Advertisement
Research Article

A Systematic Study of Gene Mutations in Urothelial Carcinoma; Inactivating Mutations in TSC2 and PIK3R1

  • Gottfrid Sjödahl,

    Affiliation: Department of Clinical Sciences, Oncology, Lund University, Skåne University Hospital, Lund, Sweden

    X
  • Martin Lauss,

    Affiliation: Department of Clinical Sciences, Oncology, Lund University, Skåne University Hospital, Lund, Sweden

    X
  • Sigurdur Gudjonsson,

    Affiliation: Department of Clinical Sciences, Urology, Skåne University Hospital, Malmö, Sweden

    X
  • Fredrik Liedberg,

    Affiliations: Department of Clinical Sciences, Urology, Skåne University Hospital, Malmö, Sweden, Section of Urology, Växjö County Hospital, Växjö, Sweden

    X
  • Christer Halldén,

    Affiliation: Department of Laboratory Medicine, Clinical Chemistry, Lund University, Skåne University Hospital, Malmö, Sweden

    X
  • Gunilla Chebil,

    Affiliation: Department of Pathology, Helsingborg Hospital, Helsingborg, Sweden

    X
  • Wiking Månsson,

    Affiliation: Department of Clinical Sciences, Urology, Skåne University Hospital, Malmö, Sweden

    X
  • Mattias Höglund mail,

    mattias.hoglund@med.lu.se

    Affiliation: Department of Clinical Sciences, Oncology, Lund University, Skåne University Hospital, Lund, Sweden

    X
  • David Lindgren

    Affiliations: Department of Clinical Sciences, Oncology, Lund University, Skåne University Hospital, Lund, Sweden, Center for Molecular Pathology, Department of Laboratory Medicine, Lund University, Skåne University Hospital, Malmö, Sweden

    X
  • Published: April 14, 2011
  • DOI: 10.1371/journal.pone.0018583

Abstract

Background

Urothelial carcinoma (UC) is characterized by frequent gene mutations of which activating mutations in FGFR3 are the most frequent. Several downstream targets of FGFR3 are also mutated in UC, e.g., PIK3CA, AKT1, and RAS. Most mutation studies of UCs have been focused on single or a few genes at the time or been performed on small sample series. This has limited the possibility to investigate co-occurrence of mutations.

Methodology/Principal Findings

We performed mutation analyses of 16 genes, FGFR3, PIK3CA, PIK3R1 PTEN, AKT1, KRAS, HRAS, NRAS, BRAF, ARAF, RAF1, TSC1, TSC2, APC, CTNNB1, and TP53, in 145 cases of UC. We show that FGFR3 and PIK3CA mutations are positively associated. In addition, we identified PIK3R1 as a target for mutations. We demonstrate a negative association at borderline significance between FGFR3 and RAS mutations, and show that these mutations are not strictly mutually exclusive. We show that mutations in BRAF, ARAF, RAF1 rarely occurs in UC. Our data emphasize the possible importance of APC signaling as 6% of the investigated tumors either showed inactivating APC or activating CTNNB1 mutations. TSC1, as well as TSC2, that constitute the mTOR regulatory tuberous sclerosis complex were found to be mutated at a combined frequency of 15%.

Conclusions/Significance

Our data demonstrate a significant association between FGFR3 and PIK3CA mutations in UC. Moreover, the identification of mutations in PIK3R1 further emphasizes the importance of the PI3-kinase pathway in UC. The presence of TSC2 mutations, in addition to TSC1 mutations, underlines the involvement of mTOR signaling in UC.

Introduction

Urothelial carcinoma (UC) of the bladder, the most common type of bladder cancer, is characterized by several gene mutations of which the most frequent is activating mutations in the FGFR3 receptor protein. FGFR3 mutations show a biased distribution of UC pathological subtypes in which low grade non-invasive tumors (Ta) shows the highest frequencies, close to 70%, whereas high grade and muscle invasive tumors (≥T2) show considerably lower frequencies, in the range of 10 to 15%. On the other hand, TP53 mutations show an opposite pattern with high frequencies in muscle invasive tumors (~50%) and low in non muscle invasive tumors (~15%). This biased distribution of FGFR3 and TP53 mutations has led to the hypothesis that UC develops along two different pathways, a FGFR3 and a TP53 pathway, respectively [1], [2]. Alternative explanations for the frequent TP53 mutations in muscle invasive tumors have however been put forward [3]. Several genes acting downstream of receptor tyrosine kinases (RTKs) have also been reported to be mutated in UC, e.g. PIK3CA [4], members of the RAS family [5], BRAF [6], and AKT1 [7]. Another characteristic of UC is the frequent LOH on chromosome 9 and this chromosome is believed to harbor more than one tumor suppressor gene of importance for UC development. At least two such loci have been established, CDKN2A and TSC1. CDKN2A shows homozygous deletions in up to 30% of UC [8] whereas inactivating sequence mutations is seen to a lesser extent. TSC1, on the other hand, has been reported to be mutated in 16% of UC [4]. TSC1 is a negative regulator of the mTOR pathway, which is important for cell proliferation and frequently found activated in tumors [9] including UC [10], [11]. Notably, TSC1 is regulated by AKT1 and is therefore a potential downstream target of the FGFR3 signaling pathway. Additional proteins in this pathway include PIK3R1, PTEN and TSC2. PIK3R1 is a negative regulator of PIK3CA while PTEN is a negative regulator of AKT1. TSC2 forms a complex with TSC1 that functions as a negative regulator of the mTOR pathway. So far no mutation data on PIK3R1 or TSC2 in UC is available. Recent reports have also implicated the APC/CTNNB1 pathway in UC [12], [13], [14]. In the present investigation we aimed to further characterize the mutational landscape of UC. In a series of 145 tumors we performed mutation analyses of 16 genes, FGFR3, PIK3CA, PIK3R1, PTEN, AKT1, KRAS, HRAS, NRAS, BRAF, ARAF, RAF1, TSC1, TSC2, APC, CTNNB1, and TP53.

Methods

Tumors and isolation of nucleic acids

Urothelial tumors were collected by cold-cup biopsies from the exophytic part of the bladder tumor from 145 patients undergoing transurethral resection at the University Hospital of Lund, Sweden, between 2001 and 2005. For detailed patient information see Table S1. To increase the statistical power, the 145 series of samples were extended with 73 samples analyzed for FGFR3, PIK3CA, TP53, HRAS, KRAS, and NRAS mutations only (Table S1). Tumor pathology, including transurethral and cystectomy specimens, were reviewed by one pathologist (GC). Written informed consent was obtained from all patients and the study was approved by the Local Ethical Committee of Lund University. Genomic DNA was extracted using the DNeasy Tissue kit (Qiagen). Genomic DNA was amplified using the Illustra GenomiPhi V2 DNA Amplification Kit (GE Healthcare) before further processing.

Mutation analysis

Coding regions in FGFR3, PIK3CA, PIK3R1, PTEN, AKT1, KRAS, HRAS, NRAS, BRAF, ARAF, RAF1, TSC1, TSC2, APC, CTNNB1, and TP53 were selected and PCR-amplified using oligonucleotide primers (Table S2). All reactions were carried out in 96-well plates in a 40 ul mixture containing PCR buffer, 1.5 mM MgCl2, 0.2 mM DNTP, 0.5 µM each of the forward and reverse primers and 1 U TrueStart Taq polymerase (Fermentas, Helsingborg, Sweden) or Platinum Taq polymerase (Invitrogen, Carlsbad, CA). The reactions were heated to 94°C for 5 min, and subjected to 38–41 amplification cycles, followed by a final elongation step of 10 min at 72°C. Each cycle consisted of a denaturation step of 30 seconds at 94°C, an annealing step of 30 seconds, and an elongation step of 1 min at 72°C. The different annealing temperatures used in the different PCRs, as well as any deviations from the standard reaction mixture are listed in Table S2. All PCR-amplifications were carried out in a MBS Satellite Thermal Cycler (Thermo Scientific, Waltham, MA). Three ml of each PCR product were run on a precast 2% agarose gel (E-Gel 96 2% agarose GP, Invitrogen) and the remaining volume was purified on AcroPrep 96 filter plate Omega 10K (Pall, Ann Arbor, MI), according to the manufacturers protocol. The purified PCR products were then sequenced using the BigDye terminator v1.1, or 3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA) on a 3130×l Genetic Analyzer (Applied Biosystems). Sequence traces were analyzed using the SeqScape v2.5 software (Applied Biosystems), and all sequence variations were validated by re-sequencing independent PCR products. A change in the DNA sequence was considered to be a mutation when it changed the amino acid sequence of the encoded protein or affected a known splice acceptor or donor site. Sequence changes reported as single nucleotide polymorphisms (SNPs) according to Database of Single Nucleotide Polymorphisms (dbSNP) at NCBI, were omitted from the analysis. All gene mutations not previously described were validated by sequencing of blood samples obtained from the same patient when available (Table S3). Previously not described mutations were analyzed using the PolyPhen 2 predictor tool [15].

Statistical analysis

Chi-2 analyses were used to establish significant differences in proportions between groups. A hypergeometric test was used to determine if the observed numbers of double mutations was significantly different from what is expected from a resample distribution assuming independence.

Results

We screened 145 UCs for sequence mutations in a total of 16 genes, FGFR3, PIK3CA, PIK3R1, PTEN, AKT1, KRAS, HRAS, NRAS, BRAF, ARAF, RAF1, TSC1, TSC2, APC, CTNNB1, and TP53. In Figure 1 the results are summarized and the cases are grouped according to tumor grade. The same data organized according to tumor stage is provided as Figure S1. Mutation frequencies are also summarized in Figure 2, along with possible interactions between the investigated genes. Detailed information on all identified mutations is given in Table S3.

thumbnail

Figure 1. Distribution of identified mutations.

In A) low grade (G1/G2) tumors and in B), high grade (G3) tumors. Red squares indicate inactivating mutation. Green squares indicate activating mutation. For PIK3CA, dark green squares indicate kinase domain mutations and light green helical domain mutations. At the right, mutation frequencies are given for each gene in the respective tumor grades.

doi:10.1371/journal.pone.0018583.g001
thumbnail

Figure 2. Schematic representation of relationships and mutation frequencies among the investigated genes.

Arrowheads, positive regulation; filled circles, negative regulation; gene names in green, genes showing activating mutations; gene names in red, genes showing inactivating mutations; gene names in black, genes with no detected mutations in the present investigation; gene names in gray, genes not investigated. TP53 is not included in the graph. The mutation frequencies (%) given are based on the 145 samples investigated for each gene.

doi:10.1371/journal.pone.0018583.g002

For FGFR3 we sequenced exons 7, 10, and 15. Mutations in these exons correspond to 96% of the FGFR3 mutations seen in tumors of the urinary tract according to the COSMIC database (www.sanger.ac.uk/genetics/CGP/cosmic). We detected mutations at 7 different amino acid positions, all previously described. Mutations were seen in 65% of low grade (G1/G2) and in 22% of high grade tumors (Table 1), and in 63%, 39%, and 22% of Ta, T1, and ≥T2, respectively (Figure S1). As expected, the FGFR3 mutation frequency was significantly higher in low grade tumors (p<0.0001, Chi2 test,) and in non-muscle invasive (NMI) compared to muscle invasive (MI) cases, 55% and 22% respectively, (p<0.0002). These results are in line with previous investigations [16], [17], [18] and clearly show that FGFR3 mutations are associated with low grade low stage tumors and with NMI tumors in particular.

thumbnail

Table 1. Mutation frequencies.

doi:10.1371/journal.pone.0018583.t001

We next performed mutation analysis of TP53 and a total of 52 TP53 mutations (36%) were detected. As expected, we found a significant difference in frequency between high grade (51%) and low grade tumors (19%) (p<0.0001). We then tested for possible negative or positive associations between TP53 and FGFR3 mutations. To increase the power we included mutation data for an additional 73 cases increasing the number of cases from 145 to 218 and a negative association was observed (p = 0.0085, hypergeometric test). This association is however lost when G1/G2 and G3 cases are tested separately, p = 0.39 and p = 0.43, respectively.

PIK3CA was screened for mutations in exons 9 and 20 that contain the hotspot positions in which close to all activating mutations occur [19]. A total of 37 (17%) mutated cases were detected in the extended series of tumors (n = 218). A significantly higher proportion of PIK3CA mutations was seen in Ta cases compared to T1 (p<0.05, Chi-2 test), but not between T1 and ≥T2, or between NMI and MI cases. PIK3CA mutations was also associated with low grade (p<0.01). The data also indicated a possible association between FGFR3 and PIK3CA mutations, with 23 detected double mutations and 14 expected. To further investigate this we added data for 92 UCs previously published by Platt et al. [4], 87 published by Lopez-Knowles et al. [20], and 257 by Kompier et al. [18]. In this combined dataset (n = 654) a significant association between FGFR3 and PIK3CA mutations could be established (hypergeometric test, p<2×10−7, 95 observed double mutants and 68 expected). We then tested for possible negative or positive associations between PIK3CA and TP53 mutations but no significant association was observed (p = 0.066, hypergeometric test, n = 218). We also performed mutation analyses of two modulators of PIK3CA activity, PTEN and PIK3R1. All exons of PTEN were sequenced in the 145 tumors but no mutations were identified. For PIK3R1 we sequenced exons 12, 14, and 15 and found one case with mutation.

We next sequenced the exons covering codons 12, 13, and 61 in HRAS, KRAS, and NRAS and detected a total of six mutations in HRAS, four in KRAS, and none in NRAS (n = 218). The overall frequency of RAS mutations was 5%, which is somewhat lower than what has been reported in recent investigations using DNA sequencing [4], [18], [21]. We further tested for a possible negative association between RAS and FGFR3 mutations but only found borderline statistical support for such an association (p = 0.059, hypergeometric test, n = 218).

BRAF has been shown to be activated by point mutations in several different tumor types [22] with the most frequent mutations located in exons 11 and 15. BRAF belongs to a gene family that also includes ARAF and RAF1, which mediates signals from RAS to downstream targets. As RAF mutations, in analogy with RAS mutations, may show tumor type specificity, we sequenced exons 11 and 15 in BRAF and the equivalent exons in ARAF (exons 10 and 13) and RAF1 (exons 11 and 14). In addition, RAF1 exon 7 was sequenced since activating mutations in this exon has been described in the Noonan Syndrome [23]. The equivalent exon was also sequenced in ARAF. No mutations were however detected in any of the RAF genes.

AKT1 is a major downstream target of PIK3CA and is known to play a key role in the regulation of cell cycle progression, survival, and mTOR signaling. AKT1 was recently shown to have activating mutations in exon 4 in UC [7]. We sequenced the complete exon 4 of AKT1 in our series of 145 tumors and found 2 tumors with an E17K mutation (1%) (Figure 1), associated with increased and constitutive kinase activity of AKT1 under conditions of growth factor withdrawal [7].

We screened all 21 coding exons of TSC1 and detected a total of 17 mutations (13%) including truncating, missense, and splice site mutations. In two additional tumors a substitution of asparagine to serine was seen at amino acid position 762. Both changes were however also detected in blood DNA from the respective patients and hence were considered to be naturally occurring polymorphisms. There was no difference in frequency between low and high grade tumors and TSC1 mutations were seen in FGFR3, PIK3CA, and in TP53 mutated cases. TSC1 functions together with TSC2 as an inhibitor of mTOR by maintaining the mTOR activator RHEB in an inactive state. We therefore sequenced all coding exons of TSC2 and found a total of 5 mutations in 145 samples (3%). The possible impacts of these mutations on protein function were investigated by the PolyPhen software. The software predicted a damaging effect for all four missense mutations, the fifth being a 1 bp frame shift deletion. TSC2 mutations were seen in both high and low grade tumors and none of the TSC2 mutated cases showed concomitant TSC1 mutations. We also tested for possible associations between TSC1 mutations, or TSC1 and TSC2 combined, and mutations in FGFR3 or TP53. No such association was however found (p>0.25 in all comparisons, hypergeometric test, n = 145).

We sequenced exon 16 in APC, known to harbor the majority of APC mutations [24], and exon 3 of CTNNB1 covering the phosphorylation sites that control ubiquitination and degradation of CTNNB1. We detected 6 APC mutations in 145 cases (4%), with no difference between low and high grade tumors. CTNNB1 mutations were seen in 3 cases (2%). We found APC/CTNNB1 mutations in both FGFR3 and TP53 wild type and mutated cases, indicating that activation of the APC/CTNNB1 signaling pathway occur independent of FGFR3 and TP53 mutations. All detected APC mutations were missense mutations. This is in line with Kastritis et al. [12], who also noted an underrepresentation of truncating mutations compared to the mutation spectrum in colorectal cancer. In fact the difference is highly significant. Our data combined with the data of Kastritis et al. show a significantly lower frequency of truncating mutations when compared with data obtained for colorectal cancer from the COSMIC database (p<0.0001, Chi-2).

Discussion

The aim of the present investigation was to further establish mutation frequencies for genes commonly mutated in UC, and to investigate possible positive or negative associations between mutations. To accomplish this we sequenced 14 genes, previously shown to be mutated in UC, in 145 cases. FGFR3, PIK3CA, HRAS, KRAS, NRAS, and TP53 were sequenced in an additional 73 cases. In addition, we sequenced PIK3R1 and TSC2 as no mutation data has been published for these two genes in UC. The observed high frequencies of FGFR3, TP53, and PIK3CA mutations were as expected. We detected a positive association between FGFR3 and PIK3CA activating mutations as reported [18], [20], and by combing our data with previously published data we could confirm that this association is highly significant. Two of our cases showed double mutations in PIK3CA; one case with a HD and a KD domain mutation, and one case with two mutations in the HD domain (E542K E545K). If these double mutations had occurred in separate alleles could not be determined as the electropherograms indicated heterozygous mutations. Irrespectively, the repeated observations of PIK3CA double mutated tumors and cell lines [4], [18], [25], [26] indicate an additive effect of multiple PIK3CA mutations.

PTEN acts as negative modulator of PIK3CA and is mutated in many tumor types as well as in UC [27], [28]. No PTEN mutations were however detected in the present set of tumors, a finding in line with previous reports of a low PTEN mutation frequency in UC (2%, n = 88) [29], [30]. This does however not exclude PTEN as an important factor in bladder cancer development since PTEN frequently shows reduced expression [31] as well as homozygous deletion in UC [29]. As mentioned, PIK3R1 also acts as a negative modulator of PIK3CA and has been shown to have tumor suppressor activity [32]. We found one case with a mutation after sequencing exons 12, 14, and 15, the most frequently mutated regions in this gene [33]. These exons cover the major part of the PIK3R1 iSH2 domain that interacts with PIK3CA. The frequent down regulation of PTEN expression and the here reported mutational inactivation of PIK3R1 further emphasized the importance of PIK3CA activity in UC.

Published data on RAS mutation is highly variable with respect to frequency, the distribution of codon 12 and 13 vs. codon 61 mutations in HRAS, and the proportion between HRAS, KRAS, and NRAS mutations [5], [34], [35], [36]. The divergent results may partly be attributed to differences in methods used and partly to differences in sample selection. In the present investigation HRAS mutations dominated over KRAS mutations and HRAS codon 61 mutations were much more frequent than codon 12 mutations, which is in line with most previous investigations.

It has been suggested that RAS mutations may substitute for FGFR3 mutations and that RAS and FGFR3 mutations therefore are mutually exclusive [21]. This hypothesis is however weakened by the fact that samples with concomitant FGFR3 and RAS mutations were identified in the present investigation as well as in the studies by Platt et al. and Kompier et al. [4], [18]. These findings may either be explained by intra-tumor heterogeneity as suggested by Platt et al., or that the biological consequences of activating FGFR3 and RAS mutations do not overlap completely and therefore double mutants may result in an additive but small selective advantage. Irrespective if these mutations are mutually exclusive or not, the evidence for a strong negative association between these two genes is compelling [4], [18], [21]. BRAF operates downstream of the RAS proteins and has been shown to be mutated in several tumor types [22]. We did not detect any mutations in any of the RAF gene family members. Hence, our data show that the previously reported low frequency or absence of BRAF mutations is not substituted for by frequent mutations in either ARAF or RAF1. Consequently, RAF mutations are not as central for UC development as FGFR3, PIK3CA, and RAS mutations.

Few and highly divergent mutation frequencies of APC and CTNNB1 have been reported for UC [12], [13], [37], [38]. Stoehr et al. found no APC or CTNNB1 mutations in 99 investigated cases, which is in stark contrast to Kastritis et al. who found 11 APC mutations in 70 cases (16%), but no CTNNB1 mutations in 35 investigated cases. Shiina et al., on the other hand, identified a total of 4 CTNNB1 mutations in 64 UC cases (6%). Our results showing APC mutations in 4% and CTNNB1 mutations in 2% of the cases are fully compatible with the combined published data resulting in overall mutation frequencies of 6.5% for APC and 2.0% for CTNNB1. As APC and CTNNB1 mutations were not seen together the frequency of APC/CTNNB1 pathway alteration was 6% in our data. As Kastritis et al. [12], we found a strong bias towards missense mutations, in contrast to truncating mutations, in APC and we could show that the mutation spectrum in UC differs significantly from the spectrum seen in colorectal cancer.

We detected a total of 17 cases (12%) with TSC1 mutations in our data, which is close to the previously reported frequency of 16% [4]. TSC1 is an established tumor suppressor gene in UC [4] that exert most of its regulatory function in a complex with TSC2. Consequently, we also screened TSC2 for mutations and detected 5 mutations in 145 cases (3%). This is a slightly higher frequency than what has been reported for other investigated solid tumors; CNS tumors 0.6%, lung cancers 0.9%, ovarian carcinomas 0.6% (COSMIC database). TSC2 acts as a dimer together with TSC1 by regulating mTOR through RHEB. The activity of TSC1/2 is, in turn, regulated by input from several upstream regulators [39] making the TSC1/2 a hub for upstream signals funneled to mTOR. The combined frequency of TSC1/TSC2 mutations of 15% indicates that a substantial proportion of UC tumors may show activation of mTOR through TSC1 or TSC2 mutational inactivation.

All in all, the present investigation emphasizes FGFR3, PIK3CA/AKT1, and TSC1/TSC2 as important nodes in intracellular signaling of transformed urothelial cells. To what extent FGFR3 activation may be directly linked to mTOR activation remain to be elucidated; we note that FGFR3 and PIK3CA mutations show a strongly skewed distribution between low and high grade tumors whereas TSC1/TSC2 mutations were seen in both categories at almost equal frequencies. Importantly, the detected alterations in the APC/CTNNB1 signaling pathway may also influence mTOR activity since GSK3B, an important member of the APC/CTNNB1 signaling pathway, is one of many modulators of TSC1 [40], [41]. There is further evidence for a crosstalk between APC/CTNNB1 and TSC1/TSC2 since wild type, but not inactive or mutated TSC2, regulates CTNNB1 negatively at the level of the CTNNB1 degradation complex [42]. In conclusion our data underscore the possible importance of mTOR activity in the development of UC. As mTOR activity is tractable for drug treatment [43], [44], and may be a possible target for various treatment regimes [43], [45], [46], [47], future investigations should be directed specifically towards mTOR activity in UC.

Supporting Information

Figure S1.

Distribution of mutations in non-invasive Ta tumors (A), in T1 tumors (B) and in (C) muscle-invasive tumors (≥T2). Red squares indicate inactivating mutation. Green squares indicate activating mutation. For PIK3CA, dark green squares indicate kinase domain mutations and light green helical domain mutations. At the right mutation frequencies are given for each gene in the respective tumor stages.

doi:10.1371/journal.pone.0018583.s001

(TIF)

Table S1.

Clinical features and mutation data.

doi:10.1371/journal.pone.0018583.s002

(XLS)

Table S2.

Primers and PCR conditions.

doi:10.1371/journal.pone.0018583.s003

(XLS)

Table S3.

Mutation details.

doi:10.1371/journal.pone.0018583.s004

(XLS)

Acknowledgments

The authors would like to thank Professor Ellen Zwarthoff for sharing details on the mutation data by Kompier et al., and for giving us useful comments.

Author Contributions

Conceived and designed the experiments: MH DL. Performed the experiments: GS CH DL. Analyzed the data: ML GS CH DL GC MH. Contributed reagents/materials/analysis tools: SG FL WM. Wrote the paper: GS MH DL.

References

  1. 1. Bakkar AA, Wallerand H, Radvanyi F, Lahaye JB, Pissard S, et al. (2003) FGFR3 and TP53 gene mutations define two distinct pathways in urothelial cell carcinoma of the bladder. Cancer Res 1;63(23): 8108–12.
  2. 2. van Rhijn BW, van der Kwast TH, Vis AN, Kirkels WJ, Boevé ER, et al. (2004) FGFR3 and P53 characterize alternative genetic pathways in the pathogenesis of urothelial cell carcinoma. Cancer Res 15;64(6): 1911–4.
  3. 3. Lindgren D, Frigyesi A, Gudjonsson S, Sjödahl G, Hallden C, et al. (2010) Combined gene expression and genomic profiling define two intrinsic molecular subtypes of urothelial carcinoma and gene signatures for molecular grading and outcome. Cancer Res 1;70(9): 3463–72.
  4. 4. Platt FM, Hurst CD, Taylor CF, Gregory WM, Harnden P, et al. (2009) Spectrum of phosphatidylinositol 3-kinase pathway gene alterations in bladder cancer. Clin Cancer Res 1;15(19): 6008–17.
  5. 5. Boulalas I, Zaravinos A, Karyotis I, Delakas D, Spandidos DA (2009) Activation of RAS family genes in urothelial carcinoma. J Urol 181(5): 2312–9.
  6. 6. Boulalas I, Zaravinos A, Delakas D, Spandidos DA (2009) Mutational analysis of the BRAF gene in transitional cell carcinoma of the bladder. Int J Biol Markers 24(1): 17–21.
  7. 7. Askham JM, Platt F, Chambers PA, Snowden H, Taylor CF, et al. (2010) AKT1 mutations in bladder cancer: identification of a novel oncogenic mutation that can co-operate with E17K. Oncogene 7;29(1): 150–5.
  8. 8. Heidenblad M, Lindgren D, Jonson T, Liedberg F, Veerla S, et al. (2008) Tiling resolution array CGH and high density expression profiling of urothelial carcinomas delineate genomic amplicons and candidate target genes specific for advanced tumors. BMC Med Genomics 31;1: 3.
  9. 9. Hidalgo M, Rowinsky EK (2000) The rapamycin-sensitive signal transduction pathway as a target for cancer therapy. Oncogene 27;19(56): 6680–6.
  10. 10. Gust KM, So AI (2009) The role of mTOR in bladder cancer. Cancer Biol Ther 8(24): 2348–50.
  11. 11. Hansel DE, Platt E, Orloff M, Harwalker J, Sethu S, et al. (2010) Mammalian Target of Rapamycin (mTOR) Regulates Cellular Proliferation and Tumor Growth in Urothelial Carcinoma. Am J Pathol 176(6): 3062–72.
  12. 12. Kastritis E, Murray S, Kyriakou F, Horti M, Tamvakis N, et al. (2009) Somatic mutations of adenomatous polyposis coli gene and nuclear b-catenin accumulation have prognostic significance in invasive urothelial carcinomas: evidence for Wnt pathway implication. Int J Cancer 1;124(1): 103–8.
  13. 13. Shiina H, Igawa M, Shigeno K, Terashima M, Deguchi M, et al. (2002) Beta-catenin mutations correlate with over expression of C-myc and cyclin D1 Genes in bladder cancer. J Urol 168(5): 2220–6.
  14. 14. Zhu X, Kanai Y, Saito A, Kondo Y, Hirohashi S (2000) Aberrant expression of beta-catenin and mutation of exon 3 of the beta-catenin gene in renal and urothelial carcinomas. Pathol Int 50(12): 945–52.
  15. 15. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, et al. (2010) A method and server for predicting damaging missense mutations. Nat Methods Apr;7(4): 248–9.
  16. 16. Hernández S, López-Knowles E, Lloreta J, Kogevinas M, Amorós A, et al. (2006) Prospective study of FGFR3 mutations as a prognostic factor in nonmuscle invasive urothelial bladder carcinomas. J Clin Oncol 1;24(22): 3664–71.
  17. 17. Van Rhijn BW, Vis AN, van der Kwast TH, Kirkels WJ, Radvanyi F, et al. (2003) Molecular grading of urothelial cell carcinoma with fibroblast growth factor receptor 3 and MIB-1 is superior to pathologic grade for the prediction of clinical outcome. J Clin Oncol 15;21(10): 1912–21.
  18. 18. Kompier LC, Lurkin I, van der Aa MN, van Rhijn BW, van der Kwast TH, et al. (2010) FGFR3, HRAS, KRAS, NRAS and PIK3CA Mutations in Bladder Cancer and Their Potential as Biomarkers for Surveillance and Therapy. PLoS One 3;5(11): e13821.
  19. 19. Kalinsky K, Jacks LM, Heguy A, Patil S, Drobnjak M, et al. (2009) PIK3CA mutation associates with improved outcome in breast cancer. Clin Cancer Res 15;15(16): 5049–59.
  20. 20. Lopez-Knowles E, Hernandez S, Malats N, Kogevinas M, Lloreta J, et al. (2006) PIK3CA Mutations Are an Early Genetic Alteration Associated with FGFR3 Mutations in Superficial Papillary Bladder Tumors. Cancer Res 66(15): 7401–4.
  21. 21. Jebar AH, Hurst CD, Tomlinson DC, Johnston C, Taylor CF (2005) FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma. Oncogene 4;24(33): 5218–25.
  22. 22. Edlundh–Rose E, Egyházi S, Omholt K, Månsson-Brahme E, Platz A, et al. (2006) NRAS and BRAF mutations in melanoma tumours in relation to clinical characteristics: a study based on mutation screening by pyrosequencing. Melanoma Res 16(6): 471–8.
  23. 23. Razzaque MA, Nishizawa T, Komoike Y, Yagi H, Furutani M, et al. (2007) Germline gain-of-function mutations in RAF1 cause Noonan syndrome. Nat Genet 39: 1013–1017.
  24. 24. Lagarde A, Rouleau E, Ferrari A, Noguchi T, Qiu J, et al. (2010) Germline APC mutation spectrum derived from 863 genomic variations identified through a 15-year medical genetics service to French patients with FAP. J Med Genet 47(10): 721–2.
  25. 25. Saal LH, Holm K, Maurer M, Memeo L, Su T, et al. (2005) PIK3CA mutations correlate with hormone receptors, node metastasis, and ERBB2, and are mutually exclusive with PTEN loss in human breast carcinoma. Cancer Res 1;65(7): 2554–9.
  26. 26. Lee JW, Soung YH, Kim SY, Lee HW, Park WS, et al. (2005) PIK3CA gene is frequently mutated in breast carcinomas and hepatocellular carcinomas. Oncogene 17;24(8): 1477–80.
  27. 27. Chalhoub N, Baker SJ (2009) PTEN and the PI3-kinase pathway in cancer. Annu Rev Pathol 4: 127–50.
  28. 28. Jaiswal BS, Janakiraman V, Kljavin NM, Chaudhuri S, Stern HM, et al. (2009) Somatic mutations in p85alpha promote tumorigenesis through class IA PI3K activation. Cancer Cell 8;16(6): 463–74.
  29. 29. Cairns P, Evron E, Okami K, Halachmi N, Esteller M, et al. (1998) Point mutation and homozygous deletion of PTEN/MMAC1 in primary bladder cancers. Oncogene 18;16(24): 3215–8.
  30. 30. Aveyard JS, Skilleter A, Habuchi T, Knowles MA (1999) Somatic mutation of PTEN in bladder carcinoma. Br J Cancer 80(5–6): 904–8.
  31. 31. Puzio-Kuter AM, Castillo-Martin M, Kinkade CW, Wang X, Shen TH, et al. (2009) Inactivation of p53 and Pten promotes invasive bladder cancer. Genes Dev 15;23(6): 675–80.
  32. 32. Taniguchi CM, Winnay J, Kondo T, Bronson RT, Guimaraes AR, et al. (2010) The phosphoinositide 3-kinase regulatory subunit p85alpha can exert tumor suppressor properties through negative regulation of growth factor signaling. Cancer Res 1;70(13): 5305–15.
  33. 33. Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, et al. (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 26;321(5897): 1807–12.
  34. 34. Olderøy G, Daehlin L, Ogreid D (1998) Low-frequency mutation of Ha-ras and Ki-ras oncogenes in transitional cell carcinoma of the bladder. Anticancer Res 18(4A): 2675–8.
  35. 35. Cattan N, Saison-Behmoaras T, Mari B, Mazeau C, Amiel JL, et al. (2000) Screening of human bladder carcinomas for the presence of Ha-ras codon 12 mutation. Oncol Rep 7(3): 497–500.
  36. 36. Przybojewska B, Jagiello A, Jalmuzna P (2000) H-RAS, K-RAS, and N-RAS gene activation in human bladder cancers. Cancer Genet Cytogenet 121(1): 73–7.
  37. 37. Stoehr R, Krieg RC, Knuechel R, Hofstaedter F, Pilarsky C, et al. (2002) No evidence for involvement of beta-catenin and APC in urothelial carcinomas. Int J Oncol 20(5): 905–11.
  38. 38. Shiina H, Igawa M, Urakami S, Shigeno K, Yoneda T, et al. (2001) Alterations of beta- and gamma-catenin in N-butyl-N-(-4-hydroxybutyl)nitrosamine-i​nducedmurine bladder cancer. Cancer Res 1;61(19): 7101–9.
  39. 39. Rosner M, Freilinger A, Hengstschläger M (2004) Proteins interacting with the tuberous sclerosis gene products. Amino Acids Oct;27(2): 119–28.
  40. 40. Inoki K, Ouyang H, Zhu T, Lindvall C, Wang Y, et al. (2006) TSC2 integrates Wnt and energy signals via a coordinated phosphorylation by AMPK and GSK3 to regulate cell growth. Cell 8;126(5): 955–68.
  41. 41. Buller CL, Loberg RD, Fan MH, Zhu Q, Park JL, et al. (2008) A GSK-3/TSC2/mTOR pathway regulates glucose uptake and GLUT1 glucose transporter expression. Am J Physiol Cell Physiol 295(3): C836–43.
  42. 42. Mak BC, Takemaru K, Kenerson HL, Moon RT, Yeung RS (2003) The tuberin-hamartin complex negatively regulates beta-catenin signaling activity. J Biol Chem 21;278(8): 5947–51.
  43. 43. Yu K, Shi C, Toral-Barza L, Lucas J, Shor B, et al. (2010) Beyond rapalog therapy: preclinical pharmacology and antitumor activity of WYE-125132, an ATP-competitive and specific inhibitor of mTORC1 and mTORC2. Cancer Res 15;70(2): 621–31.
  44. 44. Garcia JA, Danielpour D (2008) Mammalian target of rapamycin inhibition as a therapeutic strategy in the management of urologic malignancies. Mol Cancer Ther 7(6): 1347–54.
  45. 45. Seager CM, Puzio-Kuter AM, Patel T, Jain S, Cordon-Cardo C, et al. (2009) Intravesical delivery of rapamycin suppresses tumorigenesis in a mouse model of progressive bladder cancer. Cancer Prev Res 2(12): 1008–14.
  46. 46. Svirshchevskaya EV, Mariotti J, Wright MH, Viskova NY, Telford W, et al. (2008) Rapamycin delays growth of Wnt-1 tumors in spite of suppression of host immunity. BMC Cancer 21;8: 176.
  47. 47. Hou G, Zhang Q, Wang L, Liu M, Wang J, et al. (2010) mTOR inhibitor rapamycin alone or combined with cisplatin inhibits growth of esophageal squamous cell carcinoma in nude mice. Cancer Lett 28;290(2): 248–54.