Molecular Determination of Perivesical and Lymph Node Metastasis after Radical Cystectomy for Urothelial Carcinoma of the Bladder

Pathological stage and lymph node status are the common primary prognostic factors after radical cystectomy or cystoprostatectomy for urothelial (transitional) bladder cancer. Survival data from several series stratifying patients by pathological tumor stage using the 1992 UICC²/AJCC classification (1) indicate that pT_2–3a patients have a 50–65% 5-year survival (Table 27.6 in Ref. 2). In contrast, patients with pT_3b-4 tumors only have a 10–35% 5-year survival (Table 27.7 in Ref. 2). In addition, the incidence of positive pelvic lymph nodes is also a function of pathological stage ranging from <8% in pT_a-1 to 10–30% in pT_2–3a and 27–64% in pT_3b-4 (Table 27.9 in Ref. 2). Five-year survival of patients with pathologically proven pelvic lymphatic metastases have ranged from 0 to 36% (Table 27.11 in Ref. 2). Two studies (3, 4) have examined the relationship of the pathological stage of the primary tumor to survival in patients with positive nodes and have found this variable to be a strong predictor of survival. Thus, accurate determination of the pathological stage including lymph node status is critically important for deriving clinically useful information after radical cystectomy.

However, the detection of pathological stage including surgical margins and perivesical extension after radical cystectomy in bladder cancer can be cumbersome in view of the size of the specimen. This leads pathologists to focus and examine microscopically only areas of visible or palpable tumor. Because up to 15% of bladder specimens (5) do not harbor a detectable primary lesion (i.e., stage pT₀), this further compromises the accurate evaluation of pathological stage. In addition, thorough examination of the excised lymphatic tissues adds to the time and expense required for the complete pathological staging of the surgical specimen. Finally, because not all excised tissue is examined microscopically, the potential for understaging does exist. These limitations as well as the desire to enhance sensitivity of staging have led to the search for an objective, rapid, reproducible, and simple assay to determine the status of surgical margins and lymph node status after radical cystectomy.

Our approach was to use RT-PCR for UPII for the analysis of tissue samples taken from the exterior of the excised radical cystectomy specimen and from the largest palpable lymph node. Uroplakins are specific differentiation products of terminally differentiated superficial urothelial cells (6). Several recent studies have examined their expression in urothelial cancers either immunohistochemically (7, 8) or using RT-PCR techniques (9, 10, 11). In any given urothelial tumor tissue sample, despite heterogeneity, uroplakins are almost always expressed to some degree. By tailoring the RT-PCR technique (i.e., nested versus nonnested) to the clinical question, this assay can be designed to detect (nested) or not detect (nonnested) UPII-positive cells in the circulating blood of patients with no clinical evidence of metastatic bladder cancer. Thus, by applying the nonnested technique, even significant blood contamination of our tissue specimens will not induce false positivity of the assay because all our patients are clinically free of metastatic disease. Herein we report our experience with this nonnested RT-PCR technique (11) and compare UPII expression in these perivesical and lymphatic tissues excised at radical cystectomy to conventional histopathological assessment of tumor stage and lymph node status.

From November 1999 through September 2000, 27 patients with clinical stage pT_a–T₃N₀M₀ underwent radical cystectomy for bladder cancer at the University of Virginia Health Sciences Center and at the University of Western Ontario Health Sciences Center. No selection criteria were applied to exclude or include participation in this study. Clinical stage was assigned using the 1997 AJC/UICC TNM System (12). All patients had computed tomography of the abdomen and/or i.v. pyelography and pathological review of their biopsies (if done outside the two participating institutions) before surgery. All patients underwent radical cystectomy with (19 of 27) or without (8 of 28) pelvic lymphadenectomy. The 3 of 9 patients who did not have pelvic lymphadenectomy had received prior bladder irradiation for bladder cancer treatment, and lymphadenectomy was deemed technically impossible, whereas in the remaining 6 cases, the harvesting was omitted because of the lymph node specimens having no distinct palpable nodes. This prospective study was approved by the Institutional Review Boards of both participating centers.

All radical cystectomy specimens were processed in a similar fashion and pathological stage assigned according to the 1997 AJC/UICC TNM System (12). After complete excision of the radical cystectomy specimen from the patient, the specimen was placed on the side table where two to three samples/side, each measuring approximately 5 × 5 × 2 mm, were removed from its surface, as shown in Fig. 1. In addition, similarly sized samples were removed from the bladder surface overlying any palpable intravesical tumor. All biopsy samples from one side (Fig. 1) were preserved in the same vial and counted as 1 sample (Table 1). Once the bilateral lymphadenectomy was completed, a 1-mm-wide cross-section from the largest and/or most clinically suspicious (for metastatic disease) samples was removed. In all cases, this sampled lymph node was included in the final histopathological analysis. Thus, a maximum of 5 perivesical samples and 1 lymph node sample were available for each patient. All tissue samples were rinsed twice in sterile saline and then placed in 1 ml of RNALater (Ambion) and maintained on crushed ice and/or 4°C until placed in −80°C 3–48 h later.

To evaluate the specificity of the UPII RT-PCR, we also evaluated patients with pelvic malignancies using similar methodologies. Two patients with colon cancer and 2 patients with adenocarcinoma of the prostate were used as such “controls.” Tissue samples were rinsed twice in sterile saline and then placed in 1 ml of RNALater (Ambion) and maintained on crushed ice and/or 4°C until placed in −80°C 3–48 h later.

Surgical samples were removed from the RNALater solution and homogenized in excess Trizol (Life Technologies, Inc.) reagent using a PCR tissue homogenizer (Omni International, Madison, WI) with disposable blades, which eliminates the cross-contamination of tissue samples during isolation procedure. RNA quantification was determined spectrophotometrically using a Beckman DU-50 spectrophotometer (Beckman). Three hundred ng of total RNA were used as a starting material for each reaction, except the UPII positive (RT4 bladder cancer cell RNA) where 10 ng were used. The primers used in RT-PCR for UPII amplification were as described by Li et al. (11). Li et al. (11) found that the technique was able to detect 1 UPII-positive RT4 bladder cancer cell line (see below) in 5 ml of normal human blood. Using this technique, no patients with clinically localized disease, as is the case for our patient cohort, had positive signals in peripheral blood, which obviated the need and expense of parallel analysis of blood samples in our study (11). RT-PCR for UPII carried out using Qiagen OneStep RT-PCR kit (Qiagen, Valencia, CA) according to the manufacturer’s recommendation. The RT-PCR conditions were compared between the program (described below) used in the current study and the earlier report (11) using the above kit in a GeneAmp PCR System 9700 (PE Applied Biosystems) machine. On the basis of the comparison, the following program was used to analyze the samples. The continuous thermal cycles were as follows: 50°C for 30 min, 95°C for 15 min, and then 30 cycles of 94°C for 1 min, 60°C for 1 min, 72°C for 1 min, and a final extension of 72°C for 10 min.

Cyclophilin message was amplified by RT-PCR to evaluate the quality and integrity of the extracted mRNA in 2 patients with a completely negative UPII RT-PCR because this gene is ubiquitously expressed in all human cells (13). Detection of cyclophilin expression was also carried out using the same as above OneStep RT-PCR reagent at the following conditions: 50°C for 30 min, 95°C for 15 min, and then 30 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 1 min, and a final extension of 72°C for 10 min. All products were resolved by horizontal gel electrophoresis (1% agarose, 0.5× Tris acetate, 0.04 m Tris acetate, and 0.5 mm EDTA) using 20% of the reaction product. Both patients had positive cyclophilin RT-PCR signals in all extracted tissue samples (data not shown), indicating that lack of UPII positivity was not attributable to RNA degradation.

The human bladder cancer cell line RT4 was used as a positive control, whereas the human prostate cancer cell line LNCaP was used as a negative control, as described previously (11). The LNCaP and RT4 cells were grown in T and DMEM F12 medium, both supplemented with 5% fetal bovine serum (Life Technologies, Inc.), respectively (14). Total cellular RNA for RT-PCR was isolated from almost confluent cell cultures using Trizol (Life Technologies, Inc.) as per the manufacturer’s recommendations for adherent cell lines.

The characteristics of the 27 tumors, including demographic information and clinical/pathological stages, are shown in Table 1. The patient cohort consisted of 9 women and 17 men with an average age of 68 ± 11 years. Eight of 27 (30%) patients had cystectomies for clinically superficial (T_a, T_is, T₁) cancers, whereas 19 of 27 (30%) patients had muscle-invasive neoplasms. Of the patients with superficial disease, 2 of 8 (25%) had no evidence of tumor in the radical cystectomy (i.e., pT₀), and no patients had node-positive disease (pN₁). Of the patients with muscle-invasive disease, 2 of 19 (10%) had no evidence of tumor in the radical cystectomy (pT₀), whereas 3 of 19 (16%) patients had node-positive disease (pN_1–3). Overall, 1 of 4 (25%) patients with pT₀ disease had lymph node-positive disease.

In every reaction, the LNCaP cell line was used as a negative control, and human bladder cancer cell line RT4 was used as a positive control (Fig. 2). In addition, water was used as the negative control because well reactions were set up with or without any reverse transcriptase and kept at 4°C until the denaturation starts to test for primer and/or procedure-generated artifact, PCR carry-over, or laboratory contamination of the UPII template. Finally, an expected sized product of 322 bp (Fig. 2) was cloned, sequenced, and confirmed a homology of 100% with UPII gene sequence.

Using this approach, we carried out RT-PCR for UPII on all perivesical and nodal tissue samples collected after the removal of the cystectomy specimen. A total of 154 specimens were examined by RT-PCR for UPII, of which 44 (29%) were positive for UPII. In terms of individual biopsy specimens, 44 of 126 (35%) of the perivesical tissue specimens were positive, whereas 7 of 19 (37%) of the lymph node tissue specimens had detectable UPII signals. The median number (range) of positive perivesical tissue specimens in these 14 patients was 3.5 (1, 2, 3, 4, 5).

The results of RT-PCR for the UPII gene product in comparison to the clinical and pathological features of the radical cystectomy specimens are shown in Tables 1 and 2. In the 8 patients who had radical cystectomy for clinically superficial disease (cT_a–T₁), 4 (50%) had at least one perivesical biopsy positive for UPII expression (Tables 1 and 2), but no positivity was seen in the nodal tissues. In 19 patients with clinically invasive disease (cT₂–T₄), 10 (53%) had at least one perivesical biopsy positive for UPII expression and 7 of 14 (50%) had positivity in the lymph node specimens. Overall, 14 of 27 (52%) and 7 of 19 (37%) patients had positive UPII biopsies in perivesical or lymph node samples, respectively.

Correlation of pathological staging with molecular analysis for UPII expression reveals a trend toward increased frequency of UPII positivity in both the perivesical and lymph node biopsies as a function of pathological stage in patients with pathologically negative lymph nodes (Table 2). In addition, a trend toward a higher number of UPII-positive perivesical biopsies per patient as a function of stage was also present (Fig. 3). All 3 patients with nodal involvement had UPII positivity in the nodal sample, whereas only 1 (33%) had any positive perivesical biopsies. Interestingly, 1 of 12 (17%) patients with pathologically organ-confined, node-negative disease had positive lymph node UPII signals, whereas 3 of 12 (25%) patients with pT₃–T₄N₀ disease had this finding.

Two large studies have examined the impact of the number and size of positive nodes on the survival of patients after radical cystectomy and found a trend toward decreased survival in patients with increasing bulk of positive nodes (3, 4). Interestingly, the differences in survival between groups with minimal involvement [i.e., 1–5 positive nodes or pN₁ (3)] to those with more significant involvement were small. This observation serves to highlight that despite radical surgery, any nodal positivity is a dire prognostic factor. However, it has also been recognized for some time that patients with metastatic disease limited to the lymphatics may have a better prognosis compared with those harboring metastases at other sites (15, 16). This finding, coupled with the advent of novel chemotherapeutic agents with less toxicity than classic regimens (17, 18), offer the possibility for significant improvements in prognosis in patients with locally advanced or with pelvic lymph node metastases treated in an adjuvant setting (19).

Because there are limitations to the sensitivity of detecting perivesical extension and lymph node positivity in patients undergoing radical cystectomy, we sought to devise a technique that could improve on this. Thus, we proposed to biopsy the surface of the primary specimen once removed from the patient and in addition, a representative cross-sectional slice from the largest and/or most suspicious lymph node and analyze all samples using molecular techniques. This sampling procedure is simple, rapid, and safe for the patient, incurring no penalty in terms of morbidity or operative time. However, before discussing this technique any further, it must be clearly emphasized that the results presented here are clearly done both as “proof of principle” and to be hypothesis generating for further studies. They are very preliminary and thus any clinical relevance or significance of these results is purely hypothetical and speculative at this point.

Although others have used RT-PCR for UPII for the detection of floating urothelial cancer cells in blood (10, 11), the direct examination of bladder surface and lymph node biopsy samples for micrometastatic bladder cancer cells after radical cystectomy has not been described previously. Because of the novelty of this study, we have attempted to methodically eliminate factors that could confound the interpretation of our results. We have attempted to address the issue of blood contamination of the samples in two ways: (a) we have taken care to rinse the biopsy samples thoroughly, prior to analysis, to eliminate any gross blood contamination; and (b) the RT-PCR technique selected was one that was designed not to detect circulating urothelial cells in the circulation of patients with clinically localized bladder cancer (11), even when much larger quantities of blood than would be contaminating our specimens were analyzed. Furthermore, in patients with UPII-positive biopsies, not all biopsy samples were positive, suggesting that this assay was not artifactually contaminated by tumor cells circulating in the bloodstream. In addition, we found that all histologically positive lymph nodes were detected by the molecular analysis (Table 2). This latter finding suggests that the molecular methodology has a sensitivity at least equivalent to conventional pathological analysis of the nodes.

Only long term follow-up of patients with pathologically node-negative but UPII-positive signals in the lymph nodes will reveal whether these patients will have a clinical recurrence. Although infrequently seen in our small series (4 of 16 patients; 25%), this percentage is similar to the rate of recurrence observed for patients with such organ-confined disease (20), suggesting that most patients with organ-confined disease and positive UPII expression in the nodes will eventually fail. However, as mentioned earlier, in the absence of long term follow-up, such conclusions are purely speculative. Continued long-term follow-up will also reveal the clinical significance of the molecular positive perivesical biopsies in the context of either clinically or molecularly positive lymph nodes. Because the degree of positivity of these latter biopsies is higher than the expected failure rate after cystectomy (20), it is unlikely that they reflect residual cancer left inside the patient after cystectomy. Instead, they may reflect perivascular permeation of the bladder wall by cancer cells, which by definition is removed at the time of radical cystectomy. However, in view of the fact that pathological stage has been shown to be an important prognostic factor even in patients with clinically positive nodes (21), this finding may conceivably also be the case for the molecular analysis. For example, patients with UPII expression in lymph node tissue and extensive expression of UPII in most perivesical biopsies (Fig. 3) may have a higher risk of local or metastatic recurrence than a patient with only lymph node UPII expression and minimal or absent perivesical positivity.

Finally, because some series have shown a trend toward improved cure rate in patients with minimal nodal disease with radical surgery alone (3), it is conceivable that patients with molecularly positive but histologically negative lymph nodes may have an intermediate prognosis between patients with complete lack of UPII expression in the lymph nodes and those with microscopically positive lymph nodes. Conversely, if these patients do fail with a high percentage, they may be in the best prognostic group with respect to potential cure from adjuvant chemotherapy in view of the minimal bulk of disease. As mentioned earlier, in the absence of long-term follow-up, such conclusions are purely speculative.

In conclusion, this RT-PCR-mediated approach appears to detect all clinically positive lymph nodes. In addition, this methodology identifies patients with positive UPII expression in microscopically negative lymph node samples. This approach also detects UPII expression in perivesical biopsies in 54% of patients with pathologically node-negative disease. Although these very preliminary results are promising because of the close similarity of the lymph node-derived molecular results with eventual clinical failure after cystectomy in large series, only long-term follow-up of these patients will determine the validity and clinical relevance of these findings.

We thank Dr. Marguerite C. Lippert, University of Virginia Health Sciences Center, for her contribution of patients to the study and to Dr. Michael Harding, Department of Urology, University of Virginia, for helpful suggestions.