SEARCH

SEARCH BY CITATION

Keywords:

  • seminal vesicle;
  • radical prostatectomy;
  • PSA;
  • prostate cancer

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

OBJECTIVE

To investigate if the remaining seminal vesicle tips can affect serum levels of prostate-specific antigen (PSA) in patients after seminal vesicle-sparing radical prostatectomy (SVRP).

PATIENTS AND METHODS

Thirty-six patients were treated by either radical retropubic prostatovesiculectomy (23) or SVRP (13). Serum PSA was monitored in all patients before surgery, and at 6 weeks and 30 months afterward. Samples of normal seminal vesicles from radical cystectomies (six) were also snap-frozen and either processed for semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) using primers against PSA and α-actin (for normalization) or for PSA immunohistochemistry.

RESULTS

RT-PCR and sequencing showed that the seminal vesicles synthesise PSA mRNA. Furthermore, PSA peptide was detectable in the glandular epithelium of the seminal vesicle using immunohistochemical methods. There was no significant difference in serum PSA levels after standard or SVRP, with median (range) values (ng/mL) at 6 weeks of 0.04 (0.04–0.9) and 0.04 (0.04–0.66) and at 30 months of 0.17 (0.04–3.8) and 0.22 (0.04–58.2), respectively.

CONCLUSION

Although the seminal vesicles produce PSA, the PSA derived from the remaining seminal vesicle tips after SVRP has no effect on the oncological follow-up of these patients.


Abbreviations
SVRP

seminal-vesicle sparing radical prostatectomy.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Open radical retropubic prostatovesiculectomy is still followed by a high rate of distressing complications, presumably because of the partial destruction of nerves during surgery. Recently we therefore proposed a novel approach that spares the seminal vesicle tips, with the aim of preserving the transverse pelvic nerves [1]. As shown by us and others, this method gives improved early urinary continence [1] and sexual potency [2] rates. However, it is unclear whether the new technique has any implications for the subsequent serum levels of PSA. Originally, PSA was regarded as a prostate-specific protein and thus circulating PSA appeared to only originate from the prostate. However, there is increasing evidence that numerous normal and cancerous tissues produce PSA [3,4]. Using immunohistochemical methods, PSA was assessed in normal seminal vesicles, but the results were conflicting; PSA immunoreactivity was either weak and very rare [5], or strong, depending on the use of some (potentially cross-reacting) polyclonal antibodies [6]. Furthermore, it was reported that seminal vesicle fluid contains immunoreactive PSA [7], while in contrast, seminal vesicle agenesis does not seem to change the PSA content in seminal fluid [8]. To our knowledge, a possible contribution of seminal vesicle-derived PSA to PSA serum levels has not been investigated before.

As clarifying this topic is clinically very relevant for the follow-up of patients with prostate cancer treated by seminal-vesicle sparing radical prostatectomy (SVRP), we reinvestigated the expression of PSA in normal seminal vesicle using sensitive molecular biological and immunohistochemical methods. Furthermore, the possible effect of the remaining seminal vesicle tips after radical prostatectomy on serum PSA levels was evaluated.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Seminal vesicle samples were obtained during radical cystectomy (six) for localized bladder cancer, with no gross or microscopic evidence of prostate cancer (mean age of the patients 63.5 years, range 55–72). The study also included 36 patients with organ-confined prostate cancer of stages pT2a-b, pN0, M0, and Gleason score 4–7; 23 patients underwent radical retropubic prostatovesiculectomy (termed the ‘standard’ group) and 13 a SVRP using the published method [1]. None of the patients had chemotherapy or radiation therapy before surgery. The study was approved by the local ethics committee.

Serum PSA was monitored in all patients before, 6 weeks and 30 months after prostatectomy using an assay (Immulite 2000 PSA assay, DPC, Los Angeles, CA) that specifically detects total PSA. The results were analysed statistically by the Mann–Whitney U-test, with P < 0.05 accepted as indicating significant differences.

RT-PCR AND SEQUENCING

Samples of 10 seminal vesicles from the six other patients were snap-frozen and total RNA extracted using the Ultraspec Kit (Biotecx, Houston, Tx, USA). The first strand was synthesised (30 µL) with 2 µg total RNA and 150 ng random hexamer primers in the presence of 10 mmol/L dNTP, 1 × RT buffer (50 mmol/L Tris-HCl, pH 8.3, 75 mmol/L KCl, 3 mmol/L MgCl2, 10 mmol/L dithiothreitol) and 200 U M-MLV RT (Promega, Wallisellen, Switzerland). To test for DNA contamination, blank reactions were included with no RT enzyme. For PCR two primers were commercially synthesized (Microsynth, Balgach, Switzerland) according to the human PSA cDNA sequence, accession number NM_001648 at http://www.ncbi.nlm.nih.gov, with the sense primer as: 5′CACAGGCCAGGTATTTCAGG3′, located on exon 3, and the antisense primer as: 5′CCAGCACACAGCATGAACTT3′, located on exon 4. For multiplex PCR the QuantumRNA β-actin kit (Ambion, Huntington, UK) was used according to the manufacturer's recommendations. Briefly, PCR reactions (25 µL) were carried out in the presence of 1.5 µL of the RT reaction mixture, 0.4 µmol/L PSA primers, 0.4 µmol/L actin primer-competimer mix (from kit, ratio 3 : 7), 0.2 mmol/L dNTP, and 0.7 U HotstarTaq polymerase (Qiagen, Basel, Switzerland). After an initial preheating step of 15 min at 95 °C, there were 35 cycles at 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 1 min. Control reactions included replacing cDNA by water, followed by the PCR protocol as described above. PCR products were separated on ethidium bromide-stained 1.6% agarose gels and analysed by densitometry. A representative PCR fragment was gel-purified and commercially sequenced in both orientations (Microsynth).

IMMUNOHISTOCHEMISTRY

Tissue samples were either snap-frozen, cut at 6 µm and fixed with 4% paraformaldehyde for 10 min or immersion-fixed with 4% paraformaldehyde, routinely embedded in paraffin wax and cut at 5 µm. All sections were blocked for 30 min with 1% BSA and incubated for 72 h at 4 °C with a commercial premixed cocktail of two monoclonal antibodies (clones ER-PR8 and A67-B/E3) against human PSA (1 : 1000; Neomarkers, Fremont, CA, USA). After buffer washing, sections were incubated with biotinylated sheep antimouse IgG (1 : 100; Bioscience, Emmenbrücke, Switzerland), followed by streptavidin-horse radish peroxidase (1 : 100; Amersham, Heidelberg, Germany) for 30 min and 3,3′-diaminobenzidine (Sigma, Buchs, Switzerland) for 10 min. Control incubations included replacing the primary antibody by non-immune serum followed by the detection procedure as described. As positive controls, sections of prostate were incubated simultaneously.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

Using RT-PCR a transcript of the expected size of 345 base pairs was amplified from all samples studied (Fig. 1). The expression of PSA in the seminal vesicles varied only slightly among the patients, when normalized with the α-actin standard in semiquantitative PCR, with a mean (sd) optical density of 1.133 (0.127). Sequencing from both ends of one of the PCR fragments obtained with the PSA primers showed complete homology with the published sequence of the human PSA. The absence of bands in PCR reactions with either the negative control cDNA (no RT) or water instead of cDNA showed that there was no genomic DNA or contamination, respectively (data not shown).

image

Figure 1. The expression of PSA in the normal seminal vesicle, as shown by semiquantitative RT-PCR. Samples were collected from six patients (lanes 1, 7: patient 1; lanes 2, 4: patient 2; lanes 3, 5: patient 3; lane 8: patient 4; lanes 9, 10: patient 5; lane 11: patient 6). From all seminal vesicles investigated, a PSA band of the expected size of 345 bp (upper band) was amplified. Amplification of α-actin (294 bp, lower band) served for normalization. Lane 6: amplification of PSA from a pool of seminal vesicle cDNA. Lane 12: amplification of α−actin from a pool of seminal vesicle cDNA. The location of DNA marker bands of 400 and 300 bp are indicated with arrows.

Download figure to PowerPoint

Using the monoclonal antibody mix against PSA, there was an intense immunoreaction in the glandular epithelial cells of all seminal vesicles investigated. However, this pattern occurred only in cryosections (Fig. 2A,B). Although the specimens were dissected from different portions of the seminal vesicles, there was no gross difference in either intensity or the number of PSA immunoreactive glandular cells. In the paraffin-embedded sections the PSA immunoreactions were considerably weaker, with a patchy appearance and very few (data not shown). In incubations with non-immune serum there were no immunosignals, while incubation of prostate resulted in the known strong PSA immunoreaction (data not shown).

image

Figure 2. Immunohistochemical location of PSA in the normal seminal vesicle, using cryosections. For better visualization, sections were counterstained with methylene blue. Strong immunoreactions were located at the glandular epithelial cells (A), reduced from × 100. (B) is a higher magnification of the area indicated in (A); reduced from × 200.

Download figure to PowerPoint

The serum levels of total PSA are summarized in Table 1; the levels at diagnosis were comparable in both groups. There were no statistically significant differences in PSA levels between the groups at 6 weeks or 30 months after surgery. A PSA relapse was defined as a level of > 0.04 ng/mL in two consecutive analyses. As assessed after 30 months, there was no significant difference in PSA relapse between the groups, at four of 23 (standard) and three of 13 patients (SVRP).

Table 1.  A comparison of serum PSA levels with time in patients undergoing standard or SVRP
Median (range) PSA, ng/mLStandardSVRPP
Before surgery3.60 (0.20–15.60)3.30 (0.12–18.10)0.70
6 weeks0.04 (0.04–0.90)0.04 (0.04–0.66)0.60
30 months0.17 (0.04–3.80)0.22 (0.04–58.2)0.56

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

As the most important serum marker for prostate cancer currently available, PSA is used not only for screening and diagnosing, but also for assessing disease status after therapy [3,9]. The endogenous (prostatic and non-prostatic) sources that may contribute to PSA serum levels are thus of considerable clinical interest.

Historically, radical retropubic prostatovesiculectomy included complete removal of the seminal vesicles, but we recently proposed a new approach that spares the distal parts of the seminal vesicles [1]. To our knowledge it has never been unambiguously shown whether the seminal vesicles produce PSA and possibly release it into the circulation.

The present study is the first to RT-PCR to identify the expression of PSA in the seminal vesicles. Analyses of the nucleic acid sequence of the PCR fragment obtained confirmed its identity with PSA (both variant 1 and variant 2) as isolated from human prostate [10], thereby excluding nonspecific amplification of other PSA-related serine proteases [11] in our samples. We assume that the small differences among patients in PSA mRNA expression are caused by various amounts of glandular tissue in individual samples.

The present immunohistochemical results are in line with those from two previous studies indicating PSA expression in the seminal vesicle [5,6]. However, in those studies only 32%[6] and 11%[5] of the samples tested showed PSA immunoreactivity. Furthermore, the results were either only obtainable with a polyclonal but not with a monoclonal antibody against PSA [6], or with a mixture of a monoclonal and a polyclonal PSA antibody [5], leaving doubts about the specificity. Notably, one of the monoclonal antibodies used in both these studies [5,6] (‘ER-PR8’) was also one component of the antibody mix used in the present study. In three older immunohistochemical studies of the seminal vesicle no PSA immunoreactivity was found [12–14]. We propose that these discrepancies are a result of differences in sample preparation, as all previous immunohistochemical studies [5,6,12–14] used either formalin- or Bouin-fixed and paraffin-embedded samples. Using formalin-fixed and paraffin-embedded tissue we also obtained very weak and inconsistent PSA signals, hardly above background staining. Only the use of cryosections provided reliable and intense PSA immunolabelling that matched the clear expression pattern of PSA transcripts using RT-PCR. Thus we conclude that prolonged fixation and/or paraffin-embedding damages the antigenic epitopes of PSA in the seminal vesicle, rendering them (partly) ‘undetectable’ for the subsequent immunohistochemical procedure.

Previously it was reported that seminal vesicles with locally invasive prostate cancer express high levels of PSA immunoreactivity [14]. As our new surgical technique is restricted to organ-confined disease, we only used normal specimens with no signs of seminal vesicle invasion. Thus, the strong PSA immunoreactions detected in all of the present samples were not caused by prostate cancer invasion, but rather appear to be common to the normal seminal vesicles.

We are aware that from theses data it cannot be excluded that the remnants of the seminal vesicles may release small amounts of PSA into the blood. However, for the present routine PSA assay this possible error can be neglected, as the median PSA levels did not change significantly after surgery, and thus in a clinically relevant manner) between the groups.

In the present series there were ‘borderline’ high median PSA values after 30 months in both groups (Table 1). As patients with initial PSA values of up to 18 ng/mL were included in this study, there may have been a possible extracapsular tumour extension before surgery. As the increase in median serum PSA level after 30 months was not statistically different between the groups we conclude that the increase in serum PSA did not depend on the use of SVRP. The aim of the study was not to compare the equivalence of cancer control between the methods, as this is the subject of ongoing long-term studies with more patients and will be reported separately.

In summary, normal seminal vesicles produce PSA at expression levels that can be easily detected by molecular biological and immunohistochemical techniques. However, PSA from seminal vesicle tips after SVRP does not appear to be clinically significant for serum PSA levels. Consequently, PSA relapse after SVRP is not caused by remaining seminal vesicle tips but rather indicates the presence of residual prostatic tissue or cancer micrometastases. We conclude that serum PSA remains a reliable prostate tumour marker for the oncological follow-up of SVRP.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES

This study was supported by the Swiss National Foundation (32–59294.99 to H.J.), the Désirée and Niels Ydé Foundation and the Hartmann Müller-Foundation. We thank Ms Theresa Lehmann for excellent technical assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. REFERENCES