Genetic Polymorphism in EGF Is Associated with Prostate Cancer Aggressiveness and Progression-Free Interval in Androgen Blockade–Treated Patients

Prostate cancer is the most common nonskin cancer among men in most Western populations (1). Initial treatment in early stages of disease is usually prostatectomy or radiation, which removes or destroys tumoral cells that are confined within the prostate capsule. Patients diagnosed in advanced stages are frequently submitted to endocrine manipulation with androgen blockade therapy (ABT; ref. 2), although most men will eventually fail this therapy and die from recurrent hormone-resistant prostate cancer (HRPC). HRPC is a common lethal form of prostate cancer that frequently metastasizes and eventually leads to death (3). Therefore, it is important to understand the mechanisms involved in androgen-independent progression.

The androgen pathway is commonly accepted to have a critical role in survival of prostatic cells. Nevertheless, progression into advanced prostate cancer and incurable forms has also been associated with the activation of other cascades mediated by growth factors responsible for the balance between cell growth rate and apoptosis (4). Several pathways have been proposed to be involved in HRPC development, and their understanding will pave the way to more effective therapies. In fact, prostate cell growth and differentiation in the absence of androgens may be due to alternative intracellular signaling pathways, such as the ErbB pathway, which implicates epidermal growth factor (EGF) as a probable ligand (5). Additionally, cross-talk between these signaling transduction pathways has been implicated in prostate cancer survival in an androgen-poor environment (6, 7).

Cumulatively, studies by Ye et al. (8) showed that there is a positive feedback stimulation loop between androgen-induced gene transcription and EGF-stimulated signal transduction, as one could stimulate the synthesis of the receptors for the other.

The EGF receptor (EGFR) is proposed to participate in the pathogenesis or maintenance of several human cancers of epithelial origin. In prostate cancer cells, EGFR ligands are frequently elevated and EGFR itself is commonly overexpressed (9). Furthermore, EGFR expression increases during progression to a hormone-resistant stage (10).

EGF is encoded by EGF gene, located in chromosome 4q25-q27. Recently, Shahbazi et al. (11) identified a functional G>A single nucleotide polymorphism at position +61 in the 5′ untranslated region of the EGF gene. In vitro studies showed that G carriers have an increased EGF production in cultured peripheral blood mononuclear cells, glioma cells, glioblastoma cells, and breast cancer cells (11–13).

Previous studies found controversial results about the association between EGF polymorphism and cancer development and prognosis (11–16). Nevertheless, the present study is the first to evaluate the relevance the EGF functional polymorphism in HRPC patients.

Study population. One hundred and twenty-three patients (ages 71.7 ± 7.4 y) submitted to ABT between 1995 and 2006 at Lisbon Medical Centre (Central Region) were included in the study. All cases were submitted to prostate biopsy and had histopathologically confirmed diagnosis; the median follow-up time was 37 mo (range, 2-137 mo). All cases were considered with complete data for statistical analysis, from which 35% of the cases had localized prostate disease (T₁-T_2b), 36% of the cases had locally advanced disease (T₃-T₄), and 29% of the cases had metastatic disease (N⁺ and/or M⁺). The types of hormonal treatment were as follows: antiandrogens plus luteinizing hormone-releasing hormone agonist combination therapy (84.9%), luteinizing hormone-releasing hormone agonist alone (9.2%), and antiandrogens alone (5.9%). Hormone resistance was evaluated through prostate-specific antigen recurrence, which was defined as two consecutive increasing prostate-specific antigen values >1.0 ng·mL⁻¹ and differing by >0.2 ng·mL⁻¹.

Men ages >45 y, without known history of cancer, were recruited from the Portuguese Institute of Oncology-Porto Centre blood donor's bank (ages 55.6 ± 8.5 y) and included in the control group (n = 152). Study was conducted according to Helsinki Declaration principles. Antecubital peripheral venous blood sample was collected from each subject enrolled in the study followed by DNA extraction from WBC fraction according to salting out procedure (17).

EGF +61 G>A polymorphism genotyping.EGF +61 G>A variants were analyzed through PCR-RFLP method, as described in a previous report (16). Briefly, DNA was amplified in a 50-μL reaction mixture containing EGF +61 G>A primers (forward, 5′-TGTCACTAAAGGAAAGGAGGT-3′; reverse, 5′-TTCACAGAGTTTAACAGCCC-3′), 1× PCR buffer, 1 unit Taq polymerase, 1.5 mmol/L MgCl₂, 0.2 mmol/L deoxynucleotide triphosphates, and 100 ng DNA.

PCR products (242 bp) were incubated overnight with AluI restriction endonuclease at 37°C. The polymorphism was defined by the presence (A) or absence (G) of an additional restriction site.

Quality control procedures implemented for the EGF genotype analyses included double sampling in ∼10% of the samples to assess reliability and the use of negative controls to step-away false positives. Two authors obtained the results independently, and the ambiguous results were reanalyzed.

Statistical analysis. Genotype proportions among groups were compared with Pearson χ² test. Central tendency measures and comparisons between groups were evaluated through independent sample Mann-Whitney U test. Unconditional logistic regression analysis was used to compute odds ratio (OR) and 95% confidence interval (95% CI) estimating the association of genotypes to prostate cancer susceptibility, and to stratified Gleason score, metastatic disease, and hormone resistance development, while adjusting for age in all these variables [age-adjusted OR (aOR)]. The Kaplan-Meier method and log-rank test were used to compare genotype influence in the disease-free survival interval.

The functional EGF +61 G>A polymorphism genotype distribution in patients and controls is described in Table 1. Using the recessive model, frequencies for homozygous AA and AG/GG genotypes were, respectively, 0.27 and 0.73 for prostate cancer patients and 0.38 and 0.62 in the control group. Although the association is not statistically significant, there is a trend to an overrepresentation of AG/GG genotype in prostate cancer group compared with normal controls (aOR, 1.94; 95% CI, 0.99-3.79; P = 0.053).

Table 2 presents EGF +61 G>A genotype distribution and age-adjusted risk for being diagnosed with an aggressive prostate cancer phenotype (e.g., presence of metastatic disease or Gleason grade >7). Carriers of the G allele have a significantly higher age-adjusted risk for being diagnosed with metastatic prostate cancer (aOR, 2.61; 95% CI, 1.03-6.60; P = 0.043) and with higher Gleason grade tumors (≥7; aOR, 3.37; 95% CI, 1.47-7.73; P = 0.004).

Within the group of patients, EGF +61 G>A functional polymorphism was not associated with disease stage and prostate-specific antigen serum level at diagnosis (Table 3). Table 4 shows the influence of EGF +61 G>A polymorphism in the response to ABT. Prostate cancer patients who developed hormone resistance had similar genotype frequencies to hormone respondents (P = 0.468).

Kaplan-Meier survival curves (Fig. 1) and log-rank test (Mantel-Cox) show that EGF +61 G>A polymorphism is associated with progression-free interval cumulative probability. Median estimated cumulative survival was significantly lower in AG/GG carriers compared with AA homozygous (15.2 and 87.4 months, respectively; P = 0.018), showing an earlier relapse in prostate cancer patients following hormone ablation therapy.

Patients with local or distant metastatic prostate cancer are usually treated primarily through pharmacologic androgen suppression (2). This hormonal therapy is initially efficient, although the majority of patients will subsequently become unresponsive to androgen inhibition (18). HRPC is a complex and heterogeneous form of prostate cancer with a high capacity of progression and metastization (4).

EGF has been described as one of the most influent mitogenic factors in prostate function regulation (19). Furthermore, its receptor is overexpressed in prostate tumors, and the EGF-EGFR pathway is involved in hormone resistance development (9). In prostate cancers, EGFR is weakly expressed in the benign gland and in areas of low-grade prostatic intraepithelial neoplasia, whereas it is highly expressed in high-grade prostatic intraepithelial neoplasia and in neoplastic cells (20). Expression levels of EGF and EGFR in prostate cancer cells seem to be enhanced during disease progression to HRPC and metastatic prostate cancer (10).

Through overproduction of EGFR ligands such as EGF, transforming growth factor-α, amphiregulin, heparin-binding EGF, and betacellulin, cancer cells can stimulate their own proliferation through an autocrine mechanism (21–23). In vitro studies show that some EGFR ligands, such as amphiregulin, are overexpressed in prostate cancer cells (PC3; ref. 24), emphasizing the relevance of further studies in other ligands. Cumulatively, in vitro studies revealed that EGFR signaling pathway inhibition reduces androgen-independent DU145 and PC3 prostate cancer cell growth through interference in G₁-S cell cycle progression phase (25).

Although the EGF +61 G>A polymorphism is localized in a noncoding region and the exact mechanism by which it affects EGF production is still undetermined, this locus seems to be involved in EGF regulation because it affects EGF expression (11, 13). It is hypothesized that this polymorphism might influence EGF through differential EGF mRNA processing and degradation or by providing a binding site for transcription factors.

Conversely, it is expected that G carriers may have a higher EGF availability in tumor environment. EGF +61 G>A polymorphism has been the subject of investigation in several case-control studies, involving other cancer types, with controversial results (11–15).

A recent investigation using prostatic tumor cell expression signatures showed an attenuated androgen signaling signature in high-grade and metastatic prostate cancer, eventually reflecting and providing support for the clinical association of grade with prognosis (26). In our study, risk for being diagnosed with a high Gleason grade is significantly increased in G carriers (aOR, 3.37; 95% CI, 1.47-7.73). Concomitantly, we observed a trend toward an increased risk to develop prostate cancer in EGF +61 G carriers (aOR, 1.94; 95% CI, 0.99-3.79), although this association was not statistically significant. Several studies have previously shown that growth factors promote tissue proliferation and malignant transformation (27). Accordingly, results from our study support that this genetic variant, which influences EGF expression and serum levels, may have an effect in prostate cancer development and in higher grade at diagnosis. In fact, some authors (28) refer that aberrant growth and differentiation are due to inappropriate cellular environment. Our results suggest that EGF +61 G>A polymorphism, which represents a functional modification in EGF expression since birth and throughout life, exposes AG/GG carriers to higher EGF concentrations, inducing modifications in prostate normal and tumoral cell microenvironment, leading to different patterns of cell differentiation, and contributing to a higher Gleason grade at diagnosis.

We found a significantly higher risk for metastasis in G carriers (P = 0.043). Growth factors, such as EGF, have an important role in migration and invasion of cancer cells. They can disrupt different pathways and contribute to metastization. The EGFR can interact with the integrin α₆β₄ and promote cell migration through activation of phosphatidylinositol 3-kinase and other downstream pathways (29–31). Cumulatively, previous studies showed a significant association between EGFR activation and the acquisition of the invasive phenotype (32–34).

We hypothesize that in G carriers, which express significantly more EGF, the EGF-EGFR pathway may be up-regulated, inducing a metastatic tumor phenotype.

The HRPC is a multistep and multievent process with different molecular patterns throughout development (35), involving changes in growth factor signaling. Several reports support an interaction of EGFR and EGF-EGFR pathway with androgen receptor (AR) pathway in prostate cancer cells, although the role of AR in hormone resistance development is still controversial. Furthermore, AR is also overexpressed in hormone-resistant tumors (36). Recently, Bonaccorsi et al. (37) showed that expression of AR affects clathrin-mediated endocytosis pathway of EGFR, suggesting that AR-positive prostatic cells have a less aggressive phenotype due to interference in proinvasive EGFR signaling (37). Moreover, studies in prostate cancer cell cultures indicated that androgen deprivation could favor the progression to HRPC states by up-regulating the EGFR pathway (38).

Our results of earlier hormone resistance in hormone-treated prostate cancer patients agree with these previous findings because the lack of ligands to AR due to ABT will decrease AR signaling, inducing a more aggressive phenotype due to EGF-EGFR alternative signaling pathway. Alternatively, we hypothesize that because EGF AG/GG carriers have increased expression of EGF and because the AR pathway can be activated by other growth factors, including EGF, tumoral cells of G carriers may be exposed to higher EGF concentrations through both autocrine and paracrine mechanisms, inducing a faster progression to hormone resistance.

Taken together, these results suggest that simultaneous blockage of AR and EGFR pathways may be appropriate for prostate cancer treatment with effect in the time to HRPC development and in prognosis. Accordingly, in vitro and in vivo studies evidence growth suppression in HRPC cells, after tyrosine kinase inhibitor or anti-EGFR monoclonal antibody therapies combined with other chemotherapeutic drugs (39–42). Nevertheless, clinical trials of tyrosine kinase inhibitor in hormone-refractory metastatic prostate cancer patients did not effectively increase clinical response (43, 44). Further studies in larger populations and varying chemotherapeutic regimens to include tyrosine kinase inhibitor or EGFR monoclonal antibodies should be undertaken to ascertain the relevance of EGF-EGFR pathway in HRPC progression and the effectiveness of anti-EGFR treatments in these patients.

The present study suggests that the EGF functional polymorphism may contribute to the establishment of a genetic profile to ABT efficacy in prostate cancer patients before their involvement in hormonal treatment and supports the involvement of EGF as an alternative molecular pathway in hormone-resistant prostatic tumors. Further studies in larger samples are required to strengthen these results and guide future investigations.

No potential conflicts of interest were disclosed.