Aims: The association between the expression of androgen receptor (AR) or androgen-converting enzymes and malignant potential in prostate cancer (PCa) was examined.
Methods: PCa specimens from 44 cases of stage II, 10 cases of stage III, four cases of stage IV and two recurrent cases were semi-quantitatively studied with immunohistochemistry for AR and androgen-converting enzymes.
Results: The expression scores for AR, 5α-reductase type 1 (SRD5A1), 5α-reductase type 2 (SRD5A2), and aldo-keto reductase family 1 member C3 (AKR1C3) in the metastatic lesion of stage IV or recurrent cancer (n = 6) were 284.2 (30.1), 300 (0.0), 279.2 (51) and 254.2 (74.9), respectively; these scores were significantly higher than the respective scores of 121.8 (82.1), 135.1 (59.7), 167.0 (66.4) and 150.5 (62.8) for stage II and III cancer (n = 54) (p<0.001, p<0.001, p = 0.002 and p = 0.018, respectively). The expression scores for AR and SRD5A1 in stage II and III cancer with Gleason score 7 (n = 19) were 128.7 (72.3) and 150.5 (52.9); these were significantly higher than the scores of 78.8 (67.2) and 100.0 (39.6), respectively, for cancers with a Gleason score of ⩽6 (n = 20) (p = 0.032 and p = 0.002, respectively). The expression scores for AR, SRD5A1 and AKR1C3 in stage II and III cancer with primary Gleason pattern ⩾4 (n = 21) were 158.1 (84.3), 158.3 (61.1) and 173.8 (64.8); these were significantly higher than the scores of 98.6 (72.8), 120.3 (54.7) and 135.6 (57.6), respectively, for cancers with primary Gleason pattern ⩽3 (n = 33) (p = 0.011, p = 0.026 and p = 0.034, respectively). Within Gleason score 9 cancer, the expression scores for AR and SRD5A1 in the primary lesion of stage IV (n = 3) were 276.7 (5.8) and 283.3 (28.9); these scores were significantly higher than the scores of 182.1 (86.0) and 140.0 (56.6), respectively, for stage II and III cancer (n = 7) (p = 0.027 and p = 0.001, respectively).
Conclusions: Both AR and androgen-converting enzymes were upregulated in high-grade or advanced PCa.
Statistics from Altmetric.com
Prostate cancer (PCa) is the second leading cause of cancer death for men in the US. In Japan, deaths from PCa in 2020 are estimated to increase threefold compared with 2000.1 Although prostate-specific antigen (PSA) screening has led to the increased detection of early stage PCa, there are many patients with advanced disease.
Androgen and the androgen receptor (AR) play a central role in not only prostatic carcinogenesis but also disease progression.2 Androgen deprivation therapy (ADT) is the mainstay in the treatment of advanced PCa patients, and is frequently applied also to those with early disease. ADT suppresses the testicular androgen testosterone, and blocks adrenal androgens by the use of AR antagonists.3–6 The majority of patients with PCa have a clinical response to ADT, but the disease relapses in a fraction of patients with androgen-independent PCa. It has been reported that AR is detected in most patients with recurrent PCa, and the expression level of AR is inversely correlated with biochemical recurrence-free survival.2 7 8 Testosterone is converted into dihydrotestosterone (DHT), the most bioactive androgen, by 5α-reductase in the prostate gland. The 5α-reductase enzyme has two isoforms: 5α-reductase type 1 (SRD5A1) and 5α-reductase type 2 (SRD5A2).9 SRD5A2 is the predominant isoform in normal prostate tissue. Recent studies have shown that AR is further activated in the progression of PCa by a high level of DHT produced through the enhanced expression of 5α-reductase.10 11 Adrenal androgens such as androstenedione and dehydroepiandrosterone (DHEA) are converted into intraprostatic testosterone by aldo-keto reductase family 1 member C3 (AKR1C3). AKR1C3 is also known as type 5 17β-hydroxysteroid dehydrogenase.12–14 Increased expression of AKR1C3 has been shown to strongly correlate with prostatic carcinogenesis.15 16 These findings suggest that androgen-converting enzymes represented by 5α-reductase and AKR1C3 are associated with the disease progression of PCa.
We hypothesised that high-grade or advanced PCa involves an intracellular androgen environment favourable for cancer cell proliferation; however, most previous reports have underscored only one or two components of the androgen–AR pathway relevant to PCa aggressiveness.7 8 10 11 15 16 The current study is designed to clarify the distribution and expression pattern of serial components AR, SRD5A1, SRD5A2 and AKR1C3, and the detail of the androgen–AR pathway in PCa. Although Gleason scores have been shown to be an independent prognostic factor in PCa patients,17–20 the relevance of AR and androgen-converting enzyme expression to the Gleason grade has not been clarified. We have investigated the association between the Gleason grade and the expression of the AR and the aforementioned androgen-converting enzymes.
Localised PCa (stage II and III) specimens were randomly obtained from 54 patients who had undergone radical prostatectomy without prior treatments, such as hormone therapy or radiotherapy, at the Division of Urology, Niigata University Graduate School of Medical and Dental Sciences between 1999 and 2006. Clinical staging was performed according to the International Union Against Cancer guidelines (1997) based on ordinary examinations including digital rectal examination, transrectal ultrasonography, CT scan and isotoped-bone-scanning. Advanced PCa (stage IV and a recurrent case) samples were obtained from six patients by tumour removal or biopsy. Normal prostatic tissue from the peripheral zone, which did not include cancer, prostatic intraepithelial neoplasia or an atypical gland, was obtained from 20 of the patients with localised PCa described above. The procedure for this research project was approved by the ethics committee of our institution, and it conforms to the provisions of the Declaration of Helsinki. All patients gave informed consent to take part in this study.
The pathological diagnosis was determined by an experienced urological pathologist (T Kawasaki). The specimen sections examined in this study were selected according to the following criteria: (1) specimens with sufficient size of cancer area to evaluate the degree of staining (on the whole, larger than 20 mm2), (2) specimens with the highest Gleason score in multiple cancers, (3) specimens with a larger cancer area in the case of the same Gleason score in multiple cancers.
Three established human PCa cell lines LNCaP.FGC, PC-3 and DU145 were purchased from the American Type Culture Collection. It is known that LNCaP.FGC expresses AR, but neither PC-3 nor DU145 express AR. The cell lines were maintained in RPMI 1640 (Sigma-Aldrich, Irvine, UK) supplemented with 10% heat-inactivated fetal bovine serum (JRH Biosiences, Lenexa, Kansas, USA), 100 units/ml penicillin G sodium and 100 µg/ml streptomycin. For immunoblotting and cell block preparation, the cells were collected when they reached subconfluence.
The mouse monoclonal antibody for AR (H7507) was purchased from Perseus Proteomics (Tokyo, Japan). The goat polyclonal antibody for SRD5A1 (A-18) and rabbit polyclonal antibody for SRD5A2 (H-100) were purchased from Santa Cruz Biotechnology (Santa Cruz, California, USA). The mouse monoclonal antibody for AKR1C3 (NP6.G6.A6) was purchased from Sigma-Aldrich (St Louis, Missouri, USA). The chicken monoclonal antibody for actin (JLA20) was purchased from Oncogene Sciences (Uniondale, New York, USA).
Immunoblotting was performed as described previously.21 Briefly, LNCaP.FGC, PC-3 and DU145 cells were lysed in a solution containing 150 mM NaCl, 50 mM Tris/HCl (pH 8.0), 5 mM EDTA, 1% Triton X, 1 mM phenylmethylsulfonyl fluoride, 1 mM leupeptin and 1 mM aprotinin for 30 min on ice, and centrifuged at 7000 g at 4°C for 30 min. The extracted protein (20 μg) was resolved by sodium-dodecyl-sulfate–polyacrylamide-gel electrophoresis (SDS-PAGE). After electrophoresis, the protein was transferred to a polyvinylidene difluoride membrane (Amersham, Aylesbury, UK). The blots were pre-treated with 5% skimmed milk overnight, and then incubated with anti-AR and anti-actin for 1 h. Anti-AR and anti-actin antibodies were used at dilutions of 1:200 and 1:2000, respectively. The blots were incubated with secondary anti-mouse IgG horseradish-peroxidase-conjugated antibody (Amersham) in a dilution of 1:2000 for 30 min. They were washed three times with 0.01 mol/l phosphate-buffered saline (PBS), between each step of the procedure. The blots were visualised with the ECL detection system (Amersham) according to the manufacturer’s instructions.
Tissue samples were fixed at room temperature in 10% formalin for a week. They were sequentially dehydrated with an alcohol series, and embedded in paraffin. Sections (4 μm) were deparaffinised in xylene and dehydrated in descending dilutions of ethanol. Specimens were treated by incubating in citrate buffer (SRD5A1, SRD5A2 and AKR1C3: pH 6.0) at 121°C for 15 min in an autoclave. After washing in PBS, endogenous peroxidase activity was blocked by treating for 15 min with 0.3% hydrogen peroxide in absolute methanol. The sections were incubated with 10% normal goat or rabbit serum for 15 min at room temperature and reacted with the primary monoclonal or polyclonal antibody overnight at 4°C. Anti-SRD5A1, anti-SRD5A2 and anti-AKR1C3 antibodies were used for immunohistochemistry at dilutions of 1:50, 1:50 and 1:200, respectively. The sections were then incubated for 40 min at room temperature with an appropriate secondary antibody (Nichirei Corporation, Tokyo, Japan), washed three times with PBS between each step of the procedure, visualised with 0.1 mg/ml 3,3′-diaminobenzidine tetrahydrochloride (Dojindo Chemical, Kumamoto, Japan), and finally counterstained with Mayer’s haematoxylin.
The optimal condition of the anti-AR antibody (H7507) for immunohistochemical staining using paraffin sections was evaluated using a cell block prepared by the collodion bag technique. Briefly, LNCaP.FGC cells were placed in a 15 ml tube coated with collodion (Merck, Darmstadt, Germany). After the cells were centrifuged in the tube at 700 g for 10 min, the supernatant was discarded. The coated material was teased from the tube, and the coated bag was lifted with forceps. The coated bag was fixed in 10% formalin for 12 h, routinely processed with increasing grades of alcohol, cleared with xylene, and embedded in paraffin. Immunohistochemical staining using 4 μm sections was carried out using the method described above, except for the use of citrate buffer to retrieve antigens with different pH values, and anti-AR antibody (H7507) at different dilutions.
Evaluation and statistical analysis
All slides were assessed by two of authors (K Wako, and urological pathologist T Kawasaki) together. The extent of expression was scored semi-quantitatively by a combination of the staining intensity and the fraction of positive cells in the tumour area. The intensity was graded as follows: 0, no detectable signal; 1, weak signal at intermediate- to high-power field; 2, moderate signal at low- to intermediate-power field; 3, strong signal at low-power field. The score was calculated by summing each staining intensity and multiplying by the percentage of positive cells, and accordingly staining scores ranged between 0 and 300. For example, when intensity grade 3 positive cells accounted for 30% of the tumour area and intensity grade 1 positive cells represented 50%, the expression score was 140 (3×30+1×50).
Data were recorded as means (SD). Statistical comparisons were performed using Student t test. Pearson’s correlation coefficient was used in the evaluation of correlation. Values of p<0.05 were considered statistically significant.
Clinical and pathological features of localised PCa (stage II and III)
Table 1 shows the characteristics of patients with localised PCa. There were no patients with a Gleason score of ⩽4 or 10. There were no patients with high-grade prostatic intraepithelial neoplasia. None of the patients had lymph nodes or distant metastasis.
Characteristics of advanced PCa (stage IV and recurrent case)
The age of six patients with advanced PCa ranged between 63 and 78 years (mean 72.7 years), and prostate-specific antigen levels ranged between 0.37 and 1020 ng/ml (median 115.7 ng/ml) at the time of tumour removal or biopsy (table 2). The diagnosis of all cases was adenocarcinoma. Two patients (cases 1 and 2) had received no therapy, and four patients (cases 3–6) had received neoadjuvant hormone therapy. Case 4 had undergone hormone-combined radiotherapy prior to removal of a metastatic tumour. In cases 1–3, prostate biopsy specimens (ie, the primary lesion of advanced PCa) were obtained (n = 3). These three cases had lymph node or bone metastasis at the time of initial diagnosis and had not received therapy prior to prostate biopsy.
AR expression in PCa cell lines
The results for immunoblotting with H7507 showed AR protein corresponding to a molecular mass of 110 kDa in LNCaP.FGC (fig 1). No AR protein was detected in PC-3 and DU145. Immunohistochemical staining with H7507 was performed in optimal conditions, with citrate buffer condition at pH 6.5, and anti-AR (H7507) at a dilution of 1:100. A nuclear staining pattern was clearly shown in LNCaP.FGC, whereas control experiments carried out by omitting the primary antibody showed no staining of AR (fig 2). These findings suggested that the antibody is specific for human AR protein and that it can be used for paraffin-embedded tissue samples.
Expression of AR and androgen-converting enzymes
A nuclear staining pattern for AR was observed, and neither a cytoplasmic nor an anomalous staining pattern, such as a granular or a membranous pattern, was observed (fig 3A,E,I). SRD5A1, SRD5A2 and AKR1C3 expression was seen mostly in the cytoplasm (fig 3). A nuclear staining pattern was seen in some cases with high Gleason scores. There was a granular pattern for AKR1C3 in one case, but no membranous pattern was observed.
Expression of AR and androgen-converting enzymes in localised PCa (stage II and III) and advanced PCa (stage IV and a recurrent case)
The expression scores for AR, SRD5A1, SRD5A2 AKR1C3 in advanced PCa (n = 6) were 284.2 (30.1), 300 (0.0), 279.2 (51) and 254.2 (74.9), respectively; these scores were significantly higher than the respective scores of 121.8 (82.1), 135.1 (59.7), 167.0 (66.4) and 150.5 (62.8) for stage II and III cancer (n = 54) (p<0.001, p<0.001, p = 0.002 and p = 0.018, respectively) (fig 4). The expression score for each marker in advanced PCa was enhanced with or without prior neoadjuvant hormone therapy. The expression scores for these molecules were not significantly different between stage II and III PCa (data not shown). The expression score for AR in localised PCa was 121.8 (82.1), and this was significantly lower than the score of 179.0 (79.9) in normal prostatic tissue (p = 0.010). Also, AR expression in 20 cases of benign prostatic hyperplasia was higher than in cases of localised PCa (data not shown).
Association between AR or androgen-converting enzymes expression and Gleason score or primary Gleason pattern in localised PCa
The expression scores of AR, SRD5A1, SRD5A2 and AKR1C3 in localised PCa were analysed in three groups with Gleason grades of ⩽6 (n = 20), 7 (n = 19) and ⩾8 (n = 15). The scores for AR and SRD5A1 in Gleason grade 7 cancer (n = 19) were 128.7 (72.3) and 150.5 (52.9), and these were significantly higher the respective scores of 78.8 (67.2) and 100.0 (39.6) in Gleason grade ⩽6 cancer (n = 20) (p = 0.032 and p = 0.002, respectively) (fig 5A). The expression scores for AR, SRD5A1 and AKR1C3 in primary Gleason pattern ⩾4 cancer (n = 21) were 158.1 (84.3), 158.3 (61.1) and 173.8 (64.8), and these were significantly higher than the respective scores of 98.6 (72.8), 120.3 (54.7) and 135.6 (57.6) in primary Gleason pattern ⩽3 cancer (n = 33) (p = 0.011, p = 0.026 and p = 0.034, respectively) (fig 5B).
Expression of AR and androgen-converting enzymes in localised and the primary site of advanced PCa within Gleason score 9 cancer
Within Gleason score 9 cancer, the expression scores for AR and SRD5A1 in the primary lesion of stage IV (n = 3) were 276.7 (5.8) and 283.3 (28.9); these scores were significantly higher than the scores of 182.1 (86.0) and 140.0 (56.6), respectively, for localised PCa (n = 7) (p = 0.027 and p = 0.001, respectively) (fig 6).
Correlation between the expression score of AR and androgen-converting enzyme in localised PCa
It is well known that poorly differentiated or metastatic PCa is a life-threatening disease and that most of these cancers recur to be more aggressive androgen-insensitive forms during ADT.17–20 22 23 In the current study, we investigated the intracellular AR status and androgen environment in various stages of PCa.
AR is transcriptionally activated in recurrent or hormone-refractory PCa, and accelerates cancer cell proliferation, even at low androgen levels, during ADT.24 Increased expression of AR can explain the acquired anti-androgen resistance by amplifying the signal.25 Expression levels of AR have been reported to be associated with some tumour-promoting molecules and to be inversely correlated with biochemical recurrence-free survival.7 8 Our results showed increased expression level for AR in advanced PCa. We also found a relationship between the Gleason grade and AR expression in localised PCa. Interestingly, AR expression in normal prostate tissue was higher than in localised PCa. It is considered that the enhanced AR activity is associated with transcriptional activation, and also with AR co-regulators, such as steroid receptor co-activators.26–28
It has been reported that increased SRD5A1 and decreased SRD5A2 expression is observed in advanced PCa, and that increased expression of both isoforms is relevant to recurrent or metastatic PCa.29 The dominant expression level and isoform activity have been shown to shift from SRD5A2 to SRD5A1 in recurrent PCa.30 Our results showed increased expression of SRD5A1 in advanced PCa, and a relationship between the Gleason grade and SRD5A1 expression in localised PCa. These findings suggest that the high expression level of SRD5A1 promotes the conversion of testosterone into DHT in PCa with high malignant potential; however, the expression level of SRD5A2 was not different within localised PCa. SRD5A2 gene expression is consistently decreased in human PCa samples compared with benign prostatic hyperplasia or normal prostatic tissues.31 Finasteride, an SRD5A2 selective inhibitor, prevents prostatic carcinogenesis, but cancer with a high Gleason grade has been detected with higher frequency in the finasteride group than in the placebo group.32 Indeed, our study demonstrated that the expression level of AR correlated with that of SRD5A1, but not with that of SRD5A2. These findings suggest that the SRD5A2 probably has little relevance to the development of fatal PCa. Elevated expression of AKR1C3 is associated with pathogenesis of androgen-independent PCa.16 33 Enhanced intracellular conversion of adrenal androgens to testosterone is one of the mechanisms for PCa cells to adapt to androgen-deprived conditions.33 Our results showed increased expression of AKR1C3 in advanced PCa, and a relationship between Gleason grade and AKR1C3 expression in localised PCa.
It was also shown that the expression level of AR and SRD5A1 in the primary lesion of advanced PCa was higher than that in localised PCa within Gleason score 9 cancers. However, we investigated only a small number of the primary sites of advanced PCa; nevertheless, this finding implies that the androgen environment in advanced PCa is different from that in localised PCa. Advanced PCa has enhanced AR activity and androgen-converting efficiency, and this possibly produces the hormone-refractory condition. Actually, our results showed an enhanced androgen–AR pathway in advanced PCa, with or without prior neoadjuvant hormone therapy. We have previously reported the residual DHT levels in prostate tissue during ADT.34 We have also reported that adrenal androgen precursors do not directly interact with AR but are converted to DHT via the intraprostatic metabolic pathways, resulting in the induction of LNCaP activity.35 These findings suggest the relevance of the synthesis of DHT through androgen-converting enzymes in biologically aggressive PCa.
In summary, we examined the expression pattern of the components of the androgen–AR pathway. There were some limitations such as a short term follow-up issue; however, it was conclued that a poor prognostic outcome was associated with enhanced expression of the components of the androgen–AR pathway. However, the androgen–AR pathway is more complex than can be elucidated by immunohistochemical detection, and results shown in the current study relate only to the expression of the components, and not their function. Therefore, further investigations into this pathway are required.
This is thought to be the first study to demonstrate the expression of serial components of the androgen–androgen receptor (AR) pathway: AR, SRD5A1, SRD5A2 and AKR1C3, in prostate cancer (PCa).
A novel antibody for AR (H7507) was evaluated using a paraffin-embedded PCa cell line, LNCaP.FGC, which is known to express AR.
In localised PCa, AR expression correlated with SRD5A1 and AKR1C3, and these markers were associated with the Gleason grade.
In advanced PCa, expression of all markers in the metastatic lesion was higher than in localised PCa. AR and SRD5A1 expression in the primary lesion was higher than in localised PCa.
Although the androgen–AR pathway is more complex than can be elucidated by immunohistochemical detection, both AR and androgen-converting enzymes were upregulated in the high-grade or advanced PCa.
The authors thank S Momozaki, K Oyauchi and T Aoyama (Division of Cellular and Molecular Pathology, Department of Cellular Function, Niigata University Graduate School of Medical and Dental Sciences) for their excellent technical assistance.
Funding: This study was partly supported by the Program of Fundamental Studies in Health Sciences of the National Institute of Biochemical Innovation (NIBIO), by the Focus 21 project of the New Energy and Industrial Technology Development Organization (NEDO), and by the Special Coordination Fund for Science and Technology from the Ministry of Education, Culture, Sports, Science and Technology.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.