Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Somatic Copy Number Abnormalities and Mutations in PI3K/AKT/mTOR Pathway Have Prognostic Significance for Overall Survival in Platinum Treated Locally Advanced or Metastatic Urothelial Tumors

  • Joaquim Bellmunt ,

    joaquim_bellmunt@dfci.harvard.edu

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Lillian Werner,

    Affiliation Biostatistics and Computational Biology, Harvard Medical School, Dana-Farber Cancer Institute, Boston, MA, United States of America

  • Jeffrey J. Leow,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Stephanie A. Mullane,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • André P. Fay,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Markus Riester,

    Affiliation Biostatistics and Computational Biology, Harvard Medical School, Dana-Farber Cancer Institute, Boston, MA, United States of America

  • Paul Van Hummelen,

    Affiliation Center for Cancer Genome Discovery, Dana-Farber Cancer Institute, Boston, MA, United States of America

  • Mary-Ellen Taplin,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Toni K. Choueiri,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Eliezer Van Allen,

    Affiliation Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute/Brigham and Women's Hospital, Harvard Medical School, Boston, MA, United States of America

  • Jonathan Rosenberg

    Affiliation Memorial Sloan Kettering Cancer Center, New York, NY, United States of America

Abstract

Background

An integrative analysis was conducted to identify genomic alterations at a pathway level that could predict overall survival (OS) in patients with advanced urothelial carcinoma (UC) treated with platinum-based chemotherapy.

Patients and Methods

DNA and RNA were extracted from 103 formalin-fixed paraffin embedded (FFPE) invasive high-grade UC samples and were screened for mutations, copy number variation (CNV) and gene expression analysis. Clinical data were available from 85 cases. Mutations were analyzed by mass-spectrometry based on genotyping platform (Oncomap 3) and genomic imbalances were detected by comparative genomic hybridization (CGH) analysis. Regions with threshold of log2 ratio ≥0.4, or ≤0.6 were defined as either having copy number gain or loss and significantly recurrent CNV across the set of samples were determined using a GISTIC analysis. Expression analysis on selected relevant UC genes was conducted using Nanostring. To define the co-occurrence pattern of mutations and CNV, we grouped genomic events into 5 core signal transduction pathways: 1) TP53 pathway, 2) RTK/RAS/RAF pathway, 3) PI3K/AKT/mTOR pathway, 4) WNT/CTNNB1, 5) RB1 pathway. Cox regression was used to assess pathways abnormalities with survival outcomes.

Results

35 samples (41%) harbored mutations on at least one gene: TP53 (16%), PIK3CA (9%), FGFR3 (2%), HRAS/KRAS (5%), and CTNNB1 (1%). 66% of patients had some sort of CNV. PIK3CA/AKT/mTOR pathway alteration (mutations+CNV) had the greatest impact on OS (p=0.055). At a gene level, overexpression of CTNNB1 (p=0.0008) and PIK3CA (p=0.02) were associated with shorter OS. Mutational status on PIK3CA was not associated with survival. Among other individually found genomic alterations, TP53 mutations (p=0.07), mTOR gain (p=0.07) and PTEN overexpression (p=0.08) have a marginally significant negative impact on OS.

Conclusions

Our study suggests that targeted therapies focusing on the PIK3CA/AKT/mTOR pathway genomic alterations can generate the greatest impact in the overall patient population of high-grade advanced UC.

Introduction

Urothelial carcinoma (UC) is a common malignancy in the United States with nearly 75,000 cases diagnosed annually, and with more than 15,000 disease-related deaths[1]. UC is categorized into non-muscle invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC). About 70% of UC tumors present as NMIBC, which is usually low grade and indolent[2]. MIBC consists of the other 30% of UC and generally portends a poor prognosis, with 5-year overall survival rates of 50% when treated with neoadjuvant chemotherapy plus radical cystectomy[3].

NMIBC and MIBC have been described as heterogeneous tumors with different genomic landscapes. NMIBC is generally characterized by mutations in FGFR3, RAS, and deletions on chromosome 9, while MIBC is characterized by mutations in TP53, RB1, PIK3CA and PTEN[46]. The concept that most malignancies depend on driver mutations to establish and maintain the malignant phenotype has been established, although involvement of other genomic events and epigenetic alterations has been also recognized as potential drivers for carcinogenesis.

The Cancer Genome Atlas (TCGA) Research Network recently published a comprehensive molecular characterization of urothelial bladder carcinoma and recurrent mutations in 32 genes, including TP53, FGFR3, ERCC2, NF2, and CDKN1A were described to play a role in UC tumorigenesis[7]. Limited data are available from this analysis associating single or grouped genomic alterations with survival outcomes in patients with UC.

In this study, we present an integrative analysis of multiple types of genomic data, including mutations, copy number variations (CNV), and messenger RNA (mRNA) expression data from 103 metastatic UC patients treated at two institutions with first-line platinum-based combination chemotherapy. We have combined genomic events on 5 signal transduction pathways known to be important in UC tumorigenesis and maintenance of a malignant phenotype, and correlated these findings with clinical outcomes. The 5 cancer signaling pathways selected for this analysis were TP53, RTK/RAS/ /RAF, PI3KCA/AKT/mTOR, WNT/CTNNB1, and RB1.

Material and Methods

Patients

This project was approved by the local ethics committee (CEIC-IMAS) at Hospital del Mar and by the Dana-Farber/Harvard Cancer Center (DF/HCC) Institutional Review Board (IRB). A total of 103 clinically annotated patients with locally advanced or metastatic UC who were treated with first-line platinum-based combination chemotherapy were identified and tumor specimens were retrieved from the Pathology Departments. Since the majority of patients were dead at the time of sample collection, a waiver of consent was requested and given from the DF/HCC for all participants (requiring complete de-identification of the samples prior the analysis).

Tumor Samples

DNA was extracted from formalin-fixed paraffin embedded (FFPE) material using the QIAamp DNA FFPE Tissue Kit (Qiagen, Valencia, CA). We performed comparative genomic hybridization (CGH) to determine genomic imbalances using genomic DNA isolated from primary tumors as well as karyotypically normal reference genomic DNA (Promega, Madison, WI). Agilent Oligonucleotide Human Genome 180k CGH arrays were used to perform the analysis. The Genomic DNA ULS labeling kit for FFPE Samples (Agilent Technologies, Inc., Palo Alto, CA) was used to chemically label 500ng of DNA with either ULS-Cy5 (tumor) or ULS-Cy3 dye (normal/reference DNA) following the manufacturer's protocol. Samples were hybridized to the Agilent SurePrint G3 Human CGH Microarray 4x180K for 40 hours in a Robbins Scientific oven with rotation at 20 rpm at 65°C. Post-hybridization, the slides were washed and scanned using an Agilent DNA microarray scanner. CGH Analytics software (version 3.4, Agilent Technologies, CA) was used to evaluate the aCGH data. The GISTIC module was used to identify regions of the genome that were lost or gained across the set of samples. Regions with threshold of log2 ratio of ≥0.4, or ≤ -0.3 were defined as either having copy number gain or loss, respectively. Throughput mutation profiling was performed by using both mass spectroscopy-based genotyping (Oncomap 3 platform) and confirmed with hME sequencing.

mRNA expression profile was obtained by using Nanostring technology. mRNA was extracted from tumor specimens using standard protocols. Oligonucleotide probes for all genes analyzed were synthesized by Nanostring Technologies, and transcripts were counted using the automated Nanostring nCounter system. Counts were normalized with the nSolver Analysis Software (v1.0) in which mRNA expression was compared to internal Nanostring controls, several housekeeping (ACTB, GAPHD, HPRT1, LDHA, PFKP, PGAM1, STAT1, TUBA4A, VIM) and invariant genes in bladder cancer (ANGEL1, DDX19A, NAGA, RPS10, RPS16, RPS24, RPS29). These invariant genes were identified by analyzing gene expression variances in several published datasets (12, 14, 15-1q23).

Statistical Analysis

Overall survival (OS) was defined as the time period from the first chemotherapy administration to the date of death or censored on the last known alive date. Median OS rates were calculated based on date of death or last known alive date. Using the log-rank test and univariate Cox proportional regression, we tested the primary endpoint of this study which was to evaluate the association between pathway abnormalities (mutations, CNV, both genomic alterations taken together, and expression analysis in signaling pathways) and OS. For 85 patients with mutational and clinical data, mutational status of each gene identified was examined in multivariate Cox regression models. The association of CNV status (gain/amplification vs. loss) and OS was evaluated in 93 patients with CNV and clinical data available. Additionally, for the 74 patients having clinical and expression data from nanostring, we examined the association of overexpression of selected genes within a pathway with OS, adjusting for clinical variables (performance status and visceral metastases).

All statistical analyses were performed using SAS 9.3 (SAS Institute, NC). All tests were two-sided and a p-value of <0.05 was considered statistically significant.

Results

A total of 103 metastatic UC patients treated with first-line platinum-based combination chemotherapy were included in the analysis. Clinical characteristics are outlined in Table 1. Median follow up was 23 months. The median OS since the application of first chemotherapy for treatment of metastatic disease was 12 months.

Association of Mutational Status and Survival Outcome

Clinical data were available for 85 out of 93 samples (91%) scanned for gene mutations and those 85 were included in the analysis. Using the Oncomap 3 platform, we identified the proportion of patients with gene mutations: TP53 (16%), PIK3CA (11%), HRAS/KRAS (5%), FGFR3 (2%), CTNNB1 (1%), BRAF (1%) and ERBB2 (1%) (Table 2).

thumbnail
Table 2. Name of Gene & Candidate Mutation validated by hME.

https://doi.org/10.1371/journal.pone.0124711.t002

We examined whether the most frequent 3 mutations above were associated with OS. Due to the small mutation prevalence in the other genes detected with this platform (0–2%), those were not included in the analysis. TP53 mutation was associated with a shorter median OS (10 vs. 16 months) trending toward significance (p = 0.07). Mutations of PIK3CA (not reached vs. 14 months, p = 0.26) and HRAS/KRAS (17 vs. 12 months, p = 0.17) alone were not associated with shorter OS (Table 3). All other mutations were not significantly associated with OS.

Association of CNV and Survival Outcome

CNV were identified in 66% of patients. Using the GISTIC algorithm[8] we identified 96 focal (< 50% of a chromosome arm) and 22 broad (> 50% of a chromosome arm) CNV events. Association between CNV and OS were evaluated in 94 patients with corresponding clinical data available. Adverse genomic alterations were reported according to the respective signal transduction pathway: (1) TP53 (TP53 loss 13%), (2) RTK/RAS/RAF (RAF1 gain 24%, ERBB2 gain 19%, FGFR1 gain 12%, BRAF gain 5%, KRAS gain 4%, MET gain 4%, FGFR3 gain 2%, NF1 loss 1%); (3) PI3K/ATK/mTOR (TSC1 loss 13%, PIK3CA gain 10%, PTEN loss 6%, AKT1 gain 2%, MTOR gain 2%), (4) WNT/CTNNB1 (CTNNB1 gain 4%) and (5) RB1 (CCND1 gain 11%, CCNE1 gain15, CDKN2A loss 2%, E2F3 gain 28%, RB1 gain 88%). None of the CNV events alone were significantly correlated with OS (Table 4). The only individual CNV which correlated with a trend in terms of OS benefit was a copy number gain of mTOR (p = 0.07).

thumbnail
Table 4. Association of CNV of genes In TP53, RTK/RAS/RAF, PI3KCA/AKT/mTOR, WNT/CTNNB1 and RB1 pathways with OS.

https://doi.org/10.1371/journal.pone.0124711.t004

Focal gains around ERBB2 were correlated with an increase ERBB2 mRNA expression as compared with non-amplified samples. However, neither ERBB2 mRNA overexpression nor CNV (gain) were associated with survival outcomes. We also observed a slight increase in mRNA expression of RAF1 in samples with copy number gain. There were no changes in mRNA expression in patients with CNV in BRAF, FGFR3, and KRAS locus.

Integrative Pathway Analysis and Correlation with Clinical Outcome

Patients with genomic events (CNV, mutation, and gene expression) in the TP53, RTK/RAS/RAF, PI3K/AKT/mTOR, WTN/CTNNB1, or RB1 signaling pathways were analyzed according to survival outcomes.

TP53 pathway.

The MDM2 gene, which is part of the TP53 pathway, was evaluated for mutations, CNV and gene expression and no genomic events were identified. Therefore, genomic events in this pathway are represented by TP53 gene mutation. TP53 mutation was correlated with a shorter OS (p = 0.07) (S1 Table).

RTK/RAS/RAF pathway.

The association of mutational status and CNV of all genes within the RTK/RAS/RAF pathway [ERBB2 (copy number gain or mutation), FGFR3 (copy number gain or mutation), MET (copy number gain), FGFR1 (copy number gain), KRAS (copy number gain or mutation), HRAS (mutation), BRAF (copy number gain or mutation), RAF1 (copy number gain), NF1 (copy number loss)] with OS was not statistically significant (p = 0.56) (S1 Table). Genomic events within the RTK/RAS/RAF pathway were mutually exclusive (data not shown). There were overlaps of RAF1 and FGFR3 genomic events in 1 patient as well as ERBB2 with RAF1, BRAF, and KRAS in 2 patients.

PI3KCA/AKT/mTOR pathway.

Thirty-two patients (31%) had genomic events within the PI3KCA/AKT signaling pathway: PTEN (copy number loss 6%), PI3KCA (copy number gain 10% or mutation 11%), AKT1 (copy number gain 2%), TSC1 (copy number loss, 13%), and MTOR (copy number gain 2%) (Fig 1). The only mutation observed within the PI3KCA/AKT pathways was PIK3CA (E545K and H1047R) (11%) (S1 Table) (Table 5).

thumbnail
Fig 1. Association between mutations or CNV and OS.

A: PIK3CA/AKT/mTOR pathway; B: Gene expression level analysis whithin the PI3KCA/AKT pathway; C: Heat map: mutations across the PIK3CA/AKT/mTOR pathway; D: Association between PIK3CA/AKT/mTOR pathway mutations or CNV and OS.

https://doi.org/10.1371/journal.pone.0124711.g001

thumbnail
Table 5. Association between gene expression levels and OS.

https://doi.org/10.1371/journal.pone.0124711.t005

Gene expression level analysis in the PI3KCA/AKT pathway was conducted. Expression levels were dichotomized at median. Overexpression of PI3KCA was found to be statistically significant correlated with OS (p = 0.02). Similarly, a trend of longer survival in patients with lower levels of PTEN gene expression was also observed (p = 0.08). We did not see significant associations between gene expression of any other genes in this pathway and OS (Table 3).

The association of CNV or mutation with OS within the whole PI3KCA/AKT pathway was very close to statistical significance (p = 0.055) (Fig 1). All of the genomic events within the PI3KCA were mutually exclusive. We then analyzed the association of PI3KCA pathway abnormality using CNV, mutation and gene expression data with OS. We were unable to show an association of these genomic events in any of the genes from PIK3/AKT pathway and OS (Table 6). When focusing at the gene level, PI3KCA abnormalities (CNV, mutation and gene expression) were associated with shorter OS (p = 0.04) (Table 7).

thumbnail
Table 6. Association between genomic events (CNV, mutation or gene expression) in any genes of PIK3/AKT pathway and OS.

https://doi.org/10.1371/journal.pone.0124711.t006

WNT/CTNNB1 pathway.

The WNT/CTNNB1 pathway includes WNT, Axin, APC, GSK3beta, CKIα, and CTNNB1. We observed one patient (1%) with CTNNB1 mutation. When we examined the overexpression of CTNNB1, we found that CTNNB1 overexpression was associated with shorter OS (p = 0.0008) (Fig 2). No other genomic alterations were identified within this pathway.

thumbnail
Fig 2. A: WNT/CTNNB1 pathway; B: Association between CTNNB1 overexpression and OS.

https://doi.org/10.1371/journal.pone.0124711.g002

RB1 Pathway.

Genomic events in the RB1 pathway were defined as CCND1 copy number gain, CCNE1 copy number gain, CDKN2A copy number loss, E2F3 copy number gain, or RB1 copy number gain. CDK4 and CDN2A were not analyzed due to small number of patients with CNV. Gene expression level of each gene dichotomized at the median did not show any correlation with survival either alone or in combination with CNV. In an exploratory analysis, using the top or bottom 10% to categorize over- or underexpression, there was a statistically significant relationship favoring long-term survival in patients without RB1 pathway alteration (including CNV and gene expression) (p = 0.04) (Table 8; Fig 3).

thumbnail
Table 8. Association between genomic events in RB pathway (CNV with or without overexpression) and OS using 10% cutt-off.

https://doi.org/10.1371/journal.pone.0124711.t008

thumbnail
Fig 3. Association between RB1 pathway (CNV and expression) and OS.

https://doi.org/10.1371/journal.pone.0124711.g003

Discussion

We performed an integrative genomic analysis to determine the importance of individual genomic events in UC as well as the impact of those events when grouped into signaling pathways. The main finding in this study is that PIK3CA genomic events in the form of combined mutation, CNV and overexpression were statistically associated with shorter OS in patients with metastatic UC. A similar trend was observed at pathway level. We demonstrated that while there are many different genomic abnormalities within UC, there may be certain alterations which have a greater impact on survival. The correlation of genomic events with survival may help to better select targets for therapeutical intervention in UC[9,10].

Inactivating mutations or deletions in tumor suppressor genes like PTEN or TSC1, or amplifications or mutations in oncogenes like PIK3CA or AKT1 have been reported to promote constitutive activation of PI3KCA/AKT/mTOR pathway in up to 40% of UC[11,12]. PI3KCA/AKT/mTOR pathway is involved in extracellular growth signaling, metabolism and cell proliferation and are potential therapeutic targets in advanced UC[13]. We observed a PIK3CA mutation rate of 11%, which is slightly under the 15–20% reported in the literature[14]. However, the presence of this mutation alone did not correlate with OS. The only individual genomic event in the PIK3CA pathway which correlated with a trend in OS benefit was a copy number gain of mTOR (p = 0.07).

Mutually exclusive patterns of alterations in the PI3KCA/AKT/mTOR pathway were identified in 31% of patients. Taken together, different genomic events (gene expression, mutation, or CNV) into this pathway correlated with shorter OS (p = 0.055). This could indicate that targeted therapies against the genes within the PI3KCA/AKT/mTOR pathway may lead to improved OS.

Loss or reduced expression of PTEN have been also associated with activation of PI3KCA/AKT/mTOR pathway supporting the rationale for better responses to mTOR inhibitors in patients with PTEN negative tumors[15,16]. In this analysis, we observed that overexpression of PTEN and PI3KCA were correlated with shorter OS (p = 0.08, p = 0.02 respectively). However PI3KCA mutations did not correlate with overexpression of PI3KCA.

Positive prognostic significance of PIK3CA mutations have been reported in different malignancies[17,18]. Recently, Kim and colleagues reported the results from a study in which next generation sequencing was performed to identify prognostic genomic biomarkers in high-grade UC[19]. As previously reported in other tumors, PIK3CA or PI3K/AKT/mTOR pathway abnormalities were associated with a statistically significant improvement in disease-specific outcome in patients with non-metastatic UC. Interestingly, the rate of recurrence was still high among these patients (44% in 5 years). In contrast, our findings suggest that in the metastatic setting abnormalities in this pathway may predict shorter OS. Given the tumor heterogeneity in UC, prospective studies investigating the prognostic significance in both, metastatic and non-metastatic setting, should be performed in larger cohorts. In addition, these findings provide a rational to investigate the impact of agents targeting PI3K/AKT/mTOR pathway in the adjuvant and metastatic setting in a population enriched by genomic alterations in this pathway.

ERBB2 is reported to be amplified in 5% or more of tumor in UC[2022]. In our study, copy number gains in ERBB2 were identified in 19% of patients and were correlated with overexpression of the gene. However, neither overexpression nor CNV was correlated with OS. Although ERBB2 overexpression has been correlated with survival in other tumor types, such as breast cancer, there has been no previous report of this correlation in UC[23]. Notably, two patients had mutation in ERBB2. The role of these mutations as potential drivers or targets in advanced UC is still under investigation.

Less than 15% of MIBC or metastatic UC tumors harbor FGFR3 mutations[24]. In this analysis, 2 patients (2%) had activating mutation in FGFR3. It has been shown, in vivo and in vitro, that the inhibition of mutant FGFR3 leads to cell cycle arrest and/or apoptosis[25], providing a rational for the development of FGFR3 inhibitors. Dovitinib, a small molecule FGFR3 and VEGFR inhibitor, has been tested in a phase II clinical trial and limited effectiveness was observed. The lack of activity in clinical studies could be related to inefficient targeting or inadequate testing[26]. Recent reports have shown that genetic translocations and rearrangements of FGFR3 could activate the RKT/RAS/RAF pathway[27] and patients with these unusual alterations can be the population to respond to FGFR3 inhibitors. These genomic alterations are now being investigated as potential therapeutic targets with next-generation FGFR3 inhibitors and are showing more promise[28,29] as seen recently in one of our patients included in a pan FGFR inhibitor clinical trial.

In our study, we observed a large number of CNV within the RB1 pathway, but no mutations in RB1 gene were identified. We noted RB1 mutations in about 15% of advanced UC and those mutations lead to greater CNV[30]. RB1 mutations are rarely seen in low grade or low stage bladder tumors. Moreover, loss of heterozygosity at the RB locus (13q) is correlated with tumor grade and muscle invasion of bladder cancer[31]. The CNV variation seen in our analysis could indicate that RB1 loss of heterozygosity could also induces genomic instability and promotes aneuploidy. The increase in genomic instability could indicate why RB1 is correlated with higher grade and stage.

In addition, 12 patients (16%) had TP53 inactivating mutations, which was less than the expected prevalence previously reported for metastatic UC of 34%[30]30. TP53 plays essential roles in the regulation of cell proliferation, apoptosis and inhibition of angiogenesis. It is the most commonly mutated gene in cancer, including metastatic UC[7]. TP53 mutation is correlated with survival, which is consistent with other metastatic bladder cancer literature. We observed a loss of TP53 in 13% of patients. TP53 loss could be one reason we saw a lower prevalence of TP53 mutations in our cohort. We did not see any mutations or copy number alterations in MDM2, an E3 ubiquitin protein ligase, which negatively regulates TP53.

Studies investigating CTNNB1 mutations in UC have reported an average prevalence around 5%. In our cohort, only 1 patient (1%) had CTNNB1 mutation. The WNT/CTNNB1 signaling pathway plays an important role in cell differentiation and tumorigenesis. Without a WNT signal, CTNNB1 is degraded, but when there is a WNT signal, CTNNB1 is allowed to accumulate and enter the nucleus, and regulate its target genes. A mutation on the CTNNB1 allows it to escape the WNT regulation, which leads to an increase in transcription of its targeted genes[32]. In our study, CTNNB1 overexpression was significantly associated with a shorter OS. Because CTNNB1 entering the nucleus depends on the accumulation of CTNNB1, overexpression of this protein would increase the transcription of target genes, leading to increased cell proliferation and subsequently poorer survival outcomes.

Despite the fact that this is the first study correlating pathway abnormalities in a homogeneous cohort of metastatic UC patients treated with platinum-based chemotherapy, several limitations need to be considered. First, the genomic analysis was done using a targeted sequencing platform (oncomap) that can limit the capture of existing mutations that could be sensitive to therapeutic intervention[33,34]. As an example, this platform does not identify the recently described high-frequency mutations affecting chromatin-modifying genes37 or potential downstream biomarkers as previously described[35]. Second, it is possible that the prevalence of genomic alterations differ between primary and metastatic tumors. An analysis of matched primary and metastatic samples from the same patient would address this possibility and should be a focus of future research efforts. Finally, there exists some degree of uncertainty regarding the tumor heterogeneity within our samples, as we sampled a single site. In our analysis, we saw genomic heterogeneity across the patient population, but it is possible that individual tumors were heterogeneous as well. Finally, given the small numbers of patients and low prevalence of genomic events, correction for multiple comparisons or exploratory analysis (i.e. correlation of genomic events and clinical parameters) may not result in reliable results, and thus these results are hypothesis generating and require further external validation.

In conclusion, this study plays an important role in deciphering the interdependent nature of cancer genomic alterations to guide targeted therapies. Our results suggest that targeted therapies focusing on the PI3KCA/AKT/mTOR pathway genomic alterations may improve clinical outcome in patients with metastatic urothelial cancer.

Supporting Information

S1 Table. Association of mutation status and OS.

https://doi.org/10.1371/journal.pone.0124711.s001

(PDF)

Author Contributions

Conceived and designed the experiments: JB EVA LW MR PVH TKC JR. Performed the experiments: JB SAM EVA JL PVH MR JR. Analyzed the data: JB EVA SAM JL AF LW PVH TKC MR JR. Wrote the paper: JB EVA SAM JL AF LW MR PVH TKC MET JR.

References

  1. 1. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. CA Cancer J Clin. 2014;64(1):9–29.
  2. 2. Burger M, Oosterlinck W, Konety B, Chang S, Gudjonsson S, Pruthi R, et al. ICUD-EAU International Consultation on Bladder Cancer 2012: Non-muscle-invasive urothelial carcinoma of the bladder. European urology. 2013;63(1):36–44. pmid:22981672
  3. 3. Advanced Bladder Cancer Meta-analysis C. Neoadjuvant chemotherapy in invasive bladder cancer: update of a systematic review and meta-analysis of individual patient data advanced bladder cancer (ABC) meta-analysis collaboration. European urology. 2005;48(2):202–205; discussion 205–206. pmid:15939524
  4. 4. Kompier LC, Lurkin I, van der Aa MN, van Rhijn BW, van der Kwast TH, Zwarthoff EC. FGFR3, HRAS, KRAS, NRAS and PIK3CA mutations in bladder cancer and their potential as biomarkers for surveillance and therapy. PloS one. 2010;5(11):e13821. pmid:21072204
  5. 5. Juanpere N, Agell L, Lorenzo M, de Muga S, Lopez-Vilaro L, Murillo R, et al. Mutations in FGFR3 and PIK3CA, singly or combined with RAS and AKT1, are associated with AKT but not with MAPK pathway activation in urothelial bladder cancer. Human pathology. 2012;43(10):1573–1582. pmid:22417847
  6. 6. Korkolopoulou P, Levidou G, Trigka EA, Prekete N, Karlou M, Thymara I, et al. A comprehensive immunohistochemical and molecular approach to the PI3K/AKT/mTOR (phosphoinositide 3-kinase/v-akt murine thymoma viral oncogene/mammalian target of rapamycin) pathway in bladder urothelial carcinoma. BJU international. 2012;110(11 Pt C):E1237–1248.
  7. 7. The Cancer Genome Atlas Research N. Comprehensive molecular characterization of urothelial bladder carcinoma. Nature. 2014.
  8. 8. Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome biology. 2011;12(4):R41. pmid:21527027
  9. 9. Millis SZ, Bryant D, Basu G, Bender R, Vranic S, Gatalica Z, et al. Molecular Profiling of Infiltrating Urothelial Carcinoma of Bladder and Nonbladder Origin. Clinical genitourinary cancer. 2014.
  10. 10. Ross JS, Wang K, Al-Rohil RN, Nazeer T, Sheehan CE, Otto GA, et al. Advanced urothelial carcinoma: next-generation sequencing reveals diverse genomic alterations and targets of therapy. Modern pathology: an official journal of the United States and Canadian Academy of Pathology, Inc. 2014;27(2):271–280.
  11. 11. Bambury RM, Rosenberg JE. Actionable mutations in muscle-invasive bladder cancer. Current opinion in urology. 2013;23(5):472–478. pmid:23907505
  12. 12. Polivka J, Janku F. Molecular targets for cancer therapy in the PI3K/AKT/mTOR pathway. Pharmacol Ther. 2013.
  13. 13. Ching CB, Hansel DE. Expanding therapeutic targets in bladder cancer: the PI3K/Akt/mTOR pathway. Lab Invest. 2010;90(10):1406–1414. pmid:20661228
  14. 14. Platt FM, Hurst CD, Taylor CF, Gregory WM, Harnden P, Knowles MA. Spectrum of phosphatidylinositol 3-kinase pathway gene alterations in bladder cancer. Clin Cancer Res. 2009;15(19):6008–6017. pmid:19789314
  15. 15. Neshat MS, Mellinghoff IK, Tran C, Stiles B, Thomas G, Petersen R, et al. Enhanced sensitivity of PTEN-deficient tumors to inhibition of FRAP/mTOR. Proceedings of the National Academy of Sciences of the United States of America. 2001;98(18):10314–10319. pmid:11504908
  16. 16. Houede N, Pourquier P. Targeting the genetic alterations of the PI3K-AKT-mTOR pathway: Its potential use in the treatment of bladder cancers. Pharmacology & therapeutics. 2014.
  17. 17. Kalinsky K, Jacks LM, Heguy A, Patil S, Drobnjak M, Bhanot UK, et al. PIK3CA mutation associates with improved outcome in breast cancer. Clinical cancer research: an official journal of the American Association for Cancer Research. 2009;15(16):5049–5059. pmid:19671852
  18. 18. Shigaki H, Baba Y, Watanabe M, Murata A, Ishimoto T, Iwagami S, et al. PIK3CA mutation is associated with a favorable prognosis among patients with curatively resected esophageal squamous cell carcinoma. Clinical cancer research: an official journal of the American Association for Cancer Research. 2013;19(9):2451–2459. pmid:23532889
  19. 19. Kim PH, Cha EK, Sfakianos JP, Iyer G, Zabor EC, Scott SN, et al. Genomic Predictors of Survival in Patients with High-grade Urothelial Carcinoma of the Bladder. European urology. 2014.
  20. 20. Lae M, Couturier J, Oudard S, Radvanyi F, Beuzeboc P, Vieillefond A. Assessing HER2 gene amplification as a potential target for therapy in invasive urothelial bladder cancer with a standardized methodology: results in 1005 patients. Annals of oncology: official journal of the European Society for Medical Oncology / ESMO. 2010;21(4):815–819. pmid:19889613
  21. 21. Fleischmann A, Rotzer D, Seiler R, Studer UE, Thalmann GN. Her2 amplification is significantly more frequent in lymph node metastases from urothelial bladder cancer than in the primary tumours. Eur Urol. 2011;60(2):350–357.
  22. 22. Galsky MD, Von Hoff DD, Neubauer M, Anderson T, Fleming M, Nagarwala Y, et al. Target-specific, histology-independent, randomized discontinuation study of lapatinib in patients with HER2-amplified solid tumors. Invest New Drugs. 2012;30(2):695–701. pmid:20857170
  23. 23. Memon AA, Sorensen BS, Meldgaard P, Fokdal L, Thykjaer T, Nexo E. The relation between survival and expression of HER1 and HER2 depends on the expression of HER3 and HER4: a study in bladder cancer patients. Br J Cancer. 2006;94(11):1703–1709. pmid:16685269
  24. 24. van Rhijn BW, van der Kwast TH, Liu L, Fleshner NE, Bostrom PJ, Vis AN, et al. The FGFR3 mutation is related to favorable pT1 bladder cancer. J Urol. 2012;187(1):310–314. pmid:22099989
  25. 25. Lamont FR, Tomlinson DC, Cooper PA, Shnyder SD, Chester JD, Knowles MA. Small molecule FGF receptor inhibitors block FGFR-dependent urothelial carcinoma growth in vitro and in vivo. Br J Cancer. 2011;104(1):75–82. pmid:21119661
  26. 26. Milowsky MI, Dittrich C, Martinez ID, Jagdev S, Millard FE, Sweeney C, et al. Final results of a multicenter, open-label phase II trial of dovitinib (TKI258) in patients with advanced urothelial carcinoma with either mutated or nonmutated FGFR3. Journal of Clinical Oncology. (supplementary abstract) 2013.
  27. 27. Williams SV, Hurst CD, Knowles MA. Oncogenic FGFR3 gene fusions in bladder cancer. Hum Mol Genet. 2013;22(4):795–803. pmid:23175443
  28. 28. Rastislav B,Dienstmann R, Adamo B, Gazzah G, Infante JR, Zhong B,et al;. Phase 1 study of JNJ-42756493, a pan-fibroblast growth factor receptor (FGFR) inhibitor, in patients with advanced solid tumors. J Clin Oncol 32:5s, 2014 (suppl; abstr 2501). 2014.
  29. 29. Sequist LV Cassier P, Varga A, Tabernero J, Schellens JH, Delord JP, et al. Phase I study of BGJ398, a selective pan-FGFR inhibitor in genetically preselected advanced solid tumors. 2014 AACR Annual Meeting. Abstract CT326. Presented April 8, 2014. 2014.
  30. 30. Iyer G, Al-Ahmadie H, Schultz N, Hanrahan AJ, Ostrovnaya I, Balar AV, et al. Prevalence and co-occurrence of actionable genomic alterations in high-grade bladder cancer. Journal of clinical oncology: official journal of the American Society of Clinical Oncology. 2013;31(25):3133–3140.
  31. 31. Cairns P, Proctor AJ, Knowles MA. Loss of heterozygosity at the RB locus is frequent and correlates with muscle invasion in bladder carcinoma. Oncogene. 1991;6(12):2305–2309. pmid:1766677
  32. 32. Ahmad I, Singh LB, Foth M, Morris CA, Taketo MM, Wu XR, et al. K-Ras and beta-catenin mutations cooperate with Fgfr3 mutations in mice to promote tumorigenesis in the skin and lung, but not in the bladder. Disease models & mechanisms. 2011;4(4):548–555.
  33. 33. MacConaill LE, Campbell CD, Kehoe SM, Bass AJ, Hatton C, Niu L, et al. Profiling critical cancer gene mutations in clinical tumor samples. PloS one. 2009;4(11):e7887. pmid:19924296
  34. 34. Matulonis UA, Hirsch M, Palescandolo E, Kim E, Liu K, van Hummelen P. et al. High throughput interrogation of somatic mutations in high grade serous cancer of the ovary. PloS one. 2011;6(9):e24433. pmid:21931712
  35. 35. Gui Y, Guo G, Huang Y, Hu X, Tang A, Gao S. et al. Frequent mutations of chromatin remodeling genes in transitional cell carcinoma of the bladder. Nat Genet. 2011;43(9):875–878. pmid:21822268