Background
Bladder cancer is the fifth most common malignancy in most Western industrialized countries with urothelial carcinoma (UC) representing the major histological subtype. UC can be further classified into papillary and invasive carcinomas [
1]. While most cases of papillary low grade urothelial tumours have a good prognosis, high grade papillary tumours tend to progress towards invasive cancers which are often lethal. A large fraction of invasive urothelial carcinomas develop by a different route via high-grade dysplastic carcinoma in situ [
2]. Depending on tumour stage, urothelial carcinoma treatment is treated either by transurethral resection or radical cystectomy. In addition neo-adjuvant or palliative chemotherapy is used in locally advanced or metastatic urothelial carcinomas, respectively [
3,
4]. Local immunotherapy or chemotherapy is used for preventing recurrences of superficial cancers. Until now, clinical studies using molecular targeted drugs have yielded disappointing results [
5‐
7]. This failure is generally considered to reflect our insufficient knowledge of urothelial tumour biology.
The Notch signalling pathway is a promising target for cancer therapy [
8,
9]. In many tissues, Notch signalling contributes to the maintenance of stem cells or precursor cells [
10] and pathway hyperactivation caused by point mutations [
11] or translocations involving
NOTCH genes [
12,
13] is oncogenic by blocking cell differentiation, protecting against apoptosis and stimulating proliferation. Conversely, in certain epithelia, Notch signalling promotes differentiation. Accordingly, pathway inactivation, frequently caused by NOTCH1 mutations, is observed in squamous carcinomas of several organs [
14‐
16]. Therefore, depending on tissue and tumour type, drugs targeting Notch signalling may be therapeutic, useless or tumour-promoting [
17].
Notch signalling constitutes a short-range communication system between adjacent cells. Canonical Notch signalling is activated by ligand-receptor interactions releasing the receptor intracellular domain (NICD) by two proteolytic steps, including an essential cleavage by γ-secretase. The NICD translocates into the nucleus where it binds to C Promoter Binding Factor-1 (CBF-1), recruits co-activators (MAML1 and KDM5A) and typically activates the transcription of
Hairy and Enhancer of Split (HES) and
Hairy and Enhancer of Split related (HEY) target genes. Through such effectors Notch determines cell fate [
18].
To date, little information exists on the Notch signalling pathway in normal urothelium or urothelial cancers. Decreased expression of several Notch receptors and ligands in immunohistochemistry has been reported especially in papillary UC [
19]. We have investigated the expression of Notch pathway components in urothelial carcinoma tissues and cell lines and the function of NOTCH1 in UC cell lines.
Methods
Tissue samples
The benign and tumour bladder samples used for gene expression studies were a subset of those described previously [
20] comprising 11 benign bladder tissues and 30 bladder cancer tissues from 25 male and 5 female patients with ages ranging from 54 to 84 years. Tumour stages and grading according to the UICC classification were as follows: pT3 G3 in 11 cases, pT4 G3 in 6 cases, pT2 G2 in 6 cases, pT2 G3 in 3 cases and one case each of pT3 G2, pT1 G2, pTa G3 and pTa G2 tumours. The set used for immunohistochemistry comprised 4 normal tissues, 27 UC tissues with 6 pTa low grade, 5 pTa high grade, 5 pT1, 3 pT2, 3 pT3, 4 pT4, 1 CIS. The study was approved by the ethics committee of the medical faculty of the Heinrich Heine University and informed consent was obtained from all patients.
Cell lines and primary cultures
All UC cell lines (5637, 639v, 647v, BFTC905, HT1376, J82, RT4, RT112, SD, SW1710, UM-UC3, VM-Cub1, T24) and HEK293 cells were cultured in DMEM GlutaMax (Gibco, Darmstadt, Germany), supplemented with 10% fetal calf serum [
21]. They were obtained from the DSMZ (Braunschweig, Germany), except for UM-UC3, kindly provided by Dr. Grossman, Houston. The well-differentiated UC cell line BC61 was cultured and characterized as previously described [
22,
23]. Primary urothelial cells (UP) were prepared and maintained as described (20). T47D cells were maintained in RPMI1640 supplemented with 15% foetal calf serum and 10 μg/ml insulin.
Plasmids
Human N1ICD-pIRES2 was a kind gift from Prof. Giebel, Essen; human full-length wildtype NOTCH1 in a pcDNA3-HA/FLAG vector was kindly provided by Prof. Di Fiore, IFOM, Italy [
24].
Gamma secretase inhibitor treatment
Cells were treated with 0.25 - 20 μM γ-secretase inhibitors DAPT (#2634, TOCRIS/R&D Systems), L-685,458 (#2627, TOCRIS/R&D Systems) or Compound E (#56597, Calbiochem/Millipore) or, as a control, compound W (#2654, TOCRIS/R&D Systems) for 48 h when viability was analysed by MTT assay.
Transfection and reporter gene assay
For reporter gene assays, cells seeded in 6-well plates were transiently transfected with XtremeGene9 (Roche) with 1 μg of either pJH26A (Notch responsive reporter) or pJH28A (Notch non-responsive reporter), kindly provided by Prof. D. Hayward [
25], or in other experiments either pHES1 (+CBF1)-luc (addgene 41723; similar to pJH26A) or pHES1 lacking the CBF1 site (addgene 43805, similar to pJH28A). p850-luc and pGL3 (Promega) were used as positive or negative controls, respectively. NOTCH1 expression plasmids were added as indicated in the individual experiments, using the respective plasmid vectors for adjusting total amounts transfected. All cells were co-transfected with 10 ng Renilla luciferase plasmid for normalization. Cells were harvested 48 h after transfection and assayed for luciferase activities by the Dual-Luciferase reporter assay (Promega). Notch activity was determined as the ratio of wildtype to mutant plasmids and normalized to p850-luc and Renilla luciferase activity.
Ligand dependent Notch activation assay
The ligand-dependent activation assay was carried out as described [
26]. Briefly, urothelial cancer cell lines seeded in 6 well plates were co-transfected at 80% confluency with NOTCH1 full-length expression plasmid (or vector), the Notch-responsive reporters pJH26A or pJH28A and Renilla luciferase plasmid at a 1:1:0.1 ratio. After 24 h cells were replated in triplicates in 24 well plates coated with ligand or controls. For that purpose, the wells were first coated with 20 μg/ml goat anti-human IgG
1C Fc secondary antibody (ThermoFisher Scientific, Germany), washed, and then incubated for 2 h with 10 μg/ml DLL1 extracellular domain (aa 1–545) fused at its C-terminus to the Fc portion of human IgG (DLL1:Fc; AdipoGen, Switzerland) or human IgG (ThermoFisher Scientific, Germany), followed by a further wash. After a further 24 h, cells were harvested and assayed for luciferase activities by the Dual-Luciferase reporter assay (Promega). Basal activity in cells exposed to immobilized IgG was set as 1 for each cell line.
For the colony forming assay 105 cells were seeded in 6 well plates and transfected with pIRES2-N1ICD or the corresponding pIRES2 empty vector, or with pcDNA4/to-Notch1 full-length (N1-FL) or the corresponding pcDNA4/to-lacZ plasmid. After 24 h cells were seeded in triplicates into 10 cm plates and selection for stable integration of the plasmids was started the next day. Cells transfected with pIRES2 plasmids were treated with 1.0 - 1.8 mg/ml Geneticin Selective Antibiotic G418-Sulfat (Invitrogen) and cells transfected with pcDNA4/to-plasmids were treated with 100–150 μg/ml Zeocin Selection Reagent (Invitrogen) every two days for 2 weeks. After a one week recovery period with lower antibiotic concentrations, the remaining colonies were fixed with 100% methanol and stained with Giemsa solution (Sigma-Aldrich, Germany). Colonies were then counted and statistical comparisions were made by Student’s T-Test.
RNA isolation and RT-PCR
Total RNA was isolated from sub-confluent cell lines using the RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA was reverse transcribed using 200 U SuperScriptII reverse transcriptase (Invitrogen, Darmstadt, Germany), with 300 ng oligo-dT and 25 ng random hexamer primers. Real Time-PCR assays were performed with the ABI7900HT System using the QuantiTect SYBR Green PCR Kit (Qiagen) and premade or self-designed primers (Additional file
1: Tables S1–S2).
TBP was used as a reference gene. Statistical comparisons were made by Mann–Whitney U test with SPSS 21.
Immunocytochemistry
Cells grown to 70% confluence were fixed with 100% ice-cold methanol for 5 min, permeabilized in 0.1% Saponin-TBS solution and blocked in 1% bovine serum albumin in 0.2% Tween-TBS for 30 min each. Cells were then incubated with the primary antibodies NOTCH1 (Epitomics, EP1238Y), JAG1 (Santa Cruz Biotechnology, H114), DLL1 (Santa Cruz Biotechnology, H-265; Abcam, ab84620) and NOTCH2 (Abcam, ab8926) and subsequently with goat-anti-mouse Alexa Fluor488 antibody. Nuclei were stained with 1 μg/ml DAPI (Dako, Hamburg, Germany).
Western blotting
Subconfluently grown cells were lysed in RIPA buffer with protease and phosphatase inhibitors. Western blot analysis after SDS-page was performed using primary antibodies against NOTCH1 (Epitomics, EP1238Y), JAG1 (Santa Cruz Biotechnology, H114), DLL1 (Santa Cruz Biotechnology, H-265; Abcam, ab84620), NOTCH2 (Abcam, ab8926) and α-tubulin (B-1-5-2, Sigma) as a control. Secondary antibodies labelled with HRP and the ECL advanced or ECL Select kits (GE Healthcare) were used for detection.
Immunhistochemistry
Immunohistochemistry was performed by standard techniques with antigen retrieval by citrate buffer pH 9. Antibodies used were NOTCH1 (Epitomics), JAG1 (Santa Cruz) and DLL1 (Santa Cruz). Detection was achieved via a biotin-conjugated polyvalent antibody with an avidin-biotin-peroxidase reagent (Scy Tek, Logan, USA) and diamino-benzidine staining. Sections were counter-stained with hemalum. The staining results of the reference tissues for antibody specificity are provided in Additional file
1: Figure S1.
Discussion
Taken together, our results from tissues and cell lines suggest that canonical Notch signalling towards the nucleus is inactive in urothelial carcinoma, especially in the invasive subtype. NOTCH1 and DLL1 were most consistently diminished in both cell lines and tissues. Downregulation of these factors is in accord with a previous report [
19], whereas another paper reported decreased NOTCH1 expression in high-grade invasive UC tissues only, with more pervasive cytoplasmic localization [
28]. The mechanisms underlying these changes remain to be determined. Sequencing data from the TCGA consortium (
http://cancergenome.nih.gov/) indicate only few mutations in Notch pathway components in urothelial carcinoma (Additional file
1: Table S3). The frequency of NOTCH1 mutations is about 5% and is thus considerably lower than in head-and-neck squamous cell carcinomas [
16]. A more likely cause of downregulation are copy number changes at chromosome 9q34 (where
NOTCH1 is encoded) which occur in up to 30% of invasive UC. However, neither mutations nor gene loss can fully account for the high prevalence of
NOTCH1 downregulation. As both, NOTCH1 and DLL1, are predominantly localized in more differentiated cells in the normal urothelium, their decreased expression in tumours may conceivably reflect to some extent the loss of this differentiated cell population.
In addition to downregulation, we frequently observed delocalization of NOTCH1 and DLL1 to the cytoplasm in tumour cells. Delocalization of NOTCH proteins to the cytoplasm in some cells reflects lack of activation by external ligands in trans but enables cis-inhibition by ligands expressed in the same cell compartment [
29,
30]. Candidates for that function are JAG1 and especially JAG2, which appear to be retained or in the case of JAG2 even upregulated in invasive UC [
31]. JAG1 was mostly cytoplasmic in tumour cells and is thus unlikely to contribute to trans-signalling. Notably, the distribution of JAG1 in normal urothelium is quite similar to that of p63 [
32], a factor known to maintain basal cells in many epithelia and to induce JAG1 in many instances. The co-localization of DLL1 and NOTCH1 proteins in normal urothelium suggests that DLL1 could be the physiological ligand for NOTCH1 in this tissue. However, exposure of UC cell lines to immobilized DLL1, which is considered to yield superior activation compared to soluble ligand [
26] did not enhance activation of a Notch-dependent reporter by transfected NOTCH1 protein, least activate canonical signalling by itself.
We note that NOTCH2 expression was less strongly changed than that of NOTCH1 and that active N2ICD could be detected in the nuclei of some UC cell lines. Interestingly, it appeared more strongly expressed in mesenchymal UC cell lines albeit localized to the cytoplasm. Increased expression of NOTCH2 in UC cell lines with a mesenchymal phenotype as compared to lines with an epithelial phenotype has recently been reported by others [
33,
34]. The same authors observed that NOTCH2, but not NOTCH1, favors epithelial-mesenchymal transition and cell migration in UC cell lines [
33,
34]. Collectively, these findings suggest divergent functions of NOTCH1 and NOTCH2 in UC, which have also been observed in other cancers [
35,
36]. Nevertheless, despite the expression of N2CTD at a level comparable to that in normal urothelial cells, two different NOTCH-dependent reporters showed no evidence for canonical pathway activity in the UC cell lines. Likewise, γ-secretase inhibitors did not block cell proliferation in UC cell lines. These findings argue that canonical Notch signalling to the nucleus is inactivated in urothelial carcinoma. As a consequence, it seems plausible that the effects of NOTCH2 in mesenchymal UC cell lines, considering especially its cytoplasmic localisation, could be exerted by non-canonical signalling.
Across the cancer tissues, expression of HES1 mRNA was significantly increased compared to normal bladder, despite the significant downregulation of NOTCH1, NOTCH2 and DLL1. Although no upregulation of HES1 was observed in the UC cell lines compared to normal urothelial cells, this finding is somewhat puzzling, since HES1 is a common target gene of the canonical Notch pathway. However, target genes of the pathway vary strongly between tissues and even HES1 is not a universal target [
18]. Moreover, HES1 is regulated by several pathways, including Hedgehog [
37], ATF2 [
38], and signalling through JNK1 [
39]. In UC cell lines, a reporter plasmid driven by the HES1 promoter was similarly active, if the CBF sites were wild-type or mutant, supporting our conclusion that canonical Notch signalling is inactive. However, luciferase was induced by N1ICD cotransfection indicating that HES1 can be a target of an active Notch pathway in urothelial cells. Collectively, these findings suggest that the increased expression of HES1 in UC tissues could reflect its predominant regulation by a different pathway in the absence of canonical Notch signalling.
In normal urothelial cells and urothelial carcinoma cell lines, overexpression of full-length NOTCH1 or its active intracellular domain induced canonical reporters but was not compatible with long-term proliferation of UC cells. Inhibition of proliferation did not appear to occur by straightforward inhibition of cell cycling or by a major induction of apoptosis. Rather, NOTCH1 transfected cells often showed misshaped or abnormally large nuclei or became obviously binuclear. Intriguingly, in normal urothelium NOTCH1 is localized in the upper layers together with its likely ligand DLL1, where they might contribute to terminal differentiation. In particular, umbrella cells of the urothelium are usually polyploid [
40]. In squamous epithelia, Notch signalling is most important for establishing the commitment to terminal differentiation in the suprabasal layer [
41,
42]. Notch signalling is well-established as a regulator of polyploidization in Drosophila, but may also act in this fashion in certain mammalian cells [
43]. Evidently, the urothelium is one tissue, where this function should be investigated in detail. The observed effects of NOTCH1 overexpression on UC cells may thus be conceptualized by assuming that it induces a failed attempt at differentiation. Evidently, the effects of pharmacological inhibition of the Notch pathway and genetic inactivation of individual components such as NOTCH1, DLL1 and NOTCH2 should be studied in culture and animal models of urothelial differentiation.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AG and WAS conceived the study. GN was responsible for tissue acquisition. SJ performed the immunohistochemistry stainings and analyses. AG, JS, PN, and MJH carried out experiments and analysed data. AG and WAS wrote the paper. All authors had final approval of the submitted and published versions.