Urinary prostate-specific antigen and microseminoprotein-beta levels in men with and without prostate cancer: A prospective cohort study.

INTRODUCTION: The role of urinary proteomics in the diagnosis of prostate cancer (PCa) is undefined. Levels of urinary biomarkers such as prostate-specific antigen (PSA) and microseminoprotein-beta (MSMB) may differ between men with and without PCa. We tested this hypothesis using urine samples before and after digital rectal examination (DRE) in men with an indication for prostate biopsy.
MATERIALS AND METHODS: In an institutional ethics committee approved prospective cohort study, men with elevated PSA or a nodule on DRE underwent a pre- and post-DRE urine sample examination for urinary PSA and MSMB levels. Levels were compared between men who had PCa diagnosed on biopsy (Group A) and those with a negative biopsy (Group B).
RESULTS: Seventy-seven patients were recruited of whom 32 had PCa (Group A) and 45 had no cancer (Group B) on biopsy. The median (interquartile range) serum PSA was 49.6 (0.2-254) ng/ml. The median urine PSA (29.5 vs. 26.4 mg/dl) and MSMB (1.7 vs. 2.4 mg/dl) were similar in both groups at baseline. However, post-DRE, both these metabolites rose in Group B but not in Group A, resulting in significantly higher post-to-pre values in Group B versus Group A. The post-DRE urine PSA/MSMB ratio was also significantly different between the groups.
CONCLUSIONS: Urinary PSA and MSMB rose significantly after DRE only in men without PCa. Post-DRE urine PSA, MSMB, and PSA/MSMB ratio can differentiate PCa from benign pathology in men with an indication for prostate biopsy. Copyright:

INTRODUCTION

Prostate cancer (PCa) is one of the most common cancers among men.[1] The diagnosis of PCa is suspected on the basis of elevated serum prostate-specific antigen (PSA) and abnormal digital rectal examination (DRE) with definitive diagnosis established on prostate biopsy. These are invasive procedures and are associated with complications.[2] Serum PSA and biopsy have certain limitations. Serum PSA has a low predictive value for PCa, especially when PSA is between 4 and 10 ng/ml with a high number of negative prostate biopsies approaching 66%.[34] Further, all cancers detected on biopsy are not aggressive and do not need treatment. Thus, a biopsy is associated with the risk of detection of clinically indolent cancers that may not require treatment as well as missing out a fraction of clinically relevant tumors since the biopsy is not always positive.

Increasing the specificity of PSA may help decrease the number of patients who undergo a biopsy and increase the yield of cancers that require treatment. Multiparametric magnetic resonance imaging (MRI) forms an important tool in this assessment. However, MRI is an expensive modality, is not always available, and is subject to interpretation errors.[5] Other noninvasive modalities include the assessment of biomarkers in biologic fluids such as serum, semen, plasma, or urine for selecting patients for prostate biopsy.

The urine sample is noninvasive, easy to collect, and readily available. Biomarkers found in urine can be DNA, RNA and protein based.[6] Urinary exosomes have been identified as a target for new biomarker acquisition in PCa. The exosomes are secreted vesicles from various cells containing micro RNA and proteins.[7] Firm pressure with gentle systematic manipulation of the prostate gland during DRE enhances expression and drainage of prostatic secretions and detached epithelial cells into the urethra, and thus, post-DRE urine is likely to be a richer source of biomarkers for PCa.[8]

Urinary PSA is one such protein biomarker. A trend toward a lower level of expression of both PSA and prostatic acid phosphatase has been reported in post-DRE urine in PCa patients. The ratio of urinary PSA to serum PSA has been reported to be significantly lower in patients with PCa and can differentiate it from benign prostatic hyperplasia (BPH) in patients with serum PSA ranging between 2.5 and 10 ng/ml.[9]

Microseminoprotein-beta (MSMB), also called prostate secretory protein of 94 amino acids, is one of the most highly secreted proteins by the prostate gland. As its expression is lost in tumorigenesis, it has been identified as a potential biomarker of PCa risk, detection, and prognosis. A common variant, rs10993994, in the 5' region of the gene which encodes MSMB, has been proposed as a risk factor for PCa.[10] MSMB expression has potential not only as a biomarker of PCa development and its prognosis but also may be used as a molecular target in therapy.[11] Low MSMB expression is associated with PCa-specific mortality in Caucasian population[12] and may have a significant association with PCa in the Asian-Indian population.[13] However, this association is not universally documented in Asian men.[14]

Since urinary MSMB secretion and PSA levels have the potential to discriminate PCa from benign etiology and there are limited and conflicting results among the Asian population, we aimed to assess the values of urinary MSMB and PSA in men undergoing prostate biopsy for suspected PCa and compare values between men with and without PCa.

MATERIALS AND METHODS

This prospective cohort study was approved by the institutional ethics committee and all individuals provided informed consent for inclusion. Men scheduled to undergo transrectal ultrasound (TRUS)-guided biopsy of the prostate for suspicion of PCa due to either elevated PSA (>4 ng/mL) or a nodule on DRE were recruited. A sterile urine culture was confirmed before inclusion. All individuals provided a morning urine specimen followed by a DRE, including prostatic massage. Prostatic massage was performed using firm pressure, sufficient to depress prostate by about 1 cm, with three strokes for each lobe. Each stroke applied from the lateral to midline and from the base to the apex for each lobe.[15] Immediately after the DRE, urine was collected as the post-DRE sample. The pre-DRE urine collected was also used for a routine urine examination. The outcome measures included the estimation of pre-DRE and post-DRE urine samples for PSA, MSMB, and their ratio for the assessment of their discriminative ability for PCa. A predetermined sample size was not calculated and it was a based on convenient sampling.

Both the pre-DRE and post-DRE samples were used for the estimation of urinary total PSA and MSMB protein level by ELISA (Human PSA ELISA kit, catalog no. orb50199, Biorbyt, U. K; ELISA kit for MSMB, Product No. SEC628Hu, USCN Life Sciences Inc., China). Collected urine was centrifuged at 1500 rpm for 10 min to sediment the bacterial contamination, crystals, cell debris, etc. One aliquot was kept for the determination of urine creatinine. The supernatant of the urine was filtered through pressure-based amicon filters. Initially, 100 ml of the urine was filtered through 30 kD amicon filters. The flow-through of this step was filtered again using 10 kD amicon filter. The filtration was carried out until 5 ml of the urine concentrate was left. The remnant urine was highly concentrated with protein in the range of 10–30 kD. Concentrates were aliquoted to 1 ml and stored at −80°C for long term storage.

Patients with PCa on biopsy were labeled as Group A and were staged for disease on the basis of the Gleason score of their prostate biopsy, and clinical staging was done according to the 7th American Joint Committee on Cancer version.[16] The laboratory investigators were blinded to the disease and clinical grouping of the patients. All data were entered into a prospective database.

Statistical analysis

The data are presented as mean (± standard deviation [SD]) or median (interquartile range [IQR]) as appropriate. Paired or unpaired t-test or Wilcoxon signed rank-sum test were used to compare quantitative variables while Chi-square/Fisher's exact test was used for qualitative variables. The Pearson's correlation coefficient or Spearman's correlation coefficient was used for the correlation of two independent variables. A P < 0.05 was considered statistically significant. The statistical analysis was done using STATA 11.2 software (StataCorp. 2009. Stata Statistical Software: Release 11. College Station, Texas, USA).

RESULTS

Eighty patients were included in the study. Three patients had very high serum PSA values (19,005, 6283, 3359 ng/ml) and were excluded from the analysis as outliers. Of 77 patients, 32 had cancer and were included in Group A while 45 with a negative biopsy were included in Group B. The mean age (± SD) was 64.9 (±7.6) years and median (IQR) serum PSA was 49.6 (0.2–254) ng/ml. The median (IQR) serum PSA was 41.4 (1.15–769.2) ng/ml in Group A and 7.6 (0.01–37.06) ng/ml in Group B. The mean prostate volume (±SD) was 52.8 (±31.5) cc in Group A while 49.4 (±21.9) cc in Group B. Both the age and the prostate volume were comparable between the two groups [Table 1]. Twenty-four patients (75%) in Group A had abnormal DRE, while 27 patients (60%) of Group B had abnormal DRE findings. MRI was done in 20 patients (62.5%) in Group A and 21 patients (46.7%) in Group B. All patients underwent TRUS-guided biopsy of the prostate with at least 12 cores, with or without MRI fusion. One patient (3.2%) of Group A underwent TRUS-guided biopsy with MRI fusion while 12 patients (26.7%) underwent TRUS-guided biopsy with MRI fusion in Group B.

Table 1

Urinary prostate-specific antigen and β-microseminoprotein before and after digital rectal examination

Parameter	Group A	Group B	P (between groups)
Number of patients, n	32	45	-
Age (years), mean±SD	66.6±6.6	63.7±7.3	0.07
Serum PSA, median (IQR), ng/ml	41.5 (1.2-769.2)	7.6 (0.01-37)	0.001
Abnormal digital rectal examination, n (%)	24 (75)	18 (40)	0.02
Prostate volume (cc), mean±SD	52.8±31.5	49.4±21.9	0.57
Patients undergoing multiparametric MRI, n (%)	20 (62.5)	21 (46.7)	-
Patients undergoing MRI-TRUS fusion biopsy, n (%)	01 (3.2)	12 (26.7)	-
Urine PSA, median (IQR), mg/dl
Pre-DRE	29.5 (0.3-229.6)	26.4 (0.2-299.8)	0.53
Post-DRE	20.03 (0.2-181.1)	107.9 (2.4-617.5)	0.001
P (within groups)	0.21	0.001	-
Urine MSMB, median (IQR), mg/dl
Pre-DRE	1.7 (0.2-8.5)	1.5 (0.1-21.1)	0.46
Post-DRE	2.4 (0.1-9.8)	4.9 (0.4-27.5)	0.004
P (within groups)	0.06	0.001	-
Urine PSA/urine MSMB ratio, median (IQR)
Pre-DRE	17.5 (0.2-1163.8)	25.7 (0.2-509.06)	0.51
Post-DRE	8.9 (0.1-69.7)	22.3 (2.2-63.9)	0.0002
Urine PSA/serum PSA ratio, median (IQR)
Pre-DRE	0.6 (0.1-14.8)	6.6 (0.1-88.6)	0.001
Post-DRE	0.4 (0.1-19.1)	13.9 (0.4-44273.4)	0.001

PSA=Prostate-specific antigen, MSMB=Microseminoprotein-beta, DRE=Digital rectal examination, IQR=Interquartile range; MRI=Magnetic resonance imaging, TRUS=Transrectal ultrasonography, SD=Standard deviation

Comparison of urinary prostate-specific antigen and microseminoprotein-beta before and after digital rectal examination

The expression of urine PSA and MSMB before and after DRE in the two groups is shown in Table 1. Baseline urinary PSA and MSMB were similar in both groups. Both urinary PSA and MSMB values rose significantly after DRE in Group B (median PSA from 26.4 to 107.9 mg/dl [P = 0.001]; median MSMB from 1.5 to 4.9 mg/dl [P = 0.001]) but not in Group A (median PSA from 29.5 to 20.03 mg/dl [P = 0.21]; median MSMB from 1.7 to 2.4 mg/dl [P = 0.06]). In between-group comparisons, while pre-DRE urinary PSA values were similar in both groups (P = 0.53), post-DRE values were significantly higher in Group B in comparison with Group A (P = 0.001). Urinary MSMB values showed a similar trend and were similar in the two groups before DRE (P = 0.46) but were significantly higher after DRE in Group B than in Group A (P = 0.001).

The ratio of urine PSA to urine MSMB was also similar in the two groups in the pre-DRE sample (median [IQR] 17.5 [0.2–1163.8] vs. 25.7 [0.2–509]) but was significantly higher in Group B in the post-DRE sample (median [IQR] 8.9 [0.1–69.7] vs. 22.3 [2.2–63.9], P = 0.002). There was no correlation between urine PSA (both pre- and post-DRE) and serum PSA in the two groups, but the ratio of urine PSA to serum PSA was significantly different between the two groups, both before and after DRE with values in Group B being significantly higher than those in Group A.

Using receiver operating curve (ROC) analysis for PSA/MSMB ratio in post-DRE urine, the area under curve (AUC) was 75% [Figure 1]. Using Youden's index, a cutoff of 13.85 had a sensitivity and specificity of 71% for identifying PCa patients.

Figure 1

Receiver operating curve analysis for prostate specific antigen/microseminoprotein-beta ratio for differentiating prostate cancer

Subgroup analysis of patients with serum prostate-specific antigen <10 ng/mL

Eight patients in Group A and 34 patients in Group B had serum PSA <10 ng/ml [Table 2]. Similar to the trend seen in the overall cohort, post DRE urinary PSA rose significantly in Group B such that while pre-DRE values were similar in the two groups, (median [IQR] 31.3 [16.2–72.3] mg/dl vs. 26.3 [10–89.2] mg/dl, P = 0.94), post-DRE values were significantly higher in Group B (median [IQR] 45.7 [5.3–58.6] mg/dl vs. 89 [18.7–250.4] mg/dl, P = 0.05). However, unlike the overall cohort, urine MSMB values were similar in both groups before and after DRE. Urine PSA to urine MSMB ratio replicated the trend of the overall cohort and was similar in the two groups before DRE (P = 0.89) but was significantly higher in Group B after DRE (P = 0.02).

Table 2

Urinary prostate-specific antigen and β-microseminoprotein before and after digital rectal examination in patients with serum prostate-specific antigen <10 ng/ml

Parameter	Group A	Group B	P (between groups)
Number of patients, n	8	34	-
Urine PSA, median (IQR), mg/dl
Pre-DRE	31.3 (16.2-72.3)	26.3 (10.0-89.2)	0.94
Post-DRE	45.7 (5.3-58.6)	89 (18.7-250.4)	0.05
Urine MSMB, median (IQR), mg/dl
Pre-DRE	2.2 (1-2.8)	1.6 (1.1-5.6)	0.97
Post-DRE	4.2 (2.2-7.5)	4.1 (1.7-10.9)	0.78
Urine PSA/urine MSMB ratio, median (IQR)
Pre-DRE	24.3 (12.6-32)	22.3 (5.4-43.7)	0.89
Post-DRE	8.1 (1.9-13.3)	21.6 (9.2-32.3)	0.02

PSA=Prostate-specific antigen, MSMB=Microseminoprotein-beta, DRE=Digital rectal examination, IQR=Interquartile range

DISCUSSION

We found that although baseline urinary PSA and MSMB were similar in the cohorts with and without PCa, both these parameters were significantly elevated in men who did not have PCa while the change was not significant among men with PCa. The post-DRE urinary PSA, urinary MSMB, and the urinary PSA/MSMB ratio were significantly different between the two groups. The post-DRE urine PSA/MSMB ratio was able to distinguish patients with PCa and among men with PSA <10 ng/ml two parameters; post-DRE urinary PSA and PSA/MSMB ratio were significantly lower in patients with PCa, suggesting a potential role of these tests in this subgroup.

Our finding of higher post-DRE urinary PSA and MSMB as compared to pre-DRE samples in patients without PCa are similar to those reported by Drake et al.[17] They characterized the constituents in postprostatic massage urine sample using proteomic studies and reported the utility of these constituents as potential biomarkers and therapeutic targets in PCa. They also reported low levels of urine PSA in post-DRE samples in patients with PCa.

The exact mechanism of this phenomenon is yet to be determined. It is possible that DRE causes liberation of PSA into the urine. In men without PCa, cellular architecture within the prostate gland is maintained with intact cell membranes, ductal anatomy, and normal drainage of prostatic secretions into the urethra. Prostatic manipulation may stimulate secretion of proteins and other molecules/exosomes into the urethra through the intact ducts, causing a rise. However, in PCa, there is cellular disarray with compression/stenosis and disruption of prostatic ducts with neovascularity and loss of cellular polarity with the release of secreted molecules into the blood circulation across the basement membrane leading to rising in serum levels of PSA. It is known that DRE results in increased serum PSA,[1819] and thus, prostatic manipulation may not liberate any additional urinary PSA, explaining the rise in post-DRE PSA specifically in men without PCa.

Low urinary PSA among men with PSA below 10 ng/mL and having PCa have also been previously reported, and Bolduc et al.[20] found urinary PSA to be significantly lower in patients with PCa as compared to BPH for men with serum PSA between 2.5 and 10 ng/ml. They also reported urinary PSA/serum PSA ratio to be significantly higher for BPH patients than for PCa patients, both overall and in the subgroup with serum PSA between 2.5 and 10 ng/ml, suggesting that urinary PSA could identify men with PCa.

Similar results have been reported with urinary MSMB.[212223] In benign prostate, MSMB regulates cell growth by promoting apoptosis while in malignancy, there is loss or decreased MSMB expression, leading to uncontrolled growth of cells. This difference becomes more prominent in post-DRE urine samples, and the rationale appears to be similar to that for urinary PSA. Prostatic massage liberates MSMB in men with normal glands (without PCa), causing a rise while there is no additional release in men with PCa, thus heightening the difference between the two groups. Previous studies have reported higher levels of urinary MSMB in benign/normal prostatic tissue or serum than in tissue or serum from PCa patients.[212223] Whitaker et al.[21] studied urinary levels of MSMB in 215 patients without PCa (normal/BPH) and 89 patients with PCa. They reported a significant decrease in urinary MSMB in patients with PCa than those without PCa (P < 0.001), consistent with findings in prostatic tissue. They also reported urinary MSMB to be significantly better as compared to urinary PSA but not better than serum PSA for the diagnosis of PCa on analysis of ROC curves.[21]

The role of urine PSA and MSMB ratio has been scarcely reported in literature. As this ratio includes two parameters, urine PSA and urine MSMB, both of which differ significantly between cancer and controls independently, it is likely that the ratio may accentuate the discriminatory ability, and our findings of significance on all three parameters confirm these assumptions.

We found no correlation between urine and serum PSA in either group. Bolduc et al. reported urinary PSA/serum PSA ratio to be significantly higher for BPH patients than for PCa patients, overall and in the subgroup with serum PSA between 2.5 and 10 ng/ml.[20] Irani et al. have reported higher AUC for urine/serum PSA ratio than for total PSA or free to total PSA ratio for the diagnosis of PCa in the subgroup of patients with serum PSA between 4 and 10 ng/ml.[24] Similar to our data, these two studies suggest the diagnostic utility of urinary/serum PSA ratio in PCa detection when the serum PSA is between 2.5 and 10 ng/ml. On the contrary, Pannek et al., in their study, including 110 patients, did not find urinary-to-serum PSA ratio to improve the diagnostic and prognostic ability for PCa over serum PSA alone.[25]

This study has some limitations. First, we do not have a control group of patients who had lower urinary tract symptoms with no indication for biopsy or a group of asymptomatic age-matched individuals from the population. Second, the technique of DRE may vary between patients even if performed by the same individual owing to the variation in the size of the prostate gland and the degree of cooperation on the part of the patient. Third, the serum PSA values showed a wide variation that could have an impact on the results. Fourth, the small sample size comes with its inherent bias. Finally, some of the patients included in the non-PCa group may have PCa despite negative TRUS biopsy. Despite these limitations, our study identifies a potential role of urinary proteomics, which may noninvasively aid in identifying patients more likely to have PCa.

CONCLUSIONS

Urinary PSA and MSMB rose significantly after DRE in men without PCa but not in men with PCa. Post DRE levels were significantly different between the two groups. The ratio of urinary PSA to MSMB also showed similar trends. These noninvasive urinary biomarkers may have a role in identifying patients more likely to have PCa among men with an indication for prostate biopsy.