This article is mentioned in:

Introduction

Prostate cancer is the most common non-cutaneous malignant neoplasm, and the second leading cause of cancer mortalities in men. The majority of prostate cancers are sporadic, and etiological factors are largely unknown (1–3). Certain molecular pathways are important for the normal and pathological functioning of prostate cells, including those of the androgen receptor, estrogen receptor, transforming growth factor-β, insulin-like growth factor type 1 and phosphatidylinositol-4,5-biphosphate 3-kinase/protein kinase B (PI3K/Akt) (1,4,5).

The E-twenty-six (ETS) family of transcription factors are known to regulate various biological processes in benign and malignant tissues, including cell proliferation, differentiation, metastasis and angiogenesis (1,6–9). The prostate-derived ETS factor (PDEF) is limited to tissues that are of epithelial origin and are hormone-regulated, such as prostate, breast, salivary glands, ovaries, colon, airways and stomach tissues (7,10–15). In the prostate, PDEF is predominantly expressed in the luminal epithelium, acting as an activator of the transcription of prostate-specific antigen (PSA), either in the androgen-sensitive or androgen-independent setting (16–18). ETS variant 4 (ETV4) may be involved in chromosomal translocations in human prostate cancer, and in the activation of metalloproteinases, which are relevant in the processes of cell migration and tissue invasion, eventually leading to metastasis. ETV4 is overexpressed in a variety of cancers; in prostate cancer, ETV4 may be detected at any stage of the disease, and is typically associated with a poorer prognosis (8,19,20).

Thymidine phosphorylase (TP) and uridine phosphorylase-1 (UP-1) are different isoforms of the same enzyme. TP acts in a reversible way in the formation of thymidine and thymine. The conversion of thymidine into thymine generates 2-D-deoxyribose, which may affect a number of cellular functions, including the promotion of angiogenic factors (21–26). UP-1 acts in a reversible way in the transformation of uridine into uracil (27–29). These enzymes have various biological functions. UP-1 affects the activation and catabolism of numerous analogous nucleosides employed in anticancer chemotherapy, including fluorouracil (30).

Understanding the role of these molecules may provide novel insights into the diagnosis, staging, prognosis and follow-up of prostate cancer patients. The present study aimed to describe the relative expression of these enzymes and transcription factors in benign and malignant prostatic tissues obtained during transurethral resection of the prostate (TURP) or radical prostatectomy (RP).

Materials and methods

Population and tissue samples

A total of 52 tissue samples (31 of benign prostatic hyperplasia and 21 of prostatic adenocarcinoma) were analyzed. All samples were obtained from patients who underwent surgical treatment by TURP or RP at Hospital São Lucas of the Pontifical Catholic University of Rio Grande do Sul (Porto Alegre, Brazil). Prostatic tissue was obtained from specimens of RP, TURP or simple prostatectomy (SP) for benign disease. All patients received a summary of the study protocol and signed an informed consent form prior to surgery. The research protocol was registered with the National Committee for Research Ethics in Brazil (protocol no. 15212413.10000.5336).

Collection and storage of tissue samples

The surgical team collected the tissue samples during prostatic surgery (TURP, simple prostatectomy or RP). Prior to the transfer of the material to the Department of Pathology, the surgeon selected the tissue samples for the molecular analyses. In cancer cases, the prostate was sectioned at the most suspicious area, and a separate tissue sample was sent for conventional pathology in order to confirm that the tissue sent for molecular analysis was cancerous and not benign prostatic tissue. For the isolation of total RNA, the specimens were immediately stored in sterile flasks (Eppendorf®; Eppendorf Ltd., Stevenage, UK) with an RNA stabilization reagent (RNAlater™ RNA Stabilization Reagent, Ambion Life Technologies, Austin, TX, USA) at −80°C.

Isolation and purification of mRNA

Total RNA was isolated from ∼30 mg of each tissue sample using affinity chromatography and on-column DNAse I treatment according to the manufacturer's protocol (RNeasy® Protect Mini kit; Qiagen, Inc., Valencia, CA, USA), and stored at −80°C.

Optical density

The concentration of RNA in the samples was analyzed by spectrophotometry determined at 260 nm (1DO=30 µg/ml), calculated using the following equation: A260 × (1DO) × (dilution factor) = concentration (µg/ml or ng/µl). RNA purity was assessed by the ratio of absorbance at 260 and 280 nm. Samples with the A260/280<1.40, or those that required more than total RNA of 8.4 µl to acquire a concentration of 500 ng/µl for the subsequent step, were excluded and reprocessed up to three times. The purified RNA was stored at −80°C.

Reverse transcription and quantitative polymerase chain reaction (qPCR)

cDNA was synthesized from 500 ng of total RNA isolated using a reaction mixture containing random primers, dNTPs, reverse transcription buffer, MultiScribe™ reverse transcriptase (50 U/µl) and nuclease-free H2O + RNA (of each patient); the final concentration obtained was ∼25 ng/ml cDNA. The reaction mixtures were subjected to 25°C for 10 min, 48°C for 30 min, and 95°C for 5 min in a thermal cycler (TC-412, Techne®, Duxford, Cambridge, UK). The synthesized cDNA was stored at −20°C.

qPCR was performed in duplicate using 96-well plates. Each well contained 2 µl (50 ng) cDNA sample, 9.25 µl of MilliQ water, 12.5 µl Universal PCR Master Mix (TaqMan) with fluorescence marker (TaqMan® Gene Expression Assay-on-Demand, Applied Biosystems Life Technologies, Foster City, CA, USA) with the FAM-MGB dye (Assay-on-Demand, Applied Biosystems Life Technologies), 1.25 µl β-actin gene (Applied Biosystems Life Technologies) for the endogenous control, or 1.25 µl target gene (TP/UP1/ETV4/PDEF; Applied Biosystems Life Technologies). The final volume of each reaction was 25 µl per well. The amplification conditions were divided into the following stages: Stage 1, 30°C for 2 min; stage 2, 95°C for 10 min; stage 3, 50 cycles of 95°C for 15 s each; stage 4, 60°C for 1 min. Samples containing β-actin and with the genes of interest were amplified in parallel for the normalization of reverse transcription reactions.

Statistical analysis

Statistical analyses were conducted using SPSS version 21 (IBM SPSS, Armonk, NY, USA). Relative expression of the molecules of interest was assessed using the ∆-CT method (31), which indicated the variation of the target genes in the tissue relative to a calibrator (i.e., the normal control tissue or benign tissue). Data is expressed as the geometric mean. Data were analyzed using non-parametric tests and Spearman's correlation coefficients. P<0.05 was considered to indicate a statistically significant difference.

Results

Population and tissue samples

The relative expression of the different enzymes and ETS transcription factors (PDEF, ETV4, TP and UP-1) was determined in the prostatic tissue of 52 males treated at Hospital São Lucas. Among these, 31 samples were from patients with benign prostatic hyperplasia undergoing TURP or simple prostatectomy, and 21 tissue specimens came from patients with prostatic adenocarcinoma who underwent RP.

The mean age of the patients with prostatic hyperplasia was 66.4±10.4 years, whilst the mean age of the patients with prostatic adenocarcinoma was 62.9±8.2 years; no significant difference in age was identified between these groups (P=0.190).

Relative expression of TP, UP-1, PDEF and ETV4 in benign and malignant tissue

Table I shows the geometric means of the relative expression levels of the enzymes and ETS transcription factors assessed in the samples of benign and malignant prostatic tissue. No statistically significant difference was identified in the relative expression of TP or UP-1 between the benign and malignant tissues. However, the relative expression of PDEF and ETV4 was significantly higher in the malignant tissues compared with the benign tissues (Table I).

Table I. Associations between the benign and malignant prostatic tissue samples, with regard to the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4.

Table I.

Associations between the benign and malignant prostatic tissue samples, with regard to the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4.

	Mean relative expression
Gene	Benign tissue (n=31)	Malignant tissue (n=21)	P-valuea
TP	0.0074	0.0066	0.79
UP-1	0.0013	0.0016	0.49
PDEF	0.0311	3.1696	<0.001
ETV4	0.0045	0.0276	<0.001

a Spearman's correlation test. TP, thymidine phosphorylase; UP-1, uridine phosphorylase-1; PDEF, prostate-derived E-twenty-six factor; ETV4, E-twenty-six variant 4.

Association of the relative expression of TP, UP-1, PDEF and ETV4 with preoperative PSA levels

Tables II and III show the association between the geometric mean of the relative expression levels of the genes studied and the geometric mean of the pre-operative PSA levels in samples of benign or malignant prostatic tissue. There was no statistically significant difference between the geometric means of the relative expression levels of the different enzymes and ETS transcription factors in the samples of benign and malignant prostatic tissue, with regard to the geometric mean of the pre-operative PSA level.

Table II. Association between the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4 and the geometric mean pre-operative PSA values in benign prostatic tissue samples.

Table II.

Association between the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4 and the geometric mean pre-operative PSA values in benign prostatic tissue samples.

	PSA (n=31)a
Gene	rs	P-value
TP	0.037	0.857
UP-1	0.375	0.710
PDEF	0.040	0.853
ETV4	0.160	0.434

a Spearman's correlation test. TP, thymidine phosphorylase; UP-1, uridine phosphorylase-1; PDEF, prostate-derived E-twenty-six factor; ETV4, E-twenty-six variant 4; PSA, prostate-specific antigen; r_s, Spearman's correlation coefficient.

Table III. Association between the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4 and the geometric mean pre-operative PSA values in malignant prostatic tissue samples.

Table III.

Association between the geometric mean relative expression levels of TP, UP-1, PDEF and ETV4 and the geometric mean pre-operative PSA values in malignant prostatic tissue samples.

	PSA (n=21)a
Gene	rs	P-value
TP	0.461	0.084
UP-1	0.059	0.840
PDEF	-0.273	0.391
ETV4	-0.055	0.859

a Spearman's rank correlation test. TP, thymidine phosphorylase; UP-1, uridine phosphorylase 1; PDEF, prostate-derived E-twenty-six factor; ETV4, E-twenty-six variant 4; PSA, prostate-specific antigen; r_s, Spearman's correlation coefficient.

PDEF expression

The relative expression level of the ETS transcription factor PDEF was determined by qPCR in 28/31 (90.3%) samples from patients with benign hyperplasia and in 18/21 (85.7%) samples from patients with prostatic adenocarcinoma. The geometric mean relative expression level of PDEF was 0.0311 in the samples of benign prostatic hyperplasia and 3.1696 in the samples of prostatic adenocarcinoma (a 102-fold difference between the two tissue types). This difference was statistically significant (P<0.001) (Table I; Fig. 1).

Figure 1. Correlation between the geometric mean of the relative expression level of PDEF in patients with adenocarcinoma of the prostate and the geometric mean of the relative expression level of PDEF in patients with benign prostatic hyperplasia. PDEF, prostate-derived E-twenty-six factor.

The association between the values of the geometric mean pre-operative PSA level and the geometric mean relative expression level of PDEF in the benign and malignant tissues is shown in Tables II and III, respectively. In the cases of benign hyperplasia, there was no association between the geometric mean relative expression level of PDEF and the geometric mean pre-operative PSA level (Spearman's correlation coefficient = 0.040; P=0.853). Similarly, no significant association was identified between the mean PDEF expression level and the mean pre-operative PSA level in the cases of prostatic adenocarcinoma (Spearman's correlation coefficient = −0.273; P=0.391).

Among the cases of adenocarcinoma of the prostate, no association was identified between the pre-operative Gleason score and the relative PDEF mRNA expression level (P>0.05); there was also no correlation between TNM classification and the relative expression level of PDEF (P>0.05).

ETV4 expression

The relative expression level of ETV4 was determined by qPCR in 30/31 (96.8%) samples from the patients with benign hyperplasia and in 19/21 (90.5%) samples from the patients with adenocarcinoma of the prostate. The geometric mean relative expression level of ETV4 was 0.0045 in the samples of benign prostatic hyperplasia, and 0.0276 in the samples of prostatic adenocarcinoma; the mean relative expression level in the patients with prostatic adenocarcinoma was 6.1-fold greater than that of the patients with benign prostatic hyperplasia. This difference was statistically significant (P<0.001) (Table I; Fig. 2).

Figure 2. Correlation between the relative expression level of the geometric mean of ETV4 in patients with prostatic adenocarcinoma and the geometric mean of the relative expression level of the geometric mean of ETV4 in patients with benign prostatic hyperplasia. ETV4, E-twenty-six variant 4.

In the cases of benign hyperplasia, there was no correlation between the mean relative expression level of ETV4 and the mean pre-operative PSA level (Spearman's correlation coefficient = 0.160; P=0.434; Table II). In the cases of prostatic adenocarcinoma, there was also no correlation between the mean relative expression level of ETV4 and the mean pre-operative PSA level (Spearman's correlation coefficient = −0.055; P=0.859; Table III).

Among the cases of prostatic adenocarcinoma, no association was found between the Gleason score and the relative ETV4 expression level (P>0.05). Additionally, there was no association between the TNM classification and the relative ETV4 expression level (P>0.05).

TP expression

qPCR was used to determine the relative expression level of TP in 30/31 (96.8%) samples from the patients with benign hyperplasia and in all samples from the patients with prostatic adenocarcinoma (21/21; 100%). The mean relative expression level of TP was 0.0074 in the samples of benign prostatic hyperplasia and 0.0066 in the prostatic adenocarcinoma samples (12% lower in adenocarcinoma compared with benign prostatic hyperplasia). This difference was not statistically significant (P=0.79) (Table I).

In the cases of benign hyperplasia, there was no significant correlation between the mean relative expression level of TP and the mean pre-operative PSA level (Spearman correlation coefficient = 0.037; P=0.857; Table II). In the cases of prostate adenocarcinoma, there was a moderate correlation between the mean relative expression level of TP and the mean pre-operative PSA level (Spearman's correlation coefficient = 0.461), however, this difference was not statistically significant (P=0.084; Table III). No association was identified between Gleason score and TNM classification, and the relative expression level of TP (P=NS).

UP-1 expression

The relative expression level of UP-1 was determined by qPCR in 28/31 (90.3%) samples from the patients with benign hyperplasia and in 20/21 (95.2%) samples from the patients with prostatic adenocarcinoma. The mean relative expression level of UP-1 was 0.0013 and 0.0016 in the samples of benign prostatic hyperplasia and prostatic adenocarcinoma, respectively (26% higher in prostatic adenocarcinoma samples). However, there was no statistically significant difference between the means in these two tissue types (P=0.49) (Table I).

In the cases of benign hyperplasia, a moderate correlation was identified between the mean relative expression level of UP-1 and the mean pre-operative PSA level (Spearman's correlation coefficient = 0.375), however, this association was not statistically significant (P=0.071; Table II). In the cases of prostatic adenocarcinoma, the relative expression level of UP-1 was not found to be correlated with the mean pre-operative PSA level (Spearman's correlation coefficient = 0.059; P=0.840; Table III).

Among the adenocarcinoma cases, Gleason score was not associated with the relative expression level of UP-1 (P>0.05). However, the relative expression level of UP-1 differed significantly between T3 samples (0.0008) and T1/T2 samples (0.0024), exhibiting 65%/2.84-fold lower expression in the T3 samples compared with the T1/T2 samples (P=0.032).

Association between the expression levels of the enzymes and transcription factors studied

The relative expression level of the TP and UP-1 enzymes was significantly correlated in the benign prostatic hyperplasia and prostatic adenocarcinoma tissue samples. The relative expression levels of these enzymes in benign tissue showed a strong Spearman coefficient (0.620); in malignant tissues, the association was also strong, with a Spearman coefficient of 0.574. These associations were statistically significant in the cases of benign hyperplasia and prostatic adenocarcinoma (P<0.001 and P=0.008, respectively).

The relative expression level of TP was demonstrated to be moderately associated with the relative expression level of the PDEF and ETV4 transcription factors in the samples of benign prostatic hyperplasia. In this tissue type, the Spearman coefficient for TP and PDEF expression was determined to be 0.351; this association was not statistically significant (P=0.067). However, for TP and ETV4, this association was statistically significant (Spearman's coefficient = 0.394; P=0.035).

In the benign and malignant tissues, the relative expression level of PDEF was significantly associated with the expression level of ETV4 (Figs. 3 and 4). The relative expression level of PDEF and ETV4 in the benign tissues showed an extremely strong Spearman's coefficient, equal to 0.918 (P<0.001); for malignant tissues, a Spearman's coefficient of 0.600 was determined (P=0.009).

Figure 3. Correlation between the relative expression level of the PDEF transcription factor and the relative expression level of the ETV4 transcription factor in benign prostate tissues. PDEF, prostate-derived E-twenty-six factor. ETV4, E-twenty-six variant 4.

Figure 4. Correlation of the relative expression level of the PDEF transcription factor and the relative expression level of the ETV4 transcription factor in prostate adenocarcinoma tissues. PDEF, prostate-derived E-twenty-six factor. ETV4, E-twenty-six variant 4.

Discussion

In the present study, qPCR was utilized to assess the relative expression levels of TP, UP-1 and the ETS transcription factors PDEF and ETV4. qPCR is a standard technique used in the laboratory, and has also been used clinically for the detection of tumor markers and for the expression analysis of small samples, with a detection capability on the order of picograms and good accuracy. qPCR has been universally adopted due to its superiority with respect to speed, sensitivity, reproducibility and availability of reagents and instrumentation compared with other methods, such as northern blotting (32,33).

ETS transcription factors are considered to be challenging therapeutic targets due to their poor enzymatic activity, the complex regulation of their target genes, and their dependence on a wide network of connections and partnerships necessary for their proper functioning (1). In the current study, the expression of PDEF was found to be associated with that of ETV4 in benign and malignant prostatic tissues (P<0.001 and P<0.001, respectively). The mean relative expression level of PDEF was 102 times higher in the malignant tissues compared with the benign tissues, while ETV4 expression was 6.09 times higher in the malignant tissues compared with the benign tissues. Numerous studies have reported that the expression of PDEF decreases as the tumor becomes more aggressive and invasive, correlating with tumor grade and stage; this has been reported in prostate, breast, ovarian and colon cancer (7,10–12,34). However, the converse has also been observed in prostate, breast and ovarian cancer, suggesting a tumor suppressor activity for PDEF (7,10,35–38). The correlation coefficient of the expression of the two transcription factors in the benign tissues was strong, with P<0.001; in malignant tissue, this value indicated an even stronger correlation, with P=0.009. This indicates that these transcription factors are multifunctional, and may act via metalloproteinases, leading to migration, cell invasion and metastasis.

It is important to note that PDEF mRNA expression levels, although precise, may not necessarily reflect the level of expression of the protein in the laboratory or in clinical samples (7,39). Studies have suggested that PDEF is regulated at the transcriptional and post-translational levels via miRNAs (7,10,40).

The incidence of prostate cancer is associated with age, race, diet, environmental pollution and other factors, including a higher rate of gene fusions of ERG, ETV1, ETV4 and ETV5 with other genes (TMPRSS2, SLC45A3, C15orf21, CANT1, EST14, FOXP1, HERVK17, FLJ35294, HERV-K, ACSL3, NDRG1, DDX5, HNRPA2B1, KLK2) (41). It is now well established that PEA3 transcription factors are involved in prostate tumors and Ewing's sarcoma, as well as in other tumors, as a result of chromosomal translocations with ETV4 (8). Among the key features often found to be dysregulated in advanced prostate cancer, the PI3K and Ras pathways are altered in 40% of primary tumors and in 90% of metastatic tumors, and the combined signaling activity of these two pathways promotes the metastasis of prostate cancer through activation of ETV4 (42,43). In a study of breast cancer in humans, a positive association was observed between the overexpression of ETV4 and HER2/neu overexpression, tumor grade and greater recurrence rates (8). However, two other studies identified no correlation between ETV4, breast cancer and pathological clinical adverse effects (8,44,45).

No consensus has been reached on whether ETV4 stimulates or represses the Her2/neu pathway. It has been proposed that this difference may be due to the use of different cell lines between studies, and that overexpression of ETV4 may replace ETS factors that are most active in the promoter region of the gene Her2/neu, or may sequester certain co-activators causing the Her2/neu gene to be repressed rather than stimulated. Currently there is no consensus regarding whether the gene fusions of ERG, ETV1, ETV4 and ETV5 with other genes is correlated with more aggressive prostate cancer, or whether prostate tumors with gene translocations are more or less lethal than those without gene translocations (8,46).

In the present study, no statistically significant difference was identified between the relative expression levels of TP in benign and malignant prostatic tissues, despite the mean expression in the tumor tissues being 12% lower than that measured in the benign tissues. However, other studies have reported that the overexpression of TP may induce angiogenesis and tumor progression in organs such as the prostate, colon, pancreas, ovary, bladder, kidney, breast and stomach (21,47–49).

In addition, the current study found no statistically significant difference in the mean relative expression levels of UP-1 between benign and malignant prostatic tissues, although the mean expression of UP-1 was 26% higher in the tumor samples than in the benign samples. Studies have indicated that UP-1 is overexpressed in numerous tumors, including tumors of the breast, colon, kidney, lung, liver, ovary and intestine, in comparison to adjacent benign tissues; in certain studies, this difference was 2–3 times higher in tumors compared to normal tissue, and there were statistically significant differences in its expression in breast and colon tumors (27,50). However, negative results have also been reported (27,51).

The results of the present study indicated a moderate correlation between PSA levels and TP expression in the tumor samples, and also a moderate correlation between PSA levels and UP-1 expression in the benign tissues; however, these associations did not achieve statistical significance. The fact that benign and malignant prostatic tissues are associated with increased levels of serum PSA may explain these correlations and the lack of statistical significance in tumor and benign tissues with regard to pre-operative PSA levels (19,52,53).

In the present study, the relative expression of TP was strongly associated with the relative expression of UP-1 in the benign and tumor tissues (P<0.001 and P=0.008, respectively). This finding was expected, as the TP and UP-1 enzymes act on the nucleoside salvage pathway (54).

The relative expression of TP was moderately correlated with the relative expression of PDEF and also ETV4, the latter being statistically significant (P=0.035) in the benign prostatic tissues. We hypothesize that in benign, hyperplastic tissues, no significant changes occur in the concentrations of these enzymes and transcription factors or associated pathways.

In the prostatic carcinoma tissues, no correlation was found between the relative expression levels of TP and Gleason score or TNM classification, which is consistent with the majority of previous studies (55). In immunohistochemical studies of colorectal cancer, the expression of TP in stromal cells has been associated with a good prognosis; however, in breast cancer, increased TP expression was associated with a poor prognosis (56–58) It is likely that the expression of this enzyme varies according to the type of tissue, such as in tumor tissues infiltrated by macrophages, which exhibit an overexpression of TP (59).

The relative expression level of UP-1 showed a geometric mean 63% or 2.8 times lower in the T3 tumors compared with the T1 and T2 tumors (P=0.032); T3 tumors refer to those with extra-prostatic invasion, in generally larger tumors with increased vascularization. UP-1 is not associated with any angiogenic activity due to limited catalytic activity, which may explain its low levels in T3 disease (60,61).

It must be emphasized that there are numerous caveats limiting the conclusions that may be drawn with regard to these molecular mechanisms. All procedures involved, from the collection of the tissues to the final processing, may affect the quality of the biological sample. A number of variables may interfere at different stages of the study. Firstly, there may be variation during the surgical procedure, including anoxia and changes in local pH due to anesthesia, embolization or agglutination of arteries, sudden changes in systemic blood pressure and loss of intraoperative blood. All represent stress events that may alter the state of phosphorylation of various molecules, including that of TP and UP-1, and induce the activation or deactivation of molecular pathways. Tissue handling may also present challenges. This is performed by placing the sample as soon as possible into a sterile vial and DNase-free liquid under ice conditions; the temperature is extremely important, and storage must be at −80°C after the sample has been frozen in liquid nitrogen. This ensures the transcription of genes and prevents the degradation of DNA and RNA (62).

The findings that PDEF and ETV4 were significantly more highly expressed in prostate cancer than in benign tissue warrants further studies to define the role of these transcription factors as potential therapeutic targets in prostate cancer.

References