Introduction

Due to their level of incidence and recurrence, urinary bladder cancer (UBC) is a worldwide concern. Despite the significant advances in therapy, including pharmacological and surgical techniques, these tumours are the ninth most common cause of cancer death in men [1].

Over the last decades several attempts, using a variety of different methods, have been applied in order to characterize the genetic profile of these tumours identifying many oncogenes and tumour suppressor genes that contribute to tumourigenesis when activated or inactivated, respectively [2, 3]. Conventional cytogenetic studies revealed complex karyotypes with a high degree of aneuploidy and unbalanced rearrangements that could be associated to a more aggressive tumour growth [4, 5]. With comparative genomic hybridization (CGH) and, more recently, array CGH (aCGH), more detailed knowledge about the tumours’ genome concerning gained and lost regions of complex chromosomal changes was obtained, allowing to identify previously unknown genomic amplifications at the level of individual chromosomal bands [2, 6]. As a complement of conventional cytogenetic analysis and aCGH studies, fluorescence in situ hybridization (FISH) permits a more accurate definition of karyotype [7]. Multiplex ligation-dependent probe amplification (MLPA) is a sensitive and specific technique to study the gene copy number variations also applied in the carcinogenesis study [8].

The most frequent cytogenetic alterations (found in more than 50 % of all UBC), sometimes the sole alterations, involves monosomy or regional deletions at chromosome 9, and it is thus considered an early event in UBC [9]. The loss on 9p involves the CDKN2A and CDKN2B genes and the loss on 9q involves the DBC1 gene [10]. Other common described losses, which should be associated with tumour suppressor losses, include the complete or partial loss of the short arm of chromosomes: 3, 4, 5, 6, 8, 10, 11 and 17, and of the long arm of chromosomes: 2, 4, 6, 10, 13, 14 and 18 [3, 5, 10]. Chromosome 3q (PI3K), 7p (EGFR) and 17q (ERBB2) gains, associated with proto-oncogene activation, are of particular interest, given their potential diagnostic and prognostic value [11, 12]. Amplified regions are one common finding in these cancers, namely, the one located at 6p22.3 (E2F3), detected in up to 25 % of advanced stage UBC [6, 13].

Based on molecular findings, it was proposed that non-invasive and muscle-invasive tumours are pathobiologically distinct and arrive from two different pathways [13, 14]. The first pathway (relevant for 70–80 % of tumours) leads to the development of mostly low-grade non-invasive papillary tumours which often recur, but are rarely lethal with a usually favourable prognosis [4]. These tumours frequently over-express or express constitutively active mutants of HRAS and FGFR3 leading to hyperactivated Ras/MAPK signalling activity [14, 15]. The second pathway (relevant for 20–30 % of tumours) could originate from the progression of dysplasia to flat carcinoma in situ (CIS) and high-grade non-invasive lesions and leads to the formation of highly invasive and metastatic tumours with a worst prognosis [16]. This pathway is characterized by the inactivation, usually via loss-of-function of tumour suppressor genes, e.g. TP53, PTEN, RB1 and over-express EGFR and ERBB2, MMP2 and MMP9 as well as other pro-angiogenic factors, while having deleted cyclin-dependent kinase inhibitor genes CDKN2A (p16Ink4a) and CDKN2B (p15Ink4b) [17, 18].

UBC cell lines are very important research tools for studying biological and molecular mechanisms in UBC [16, 17, 19]. There are approximately 50 human UBC cell lines that have been used in research studies [20, 21]. There has been controversy surrounding cell line classification, and doubts have been raised over the possibility of cross-contamination of cell lines [22, 23]. To ensure the validity and reproducibility of the experimental results, it is essential that cell lines are thoroughly characterized. T24, 5637 and HT1376 cell lines have been successfully used in several studies of UBC, namely, to identify potential therapeutic targets [24, 25].

In this study, we have characterized the genomic profile of three established UBC cell lines: one non-muscle-invasive UBC cell line (5637) and two muscle-invasive UBC cell lines (T24 and HT1376), through a combination of classical cytogenetics techniques (cell culture synchronization and G banding), FISH (subtelomeric, painting, alpha satellite and unique sequence probes) and aCGH analysis. Additionally, some important genes related to carcinogenesis were studied by MLPA in order to access copy number variations. The pathways and biological functions of the altered genes involved were also assessed. We accomplished a more complete picture of genomic characterization of these cell lines, obtained a good correlation with the profiles of UBC, thus validating 5637, T24 and HT1376 cell lines as representative models for the study of human urinary bladder carcinogenesis to assay new methodologies and therapeutic approaches.

Material and methods

Cell lines

The cell lines characterized in this work were 5637, T24 and HT1376. 5637, T24 and HT1376 cell lines were provided by DSMZ, Düsseldorf, Germany; 5637 and HT1376 cell lines were a kind gift from Dr. Paula Videira of the Universidade Nova de Lisboa, Portugal. Cells were cultured in RPMI 1640 medium (PAA, Pasching, Austria), supplemented with 10 % heat-inactivated fetal bovine serum (Biological Industries, Kibbutz Beit Haemek, Israel), 100 U/mL penicillin (Biological Industries), 100 μg/mL streptomycin (Biological Industries) and 2 mM l-glutamine (Sigma Aldrich, St Louis, MO, USA). The cell lines were cultured as a monolayer at 37 °C in 5 % CO2 in a humidified atmosphere, and the cells were routinely subcultured by trypsinization.

Karyotyping and FISH

The cell culture was performed with optimized protocols, culture with 17 h of 5-bromo-2′-deoxyuridine (BrdU; Sigma Aldrich, St Louis, MO, USA) followed by 2 h of colcemid (Biological Industries, Beit Haemek, Israel) incubation. The cells were harvested, and the slides were banded according to routine protocols. Chromosomes of 50 metaphases were counted in order to determine the modal number and at least 25 metaphasic cells were analyzed for each cell line. Metaphase were digitally imaged and karyotyped with a CW4000 Karyotyping system (Leica). Karyotype description followed the International System for Human Cytogenetic Nomenclature 2013 recommendations [26].

FISH was applied to freshly prepared slides from cells fixed in methanol-acetic acid. Subtelomeric probes for all chromosomes (Vysis ToTelVysion), painting probes (Star FISH probes, Cambio), alpha satellite (Vysis) and unique sequence probes for RB1 and TP53 genes (Vysis) were performed according to the manufacturer’s instructions. The FISH images were digitally imaged and analyzed using Leica software.

DNA extraction

DNA was extracted from the three cell lines using the High Pure PCR Template Preparation Kit (Roche Applied Science, Indianapolis, USA) according to the manufacturer’s instructions.

aCGH

aCGH was carried out using an Agilent SurePrint G3 Human Genome microarray 180 K (Agilent Technologies, Santa Clara, CA, USA), an oligonucleotide microarray containing approximately 180,000 60-mer probes with a 17-kb average probe spacing. Briefly, DNA from each cell line was labelled with cyanine 5-dUTP (Cy5) by random primer labelling, according to the manufacturer’s instructions. Control DNA was obtained from peripheral blood lymphocytes of sex-matched donors (previously studied and without significant genomic imbalances) and labelled with cyanine 3-dUTP (Cy3). Excess primers and nucleotides were removed using Amicon 30-kDA individual filters (Millipore, Billerica, MA, USA). Cy5-labelled tumour DNA and Cy3-labelled reference DNA were combined with Human Cot-1 DNA (Kreatech Diagnostics, Amsterdam, Netherlands), treated with Agilent blocking agent and 2× Hi-RPM buffer and hybridized in a 4× 180 K oligonucleotide slide, at 65 °C for 24 h in a hybridization oven (Agilent Technologies) at a constant rotation of 20 rpm. Then, the slide was washed with buffers and read (scanner C, Agilent). Data was processed with the Feature Extraction software (v10.7), and the results were analyzed using Agilent Genomic Workbench v6.5.

MLPA

MLPA, using the MLPA mixes: P005, P006 and P007 (MRC-Holland, Amsterdam, Netherlands), was performed on the three cell lines and controls as described by Schouten and colleagues [8]. Briefly, 100 ng of DNA samples were heated at 98 °C for 10 min. After the addition of the probe mix, samples were heated for 1 min at 95 °C and then incubated for 16 h at 60 °C. Ligation of the annealed oligonucleotide probes was performed for 15 min at 54 °C in buffer containing Ligase-65 enzyme. After inactivating the ligase enzyme by heating at 98 °C for 5 min, the multiplex PCR was carried out using FAM-labelled primers, dNTPs and SALSA polymerase. PCR was performed for 35 cycles of 30 s at 95 °C, 30 s at 60 °C and 1 min at 72 °C. All the reactions were carried out in a thermal cycler equipped with a heat lid (ABI 2720, Applied Biosystems, Foster City, CA, USA). DNA from three healthy controls (previously studied and without significant genomic imbalances for the studied loci) and a negative control (without DNA) were always included in the same reaction. PCR products (1 μL each) were heat denatured and analyzed on a GeneScan ABI PRISM 3130 capillary electrophoresis system (Applied Biosystems, Foster City, CA, USA). The sample results were displayed as a ratio between signals obtained for the reference and experimental samples. For each MLPA probe, we determined specific cut-off values for gains and losses, using the values limiting the 95 % confidence interval (CI) as determined on non-cancer subjects. A numerical gain was scored when the values exceeded 1.2, and a numerical loss was defined when the values were lower than 0.8. Ingenuity Pathways Analysis (IPA; http://www.ingenuity.com/) was used to analyze the major cellular and biologic functions of those genes.

Results

Karyotyping, FISH and aCGH evaluation

The 5637 cell line

This is a hypotriploid cell line with a modal number of 59 chromosomes (Table 1). It possesses only two copies of chromosomes 4, 10, 16, 19, 21, 22 and X and a total loss of the Y chromosome. An extra copy of chromosomes 7 and 20 were detected. Only nine chromosomes were preserved and not implicated in rearrangements in this cell line (1, 7, 14, 16, 18, 19, 20, 22 and X). Most rearrangements found were simple, involving just two chromosomes, except for those concerning chromosome 3 (Fig. 1(a–d), Table 1). One reciprocal translocation was detected, between chromosomes 4 and 8 der(4)(4pter→4q25::8p21→8pter) and der(8)(4qter→4q25::8p21→8qter). Chromosome 8 was also involved in a rearrangement with chromosome 9 with a deletion of 9p (Fig. 1(e)). Chromosome 3 was involved in the most complex rearrangements, giving origin to two derivative chromosomes: one derivative chromosome involving several amplifications of 3p25.3p25.2 (log ratio indicative of approximately 9 copies) and 6p22.3 (log ratio indicative of approximately 17 copies) in a ladder-like manner that had already been described to some extent by others [28, 29] (Table 1, Fig. 2).

Table 1 Comparison between the published literature and the results of the present work for 5637 cell line
Full size table
Fig. 1
figure 1

Some rearrangements identified in the 5637 (a–e), T24 (f–i) and HT1376 (j–m) cell lines. Each rearrangement of the G banding is presented: ideogram with colours between “square brackets” and FISH with probes signals between “parentheses”, and the derivative chromosomes are pointed out with arrows. a Derivative containing chromosomes 3 and 6 [blue and orange] and painting probes for chromosomes 3 (red) and 6 (green). b Derivative containing chromosomes 3, 12, 15 and 21 [dark blue, brown, pink and light blue] and with painting probes for chromosomes 3 (red) and 15 (green). c With subtelomeric probes for 12p (green), 12q (red) and 18q (yellow) and alpha satellite probe for chromosome 18 (aqua). d With subtelomeric probes for 4p (green), 4q (red) and 21q22 region (aqua). e Derivative containing chromosomes 8 and 9 [purple and blue] and painting probes for chromosomes 8 (red) and 9 (green). f Derivative chromosome containing chromosomes 10 and 11 [brown and yellow] and subtelomeric probes for 11p (green), 11q (red) and 18p (yellow) and alpha satellite for chromosome 18 (aqua). g Painting probes for chromosomes 10 (green) and 11 (red). h Derivative involving chromosomes 4/17 [blue/green] and 6/17 [orange/green] and subtelomeric probes for 6p (green), 6q (red) and 17q (yellow) and alpha satellite for chromosome 17 (aqua). i Painting probes for chromosomes 6 (red) and 17 (green). j Derivative containing chromosomes 2 and 12 [purple and brown] and painting probes for chromosomes 2 (red) and 12 (green). k Derivative containing chromosomes 7 and 20 [light red and dark red] and painting probes for chromosomes 7 (red) and 20 (green). l Derivative containing chromosomes 1, 7, 8 and 13 [green, red, purple and orange] and painting probes for chromosomes 1 (red) and 8 (green). m Painting probes for chromosomes 7 (red) and 20 (green)

Full size image
Fig. 2
figure 2

aCGH profile of all chromosomes from the three cell lines (a 5637 cell line, b T24 cell line and c HT1376 cell line). Sample DNA was labelled in red (Cy5) and control DNA in green (Cy3). Gains and losses are represented by deviations to the right (red fluorescence) and left (green fluorescence), respectively

Full size image

The T24 cell line

The modal number of chromosomes was 73, which makes it a hypertriploid cell line (Table 2). Regarding aneuploidies, only the chromosome 4 was lost, whereas extra chromosomes 14, 20 and 22 had already been described by others [6, 27, 29]. Only the chromosomes 2, 3, 21 and 22 exhibited a normal band pattern. No reciprocal translocations were observed. Partial deletions and gains were also confirmed in this study, and it was possible to identify which rearrangements they were involved in. For example, in our study, the loss of 10pter to p11.21 and 10q24 to 10qter [27] corresponded to the der(10)(10p11.21::10p11.21→q25.1::11p15.1→11pter) (Fig. 1(f, g)), with the gain of 10p11.21 confirmed by the aCGH (Fig. 2). The gain of 17q22qter was associated with der(4)(4pter→4q32.1::17q22→17qter) and der(6)(17qter→17q22::6p11.2→6qter) (Fig. 1(h, i)). This cell line was the one in which we were able to detect more gains than losses in band chromosomes.

Table 2 Comparison between published literature and the results of the present work for T24 cell line
Full size table

HT1376 cell line

The cytogenetic characterization and aCGH analysis of the HT1376 cell line was a challenge considering the number of chromosomes and the complexity of the rearrangements found (Table 3, Fig. 2). Additionally, there are few available reports on the HT1376 genomic profile [6, 27]. In this study, the modal number of chromosomes was 96, and if we consider it to be a near-tetraploid cell line, there is a loss of chromosomes 14 and 21 and an extra chromosome 20. All chromosomes, except 14, 18 and 21, were involved in unbalanced structural rearrangements. Regarding certain anomalies, it was not possible to correlate the cytogenetics with aCGH data, such as an extra copy of 5q arm, two extra copies of 5q31qter, loss of a 6q arm, an extra copy of 7pterp11.1, gain of 13q21.33q31.1 and 13q33.1qter. The cell line supplier description had several add chromosomes, such as add(12), that we found in the present study in der(2)(2pter→2q12.3::12q13.11→12qter) (Fig. 1(j)). Likewise, add(7)(p22) was found to be der(7)(20qter→20q11.21::7pter→7qter) (Fig. 1(k)). Chromosome 1 presented the highest number of structural rearrangements such as the one described as der(13)t(13;?)(p11;?)t(?;1)(?;q11) that, in fact, is a derivative of chromosomes 1 and 13 with two more chromosomes (7 and 8) involved (Fig. 1(l, m)). The double min or ring chromosomes were not observed in any of the three cell lines studied. Overall, there was a very high correlation between the results of classical cytogenetics, FISH and aCGH.

Table 3 Comparison between the published literature and the results of the present work for HT1376 cell line
Full size table

Analysis of RB1 and TP53 genes

T24 cell line presented a normal RB1 gene pattern, and the other two cell lines showed losses in this gene. TP53 deletion was identified in the 5637 and HT1376 cell lines. In the T24 cell line, despite the fact that three rearrangements were found where chromosome 17 was involved, it did not show any loss or gain in TP53 (Tables 1, 2 and 3).

MLPA interpretation

A total of 107 genes from all chromosomes (except sexual ones) were analyzed for gains and losses (Fig. 2). It was detected that 16.8 % of the genes studied showed a normal pattern in the three cell lines (Fig. 3). Some gene gains were present in the three cell lines, such as gains on HRAS, EHF, THBD, BCL21 and PTPN1. These genes are all involved in tumourigenesis. All of them have roles in cell proliferation, growth, apoptosis, senescence and migration. Comparing the loss and gain that were common in 5637 and HT1376 cell lines, we found, besides the loss of RB1 and TP53, losses on genes IL1A, TANK, CREM and LRMP, which are all related with tumour development. Gains on the genes RAD17, PTP4A3, LMO2, ING1, DCC and AURKA were also common in both cell lines. These genes are associated with DNA damage response, invasion, differentiation, proliferation and apoptosis, key functions in the tumour initiation and progression. In the 5637 and T24 cell lines, the losses involved the genes ABCG2, NFKB1, BAK1, CDKN1A and CDKN2B, mainly related with proliferation, cell cycle progression and apoptosis. Gains were detected in the CASP2, AXIN2 and CTTN genes. The first two involved in cell apoptosis and proliferation and the last one in cell migration, adhesion and invasion.

Fig. 3
figure 3

Gene gains and losses in the three cell lines detected using the MLPA technique. IL4, IL12B, TNF, MET, RECOL4, PTEN, HIPK3, IL18, CCND2, CD27/TNFRSF7, CDH2, BCL2, DNMT1, CDKN2D and SIM2 are the genes analyzed and not altered

Full size image

Using Ingenuity Pathways Analysis software, the major cellular and biologic functions of these 107 genes were analyzed, including their participation in biological pathways associated with metastization, angiogenesis, differentiation, proliferation, apoptosis, metabolic processes and cell cycle (Fig. 4). In terms of diseases, we evaluated the existence of a relationship with cancer in general and UBC. The 5637 cell line showed more gains in genes associated with cancer and differentiation and losses in metastasis, proliferation, apoptosis, metabolic process and cell cycle. In the case of the T24 cell line, gene gains were associated with cancer and UBC, differentiation, protein binding and metabolic process, and gene losses predominated in apoptosis and cell cycle. Gains and losses in metastasis and proliferation were almost equivalent. In the HT1376 cell line, gene losses were predominant in all cellular processes, being more common in differentiation, proliferation and apoptosis-related genes. The genes associated with these biological processes were selected for discussion.

Fig. 4
figure 4

Relative frequencies of losses and gains of genes involved in several biological processes related to cancer for the three cell lines. Each bar represents the number of genes associated with a given biological function divided by the total number of genes with gains or losses

Full size image

Discussion

The aim of this study was to identify the genomic imbalances and their biological significance in three human UBC cell lines in order to validate it as elective models to test new therapeutic approaches. To be representative models, these cell lines should bear the most commonly found alterations in human UBC.

We thoroughly characterized, for each cell line, the chromosomal modal number in 50 metaphases and the structural chromosomal anomalies consistently observed in at least 25 metaphases. It is well established that high-grade tumours show an increased number of genomic alterations, chromosomal instability and complex karyotypes that we effectively found, but the scope of our work was to characterize the cytogenetic alterations that were consistent and representative. The results obtained showed that the 5637, T24 and HT1376 cell lines maintained their previously described genomic profile [5–7, 27–30], and an improvement in their cytogenetic and genomic characterization was accomplished.

The different gene distributions with copy number variations of the three cell lines in the various biological processes were analyzed, and in general, all the reported genes were related to cellular functions such as cell cycle, metabolic process, apoptosis, proliferation, differentiation, angiogenesis and metastasis, all hallmarks of malignant transformation and cancer development [33, 34].

Correlation between the results of 5637, T24 and HT1376 cell lines with previous works

5637 cell line

Concerning the 5637 cell line, by comparing the chromosomal alterations described herein with those described in other works [5, 6, 27–29], we confirmed a p arm deletion of chromosomes 8 and 17 as well as several gains, like the ones in 5p and 17q. The structural rearrangement der(15)t(15;17) [5, 28, 29] was not found in the present study, and some anomalies were now identified. For example, instead of i(8)(q10) [27], der(8;9)(8qter→8q11::9q11→9qter) was observed, and we further characterize the der(3)t(3;21) [5, 28, 29] as also containing chromosomes 12 and 15. Comparing array data already published with that from this present analysis, more imbalances were detected for this cell line and allowed us to confirm and redefine the amplifications of 6p22.3 and 3p25.3p25.2 [6, 35]. The band losses in chromosome 13 were similar to those found by Hurst et al. [6], but we also had band gains (13q12.2q12.3, 13q14.11q14.13, 13q21.32qter). We observed a gain of 13q21.32pter, a region that includes the gene found in high copy number by MLPA, ABCC4. The other similarity as far as losses are concerned refers to the one detected by Padilla-Nash et al. [28] in del(2)(q31) that we found in 2q12.2q33.3. The gain of 17q is confirmed by MLPA by the increased copy number of LHX1 (17q12).

T24 cell line

Regarding the T24 cell line, we redefined the described del(5), der(15) and der(8) [29, 31, 32] (Table 1), while the rearrangements of der(3)t(1;3) [29], t(8;13) [29, 32], i(10)(p10) [7], i(10)(q10) [7], t(3;16)(p21;p13.3) [32] and t(8;17)(q21;q22) [32] were not detected in the present study. Some losses have been already reported [27] and confirmed in this work by aCGH, such as 1q21qter, 6p21.32p12.1 and Xp21.1qter; a few gains were also in agreement with the already published data such as 17q22qter [27]. Other gains and losses were in bands near those described [27], like those concerning chromosome 7 (gains in q34q36.1 and q36.2qter). On chromosome 16, we observed a gain in 16q11q23.1 that is supported by the increase in copy number of CDH1 (16q22.1) by MLPA technique.

HT1376 cell line

There are few available reports on the HT1376 genomic profile. Some gains and losses have been described with poor cytogenetic characterization and without correlation to aCGH analysis [6, 27, 30]. In this cell line, our results confirmed the observation of amplification in 6p22.3 [6, 35] and reported a new one at 16p13.13p13.12. The described loss 1q21.3q21.2 [6] was different from that we found (1pterp35.2;1p21.1p11.2), but the gains in the p arm of chromosome 1(1p22.2p21.3 (x2)) were similar and confirmed by the augmented copy number of F3 (1p21.3) and BCAR3 (1p22.1). Several losses were also similar (10q25.1qter) or equal with other studies, such as the 12p12.3pter, 13q13.2q14.3 and 22q13.2qter [6, 27]. A gain in 5p and 7p21.2p21.1 was found in this study and in those of Hurst and collaborators [6]. Gain on q arm of chromosome 8 has been described by Hurst and collaborators [6] and confirmed by us, but we observed also an 8q23.3qter and PTP4A3 (8q24.3) gain.

Correlation between genomic, gene and cellular functions altered in 5637, T24 and HT1376 cell lines and in UBC

Overall, the observed rearrangements are in agreement with those described in human UBC. A major concern with tumour cell lines is whether they retain an entity-specific pattern of chromosomal imbalance [27]. As shown by cytogenetics and aCGH, the established cell lines analyzed here present many of the changes observed in primary tumours.

The chromosome 20 aneuploidy, associated with primary UBC [5, 36], was identified in all three cell lines as well as the commonly loss of almost the genomic region of chromosome 4 [3, 37]. The deletion of 8p is an early change observed during tumour progression in vivo [38], and the three cell lines exhibited a loss of 8p. The most common anomaly in UBC is the loss of chromosome 9, particularly the loss of 9p21, which was also detected in all cell lines studied. Conversely, the deletion of 14q locus described in primary tumours [3, 5] was observed by us only in the HT1376 cell line, with loss of NFKBIA (14q13.2).

Concerning cytogenetic alterations seen in advanced stages of UBC patients, the gain of chromosome 7 is another common finding, and it is associated with a more aggressive biological behaviour [3, 39]. Surprisingly, the non-muscle-invasive cell line was the only one with an extra copy of this chromosome; in the T24 cell line, an extra segment (7q33qter) was present, and in the HT1376 cell line, a 7p21.2p21.1 amplification was found. These results are in accordance with previous ones for the 5637 [6, 28, 29], T24 [6] and HT1376 [6] cell lines. Rearrangements involving chromosome 3 are also described in high-grade UBC, and they were present in the 5637 and HT1376 cell lines. Also, alterations in the chromosome 10 were present in the three cell lines making a relation with high-grade tumours [13, 40]. Gains in 5p have been found mainly in aggressive tumours and associated with UBC cell lines [2, 41]. We have found several gains, such as 5p, in all analyzed cell lines, and it was also possible to localize which chromosomal rearrangements they were in.

Tumour suppressors and tumour progression

When RB1 and TP53 genes are mutated, a worst prognosis is expected with higher recurrences and/or rapid progression [42]. An altered RB1 tumour suppressor gene (13q14.2) is frequently seen in high-stage, invasive UBC and associated with prognostic value, namely, a short-time patient survival [43]. The RB1 locus was shown to be loss in the 5637 and HT1376 cell lines by FISH and by MLPA (Fig. 3). RB1 is described as mutated in homozigoty in these two cell lines [44], and so, this loss of heterozigoty is expected to be associated with a complete loss of function of this well-known tumour suppressor gene. Sanchez-Carbayo and collaborators [45] described undetectable pRB levels in the 5637 and HT1376 cell lines. In the T24 cell line, we did not find any loss of this gene, as it is described as non-mutated [44] and the protein detected at normal levels [45]. This protein is a central player in cell cycle control as well as an integrator of the TGF-beta signalling for cell cycle repression [46]. The loss of 17p and, consequently, of the tumour suppressor TP53 (17p13.1) is a late event in UBC progression and usually associated with a worse prognosis for the patient [42]. The TP53 functions essentially as a transcription factor, protecting cells from several cellular stresses including DNA damage and oncogene activation by inducing cell cycle arrest, senescence, apoptosis, DNA repair or autophagy [47]. In the present study, loss of TP53 locus was observed in the 5637 and HT1376 cell lines by cytogenetic and specific gene FISH analysis, which is in accordance with loss of function associated with a mutation in the remaining allele. In fact, mutation in this gene is described in homozigoty in both cell lines [34, 44, 45]. In the T24 cell line, despite the fact that three rearrangements were found where chromosome 17 was involved, only a gain of 17q22qter was observed by aCGH, and also, we were unable to detect any loss or gain in TP53 by the FISH and MLPA techniques. However, it is described as a homozygous mutation on TP53 in this cell line that abolishes its function [44]. This suggests that the multiple rearrangements involving chromosome 17 could led to the loss of the wild-type allele and to duplications of the regions containing the mutant one, making it undetectable by the approaches we used.

The highly frequent loss of 9p21.3 in UBC, which we found in all cell lines studied, is usually associated in the literature with the loss of CDKN2A gene, known as p16, and CDKN2B gene, known as p14, both key cell cycle inhibitors [47]. These proteins are involved in the regulation of proliferation, apoptosis, cell cycle progression, growth and G1 phase, and its loss has been associated with several malignancies, including UBC, leukaemia and melanoma, among others [48]. In this work, we confirmed a loss of the CDKN2A in the three cell lines as well as a loss of CDKN2B in the 5637 and HT1376 cell lines. In 2010, Forbes and collaborators [44] described that these two genes (CDKN2A and CDKN2B) were found as non-mutated in these cell lines, which suggests that there was not a loss of function of these tumour suppressors genes and interrogates the role of these genes loss in tumour progression. Nevertheless, there is another gene located in 9p21.3, which is deleted in the three cell lines, the miR-31, a microRNA recently described as a suppressor of breast cancer metastases [49]. In whole-genome profiling of miR-31 transcriptional effects identified as its targets, factors orchestrating DNA repair, replication and cell cycle progression and reintroduction of miR-31 suppressed the cell cycle and inhibited expression of factors involved in DNA replication and cell cycle [50].

An entire chromosome 4 is lost in the 5637 and T24 cell lines, and loss of 4p was seen in the HT1376 cell line. Tatarano and collaborators [37] described in this region the miRNA-218 (4p15.31) as specifically down-regulated on UBC and showed that the transfection of this miRNA results in a decrease in cell proliferation, migration and invasion activity and an increase of apoptosis, suggesting miR-218 as a tumour-suppressive miRNA in UBC.

The loss of 8p was seen in the three cell lines studied. In this region, the DEFB1 (8p23.1) and MTSU1 (8p22) were found to be associated with UBC [3, 51].

Proliferation, oncogenes and tumour progression

We observed a gain of 5p in all the cell lines. However, in the T24 cell line, it was confined to 5p15.31p14.2. In this region, CTNND2 (5p15.2), the gene for delta-catenin, frequently over-expressed in cancer [52] and specifically over-expressed in a cell model of experimentally induced UBC [53] was located. Delta-catenin acts in a synergistic manner with beta-catenin to activate MMP7 (11q22.3), a matrix metalloproteinase correlated with tumour progression and metastasis. Accordingly, MMP7 is frequently up-regulated and augmented in the serum of UBC patients [54]. We observed a gain of the MMP7 locus on the 5637 and HT1376 cell lines.

HRAS (11p15.5) gene plays a pivotal role in urothelial tumourigenesis, being frequently activated in UBC [14, 18]. Activating mutations in this gene lead to activation of MAPK and PI3K pathways, promoting cell survival and angiogenesis [55]. Here, we found a moderate high copy number of this gene in all cell lines studied, and observed a gain of 11pterp15.1 (x2) only in the T24 cell line. It is already described that HRAS is activated by mutation in the T24 cell line (HRAS p.G12V (c.35G → T)). However, we suggest that the state of HRAS is not sufficient to explain the tumourigenesis of the 5637 and HT1376 cell lines.

Hurst and collaborators [6] detected a gain of 6p22 in the 5637 and HT1376 cell lines that was corroborated by Blaveri and collaborators [56] in UBC patients and confirmed in this work. We found it amplified in the 5637 cell line with a log ratio indicative of approximately 17 copies and in the HT1376 cell line with a log ratio indicative of approximately 12 copies and not in the T24 cell line, which is in agreement with the study of Chekaluk [35]. This genetic location is frequently amplified in advanced stage UBC and targets several genes [57]. Expression of E2F3, DEK and ID4 was shown to be augmented [57], and over-expression of E2F3 was shown to be a key factor on urinary bladder tumourigenesis, being a transcription factor that promotes cell cycle progression [15].

In addition, we found an amplification of the region 3p25.3p25.2 (log ratio indicative of approximately nine copies) in the 5637 cell line. An amplification of this same region was also observed in colorectal cancer [58] and involves the amplification of RAF1 (3p25.2), known as proto-oncogene c_RAF involved in MAPK signalling pathway and ERBB signalling pathway [59].

The gain of chromosome 20 in all cell lines studied is also usually observed in UBC. It was confirmed by MLPA, with the increase in copy number of AURKA (20q13), a cell cycle check point protein frequently over-expressed in cancers and PTPN1 (20q13.1-q13.2), which is associated with oncogenic transformation [54]. Also in chromosome 20 is the locus of MMP-9 (20q12-q13), a matrix metalloproteinase which plays an essential role in local proteolysis of the extracellular matrix, in leukocyte migration, and also as a pro-angiogenic factor, shown to be over-expressed in high-grade urothelial carcinomas.

By MLPA, we detected amplification of CTTN (11q13.3), cortactin, in the 5637 and T24 cell lines. This gene is over-expressed in cancers and contributes to tumour cell invasion and metastasis [54].

Multidrug resistance

There is another gene that localizes to the amplified region, 6p22.3, which is TPMT enrolled in the metabolic inactivation of thiopurine drugs [60]. This gene was found to be amplified also in a number of other tumour types [61] and may be involved in chemoresistance of the tumour [60].

A gain of 13q21.32pter was seen in the 5637 cell line comproved by the finding of a high copy number by MLPA of ABCC4 (13q31), an ABC transporter also known as multidrug resistance gene [54]. Curiously, the HT1376 cell line also showed this region with a cytogenetic gain, but MLPA technique failed to detect a higher copy number of ABCC4.

Apoptosis inhibition

The gene coding for p21, CDKN1A (6p21.1), was shown to be down-represented by the MLPA technique loci in all the cell lines studied, an observation confirmed by the cytogenetic results. The p21 is a potent cyclin-dependent kinase inhibitor, a protein that binds to and inhibits the activity of cyclin-CDK2 or cyclin-CDK1 complexes and thus functions as a regulator of cell cycle progression at G1 phase [60]. The expression of this gene is tightly controlled by the tumour suppressor protein p53, through which this protein mediates the p53-dependent cell cycle G1 phase arrest in response to a variety of stress stimuli [62]. On a nearby region, BAK (6p21.31), coding for a pro-apoptotic protein, is also down-represented. Although the loss of these regions does not really means the loss of these tumour suppressors, it is possible that its reduced expression would be related with defects in the regulation of apoptosis [60].

We found in the 5637 cell line a gain in 11q14.1q22.3 and increased copy number of BIRC2 (11q22.2), an apoptosis inhibitor [54]. Very recently, Song and collaborators [63] described an association between other genes on this region, miR-708 (11q14.1) with the tumourigenicity of the 5637 cell line. They also showed that caspase-2 is one target of this miRNA and that silencing of miR-708 could promote the T24 and 5637 cells to apoptosis. Remarkably, the miR-708 loci (11q14.1) is gained in all the cell lines that we studied, which suggest that could be considered as a good candidate for a common event in UBC development and when associated with apoptosis inhibition.

Activation of PI3K-AKT pathway

The PI3K-AKT pathway is commonly activated in tumours, and it is also true for UBC [64]. PTEN, FOXO1, MAPK1 and PDK1 were found to be down-regulated, while AKT1, AKT2, HRAS and RAC1 were up-regulated in an experimental model of UBC [64]. In our cell lines, we found a loss of the region where PTEN gene localizes (10q23.31) by cytogenetics in the 5637 and HT1376 (not confirmed by the MLPA technique) and a duplication of chromosome 14 in the T24 cell line, where it included the loci of AKT1 (14q32.33). We also observed a gain of chromosome 7 in the 5637 cell line, where the EGFR gene (7p11.2), an upstream activator of this pathway, in response to epidermal growth factor [64] is localized.

Additional alterations in the 5637 cell line

We found a gain of 17q in the 5637 cell line, confirmed by MLPA, by the increased copy number of LHX1 (17q12), a LIM homeobox transcription regulator involved in urogenital development, and of ERBB2 (17q11.2-q12) also known as HER2, another oncogene involved in the metastasis process and shown to be associated to aggressiveness on other cancers, such as breast cancer [65]. The C-ERBB2 (HER2/Neu) is amplified and over-expressed in 10–26 % of UBC and is associated with a poor prognosis [66].

Additional alterations in the T24 cell line

We found a gain of 10p11.21 in the T24 cell line. A gain of 10p11.21 was previously related with breast cancer recurrence [65]. In this region, we found by MLPA an increase in copy number of CREM (10p11.21), a bZIP transcription factor that binds to the cAMP responsive element found in many viral and cellular promoters. CREM expression is seen augmented in cancers [67] and it seems to be associated with IL2 expression by epigenetic de-repression of its promoter [67]. Also in the T24 cell line we described a gain of 11p15.1q14.1 region which includes the genes found over-represented by MLPA, FGF3 (11q13.3), a proto oncogene frequently activated in cancer; CCND1 (11q13.3) cyclin D1, a promoter of cell cycle progression and RELA (11q13.1) the P65 subunit of NF-kB transcription factor [67]. Another gene associated with tumour progression in primary UBC [14], the TNFRSF1B (1p36.2), locates to a gained region on the T24 cell line. MIR29C (1q32.2) was found significantly under-expressed in tumours that progressed and proposed to stratify patients with T1 disease into risk groups [68]. In the T24 cell line there is a loss of 1p21qter involving this gene.

Additional alterations in the HT1376 cell line

In the HT1376 cell line, we found much more genomic alterations that were only seen in this cell line, such as the genomic gain of 1p22.2p21.3 (x2) correlated with an augmented copy number of F3 (TF) (1p22-p21) and of BCAR3 (1p22.1), an extracellular matrix (ECM) adhesion factor. The elevated expression of TF on the surface of tumour cells can trigger coagulation and platelet activation, resulting in the formation of a platelet coat that protects tumour cells travelling on the blood vessels and helps in the extravasation to new locations. There is a tight correlation between elevated TF expression and poor prognosis in multiple tumour types. Interestingly, it was found that the microRNA MIR19A (13q31.3) that targets TF to inhibit migration, invasion and metastization, and its locus is located on a cytogenetic gain we found in the 5637 cell line, which could help to explain the less aggressive properties of the 5637 when compared to the HT1376 cell line.

Also in the HT1376 cell line, we found an amplification of 6p22.1 p21.33 (x3) and high copy number of HLAF (6p21.3); MDC1 (6p21.3), from DNA repair pathways; and IER3 (6p21.3) that protects cells from TNF-alpha-induced apoptosis [54]. A gain of 1p34.2 was related with the augment of CTPS1 (1p34.1), an enzyme involved in the biosynthesis of phospholipids and nucleic acids, which is proposed to play a key role in cell growth, development and tumourigenesis [69].

Additional genomic alterations closely related with altered copy number of genes assessed by the MLPA technique were found in the HT1376 cell line, which are also associated with different steps of tumour progression and metastasis capability. These are the loss of NFKBIA (14q13.2), the NF-kB inhibitor, the gain of 8q23.3qter and PTP4A3 (8q24.3), a gene associated with cell growth; the gain of 16p12.2 p11.2 and MVP (16p11.2), a multidrug resistance gene and also a prognostic marker for cancer; and the loss of 19p12qter and BAX (19q13.3-q13.4), a pro-apoptotic protein p53-induced.

Two main genomic types of UBC

Lindgren et al. [15] integrated genomic and gene expression profiling, in combination with gene mutation and protein expression data to 146 cases of UBC and identified two main genomic pathways operating in this cancer. Central alterations within these pathways were alterations of FGFR3/CCND1 and E2F3/RB1, respectively. The authors also suggest that alterations in RAS or RAF1, which are found near mutually exclusive to FGFR3, could define a third pathway, an alternative to the former one.

In our work, we can categorize the T24 cell line as belonging to the alternative pathway of FGFR3/CCND1 by presenting a mutated HRAS and over-represented CCND1, by the gain of 11q13 band in the 11p15.1q14 gain. The 5637 and HT1376 cell lines can be categorized as belonging to the E2F3/RB1 circuit due to amplification of 6p22.3, concomitant with a loss of one copy of RB1 on 13q14.2 and mutation of the remaining copy.

Additionally, the HT1376 cell line presented 10q deletion of PTEN region (10q23.31) and no alteration of PIK3CA region which, in combination with the inactivation of TP53, as also seen, leads to invasive tumour growth. In contrast, the 5637 cell line does not present any loss of PTEN region nor gain of PIK3CA region; in fact, it bears a loss of this late region, which could indicate a less-invasive phenotype [15]. Another group of tumours found by Lindgren et al. [15], which is very rare and with the worst prognosis, is characterized uniquely by enrichment of CCND1 amplifications, which we did not found in our work.

In summary, the cell lines that we characterized cover the more frequent subtypes of UBC; the T24 represents a FGFR3/CCND1 subtype, while the 5637 and HT1376 represent the E2F3/RB1 circuit mutational profile, with the former representing a less aggressive phenotype and the later bearing more invasive and metastatic properties. We conclude that these three cell lines can be used, as a group, to assay new methodologies and drug-therapeutic approaches to treat patients with UBC, because they represent the more frequent subtypes of this kind of tumour.