SEARCH

SEARCH BY CITATION

Keywords:

  • 17β-estradiol;
  • immunohistochemistry;
  • ovariectomy;
  • SVHUC cell line;
  • urothelial neoplasm

Abstract

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

Although UDP-glucuronosyltransferase 1A (UGT1A) plays an important role in preventing bladder cancer initiation by detoxifying carcinogenic compounds, its contribution to bladder cancer progression is poorly understood. We immunohistochemically stained for UGT1A in bladder specimens. UGT1A was positive in 130/145 (90%; 28 [19%] weak, 53 [37%] moderate, and 49 [34%] strong) urothelial neoplasms, which was significantly weaker than in matched non-neoplastic urothelial tissues (100/101 [99%]; 2 [2%] weak, 17 [17%] moderate, and 81 [80%] strong). Fifty (98%) of 51 low-grade/79 (99%) of 80 non-muscle-invasive tumors were immunoreactive to UGT1A, whereas 80 (85%) of 94 high-grade/51 (78%) of 65 muscle-invasive tumors were UGT1A-positive. Kaplan-Meier analysis showed strong associations between lower UGT1A expression versus the risk of recurrence in high-grade non-muscle-invasive tumors (P = 0.038) or disease-specific mortality in muscle-invasive tumors (P = 0.016). Multivariate analysis further revealed UGT1A loss as an independent prognosticator for disease-specific mortality in patients with muscle-invasive tumor (P = 0.010). Additionally, the expression of UGT1A was positively and negatively correlated with those of estrogen receptor-α and estrogen receptor-β, respectively. We then assessed UGT1A/Ugt1a levels in human cell lines/mouse tissues. 17β-Estradiol increased and decreased UGT1A expression in normal urothelium and bladder cancer lines, respectively, and an anti-estrogen abolished these effects. Ovariectomy in mice resulted in down-regulation of Ugt1a subtypes. These results suggest the involvement of UGT1A in not only bladder carcinogenesis but tumor progression. Moreover, UGT1A is likely regulated by estrogens in non-neoplastic urothelium versus bladder tumor in opposite manners, which could be underlying mechanisms of gender-specific differences in bladder cancer incidence and progression. © 2012 Wiley Periodicals, Inc.


Abbreviations
NNAL

4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol

UGT

glucuronosyltransferase

AR

androgen receptor

DHT

dihydrotestosterone

ER

estrogen receptor

TMA

tissue microarray

PUNLMP

papillary urothelial neoplasm of low malignant potential

FBS

fetal bovine serum

E2

17β-estradiol

TAM

tamoxifen

HF

hydroxyflutamide

RT

reverse transcription

PCR

polymerase chain reaction

BBN

N-butyl-N-(4-hydroxybutyl)nitrosamine

INTRODUCTION

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

It is estimated that 73,510 individuals living in the United States will develop urinary bladder cancer in 2012 and 14,880 will die of the disease [1]. Data from comparative studies have demonstrated that bladder cancer affects men three to four times more often than women, while female patients tend to present with more aggressive tumor than male patients [1, 2].

Bladder cancer is one of the first neoplasms recognized to be caused by exposure to carcinogenic compounds, such as industrial chemicals and cigarette smoke. Aromatic amines, well-known industrial bladder carcinogens, and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), a metabolite of the most potent and abundant procarcinogens derived from tobacco and its smoke, are glucuronidated in the liver and excreted into the urinary system [3, 4]. It has been shown that UDP-glucuronosyltransferases (UGTs), belonging to the superfamily of major phase II drug metabolism enzymes, play a vital role in catalyzing the glucuronidation of carcinogens, including aromatic amines and NNAL [5].

Human UGTs are composed of UGT1A, UGT2A, and UGT2B, based on gene sequence homology [6]. In addition to the liver, the expression of UGTs has been detected in other organs, including the aerodigestive tract [7], gastrointestinal tract [8], and kidney [9]. It has also been reported that normal bladder expresses all the UGT subtypes except UGT2B17 [10] and that, compared with normal urothelium, UGT1A expression is down-regulated in several bladder cancer tissue samples [11, 12]. Recently, a genome-wide association study revealed a linkage between UGT1A gene locus and bladder cancer susceptibility [13]. Indeed, UGT1A, rather than the UGT2 family, has been suggested to contribute to metabolism of aromatic amines [5]. Down-regulation of Ugt1a, although there is no clear consensus about the functional homology between human UGT and mouse Ugt, was associated with higher incidence of chemically induced bladder cancer in mice [14]. Thus, UGT1A, consisting of nine functional proteins and four pseudogenes generated by alternative splicing, and Ugt1a, consisting of nine functional protein and five pseudogenes [15], are likely key enzymes involved in bladder carcinogenesis.

Although excessive exposure to industrial chemicals and cigarette smoke may have contributed to male dominance in bladder cancer, men remain at a substantially higher risk of bladder cancer than women even after controlling for these carcinogenic factors [16]. We have shown molecular evidence for this gender-specific difference by implicating androgen receptor (AR) signals in bladder cancer development [17]. There is also increasing evidence suggesting that other steroid hormone receptors are involved in bladder carcinogenesis and cancer progression [18]. Recently, we showed that dihydrotestosterone (DHT) reduced the expression of UGT1A subtypes via the AR pathway in normal urothelial cells [19]. In the previous report [19], we also showed that UGT1A was down-regulated in high-grade urothelial carcinoma tissues and strong expression of UGT1A correlated with favorable prognosis. However, the results were not conclusive presumably due to a relatively small number of cases (n = 24) including no low-grade tumors. The purpose of the current study is to validate the previous findings in larger patient cohorts with longer follow-up. We also assessed possible associations of UGT1A expression with AR and estrogen receptor (ER) signals in non-neoplastic urothelium and urothelial carcinoma.

MATERIALS AND METHODS

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

Tissue Samples

We retrieved 145 bladder tissue specimens obtained by transurethral resection or cystectomy performed at the University of Rochester Medical Center or the Johns Hopkins Hospital. All the sections were reviewed for confirmation of original diagnoses, according to the 2004 WHO/ISUP classification system for urothelial neoplasms [20], by two urologic pathologists (J.L.Y. and G.J.N.) at respective institutions. Appropriate approval from the Institutional Review Board at each institution was obtained prior to construction and use of the tissue microarray (TMA). Bladder TMAs were constructed from formalin-fixed paraffin-embedded specimens (145 tumor tissues and 101 benign appearing tissues from bladders of patients with tumors), as previously described [21]. These patients included 110 men and 35 women, with a mean age of 66.0 yr (range: 30–89 yr) at the time of surgery and a mean follow-up of 31.6 months (range 2–164 months) after the surgery. The tumors included 11 papillary urothelial neoplasms of low malignant potential (PUNLMPs), 40 non-invasive (pTa) low-grade urothelial carcinomas, 29 non-muscle-invasive (≤pT1) high-grade urothelial carcinomas, and 65 muscle-invasive (≥pT2) high-grade urothelial carcinomas. All 65 patients with muscle-invasive tumor underwent cystectomy. None of the patients had received therapy with radiation or anticancer drugs pre-operatively, except for 17 cases with intravesical bacillus Calmette-Guérin treatment prior to radical cystectomy. All of these 145 cases were included in our prior study analyzing 188 cases for the expression of AR, ERα, and ERβ [21].

Immunohistochemistry

Immunohistochemical staining was performed at the University of Rochester Medical Center, using the primary antibody to UGT1A (H300 clone; diluted 1:100; Santa Cruz Biotechnology, Santa Cruz, CA), as described previously [19, 22]. An optimal condition for the stains was determined in control tissues. Each TMA contained orientation cores (tissues from multiple organs) that also served as internal positive and negative controls. All the stains were manually scored by one pathologist (H.M.) blinded to patient identity. German Immunoreactive Score (0–12) was calculated by multiplying the percentage of immunoreactive cells (0% = 0; 1–10% = 1; 11–50% = 2; 51–80% = 3; 81–100% = 4) by staining intensity (negative = 0; weak = 1; moderate = 2; strong = 3). Cores with the immunoreactive score of 0–1, 2–4, 6–8, and 9–12 were considered negative, weakly positive (1+), moderately positive (2+), and strongly positive (3+), respectively.

Cell Culture and Chemicals

Human urothelium cell line (SVHUC) and bladder cancer cell lines (UMUC3, 5637, and J82; all obtained from the American Type Culture Collection, Manassas, VA) were maintained in appropriate media (Mediatech, Manassas, VA; Kaighn's Modification of Ham's F-12 for SVHUC and Dulbecco's modified Eagle's medium for other cell lines) supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere of 5% CO2. Cells were cultured in phenol-red free medium supplemented with 5% charcoal-stripped FBS at least 18 h before experimental treatment. We obtained DHT, 17β-estradiol (E2), and tamoxifen (TAM) from Sigma (St. Louis, MO), and hydroxyflutamide (HF) from Schering (Kenilworth, NJ).

Western Blot

Protein extraction and Western blot were performed, as described previously [19, 22] with minor modifications. Briefly, equal amounts of protein (20 µg) obtained from cell extracts were separated in 10% sodium dodecylsulfate-polyacrylamide gels and transferred to polyvinylidene difluoride membrane (Millipore, Billerica, MA) by electroblotting using a standard protocol. Specific binding of primary antibodies to ERα (E115 clone; diluted 1:1,000; Epitomics, Burlingame, CA), ERβ (14C8 clone; diluted 1:1,000; Abcam, Cambridge, MA), UGT1A (H300 clone; diluted 1:1,000), and GAPDH (6C5 clone; diluted 1:1,000; Santa Cruz Biotechnology) were detected, using horseradish peroxidase detection system (SuperSignal West Pico Chemiluminescent Substrate; Thermo Scientific, Rockford, IL).

Ovariectomized Mice

Female CL57BL/6-129SV mice were housed according to the institutional guidelines and were allowed food and water ad libitum. Mice received bilateral ovariectomy (n = 9) or sham surgery (n = 6) at 5 weeks of age. Among the mice undergoing oophorectomy, 0.1 ml of peanut oil with (n = 3) or without (n = 6) 20 µg of E2 was injected subcutaneously every 2 d. One week after the surgery, all the mice were sacrificed and urinary bladders were harvested. These specimens were rapidly frozen in liquid nitrogen and stored at −80°C for subsequent RNA analysis.

Reverse Transcription (RT) and Real-Time Polymerase Chain Reaction (PCR)

Total RNA (1.0 µg) isolated from harvested cell lines or mouse tissues, using TRIzol (Invitrogen, Carlsbad, CA), was reverse transcribed using 1 µmol/L oligo (dT) primers and four units of Omniscript reverse transcriptase (Qiagen, Valencia, CA) in a total volume of 20 µl. Real-time PCR was then performed in 15 µl system by using SYBR GreenER qPCR SuperMix for iCycler (Invitrogen), as described previously [19, 22]. Primers used to amplify UGT1A mRNA were: 5′-TGATTGGTTTCCTCTTGGC-3′ and 3′-GGGTCTTGGATTTGTGGG-5′. The primer sequences for Ugt1a subtypes are given elsewhere [19]. Due to the high degree of homology, we were unable to design primers that separately amplify Ugt1a6a and Ugt1a6b.

Statistical Analyses

The Fisher's exact test was used to evaluate the association between categorized variables. Non-parametric two-group comparisons were carried out using Mann–Whitney U-test to assess differences in variables with ordered distribution across dichotomous categories. Survival rates in patients were calculated by the Kaplan-Meier method, and comparisons were made by log-rank test. These included comparisons among patients with non-invasive low-grade tumor (PUNLMP + low-grade carcinoma) with a mean follow-up of 37.1 months (range: 5–164), patients with non-muscle-invasive high-grade tumor with a mean follow-up of 26.8 months (range: 3–117), or patients with muscle-invasive tumor with a mean follow-up of 29.4 months (range: 2–141). Tumor recurrence was separately evaluated in 51 patients with non-invasive low-grade tumor and 29 patients with non-muscle-invasive high-grade tumor. Tumor progression was separately evaluated in 51 patients with non-invasive low-grade tumor (development of high-grade or invasive tumor), 29 patients with non-muscle-invasive high-grade tumor (development of muscle-invasive or metastatic tumor), and 65 patients with muscle-invasive tumor (development of local recurrence or metastatic tumor). Disease-specific survival was evaluated in 65 patients with muscle-invasive tumor. Correlation analyses were also performed, using the Spearman's correlation test. The Cox proportional hazards model was used to determine statistical significance of predictors in a multivariate setting. P values less than 0.05 were considered to be statistically significant.

RESULTS

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

Immunoreactivity in Benign and Carcinoma Tissues

We immunohistochemically investigated the expression of UGT1A in 145 bladder tumor specimens as well as corresponding 101 benign bladder tissues. Positive signals were detected predominantly in cytoplasms of non-neoplastic and neoplastic epithelial cells (Figure 1). Correlations of the expression status with different benign and tumor tissues are summarized in Table 1.

image

Figure 1. UGT1A immunohistochemisty in non-neoplastic urothelium (A) and bladder tumors (B and C). Strong stains for UGT1A were observed in non-neoplastic (A) and neoplastic (B) urothelial cells, whereas an invasive urothelial carcinoma (C) showed no immunoreactivity to UGT1A.

Download figure to PowerPoint

Table 1. Expression of UGT1A in Bladder Tissue Microarrays
 NPositivityFisher test (P value)Score (mean ± SD)Mann–Whitney U-test (P value)
0 (%)1+ (%)2+ (%)3+ (%)0 vs. 1+/2+/3+0/1+ vs. 2+/3+0/1+/2+ vs. 3+
  1. a

    All cases.

  2. b

    Males only.

  3. c

    Females only.

  4. d

    Student's t-test.

  5. e

    PUNLMP + Low-grade carcinoma vs. High-grade carcinoma.

  6. f

    Non-muscle-invasive (pTa + pT1) vs. Muscle-invasive (≥pT2).

Tissue type     0.003a/0.016b/0.250c<0.001a/<0.001b/0.004c<0.001a/<0.001b/<0.001c <0.001a/<0.001b/<0.001c
Benign urothelium1011 (1)2 (2)17 (17)81 (80)   2.81 ± 0.46 
(Age: mean ± SD, years) (64.00)(61.50 ± 0.71)(70.47 ± 10.81)(65.36 ± 12.64)Not availablea,d0.589a,d0.206a,d  
Urothelial neoplasm14515 (10)28 (19)53 (37)49 (34)   2.01 ± 0.91 
(Age: mean ± SD, years) (63.73 ± 10.38)(66.21 ± 10.61)(68.15 ± 12.02)(64.27 ± 14.39)0.459a,d0.682a,d0.232a,d  
Tumor grade     0.020a,e/0.054b,e/0.279c,e<0.001a,e/0.003b,e/0.002c,e0.099a,e/0.680b,e/0.006c,e <0.001a,e/0.013b,e/<0.001c,e
PUNLMP111 (9)1 (9)7 (64)2 (18)   2.09 ± 0.83 
Low-grade carcinoma400 (0)2 (5)18 (45)20 (50)   2.54 ± 0.47 
PUNLMP + Low grade511 (2)3 (6)25 (49)22 (43)   2.44 ± 0.59 
High-grade carcinoma9414 (15)25 (27)28 (30)27 (29)   1.77 ± 0.97 
Tumor stage     <0.001a,f/0.001b,f/0.056c,f<0.001a,f/<0.001b,f/<0.001c,f0.112a,f/0.431b,f/0.027c,f <0.001a,f/<0.001b,f/<0.001c,f
pTa751 (1)7 (9)37 (49)30 (40)   2.40 ± 0.59 
pT150 (0)0 (0)3 (60)2 (40)   2.43 ± 0.44 
Non-muscle-invasive801 (1)7 (9)40 (50)32 (40)   2.40 ± 0.58 
pT2235 (22)9 (39)3 (13)6 (26)   1.50 ± 0.97 
pT3317 (23)10 (32)7 (23)7 (23)   1.46 ± 1.04 
pT4112 (18)2 (18)3 (27)4 (36)   1.79 ± 1.11 
Muscle-invasive6514 (22)21 (32)13 (20)17 (26)   1.53 ± 1.01 
Lymph node involvement     0.749a/0.708b/1.000c0.605a/0.224b/1.000c0.228a/0.102b/1.000c 0.607a/0.377b/0.829c
pN0479 (19)15 (32)14 (30)9 (19)   1.52 ± 0.96 
pN+215 (24)4 (19)5 (24)7 (33)   1.66 ± 1.06 

UGT1A was positive in 100 of 101 (99%; 2 [2%] 1+, 17 [17%] 2+, and 81 [80%] 3+) benign urothelial tissues and 130 of 145 (90%; 28 [19%] 1+, 53 [37%] 2+, and 49 [34%] 3+) urothelial neoplasms. Overall, UGT1A expression was significantly weaker in tumors than in benign tissues (score, P < 0.001).

Immunoreactivity and Clinicopathologic Features

We evaluated the correlation of expression levels of UGT1A stains with clinicopathologic features available for our patient cohort (Table 1). There were no statistically significant differences in UGT1A expression pattern in benign or neoplastic bladders between ages of the patients. Fifty (98%) of 51 low-grade tumors were immunoreactive to UGT1A (3 [6%] 1+, 25 [49%] 2+, and 22 [43%] 3+), and 80 (85%) of 94 high-grade tumors were UGT1A-positive (25 [27%] 1+, 28 [30%] 2+, and 27 [29%] 3+). Thus, UGT1A expression was significantly lower in high-grade carcinomas than in PUNLMPs + low-grade carcinomas (score, P < 0.001). Similarly, UGT1A expression was significantly lower in muscle-invasive tumors (51/65 [78%] positive; 21 [32%] 1+, 13 [20%] 2+, and 17 [26%] 3+) than in non-muscle-invasive tumors (79/80 [99%] positive; 7 [9%] 1+, 40 [50%] 2+, and 32 [40%] 3+; score, P < 0.001). However, among 68 cases with regional lymph node dissection, there was no significant difference in UGT1A levels between node-negative tumors (38/47 [81%] positive; 15 [32%] 1+, 14 [30%] 2+, and 9 [19%] 3+) and node-positive tumors (16/21 [76%] positive; 4 [19%] 1+, 5 [24%] 2+, and 7 [33%] 3+; score, P = 0.607).

We next performed Kaplan-Meier analysis coupled with log-rank test to assess possible associations of UGT1A staining with tumor recurrence or progression (Figure 2). For these analyses, we dichotomized UGT1A expression as 0/1+ vs. 2+/3+ in non-muscle-invasive tumors and 0 vs. 1+/2+/3+ in muscle-invasive tumors. Of the 51 patients with PUNLMP or low-grade carcinoma, 17 (33%) and 3 (6%) had recurrence and progression, respectively. There were no statistically significant correlations between UGT1A expression and recurrence (P = 0.236) or progression (P = 0.683). Of the 29 patients with non-muscle-invasive high-grade carcinoma, 14 (48%) and 7 (24%) had recurrence and progression, respectively. Low UGT1A levels were strongly associated with recurrence (P = 0.038), but not with progression (P = 0.281). Finally, of the 65 patients with muscle-invasive tumor, 35 (54%) had disease progression and 22 (34%) died of bladder cancer. Loss of UGT1A expression strongly correlated with disease-specific mortality (P = 0.016), but not with progression (P = 0.168).

image

Figure 2. Kaplan-Meier analysis according to the levels of UGT1A expression. Recurrence/progression-free survivals in PUNLMPs and low-grade carcinomas (A) or non-muscle-invasive high-grade carcinomas (B) and progression-free/disease-specific survivals in muscle-invasive tumors (C). Comparisons were made by log-rank test.

Download figure to PowerPoint

To see whether UGT1A expression was an independent predictor of survival in patients with high-grade muscle-invasive tumor, multivariate analysis was performed with Cox model, including dichotomized pT stage (pT2 vs. pT3 + pT4), lymph node metastasis (pN0 vs. pN+), and UGT1A expression (0 vs. 1+/2+/3+; Table 2). In this subgroup, UGT1A expression was found to correlate with better cancer-specific survival (HR = 0.293; 95%CI = 0.116–0.745; P = 0.010), but not with tumor progression (HR = 0.574; 95%CI = 0.267–1.233; P = 0.155).

Table 2. Multivariate Cox Model in High-Grade Muscle-Invasive Tumors
 HR95%CIP value
  1. a

    pT2 vs. pT3/pT4.

  2. b

    pN0 vs. pN+.

  3. c

    0 vs. 1+/2+/3+.

Tumor progression
pTa3.3381.337–8.3340.010
pNb1.4900.718–3.0900.284
UGT1Ac0.5740.267–1.2330.155
Cancer-specific mortality
pTa4.1601.182–14.6390.026
pNb1.0450.398–2.7400.929
UGT1Ac0.2930.116–0.7450.010

Association of UGT1A Expression With Gender or Expression of Sex Hormone Receptors

We investigated gender differences in UGT1A expression both in benign and tumor tissues (Table 3). UGT1A was positive in 73 of 74 (99%; 1 [1%] 1+, 14 [19%] 2+, and 58 [78%] 3+) male vs. 27 of 27 (100%; 1 [4%] 1+, 3 [11%] 2+, and 23 [85%] 3+) female benign tissues and 98 of 110 (89%; 19 [17%] 1+, 37 [34%] 2+, and 42 [38%] 3+) male vs. 32 of 35 (91%; 9 [26%] 1+, 16 [46%] 2+, and 7 [20%] 3+) female tumors. Although there were no statistically significant differences in UGT1A expression between males and females, its strong positivity (3+) was more often seen in male tumors than in female tumors (P = 0.064). Interestingly, these differences were more significant when separately analyzed in high-grade (P = 0.006), muscle-invasive (P = 0.014), or pN+ (P = 0.046; 0/1+ vs. 2+/3+) tumors, but not in low-grade (P = 1.000), non-muscle-invasive (P = 0.606), or pN0 (P = 0.318) tumors. However, no significant differences in UGT1A levels in either non-neoplastic or neoplastic bladders were seen between the two age groups of males only, females only, or all patients (e.g., ≤50 vs. ≥51 yr, ≤55 vs. ≥56 yr). In addition, in female patients, UGT1A expression still showed prognostic significance (i.e., recurrence of non-muscle-invasive high-grade tumor [0/1+ vs. 2+/3+, P = 0.014], progression [P < 0.001] or survival [P = 0.055] of muscle-invasive tumor [0 vs. 1+/2+/3+]; figures not shown). By contrast, in males, there were no statistically significant associations between UGT1A levels and patients' outcomes.

Table 3. Gender Difference in UGT1A Expression
 NPositivityFisher test (P value)Score (mean ± SD)Mann–Whitney U-test (P value)
0 (%)1+ (%)2+ (%)3+ (%)0 vs. 1+/2+/3+0/1+ vs. 2+/3+0/1+/2+ vs. 3+
Benign     1.0001.0000.578 0.288
Male741 (1)1 (1)14 (19)58 (78)   2.79 ± 0.48 
Female270 (0)1 (4)3 (11)23 (85)   2.85 ± 0.43 
Neoplasm     0.7650.5270.064 0.267
Male11012 (11)19 (17)37 (34)42 (38)   2.04 ± 0.92 
Female353 (9)9 (26)16 (46)7 (20)   1.91 ± 0.87 
PUNLMP + Low-grade carcinoma     1.0000.3421.000 0.461
Male381 (3)3 (8)18 (47)16 (42)   2.39 ± 0.64 
Female130 (0)0 (0)7 (54)6 (46)   2.59 ± 0.39 
High-grade carcinoma     1.0000.2160.006 0.078
Male7211 (15)16 (22)19 (26)26 (36)   1.86 ± 1.00 
Female223 (14)9 (41)9 (41)1 (5)   1.50 ± 0.83 
Non-muscle-invasive tumor     1.0001.0000.606 0.947
Male591 (2)5 (8)28 (47)25 (42)   2.40 ± 0.60 
Female210 (0)2 (10)12 (57)7 (33)   2.42 ± 0.53 
Muscle-invasive tumor     1.0000.2260.014 0.133
Male5111 (22)14 (27)9 (18)17 (33)   1.64 ± 1.07 
Female143 (21)7 (50)4 (29)0 (0)   1.14 ± 0.71 
pN0 tumor     0.6050.4160.318 0.132
Male407 (18)12 (30)12 (30)9 (23)   1.62 ± 0.99 
Female72 (29)3 (43)2 (29)0 (0)   1.00 ± 0.64 
pN+ tumor     1.0000.0460.120 0.092
Male154 (27)0 (0)4 (27)7 (47)   1.87 ± 1.12 
Female61 (17)4 (67)1 (17)0 (0)   1.11 ± 0.71 

We then analyzed the correlations between expressions of UGT1A and AR/ERα/ERβ (Table 4). In our cohort of 101 benign and 145 malignant bladders where the expression of all these four proteins were examined, AR/ERα/ERβ was positive in 79 (78%)/54 (54%)/49 (49%) and 64 (44%)/38 (26%)/74 (51%), respectively. In benign tissues, there were no significant correlations between UGT1A and each hormone receptor. In neoplastic tissues, UGT1A showed a weak positive correlation (0.2 < CC < 0.4) with ERα and a weak negative correlation (−0.4 < CC < −0.2) with ERβ. When analyzed separately in men and women, these correlations were more significant in women (i.e., moderate positive correlation [0.4 < CC < 0.7] with ERα and moderate negative correlation [−0.7 < CC < −0.4] with ERβ).

Table 4. Correlation of UGT1A Expression With AR, ERα, or ERβ
 NARERαERβ
CCP valueCCP valueCCP value
  1. CC, correlation coefficient.

Benign
All cases101−0.0880.3790.0780.4380.0760.452
Male740.1540.4420.0870.6670.1050.601
Female27−0.1520.1950.0860.4680.0670.568
Tumor
All cases1450.1060.2040.2240.007−0.292<0.001
Male1100.0920.3390.1650.085−0.2190.021
Female350.1780.3070.4120.014−0.578<0.001

Regulation of UGT1A/Ugt1a Expression by Sex Hormones

We assessed the effects of estrogen and androgen on UGT1A expression in human bladder cell lines. UMUC3 urothelial carcinoma line was shown to be AR-positive, whereas AR was undetectable in other bladder cancer lines examined (5637, J82) and in SVHUC normal urothelial line [17, 19, 22]. We also examined the expression of ERα and ERβ in these cell lines. As shown in Figure 3A, all these lines expressed the ERβ, but not ERα. We then compared the levels of UGT1A protein expression in these cells treated with estrogen (E2) ± ER antagonist (TAM) or androgen (DHT) ± AR antagonist (HF). E2 clearly induced UGT1A expression in SVHUC, and TAM showing a partial agonist activity at least partially restored the E2 effect (Figure 3B). In contrast, E2 reduced UGT1A expression in three cancer cell lines, and TAM showing marginal agonist activities antagonized the E2 effect. We have shown that DHT treatment in SVHUC stably expressing AR resulted in decreases in UGT1A expression and HF antagonized the DHT effect [19]. However, DHT showed only marginal effects on UGT1A in UMUC3 (Figure 3C). We also tested mRNA expression of UGT1A in these cell lines treated with E2 or DHT by a quantitative real-time RT-PCR method. As expected, similar changes in UGT1A levels were observed (Figure 3D). These findings suggest that ERβ signals induce and repress UGT1A expression in normal urothelium and bladder cancer cells, respectively.

image

Figure 3. Regulation of UGT1A expression by sex steroids in benign and malignant urothelial cell liens. A: Protein extracts from SVHUC, UMUC3, 5637 and J82 cells were immunoblotted for ERα (66 kDa) and ERβ (56 kDa). MCF7 breast cancer line served as the positive control. B: Protein extracts from respective cell lines cultured in the presence of ethanol (mock), 1 nM E2, and/or 1 µM TAM for 24 h were immunoblotted for UGT1A (56 kDa). C: Protein extracts from UMUC3 cultured in the presence of ethanol (mock), 1 nM DHT, and/or 1 µM HF for 24 h were immunoblotted for UGT1A. In these western blots, GAPDH expression (37 kDa) served as the internal control. In (B) and (C), densitometry values for specific bands standardized by GAPDH that are relative to those of mock treatment (first lane in each cell line; set as onefold) are included below the lanes. Each value represents the mean from at least two independent experiments. D: Cell lines (SVHUC, UMUC3, 5637, and J82) cultured in the presence of ethanol (mock), 1 nM E2/DHT, and/or 1 µM TAM/HF for 24 h, as indicated, were analyzed on real-time RT-PCR for UGT1A. Expression of UGT1A was normalized to that of GAPDH. Transcription amount is presented relative to that of mock treatment in each line (first lane; set as 100%). Each value represents the mean + SD of triplicates. *P < 0.05 (vs. mock treatment in the same cell line). #P < 0.01 (vs. mock treatment in the same cell line).

Download figure to PowerPoint

We found that female mice have higher levels of Ugt1a in their bladders, compared with male mice [19]. To further confirm the stimulatory effect of E2 on UGT1A expression in SVHUC, we evaluated Ugt1a expression in the bladders from wild-type female mice undergoing bilateral ovariectomy or sham surgery followed by E2 or mock treatment. As shown in Figure 4, ovariectomy down-regulated all the subtypes of Ugt1a (18–50% decrease), and E2 supplement at least partially restored the effects of ovariectomy.

image

Figure 4. Effects of ovariectomy on Ugt1a expression in mouse bladder tissues. Wild-type female mice underwent sham surgery (n = 6), ovariectomy only (n = 6), or ovariectomy followed by E2 supplement (n = 3) at 5 weeks of age. One week after surgery, urinary bladders were harvested and analyzed on real-time RT-PCR for all the subtypes of Ugt1a. Expression of each specific gene was normalized to that of GAPDH. Transcription amount is presented relative to that in wild-type females with sham surgery (first lanes; set as onefold). Each value represents the mean + SEM. *P < 0.05 (vs. control). #P < 0.01 (vs. control).

Download figure to PowerPoint

DISCUSSION

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

UGTs are enzymes that contribute to detoxifying bladder carcinogens including aromatic amines and NNAL [5]. A recent genome-wide association study identified UGT1A as one of the susceptibility loci of bladder cancer [13]. It has also been reported, using mouse bladder cancer models, that a chemical carcinogen N-butyl-N-(4-hydroxybutyl)nitrosamine (BBN) reduces Ugt1a expression and that knockout of Nfr2 (nuclear factor-like 2) results in decreased Ugt1a and increased incidence of BBN-induced bladder cancer [14]. Together with other findings demonstrating down-regulation of UGT1A in bladder cancer [11, 12, 19], UGT1A likely plays an important role in prevention of bladder carcinogenesis.

Down-regulation of UGT1A mRNA was first reported in 9 of 10 bladder tumor samples (complete loss in four tumors), compared with matched benign bladder tissues [11]. Then, the same group immunohistochemically stained for UGT1A in 19 bladder tumors and found 6 of the tumors, mostly high-grade and/or invasive, were virtually negative for UGT1A whereas benign tissues consistently expressed it [12]. In our previous study involving 24 high-grade urothelial carcinomas [19], we showed down-regulation of UGT1A in 13 (54%) tumors as well as an association of strong UGT1A staining with a lower progression rate. In the current study, we analyzed 145 bladder tumors that included PUNLMPs and low-grade carcinomas and found inverse correlations between UGT1A levels versus tumor grade or pT stage. We also showed that decreased UGT1A expression was strongly associated with progression in high-grade non-muscle-invasive tumors and disease-specific mortality in muscle-invasive tumors. In addition, patients with UGT1A-negative muscle-invasive tumor tended to have a risk of disease progression. Multivariate analysis further revealed that loss of UGT1A in muscle-invasive tumors was an independent prognosticator of disease-specific mortality. In contrast, no significant correlations were seen between UGT1A expression status and recurrence/progression of low-grade tumors or progression of high-grade non-muscle-invasive tumors, probably because only few of low-grade and/or non-muscle-invasive tumors were negative or weakly positive for UGT1A. Of note was that, as we recently showed in the identical 68 cases undergoing lymph node dissection [23], pN status was not a prognostic factor (P = 0.284–0.929) in our cohort. Our results thus suggest that UGT1A has protective effects on not only bladder cancer development but also tumor progression.

Epidemiological and clinical evidence indicates that men have a substantially higher risk of bladder cancer, whereas women with bladder cancer have less favorable prognosis [1, 2]. Recent experimental data have suggested the involvement of AR and ER signaling pathways in bladder tumorigenesis and cancer progression and, therefore, urothelial carcinoma, like prostate and breast cancers, is considered as an endocrine-related neoplasm [18]. Nevertheless, no significant gender difference in the expression of AR, ERα, or ERβ in bladder tumors has been found [21, 24-28]. In the current study, although there were no statistically significant differences in UGT1A expression between male versus female tissues (both benign and malignant bladders), strong positivity of UGT1A was more often detected in male tumors than in female tumors (P = 0.064). It was likely that UGT1A was considerably down-regulated in potentially aggressive tumors from females compared with male tumors, but UGT1A levels in low-grade and/or superficial tumors were similar between genders of the patients. In addition, UGT1A tended to co-express with ERα and to express inversely with ERβ in bladder tumors, especially female tumors, whereas no significant correlation between expressions of UGT1A versus AR was found. In previous studies [21, 25-27], expression levels of AR or ERα alone in bladder cancer exhibited no prognostic significance. However, some of these studies showed an association between higher ERβ expression and poorer patients' outcomes [21, 27]. Furthermore, in a study using specific small interfering RNAs and selective agonists for ERα and ERβ, estrogens were shown to promote bladder cancer cell proliferation through the ER pathways [29]. Based on these findings, we hypothesized that in bladder cancer cells estrogens down-regulated UGT1A which may have a protective role in tumor progression. As expected, E2 repressed UGT1A expression at both mRNA and protein levels in all three bladder cancer lines expressing ERβ, but not ERα, and TAM antagonized the E2 effect. Indeed, up to 76% of bladder tumors were reported to be positive for ERβ, while positive rates of ERα in immunohistochemical studies ranged from 1% to 27% [18, 21, 24, 25, 27, 30, 31]. In AR-positive UMUC3 cells, DHT did not alter the UGT1A level. Thus, estrogen-induced ERβ signals may promote bladder cancer progression via down-regulating UGT1A. Further analyses of UGTs in vitro and in vivo are necessary to determine its biological functions in bladder cancer growth, which may provide new insights into not only prognosis and progression but also novel therapeutic approaches.

We previously proposed that down-regulation of UGT1A by androgens in urothelial cells was a potential mechanism for male dominance in bladder cancer development [19]. We also showed significantly lower levels of Ugt1a in male mouse bladders, which could be augmented by bilateral orchiectomy, than in those from females [19]. In addition, the levels of Ugt1a in castrated males or AR knockout males were still lower, compared with intact female mice. As shown in the liver [32], our previous results suggested the role of estrogen/ER signals in regulation of Ugt1a in the bladder. In the present study, we further demonstrated that bilateral ovariectomy led to decreases in Ugt1a expression in mouse bladders and estrogen supplement in ovariectomized mice restored the levels of Ugt1a. Consistent with these findings, E2 up-regulated the expression of UGT1A mRNA and protein in SVHUC normal urothelial cells expressing ERβ, but not ERα, which was at least partially abolished by an ER antagonist. Thus, it is likely that both androgens/AR and estrogens/ER signals are able to modulate UGT1A expression in the bladder, which may in turn affect the susceptibility to bladder carcinogens.

It is worth pointing out that bladder cancer is primarily a disease of advanced age. Thus, most of female bladder cancers are diagnosed after menopause in that serum levels of estrogens may not be different from those in male patients. Additionally, as aforementioned, studied have failed to show differential expression of ERs in bladder tumors between genders or ages of the patients [21, 24, 25, 27]. These may not readily support important roles of estrogens and ER signals in bladder carcinogenesis and cancer progression. Although menopausal status of female patients was uncertain in our study, there was no significant difference in UGT1A levels in the younger cohorts compared with their older counterparts, possibly due to a relatively small number of women aged ≤50 (n = 4) or ≤55 (n = 8) yr. However, it is still possible that high levels of UGT1A maintained by estrogens in women until menopause have reduced the exposure to bladder carcinogens excreted in the urine, which subsequently prevents or delays the development of bladder cancer even after menopause. We also showed that UGT1A levels, predominantly in female tumors, were associated with expression status of ERs and the prognosis. While estrogen levels in our cohort of patient were undetermined and might have varied, these results along with our in vitro data suggest the involvement of ER-mediated UGT1A signals in the growth of bladder cancer cells.

In conclusion, we showed down-regulation of UGT1A expression in bladder cancer and its inverse correlations with tumor grades and stages. Loss of UGT1A was an independent prognosticator for cancer-specific mortality in patients with muscle-invasive tumors. In addition, estrogen was found to up-/down-regulate UGT1A expression in normal urothelium/bladder cancer, respectively. These results not only suggest protective roles of UGT1A in both the development and progression of bladder cancer but may also provide potential underlying mechanisms responsible for gender-specific differences in the incidence and outcomes of bladder cancer. Further functional analyses of UGTs in bladder cancer are necessary to determine their biological significance.

REFERENCES

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