Effect of Genetic Polymorphisms on Therapeutic Response and Clinical Outcomes in Pancreatic Cancer Patients Treated With Gemcitabine

Hye In Woo; Ka-Kyung Kim; Hangseok Choi; Kee-Taek Jang; Jun Ho Yi; Young Suk Park; Joon Oh Park; Soo-Youn Lee

Pharmacogenomics. 2012;13(9):1023-1035.

Patients & Methods

Study Population

We initially included 298 Korean patients with advanced pancreatic cancer who were treated with first-line gemcitabine-based chemotherapy at the Samsung Medical Center (Seoul, Korea) between January 1999 and November 2009, retrospectively.^[22] Among them, a total 102 patients who had stored DNA, and clinical and laboratory data were included in the current study. A complete set of clinical data including age at diagnosis, sex, smoking history, diabetes, tumor location, tumor status, carbohydrate antigen (CA)19-9 level before the start of chemotherapy, chemotherapeutic regimens, performance status, and toxicity was obtained from the medical record of each patient (Table 1). The patients had biopsy-proven pancreatic ductal adenocarcinoma or adenosquamous carcinoma (six patients). All of patients were treated with gemcitabine alone or gemcitabine in combination with either erlotinib or fluoropyrimidines. A clinical and laboratory assessment was performed once or twice a month, and tumor size was measured by computed tomography scan at 8 weeks after the initiation of chemotherapy. Tumor response was evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST) and the best response was recorded, retrospectively.^[23] Therapeutic response data were screened for two categories, response versus nonresponse and progressive versus nonprogressive. In this study, we chose to focus on the progressive versus nonprogressive groups since this analysis provided a more balanced distribution of the number of patients. Overall survival (OS) was calculated from the time that a patient started treatment until the last follow-up or death, and time to progression (TTP) was calculated from the time that a patient started treatment to disease progression. Owing to the fact that relatively few patients experienced adverse effects from the chemotherapeutic agents, no further analysis of the association between toxicity and genetic polymorphisms was undertaken. This study was approved by the Samsung Medical Center institutional review board.

Selection of Target Gene Polymorphisms

Twenty four genes related to metabolism and action sites of gemcitabine were extracted using a public database,^[101] and through literature review using the PubMed database.^[102] Search keywords used in various combinations were 'pancreatic cancer', 'gemcitabine' and 'polymorphism'. For 24 genes, a total of 197 SNPs from previous studies using the above keywords in the PubMed database,^[102] from public database,^[101] and from Haploview version 4.2 were selected. An r² ≥ 0.80 and minor allele frequency (MAF) ≥0.05 were used for tagSNPs selection using Haploview based on Korean HapMap database,^[103] or data of Japanese and Chinese populations from the International HapMap Project^[104] in cases of lack of Korean data. Next, SNPs with MAF <0.01 were removed, including all five SNPs in the RRM2 gene, and overlapping SNPs from each selection method. Finally, 124 genetic polymorphisms in 23 genes (ABCB1, AICDA, CDA, CDC5L, CMPK1, DCK, DCTD, EPC2, ESR2, FKBP5, MYBBP1A, NME7, NT5C3, PARP1, POLS, RRM1, RRM2B, SH2D5, SLC28A1, SLC28A3, SLC29A1, TLE4 and TYMS) were selected for genetic analyses.

SNP Genotyping

Genomic DNA was extracted from peripheral blood leukocytes of all patients using the Wizard® Genomic DNA Purification Kit according to the manufacturer's instructions (Promega, WI, USA). Extracted DNA samples were stored at −70°C until analyzed.

The DNA concentration and purity were measured as the optical density at 260 nm and the ratio of the optical densities at 260/280 nm, respectively, using a NanoDrop (Thermo Fisher Scientific, DE, USA). The required concentration and purity were 5–10 ng/µl and 1.7–1.9. SNPs were genotyped using the MassARRAY® system (Sequenom, Inc., CA, USA) and genotypes were called by the Birdseed calling algorithm in the SpectroTYPER™ software (Sequenom, Inc.). The TYMS 28-bp repeat was detected using a protocol previously described.^[24] After excluding 22 SNPs with a MAF <0.01, a call rate <85% or Hardy–Weinberg equilibrium (HWE) p-value <0.001, 101 autosomal SNPs in 22 genes and a copy number variation in the TYMS gene were analyzed (Table 2).

Statistical Analysis

The OS was the primary end point considered in this study. The TTP and tumor response to chemotherapy were also evaluated. The Cox proportional hazards regression model was performed under an additive, dominant, and recessive genetic models to evaluate the association of each SNP with OS and TTP.^[25] For clinical factors, Kaplan–Meier curves and the log-rank test were performed. The prechemotherapeutic CA19-9 level and tumor status were included as an adjusted variable for the multivariable analysis of OS and TTP. Patients with complete remission, partial remission or stable disease were combined in the nonprogressive group and patients with progressive disease were included in the progressive group.^[26] An association between each SNP and patient group according to tumor progression was evaluated using allelic test and additive, dominant, and recessive genetic models. For SNPs with raw p-values less than 0.05, age, diabetes, prechemotherapeutic CA19-9 level, and tumor status were included as confounder variables in the multiple logistic regression. p-values were corrected with the false-discovery rate for candidate SNPs and the Bonferroni method for the number of genetic models (additive, dominant and recessive). p-values less than 0.05 were considered statistically significant.

Haplotype analyses using PLINK, version 1.06 and R, version 2.11.1 (R Foundation for Statistical Computing, Vienna, Austria) was performed for each gene and combinations of genes involved in identical action sites according to tumor response group.

1 2 3 4 5 6

Next Section

Table 1. Patient characteristics (n = 102).
Variable	n (%)
*Age (years)*
Median (range)	60 (34–80)
Less than 60	53 (52.0)
60 or more	49 (48.0)
*Gender*
Male	71 (69.6)
Female	31 (30.4)
*Smoking*
Current/ever	34 (33.3)
Never	68 (66.7)
*Diabetes status*
Negative	79 (77.5)
Positive	23 (22.5)
*CA19-9 (units/ml)*
Less than 200	53 (52.0)
200 or more	49 (48.0)
*Tumor status*
Locally advanced	25 (24.5)
Metastasis	77 (75.5)
*Tumor grade*
WD/MD	36 (53.7)
PD/UD	31 (46.3)
*Tumor location*
Head	51 (50.0)
Others	51 (50.0)
*Chemotherapy regimen*
G only	56 (54.9)
G combination	46 (45.1)
*ECOG performance status*
0, 1	100 (98.0)
2	2 (2.0)
*Response to treatment*
Complete remission	4 (3.9)
Partial remission	22 (21.6)
Stable disease	46 (45.1)
Progressive disease	30 (29.4)
*Survival, median (95% CI)*
Median TTP, days	5.5 (4.2–6.3)
Median OS, months	10.0 (8.6–12.7)

G: Gemcitabine; MD: Moderately differentiated; OS: Overall survival; PD: Poorly differentiated; TTP: Time to progression; UD: Undifferentiated; WD: Well differentiated.

Table 2. List of 102 genetic polymorphisms in 23 genes used in this study.
Gene	Ch	SNP ID	MAF	Call rate
CDA	1	rs532545	0.15	0.99
CDA	1	rs602950	0.17	0.96
CDA	1	rs3215400	0.48	0.99
CDA	1	rs2072671	0.15	0.95
CDA	1	rs818202	0.33	0.97
CDA	1	rs1048977	0.24	0.99
CMPK1	1	rs3925058	0.38	0.96
CMPK1	1	rs7543016	0.35	0.93
CMPK1	1	rs4600090	0.20	0.96
CMPK1	1	rs9436883	0.13	0.95
CMPK1	1	rs35687416	0.10	0.95
CMPK1	1	rs1044457	0.20	0.99
NME7	1	rs1320964	0.42	0.95
NME7	1	rs12753710	0.28	0.96
PARP1	1	rs2271347	0.04	0.97
PARP1	1	rs3219110	0.11	0.98
PARP1	1	rs3219090	0.17	0.90
PARP1	1	rs2293464	0.34	0.96
PARP1	1	rs932002	0.40	0.85
PARP1	1	rs10799349	0.33	0.96
SH2D5	1	rs10916852	0.13	0.97
SH2D5	1	rs733581	0.03	0.99
EPC2	2	rs16829119	0.24	0.96
EPC2	2	rs6724787	0.12	0.99
EPC2	2	rs6739555	0.49	0.89
EPC2	2	rs12614977	0.23	0.95
DCK	4	rs7684954	0.07	0.98
DCK	4	rs4694362	0.21	0.89
DCTD	4	rs12507552	0.13	0.91
DCTD	4	rs3886768	0.16	0.99
DCTD	4	rs13147196	0.12	0.97
DCTD	4	rs13137332	0.11	0.97
DCTD	4	rs13116598	0.26	0.97
DCTD	4	rs6823077	0.13	0.96
DCTD	4	rs11730323	0.16	0.99
DCTD	4	rs17074255	0.24	0.90
POLS	5	rs274717	0.41	0.99
POLS	5	rs274714	0.50	0.94
POLS	5	rs274713	0.36	0.98
POLS	5	rs28363340	0.11	0.99
POLS	5	rs2279655	0.11	0.95
POLS	5	rs2279653	0.23	0.98
CDC5L	6	rs2297334	0.46	1.00
CDC5L	6	rs2297333	0.23	0.96
CDC5L	6	rs9369480	0.05	0.98
CDC5L	6	rs12527053	0.13	0.99
CDC5L	6	rs525043	0.22	0.88
CDC5L	6	rs9381324	0.19	0.99
CDC5L	6	rs992160	0.09	1.00
FKBP5	6	rs1360780	0.23	0.98
FKBP5	6	rs7751693	0.07	0.99
SLC29A1	6	rs1057985	0.45	0.98
SLC29A1	6	rs3778504	0.04	0.99
SLC29A1	6	rs693955	0.30	0.99
SLC29A1	6	rs760370	0.26	0.97
ABCB1	7	rs2032582	0.49	0.96
ABCB1	7	rs1045642	0.36	0.87
NT5C3	7	rs10280692	0.47	0.96
NT5C3	7	rs4720100	0.13	0.96
NT5C3	7	rs12668520	0.31	0.89
NT5C3	7	rs3750117	0.44	0.92
RRM2B	8	rs2853228	0.46	0.95
RRM2B	8	rs2607660	0.11	0.98
RRM2B	8	rs2853229	0.50	0.96
SLC28A3	9	rs56350726	0.09	0.98
SLC28A3	9	rs10868138	0.08	0.99
TLE4	9	rs2807312	0.17	0.97
TLE4	9	rs10125657	0.13	1.00
TLE4	9	rs12337256	0.15	0.95
TLE4	9	rs11138328	0.34	0.99
TLE4	9	rs7039267	0.41	0.90
RRM1	11	rs183484	0.40	0.97
RRM1	11	rs9937	0.37	0.76
AICDA	12	rs3794318	0.21	0.98
AICDA	12	rs2580874	0.16	0.99
AICDA	12	rs2518144	0.14	0.92
AICDA	12	rs2580873	0.28	0.99
AICDA	12	rs2028373	0.45	0.96
AICDA	12	rs11046349	0.12	0.93
ESR2	14	rs944459	0.21	0.94
ESR2	14	rs944050	0.34	0.98
ESR2	14	rs1256064	0.41	0.97
ESR2	14	rs1256061	0.48	0.96
ESR2	14	rs944461	0.27	0.96
ESR2	14	rs1952585	0.15	0.97
ESR2	14	rs8003490	0.08	0.98
ESR2	14	rs1256033	0.45	0.97
ESR2	14	rs3783736	0.27	0.93
ESR2	14	rs12436302	0.06	0.99
ESR2	14	rs3020449	0.37	0.91
ESR2	14	rs3020445	0.30	0.95
ESR2	14	rs1256065	0.35	0.95
ESR2	14	rs1048315	0.34	0.96
ESR2	14	rs2772163	0.18	0.91
SLC28A1	15	rs2290272	0.44	0.97
SLC28A1	15	rs8187758	0.34	0.94
SLC28A1	15	rs3803390	0.49	0.97
SLC28A1	15	rs2305367	0.12	0.96
SLC28A1	15	rs2242047	0.31	0.93
SLC28A1	15	rs2242046	0.10	0.98
MYBBP1A	17	rs3809849	0.23	1.00
TYMS	18	^†	^†	1.00

^†Copy number variation of a 28-bp repeated sequence.Ch: Chromosome number; MAF: Minor allele frequency.

Table 3. Associations between genetic polymorphisms and overall survival ^†.
Genotype	n	Median, months (95% CI)	Additive		Dominant		Recessive p-value
			HR (95% CI)	p-value	HR (95% CI)	p-value
CMPK1 rs35687416
TT + GT	17	6.8 (4.1–9.4)	2.18 (1.32–3.97)	0.003	2.15 (1.22–3.81)	0.009	0.005
GG	80	10.3 (8.1–12.5)	2.41 (1.39–4.18)^‡	0.001^‡	2.29 (1.30–4.08)^‡	0.005^‡	0.010^‡
CMPK1 rs1044457
TT + CT	37	8.5 (4.9–12.1)	1.60 (1.05–2.42)	0.029	1.56 (0.98–2.50)	0.042	0.068
CC	64	11.1 (8.3–13.9)	1.90 (1.26–2.87)^‡	0.002^‡	2.07 (1.27–3.36)^‡	0.003^‡	0.128^‡
DCTD rs12507552
TT + CT	23	9.1 (7.1–11.0)	1.83 (1.15–2.90)	0.010	2.09 (1.23–3.56)	0.006	0.797
CC	70	12.7 (7.9–17.5)	2.53 (1.51–4.25)^‡	<0.001^‡	2.60 (1.50–4.50)^‡	<0.001^‡	0.167^‡
POLS rs274713
TT + GT	54	10.0 (8.4–11.6)	1.39 (1.00–1.94)	0.051	1.20 (0.75–1.92)	0.452	0.003
GG	46	11.8 (8.5–15.0)	1.49 (1.04–2.15)^‡	0.031^‡	1.11 (0.69–1.81)^‡	0.661^‡	<0.001^‡
POLS rs274717
GG + AG	68	11.8 (8.8–14.8)	0.70 (0.50–1.00)	0.050	0.57 (0.35–0.91)	0.020	0.462
AA	33	8.2 (5.5–10.9)	0.71 (0.50–1.00)^‡	0.050^‡	0.61 (0.37–1.00)^‡	0.050^‡	0.245^‡
CDC5L rs992160
TT + CT	18	20.1 (7.4–NA^§)	0.55 (0.28–1.07)	0.079	0.49 (0.24–0.99)	0.046	0.269
CC	84	9.8 (8.5–11.0)	0.49 (0.25–0.96)^‡	0.039^‡	0.44 (0.21–0.89)^‡	0.022^‡	0.434^‡

Three genetic models – additive, dominant and recessive – were used to assess the differences in overall survival.
^†The numbers of genotype and median months of overall survival based on the classification of dominant homozygote and other genotype, and p-values according to each genetic model, are listed.
^‡Multivariate HR and p-value. §Upper confidence interval was not estimated in our data.
HR: Hazard ratio; NA: Not applicable.

Table 4. Associations between genetic polymorphisms and time to progression ^†.
Genotype	n	Median, months (95% CI)	Additive		Dominant		Recessive p-value
			HR (95% CI)	p-value	HR (95% CI)	p-value	Recessive p-value
SH2D5 rs10916852
AA + CA	25	6.8 (0.3–13.4)	0.61 (0.38–1.00)	0.051	0.58 (0.35–0.98)	0.041	0.806
CC	74	5.4 (3.7–7.1)	0.57 (0.35–0.94)^‡	0.028^‡	0.54 (0.32–0.91)^‡	0.022^‡	0.720^‡
CMPK1 rs3925058
GG + AG	62	5.4 (3.9–6.9)	1.38 (1.00–1.91)	0.051	1.72 (1.07–2.74)	0.023	0.672
AA	36	5.9 (4.3–7.6)	1.38 (0.99–1.94)^‡	0.059^‡	1.73 (1.08–2.75)^‡	0.022^‡	0.826^‡
CMPK1 rs7543016
GG + CG	58	4.4 (2.6–6.3)	1.35 (0.98–1.87)	0.065	1.78 (1.12–2.82)	0.014	0.884
CC	38	6.3 (3.9–8.7)	1.41 (1.00–1.97)^‡	0.047^‡	2.01 (1.25–3.24)^‡	0.004^‡	0.560^‡
CMPK1 rs35687416
TT + GT	17	3.9 (1.6–6.2)	1.90 (1.14–3.19)	0.014	1.85 (1.07–3.18)	0.026	0.053
GG	80	5.9 (5.1–6.8)	1.94 (1.16–3.24)^‡	0.011^‡	1.91 (1.11–3.30)^‡	0.019^‡	0.099^‡
CMPK1 rs1044457
TT + CT	37	3.9 (1.5–6.2)	1.67 (1.16–2.42)	0.006	1.81 (1.18–2.79)	0.007	0.283
CC	64	6.0 (4.2–7.8)	1.84 (1.28–2.66)^‡	0.001^‡	2.15 (1.38–3.34)^‡	<0.001^‡	0.390^‡
SLC29A1 rs3778504
AA + GA	9	8.3 (0.1–16.5)	0.34 (0.15–0.82)	0.015	0.34 (0.15–0.82)	0.015	–
GG	92	5.4 (4.1–6.8)	0.56 (0.23–1.38)^‡	0.208^‡	0.56 (0.23–1.38)^‡	0.208^‡	–
ESR2 rs944050
GG + AG	58	5.2 (3.5–6.9)	1.47 (1.03–2.09)	0.034	1.37 (0.88–2.14)	0.160	0.020
AA	42	6.0 (3.3–8.6)	1.72 (1.21–2.43)^‡	0.002^‡	1.76 (1.12–2.78)^‡	0.014^‡	0.012^‡

Three genetic models – additive, dominant and recessive – were applied to estimate the duration to tumor progression.
^†The numbers of genotype and median months of time to progression based on the classification of dominant homozygote and other genotype, and p-values according to each genetic model, are listed.
^†Multivariate HR and p-value.
HR: Hazard ratio.

Table 5. Associations between genetic polymorphisms and tumor progression.
Genotype	Total (n)	Progressive, n (%)	Nonprogressive, n (%)	Additive p-value	Dominant p-value	Recessive p-value
CMPK1 rs1044457
TT	3	2 (6.7)	1 (1.4)	0.015	0.040	0.209
TC	34	14 (46.7)	20 (28.2)	0.017^†	0.014^†	0.391^†
CC	64	14 (46.7)	50 (70.4)
SLC28A1 rs2242047
TT	9	4 (16.0)	5 (7.1)	0.020	0.035	0.236
TC	41	14 (56.0)	27 (38.6)	0.056^†	0.033^†	0.466^†
CC	45	7 (28.0)	38 (54.3)

Three genetic models – additive, dominant and recessive – were used to quantify the genetic associations.
^†Multivariate p-value, analyzed with clinical factors including age, diabetes, prechemotherapeutic CA19-9 level and tumor status.

Table 6. Haplotype and tumor progression.
Gene	Haplotype	Progressive, n (%)	Nonprogressive, n (%)	p-value^†
CMPK1	GGCCTT^‡	10 (16.7)	9 (6.3)	0.032
DCTD	CGATAACA	5 (8.3)	2 (1.4)	0.024
TLE4	CTCGA	6 (10.0)	1 (0.7)	0.003 (0.025)^§
SLC28A1	GCCACG	1 (1.7)	19 (13.2)	0.009
SLC29A1, SLC28A3, SLC28A1	TGAAATGCTATG	10 (16.7)	5 (3.5)	0.002
CMPK1, DCK	GGCCGTGT	9 (15.0)	7 (4.9)	0.021
CDA, DCTD	CTCAACCGATAACA	5 (8.3)	1 (0.7)	0.009
Total		60	144

^†Fisher's exact test.
^†Haplotype sequence was designated as the order of SNPs in Table 1. For example, GGCCTT haplotype in CMPK1 gene means G allele in rs3925058, G allele in rs7543016, C allele in rs4600090, C allele in rs9436883, T allele in rs35687416 and T allele in rs1044457. GGCCGTGT haplotype in CMPK1 and DCK genes, G allele in rs3925058, G allele in rs7543016, C allele in rs4600090, C allele in rs9436883, G allele in rs35687416, T allele in rs1044457, G allele in rs7684954 and T allele in rs4694362.
^§False-discovery rate p-value < 0.05.

Comments

Commenting is limited to medical professionals. To comment please Log-in.

Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.