heterozygotes
Homozygotes
5 (20%)
1 (4%)
n.d.
n.d.
n.d.
n.d.
IVS14 + 1G>A
• 24-29% of patients with grade 3-4 5-FU toxicity were heterozygous or
homozygous for the IVS14 + 1G>A mutation
• Almost half of the patients with decreased DPD activity were carriers of the
IVS14 + 1G>A mutation
• Applying Bayes’ theorem we can estimate the risk of 5-FU grade 3-4 toxicity
for a carrier of IVS14+1G>A mutation to 87%
van Kuilenburg et al. Eur J Cancer. 2004;40:939-950; van Kuilenburg et al. Pharmacogenetics. 2002;12:555-558;
Raida et al. Clin Cancer Res. 2001;7:2832-2839.
5-FU Toxicity and the Prevalence
of DYPD polymorphism
14 IVS14 + 1G>A Exon 14 skipping; 2846A>T, D949V; 1679T>G,I560S
Sensitivity, specificity, and PPV and NPV of the detection of these three
major SNPs as toxicity predictive factors were 0.31, 0.98, and 0.62 and
0.94, respectively.
Morel A. Mol Cancer Ther 2006:5:2895-904.
TS Polymorphisms
Functional consequences
Polymorphisms
• One in the enhancer region of TS consisting • In vitro studies demonstrated that an increasing
in a double or triple repeat of a 28-base pair
number of tandem repeats leads to an increase in
sequence (2R or 3R)
TS gene expression and TS enzyme activity
• G>C SNP in the second of the three 28-bp
• The SNP in the second repeat of the 3R allele disrupts
repeats produces two additional alleles (3RG
the upstream stimulatory factor (USF) consensus
or 3RC)
element and therefore decreases the in vitro
• One in the 3’UTR region consisting of an
transcriptional activation of TS
insertion or a deletion of 6 bp
•
Desai et al. Oncogene. 2003;22:6621-6628.
The TS 3’UTR del6 allele may determine low TS mRNA
stability and low TS expression in comparison with TS
3’UTR ins 6 allele
N = 86 CRC patients
5-FU based chemotherapy
• TS promoter genotype is
predictive of grade 3/4
toxicities with 5-FU
• 3R/3R genotype, overexpressing TS, has fewer
toxicities
• No association with a
response to 5-FU and survival
Grade 3 / 4 toxicities (%)
5-FU Toxicity (all) and Polymorphism
of TS Gene Promoter in CRC
50
43 %(6/14)
40
30
18 % (8/44)
20
10
4 % (1/28)
0
2R/2R
(16%)
2R/3R
(49%)
P = 0.02
Lecomte et al. Clin Cancer Res. 2004;10:5880-5888.
* 2R/2R vs 2R/3R, 3R/3R
3R/3R
(31%)
MTHFR Polymorphisms
A222V
E428A
• Two polymorphisms may alter enzyme activity
– 677C->T (Ala222Val) increases MTHFR thermolability
– 1298A->T (Glu428Ala) decreases MTHFR activity
• Since a loss in MTHFR enzymatic activity may favor an increase in
intracellular CH2FH4 concentrations, it can be hypothesized that tumors
exhibiting mutated MTHFR genotypes may be more sensitive to 5-FU
cytotoxicity.
MTHFR Polymorphisms & Clinical
Outcomes
References
5-FU Dosage
Sample size
Finding
Cohen et al. Clin
Cancer Res 2003
Different regimen
of p.o or i.v. 5-FU
43
Significant increase
response rate for
pts with at least
one 222 mutant
allele
Etienne et al.
Pharmacogenetics
2004
Different regimen
of i.v. 5-FU
98
Significant increase
response rate for
pts homozygous
for 222 mutant
allele
Jakobsen et al.
J Clin Oncol
2005
Diffenrent regimen
of i.v. 5-FU
88
Significant increase
response rate for
pts homozygous
for 222 mutant
allele
MTHFR and 5-FU response
Responder
Non
responder
OR [95% CI]*
A222A
37 (39%)
57
1 ref
A222V
29 (31%)
64
0.7 [0.36-1.3]
V222V
21 (62.5%)
11
2.94 [1.18-7.53]
219 patients with advanced colorectal cancer receiving 5-FU
*unadjusted OR
Cohen et al. Clin Cancer Res 2003;9:1611-1615; Etienne et al. Pharmacogenetics 2004;14:785-792; Jakobsen
et al. J Clin Oncol. 2005;23:1365-1369.
Two types of colorectal cancer
Tumor LOH + (85%)
Tumor MSI + (15%)
•Hyperploid
•Diploid
•Recurrent allelic losses
on chromosomes 17p, 18q,
5q, 8p, 22q
•No allelic losses on
chromosome 17p, 18q,
5q, 8p, 22q
•Frequent p53 and APC gene
mutations
•Rare p53 and APC gene mutations
•Frequent KRAS and
PIK3CA gene mutations
•Frequent BRAF and PIK3CA
gene mutations
•Up to 20 different
genes were found
mutated
•Mainly in distal colon
•Frequent mutation of TGFß receptor
type II, Caspase 5, Bax and TCF4 genes
•Alteration of MLH1 MSH2, MSH6 and
MSH3 genes
•Mainly in proximal colon
Chromosomal instability
Genetic instability
Paradigm: FAP tumors owing to germline
Paradigm: HNPCC tumors owing to
mutation of APC gene
germline mutation of MMR genes
Treatment
Role of MSI status
in adjuvant 5FU
treatment response
Control
MSS tumours
Control
MSI tumours
Treatment
A significant interaction was observed between microsatellite instability status
and the benefit of treament p=0.01 Ribic et al N Engl J Med 2003; 349:247-57.
Irinotecan Pathway
P-gp
CPT-11
CPT-11
ABCB1
CES2
CYP3A4
APC
CPT-11
CYP3A5
SN-38
NPC
CES2
ABCB1
SN-38
UGT1A1
SN-38G
P-gp
SN-38
TOPI
death
UGT1A1 promoter polymorphisms
Promoter
Exons
UGT1A1
Variant (TA)7 TAA
(TA)6 TAA
Decrease gene expression
Normal gene expression
Low glucuronidation
leads to 1.8 to 3.9 fold
lower glucuronidation of SN-38
Normal glucuronidation
Iyer et al. Pharmacogenomics J. 2002;2:43-47.
Published Data on UGT1A1 & grade 4
neutropenia
n/N (%)
Author
7/7
95% CI
21/58 (36%)
--
--
Carlini
0/6
Innocenti
3/6 (50%)
3/53
(6%)
16.7
2.3 – 120.6
Marcuelloa
1/10 (10%)
2/85
(2%)
4.6
0.4 – 56.0
Rouits
4/7 (57%)
10/66 (15%)
7.5
1.4 – 38.5
Andob
4/7 (57%)
22/111 (20%)
5.4
1.1 – 25.9
aOriginally
(0%)
6/6 + 6/7
Est.
Odds
Ratio
reported Gr 3+ Gr 4 values from personal communication.
bGr 4 leukopenia and/or Gr 3+ diarrhea.
UGT1A1 promoter genotype and
Irinotecan toxicity (in combination)
•
•
400 patients with high risk stage III colorectal cancer included in a
randomised trial FNCLCC Accord02 / FFCD9802 comparing
LV5FU2 alone versus Folfiri regimen.
FOLFIRI regimen
– Irinotecan 180mg/m2, 90 min iv day 1
– Leucovorin 200mg/m2 during irinotecan day 1
– 5FU 400mg/m2 iv bolus followed by 2400mg/m2, during 46 hours
•
Clinical evaluation
– Toxicity
– Survival
•
UGT1A1 TA6/TA7 and -3156 G->A polymorphisms were studied in
94 patients receiving FOLFIRI
Hematological toxicity grade 3-4
according to UGT1A1 genotypes
TA6>TA7
-3156G>A
100%
100%
80%
80%
60%
60%
40%
40%
20%
20%
0%
(TA6)2
TA6TA7
Toxicity >3
(TA7)2
Toxicity 0-2
p=0.15
0%
GG
GA
Toxicity >3
AA
Toxicity 0-2
p=0.02
Survival without grade 3-4
hematological toxicity (-3156G>A)
Proportion of patient
without hematological
grade 3-4 toxicity
1
0.8
GG
0.6
AG
AA
0.4
0.2
0
0
2
Côté et al. Clinical Cancer Res accepted
4
6
8
Number of cycles
10
12
p=0.012
Response to chemotherapy WHO and
UGT1A1 *28 polymorphism (n=238)
UGT1A1
CR
PR
SD
PD
R
Risk for PD
OR 95%CI
Risk for PD+SD
OR 95%CI
6/6
N=109
10
34
39
36
40.3%
1
1
6/7
N=108
5
40
32
31
41.6%
0.77
[0.4-1.4]
0.92
[0.5-1.6]
7/7
N=21
3
11
5
2
66.6%
0.19
[0.04-0.9]
0.32
[0.12-0.56]
Toffoli presentation the ASCO GI Meeting in January 2006
Oxaliplatin Pathway
Platinum
Extracellular
Cell membrane
Intracellular
ABCG2
ABCC2
Platinum
Detoxify
GSTP1
Platinum
Translesional
replication
Platinum
Pt
G
Platinum
G
ERCC1
Damage
recognition
XPA
XRCC1
ERCC2
Mismatch
repair
Excision
repair
Cell death
GSTP1 Polymorphisms
I105V
A114V
•
Two non synonymous coding polymorphisms
– Single nucleotide polymorphism (SNP) at residue 105 Ile105Val
substitution (39%)
– Single nucleotide polymorphism (SNP) at residue 114 Ala114Val (12%)
– Four haplotypes *A Ile105+Ala114; *B Val105+Ala114; *C
Val105+Val114; *D Ile105+Val114
•
Variability in enzymatic activities depending of the substrate
•
Lower thermal stability for the 105 Val allele
 Variability in detox properties according to the haplotypes
Ishimoto TM, Ali-Osman F. Pharmacogenetics 2002;12:543-553.
GSTP1 Polymorphisms & Clinical
Outcomes
References
Chemotherapy
regimen
Sample size
Finding
GSTP1 and neurotoxicity
Lecomte et al.
Clin Cancer Res
2006
Accepted
Grothey et al.
A3509 ASCO 2005
FOLFOX regimens
FOLFOX regimen
64
Early neurotoxicity
occurs more
frequently in pts
homozygous for wild
type allele
288
Early neurotoxicity
occurs more
frequently in pts with
at least one Val Allele
107
Better prognosis for
patients with mutant
allele
GSTP1 and survival
Stoehlmacher et al.
JNCI 2004
Combination of 5FU
and Oxaliplatin
GSTP1, Oxaliplatin and
survival
107 patients with metastatic colorectal cancer who received 5-FU/oxal
VAL/VAL (n=10)
ILE/VAL (n=45)
ILE/ILE (n=45)
P < 0.001
After adjustment for performance status the relative risk of dying for patients
with ILE/VAL and ILE/ILE genotypes was 2.73 and 3.25 respectively P = 0.072
Stoehlmacher et al. JNCI. 2002;94:936-941; Stoehlmacher et al. Br J Cancer. 2004;94: 944-954.
ERCC1 polymorphism
N118N
•
ERCC1 codes for a protein in the nucleotide excision repair pathway
•
High ERCC1 level is associated with resistance to oxaliplatin
•
Polymorphism in codon 118 is silent
•
118 TT genotype is associated with reduced translation of the gene, and
presumably reduced DNA repair capability (ovarian cancer cell lines).
But in colon cancer, patients with ERCC1 118 TT genotype have a higher
expression of ERCC1 mRNA.
ERCC1 Polymorphisms & Clinical
Outcomes
2 studies with conflicting results:
– Stoehlmacher et al. Br J Cancer 2004;94: 944-54
• 107 metastatic colorectal cancer patients receiving
oxaliplatin plus 5FU
• Relative risk of dying was 2.05, 95%CI [1.00-4.20] for CT
and TT genotypes as compared to CC genotype patients
– Viguier et al. Clin Cancer Res 2005;11:6212-7
• 61 metastatic colorectal cancer patients receiving FOLFOX
regimen
• The objective response rate was 21.4% , 42.3% and
61.9% for CC, CT and TT genotypes
Summary
Polymorphism
TS
Genotype
or Allele
Frequency
Toxicity
Efficacy
Comments
3R/3R
30 to 40%
Lower
Lower?
2R/2R
18 to 25%
Higher
Higher?
Exact role of 3RG
and 3RC allele, role
of allelic losses in
tumor DNA
Role of TS
haplotype
IVS14+1G>
A
0.5 to 2%
Higher
222: Val/Val
4 to 6%
Higher
428: Ala/Ala
6 to 8%
Higher
7/7
10%
Higher
Higher or
Equivalent
Role of other
polymorphisms
105: Val/Val
15 to 20%
Higher
Lower
105: Ala/Ala
35 to 40%
Higher
Contradictory
on toxicity
118: CC
38 to 42%
5-FU
DPD
MTHFR
Irino
UGT1A1
GSTP1
Oxali
ERCC1
118: TT
10 to 15%
Promoter
methylation of DPD
Better
survival
Better
response
Role of haplotype
Contradictory for
efficacy
EGFR pathways
EGF ligands
TGF
Dimerisation
membrane
EGFR
TK
TK
TK P Phosphorylation
=
P
Activation
angiogenesis
proliferation
Resistance to
apoptosis
VEGF, IL8
Cycline D1
Bad, caspase 9
Ras/MAPK pathways
membrane
TK TK P Grb
hSos Ras
GTP
P
Intracellular phosphorylation cascade
(Raf, Erk, Mek) = MAP Kinases
cycline D1
 cdk6
P
MEK
ERK
P
ERK
P
Serine/Threonine Kinases
C-MYC,
JUNB,
c-JUN
MEK
Raf
c-fos
P
ERK
P
phase G1 cell cycle
nucleus
PI3K/Akt pathways
Phosphatidyl-inositol
membrane
PI 3,4,5-P3
TK TK P
P
P
PI3K
Pdk1
Pdk2
PI3K
Akt
Akt
P P
Activation of eIF-4E
inhibition of 4E-BP1 S6
kinase
Bad, caspase-9,
Fkhrl-1
P
Gsk3
P
Stabilisation of Cyclin D1
Protein synthesis 
Resistance to apoptosis
Cell cycle G1 transition
Response to Cetuximab
• 30 patients treated by cetuximab for a stage
IV colorectal cancer
– 3 in first line with folfiri
– 3 in second line
– 24 in third line
• 11 responders (1 with complete response )
• Sequencing of KRAS (ex1), BRAF (ex11&15),
PIK3CA (ex1,ex2, ex9, ex11)
• Measure of EGFR amplification by CISH
(F.Penault-Llorca)
Lievre et al. Can Res 2006;66:3992-5
Prevalence of alterations
• KRAS is mutated in 13 cases (48%)
– 10 cases at codon 12
– 3 cases at codon 13
• PIK3CA is mutated in 2 cases (7%)
– 2 cases in exon 9
– these 2 tumours are also mutated for KRAS
• BRAF is not mutated
• EGFR is amplified in 3 cases
– >20 copies in 1 case
– >10 copies in 2 cases
Response rate to cetuximab therapy
according to KRAS mutation
100%
80%
60%
40%
20%
0%
Non responders
patients
responder patients
P=0.0003
non mutated KRAS
Mutated KRAS
6
13
11
0
Lievre et al. Can Res 2006;66:3992-5
Overall survival according to KRAS
mutation for patients treated with
cetuximab
Lievre et al. Can Res 2006;66:3992-5
Conclusions
•Molecular predictive factors of response to
chemotherapy for colorectal cancer patients have
been identified
•From the host or from the tumor
•From different pathways
–Xenobiotic metabolizing enzyme , Drug target
•It is time to study more than one predictive factor
at the same time in order to choose the best
– The development of “omic” approaches will allow the
“a la carte” treatment of cancer patient with the most
efficient and the less toxic drugs