Prostate-specific antigen (PSA), first identified in 1970 (1), is a serine protease widely used for the early detection and monitoring of prostate cancer (2). Although previously thought to be produced exclusively by the epithelial cells of the prostate, PSA at present is considered a widespread biochemical marker present in many nonprostatic tissues and fluids (3). Studies indicate that the PSA expression in both physiological and pathological conditions is not organ- or sex-specific, but rather a steroid hormone-mediated response (4). In 1985 PSA immunoreactivity was demonstrated in urine, even though the source of urinary PSA appeared to be a local production rather than the renal filtering of serum PSA into urine (5)(6).
Considering the recent studies on PSA immunoreactivity in normal and tumor kidney tissues (7)(8), we undertook the present study on PSA content in urine samples (in particular on the first 5 mL of voided urine) collected from 50 healthy female volunteers, ages 25–39 years (mean 31 ± 3). All of these women did not have sexual intercourse for at least 1 or 2 weeks before the sample collection. The urine specimens were centrifuged at 800g for 10 min at 4 °C and the supernatants immediately stored at −30 °C until processed. Blood samples were also drawn and, after clotting, were centrifuged at 500g for 5 min at 4 °C and stored at −30 °C until assay. Another 50 healthy women, ages 27–41 years (mean 34 ± 4), had been taking for at least the preceding 6 months a widely prescribed estro-progestin oral contraceptive (Milvane®; Schering), containing both gestoden (as a progestin) and ethinylestradiol in each tablet.
The PSA determinations were carried out with two commercially available kits: a solid-phase two-site IRMA (PSA-RIACTTM; CIS Bio International) and a microparticle enzyme immunoassay (MEIA; IMx® from Abbott Labs.). The procedures of the PSA determinations, performed according to manufacturers’ recommendations, were described in detail elsewhere (9). Results, expressed as mean ± SE, were considered statistically significant when P <0.01. All statistical analyses were performed by using the StatView v.4.1 package (Abacus Concepts) on a Macintosh Power PC (Apple Computers). The present work was carried out in accordance to the ethical standard of the Helsinki Declaration of 1975, as revised in 1983.
Because the pH seemed to influence the measured PSA concentration as well as the ionic strength (2), we excluded urines with a pH <6.0 and >7.0; occasionally, we diluted the urine samples to optimize the conditions for the immunological reaction. To exclude possible matrix effects in the PSA assays (from the presence of divalent cations, proteins, and bilirubin), we performed dilutions of urines having high PSA content. The relation between PSA content and dilution showed good linearity (y = −0.139 78.6x, r = 0.962), confirming that other urinary constituents do not affect the assays’ performance. Comparing the urinary PSA values obtained with PSA-RIACT (y) and IMx-PSA (x) gave a regression equation of y = 1.11x − 0.29 (r = 0.983, P <0.001).
In agreement with recent literature (10), all serum samples examined contained PSA concentrations <0.06 μg/L and showed no statistically significant difference between the control women and the group who had been taking hormonal contraceptives. Of the 50 patients examined for each group, 80% of healthy control women contained detectable amounts of urinary PSA, with a median value of 0.035 μg/L and a mean ± SE of 0.038 ± 0.005 μg/L (range 0.02–0.15 μg/L), whereas 92% of the women taking hormonal contraceptives were positive for urinary PSA (median 0.451 μg/L, mean 0.521 ± 0.05 μg/L, range 0.09–1.239 μg/L)—a difference that was significant (t = −8.685, P <0.0001). In the contraceptive takers, the mean PSA concentration in urine samples was higher in first voided urine than in serum (t = −9.924, P <0.0001); no statistically significant difference was found between serum and urine samples in healthy control women (t −2.296, P = 0.0467), even though 8% (n = 4) of the voided urines from the contraceptive takers and 20% (n = 10) of the control women’s urines were PSA-negative. No statistically significant correlation between PSA concentrations and the women’s ages was found, even after log-transformation. Moreover, we found no significant correlation with serum PSA values in any of our urine tests.
Western blot analysis (with an anti-human PSA monoclonal antibody supplied by Dako) of the urines detected a major 33-kDa band corresponding to free PSA and occasionally a minor band of 100 kDa for α1-antichymotrypsin-bound PSA, with no additional spurious bands (data not shown).
Although previous reports demonstrated that PSA is present in undetectable or very low concentrations in nephrostomy urine specimens (6), recent data have revealed immunoreactive PSA in kidney tumor tissue extracts and sera of patients with renal cell carcinoma (7)(11). Current opinion, however, is that the human kidney does not play a significant role in eliminating (by filtration or secretion) PSA from serum into the urine, as has been confirmed by the absence of renal PSA clearance in normal and hemodialysis conditions (6)(14)(15)(16). The source of urinary PSA is probably a local production (through the periurethral Skene’s glands, the kidney and urachal structures, and the urothelium lining the bladder neck and trigone) (8)(12)(13). Moreover, the androgen control of extraprostatic production/secretion of female PSA into the urine (17) may represent another possible molecular mechanism of enhanced urinary PSA expression, related to the reported modulation of this serine protease by several steroid hormones (4)(9)(18)(19).
To our knowledge (after a careful review of literature), this is the first report concerning the oral contraceptive-induced expression of PSA in female urine at quantities measurable by commercial methods. The detectable amounts of PSA in first-voided urine samples give further evidence of the distinct presence of this serine protease in nonprostatic sources, even though its biological effects and physiological role are currently unknown.
- © 1998 The American Association for Clinical Chemistry
References
In this issue
