Gynecologic Oncology 97 (2005) 514 – 518 www.elsevier.com/locate/ygyno
Common death receptor 4 (DR4) polymorphisms do not predispose to ovarian cancer Peter Horaka, Dietmar Pilsa, Max Roesslera, Sandra Tomeka, Katarzyna Elandta, Robert Zeillingerb, Christoph Zielinskia, Michael Krainera,T a
Clinical Division of Oncology, Department of Medicine I, Medical University Vienna, Austria b Department of Gynecology and Obstetrics, Medical University Vienna, Austria Received 25 October 2004 Available online 16 March 2005
Abstract Objective. Polymorphisms of death receptor 4 (DR4) might impair the apoptotic signal transduction and lead to dysregulation of the homeostasis between cell survival and cell death, promoting tumor development and progression. Methods. We performed an analysis of known DR4 polymorphisms, namely G442A, C626G, and A1322G, in germ line DNA of 97 ovarian cancer patients and controls as well as in established ovarian cancer cell lines. Results. Patient and matched control populations were not differing significantly in case of G442A ( P = 0.736) and C626G alterations ( P = 0.699). For the A1322G transversion, we generated population data for the first time and could find a rate of 19% heterozygotes and 3% homozygotes. Again, we could not detect any significant difference between patients and controls ( P = 0.326). Conclusion. To summarize, alterations of the DR4 gene do not lead to clinically relevant ovarian cancer predisposition and are therefore most unlikely to contribute to familial ovarian cancer. D 2005 Elsevier Inc. All rights reserved. Keywords: Ovarian cancer; DR4 protein; Polymorphism; RFLP
Introduction Ovarian cancer is the most lethal gynecologic malignancy and the fourth most frequent cause of cancer-related death in women. Mostly to late diagnosis, most patients have a poor prognosis despite frequent initial response to chemotherapeutic regimens. Ovarian cancer cells were shown to be sensitive to TRAIL-induced apoptosis and combination with chemotherapeutic drugs enhances the apoptotic effect of TRAIL [1–3]. TRAIL induces apoptosis by ligation to death receptor DR4 and death receptor 5 (DR5, TRAIL-R2, KILLER) [4–6] and subsequent activation of the apoptotic cascade through caspase 8 and FADD T Corresponding author. Clinical Division of Oncology, Department of Medicine I, University Hospital, Wa¨hringer Gu¨rtel 18-20, A-1090 Vienna, Austria. Fax: +43 1 40400 4451. E-mail address:
[email protected] (M. Krainer). 0090-8258/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2005.01.021
[7], forming the death inducing signaling complex (DISC). Two other, truncated, TRAIL receptors, decoy receptor 1 (DcR1, TRAIL-R3, TRID) and decoy receptor 2 (DcR2, TRAIL-R4, TRUNDD), can interfere with the programmed cell death induction by TRAIL [4,6,8,9]. Latest evidence for the involvement of the TRAIL system in ovarian cancer is the fact that TRAIL expression was shown to correlate closely with the prognosis of ovarian cancer patients [10]. DR4 may play an important role during tumor development in ovarian cancer either as an endogenous factor in the regulation of apoptosis of tumor cells or as an exogenous factor in the immune escape of tumor cells [11]. Moreover, all four receptors for TRAIL were mapped to chromosome 8p22–p21 [8,9], a region we and other research groups found frequently deleted in ovarian cancer [12,13], suggesting that genes located in this area may be crucially involved in the pathogenesis of this disease.
P. Horak et al. / Gynecologic Oncology 97 (2005) 514–518
Missense mutations at the 5V end of the DR4 gene may interfere with the ligand–receptor interaction of TRAIL, whereas 3V terminal mutations may influence DISC formation. Identification of these mutations in an ovarian cancer population and definition of their impact on TRAIL sensitivity in this population would be needed in terms of future TRAIL-based therapy. Mutations in the ectodomain of DR4, namely G422A (R141H) and C626G (T209R), were found in a higher frequency in primary tumors of different origin (SCLC, NSCLC, HNSCC, gastric adenocarcinoma) as compared to matched controls [14]. Additionally, these alterations were suggested to be mostly germline in origin and having influence on tumorigenesis of some cancer entities, particularly in association with additional environmental influences, like smoking [15]. In addition, another DR4 polymorphism (A1322G, K441R) in the death domain of DR4 was reported, and this time the genotype was functionally linked to resistance against TRAIL-induced apoptosis in several cancer cells [16]. This prompted us to investigate the possible influence of DR4 polymorphisms on genetic predisposition to ovarian cancer in a population-based study comparing the frequency of genotypes in ovarian cancer patients and matched controls as well as in ovarian cancer cell lines.
Materials and methods Study population and statistical analysis Ninety-seven patients (mean age at diagnosis 55.8 (SD F 12.0) years) with pathologically confirmed epithelial ovarian cancer diagnosed between 1981 and 2000 were included in our observation (Table 1). Written informed consent was obtained from all patients before drawing blood samples. Table 1 Patients’ characteristics (N = 92)a,b Histology Serous Mucinous Endometrioid Clear cell Undifferentiated Mixed Grading 1 2 3 FIGO I II III IV a
n
%
51 9 15 3 11 3
55.4 9.8 16.3 3.3 12.0 3.3
26 28 39
28.3 30.4 41.3
33 5 45 9
35.9 5.4 48.9 9.8
All patients were of Caucasian ethnicity. For 5 patients, information regarding the above criteria was not available (except for epithelial ovarian cancer).
b
515
One hundred subjects without a history of malignant disease, but corresponding by age, sex, and race, were chosen as controls. Deviations from Hardy Weinberg equilibrium were calculated for all groups and the goodness of fit v 2 test was used to compare the observed allele frequencies between patient and control populations. Risk of ovarian cancer associated with DR4 polymorphisms was assessed using odds ratio estimate (OR) and 95% confidence interval (CI). Our study had N80% power to detect the protective effect (OR = 0.45) of the homozygous C626G alteration as described by Hazra et al. [15]. The G422A allele cosegregates with C626G in about 96%, so similar odds ratios were expected. For the A1322G polymorphism, we had a more than 80% power to detect an OR = 2 for the heterozygous and an OR = 6 for the homozygous population as based on our results. Cell culture and genotype analysis Epithelial ovarian cancer cell lines ES-2 and Caov-3 were obtained from ATCC (Rockville, MD, USA), cell lines A2780 and A2780ADR from ECACC (Salisbury, Wiltshire, UK), and cell lines OV-MZ-15 and OV-MZ-26 were established from ascitic fluid of ovarian cancer patients. All cell lines were cultured with appropriate media and kept at standard conditions. Genomic DNA was extracted from peripheral blood lymphocytes of ovarian cancer patients and controls. For the A1322G polymorphism, we sequenced the death domain of DR4 using an automated fluorescence-based cycle sequencer (ABI PRISMR 310 Genetic Analyzer, Applied Biosystems). PCR products generated with primers DR4-11 and DR4-12 [16] were subjected to sequencing. To detect the G422A and the C626G polymorphism, a PCR with subsequent RFLP analysis was performed according to Fisher et al. [14]. The A1322G polymorphism was sequenced from both sides in all patients and the RFLP results were also validated by sequencing in ten cases.
Results G422A polymorphism The G to A transversion results in an amino acid change of a histidine for an arginine (R141H) and creates a new and unique FokI restriction site and therefore yields a 160-bp and a 70-bp fragment upon incubation with FokI. From the six cell lines studied, two had the homozygous GG variant (OV-MZ-26 and ES-2), one was heterozygous (OV-MZ-15) and three had the homozygous CC variant (A2780, A2780ADR and Caov-3) (Fig. 1). In the study population, both cohorts fulfilled the HW equilibrium criteria. In direct comparison, the GG, GA, and AA alleles were almost equally distributed in both, patients and controls ( P = 0.736; Table 2).
516
P. Horak et al. / Gynecologic Oncology 97 (2005) 514–518
respectively. It did not differ significantly from the control group ( P = 0.699, Table 2) nor from the HW equilibrium, estimated for each population. Odds ratio for the ovarian cancer population carrying the homozygous GG alteration, which was found to reduce bladder cancer risk [15], was 0.778 (95% CI, 0.425–1.422), displaying a slight but not significant trend. Not unexpectedly, we could observe a strong cosegregation of both polymorphisms in our study population (93%) with a linkage disequilibrium between the two loci (DV= 0.886, P b 0.01). A1322G polymorphism
Fig. 1. PCR restriction fragment length polymorphism analysis of the G422A and C626G alleles in ovarian cancer cell lines using the restriction enzymes FokI for G422A (cleaving the A allele) and DraIII for C626G (cleaving the C allele). The fragments were separated on a 3% agarose gel and visualized using ethidium bromide.
C626G polymorphism The C to G transversion at position 626, resulting in a substitution of arginine for threonine (T209R), eliminates a unique DraIII restriction site. Incubation with DraIII endonuclease yields two fragments (164 and 56 bp) for the cytosine allele and the 220-bp fragment if the allele was homozygous for guanine. The ovarian cancer cell lines showed 100% cosegregation of the G422A and C626G polymorphisms, the cell lines OV-MZ-26 and ES-2 being homozygous for the CC alteration, OV-MZ-15 being heterozygous, and A2780, A2780ADR, and Caov-3 displaying the GG genotype (Fig. 1). In the ovarian cancer population studied, the distribution was 13 (14%), 51 (55%), and 28 (31%) for CC, CG, and GG genotypes,
For the A1322G (K441R) polymorphism, we sequenced the DR4 death domain of 97 ovarian cancer patients and 97 controls. Additionally, we analyzed six ovarian cancer cell lines for the A1322G alteration. The sequencing revealed one homozygous GG variant (Caov-3) and one heterozygous GA variant (OV-MZ-15), the other four cell lines displaying the wild-type AA genotype. Although the A1322G polymorphism was shown to decrease TRAIL sensitivity [16], the cell line carrying the homozygous A1322G alteration (Caov-3) did not display reduced TRAIL-induced apoptosis in comparison to other ovarian cancer cell lines and the observed TRAIL resistance in the OV-MZ-15 cell line was rather caused by FLIPL overexpression than by A1322G heterozygosity [3]. Consequently, in the study population, group comparison did not reveal any significant differences between the cohorts ( P = 0.326), although the patients’ cohort did not match the HW equilibrium. Risk for ovarian cancer in the population carrying the homozygous GG allele was not convincing with OR of 3.948 (95% CI, 0.431–36.137) and even reversed in the AG/GG population with OR of 0.940 (95% CI, 0.472–1.873).
Table 2 Allelic variance of the polymorphism at positions 422, 626, and 1322 of the DR4 gene does not show any significant differences between ovarian cancer patients and matched controls G422A GG GA AA v 2 test C626G CC CG GG v 2 test A1322G AA AG GG AG + GG v 2 test a
Cases
%
Controls
%
Odds ratio (95% CI)a
18 48 26 P = 0.736
20 52 28
20 47 33
20 47 33
1 1.135 (0.534–2.410) 0.875 (0.386–1.984)
1
13 51 28 P = 0.699
14 50 31
14 50 36
14 50 36
1 1.098 (0.470–2.569) 0.838 (0.340–2.065)
1
77 16 4
79 17 4
76 20 1
78 21 1
1 0.790 (0.381–1.638) 3.948 (0.431–36.137) 0.940 (0.472–1.873)
P = 0.326
Results of an univariate analysis.
0.825 (0.445–1.530)
0.778 (0.425–1.422)
P. Horak et al. / Gynecologic Oncology 97 (2005) 514–518
Discussion Elevated TRAIL expression in primary ovarian cancer specimen was correlated with favorable ovarian cancer survival [10]. The molecular mechanisms of this effect are largely unknown, so further evaluation of TRAIL signaling and TRAIL receptors in ovarian cancer could be necessary. Moreover, the four TRAIL receptors are located on chromosome 8p21.2, which is often subject to loss of heterozygosity (LOH) in various solid tumors, including ovarian cancer. The frequent LOH in this region may signalize the presence of one or more tumor suppressor genes around 8p21. TRAIL receptors are highly promising candidates for this position, as apoptosis signaling and especially the loss of it are often involved in carcinogenesis. In addition to LOH, epigenetic changes as well as single nucleotide polymorphisms may further be responsible for loss of function or downregulation of these proteins. Currently, three polymorphisms in the DR4 gene are considered to be of functional relevance in some solid tumors. The A1322G mutation was described to cause impaired TRAIL signaling and hereby TRAIL resistance in cancer cell lines in a dominant-negative fashion [16]. The G422A and C626G polymorphisms were found more frequently in some primary tumors [14]. A significantly lower risk of bladder cancer was attributed to individuals carrying the homozygous germline C626G alteration, who were exposed to environmental carcinogens [15]. While TRAIL seems to be a putative therapeutic agent for ovarian cancer, a population-based assessment of the known DR4 polymorphisms in ovarian cancer patients was missing. Our study focused on the significance of germline DR4 mutations in ovarian cancer and their role in the predisposition to this disease based on observations in other tumor entities and in vitro observations in ovarian cancer. We could not detect any significant deviation for the three polymorphisms (G422A, C626G, and A1322G) from the normal, age-, and race-matched population. For the mutation in the ectodomain of DR4, C626G, which was shown to be involved in the pathogenesis of bladder cancer in smokers, there was no significant difference between ovarian cancer patients and age-matched controls as shown in Table 2, which rules out an influence of this polymorphism on the genetic predisposition for ovarian cancer. The difference between the two tumor entities may partly be explained by the fact that environmental agents do not play a major role in ovarian cancer and cofactors might be necessary for the penetrance of the genotype. Significant differences between the patient and the control population could not be observed neither for the G422A nor for the A1322G polymorphism. We observed a prevalence of 19% heterozygotes (AG) and 3% homozygotes (GG) for the A1322G polymorphism in the study population. This polymorphism in the death domain of DR4 was previously shown to decrease TRAIL sensitivity in vitro in a
517
dominant-negative manner [16]. In consideration of the proposed diminished TRAIL sensitivity of these genotypes, one could expect a reduced TRAIL response rate in vivo. However, we could not confirm any correlation between the A1322G allele status and TRAIL resistance in vitro, which is most probably due to disturbances of other components of the TRAIL pathway, and consequently there was no significant difference between patients and controls, though due to the low prevalence of the GG genotype, no definite conclusion about this polymorphism can be made. To summarize, our investigation of germline DR4 alterations could not reveal any significant allelic difference between ovarian cancer patients and controls. Therefore, we cannot claim that the so far described DR4 mutations may be a risk factor for ovarian cancer development. However, since TRAIL expression does influence ovarian cancer prognosis, genetic or epigenetic dysregulations of DR4 might be important co-factors in the later course of the disease. Given the TRAIL sensitivity of ovarian cancer cells and its synergistic effects in combination with chemotherapy [1–3], tumor necrosis factor-related apoptosisinducing ligand (TRAIL) alone or in combination with cytotoxic drugs could constitute one of the novel therapeutic approaches to this deadly disease. Acknowledgment This work was supported by FWF (Fonds zur Ffrderung der wissenschaftlichen Forschung) Grant 13473. References [1] Cuello M, Ettenberg SA, Nau MM, Lipkowitz S. Synergistic induction of apoptosis by the combination of TRAIL and chemotherapy in chemoresistant ovarian cancer cells. Gynecol Oncol 2001; 81:380 – 90. [2] Vignati S, Codegoni A, Polato F, Broggini M. Trail activity in human ovarian cancer cells: potentiation of the action of cytotoxic drugs. Eur J Cancer 2002;38:177 – 83. [3] Tomek S, Horak P, Pribill I, Haller G, Rossler M, Zielinski CC, et al. Resistance to TRAIL-induced apoptosis in ovarian cancer cell lines is overcome by co-treatment with cytotoxic drugs. Gynecol Oncol 2004; 94:107 – 14. [4] Pan GH, Ni J, Wei YF, Yu GL, Gentz R, Dixit VM. An antagonist decoy receptor and a death domain-containing receptor for TRAIL. Science 1997;277:815 – 8. [5] Pan GH, ORourke K, Chinnaiyan AM, Gentz R, Ebner R, Ni J, et al. The receptor for the cytotoxic ligand TRAIL. Science 1997;276: 111 – 3. [6] MacFarlane M, Ahmad M, Srinivasula SM, Fernandes Alnemri T, Cohen GM, Alnemri ES. Identification and molecular cloning of two novel receptors for the cytotoxic ligand TRAIL. J Biol Chem 1997; 272:25417 – 20. [7] Sprick MR, Weigand MA, Rieser E, Rauch CT, Juo P, Blenis J, et al. FADD/MORT1 and caspase-8 are recruited to TRAIL receptors 1 and 2 and are essential for apoptosis mediated by TRAIL receptor 2. Immunity 2000;12:599 – 609. [8] Degli-Esposti MA, Dougall WC, Smolak PJ, Waugh JY, Smith CA,
518
[9]
[10]
[11]
[12]
P. Horak et al. / Gynecologic Oncology 97 (2005) 514–518 El Deiry WS. The novel receptor TRAIL-R4 induces NF-kappa B and protects against TRAIL-mediated apoptosis, yet retains an incomplete death domain. Immunity 1997;7:813 – 20. Marsters SA, Sheridan JP, Pitti RM, Huang A, Skubatch M, Baldwin D, et al. A novel receptor for Apo2L/TRAIL contains a truncated death domain. Curr Biol 1997;7:1003 – 6. Lancaster JM, Sayer R, Blanchette C, Calingaert B, Whitaker R, Schildkraut J, et al. High expression of tumor necrosis factor-related apoptosis-inducing ligand is associated with favorable ovarian cancer survival. Clin Cancer Res 2003;9:762 – 6. Wajant H, Pfizenmaier K, Scheurich P. TNF-related apoptosis inducing ligand (TRAIL) and its receptors in tumor surveillance and cancer therapy. Apoptosis 2002;7:449 – 59. Pribill I, Speiser P, Leary J, Leodolter S, Hacker NF, Friedlander ML, et al. High frequency of allelic imbalance at regions of chro-
[13]
[14]
[15]
[16]
mosome arm 8p in ovarian carcinoma. Cancer Genet Cytogenet 2001;129:23 – 9. Wiley SR, Schooley K, Smolak PJ, Din WS, Huang CP, Nicholl JK, et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 1995;3:673 – 82. Fisher MJ, Virmani AK, Wu L, Aplenc R, Harper JC, Powell SM, et al. Nucleotide substitution in the ectodomain of TRAIL receptor DR4 is associated with lung cancer and head and neck cancer. Clin Cancer Res 2001;7:1688 – 97. Hazra A, Chamberlain RM, Grossman HB, Zhu Y, Spitz MR, Wu XF. Death receptor 4 and bladder cancer risk. Cancer Res 2003;63: 1157 – 9. Kim KH, Fisher MJ, Xu SQ, El Deiry WS. Molecular determinants of response to TRAIL in killing of normal and cancer cells. Clin Cancer Res 2000;6:335 – 46.
