Hypoxia-inducible factor and mammalian target of rapamycin pathway markers in urothelial carcinoma of the bladder: possible therapeutic implications


Satish K. Tickoo, Department of Pathology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA. e-mail: tickoos@mskcc.org


What’s known on the subject? and What does the study add?

The hypoxia-inducible factor (HIF) and mammalian target of rapamycin (mTOR) pathways are important in tumorigenesis and novel agents targeting these respective pathways have shown promising activity in several malignancies. The current study demonstrates the expression of HIF and mTOR related pathway markers in urothelial carcinoma providing a rationale for clinical trials evaluating agents targeting these pathways.


To investigate the rationale for using targeted therapies against hypoxia-inducible factor (HIF) and mammalian target of rapamycin (mTOR) pathways in urothelial carcinoma of the bladder, by studying the immunohistochemical expression of molecules of these pathways in urothelial carcinoma, as recent pre-clinical studies and clinical trials have shown the potential utility of such targeted therapies.


Immunohistochemical stains were performed on a tissue microarray prepared from 92 cases of ≥ pT2 urothelial (transitional cell) carcinoma of bladder, using antibodies against HIF-1α and VEGF-R2, and phospho-S6 and phospho-4E BP1, molecules of HIF and activated mTOR pathways, respectively. Immunoreactivity was graded from 0 to 3+ (0, 0–5%; 1+, 6–25%; 2+, 26–50%; 3+, > 50% tumour cells positive).


In all, 58, 34, 35 and 17% of the tumours showed grade 2–3+ expression of phospho-4E BP1, phospho-S6, HIF-1α and VEGF-R2, respectively. Moderate correlation for immunoreactivity was observed between molecules within the same pathway [(phospho-4E BP1 with phospho-S6 (rho = 0.411), and HIF-1α with VEGF-R2 (rho = 0.265)], but not between molecules across pathways.


Urothelial carcinomas of the bladder express molecules of the HIF and mTOR pathways, providing a rationale for clinical trials evaluating agents targeting these pathways. Correlation between molecules within the same pathway, and not across pathways, suggests that investigating the usefulness of a specific targeted agent might benefit from pre-treatment evaluation of pathway marker expression.


hypoxia-inducible factor




metastatic RCC


mammalian target of rapamycin


platelet-derived growth factor


phosphoinositol 3-kinase


phosphatase and tensin homologue


urothelial carcinoma


vascular endothelial growth factor.


Urothelial (transitional cell) carcinoma (UC) of the bladder is the fifth most common cancer in humans, with an estimated 357 000 new cases worldwide and 145 000 deaths in 2002 [1]. In the USA, 68 810 new cases and 14 100 deaths were expected in the year 2008 [2]. Although UC is chemo-sensitive, the response durations are short and the median survival of patients with metastatic disease is ≈ 14 months [3]. Second-line trials with standard cytotoxic agents have generally not yielded encouraging results, with time to progression in the 2–3 month range and median survival of 6–9 months [4]. Furthermore, many patients receive these agents as first-line therapy, leaving few available therapeutic options for patients with progressive or recurrent disease. In patients with advanced UC that has progressed after first-line platinum-based therapy, there are no US Food and Drug Administration-approved agents for second-line treatment. The development of new therapies for treating patients with metastatic UC is desperately needed.

Over the past several years, molecularly targeted agents have been investigated in multiple tumour types. In UC, many initial pre-clinical studies suggested potential therapeutic roles for agents targeting epidermal growth factor receptors, their tyrosine kinases and erb-B2 (HER-2) [5,6]. However, to date, initial results on these have been disappointing [5–8]. Other therapies targeting hypoxia-inducible factor (HIF) and mammalian target of rapamycin (mTOR) pathway molecules, which have shown promising results in the management of patients with metastatic RCC (mRCC), have recently also been the focus of interest and investigation in UC [9–14].

Over-expression of HIF can result from tissue hypoxia, alterations in the von Hippel–Lindau (VHL) tumour suppressor gene leading to loss of pVHL (as in most cases of clear-cell RCC) or translational activation [15]. HIF over-expression leads to induction/activation of the hypoxia-response element (HRE) of a number of genes, including those for vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), glucose transporter 1 (GLUT1), and carbonic anhydrase IX (CA9) [15]. Many of these molecules are known to play a substantial role in tumourigenesis. mTOR, a ubiquitous serine–threonine kinase and a downstream component of the phosphoinositol 3-kinase (PI3K)/AKT/phosphatase and tensin homologue (PTEN)-signalling pathway, plays a critical role in the regulation of protein synthesis, cell growth, proliferation, apoptosis, survival and angiogenesis [16]. mTOR also acts as a translational activator of HIF through its activated downstream molecules, S6 and 4E BP1. Both the HIF and mTOR pathways are known to be activated in clear-cell RCC. In UC, angiogenesis and VEGF also seem to possess key roles in tumour initiation, progression and invasion [1,17,18].

Sunitinib, a potent inhibitor of the tyrosine kinase activity of the HIF pathway receptor tyrosine kinases, VEGF-R2 and PDGFR, has become a standard of care in the first-line treatment of patients with mRCC [19]. Similarly, rapamycin analogues/mTOR inhibitors such as temsirolimus and everolimus (RAD001) have shown marked activity in metastatic kidney cancer [19]. Also in UC, inhibitors of angiogenesis are reported to show activity in both pre-clinical models and some recent preliminary clinical trials [1,9–14]. Thus, sunitinib has been shown to enhance the activity of cisplatin – the most active agent in treating patients with urothelial malignancies [6] against UC in a pre-clinical murine xenograft model [11]. In another recent study, sunitinib was shown to decrease tumour growth rate, tumour stage and VEGF-R2 expression in an orthotopic mouse bladder cancer model [12]. Sunitinib has shown single-agent activity in a phase II trial in patients with relapsed or refractory UC [10]. mTOR inhibitors are also known to sensitize tumours to cisplatin, with RAD001 (everolimus) enhancing cisplatin-induced apoptosis in cells with high AKT/mTOR activity [20]. With regard to the anti-angiogenic effects of mTOR inhibition, in vitro studies have shown that rapamycin delays proliferation of transitional-cell carcinoma cell lines and decreases hypoxia-induced synthesis of VEGF [21]. In pre-clinical models, activation of the PI3K pathway through loss of the tumour suppressor PTEN or activation of AKT sensitizes tumour cells to mTOR inhibition, and in a phase I trial of patients with recurrent glioblastoma lacking PTEN expression, rapamycin showed anti-cancer activity [22]. PTEN mutations occur in approximately 30% of bladder cancers and PI3K has been shown to regulate bladder cancer cell invasion, with over half of primary human bladder tumours demonstrating high AKT phosphorylation [23]. In two recent studies [12,14], everolimus was shown to inhibit the growth of multiple human bladder cancer cell lines, including UMUC-3, which has a PTEN mutation and constitutively active AKT. Protein synthesis inhibition via the S6K and 4EBP1 pathway appeared to be the main mechanism of cell growth inhibition by everolimus, with marked inhibition of phosphorylation of S6 downstream of mTOR and VEGF [12]. Additionally, tumour weights from nude mice bearing human KU-7 subcutaneous xenografts treated with everolimus showed significant reduction as compared with placebo-treated mice [14].

At immunohistochemical (IHC) levels, over-expression of HIF-1α is reported to be a significant predictor of the time to first recurrence in T1 bladder cancer [24], and shows significantly greater expression in muscle-invasive than in more superficial tumours [25]. HIF-1α over-expression is associated with worse disease-free survival [24,26], but not overall survival [24]. Thus, there is accumulating justification for further investigation of the HIF and mTOR pathways in UC, with many potential therapies available [10,16].

To further explore the rationale for targeting the HIF and mTOR pathways in UC, we performed an IHC study to assess the expression of HIF and mTOR pathway markers, as well as the inter-relationship of the respective markers, in high-stage UC of the bladder.


The study was approved by the institutional review board. A tissue microarray was constructed from one representative formalin-fixed, paraffin-embedded tissue block from 92 consecutive cases of cystectomy or transurethral resection specimens with invasive UC showing muscularis propria or deeper invasion (≥pT2). No tumours with aberrant differentiation, i.e., adenocarcinoma, squamous cell carcinoma, small cell carcinoma, and the like, were included in the study, and all tumours were histologically high-grade, as is common in invasive UC. Three core tissue biopsies, 0.6 mm in diameter, were taken from morphologically representative regions from each block and precisely arrayed using an automated tissue arrayer ATA-27 (Beecher Instruments, Inc, Sun Prairie, WI, USA). IHC stains were performed using antibodies against phospho-S6 and phospho-4E BP1 (markers for activated mTOR pathway), and HIF-1α and VEGF-R2 (markers for HIF pathway). Positivity was graded as 0 to 3+ (0, 0–5%; 1+, 6–25%; 2+, 26–50%; and 3+, >50% tumour cells positive) for phospho-S6 (cytoplasmic), phospho-4E BP1 (cytoplasmic and/or nuclear), HIF-1α (nuclear) and VEGF-R2 (cytoplasmic). Only 2 or 3+ positivity for any marker was regarded as a positive expression. The presence of at least two evaluable cores from each tumour was considered necessary for inclusion in the final evaluation. Appropriate positive and negative controls were used with each run of immunostaining. The immunostaining procedures and interpretations are well established and standardized in our laboratory, mainly based on the expression profiles in molecularly well-characterized cell lines, and human kidney, breast and prostate cancer tissues [27,28]. For HIF-1α and VEGF-R2, the positive controls consisted of tissue sections from clear-cell RCC with known positivity for these two markers, and for phospho-S6 and phospho-4E BP1, tissue sections from prostatic adenocarcinoma previously shown/reported to express these markers were used as positive controls. Results on evaluable cores were averaged for each patient. The details of the antibodies used are described in Table 1.

Table 1.  Antibodies used for the IHC analysis
AntibodySourceClone and other characteristicsAntigen retrievalFinal dilution
  • *

    Using Catalyzed Signal Amplication (DakoCytomation) system

HIF-1α*Novus Biologicals, Littleton, CONB100-123, mouse monoclonalHeat (oven) at 62o F for 60 min1 : 1600
VEGF-R2Cell Signaling Technology, Danvers, MA55B11, rabbit monoclonalEDTA, pH 8.01 : 125
p-S6Cell Signaling Technology, Danvers, MASer240/244, rabbit polyclonal10 mM citrate buffer, pH 6.01 : 1000
p-4E BP1Cell Signaling Technology, Danvers, MAThr37/46, rabbit polyclonal10 mM citrate buffer, pH 6.01 : 500

The non-parametric Spearman’s correlation coefficient test was utilized to assess the measure of correlation between IHC expression of the individual pathway markers in the tumour specimens.


The immunoreactivity results in the tumour cells are summarized in Table 2. Phospho-4E BP1 showed positivity (2–3+) in 49/84 (58%) tumour specimens. HIF-1α and phospho-S6 showed similar grades of expression in 30/85 (35%) and 29/85 (34%) tumour specimens, respectively (Figs 1,2). VEGF-R2 tumour cell positivity was only seen in 14/86 (17%) specimens; however, 38/86 (44%) of the tumours showed positive results for VEGF-R2 in the tumour vasculature.

Table 2.  Overall grading of immunoreactivity for the four antibodies in the tumours
Antibody(N)Grade 0 (%)Grade 1+ (%)Grade 2+ (%)Grade 3+ (%)
HIF-1α (85)41 (48)14 (17) 8 (9)22 (26)
VEGF-R2 (86)66 (77) 6 (7) 6 (7) 8 (9)
p-S6 (85)45 (53) 11 (13) 9 (11)20 (23)
p-4E BP1 (84)25 (30)10 (12)17 (20)32 (38)
Figure 1.

A high-grade invasive UC showing 3+ positivity for phospho-4E BP1 (A; predominantly nuclear) and phospho-S6 (B; cytoplasmic). Staining for both HIF-1α (C) and VEGF-R2 (D) is negative in this case.

Figure 2.

Tumour showing 1+ staining for phospho-4E BP1 (A) and phospho-S6 (B). However, HIF-1α (C; nuclear) and VEGF-R2 (D; cytoplasmic) are both 3+ immunoreactive.

Comparison of the grading of immunoreactivity in individual tumours showed that, of 29 tumours with 2–3+ positivity for phospho-S6, similar grades (2–3+) of staining were present in 25 (86%) tumours for phospho-4E BP1, 10 (34%) for HIF-1α and four (14%) for VEGF-R2. On the other hand, of the 56 tumours with 0 or 1+ staining for phospho-S6, 32 (57%), 36 (64%) and 46 (82%) tumours were 0 or 1+ for phospho-4E BP1, HIF-1α and VEGF-R2, respectively. Similarly, of the 49 tumours with 2–3+ positivity for phospho-4E BP1, 25 (51%), 14 (28%) and eight (16%) tumours stained 2–3+ for phospho-S6, HIF-1α and VEGF-R2, respectively. On the other hand, of the 35 tumours with 0 or 1+ staining for phospho-4E BP1, 24 (69%), 20 (57%) and 30 (86%) tumours were also 0 or 1+ for phospho-S6, HIF-1α and VEGF-R2, respectively.

Analyses using Spearman’s correlation coefficient indicated a moderate correlation between the expression of phospho-S6 and phospho-4E BP1 (rho = 0.411), and between HIF-1α and VEGF-R2 in tumour cells (rho = 0.265). The correlations between HIF-1α and phospho-S6 (rho = 0.117), HIF-1α and phospho-4E BP1 (rho =−0.149), VEGF-R2 tumour cell and tumour vasculature expression (rho = 0.047), VEGF-R2 tumour cell and phospho-S6 (rho = 0.136), and VEGF-R2 tumour cell and phospho-4E BP1 (rho =−0.054) were either very weak or there was little to no correlation between them.


The present study shows that HIF and mTOR pathway markers are often expressed in high-stage (≥ pT2) UC by IHC. Over-expression of these pathway markers at the protein level lends support for the role of targeted therapies against HIF and mTOR pathways in high-stage UC. This is consistent with the observations that sunitinib is active against multiple human bladder cancer cell lines [11], and that mTOR inhibition depresses the growth of multiple human bladder cancer cell lines [12,14], including those with PTEN mutations/active AKT, primarily through inhibition of protein synthesis in the mTOR pathway [12].

Thus, there are clear potential therapeutic implications for the over-expression of HIF and mTOR pathway markers in UC, with several recently reported and ongoing/planned clinical trials using therapies targeting these pathways. Sunitinib, a potent inhibitor of HIF pathway receptor tyrosine kinases, has shown single agent activity in patients with advanced UC [10,29]. A randomized, double-blind, placebo-controlled phase II trial of maintenance sunitinib vs placebo after chemotherapy for patients with advanced UC is ongoing [9], and a phase II trial of sunitinib in combination with gemcitabine and cisplatin chemotherapy in the neoadjuvant setting is planned. The VEGF-targeted monoclonal antibody, bevacizumab, in combination with chemotherapy is under investigation in patients with advanced UC, with an ongoing phase II trial evaluating gemcitabine and carboplatin plus bevacizumab in the first- line treatment of patients with an Easter Cooperative Oncology Group performance status of 0–1 [30]; and a planned phase III randomized co-operative group trial (CALGB) evaluating gemcitabine and cisplatin with or without bevacizumab in the first-line setting for patients with metastatic UC. With regard to the mTOR pathway, a phase II trial of the mTOR inhibitor, RAD001, in the second-line setting for patients with advanced UC is ongoing. The current IHC study provides a further rationale for the above clinical trials to target these respective pathways.

There is a relationship between the expression of markers within the same pathways (i.e. HIF-1α with VEGF R2; and phospho-S6 with phospho-4E BP1), but not between the HIF and mTOR pathways. Although the activated mTOR pathway is a prominent translational activator of HIF, the present study suggests that a one-to-one relationship between the two might not exist in some high-stage UCs, and that mechanistically alternative routes for activation of each of these pathways could exist. Given the technical variability, including those related to tissue procurement, type and time of fixation, tissue processing as well as others, IHC alone might not be adequate to investigate the relationship between the HIF and mTOR pathways. However, we have found a close association using IHC between molecules of these two pathways in our studies in RCC tumour specimens (unpublished data), suggesting that technical reasons alone might not be the only reason for the lack of correlation between these pathways in UC.

Another aspect of IHC staining for phospho-S6 and phospho-4E BP1 that needs to be taken into account is the fact that delayed fixation of tissues has been reported in some cases to undergo spontaneous dephosphorylation of some phosphorylated antigens. However, the reported results have been quite inconsistent [31]. There are no reports of such an occurrence for phospho-S6 and phospho-4E BP1 in the current literature. This could be because these antibodies have not been in use for very long. However, in this group of tumours, we did not observe any significant differences in expression of phospho-S6 and phospho-4E BP1 between transurethral resection (usually immediate fixation) and cystectomy (often relatively delayed fixation) specimens. Prospective studies to investigate this issue are merited.

The combination of HIF and mTOR pathway targeted agents is being investigated in patients with advanced RCC in an ongoing trial evaluating the combination of sunitinib and RAD001. One of the potential therapeutic implications of the differential expression of HIF and mTOR pathway markers in UC is that selecting a specific targeted therapy may be best determined by pre-treatment evaluation of pathway marker expression. However, marker expression by IHC does not necessarily correlate with predictors of outcome, as evidenced by the discordance between the presence of epidermal growth factor receptor-activating mutations in lung cancer and epidermal growth factor receptor positivity by IHC [32]. Several of the ongoing trials evaluating HIF and mTOR pathway targeted agents in UC are attempting to determine the predictive value of IHC for pathway marker expression in pre-treatment tumour specimens.

In conclusion, the expression of HIF and mTOR pathway markers in UC of the bladder provides further justification for the evaluation of targeted therapies against these pathway markers in patients with advanced UC. Additional studies are needed to determine if marker expression by IHC has predictive or prognostic value in UC. Clinical trials can then focus on the development of rational combination strategies that target specific molecular alterations in these tumours. For patients with advanced disease, there is no curative therapy and no standard of care for treating those who progress after first-line platinum-based chemotherapy. Trials designed to evaluate novel targeted agents in patients who are most likely to respond will accelerate the development of promising therapies.


None declared.