- Email: [email protected]
Semi-quantitative analysis of telomerase activity of exfoliated cells in urine of patients with urothelial cancers:
Semi-Quantitative Analysis of Telomerase Activity of Exfoliated Cells in Urine of Patients with Urothelial Cancers: Causative Factors Affecting Sensitivity and Specificity Toshiya Akao,1 Yoshiyuki Kakehi,2 Xiu-Xian Wu,2 Hidefumi Kinoshita,2 Takeshi Takahashi,2 Osamu Ogawa,1 Tetsuro Kato,1 and Osamu Yoshida2 1
Department of Urology, Faculty of Medicine, Akita University, Akita, Japan; 2Department of Urology, Faculty of Medicine, Kyoto University, Shogoin, Sakyo-ku, Japan
We previously reported that detection of telomerase activity in urinary exfoliated cells obtained from urothelial cancer patients by telomeric repeat amplification protocol (TRAP) assay is a more sensitive method of diagnosis than conventional urine cytologic examination, particularly in patients with grade 1 tumors. To establish this method as a noninvasive screening test for the diagnosis of urothelial cancers, we performed the semi-quantitative analysis of telomerase activity using Telomerase PCR ELISA™. Spontaneous voided urine samples were obtained from 65 urothelial and 58 non-urothelial cancer patients. When the mean 1 2 standard deviation of telomerase activity in urine sediments of non-urothelial cancer patients was arbitrarily determined as a cut-off, the sensitivity of TRAP enzyme-linked immunosorbent assay (ELISA) and the conventional cytology were 57% and 35%, respectively. Detection rate was significantly higher in semi-quantitative TRAP assay than in conventional cytologic examination in grade 1 cancer patients (52% vs. 5%, p 5 0.00195). False positives were detected in 5% of non-urothelial cancer patients without pyuria and in 11% of non-urothelial cancer patients with pyuria (p 5 0.395). Telomerase activity was enhanced in some cases after phenol extraction or extracting epithelial cells by using Dynabeads of macroscopic hematuria and pyuria, indicating that hematuria and pyuria might contribute to false negatives. In conclusion, the TRAP-ELISA method is superior to the standard TRAP assay in quantitativeness and simplicity of the experimental procedure. Detection of telomerase activity in urine sediments is particularly useful for the diagnosis of low-grade tumors. However, telomerase activity in patients with high grade tumors often might be underestimated due to the excessive amount of exfoliated epithelia with necrotic tissues, hematuria, and pyuria. (Urol Oncol 1997;3: 118 –124) © 1998 Elsevier Science Inc.
Address correspondence to: Dr. Osamu Yoshida, Department of Urology, Faculty of Medicine, Kyoto University, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606, Japan. Urol Oncol 1997;3:118 –124 © 1998 Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
KEY WORDS: Urothelial cancer, telomerase, semi-quantitative, causative factor
T
elomerase is expected to be a powerful tool for cancer diagnosis1–5 because of its selective expression exclusively in various cancers,6 including urothelial cancers,7 at a high frequency. Telomeric repeat amplification protocol (TRAP) assay is the convenient and time saving assay of detecting telomerase activity.6 We previously reported that telomerase activity can be detected by this assay in exfoliated cells in urine from patients with urothelial cancers and that measurement of this activity appears to be more sensitive in detecting the presence of cancer cells than standard urine cytologic examination, especially in low-grade tumors.8 Detection of telomerase activity in exfoliated cells in urine might be an alternative and more useful method for follow-up of urothelial cancers. However, there are some problems to be resolved in this new assay for clinical use. The pitfalls of this assay are quantitativeness and causative factors in urine affecting the sensitivity of detecting telomerase activity. Because the number of epithelial cells in urine varies, a quantitative detection of telomerase activity or determination of clear borderline should be performed. There also exist some elements in urine that might affect the sensitivity of detecting telomerase activity. For example, pyuria may show a false-positive result because normal lymphocytes have a subpopulation showing limited telomerase activity.6,9,10 Hematuria might be a cause of false negatives because hemes in hematuria is a major inhibitor of Taq polymerase.10 Therefore, we performed the semi-quantitative analysis of detecting telomerase activity by using the enzymelinked immunosorbent assay (ELISA) technique in urine exfoliated cells of urothelial and non-urothelial cancer patients and determined the cut-off point. Moreover, we examined the causative factors in hematuria and pyuria that affect the sensitivity of detecting telomerase activity. 1078-1439/97/$17.00 PII S1078-1439(98)00009-X
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
119
Materials and Method Materials Exfoliated cells were obtained from spontaneous voided urine of 65 urothelial cancer patients (60 bladder cancers, 4 renal pelvic cancers, and 1 ureteral cancer) and 58 non-urothelial cancer patients, including patients having benign urologic diseases. The 65 urothelial cancers consisted of 21 grade 1 tumors, 33 grade 2 tumors, and 11 grade 3 tumors. The stages of the bladder cancers were unknown. The 58 non-urothelial cancer patients consisted of 25 normal healthy volunteers and 33 patients with benign urologic disease (15 benign prostate hypertrophy, 7 urolithiasis, 7 urinary tract infection, 2 neurogenic bladder, and 2 chronic renal failure). In urothelial cancer patients, 100 mL of voided urine was collected before operations: 50 mL was subjected to TRAP assays and 50 mL was subjected to cytologic examinations. As a positive control, the human bladder cancer cell line T24 was used. This cell line was obtained from the American Type Culture Collection (ATCC) and maintained according to the supplier’s recommendation. Tumors were graded according to World Health Organization (WHO) classification.11
Semi-quantitative Analysis of Telomerase Activity Urine samples of 50 mL were centrifuged at 800 g at 4°C for 5 minutes and the pelleted materials were washed twice with 15 mL of calcium-free, magnesium-free phosphate buffered saline [PBS(2)], and were pelleted again by centrifugation. In hematuria cases, the samples were treated with Triton X-100. After adding 1/10 volume of 1% Triton X-100 and mixing gently, the samples were centrifuged. The pellets were subjected to protein extraction and then to semi-quantitative TRAP procedure. Semi-quantitative analysis of telomerase activity was performed by using Telomerase PCR ELISA™ (Boehringer, Mannheim, Germany). Telomerase PCR ELISA is based on the original TRAP method described by Kim et al.6 In this assay, telomerase activity is detected by ELISA and expressed by an absorbance (A450 nm 2 A690 nm). The details of the procedure are described in the manual of the kit. Briefly, a biotinylated polymerase chain reaction (PCR) primer is immobilized to the streptavidin-coated microtiter plate. The TRAP product immobilized to the microtiter plate is hybridized with a digoxigenin-labeled, telomeric repeat specific probe and detected with an antibody for digoxigenin that is conjugated to peroxidase. Finally, the probe is visualized by peroxidase metabolizing 3,39,5, 59-tetramethyl benzidine to form a colored reaction product. Each sample treated with RNase was also subjected to semi-quantitative TRAP as a negative control.
FIGURE 1.
The scheme of modified telomeric repeat amplification protocol (TRAP) assay with Triton X or Dynabeads (A) and with phenol extraction (B) in macroscopically hematuria and pyuria samples. A comparison between standard TRAP assay and these modified TRAP assays was performed.
and the PCR amplification step. Before proceeding to the amplification step, phenol extraction was performed. Briefly, after elongation step, equal volume of phenol/ chloroform was added and then mixed vigorously. Then the samples were centrifuged and aqueous phase was collected. Finally, the samples were ethanol precipitated, dissolved in a suitable volume of TE, and subjected to the PCR amplification step.
Extraction of Epithelium from Hematuria and Pyuria Fifty milliliters of macroscopic hematuria and pyuria samples were centrifuged at 800 g at 4°C for 5 minutes and the pelleted materials were washed twice with 15 mL of calcium-free, magnesium-free PBS(2) and were pelleted again by centrifugation. The new pellets were resuspended in 1 mL of PBS(2)/1% bovine serum albumin (BSA) and added to 50 ml of prewashed Dynabeads Anti Epithelial Cell (Dynal, Oslo, Norway) containing 3 3 107 magnetic beads coated with Ber-Ep4, which is a specific antibody of almost all kinds of epithelial cells.12 After incubating at 4°C for 30 minutes, epithelial cells attached to magnetic beads were separated from erythrocytes, lymphocytes, and other nonepithelial elements by using a magnetic device (MPC-M, Dynal). This separating procedure was repeated five times, and finally the pellets were washed with PBS(2) and subjected to standard TRAP assay.10 Finally, a comparison of telomerase activities detected by standard TRAP assay and modified TRAP assay with Dynabeads was performed in selected cases of urothelial cancer patients with macroscopic hematuria and pyuria.
Standard TRAP Assay Phenol Extraction A comparison of telomerase activities detected by standard TRAP assay10 and modified TRAP assay with phenol extraction was performed in selected cases of urothelial cancer patients with macroscopic hematuria and pyuria (Figure 1). TRAP reaction consists of the elongation step
In the examination of causative factors in hematuria and pyuria, telomerase activity was examined by standard TRAP assay as described previously.10 The amplified products were separated in 10% nondenaturing acrylamide gels and transferred to nylon membranes (HybondN1, Amersham, Bucks, UK). Detection of telomerase reaction prod-
120
Akao et al
Urol Oncol 1997;3:118 –124
TABLE 1. POSITIVE RATES OF TRAP ELISA IN URINE SEDIMENTS Tumor
37/65 (57%)*
Normal BPH Urolithiasis UTI Others† NHV
FIGURE 2. Distribution of telomeric repeat amplification protocol (TRAP) enzyme-linked immunosorbent assay (ELISA) in exfoliated cells. There was statistically significant correlation between absorbances of urothelial and non-urothelial cancer patients (Mann-Whitney test, p , 0.0001). The mean 1 2 standard deviation (0.232) of non-urothelial cancer patients was arbitrarily determined as a cut-off point of telomerasepositive. N 5 58 non-urothelial cancer patients; T 5 65 urothelial cancer patients; OD 5 absorbance detected by TRAP ELISA; dotted line 5 cut-off point (0.232); –– 5 means of telomerase activity in cancer and noncancer patients (non-urothelial cancer patients, 0.082; urothelial cancer patients, 0.542). ucts was performed by chemiluminescence methods (Imaging High, Toyobo, Osaka, Japan).8
Lysate Dilution Experiment We examined the relationship between the telomerase activity and lysate concentration by serial attenuation of lysate in T24 bladder cancer cell line. The lysate of 9 mg/mL, obtained from 107 cells of T24, was attenuated by lysis buffer into some concentrations (9 –1.5 mg/mL), and then telomerase activity of each lysate was detected by standard TRAP assay. The lysate of 0.6 mg/mL, obtained from 106 cells of T24, was also attenuated by lysis buffer (106 to 101 cells), and then each lysate was subjected to standard TRAP assay.
Statistical Analysis Comparison of the absorbance obtained in two groups was carried out by use of Mann-Whitney test. Comparison of the positive rates detected by the two procedures was performed by McNemar’s or Chi-square test.
Results
Note: The mean 1 2 standard deviation of telomerase activity in normal urine sediments was arbitrarily determined as a cut-off point. BPH 5 benign prostate hypertrophy; UTI 5 urinary tract infection; NHV 5 normal healthy volunteers. *Significantly different from one another (Chi-square test, p < 0.001). †Others consist of two neurogenic bladder patients and two chronic renal failure patients.
urothelial cancer patients (Mann-Whitney test, p , 0.0001). When the mean 1 2 standard deviation (m 1 2SD, 0.232) was arbitrarily determined as a cut-off point, telomerase activity was detected in 57% of urothelial cancer patients, as shown in Table 1, which was similar to the ratio that we previously reported.8 In non-urothelial cancer patients, telomerase activity was detected in 7% (false positives). These false-positive samples consisted of two benign prostate hypertrophy patients, one renal failure patient, and one normal healthy volunteer. Table 2 shows the positive rates of TRAP ELISA and cytology according to histologic grade of tumors. In grade 1 tumors, the positive rate of TRAP assay was significantly higher than that of cytologic examination (McNemar’s test, p 5 0.00195). In grade 2 tumors, the positive rate of TRAP assay was also higher than that of cytology, although it was not significant (McNemar’s test, p 5 0.118). In grade 3 tumors, however, cytologic examination revealed the higher positive rate than TRAP ELISA assay (McNemar’s test, p 5 0.25). The absorbances of exfoliated cells in grade 3 patients also revealed lower value than grades 1 and 2 patients (data not shown).
Causative Factors Affecting False Positives and Negatives Eleven percent of non-urothelial cancer patients with macroscopic hematuria and/or pyuria and 5% of non-urothelial
TABLE 2. SENSITIVITY OF TRAP ELISA AND CYTOLOGY ACCORDING TO HISTOLOGIC GRADE
Semi-quantitative Analysis of Telomerase Activity in Urine Exfoliated Cells Telomerase activity was measured by ELISA with Telomerase PCR ELISA and expressed by an absorbance (A450 nm 2 A690 nm). The distribution of the absorbances of urine samples is shown in Figure 2. The absorbances of TRAP ELISA in urothelial cancer patients and non-urothelial cancer patients ranged from 0.028 to 2.255 (mean absorbance, 0.542) and from 0.003 to 0.395 (mean absorbance, 0.082), respectively. There was statistically significant correlation between absorbances of urothelial and non-
4/58 (7%)* 2/15 (13%) 0/7 (0%) 0/7 (0%) 1/4 (25%) 1/25 (4%)
G1 G2 G3 Total
TRAP (%; 95% CI)
Cytology (%; 95% CI)
11/21 (52%; 32–77)* 21/33 (64%; 45–80)† 5/11 (45%; 17–70)‡ 37/65 (57%; 45–70)
1/21 (5%; 0–24)* 14/33 (42%; 25–61)† 8/11 (73%; 39–94)‡ 23/65 (35%; 24–48)
*Significantly different from one another (McNemar’s test, p 5 0.00195). †p 5 0.118. ‡p 5 0.25 (McNemar’s test). TRAP 5 telomeric repeat amplification protocol; ELISA 5 enzymelinked immunosorbent assay; 95% CI 5 95% confidence interval.
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
TABLE 3. RELATIONSHIP BETWEEN URINALYSIS AND TRAP ELISA IN PATIENTS WITHOUT MALIGNANCY TRAP Urinalysis Abnormal Pyuria Hematuria Hematopyuria Normal
Positive (%)
Negative
Total
2 (11%)* 1 0 1 2 (5%)*
16 11 2 3 38
18 12 2 4 40
*Chi-square test, p 5 0.395. TRAP 5 telomeric repeat amplification protocol.
cancer patients with normal urinalysis showed substantial levels of telomerase activity (Table 3). Although the numbers of these false-positive results were small, hematuria and pyuria did not significantly influence the false-positive rate in this series (Chi-square test, p 5 0.395). We then examined candidate causative factors that might have affected false-negative rates including protein concentration of the lysates, hematuria, and pyuria. The changes of telomerase activity of the bladder cancer cell line T24 according to the serial attenuation of lysate is shown in Figure 3. The telomerase activity at the lysate concentra-
121
tion of 0.6 mg/mL showed maximal activity. Namely, below the concentration of 0.6 mg/mL, telomerase activity was increased with the lysate concentration, whereas it was decreased with lysate concentration above 1.5 mg/mL. Figure 4 shows the results of telomerase activity of urothelial cancer patients with macroscopic pyuria detected by original TRAP assay. Samples treated with Dynabeads showed higher telomerase activity than untreated samples (cases 1 and 2), whereas the telomerase activity was decreased by the treatment with Dynabeads in case 3. Figure 5 represents the telomerase activity of urothelial cancer patients with macroscopic hematuria. Cases 1 and 3 show that false-negative results were avoided by the treatment with Dynabeads. In case 2, both Dynabeads and Triton X were preventive of false negative. Figure 6 shows the effect of phenol extraction. In the bladder cancer patient with pyuria, a false-negative result was avoidable by phenol extraction. In the bladder cancer patient without pyuria, no increase of telomerase activity was observed by Phenol extraction or Dynabeads.
Discussion Telomerase is expected to be a powerful diagnostic tool to detect cancers because of its frequent expression in vari-
FIGURE 3. Relationship between telomerase activity and protein concentration or cell numbers of lysate. Telomerase activity at the lysate concentration of 0.6 mg/mL (106 cells for cell lysis) showed the maximal activity. Above the concentration of 1.5 mg/mL, telomerase activity showed gradual decrease, suggesting that inhibitory substances of telomeric repeat amplification protocol assay exist.
122
Akao et al
Urol Oncol 1997;3:118 –124
FIGURE 4. Representative figure of telomeric repeat amplification protocol assay of bladder cancer patients with macroscopic pyuria. In cases 1 and 2, higher telomerase activities were detected by extracting epithelial cells with Dynabeads. In case 3, telomerase activity was decreased by treatment of Dynabeads.
FIGURE 5.
Representative figure of telomeric repeat amplification protocol assay of bladder cancer patients with macroscopic hematuria. In cases 1 and 3, higher telomerase activities were detected by extracting epithelial cells with Dynabeads. In case 2, increased telomerase activity was detected by adding Triton-X.
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
123
FIGURE 6. Representative figure of bladder cancer (BT) patients treated with phenol extraction. (A) Bladder cancer patients with macroscopic pyuria that was not treated with Dynabeads showed the obvious increase of telomerase activity by phenol extraction. Dynabeads also increased the sensitivity of detecting telomerase activity in bladder cancer patients with pyuria although the effect of phenol extraction was not observed after the treatment with Dynabeads. (B) No increase of telomerase activities by phenol extraction or Dynabeads was observed in bladder cancer patients without pyuria.
ous cancers.1–7 Preliminary studies by us8 and others13–16 clearly indicate the usefulness of detecting telomerase activity in urinary exfoliated cells obtained from urothelial cancer patients. The reported detection rates vary from 0 to 62% in voided urine and from 73 to 95% in bladder washing samples.8,13–16 Although the detection rates were higher in bladder washing samples than in voided urine, catheterization through the urethra to collect bladder washing samples is undesirable in terms of noninvasiveness as a screening test. Therefore, to increase sensitivity of detection rates, it is imperative that we investigate telomerase activity in voided urine sample. One disadvantage in applying the standard TRAP assay to clinical use is its lack of quantitativeness. The numbers of exfoliated cells in 50 mL voided urine vary a great deal,17 which results in fluctuation of the concentration of the templates for PCR amplification in TRAP assay. Ideally, quantitative PCR amplification in TRAP assay should be performed using an internal control and the amplified products should be expressed by objective values. From this viewpoint, the TRAP ELISA method used here is superior to the standard TRAP assay in that the telomerase activity is expressed by an objective value—absorbance—although its PCR amplification is not quantitative. Substances other than epithelial cells in urine may affect the detection rate of telomerase activity. Because peripheral leukocytes and lymphocytes have been re-
ported to exhibit weak telomerase activity,9,10 inflammation of urothelium (e.g., as in urinary tract infection) may result in false positives. We detected the telomerase activity in 11% of pyuria patients without urothelial cancers, which is a slightly higher rate than the false-positive rate (5%) of non-urothelial cancer patients with normal urinalysis. Although the numbers of these false-positive results were small, there was no statistically significant difference between false-positive ratios and pyuria, indicating that pyuria may not contribute to false positives. On the other hand, there may exist substances that inhibit detection of telomerase activity, which result in false negatives. In the present study, the telomerase activity was decreased in samples with highly-concentrated lysate, indicating that inhibitory substances of TRAP assay do exist. Most of these inhibitory substances were thought to be Taq polymerase inhibitors, such as heme in hematuria.10 Other substances such as proteases, urea, salts, and RNAses in urine, which may degradate the ribonucleoprotein enzyme telomerase, are considered to repress telomerase activity.13 Therefore, it is necessary to exclude those substances that cause false positives and negatives. By the semi-quantitative TRAP assay performed in this study, telomerase activity was detected in 37 of 65 urothelial cancer patients (sensitivity, 57%) and in 4 of 58 nonurothelial cancer patients (specificity, 93%) when the mean 1 2 standard deviation of telomerase activity in
124
Akao et al
noncancer patients was arbitrarily determined as a cut-off point. These ratios were similar to those previously reported.8,13–16 In grades 1 and 2 tumors, the positive ratios in semi-quantitative TRAP assay were higher than those of cytologic examination and there was a statistically significant difference between TRAP assay and cytologic examination in grade 1 tumors. In grade 3 tumors, however, semi-quantitative TRAP assay revealed a lower positive ratio than cytologic examination. The mean absorbance of TRAP ELISA in grade 3 tumors was also lower than those in grades 1 and 2 tumors. These results might be explained by the existence of inhibitory factors of TRAP assay in urine because grade 3 tumors contained hematuria and pyuria more frequently than those with grades 1 and 2 tumors. It is quite possible that with a larger number of patients, the results may be different. These results, however, suggest the possible false negatives in urothelial cancer patients with hematuria and pyuria. In some cases of bladder cancer patients with macroscopic pyuria, Dynabeads increased the sensitivity in detecting telomerase activity, although in one case, telomerase activity was decreased by treating with Dynabeads. This may result from degradation of telomerase due to excess of time in treating with Dynabeads. In some cases of bladder cancers with macroscopic hematuria, both Dynabeads and Triton X also enhanced sensitivity. Phenol extraction was effective only in bladder cancer samples with pyuria. Thus, in selected cases these modifications increased the sensitivity in detecting telomerase, which may result in excluding false negatives. In conclusion, semi-quantitative TRAP assay is promising as a new diagnostic tool for detecting urothelial cancers, especially in low-grade tumors. The modification of TRAP assay, such as in conjunction with Dynabeads and Triton X treatment or phenol extraction, has the possibility to increase the detection rate, which may result in reducing false negatives. More quantitative and convenient modification of TRAP assay and further studies on a larger number of samples are desirable before we apply this new diagnostic tool to clinical use.
Urol Oncol 1997;3:118 –124
References 1. Sugino T, Yoshida K, Bolodeoku J, et al. Telomerase activity in human breast cancer and benign breast lesions: Diagnostic applications in clinical specimens, including fine needle aspirates. Int J Cancer 1996;69:301– 6. 2. Nouso K, Urabe Y, Higashi T, et al. Telomerase as a tool for the differential diagnosis of human hepatocellular carcinoma. Cancer 1996;78:232– 6. 3. Sommerfeld HJ, Meeker AK, Piatyszek MA, et al. Telomerase activity: A prevalent marker of malignant human prostate tissue. Cancer Res 1996;56:218 –22. 4. Dickman S. Telomerase: Optimism for cancer marker grows. Ann Oncol 1995;6:624 –30. 5. Sharma HW, Maltese JY, Zhu X, et al. Telomeres, telomerase and cancer: Is the magic bullet real? Anticancer Res 1996;16: 511–5. 6. Kim NW, Piatyszek MA, Prowse KR, et al. Specific association of human telomerase activity with immortal cells and cancer. Science 1994;266:2011–5. 7. Lin T, Miyamoto H, Fujinami K, et al. Telomerase activity in human bladder cancer. Clin Cancer Res 1996;2:929 –932. 8. Kinoshita H, Ogawa O, Kakehi Y, et al. Detection of telomerase activity in exfoliated cells in urine from patients with bladder cancer. J Natl Cancer Inst 1997;89:724 –30. 9. Counter CM, Gupta J, Harley CB, et al. Telomerase activity in normal leukocytes and in hematologic malignancies. Blood 1995;85:2315–20. 10. Piatyszek MA, Kim NW, Weinrich SL, et al. Detection of telomerase activity in human cells and tumors by a telomeric repeat amplification protocol (TRAP). Methods Cell Sci 1995; 17:1–15. 11. Mostofi FK, Sobin LH, Torloni H. Histological Typing of Urinary Bladder Tumors Geneva: World Health Organization, 1973. 12. Latza U, Niedobitek G, Schwarting R, et al. Ber-EP4: New monoclonal antibody which distinguishes epithelia from mesothelial. J Clin Pathol 1990;43:213–9. 13. Mu ¨ ller M, Heine B, Heicappell R, et al. Telomerase activity in bladder cancer, bladder washings and in urine. Int J Oncl 1996;9:1169 –73. 14. Landman J, Kavaler E, Droller MJ, et al. Applications of telomerase in urologic oncology. World J Urol 1997;15:120 – 4. 15. Yang SC, Lee DH, Hong SJ, et al. Telomerase activity: A potential marker of bladder transitional cell carcinoma in bladder washes. Yonsei Med J 1997;38:155–9. 16. Yoshida K, Sugino T, Tahara H, et al. Telomerase activity in bladder carcinoma and its implication for noninvasive diagnosis by detection of exfoliated cancer cells in urine. Cancer 1997;79:362–9. ¨ zdemir E, et al. Association of in vitro 17. Okuno H, Kakehi Y, O growth potential of urinary exfoliated cells with tumor localization and intraluminal recurrence rates of urothelial cancers. J Urol 1997;158:1996 –9.
Department of Urology, Faculty of Medicine, Akita University, Akita, Japan; 2Department of Urology, Faculty of Medicine, Kyoto University, Shogoin, Sakyo-ku, Japan
We previously reported that detection of telomerase activity in urinary exfoliated cells obtained from urothelial cancer patients by telomeric repeat amplification protocol (TRAP) assay is a more sensitive method of diagnosis than conventional urine cytologic examination, particularly in patients with grade 1 tumors. To establish this method as a noninvasive screening test for the diagnosis of urothelial cancers, we performed the semi-quantitative analysis of telomerase activity using Telomerase PCR ELISA™. Spontaneous voided urine samples were obtained from 65 urothelial and 58 non-urothelial cancer patients. When the mean 1 2 standard deviation of telomerase activity in urine sediments of non-urothelial cancer patients was arbitrarily determined as a cut-off, the sensitivity of TRAP enzyme-linked immunosorbent assay (ELISA) and the conventional cytology were 57% and 35%, respectively. Detection rate was significantly higher in semi-quantitative TRAP assay than in conventional cytologic examination in grade 1 cancer patients (52% vs. 5%, p 5 0.00195). False positives were detected in 5% of non-urothelial cancer patients without pyuria and in 11% of non-urothelial cancer patients with pyuria (p 5 0.395). Telomerase activity was enhanced in some cases after phenol extraction or extracting epithelial cells by using Dynabeads of macroscopic hematuria and pyuria, indicating that hematuria and pyuria might contribute to false negatives. In conclusion, the TRAP-ELISA method is superior to the standard TRAP assay in quantitativeness and simplicity of the experimental procedure. Detection of telomerase activity in urine sediments is particularly useful for the diagnosis of low-grade tumors. However, telomerase activity in patients with high grade tumors often might be underestimated due to the excessive amount of exfoliated epithelia with necrotic tissues, hematuria, and pyuria. (Urol Oncol 1997;3: 118 –124) © 1998 Elsevier Science Inc.
Address correspondence to: Dr. Osamu Yoshida, Department of Urology, Faculty of Medicine, Kyoto University, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606, Japan. Urol Oncol 1997;3:118 –124 © 1998 Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
KEY WORDS: Urothelial cancer, telomerase, semi-quantitative, causative factor
T
elomerase is expected to be a powerful tool for cancer diagnosis1–5 because of its selective expression exclusively in various cancers,6 including urothelial cancers,7 at a high frequency. Telomeric repeat amplification protocol (TRAP) assay is the convenient and time saving assay of detecting telomerase activity.6 We previously reported that telomerase activity can be detected by this assay in exfoliated cells in urine from patients with urothelial cancers and that measurement of this activity appears to be more sensitive in detecting the presence of cancer cells than standard urine cytologic examination, especially in low-grade tumors.8 Detection of telomerase activity in exfoliated cells in urine might be an alternative and more useful method for follow-up of urothelial cancers. However, there are some problems to be resolved in this new assay for clinical use. The pitfalls of this assay are quantitativeness and causative factors in urine affecting the sensitivity of detecting telomerase activity. Because the number of epithelial cells in urine varies, a quantitative detection of telomerase activity or determination of clear borderline should be performed. There also exist some elements in urine that might affect the sensitivity of detecting telomerase activity. For example, pyuria may show a false-positive result because normal lymphocytes have a subpopulation showing limited telomerase activity.6,9,10 Hematuria might be a cause of false negatives because hemes in hematuria is a major inhibitor of Taq polymerase.10 Therefore, we performed the semi-quantitative analysis of detecting telomerase activity by using the enzymelinked immunosorbent assay (ELISA) technique in urine exfoliated cells of urothelial and non-urothelial cancer patients and determined the cut-off point. Moreover, we examined the causative factors in hematuria and pyuria that affect the sensitivity of detecting telomerase activity. 1078-1439/97/$17.00 PII S1078-1439(98)00009-X
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
119
Materials and Method Materials Exfoliated cells were obtained from spontaneous voided urine of 65 urothelial cancer patients (60 bladder cancers, 4 renal pelvic cancers, and 1 ureteral cancer) and 58 non-urothelial cancer patients, including patients having benign urologic diseases. The 65 urothelial cancers consisted of 21 grade 1 tumors, 33 grade 2 tumors, and 11 grade 3 tumors. The stages of the bladder cancers were unknown. The 58 non-urothelial cancer patients consisted of 25 normal healthy volunteers and 33 patients with benign urologic disease (15 benign prostate hypertrophy, 7 urolithiasis, 7 urinary tract infection, 2 neurogenic bladder, and 2 chronic renal failure). In urothelial cancer patients, 100 mL of voided urine was collected before operations: 50 mL was subjected to TRAP assays and 50 mL was subjected to cytologic examinations. As a positive control, the human bladder cancer cell line T24 was used. This cell line was obtained from the American Type Culture Collection (ATCC) and maintained according to the supplier’s recommendation. Tumors were graded according to World Health Organization (WHO) classification.11
Semi-quantitative Analysis of Telomerase Activity Urine samples of 50 mL were centrifuged at 800 g at 4°C for 5 minutes and the pelleted materials were washed twice with 15 mL of calcium-free, magnesium-free phosphate buffered saline [PBS(2)], and were pelleted again by centrifugation. In hematuria cases, the samples were treated with Triton X-100. After adding 1/10 volume of 1% Triton X-100 and mixing gently, the samples were centrifuged. The pellets were subjected to protein extraction and then to semi-quantitative TRAP procedure. Semi-quantitative analysis of telomerase activity was performed by using Telomerase PCR ELISA™ (Boehringer, Mannheim, Germany). Telomerase PCR ELISA is based on the original TRAP method described by Kim et al.6 In this assay, telomerase activity is detected by ELISA and expressed by an absorbance (A450 nm 2 A690 nm). The details of the procedure are described in the manual of the kit. Briefly, a biotinylated polymerase chain reaction (PCR) primer is immobilized to the streptavidin-coated microtiter plate. The TRAP product immobilized to the microtiter plate is hybridized with a digoxigenin-labeled, telomeric repeat specific probe and detected with an antibody for digoxigenin that is conjugated to peroxidase. Finally, the probe is visualized by peroxidase metabolizing 3,39,5, 59-tetramethyl benzidine to form a colored reaction product. Each sample treated with RNase was also subjected to semi-quantitative TRAP as a negative control.
FIGURE 1.
The scheme of modified telomeric repeat amplification protocol (TRAP) assay with Triton X or Dynabeads (A) and with phenol extraction (B) in macroscopically hematuria and pyuria samples. A comparison between standard TRAP assay and these modified TRAP assays was performed.
and the PCR amplification step. Before proceeding to the amplification step, phenol extraction was performed. Briefly, after elongation step, equal volume of phenol/ chloroform was added and then mixed vigorously. Then the samples were centrifuged and aqueous phase was collected. Finally, the samples were ethanol precipitated, dissolved in a suitable volume of TE, and subjected to the PCR amplification step.
Extraction of Epithelium from Hematuria and Pyuria Fifty milliliters of macroscopic hematuria and pyuria samples were centrifuged at 800 g at 4°C for 5 minutes and the pelleted materials were washed twice with 15 mL of calcium-free, magnesium-free PBS(2) and were pelleted again by centrifugation. The new pellets were resuspended in 1 mL of PBS(2)/1% bovine serum albumin (BSA) and added to 50 ml of prewashed Dynabeads Anti Epithelial Cell (Dynal, Oslo, Norway) containing 3 3 107 magnetic beads coated with Ber-Ep4, which is a specific antibody of almost all kinds of epithelial cells.12 After incubating at 4°C for 30 minutes, epithelial cells attached to magnetic beads were separated from erythrocytes, lymphocytes, and other nonepithelial elements by using a magnetic device (MPC-M, Dynal). This separating procedure was repeated five times, and finally the pellets were washed with PBS(2) and subjected to standard TRAP assay.10 Finally, a comparison of telomerase activities detected by standard TRAP assay and modified TRAP assay with Dynabeads was performed in selected cases of urothelial cancer patients with macroscopic hematuria and pyuria.
Standard TRAP Assay Phenol Extraction A comparison of telomerase activities detected by standard TRAP assay10 and modified TRAP assay with phenol extraction was performed in selected cases of urothelial cancer patients with macroscopic hematuria and pyuria (Figure 1). TRAP reaction consists of the elongation step
In the examination of causative factors in hematuria and pyuria, telomerase activity was examined by standard TRAP assay as described previously.10 The amplified products were separated in 10% nondenaturing acrylamide gels and transferred to nylon membranes (HybondN1, Amersham, Bucks, UK). Detection of telomerase reaction prod-
120
Akao et al
Urol Oncol 1997;3:118 –124
TABLE 1. POSITIVE RATES OF TRAP ELISA IN URINE SEDIMENTS Tumor
37/65 (57%)*
Normal BPH Urolithiasis UTI Others† NHV
FIGURE 2. Distribution of telomeric repeat amplification protocol (TRAP) enzyme-linked immunosorbent assay (ELISA) in exfoliated cells. There was statistically significant correlation between absorbances of urothelial and non-urothelial cancer patients (Mann-Whitney test, p , 0.0001). The mean 1 2 standard deviation (0.232) of non-urothelial cancer patients was arbitrarily determined as a cut-off point of telomerasepositive. N 5 58 non-urothelial cancer patients; T 5 65 urothelial cancer patients; OD 5 absorbance detected by TRAP ELISA; dotted line 5 cut-off point (0.232); –– 5 means of telomerase activity in cancer and noncancer patients (non-urothelial cancer patients, 0.082; urothelial cancer patients, 0.542). ucts was performed by chemiluminescence methods (Imaging High, Toyobo, Osaka, Japan).8
Lysate Dilution Experiment We examined the relationship between the telomerase activity and lysate concentration by serial attenuation of lysate in T24 bladder cancer cell line. The lysate of 9 mg/mL, obtained from 107 cells of T24, was attenuated by lysis buffer into some concentrations (9 –1.5 mg/mL), and then telomerase activity of each lysate was detected by standard TRAP assay. The lysate of 0.6 mg/mL, obtained from 106 cells of T24, was also attenuated by lysis buffer (106 to 101 cells), and then each lysate was subjected to standard TRAP assay.
Statistical Analysis Comparison of the absorbance obtained in two groups was carried out by use of Mann-Whitney test. Comparison of the positive rates detected by the two procedures was performed by McNemar’s or Chi-square test.
Results
Note: The mean 1 2 standard deviation of telomerase activity in normal urine sediments was arbitrarily determined as a cut-off point. BPH 5 benign prostate hypertrophy; UTI 5 urinary tract infection; NHV 5 normal healthy volunteers. *Significantly different from one another (Chi-square test, p < 0.001). †Others consist of two neurogenic bladder patients and two chronic renal failure patients.
urothelial cancer patients (Mann-Whitney test, p , 0.0001). When the mean 1 2 standard deviation (m 1 2SD, 0.232) was arbitrarily determined as a cut-off point, telomerase activity was detected in 57% of urothelial cancer patients, as shown in Table 1, which was similar to the ratio that we previously reported.8 In non-urothelial cancer patients, telomerase activity was detected in 7% (false positives). These false-positive samples consisted of two benign prostate hypertrophy patients, one renal failure patient, and one normal healthy volunteer. Table 2 shows the positive rates of TRAP ELISA and cytology according to histologic grade of tumors. In grade 1 tumors, the positive rate of TRAP assay was significantly higher than that of cytologic examination (McNemar’s test, p 5 0.00195). In grade 2 tumors, the positive rate of TRAP assay was also higher than that of cytology, although it was not significant (McNemar’s test, p 5 0.118). In grade 3 tumors, however, cytologic examination revealed the higher positive rate than TRAP ELISA assay (McNemar’s test, p 5 0.25). The absorbances of exfoliated cells in grade 3 patients also revealed lower value than grades 1 and 2 patients (data not shown).
Causative Factors Affecting False Positives and Negatives Eleven percent of non-urothelial cancer patients with macroscopic hematuria and/or pyuria and 5% of non-urothelial
TABLE 2. SENSITIVITY OF TRAP ELISA AND CYTOLOGY ACCORDING TO HISTOLOGIC GRADE
Semi-quantitative Analysis of Telomerase Activity in Urine Exfoliated Cells Telomerase activity was measured by ELISA with Telomerase PCR ELISA and expressed by an absorbance (A450 nm 2 A690 nm). The distribution of the absorbances of urine samples is shown in Figure 2. The absorbances of TRAP ELISA in urothelial cancer patients and non-urothelial cancer patients ranged from 0.028 to 2.255 (mean absorbance, 0.542) and from 0.003 to 0.395 (mean absorbance, 0.082), respectively. There was statistically significant correlation between absorbances of urothelial and non-
4/58 (7%)* 2/15 (13%) 0/7 (0%) 0/7 (0%) 1/4 (25%) 1/25 (4%)
G1 G2 G3 Total
TRAP (%; 95% CI)
Cytology (%; 95% CI)
11/21 (52%; 32–77)* 21/33 (64%; 45–80)† 5/11 (45%; 17–70)‡ 37/65 (57%; 45–70)
1/21 (5%; 0–24)* 14/33 (42%; 25–61)† 8/11 (73%; 39–94)‡ 23/65 (35%; 24–48)
*Significantly different from one another (McNemar’s test, p 5 0.00195). †p 5 0.118. ‡p 5 0.25 (McNemar’s test). TRAP 5 telomeric repeat amplification protocol; ELISA 5 enzymelinked immunosorbent assay; 95% CI 5 95% confidence interval.
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
TABLE 3. RELATIONSHIP BETWEEN URINALYSIS AND TRAP ELISA IN PATIENTS WITHOUT MALIGNANCY TRAP Urinalysis Abnormal Pyuria Hematuria Hematopyuria Normal
Positive (%)
Negative
Total
2 (11%)* 1 0 1 2 (5%)*
16 11 2 3 38
18 12 2 4 40
*Chi-square test, p 5 0.395. TRAP 5 telomeric repeat amplification protocol.
cancer patients with normal urinalysis showed substantial levels of telomerase activity (Table 3). Although the numbers of these false-positive results were small, hematuria and pyuria did not significantly influence the false-positive rate in this series (Chi-square test, p 5 0.395). We then examined candidate causative factors that might have affected false-negative rates including protein concentration of the lysates, hematuria, and pyuria. The changes of telomerase activity of the bladder cancer cell line T24 according to the serial attenuation of lysate is shown in Figure 3. The telomerase activity at the lysate concentra-
121
tion of 0.6 mg/mL showed maximal activity. Namely, below the concentration of 0.6 mg/mL, telomerase activity was increased with the lysate concentration, whereas it was decreased with lysate concentration above 1.5 mg/mL. Figure 4 shows the results of telomerase activity of urothelial cancer patients with macroscopic pyuria detected by original TRAP assay. Samples treated with Dynabeads showed higher telomerase activity than untreated samples (cases 1 and 2), whereas the telomerase activity was decreased by the treatment with Dynabeads in case 3. Figure 5 represents the telomerase activity of urothelial cancer patients with macroscopic hematuria. Cases 1 and 3 show that false-negative results were avoided by the treatment with Dynabeads. In case 2, both Dynabeads and Triton X were preventive of false negative. Figure 6 shows the effect of phenol extraction. In the bladder cancer patient with pyuria, a false-negative result was avoidable by phenol extraction. In the bladder cancer patient without pyuria, no increase of telomerase activity was observed by Phenol extraction or Dynabeads.
Discussion Telomerase is expected to be a powerful diagnostic tool to detect cancers because of its frequent expression in vari-
FIGURE 3. Relationship between telomerase activity and protein concentration or cell numbers of lysate. Telomerase activity at the lysate concentration of 0.6 mg/mL (106 cells for cell lysis) showed the maximal activity. Above the concentration of 1.5 mg/mL, telomerase activity showed gradual decrease, suggesting that inhibitory substances of telomeric repeat amplification protocol assay exist.
122
Akao et al
Urol Oncol 1997;3:118 –124
FIGURE 4. Representative figure of telomeric repeat amplification protocol assay of bladder cancer patients with macroscopic pyuria. In cases 1 and 2, higher telomerase activities were detected by extracting epithelial cells with Dynabeads. In case 3, telomerase activity was decreased by treatment of Dynabeads.
FIGURE 5.
Representative figure of telomeric repeat amplification protocol assay of bladder cancer patients with macroscopic hematuria. In cases 1 and 3, higher telomerase activities were detected by extracting epithelial cells with Dynabeads. In case 2, increased telomerase activity was detected by adding Triton-X.
Urol Oncol 1997;3:118 –124
Telomerase Activity of Exfoliated Cells
123
FIGURE 6. Representative figure of bladder cancer (BT) patients treated with phenol extraction. (A) Bladder cancer patients with macroscopic pyuria that was not treated with Dynabeads showed the obvious increase of telomerase activity by phenol extraction. Dynabeads also increased the sensitivity of detecting telomerase activity in bladder cancer patients with pyuria although the effect of phenol extraction was not observed after the treatment with Dynabeads. (B) No increase of telomerase activities by phenol extraction or Dynabeads was observed in bladder cancer patients without pyuria.
ous cancers.1–7 Preliminary studies by us8 and others13–16 clearly indicate the usefulness of detecting telomerase activity in urinary exfoliated cells obtained from urothelial cancer patients. The reported detection rates vary from 0 to 62% in voided urine and from 73 to 95% in bladder washing samples.8,13–16 Although the detection rates were higher in bladder washing samples than in voided urine, catheterization through the urethra to collect bladder washing samples is undesirable in terms of noninvasiveness as a screening test. Therefore, to increase sensitivity of detection rates, it is imperative that we investigate telomerase activity in voided urine sample. One disadvantage in applying the standard TRAP assay to clinical use is its lack of quantitativeness. The numbers of exfoliated cells in 50 mL voided urine vary a great deal,17 which results in fluctuation of the concentration of the templates for PCR amplification in TRAP assay. Ideally, quantitative PCR amplification in TRAP assay should be performed using an internal control and the amplified products should be expressed by objective values. From this viewpoint, the TRAP ELISA method used here is superior to the standard TRAP assay in that the telomerase activity is expressed by an objective value—absorbance—although its PCR amplification is not quantitative. Substances other than epithelial cells in urine may affect the detection rate of telomerase activity. Because peripheral leukocytes and lymphocytes have been re-
ported to exhibit weak telomerase activity,9,10 inflammation of urothelium (e.g., as in urinary tract infection) may result in false positives. We detected the telomerase activity in 11% of pyuria patients without urothelial cancers, which is a slightly higher rate than the false-positive rate (5%) of non-urothelial cancer patients with normal urinalysis. Although the numbers of these false-positive results were small, there was no statistically significant difference between false-positive ratios and pyuria, indicating that pyuria may not contribute to false positives. On the other hand, there may exist substances that inhibit detection of telomerase activity, which result in false negatives. In the present study, the telomerase activity was decreased in samples with highly-concentrated lysate, indicating that inhibitory substances of TRAP assay do exist. Most of these inhibitory substances were thought to be Taq polymerase inhibitors, such as heme in hematuria.10 Other substances such as proteases, urea, salts, and RNAses in urine, which may degradate the ribonucleoprotein enzyme telomerase, are considered to repress telomerase activity.13 Therefore, it is necessary to exclude those substances that cause false positives and negatives. By the semi-quantitative TRAP assay performed in this study, telomerase activity was detected in 37 of 65 urothelial cancer patients (sensitivity, 57%) and in 4 of 58 nonurothelial cancer patients (specificity, 93%) when the mean 1 2 standard deviation of telomerase activity in
124
Akao et al
noncancer patients was arbitrarily determined as a cut-off point. These ratios were similar to those previously reported.8,13–16 In grades 1 and 2 tumors, the positive ratios in semi-quantitative TRAP assay were higher than those of cytologic examination and there was a statistically significant difference between TRAP assay and cytologic examination in grade 1 tumors. In grade 3 tumors, however, semi-quantitative TRAP assay revealed a lower positive ratio than cytologic examination. The mean absorbance of TRAP ELISA in grade 3 tumors was also lower than those in grades 1 and 2 tumors. These results might be explained by the existence of inhibitory factors of TRAP assay in urine because grade 3 tumors contained hematuria and pyuria more frequently than those with grades 1 and 2 tumors. It is quite possible that with a larger number of patients, the results may be different. These results, however, suggest the possible false negatives in urothelial cancer patients with hematuria and pyuria. In some cases of bladder cancer patients with macroscopic pyuria, Dynabeads increased the sensitivity in detecting telomerase activity, although in one case, telomerase activity was decreased by treating with Dynabeads. This may result from degradation of telomerase due to excess of time in treating with Dynabeads. In some cases of bladder cancers with macroscopic hematuria, both Dynabeads and Triton X also enhanced sensitivity. Phenol extraction was effective only in bladder cancer samples with pyuria. Thus, in selected cases these modifications increased the sensitivity in detecting telomerase, which may result in excluding false negatives. In conclusion, semi-quantitative TRAP assay is promising as a new diagnostic tool for detecting urothelial cancers, especially in low-grade tumors. The modification of TRAP assay, such as in conjunction with Dynabeads and Triton X treatment or phenol extraction, has the possibility to increase the detection rate, which may result in reducing false negatives. More quantitative and convenient modification of TRAP assay and further studies on a larger number of samples are desirable before we apply this new diagnostic tool to clinical use.
Urol Oncol 1997;3:118 –124
References 1. Sugino T, Yoshida K, Bolodeoku J, et al. Telomerase activity in human breast cancer and benign breast lesions: Diagnostic applications in clinical specimens, including fine needle aspirates. Int J Cancer 1996;69:301– 6. 2. Nouso K, Urabe Y, Higashi T, et al. Telomerase as a tool for the differential diagnosis of human hepatocellular carcinoma. Cancer 1996;78:232– 6. 3. Sommerfeld HJ, Meeker AK, Piatyszek MA, et al. Telomerase activity: A prevalent marker of malignant human prostate tissue. Cancer Res 1996;56:218 –22. 4. Dickman S. Telomerase: Optimism for cancer marker grows. Ann Oncol 1995;6:624 –30. 5. Sharma HW, Maltese JY, Zhu X, et al. Telomeres, telomerase and cancer: Is the magic bullet real? Anticancer Res 1996;16: 511–5. 6. Kim NW, Piatyszek MA, Prowse KR, et al. Specific association of human telomerase activity with immortal cells and cancer. Science 1994;266:2011–5. 7. Lin T, Miyamoto H, Fujinami K, et al. Telomerase activity in human bladder cancer. Clin Cancer Res 1996;2:929 –932. 8. Kinoshita H, Ogawa O, Kakehi Y, et al. Detection of telomerase activity in exfoliated cells in urine from patients with bladder cancer. J Natl Cancer Inst 1997;89:724 –30. 9. Counter CM, Gupta J, Harley CB, et al. Telomerase activity in normal leukocytes and in hematologic malignancies. Blood 1995;85:2315–20. 10. Piatyszek MA, Kim NW, Weinrich SL, et al. Detection of telomerase activity in human cells and tumors by a telomeric repeat amplification protocol (TRAP). Methods Cell Sci 1995; 17:1–15. 11. Mostofi FK, Sobin LH, Torloni H. Histological Typing of Urinary Bladder Tumors Geneva: World Health Organization, 1973. 12. Latza U, Niedobitek G, Schwarting R, et al. Ber-EP4: New monoclonal antibody which distinguishes epithelia from mesothelial. J Clin Pathol 1990;43:213–9. 13. Mu ¨ ller M, Heine B, Heicappell R, et al. Telomerase activity in bladder cancer, bladder washings and in urine. Int J Oncl 1996;9:1169 –73. 14. Landman J, Kavaler E, Droller MJ, et al. Applications of telomerase in urologic oncology. World J Urol 1997;15:120 – 4. 15. Yang SC, Lee DH, Hong SJ, et al. Telomerase activity: A potential marker of bladder transitional cell carcinoma in bladder washes. Yonsei Med J 1997;38:155–9. 16. Yoshida K, Sugino T, Tahara H, et al. Telomerase activity in bladder carcinoma and its implication for noninvasive diagnosis by detection of exfoliated cancer cells in urine. Cancer 1997;79:362–9. ¨ zdemir E, et al. Association of in vitro 17. Okuno H, Kakehi Y, O growth potential of urinary exfoliated cells with tumor localization and intraluminal recurrence rates of urothelial cancers. J Urol 1997;158:1996 –9.