Association between glutathione S-transferase π polymorphisms and survival in patients with advanced nonsmall cell lung carcinoma

Authors

  • Charles Lu M.D.,

    Corresponding author
    1. Department of Thoracic/Head and Neck Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    • Department of Thoracic/Head and Neck Medical Oncology, Unit 432, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030-4009
    Search for more papers by this author
    • Fax: (713) 796-8655

    • Charles Lu is a recipient of the Clinical Oncology Research Career Development Award from the National Cancer Institute.

  • Margaret R. Spitz M.D., M.P.H.,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Hua Zhao Ph.D.,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Qiong Dong M.S.,

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Mylene Truong M.D.,

    1. Department of Diagnostic Radiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Joe Y. Chang M.D.,

    1. Department of Radiation Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • George R. Blumenschein Jr. M.D.,

    1. Department of Thoracic/Head and Neck Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Waun K. Hong M.D.,

    1. Department of Thoracic/Head and Neck Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Xifeng Wu M.D., Ph.D.

    1. Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author

Abstract

BACKGROUND

Glutathione S-transferase (GST) π (GSTP1) is a detoxification enzyme with substrate specificity for both exogenous carcinogens and chemotherapy agents. Genetic polymorphisms of GSTP1 exon 5 (Ile105Val) and exon 6 (Ala114Val) appear to reduce this enzyme's activity. Previously, the authors reported that the exon 6 variant was associated with an increased risk of lung carcinoma, particularly among men, younger patients, and ever smokers. In this study, the authors hypothesized that variant GSTP1 genotype would result in reduced inactivation of chemotherapy agents and improved survival in patients with advanced-stage nonsmall cell lung carcinoma (NSCLC), a population that is likely to receive platinum-based chemotherapy.

METHODS

Patients with Stage III and IV NSCLC who were enrolled in a molecular epidemiology study were identified, and a polymerase chain reaction-restriction fragment length polymorphism assay was used to genotype GSTP1 exons 5 and 6 in 424 patients and 425 patients, respectively.

RESULTS

Patients who had the exon 6 variant genotype (Ala/Val or Val/Val) had significantly better survival compared with patients who had the wild type genotype (Ala/Ala; P = 0.037), with median survival of 16.1 months and 11.4 months, respectively. Multivariate analysis revealed a reduced adjusted hazard ratio (HR) of death associated with the exon 6 variant genotype of 0.75 (95% confidence interval [95% CI], 0.54–1.05). This protective association was observed in younger patients (younger than age 62 yrs; HR, 0.59; 95% CI, 0.57–0.97) and in males (HR, 0.64; 95% CI, 0.41–0.99). GSTP1 exon 5 genotype was not associated with survival.

CONCLUSIONS

GSTP1 exon 6 variant genotypes may be associated with improved survival among patients with Stage III and IV NSCLC. Cancer 2006. © 2005 American Cancer Society.

Lung carcinoma remains a worldwide public health issue of immense proportions. In 2004, carcinomas of the lung and bronchus are expected to continue to account for the highest proportion of cancer deaths in the United States (160,440 deaths or 28.5%), more than the estimated total number of deaths due to carcinomas of the breast, prostate, colon, and rectum combined.1 Approximately 80% of lung carcinomas will have nonsmall cell carcinoma histology.2

The majority of patients with nonsmall cell lung carcinoma (NSCLC) present with locally advanced or metastatic disease3: In this population, standard treatment strategies include platinum-based chemoradiotherapy4 and chemotherapy,5 respectively. Despite the use of these aggressive and toxic therapies, patients with NSCLC have poor overall outcomes.6, 7 In the absence of validated biomarkers that can be used reliably to predict response to therapy, oncologists lack the tools needed to individualize and optimize patient therapies.8, 9

Glutathione S-transferases (GSTs) are Phase II metabolic enzymes that are involved in the detoxification of mutagenic and cytotoxic, DNA-reactive molecules mediated by glutathione conjugation.10, 11 Various cytosolic GST subclasses have been classified according to their biochemical properties, including GST α, GST μ (GSTM), GST π (GSTP1), and GST θ (GSTT).12 GSTP1 is expressed in many human epithelial tissues and is the most abundant GST isoform in the lung.13, 14GSTP1 is a polymorphic gene located on chromosome 11 with 2 single-nucleotide substitutions in exon 5 and exon 6 that give rise to Ile105Val and Ala114Val amino acid substitutions, respectively.15 These amino acid substitutions appear to be within the GSTP1 active site for binding of hydrophobic electrophiles,15, 16 and studies suggest that the variant genotype results in diminished enzymatic activity.17, 18 Because reduced enzyme activity may lead to a decreased ability to detoxify carcinogenic and mutagenic compounds, it is plausible that these polymorphisms may confer an increase in cancer susceptibility. A number of epidemiologic studies have explored possible associations between GSTP1 polymorphisms and the risk of lung carcinoma, often yielding conflicting results.19–22 Our group recently reported the largest case–control study to examine the association of GSTP1 exon 5 and 6 polymorphisms and NSCLC risk.23 That study included 582 patients (cases) and 600 matched controls. The primary finding was that the exon 6 polymorphism was associated with an elevated risk of NSCLC (odds ratio, 1.4) that was not observed with the exon 5 polymorphism. It is noteworthy that subgroup analyses suggested that men, younger individuals, and ever smokers with the exon 6 polymorphism had the highest associated risk.

In addition to carcinogens, a number of chemotherapeutic agents appear to be substrates for GSTP1, including cisplatin.24–27 Several studies have reported that elevated levels of GSTP1 in serum or tumor tissue from patients with NSCLC were correlated with resistance to platinum-based chemotherapy.28–31 Building on the results of our case–control study, we hypothesized that the GSTP1 exon 6 variant genotype would be associated with improved clinical outcome in the case group of patients with NSCLC who were treated with chemotherapy because of reduced chemotherapy inactivation. Because it is the standard of care for patients with locally advanced and metastatic NSCLC (Stages III and IV32) to receive platinum-based chemotherapy, we restricted our analysis to this patient population.

MATERIALS AND METHODS

Patients

The patients for this study were accrued from an ongoing, hospital-based, case–control study of epidemiologic and genetic risk factors for the development of lung carcinoma, as described previously.33 The overall participation rate was 86%. Among the patients who were enrolled between August 1995 and June 2000, 425 patients were identified with Stage III or IV, pathologically confirmed NSCLC diagnosed within the 4 months prior to enrollment. Follow-up was complete through September 17, 2001. All patients were accrued at the University of Texas M. D. Anderson Cancer Center (Houston, TX). On entry into the study, each patient had a personal interview based on a structured questionnaire in which they were asked to provide information that included sociodemographic variables and smoking history. An individual who had smoked at least 100 cigarettes was defined as an ever smoker. A former smoker had quit smoking at least 1 year prior to diagnosis. Ethnicity was self-reported. Each patient had a 30-mL sample of blood drawn into coded, heparinized tubes for immediate DNA isolation. All patients provided written informed consent, and the protocol was approved by The University of Texas M. D. Anderson Cancer Center Institutional Review Board. The American Joint Committee on Cancer staging system was used. Vital status was obtained from the medical record, the institution's tumor registry, or the Social Security Death Index.

GSTP1 Genotyping

Genomic DNA was isolated from the lymphocytes of whole blood samples using a proteinase-K, sodium dodecyl sulfate, and ethylene diamine tetraacetic acid-Tris approach. The polymorphic sites in exon 5 (Ile105Val) and exon 6 (Ala114Val) of the GSTP1 gene were amplified by using polymerase chain reaction (PCR) analysis, as described previously.23 Briefly, PCR was performed in a 25 μL volume that contained 1 × PCR buffer, 3.0 mM MgCl2, 0.25 mM dioxyribonucleoside triphosphates, 1.5 units of Taq polymerase (Promega, Madison, WI), and 0.3 μM of primers GSTP1-5 forward (5′-GTAGTTTGCCCAAGGTCAAG-3′) and GSTP1-5 reverse (5′-AGCCACCTGAGGGGTAAG-3′) for exon 5 or primers GSTP1-6 forward (5′-GGGAGCAAGCAGAGGAGAAT-3′) and GSTP1-6 reverse (5′-CAGGTTGTAGTCAGCGAAGGAG-3′) for exon 6. The PCR conditions were 94 °C for 5 minutes, followed by 5 cycles in which the annealing temperature decreased by 1 °C each cycle (cycle 1: 30 sec at 94 °C, 30 sec at 64 °C, 30 sec at 72 °C). This was followed by another 25 cycles at 94 °C for 30 seconds, 59 °C for 30 seconds, and 72 °C for 30 seconds, with a final extension step at 72 °C for 5 minutes. A 433-base pair (bp) DNA fragment was amplified for exon 5, and a 420 bp fragment was amplified for exon 6; this was followed by overnight digestion with 5 units of BsmAI for exon 5 and AciI for exon 6 (New England Biolabs, Beverly, MA). The fragments were separated on a 3% agarose gel stained with ethidium bromide. The wild type (AA), heterozygous genotype (A/G), and mutant genotype (G/G) yielded 2 bands (328 bp and 105 bp), 4 bands (328 bp, 222 bp, 106 bp, and 105 bp), and 3 bands (222 bp, 106 bp, and 105 bp), respectively, for the polymorphism in exon 5. For exon 6, the digested product resolved bands at 246 bp, 116 bp, and 58 bp for the wild type (C/C); 362 bp, 246 bp, 116 bp, and 58 bp for the heterozygous genotype (C/T); and 362 bp and 58 bp for the mutant genotype (T/T).

Statistical Analysis

Chi-square analyses and Fisher exact tests were used to compare the distribution of genotypes between subgroups based on gender, ethnicity, smoking status, age, disease stage, and tumor histology. Survival analysis methods were used to evaluate the effects of the GSTP1 exon 5 and exon 6 genotypes on survival in patients with lung carcinoma. We calculated survival as the period from diagnosis to the date of death or the date of last follow-up for each patient. We calculated person-years at risk within each genotype category as the sum of the survival of all patients in that category. Overall survival in relation to GSTP1 exon 5 and exon 6 genotypes was evaluated by the Kaplan–Meier survival function and log-rank tests. Hazard ratios (HRs) were estimated from a multivariate Cox proportional hazards model, with adjustment for age, gender, ethnicity, smoking status, and tumor stage. Stratified analysis was also carried out according to these adjusted variables. STATA software (STATA Corp., College Station, TX) was used for statistical analyses.

RESULTS

Among the 425 patients in this study, 55.5% were male, and 53.2% had Stage III disease, and their mean age was 61.6 years. The distributions of GSTP1 genotypes according to patient characteristics are shown in Table 1. Genotypes for GSTP1 exon 5 and exon 6 were available for 424 patients and 425 patients, respectively. The distribution of genotypes was not related to age, gender, smoking status, stage at diagnosis, or tumor histology. There was a significant association between exon 5 genotype and ethnicity: Hispanics had a higher frequency of the homozygous exon 5 variant allele (Val/Val). In light of the multiple subgroup comparisons and the small number of Hispanic patients (n = 18 patients), the importance of this observation remains to be determined. Because there were only 3 patients who were homozygous for the variant exon 6 allele (Val/Val), this group was combined with the heterozygous (Ala/Val) group for all future analyses.

Table 1. Selected Characteristics of Patients with Lung Carcinoma by Glutathione S-Transferase π Genotype
CharacteristicNo. of patients (%)
All patientsGSTP1 exon 5aGSTP1 exon 6
Ile/IleIle/ValVal/ValP valueAla/AlaAla/ValVal/ValP value
  • GSTP1: glutathione s-transferase π; CA: carcinoma.

  • a

    GSTP1 exon 5 genotype data were available for 424 patients.

Total425166 (39.2)207 (48.8)51 (12.0) 358 (84.2)64 (15.1)3 (0.7) 
Gender         
 Men236 (55.5)94 (56.6)111 (53.6)30 (58.8) 196 (54.8)37 (57.8)3 (100.0) 
 Women189 (44.5)72 (43.4)96 (46.4)21 (41.2)0.74162 (45.3)27 (42.3)0 (0.00)0.35
Ethnicity         
 Caucasian360 (84.7)146 (88.0)176 (85.0)37 (72.6) 296 (82.7)61 (95.3)3 (100.0) 
 Mexican American18 (4.2)4 (2.4)8 (3.9)6 (11.8) 17 (4.8)1 (1.6)0 (0.00) 
 African American47 (11.1)16 (9.6)23 (11.1)8 (15.7)0.0345 (12.5)2 (3.1)0 (0.00)0.11
Smoking status         
 Never44 (10.4)18 (10.8)24 (11.6)2 (3.9) 38 (10.6)6 (9.4)0 (0.00) 
 Former185 (43.5)71 (42.8)90 (43.5)23 (45.1) 151 (42.2)32 (50.0)2 (66.7) 
 Current196 (46.1)77 (46.4)93 (44.9)26 (51.0)0.60169 (47.2)26 (40.6)1 (33.3)0.74
Age         
 ≤ 49 yrs47 (11.1)15 (9.0)29 (14.0)3 (5.9) 36 (10.1)11 (17.2)0 (0.00) 
 50–59 yrs118 (27.8)44 (26.5)63 (30.4)11 (21.6) 98 (27.4)18 (28.1)2 (66.7) 
 60–69 yrs171 (40.2)66 (40.0)78 (37.7)26 (51.0) 146 (40.8)25 (39.1)0 (0.00) 
 ≥ 70 yrs89 (20.9)41 (24.7)37 (17.9)11 (21.6)0.1978 (21.8)10 (15.6)1 (33.3)0.28
Stage at diagnosis         
 Stage III226 (53.2)83 (50.0)111 (53.6)31 (60.8) 187 (52.2)39 (60.9)0 (0.00) 
 Stage IV199 (46.8)83 (50.0)96 (46.4)20 (39.2)0.39171 (47.8)25 (39.1)3 (100.0)0.07
Histology         
 Adeno CA215 (50.5)82 (49.4)110 (53.4)23 (45.1) 181 (50.7)31 (48.4)3 (100.0) 
 Squamous CA106 (24.9)45 (27.1)51 (24.8)15 (29.4) 86 (24.1)20 (31.3)0 (0.00) 
 Other103 (24.2)39 (23.5)45 (21.8)13 (25.5)0.6890 (25.2)13 (20.3)0 (0.00)0.45

The Kaplan–Meier survival functions for overall survival according to GSTP1 genotypes are presented in Figure 1. In total, 312 deaths were observed during follow-up. The median follow-up among patients who were alive at the end of observation was 18 months after diagnosis. Figure 1A shows that the exon 6 polymorphism was associated with improved overall survival (log-rank P = 0.037), with a median survival of 16.1 months and 11.4 months for patients with and without the variant allele, respectively. The Kaplan–Meier estimate of overall survival at 1 year, 2 years, and 3 years was 0.64 (95% CI, 0.50–0.75), 0.32 (95% CI, 0.20–0.45), and 0.21 (95% CI, 0.09–0.36) for patients with the GSTP1 variant genotype and 0.47 (95% CI, 0.41–0.52), 0.22 (95% CI, 0.18–0.27), and 0.13 (95% CI, 0.09–0.18) for patients with the wild type genotype. When they were stratified by disease stage, patients who had the exon 6 variant demonstrated improved survival, but the association was no longer statistically significant (log-rank P = 0.14 and P = 0.19 for the Stage III and IV subgroups, respectively). There was no evidence of a survival difference by GSTP1 exon 5 genotype (Fig. 1B) (log-rank P = 0.35).

Figure 1.

Overall survival is illustrated for patients with nonsmall cell lung carcinoma according to glutathione S-transferase π (GSTP1) genotype (Kaplan–Meier function. (A) The GSTP1 exon 6 genotype. (B) The GSTP1 exon 5 genotype. Numbers in parentheses represent the number of deaths/the number of patients.

The adjusted HRs of death associated with GSTP1 genotypes are shown in Table 2. After adjustment for age, gender, ethnicity, smoking status, disease stage, and tumor histology, patients who had the GSTP1 exon 6 variant genotype had marginally longer survival, with an adjusted HR of 0.75 (95% CI, 0.54–1.05), compared with patients who had the wild type genotype. In stratified analyses, the protective effect was evident among younger patients (younger than age 62 yrs), with an adjusted HR of 0.59 (95% CI, 0.57–0.97), but it was less evident among older patients (age 62 yrs and older; adjusted HR, 0.89; 95% CI, 0.56–1.42). The protective effect also was evident in men (adjusted HR, 0.64; 95% CI, 0.41–0.99) but not in women (adjusted HR, 0.93; 95% CI, 0.54–1.60). For the GSTP1 exon 5 variant genotype, the adjusted Cox model indicated no significant association with overall survival.

Table 2. Glutathione S-Transferase π Genotypes in Patients with Lung Carcinoma
GenotypeSurvival status (no. of patients)HR (95% CI)a
DeadAlive
  • HR: hazard ratio; 95% CI: 95% confidence interval; GSTP1: glutathione S-transferase π.

  • a

    HR and 95% CI values were determined by using a Cox proportional hazards model that was adjusted for age, gender, ethnicity, smoking status, tumor stage, and histology.

GSTP1 exon 5   
 Ile/Ile12046Reference
 Ile/Val158491.24 (0.97–1.58)
 Val/Val34170.88 (0.60–1.30)
GSTP1 exon 6   
 Ala/Ala26989Reference
 Ala/Val + Val/Val43240.75 (0.54–1.05)

Patients with Stage III and IV NSCLC were selected for this study, because the accepted standard of care for these patients incorporates platinum-based chemotherapy for those who are deemed eligible for therapy. Although baseline patient data and blood specimens were collected meticulously for patients who were enrolled in this prospective molecular epidemiologic study, subsequent clinical follow-up and treatment data remained incomplete. Initial treatment-related data, however, were available for a subgroup of 310 patients (73%). In this subgroup, 204 patients received chemotherapy as a component of their initial therapy, and 106 patients did not receive chemotherapy. Among the former group, there appeared to be a trend toward improved survival for patients who had the variant exon 6 allele (log-rank P = 0.13), whereas this trend was not evident in the latter group (log-rank P = 0.88). A similar exploratory analysis for the exon 5 polymorphism did not demonstrate any associations with survival among patients who received or did not receive initial chemotherapy (log-rank P = 0.74 and P = 0.80, respectively). A possible factor that may have influenced the availability of initial treatment data was the site of treatment, because it was more difficult to obtain this information for patients who were not treated at our institution.

Data regarding whether or not patients ever received chemotherapy were available for 423 patients (99.5%). Two hundred thirty-seven patients received chemotherapy, whereas 186 patients never received chemotherapy. Among the former group, there appeared to be a nonsignificant trend toward improved survival for patients who had the variant exon 6 allele (log-rank P = 0.17), whereas this trend was not evident in the latter group (log-rank P = 0.46). A similar exploratory analysis for the exon 5 polymorphism did not demonstrate significant associations with survival among patients who received or did not receive chemotherapy (log-rank P = 0.99 and P = 0.17, respectively).

DISCUSSION

Based on data indicating that variant alleles are associated with reduced enzymatic activity, we hypothesized that GSTP1 genotype also would affect clinical outcome in the patients with NSCLC who were most likely to receive chemotherapy. Because it is the standard of care for patients with advanced NSCLC (Stage III and IV) to receive platinum-based chemotherapy, we decided to focus our analysis on this population. We found that the exon 6 polymorphism was associated with improved survival. The adjusted HR similarly indicated a reduced risk of death, although the 95% CI included the null. It is noteworthy that our subset analyses suggested that this association with survival was restricted to younger patients and males, the same groups that demonstrated increased lung carcinoma risk with the variant genotype.

Other groups have investigated GSTP1 polymorphism in patients with lung carcinoma, although no studies have been reported with data on the exon 6 variant allele to our knowledge. Sweeney et al. examined 274 patients with lung carcinoma (all histologies) who were identified through a Washington State population-based cancer registry.34 Those authors found no association between the GSTP1 exon 5 (Ile105Val) or GSTT1 null genotype and survival in the overall population nor in a subgroup of patients who were treated with chemotherapy. An association between the GSTM1 null genotype and shorter survival was found; however, this effect was not modified by chemotherapy but appeared to be strongest among patients who received radiotherapy. The biologic significance of these findings was unclear. One possible explanation offered was that carcinomas in GSTM1-null patients may be more aggressive due to increased carcinogen-DNA adducts and mutations in p53, K-ras, and other genes.35 Yang et al. examined 250 patients with lung carcinoma in all stages of disease36 but found no association between GSTP1 exon 5 (Ile105Val), GSTT1 null, or GSTM1 null genotypes and 1-year survival. The lack of an association between GSTP1 exon 5 genotype and survival in patients with lung carcinoma appears to be a consistent finding in our study and in the 2 aforementioned reports. In a study of patients with advanced colorectal carcinoma who received oxaliplatin-based chemotherapy, however, the exon 5 variant allele was associated with improved survival.37

The current data also suggest that the exon 6 polymorphism is associated with a gender and age effect. The biologic significance of these findings remains unclear. It is intriguing that, in our prior case–control study, we also found an increased risk of lung carcinoma associated with the exon 6 variant allele that was especially evident in males and younger individuals and not in females and older individuals. The consistency of these results suggests that the exon 6 polymorphism may have a real biologic effect that can affect clinical outcome. It is noteworthy that gender has been identified as a prognostic factor in patients with NSCLC.38 and it is possible to speculate that gender-specific differences in metabolic enzymes involved in the detoxification of chemotherapy may be a possible explanation for this observation. These data must be interpreted with caution, however, because it remains possible that subgroup analyses may result in misleading artifactual findings secondary to multiple comparisons. Future prospective validation studies in other patient populations are needed.

There are some important limitations of this research. The patients in this study were part of a larger molecular epidemiologic study of risk factors for lung carcinoma development. Although blood samples and initial clinical, stage, and demographic data were collected prospectively, follow-up clinical and treatment information were not primary endpoints; thus, the data were incomplete, and the study's primary endpoint was overall survival. Therefore, we cannot conclude that the GSTP1 exon 6 polymorphism is associated with differences in response to chemotherapy. An alternative explanation is that the polymorphism is a favorable prognostic factor that influences clinical outcome independent of treatment status. We did explore this issue further by examining patients with early-stage NSCLC, because, during the study period, the standard of care for these patients was surgical resection alone. In a cohort of 162 patients with Stage I and II NSCLC, we did not detect an association between the exon 6 polymorphism and survival (data not shown). This finding would support the hypothesis that this polymorphism may function as a predictor of response to chemotherapy. Because, currently, there are no useful biomarkers to guide therapeutic decisions for patients with advanced NSCLC, the development of validated predictive biomarkers would have significant clinical importance.

In summary, the current results suggest that the GSTP1 exon 6 variant genotype is associated with improved survival among patients with Stage III and IV NSCLC. Further studies incorporating treatment and response data are needed to validate these findings and to determine the predictive and prognostic significance of GSTP1 genotype in patients with advanced NSCLC. Ultimately, future pharmacogenomic studies analyzing multiple functional polymorphisms will be required to identify biomarkers that will guide the selection and delivery of individualized therapy.

Ancillary