Fax: (011) 31-50-3614862
Association of PAI-1 gene polymorphism with survival and chemotherapy-related vascular toxicity in testicular cancer†
Article first published online: 24 AUG 2010
Copyright © 2010 American Cancer Society
Volume 116, Issue 24, pages 5628–5636, 15 December 2010
How to Cite
de Haas, E. C., Zwart, N., Meijer, C., Suurmeijer, A. J.H., Meijer, K., Guchelaar, H.-J., Hoekstra, H. J., van Leeuwen, F. E., Sleijfer, D. Th., Boezen, H. M. and Gietema, J. A. (2010), Association of PAI-1 gene polymorphism with survival and chemotherapy-related vascular toxicity in testicular cancer. Cancer, 116: 5628–5636. doi: 10.1002/cncr.25300
Presented in part as a poster at the 2009 ASCO Annual Meeting, Orlando, FL, May 29-June 2, 2009.
- Issue published online: 3 DEC 2010
- Article first published online: 24 AUG 2010
- Manuscript Accepted: 8 FEB 2010
- Manuscript Revised: 5 FEB 2010
- Manuscript Received: 4 NOV 2009
- testicular cancer;
- gene polymorphism;
- vascular toxicity
High Plasminogen-Activator Inhibitor 1 (PAI-1) expression by tumors has been associated with poor prognosis in several cancer types, and high systemic PAI-1 levels with increased thrombosis risk. The authors investigated whether the germline 4G/5G deletion/insertion polymorphism in the PAI-1 promoter (rs1799889), which may influence PAI-1 expression, is associated with survival and chemotherapy-related vascular toxicity in testicular cancer (TC).
Data were collected on PAI-1 4G/5G polymorphism, survival, venous thromboembolism (VTE), and coronary heart disease (CHD) for 324 non-seminomatous TC patients treated with platinum-based chemotherapy. Genotypes were compared regarding survival and disease outcome. VTE and CHD incidence were compared with adjustment for cardiovascular risk factors and prothrombotic gene polymorphisms of coagulation factors II/prothrombin (G20210A) and V (G1691A).
The 4G/4G variant of PAI-1 4G/5G polymorphism shows a higher prevalence of International Germ Cell Cancer Classification (IGCCC) poor prognosis compared with 4G/5G and 5G/5G (24% vs 8% and 15%; chi-square P = .003). In addition, the 4G/4G variant shows reduced TC-related survival with a hazard ratio of 2.69 (95% CI, 1.26-5.73; P = .010) for TC-related death (adjusted for IGCCC). This is related to an increased risk for refractory disease and early relapses (odds ratio, 3.35; 95% CI, 1.48-7.59; P = .004). PAI-1 4G/5G polymorphism is not associated with VTE and CHD risk.
The 4G/4G variant of PAI-1 4G/5G polymorphism may be an unfavorable prognostic as well as predictive factor for response to chemotherapy in TC patients. If confirmed, it may contribute to the identification of patients with increased risk for refractory disease. Cancer 2010. © 2010 American Cancer Society.
The introduction of cisplatin for the treatment of disseminated testicular cancer (TC) in the late 1970s and subsequent development of cisplatin-based regimens have led to considerable improvement in survival.1, 2 Current, long-term survival rates for this most common malignancy in adult young men exceed 80%.3
With improved survival, long-term treatment-related toxicity has emerged. Together with secondary malignancies, cardiovascular disease is one of the most important treatment-related risks in TC survivors, threatening life expectancy and quality of life.4-6 Chemotherapy-related vascular toxicity can occur as an acute as well as a long-term complication. During chemotherapy cardiovascular complications have been observed as venous thromboembolism (VTE) and arterial disease.7-9 With longer follow-up, an increased risk of coronary heart disease (CHD) appears to prevail.5, 10, 11
Attention is increasingly drawn to designing chemotherapy regimens that will improve prognosis for patients with unfavorable risk features as well as reduce chemotherapy-related toxicity.12-15 For both purposes, exploratory analyses of germline polymorphisms that influence chemotherapy efficacy and/or toxicity by involvement in metabolism and/or target pathways of cytotoxic drugs may contribute. We selected as gene of interest the gene for Plasminogen-Activator Inhibitor 1 (PAI-1).
PAI-1 is a major inhibitor of the urokinase plasminogen activator (uPA)/uPA receptor (uPAR) pathway, which is increasingly associated with tumor growth, invasion, and metastasis.16, 17 High expression of uPA and uPAR by tumors is related to poor prognosis in several cancer types.17, 18 Although PAI-1 might initially be expected to have an opposite effect to uPA, high expression of PAI-1 by tumors has been associated with poor prognosis.17
Moreover, PAI-1 is the most important circulating inhibitor of intravascular fibrinolysis. Binding of PAI-1 to the active site of endothelial-derived tissue-type Plasminogen Activator (t-PA) inhibits the conversion of plasminogen to the active protease plasmin and the subsequent plasmin-mediated degradation of fibrin.19, 20 Elevated PAI-1 plasma levels may contribute to the development of thrombosis and have been associated with both VTE and CHD.21-24
The PAI-1 gene promoter contains at -675 bp a common, single base pair deletion/insertion polymorphism, the 4G/5G polymorphism (PAI-1 -675 4G/5G, rs1799889).25 The 5G allele creates an additional binding site for a transcription inhibitor, which may lead to differences in PAI-1 expression.25, 26 The 4G allele has been associated with higher plasma PAI-1 levels, and with VTE and CHD.26-28 However, studies on the association of the PAI-1 4G/5G polymorphism with prognosis in different cancer types are inconclusive.17 So far, it is unknown whether PAI-1 expression and the PAI-1 4G/5G polymorphism are related to prognosis and survival in TC.
We investigated whether the PAI-1 4G/5G polymorphism is associated with differences in survival and incidence of chemotherapy-related cardiovascular disease in non-seminomatous TC patients treated with platinum-based chemotherapy. For the analysis of cardiovascular disease, the PAI-1 4G/5G polymorphism was combined with the less common, but potentially confounding, prothrombotic gene polymorphisms of coagulation factors: factor II/prothrombin (G20210A, rs1799963) and factor V (G1691A, also known as factor V Leiden, rs6025).29, 30
MATERIALS AND METHODS
Patients with non-seminomatous TC, who were treated with platinum-based chemotherapy at the University Medical Center Groningen between January 1977 and April 2004, were eligible. Patients with an extragonadal tumor were excluded. Data were collected on baseline characteristics, disease characteristics (International Germ Cell Consensus Classification (IGCCC) prognosis group, dissemination pattern, and tumor marker levels), chemotherapy regimen, disease outcome, cardiovascular risk factors, and development of cardiovascular disease (VTE and CHD).
DNA Collection and Genotyping
After informed consent, germline DNA was isolated from EDTAblood collected at the general practitioner's or at the outpatient clinic. For the patients who died, DNA was isolated from routinely stored serum. Serum-derived DNA was amplified and quality checked (REPLI-g Service; Qiagen, Hilden, Germany). Samples deemed ‘usable’ or ‘highly usable’ were used for genotyping. Genotypes were assessed with an allelic discrimination assay on an ABI Prism7900HT sequence detection system (Applied Biosystems [Life Technologies], Carlsbad, California). For analysis of the PAI-1 -675 4G/5G polymorphism, a 91 bp DNA fragment was created with a forward primer 5′-GCCAGACAAGGTTGTTGACACA-3′ and a reverse primer 5′-GCCGCCTCCGATGATACA-3′. The 4G allele was identified with a 5′-VIC-TCCCCACGTGTCCA-MGB-NFQ probe and the 5G allele with a 5′-FAM-CTCCCCCACGTGTC-MGB-NFQ probe (Applied Biosystems). Genotyping of factor II G20210A and factor V G1691A was performed with a validated TaqMan Genotyping Assay from Applied Biosystems (ID C_8726802_20 and ID C_11975250_10, respectively).
Assessment of Survival
The last follow-up date, including vital status and, if applicable, date and cause of death, were available from the medical record or the general practitioner's files. Survival time was calculated from start of chemotherapy to death or last follow-up date. TC-related death was defined as death due to TC and not to chemotherapy-induced toxicity. As described previously31, disease outcome was classified as refractory disease (persistence of tumor markers lacto-dehydrogenase (LDH), α-fetoprotein (αFP), and β-human chorionic gonadotropin (βHCG), or renewed elevation within 4 weeks after completion of chemotherapy); early relapse (within 2 years after start of treatment after an initial, complete response), late relapse (more than 2 years after start of treatment), or no evidence of disease (complete remission without relapse).
VTE (deep vein thrombosis and/or pulmonary embolism) occurring during chemotherapy was analyzed with adjustment for the potential confounding risk factors age at start of chemotherapy and IGCCC prognosis.
For analyzing CHD, all patients who experienced a myocardial infarction or coronary artery disease (proven by coronary angiography or by treatment with coronary angioplasty or coronary bypass surgery) during chemotherapy or follow-up were included. CHD before start of chemotherapy was an exclusion criterion. Data were collected on age and follow-up duration at diagnosis of CHD, and the presence or development of the following cardiovascular risk factors: positive family history for cardiovascular disease (a parent or brother/sister with proven coronary artery disease, myocardial infarction, cardiac death, or cerebrovascular accident before the age of 60), smoking status, hypercholesterolemia (non-fasting plasma cholesterol levels >6.5 mmol/L at ≥3 separate time points or use of cholesterol lowering medication), hypertension (blood pressure systolic >150 mm Hg and/or diastolic >95 mm Hg and/or use of antihypertensive medication), overweight (defined as body mass index [BMI] >27.8 kg/m2), and diabetes mellitus.
Hardy-Weinberg equilibrium of the genotypes was determined by using the chi-square test. Genotype groups were compared for patient characteristics, disease characteristics, and received chemotherapy regimen using the Mann-Whitney U and Kruskal-Wallis tests for continuous variables, as appropriate, and the chi-square test for categorical variables.
Overall and TC-related survival were analyzed with Kaplan-Meier curves and tested with the log-rank test. To adjust for potential confounding factors, multivariate Cox regression analysis was performed for TC-related survival and multiple logistic regression analysis for disease outcome.
The risk for VTE during chemotherapy was analyzed using logistic regression analysis. A Cox proportional hazard model was used to analyze the effect of the genotypes on CHD risk during chemotherapy and follow-up.
All statistical analyses were performed with SPSS for Windows 16.0 (SPSS, Chicago, IL).
Patient Characteristics and Genotypes
From a total cohort of 439 patients, DNA from 334 patients was available for genotyping. Fifty-two persons did not give informed consent. From 53 deceased patients, either no serum was available for DNAisolation (n = 48) or serum-isolated DNA was found to be ‘unusable’ after quality check by Qiagen REPLI-g Service (n = 5). PAI-1 4G/5G polymorphism was genotyped in 324 (97%) samples. Of these 324, factor II G20210A polymorphism was genotyped in 321 and factor V G1691A polymorphism in 320. Patient characteristics, disease characteristics, chemotherapy regimens, and genotype frequencies of the 324 patients are presented in Table 1. For 90% of the patient cohort, follow-up was complete till at least 2006.
|Patient characteristics||Median [range]|
|Age at start chemotherapy, y||28 [16-64]|
|Follow-up duration, y||10 [0-28]|
|Age at end of follow-up, y||40 [19-73]|
|Disease characteristics||No. (%)|
|Chemotherapy characteristics||No. (%)|
|PAI-1 4G/5G polymorphism|
|Factor II G20210A polymorphism|
|Wildtype, GG||314 (98)|
|Heterozygous variant, GA||7 (2)|
|Factor V G1691A polymorphism|
|Wildtype, GG||302 (94)|
|Heterozygous variant, GA||18 (6)|
PAI-1 4G/5G allele frequencies are in Hardy-Weinberg equilibrium with the following distribution: 84 (26%) 4G/4G, 164 (51%) 4G/5G, and 76 (23%) 5G/5G (P > .05). There was no apparent difference in distribution of the potential confounding polymorphisms of factor II and factor V over the PAI-1 genotype groups (chi-square test P > .05). Because none of the patients experienced CHD before start of chemotherapy, association of PAI-1 4G/5G polymorphism with survival and chemotherapy-related cardiovascular disease could be analyzed in 324 (74%) of the total cohort of 439 patients.
The 3 PAI-1 genotype groups did not differ with respect to age and received chemotherapy regimen. However, the 4G/4G genotype group contained more patients with an IGCCC poor prognosis, which appears to be largely attributable to high βHCG levels (Table 2).
|4G/4G (84)||4G/5G (164)||5G/5G (76)||Pa|
|Disease Characteristics||No. (%)||No. (%)||No. (%)|
|Good||39 (46%)||90 (55%)||42 (56%)||.359|
|Intermediate||25 (30%)||60 (37%)||22 (29%)||.385|
|Poor||20 (24%)||13 (8%)||11 (15%)||.003|
|Nonpulmonary visceral metastases||6 (7.1%)||4 (2.4%)||5 (6.6%)||.162|
|αFP, ug/L, n=322|
|Level, median [range]||89 [1-540,000]||34 [1-70,600]||22 [1-19,570]||.377|
|Poor risk||6 (7.2%)||3 (1.8%)||5 (6.7%)||.077|
|βHCG, IU/mL, n=321|
|Level, median [range]||78 [1-3,835,200]||68 [0-4,207,160]||37 [1-19,652,000]||.257|
|Poor risk||13 (15.7%)||7 (4.3%)||5 (6.7%)||.007|
|LDH, U/L, n=323|
|Level, median [range]||345 [99-3,393]||300 [130-18,613]||279 [170-3,204]||.059|
|Poor risk||2 (2.4%)||3 (1.8%)||2 (2.6%)||.910|
PAI-1 4G/5G Polymorphism and Survival
Of the 324 analyzed patients, 35 (10.8%) died. In 28 of 324 (8.6%) patients TC was the cause of death. Other causes of death were bleomycin-induced pulmonary toxicity (n = 1), infectious complications (n = 2), myocardial infarction (n = 1), ruptured abdominal aortic aneurysm (n = 1), pulmonary emphysema (n = 1), and unknown cause (n = 1).
The PAI-1 4G/5G polymorphism showed an association with overall as well as TC-related survival. Overall survival was worse for the 4G/4G genotype group, mainly because of worse TC-related survival within the first years after chemotherapy (Fig. 1). In the 4G/4G genotype 14 (16.7%) patients died of TC compared with 8 (4.9%) in the 4G/5G group and 6 (7.9%) in the 5G/5G group (log-rank test P = .003). TC-related survival was comparable for the 4G/5G and 5G/5G genotype, suggesting a recessive effect of the 4G allele. In multivariate Cox-regression analysis, with adjustment for the potential confounding variables age at start of chemotherapy and IGCCC prognosis group, both prognosis group and PAI-1 4G/5G polymorphism seemed to be independent predictive factors for TC-related survival. With adjustment for prognosis group, patients with the 4G/4G variant showed an increased risk for TC-related death with a hazard ratio (HR) of 2.69 (95% confidence interval [CI], 1.26-5.73; P = .010) compared with 4G/5G and 5G/5G variants combined (Table 3 and Fig. 2).
|Prognostic Factor||No.||Death of TC||HR||95% CI||P|
|4G/5G or 5G/5G||240||14||1||ref.||—|
Although residual confounding by prognosis group cannot be completely excluded, we performed additional analyses that showed the unfavorable effect of the 4G/4G variant on TC-related survival remains. We performed Cox-regression analysis, stratified for IGCCC prognosis group. In addition, multivariate Cox-regression analysis was performed with tumor markers LDH, aFP, and bHCG as continuous variables and the presence of nonpulmonary visceral metastases as dichotomous variable. Both analyses showed a comparable, independent, unfavorable effect of the 4G/4G variant: hazard ratio [HR] for TC-related death 2.66; 95% CI, 1.24-5.66; P = .012 and HR, 2.97; 95% CI, 1.37-6.42; P = .006, respectively. The unfavorable effect of the 4G/4G variant on TC-related survival did not show predominance for a particular IGCCC prognosis group, βHCG-producing disease, or the presence of choriocarcinoma (as βHCG-producing component) in the primary testicular tumor. We found no significant interaction between the PAI-1 4G/4G variant and IGCCC prognosis group, βHCG-producing disease, or choriocarcinoma (component) in the primary tumor (data not shown). Looking specifically at disease outcome after completion of chemotherapy, patients with 4G/4G genotype showed a higher prevalence of refractory disease and early relapses with an odds ratio (OR) of 3.35 (95%, CI, 1.48-7.59; P = .004) compared with patients with 4G/5G or 5G/5G genotype (adjusted for IGCCC prognosis group). The 4G/4G genotype showed no significant increased OR for late relapses(Table 4). After completion of first-line chemotherapy, 225 patients were treated with surgery for residual mass(es). Of the patients with the 4G/4G variant, 7 of 56 (12.5%) had viable tumor compared with 11 of 169 (6.5%) viable tumor in patients with 4G/5G or 5G/5G genotype (chi-square test, P = 0.162).
|Prognostic Factor||No. (%)||OR||95% CI||P|
|Refractory disease or early relapse <2 y|
|Good (171)||4 (2.3%)||1||ref.||—|
|Intermediate (107)||14 (13.1%)||6.50||2.06-20.56||.001|
|Poor (42)||12 (28.6%)||14.03||4.16-47.35||<.001|
|4G/5G or 5G/5G (240)||14 (5.8%)||1||ref.||—|
|4G/4G (82)||16 (19.5%)||3.35||1.48-7.59||.004|
|Late relapse >2 y|
|Good (171)||10 (5.8%)||1||ref.||—|
|Intermediate (107)||10 (9.3%)||1.66||0.67-4.13||.277|
|Poor (42)||4 (9.5%)||1.64||0.48-5.60||.429|
|4G/5G or 5G/5G (240)||17 (7.1%)||1||ref.||—|
|4G/4G (82)||7 (8.5%)||1.17||0.46-2.97||.743|
Factor II G20210A polymorphism and factor V G1691A polymorphism were not associated with overall and TC-related survival(data not shown).
A total of 26 (8.1%) patients developed VTE during chemotherapy, including deep vein thrombosis of a leg (n = 6), central venous access port-related deep vein thrombosis of an arm (n = 17), and pulmonary embolism (n = 3).
PAI-1 4G/5G polymorphism did not show an association with VTE. Potential confounding factor V G1691A polymorphism was not associated with VTE either, whereas the heterozygous variant of factor II G20210A polymorphism showed increased incidence (Table 5). When excluding central venous access port-related deep vein thrombosis of an arm, 1 of 7 (14.3%) patients with the heterozygous variant of factor II G20210A developed VTE during chemotherapy as did 8 of 314 (2.5%) patients with wildtype factor II G20210A (chi-square test P = .162).
|A. Venous thromboembolism|
|4G/4G (84)||7 (8.3%)||.867|
|4G/5G (164)||14 (8.5%)||—|
|5G/5G (76)||5 (6.6%)||—|
|Factor V G1691A polymorphism|
|Wild-type (302)||24 (7.9%)||.648|
|Heterozygous variant (18)||2 (11.1%)||—|
|Factor II G20210A polymorphism|
|Wild-type (314)||22 (7.0%)||.001|
|Heterozygous variant (7)||4 (57.1%)||—|
|B. Coronary heart disease|
|4G/4G (84)||5 (6.0%)||.594|
|4G/5G (164)||8 (4.9%)||—|
|5G/5G (76)||2 (2.6%)||—|
|Factor V G1691A polymorphismb|
|Wild-type (302)||11 (3.6%)||.037|
|Heterozygous variant (18)||3 (16.7%)||—|
|Factor II G20210A polymorphism|
|Wild-type (314)||14 (4.5%)||.287|
|Heterozygous variant (7)||1 (14.3%)||—|
When considering all VTE combined, both IGCCC prognosis and factor II G20210A polymorphism were independently associated with an increased risk for VTE. When combined in multiple logistic regression analysis, patients with an intermediate or poor prognosis had an odds ratio [OR] of 2.66 (95% CI,1.09-6.53; P = .032) compared with patients with a good prognosis. The presence of the heterozygous variant of factor II G20210A polymorphism led to an OR of 14.21 (95% CI, 2.57-78.78; P = .002) when compared with the wildtype.
Coronary heart disease
CHD was diagnosed in 15 (4.6%) patients after start of chemotherapy. Median age at CHD diagnosis was 51 years (range, 30-62 years) and median follow-up duration was 15 years (range, 0-28 years). Although the incidence of CHD was relatively higher in the 4G/4G and 4G/5G variants of PAI-1 4G/5G polymorphism (6.0% and 4.9% compared with 2.6% in the 5G/5G variants), these differences were not significant(Table 5).
In multivariate Cox proportional hazard analysis with cardiovascular risk factors (age, positive family history, smoking status, hypercholesterolemia, hypertension, overweight, and diabetes mellitus), there was no significant effect of PAI-1 4G/5G polymorphism on CHD risk (4G/4G HR, 6.79; 95% CI, 0.54-84.94; P = .137; and 4G/5G HR, 2.43; 95% CI, 0.26-22.78; P = .436). The heterozygous variant of factor V G1691A polymorphism showed increased risk for the development of CHD with an HR of 6.25 (95% CI, 1.12-34.93; P = .037). Factor II G20210A polymorphism did not affect CHD risk (heterozygous variant HR, 2.56; 95% CI, 0.26-24.73; P = .416, based on 1 patient with CHD).
PAI-1 may have a pleiotropic role in cancer patients. On the one hand, PAI-1 is increasingly recognized as a key player in cancer progression as well as a potential therapeutic target.16-18 Conversely, PAI-1 is a prothrombotic factor that may contribute to cardiovascular complications. We investigated whether the PAI-1 4G/5G polymorphism, which may influence PAI-1 expression, is associated with survival and chemotherapy-related cardiovascular disease in non-seminomatous TC patients treated with platinum-based chemotherapy.
The current results show that the homozygous 4G/4G variant of the PAI-1 4G/5G deletion/insertion polymorphism is associated with reduced TC-related survival. This unfavorable effect of the 4G allele appears to be recessive and at least partially independent from the commonly used prognosis classification (IGCCC). The increased risk for TC-related death (HR 2.69 (1.26-5.73)) may be related to an increased risk for refractory disease. Remarkably, the PAI-1 4G/5G polymorphism is also associated with IGCCC prognosis group. Patients with the 4G/4G variant more often belong to the poor prognosis group, which suggests more aggressive tumor behavior.
Available data on PAI-1 and cancer prognosis are most consistent with respect to expression data. High expression of PAI-1 in tumor tissue has been associated with poor prognosis in breast cancer, cervical cancer, and kidney cancer.32-35 Although the exact mechanism behind PAI-1's effect on prognosis is unknown, PAI-1 is increasingly recognized as a multifunctional protein with several tumor-promoting features, with various effects on cell adhesion, migration, and invasion to roles in apoptosis and angiogenesis.16, 17, 36, 37
However, studies on the association of the germline PAI-1 4G/5G polymorphism with prognosis are inconclusive.17 The 5G allele contains an additional protein binding site, which may bind a repressor to the PAI-1 promoter and, thus, lead to reduced PAI-1 transcription. Consequently, some data suggest the 4G allele is associated with a higher basal transcription of the PAI-1 gene26, whereas other data suggest an increased transcription in response to an acute phase stimulus25. Whether the germline PAI-1 4G/5G polymorphism affects PAI-1 expression by tumor tissue has not been established. For breast cancer tissue both an association of the PAI-1 4G/5G polymorphism with PAI-1 expression38 and a lack of association39 have been described. Moreover, tumor-related mutations, copy number variants, and epigenetic changes may influence PAI-1 expression by tumors.
As to how the germline PAI-1 4G/5G polymorphism affects tumor characteristics and cancer survival, study results are also inconclusive. For example, in breast cancer, results vary from a negative effect of the 4G allele38 to a negative effect of the 5G allele,40 or no effect at all.39 No data are available on tumor PAI-1 expression and the effect of germline PAI-1 4G/5G polymorphism in TC patients.
We analyzed germline DNA for the PAI-1 4G/5G polymorphism in TC patients. It is not known whether the analyzed PAI-1 4G/5G polymorphism is associated with differences in PAI-1 expression by tumor tissue. Although expression of PAI-1 by host stromal cells may be important for invasion by tumor cells and angiogenesis,41 the effect of the polymorphism on PAI-1 expression by surrounding stromal tissue is also not known. Nevertheless, the common 4G/4G variant of the PAI-1 4G/5G polymorphism seems to be an unfavorable prognostic factor, in addition to the commonly used IGCCC system. Due to its association with poor disease outcome, the 4G/4G variant may represent an unfavorable predictive factor for response to chemotherapy as well.
Although the current study is an association study that cannot exclude the effect of linkage disequilibrium and was performed in a mainly white patient group, it may function as hypothesis generating. Further research on the expression and role of the uPA/PAR pathway components and PAI-1 in TC is warranted.
As to the role of PAI-1 as intravascular fibrinolysis inhibitor (tPA pathway), we have not found a significant effect of the PAI-1 4G/5G polymorphism on the incidence of cardiovascular disease during or after chemotherapy. Analysis of CHD is complicated by the small number of patients with CHD, the short follow-up duration of patients with the 4G/4G variant (median, 7 years; range, 0-25 years) compared with 10 years (range, 0-27 years) for the 4G/5G group, and 12 years (range, 1-28 years) for the 5G/5G group, and the lower attained age in patients with 4G/4G (median age, 38 years; range, 20-65 years of age) versus 40 years of age (range, 20-73 years) for the 4G/5G group and 44 years of age (range, 19-70 years) for the 5G/5G group. Moreover, a considerable number of VTE is probably related to a central venous access port used for chemotherapy administration (17 of 26 [65%] VTE). Since 1998,central venous access ports have been used less frequently. A larger cohort of TC patients without central venous access ports is needed to study the effect of gene polymorphisms on VTE risk. However, treatment with chemotherapy per se is a risk factor for development of VTE as well as malignancy,42 known to be associated with circulating, tumor-derived, tissue factor-bearing microparticles.43 Coagulation activation by these factors, in combination with prothrombotic gene polymorphisms such as factor II G20210A and factor V G1691A, may further increase the risk for VTE.42 Of the potentially confounding polymorphisms factor II G20210A and factor V G1691A, the heterozygous variant of factor II G20210A shows an increased risk for all VTE combined, including central venous access port related VTE, during chemotherapy. The heterozygous variant of factor V G1691A (factor V Leiden) shows increased risk for CHD. These observations are in agreement with the established role of these polymorphisms in VTE and their potential, moderate role in CHD.28, 44 However, the variants of factor II G20210A polymorphism and factor V G1691A polymorphism occur at low frequencies and are present in a relatively small fraction of the TC patients experiencing VTE and CHD. In addition, due to the combination of a low frequency of polymorphic variants and relatively low VTE and CHD incidence rates, the power of the current study to assess the effects of factor II G20210A polymorphism and factor V G1691A polymorphism on risk for chemotherapy-related vascular toxicity is low.
In conclusion, we have found an association between the PAI-1 4G/5G polymorphism and TC-related survival, but not with chemotherapy-related cardiovascular toxicity. Because the 4G/4G variant of PAI-1 4G/5G polymorphism is associated with IGCCC poor prognosis, reduced TC-related survival, and higher prevalence of refractory disease, it may be an unfavorable prognostic as well as predictive factor for response to chemotherapy in TC patients. If the unfavorable effect of the 4G/4G variant can be confirmed, it may help identify patients at increased risk of not responding to chemotherapy, with curative intent. Because our observations suggest that PAI-1 may have a role in the prognosis of TC, future research entailing the uPA/PAR pathway, PAI-1, and the PAI-1 4G/5G polymorphism may enable fine tuning of the current prognosis classification system and the development of alternative treatment strategies.
CONFLICT OF INTEREST DISCLOSURES
Financial support: grant RUG 2004-3157 from the Dutch Cancer Society, The Netherlands.
- 21High plasminogen activator inhibitor and tissue plasminogen activator levels in plasma precede a first acute myocardial infarction in both men and women: evidence for the fibrinolytic system as an independent primary risk factor. Circulation. 1998; 98: 2241-2247., , , et al.