|Home | About | Journals | Submit | Contact Us | Français|
To assess the expression status of steroid hormone receptors in upper urinary tract urothelial carcinoma (UUTUC), we immunohistochemically stained for androgen receptor (AR), estrogen receptor-α (ERα), ERβ, glucocorticoid receptor (GR), and progesterone receptor (PR) in 99 UUTUC specimens and paired non-neoplastic urothelial tissues. AR/ERα/ERβ/GR/PR was positive in 20%/18%/62%/63%/16% of tumors, which was significantly lower (except PR) than in benign urothelial tissues [57% (P < 0.001)/40% (P = 0.001)/85% (P = 0.001)/84% (P = 0.002)/13% (P = 0.489)]. There were no significant associations between each receptor expression pattern and histopathological characteristic of the tumors including tumor grade/stage. Kaplan-Meier and log-rank tests revealed no significant prognostic value of each receptor expression in these 99 patients. However, patients with UUTUC positive for either ERα or PR had a significantly higher risk of disease-specific mortality (P = 0.025), compared with those with UUTUC negative for both. PR positivity alone in pT3 or pT4 tumors was also strongly associated with the risk of disease-specific mortality (P = 0.040). Multivariate analysis further identified the expression of ERα and/or PR as a strong predictor for disease-specific mortality in the entire cohort of the patients (hazard ratio, 2.434; P = 0.037). Thus, in accordance with previous observations in bladder specimens, significant decreases in the expression of AR/ERα/ERβ/GR in UUTUC, compared with that in non-neoplastic urothelium, were observed. Meanwhile, the negativity of both ERα and PR in UUTUC as well as the negativity of PR alone in deeply invasive tumor was suggested to serve as a prognosticator.
Upper urinary tract urothelial carcinoma (UUTUC) is a relatively rare but often aggressive urologic malignancy.1 Due to its lower incidence compared with that of bladder cancer (BC), the pathogenesis of UUTUC remains far from fully understood, while there are some similarities between molecular/genetic features of UUTUC and BC.2,3 In addition, there are currently no reliable prognostic biomarkers of UUTUC in clinical use.
Men have a substantially higher risk of BC than women.4 Gender-specific differences in the incidence and progression of UUTUC have also been reported.1 Meanwhile, a growing body of preclinical evidence has indicated the involvement of sex hormones and their receptor signals in urothelial cancer outgrowth,5 which may clearly explain some of the gender disparity. Specifically, androgen receptor (AR) activation correlates with induction of urothelial tumorigenesis and tumor progression. On the other hand, both stimulatory and inhibitory effects of estrogens/estrogen receptors (ERs) on these, which appear to be cell-specific and dependent on the functional activity of ERα and ERβ, have been documented. In contrast, little is known about the functions of another class of sex hormones, progestogens, including progesterone that generally counteracts estrogen action, and progesterone receptor (PR) signals in urothelial cancer.
In addition to these sex hormone receptors, a steroid hormone receptor, glucocorticoid receptor (GR), has been studied in BC and found to function as a tumor suppressor.6-8 We have demonstrated that glucocorticoids directly mediate GR activity in BC cells and strongly inhibit BC cell invasion and metastasis. It is noteworthy, however, that some natural or synthetic glucocorticoids, such as hydrocortisone and dexamethasone, contradictorily promote BC cell proliferation via inhibiting apoptosis particularly that induced by a chemotherapeutic drug cisplatin.6,7
Previous immunohistochemical studies have shown the levels of AR, ERα, ERβ, GR, and PR expression in BC and adjacent or separate non-neoplastic urothelial tissue specimens.9 However, there were only a few studies demonstrating the expression of steroid hormone receptors, such as AR10-12 and ERβ,11 in UUTUCs. In the current study, we aimed to determine the expression status of these 5 steroid hormone receptors in UUTUCs and corresponding non-neoplastic urothelial tissues as well as to assess their correlations with clinicopathological features of the tumors.
We immunohistochemically stained for AR, ERα, ERβ, GR, and PR in a set of TMA consisting of 99 UUTUC samples and corresponding 80 normal-appearing urothelial tissues. Clinicopathological characteristics of the patient cohort were summarized previously.13 Positive signals for all 5 receptors were detected predominantly in nuclei of benign and malignant epithelial cells (Fig. 1).
Table 1 summarizes the status of receptor expression in non-neoplastic urothelium versus urothelial tumor tissues. AR, ERα, ERβ, GR, and PR were positive in 57%, 40%, 85%, 84%, and 13% of benign urothelial tissues and 20%, 18%, 62%, 63%, and 16% of UUTUCs, respectively. Thus, the positive rate of AR (P < 0.001), ERα (P = 0.001), ERβ (P = 0.001), or GR (P = 0.002) was significantly lower in tumor than in non-neoplastic urothelium. There was no significant difference in PR expression pattern between benign and carcinoma.
Next we analyzed the correlations of AR, ERα, ERβ, GR, or PR expression in UUTUCs with the clinicopathological profile available for our patient cohort (Table 2). The positive rates of AR and PR tended to be higher (P = 0.072) and lower (P = 0.052) in male tumors (27% and 10%) than in female tumors (10% and 26%), respectively, while no significant difference in the positivity of ERα (17% vs. 21%; P = 0.790), ERβ (58% vs. 69%; P = 0.296), or GR (67% vs. 56%; P = 0.395) between genders was seen. The positive rates of AR (11% vs. 28%; P = 0.070) and ERβ (51% vs. 72%; P = 0.056) were also marginally reduced in renal pelvic tumors, compared with ureteral tumors, whereas differences in the positivity of ERα (22% vs. 12%; P = 0 .272), GR (58% vs. 68%; P = 0.395), or PR (9% vs. 20%; P = 0.155) between the 2 groups were not statistically significant. In addition, there were no strong correlations of the status of each receptor expression with the laterality of the tumor, tumor grade, pT stage, or lymph node involvement.
We then performed Kaplan-Meier analysis coupled with the log-rank test to assess possible associations between each receptor staining and patient outcomes. There were no statistically significant differences in tumor progression (Fig. 2) or cancer-specific mortality (Fig. 3) between AR, ERα, ERβ, GR, or PR positivity and negativity. Nonetheless, patients with ERα-positive or PR-positive tumor had a significantly higher risk of cancer-specific mortality (P = 0.025), but not that of tumor progression (P = 0.102), compared with those with ERα-negative and PR-negative tumor. In addition, in a total of 54 patients with pT3 or pT4 tumor, PR positivity correlated with a significantly higher risk of cancer-specific mortality (P = 0.040) (Fig. 4). Patients with pT3 or pT4 tumor negative for both ERα and PR had also a significantly lower risk of cancer-specific mortality (P = 0.023). There were still no strong associations between each of AR, ERα, ERβ, or GR expression and cancer-specific survival in this subgroup.
To determine whether ERα/PR status was an independent prognosticator in patients with UUTUC, multivariate analysis was performed with the Cox model (Table 3). In the entire cohort of the patients, positivity of either ERα or PR, or both was found to correlate with cancer-specific survival [hazard ratio (HR), 2.434; 95% confidence interval (CI), 1.055–5.617; P = 0.037], but not with tumor progression (HR, 1.011; 95% CI, 0.514–1.989; P = 0.197).
In this study, we immunohistochemically determined the expression levels of steroid hormone receptors, including AR, ERα, ERβ, GR, and PR, in a set of TMA consisting of 99 nephroureterectomy specimens. We obtained some results similar to previous observations in UUTUC and/or bladder urothelial tumor and others dissimilar to them.
Previous immunohistochemical studies have demonstrated that the positive rates of AR expression range 16–55% in UUTUC11,12 and 13–53% in BC.12,14-20 In accordance with our current findings in UUTUCs, significant decreases in the AR-positive rate have been reported in BCs, compared with non-neoplastic urothelial tissues.14-16 In contrast, 2 other studies showed no AR expression in normal bladder urothelium.18,20 In any of the studies involving UUTUCs,10-12 AR staining was not performed simultaneously in normal/benign urothelial tissues. Significant or marginal reduction of the AR positive rate has also been found in high-grade and/or muscle-invasive UUTUCs10,11 or BCs,12,14-16,18 compared with low-grade and/or superficial tumors, while some of other studies19,21 as well as our current study have shown no considerable correlations between AR positivity and tumor grade or stage. Prognostic significance of AR expression in BCs remains controversial: AR expression correlates with better18,21 or worse14,20 outcomes, whereas no strong correlations have been demonstrated in other studies.16,17 Additionally, previous10 and current studies assessed for AR expression in UUTUCs have failed to reveal its values as a prognosticator.
No previous immunohistochemical studies have examined ERα expression in UUTUC. Instead, ERα protein has been immunohistochemically detected only in a subset (e.g. 1–5%) of BC specimens,20,22-24 while our immunohistochemical analyses have shown relatively high positive rates in BC (27%14) and UUTUC (18%, current results). Although the exact reasons for the difference in the positive rate of ERα expression in BC remain uncertain, PCR-based analyses have detected ERα transcripts in virtually all the BCs examined.25 When compared with BC14,23 as well as UUTUC (current study), these studies showed elevated levels or positive rates of ERα protein expression in non-neoplastic urothelium. At least 2 of the studies have also demonstrated down-regulation of ERα expression in higher grade or stage BCs,14,23 while we failed to identify such a tendency in UUTUC. Additionally, there have been no published data suggesting the prognostic values of ERα expression in urothelial tumors. Nonetheless, the current study revealed that ERα negativity in combination with PR negativity was able to predict better patient outcomes as an independent indicator.
Shyr et al.11 showed that 36 (43%) of 83 UUTUCs, including 44% of low-grade vs. 43% of high-grade tumors (P = 1.000) as well as 47% of superficial vs. 51% of muscle-invasive tumors (P = 0.815), were immunoreactive for ERβ. In BCs, the positive rates of ERβ expression ranged from 22 to 76%,14,18,21,22,26,27 which was significantly lower than those in non-neoplastic bladder tissues.14,27 A significant elevation in ERβ positivity in BC, compared with benign urothelium, was also documented.28 More recently, all the 410 BCs examined were found to be positive for ERβ.24 Additionally, there have been considerable increases,14,21,22,26 considerable decreases,27 or no significant changes18,24 in the ERβ positive rate in higher grade and/or more invasive BCs. Furthermore, associations between ERβ overexpression and favorable21,24 or unfavorable14,18.28 prognosis of BC patients have been demonstrated. Our current data show significant reduction of ERβ positivity in UUTUC, compared with non-neoplastic urothelium, as well as no strong correlations between ERβ expression and tumor grade/stage or patient outcomes.
Our previous immunohistochemical analyses in the bladder TMA showed that most of non-neoplastic urothelial tissues (e.g., 96%) expressed the GR, which was significantly lower in BCs (e.g. 87%).6,29 Similar tendency was seen in our UUTUC specimens in the current study. To the best of our knowledge, no other studies have assessed the expression of GR in UUTUC or BC specimens. Meanwhile, loss of GR expression was strongly associated with higher grade/stage BCs as well as tumor recurrence/progression.29 However, these findings were not observed in UUTUCs in the current study.
An immunohistochemical study revealed that PR was positive in non-neoplastic bladder urothelium from 18 of 20 male children aged 1–12.30 However, subsequent studies showed no PR signals in 19823 and 41024 BCs or PR expression in only 2%/4% of non-neoplastic urothelium/BC tissues, respectively.20 We here demonstrate that 13% of normal-appearing urothelial tissues and 16% of UUTUCs, including 7% of low-grade vs. 18% of high-grade tumors (P = 0.453) as well as 14% of superficial vs. 18% of muscle-invasive tumors (P = 0.779) are immunoreactive for PR. More interestingly, PR positivity in pT3 or pT4 UUTUCs, but not in entire 99 cases, was strongly associated with disease-specific mortality. Indeed, in breast and endometrial cancer cell lines, progesterone has been shown to activate the mitogen-activated protein kinase signaling pathway and thereby up-regulate human telomerase reverse transcriptase31 whose activity is known to correlate with the progression of urothelial tumors.32,33 Inconsistent with these observations, in a study using a transgenic mouse model for BC, multiparous females were shown to develop significantly smaller tumors than nulliparous females, implying the preventing effects of progesterone (and estrogens) on BC outgrowth.34 Further studies are required not only to correlate between PR expression and the prognosis of UUTUC and/or BC in larger patient cohorts but also to determine the role of progestogens-mediated PR signals in urothelial tumor cell growth.
None of previous studies have demonstrated significant sex differences in the expression pattern of AR, ERα, ERβ, GR, or PR in urothelial tumors.9 It is therefore assumed that the disparities in the levels of circulating sex hormones and the systemic hormonal milieu lead to differential activation of, for instance, AR and ERs, in male vs. female urothelial cells. We observed that the positive rates of AR and PR tended to be higher in male and female UUTUCs, respectively. The significance of these findings needs to be further investigated.
Remarkably, ureteral tumors were inclined to exhibit higher positive rates of the expression of steroid hormone receptors, such as AR and ERβ, compared with renal pelvic tumors. A similar trend was indeed seen in a previous study assessing AR and ERβ expression in UUTUCs.11 The underlying reasons for these findings remain unclear but may include a difference in tissue preservation between ureteral and renal pelvic tumors. Delay to formalin fixation of surgical specimens is known to lead to false-negative results in steroid hormone receptor immunohistochemistry (e.g., ER/PR expression in breast cancer35). It may have taken a longer time to complete tissue fixation of renal pelvic tumors, especially deep lesions, compared with ureteral tumors, due to their anatomic locations/thickness of the specimens around the tumors. Interestingly, when analyzed our current data in ureteral tumors only, there was a significant decrease in AR positivity between low-grade [4 (67%) of 6] versus high-grade [10 (23%) of 44] carcinomas (P = 0.044), as seen in previous studies in ureteral tumor11 and BC.14,18 However, there were no statistically significant differences in AR positivity between non-muscle-invasive vs. muscle-invasive ureteral tumors as well as in the expression pattern of ERα, ERβ, GR, or PR between low-grade or non-muscle-invasive versus high-grade or muscle-invasive ureteral tumors.
Emerging evidence has indicated that steroid hormone receptor-mediated signals play a critical role in urothelial tumorigenesis as well as tumor progression.5,9 In particular, androgens, estrogens, and glucocorticoids have been shown to modulate tumor outgrowth in a variety of cell line and animal models. Accordingly, in these preclinical models, treatment with clinically available drugs, including AR antagonists or inhibitors being used in prostate cancer patients, ER modulators being used in breast cancer patients (or estrogens), or glucocorticoids being used as anti-inflammatory, anti-immune, or anti-tumor agents, resulted in inhibition of cell proliferation and invasion in urothelial tumor lines as well as tumor development and growth in animals. Recent retrospective clinical studies have also implied that androgen deprivation therapy for prostate cancer prevents bladder cancer development and recurrence.36,37 These findings further suggest potential application of currently available therapeutic options used for other pathological conditions to the treatment of urothelial tumors.
In conclusion, we observed that, consistent with prior observations in bladder specimens, the positive rates of AR, ERα, ERβ, and GR expression were significantly reduced in UUTUC, compared with non-neoplastic urothelium included in the same surgical specimens. PR was found to express in a subset of non-neoplastic urothelial tissues and UUTUCs. However, we failed to show significant differences in the expression of these 5 steroid hormone receptors between low-grade/superficial vs. high-grade/muscle-invasive tumors. Meanwhile, PR positivity alone was suggested to serve as a prognosticator in patients with deeply invasive UUTUC. Furthermore, loss of both ERα and PR in UUTUC, as an independent predictor, was strongly correlated with better patient outcomes.
UUTUC TMA was constructed with spotted triplicate tumor samples and paired normal-appearing urothelial tissues from 99 patients who underwent radical nephroureterectomy performed at Osaka General Medical Center, Osaka, Japan, as described previously.13 Appropriate approval was obtained from the institutional review board before construction and use of the TMA.
Immunohistochemical staining was performed, using each primary antibody [i.e. AR (clone N-20, dilution 1:200; Santa Cruz Biotechnology), ERα (clone 6F11, dilution 1:100; Ventana), ERβ (clone H-150, dilution 1:200; Santa Cruz Biotechnology), GR (clone H-300, dilution 1:200; Santa Cruz Biotechnology), PR (clone 16, dilution 1:100; Leica Biosystems)], as described previously.14,29 All the stains were manually scored by a single pathologist (H.M.) who was blinded to patient identity. The German Immunoreactive Score (range: 0–12) 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 (0, negative; 1, weak; 2, moderate; 3, strong) was considered negative (0; 0–1), weakly positive (1+; 2–4), moderately positive (2+; 6–8), or strongly positive (3+; 9–12).
The Fisher's exact test was used to evaluate the association between categorized variables. The survival rates were determined using the Kaplan-Meier method and comparison was made by the log-rank test. In addition, the Cox proportional hazards model was used to determine statistical significance of prognostic indicators in a multivariate setting. Tumor progression was defined as the development of non-bladder lesions, including recurrence at the nephroureterectomy site and lymph node or visceral metastasis. Patients were followed up from initial diagnosis to the appearance of the event of interest or the end of the study. P values less than 0.05 were considered statistically significant.
The authors have no conflicts of interest to disclose.