Abstract
Recent data suggest that erythropoietin (EPO) plays a substantial role in cancer development and clinical outcome by stimulating cell proliferation, invasion and angiogenesis. A functional polymorphism (rs1617640 G>T) in the promoter region of the EPO gene increases EPO protein expression. In the present study, we investigated the association of EPO rs1617640 G>T with susceptibility and clinical outcome of early-stage breast cancer. Genomic DNA of 539 female patients with histologically confirmed early-stage breast cancer and 804 controls was genotyped for EPO rs1617640 G>T. No association was found between EPO rs1617640 G>T and early-stage breast cancer susceptibility and clinical outcome (hazard ratio=1.24, 95% confidence interval=1.82-1.90, p=0.31). In conclusion, our findings suggest a lack of influence of EPO rs1617640 G>T on early-stage breast carcinogenesis and clinical outcome.
Erythropoietin (EPO) is a glycoprotein hormone mainly produced in the adult kidney and fetal liver (1). EPO gene expression is regulated by inhibitory [GATA binding protein 2 (GATA2), nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB)] and stimulatory [hypoxia-inducible factors (HIFs), hepatocyte nuclear factor] transcription factors (2). The EPO gene is activated under hypoxic conditions through binding of the HIF1α and HIF1β to the promoter, respectively (3). Its main function is to stimulate erythropoiesis by promoting proliferation and differentiation, or by inhibiting apoptosis of erythroid progenitors (4). Its biological effects are mediated through the binding to the erythropoietin receptor (EPOR), which belongs to the cytokine receptor type I superfamily (1, 5). EPO was long considered to be a specific stimulator of erythropoiesis, but experimental findings have shown that its expression is not restricted to the hematopoietic system. Differential expression of EPO has been shown in other non-hematopoietic tissues, such as brain astrocytes, breast epithelial cells and human reproductive organs, suggesting a pleiotropic mechanism of action, extending well beyond the maintenance of red cell mass (6-8). Furthermore, multiple investigators have documented EPO production in various tumor cell lines and cancer entities, including breast cancer (9-13). Major signal-transduction pathways activated by EPO involve the Janus kinase (JAK)/signal transducer and activator of transduction (STAT) and rat sarcoma (RAS)/mitogen-activated protein kinase pathways, which are essential for the inhibition of apoptosis and the stimulation of cell proliferation in response to this hormone (5, 14). Moreover, it has been demonstrated that exposure of tumor cells to EPO-stimulated tyrosine phosphorylation, DNA synthesis and proliferation, suggesting that EPO signaling is biologically active in malignant cells (9, 10). Additionally, recent data indicate that EPO is involved in angiogenesis and invasion of neoplastic cells and may be important in breast tumorigenesis (12, 15, 16). Mohyeldin et al. reported an EPO-mediated invasion through the JAK/STAT pathway in head and neck squamous carcinoma cells (17). Furthermore, it has been demonstrated that EPO stimulates proliferation in renal cancer cells (18). In endometrial carcinoma, EPO signaling was shown to contribute to tumor progression and increased aggressiveness, identifying increased EPO expression as an independent prognostic factor (19). In prostate cancer, EPO was shown to regulate an autocrine/paracrine signaling pathway that influences growth and survival of cancer cells (20). In a recent study, Acs et al. demonstrated that EPO signaling inhibits hypoxia-induced apoptosis in human breast carcinoma cells and was associated with poor prognosis (21). The TT genotype of a common gene variant in EPO, rs1617640 G>T, located in the promoter region, leads to increased EPO protein expression through an ecotopic viral integration site 1 (EVI1)/melatonin 1 (MEL1) or activator protein 1 (AP1) enhancer binding site and may be associated with cancer susceptibility and clinical outcome (22). In the present study, we investigated the association of EPO rs1617640 G>T with susceptibility and clinical outcome in patients with early-stage breast cancer.
Materials and Methods
Study participants. A total of 539 female patients with histologically confirmed breast cancer without distant metastases were included in this retrospective study. All patients were included in a breast cancer surveillance program between 2001 and 2009 at the Division of Clinical Oncology, Department of Internal Medicine, Medical University of Graz, Austria, providing follow-up care at regular intervals (3-month intervals in years 1-3, 6-month intervals in years 4-5, and 12-month intervals in years 6-10 after curative surgery). Follow-up investigations included clinical check-up, laboratory, radiological (bone scan, liver scan, chest X-ray, and mammograms), and gynecological examination. A total of 804 healthy female controls were recruited from local health screening studies, the Salzburger Atherosklerose Präventionsprogramm bei Personen mit hohem Infarkt Risiko and the Grazer Diabetes Screening Programm. Controls with previous malignant disease were excluded from the present study. Blood samples were taken from all patients and controls, and stored at −20°C. The present study was performed according to the Austrian Gene Technology Act, and has been approved by the Institutional Reviewer Board of the Medical University of Graz. All participants gave written informed consent and were Caucasian.
Candidate polymorphisms. The following pre-defined criteria for the polymorphism selection within the EPO gene were used: (i) polymorphisms that modulate the function of the gene (based on published data); (ii) minor allele frequency ≥10% in Caucasians (based on the population genetics section in the Ensembl Genome Browser: http://uswest.ensembl.org/index.html); (iii) reported in published clinical association studies.
Genotyping. DNA was extracted using the GenElute™ Blood Genomic DNA Kit from SIGMA (Saint Louis, Missouri, USA). Genotyping was performed using the 5’-exonuclease (TaqMan™) technology with primers and probes designed and manufactured by Applera's Assay by-Design custom service (Applied Biosystems, Vienna, Austria). Polymerase chain reaction (PCR) and evaluation of fluorescence data were carried out as described elsewhere (23). For each sample, one negative control containing water instead of DNA was added to check for contamination. For purposes of quality control of genotyping, a total of 10% of the analyzed samples were re-analyzed. The investigator responsible for genotyping was blinded to the clinical dataset.
Statistics. The primary endpoint of the study was disease-free survival (DFS). DFS was calculated from the date of diagnosis of early-stage breast cancer to the date of the first observation of tumor recurrence. DFS was censored at the last follow-up if the patient remained tumor-free at that time. Allelic distribution of the polymorphism was tested for deviation from Hardy-Weinberg equilibrium using χ2-test. The true mode of inheritance of the polymorphism tested has not been established yet and we assumed a co-dominant, additive, dominant, or recessive genetic model was appropriate. The association of the polymorphism with DFS was analyzed using Kaplan Meier curves and log-rank test. In the multivariate Cox regression analysis, the model was adjusted for menopausal status, stage, histopathological grading, receptor status, human epidermal growth factor receptor 2 (HER2)/neu status and adjuvant treatment modalities. Case-wise deletion for missing genotypes was used in univariate and multivariate analyses. Genotype frequencies in patients and controls were compared using χ2 tests. All analyses were performed using the SPSS statistical software package (SPSS Inc., Sunnyvale, CA, USA).
Results
Baseline patient characteristics, tumor biological factors, treatment modalities and their association with clinical outcome are shown in Table I. The median age at time of diagnosis was 57 years (range 29 to 84 years). The median follow-up duration was 61.1 months (range 12 to 107 months). Follow-up data was missing for 38 (7.1%) patients. Genotyping was successful in 520 (96.5%) patients and 799 (99.4%) controls. In failed cases, genotyping was not successful due to limited quantity and/or quality of extracted genomic DNA. The genotyping quality control by subsample re-analysis provided a genotype concordance of >99%. The allelic frequencies for the polymorphism were within the probability limits of Hardy-Weinberg equilibrium. A statistically significant correlation was found between clinical outcome and tumor size, lymph node involvement and tumor stage (Table I). There was no significant association between EPO rs1617640 G>T and breast cancer susceptibility (Table II). When EPO rs1617640 G>T was analyzed for predicting clinical outcome, no association was found with DFS (hazard ratio (HR)=1.24, 95% confidence interval (CI)=1.82-1.90, p=0.31; multivariate analysis: HR=1.30, 95% CI=0.85-1.98, p=0.23) (Figure 1).
Discussion
In the present study, we investigated the clinical impact of a functional promoter polymorphism of the EPO gene on the risk and outcome of early-stage breast cancer. Our findings indicated no association between EPO rs1617640 G>T and early-stage breast cancer susceptibility and outcome. EPO signaling is known to activate several intracellular kinase pathways (5). These pathways were investigated particularly in hematopoietic cells and may also contribute to its non-hematopoietic/tumor cell action (24). Elevated EPO levels have been found in patients with various tumor entities, including breast carcinomas, and numerous reports link EPO signaling to the modulation of tumor cell proliferation, apoptosis and tumorigenesis (9, 10, 12, 18, 21). Acs et al. showed that EPO promotes tyrosine phosphorylation and proliferation of breast cancer cells in vitro (9). However, in other studies, no biological effect to EPO was observed, as measured by proliferation or clonogenic growth in different tumor cell lines (25-27). Another report suggested that increased autocrine EPO signaling induced by moderate levels of hypoxia inhibits hypoxia-induced apoptosis and promotes survival in MCF-7 human breast cancer cells (28). In a study by Arcasoy et al., EPO expression was found in 60% of breast cancer cells and EPO and EPOR co-localization in tumor cells was present in many cases. Moreover, local administration of a neutralizing anti-EPO antibody, soluble EPOR, or an inhibitor of JAK2, resulted in a delay in tumor growth, demonstrating functional EPO signaling in breast cancer cells (12). In contrast, Westphal et al. showed that even though various cancer cell lines, including breast cancer cells, expressed EPOR, EPO signaling was not biologically active, and EPO was not essential for growth of these tumor cells in culture (29). This observation was confirmed by LaMontagne et al., who revealed in two well-established breast carcinoma models that EPOR was predominately cytosolic and not present as an active surface receptor on these cell lines, supported by the findings that there was no significant measurable EPO-specific binding activity on the surface of either cell line (30). In vitro, EPO reportedly stimulated angiogenesis by induction of proliferation of primary endothelial cells and endothelial cell lines (31-33). Moreover, it was reported that high doses of EPO induced the release of vascular endothelial growth factor (VEGF) in some cultured tumor cell lines (34). In vivo, using both rat mammary adenocarcinoma and mouse colon carcinoma models, no difference was observed in angiogenesis between EPO-treated and placebo-treated tumors (35). The effect of EPO on tumor vascular density was investigated in two different glioma xenograft models with no increased density reported (36). Biological disorders that alter EPO levels in humans offer additional insights into the potential role of EPO in tumor induction and progression. Primary congenital disorders associated with increased EPO production and mutations in EPOR that result in hypersensitivity to EPO are associated with erythrocytosis in humans (37, 38). However, a higher cancer incidence has not been observed in patients with familial and congenital polycythemias. In a study carried out on transgenic mice that express high levels of EPO, no incidence of erythroleukemia was observed during the course of two years, indicating that EPO stimulation alone does not induce tumorigenesis (39). The conflicting results in these studies might be due to variable methodological approaches, with some limited to histopathological and biochemical evaluation, or differences in the cell lines used. Furthermore, most of these studies were performed in vitro and may not reflect the affection of EPO on tumor cell function in vivo. A recent study demonstrated that the TT genotype of EPO rs1617640 G>T is associated with severe diabetic microvascular complications, such as diabetic retinopathy and end-stage renal disease (22). It has been reported that in non-diabetic individuals with the TT genotype, EPO protein concentration was 7.5-fold higher in vitreous samples as compared to those with the GG genotype (22). In vitro experiments showed a 25-fold higher expression of EPO in constructs containing the T allele in the promoter region of the EPO gene (22). It has been suggested that the increase in promoter activity might be due to an EVI1/MEL1 or AP1 binding site created by the T allele. Ma et al. demonstrated a strong association between the GG genotype of EPO rs1617640 G>T with myelodysplastic syndrome (MDS), relative to the control group and other types of acute and chronic leukemia, implicating low levels of EPO in the development of MDS. In addition, the MDS group with the GG genotype displayed significantly shorter complete remission duration compared to patients with the TT genotype, but there was no correlation between genotypes and survival (40).
A limitation of our study is its retrospective design, therefore a selection bias cannot be excluded. Furthermore, the control group was not age matched; however, single nucleotide polymorphisms (SNPs) are constant during life. Frequencies of polymorphisms have been shown to vary between different ethnic cultures; therefore our results may not be attributable to ethnicities other than Caucasian. In conclusion, we found no association between EPO rs1617640 G>T and early-stage breast cancer risk and outcome, but as the current knowledge on the impact of EPO on carcinogenesis and cancer progression is inconclusive, further studies designed to specifically evaluate the effects of EPO on tumor susceptibility and survival in cancer patients are warranted.
Footnotes
-
Grant Support
This work was funded by a Styrian Cancer Aid research grant.
- Received April 15, 2012.
- Revision received June 25, 2012.
- Accepted June 26, 2012.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved