Lucie C. Kompier

FGFR3, HRAS, KRAS, NRAS AND PIK3CA mutations in bladder cancer and their potential as biomarkers for surveillance and therapy

Background Fifty percent of patients with muscle–invasive bladder cancer (MI-BC) die from their disease and current chemotherapy treatment only marginally increases survival. Novel therapies targeting receptor tyrosine kinases or activated oncogenes may improve outcome. Hence, it is necessary to stratify patients based on mutations in relevant oncogenes. Patients with non-muscle-invasive bladder cancer (NMI-BC) have excellent survival, however two-thirds develop recurrences. Tumor specific mutations can be used to detect recurrences in urine assays, presenting a more patient-friendly diagnostic procedure than cystoscopy. Methodology/Principal Findings To address these issues, we developed a mutation assay for the simultaneous detection of 19 possible mutations in the HRAS, KRAS, and NRAS genes. With this assay and mutation assays for the FGFR3 and PIK3CA oncogenes, we screened primary bladder tumors of 257 patients and 184 recurrences from 54 patients. Additionally, in primary tumors p53 expression was obtained by immunohistochemistry. Of primary tumors 64% were mutant for FGFR3, 11% for RAS, 24% for PIK3CA, and 26% for p53. FGFR3 mutations were mutually exclusive with RAS mutations (p = 0.001) and co-occurred with PIK3CA mutations (p = 0.016). P53 overexpression was mutually exclusive with PIK3CA and FGFR3 mutations (p≤0.029). Mutations in the RAS and PIK3CA genes were not predictors for recurrence-free, progression-free and disease-specific survival. In patients presenting with NMI-BC grade 3 and MI-BC, 33 and 36% of the primary tumors were mutant. In patients with low-grade NMI-BC, 88% of the primary tumors carried a mutation and 88% of the recurrences were mutant. Conclusions/Significance The mutation assays present a companion diagnostic to define patients for targeted therapies. In addition, the assays are a potential biomarker to detect recurrences during surveillance. We showed that 88% of patients presenting with low-grade NMI-BC are eligible for such a follow-up. This may contribute to a reduction in the number of cystoscopical examinations.

Genome-wide Analysis of CpG Island Methylation in Bladder Cancer Identified TBX2, TBX3, GATA2, and ZIC4 as pTa-Specific Prognostic Markers

BACKGROUND DNA methylation markers could serve as useful biomarkers, both as markers for progression and for urine-based diagnostic assays. OBJECTIVE Identify bladder cancer (BCa)-specific methylated DNA sequences for predicting pTa-specific progression and detecting BCa in voided urine. DESIGN, SETTING, AND PARTICIPANTS Genome-wide methylation analysis was performed on 44 bladder tumours using the Agilent 244K Human CpG Island Microarray (Agilent Technologies, Santa Clara, CA, USA). Validation was done using a custom Illumina 384-plex assay (Illumina, San Diego, CA, USA) in a retrospective group of 77 independent tumours. Markers for progression were identified in pTa (n = 24) tumours and validated retrospectively in an independent series of 41 pTa tumours by the SNaPshot method (Applied Biosystems, Foster City, CA, USA). MEASUREMENTS The percentage of methylation in tumour and urine samples was used to identify markers for detection and related to the end point of progression to muscle-invasive disease with Kaplan-Meier models and multivariate analysis. RESULTS AND LIMITATIONS In the validation set, methylation of the T-box 2 (TBX2), T-box 3 (TBX3), GATA binding protein 2 (GATA2), and Zic family member 4 (ZIC4) genes was associated with progression to muscle-invasive disease in pTa tumours (p = 0.003). Methylation of TBX2 alone showed a sensitivity of 100%, a specificity of 80%, a positive predictive value of 78%, and a negative predictive value of 100%, with an area under the curve of 0.96 (p<0.0001) for predicting progression. Multivariate analysis showed that methylation of TBX3 and GATA2 are independent predictors of progression when compared to clinicopathologic variables (p = 0.04 and p = 0.03, respectively). The predictive accuracy improved by 23% by adding methylation of TBX2, TBX3, and GATA2 to the European Organisation for Research and Treatment of Cancer risk scores. We further identified and validated 110 CpG islands (CGIs) that are differentially methylated between tumour cells and control urine. The limitation of this study is the small number of patients analysed for testing and validating the prognostic markers. CONCLUSIONS We have identified four methylation markers that predict progression in pTa tumours, thereby allowing stratification of patients for personalised follow-up. In addition, we identified CGIs that will enable detection of bladder tumours in voided urine.

The development of multiple bladder tumour recurrences in relation to the FGFR3 mutation status of the primary tumour.

Non‐muscle invasive bladder cancers (NMI‐BCs) represent 75% of bladder cancers upon presentation. After removal of the primary tumour by transurethral resection, multiple recurrences continue to develop in 70% of patients. Consequently, prolonged and costly surveillance by cystoscopy is required. Mutations in the FGFR3 oncogene are common in NMI‐BCs and are associated with a lower chance of progression to muscle‐invasive disease. Here we analysed the consistency of FGFR3 mutations in primary and recurrent tumours. This knowledge is of crucial importance if FGFR3 mutation analysis on urinary cells is to be used as an alternative for cystoscopical surveillance. To this end, we monitored the disease process and FGFR3 mutation status of primary and recurrent tumours in 118 patients with NMI‐BC. During median follow‐up of 8.8 years, these patients underwent 2133 cystoscopies and 80 patients developed 414 recurrences. FGFR3 mutations were equally prevalent in primary and recurrent tumours (63%). Patients can have different types of FGFR3 mutations in different tumours. Recurrence risk was not significantly different for patients with a mutant or wild‐type primary tumour. Recurrence rates varied widely between patients but were constant for a patient and were unrelated to FGFR3 status. In the mutant patient group, in contrast to the wild‐type group, recurrences continued to develop after 10 years. In 81% of the recurrences of patients with a mutant primary tumour, a mutation was found. Moreover, recurrences in this patient group were of lower stage and grade than those of patients with a wild‐type primary tumour (p < 0.001). These results suggest that surveillance by FGFR3 mutation analysis on voided urine in combination with a reduced cystoscopy frequency of patients presenting with an FGFR3 mutant tumour is worth investigating. Copyright

FGFR3 Mutation Analysis in Voided Urine Samples to Decrease Cystoscopies and Cost in Nonmuscle Invasive Bladder Cancer Surveillance: A Comparison of 3 Strategies

PURPOSE We determined whether FGFR3 mutation analysis of voided urine samples would be cost-effective to partly replace cystoscopy in the surveillance of patients treated for nonmuscle invasive urothelial carcinoma. MATERIALS AND METHODS In this decision analytical study we analyzed data on 70 Dutch patients with FGFR3 positive primary tumors and a median followup of 8.8 years. Surveillance strategies were compared in a Markov model. Modified surveillance consisted of FGFR3 mutation analysis of voided urine samples every 3 months, and cystoscopy at 3, 12 and 24 months. Standard surveillance was defined as cystoscopy every 3 months and minimal surveillance was defined as cystoscopy at 3, 12 and 24 months. Analysis was stratified for 3 risk profiles, including surveillance after 1) the primary tumor, 2) the first to third recurrence and 3) the fourth recurrence or more. Sensitivity analysis was performed to evaluate the impact of variations in cost, sensitivity and specificity. RESULTS The probability of no recurrence after 2 years of surveillance after a primary tumor was higher for modified surveillance than for standard and minimal surveillance, eg after primary tumors (95.7% vs 95.0% and 93.9%, respectively). The total cost of surveillance after the primary tumor was lower for minimal and modified surveillance (€2,254 and €2,558, respectively) than for standard surveillance (€5,861). Results were robust to changing inputs over plausible ranges and consistent for each of the 3 risk profiles. CONCLUSIONS Surveillance in which cystoscopy is partly replaced by FGFR3 mutation analysis of urine seems a safe, effective and cost-effective surveillance strategy. Further validation in larger cohorts is required.

Bladder cancer: novel molecular characteristics, diagnostic, and therapeutic implications.

Bladder cancer (BC) comes in two flavors: as non-muscle invasive (NMI) and as muscle invasive (MI) disease. These two subtypes differ in their genetic aberrations. In NMI-BC mutations in the FGFR3 oncogene are found with a frequency of 75%, whereas mutations in the TP53 tumor suppressor gene prevail in MI-BC. Mutations in the RAS genes occur in 15% of BC of all stages and are mutually exclusive with FGFR3 mutations. Mutations in the PIK3CA gene are found in about 13% and these almost exclusively co-occur with FGFR3 mutations. NMI-BC with FGFR3 mutations are genetically stable, but FGFR3 wild type NMI-BC and MI tumors are genetically unstable. In this paper, we discuss the use of these genetic aberrations in relation to recurrence, progression, surveillance, and therapeutic options. As of yet, there is no biomarker that can predict recurrences or the rate of recurrences when they occur. We show that FGFR3 mutations are associated with a decreased risk of progression, and a better survival both in BC and in upper urinary tract cancer. Microsatellite analysis (MA) in order to detect loss-of-heterozygosity can be used to detect recurrences in urinary cells of patients under surveillance. The results of a Dutch randomized trial show that consecutive positive MA results are a strong predictor for future recurrences. Using FGFR3 mutation analysis for those patients who have an FGFR3 mutant tumor will enhance performance of urine-based surveillance. Although FGFR3 mutations occur in only 20% of MI tumors, these tumors often have a high expression of the FGFR3 protein. This suggests that this receptor could present a target for adjuvant therapy in MI-BC. However, whether the FGFR3 pathway is active in these tumors remains to be established.

Selection of Microsatellite Markers for Bladder Cancer Diagnosis without the Need for Corresponding Blood

Microsatellite markers are used for loss-of-heterozygosity, allelic imbalance and clonality analyses in cancers. Usually, tumor DNA is compared to corresponding normal DNA. However, normal DNA is not always available and can display aberrant allele ratios due to copy number variations in the genome. Moreover, stutter peaks may complicate the analysis. To use microsatellite markers for diagnosis of recurrent bladder cancer, we aimed to select markers without stutter peaks and a constant ratio between alleles, thereby avoiding the need for a control DNA sample. We investigated 49 microsatellite markers with tri- and tetranucleotide repeats in regions commonly lost in bladder cancer. Based on analysis of 50 blood DNAs the 12 best performing markers were selected with few stutter peaks and a constant ratio between peaks heights. Per marker upper and lower cut off values for allele ratios were determined. LOH of the markers was observed in 59/104 tumor DNAs. We then determined the sensitivity of the marker panel for detection of recurrent bladder cancer by assaying 102 urine samples of these patients. Sensitivity was 63% when patients were stratified for LOH in their primary tumors. We demonstrate that up-front selection of microsatellite markers obliterates the need for a corresponding blood sample. For diagnosis of bladder cancer recurrences in urine this significantly reduces costs. Moreover, this approach facilitates retrospective analysis of archival tumor samples for allelic imbalance.

1104 FGFR3 mutation analysis in voided urine samples to decrease cystoscopies and cost in nonmuscle invasive bladder cancer surveillance: A comparison of 3 strategies

Purpose: We determined whether FGFR3 mutation analysis of voided urine samples would be cost-effective to partly replace cystoscopy in the surveillance of patients treated for nonmuscle invasive urothelial carcinoma.Materials and Methods: In this decision analytical study we analyzed data on 70 Dutch patients with FGFR3 positive primary tumors and a median followup of 8.8 years. Surveillance strategies were compared in a Markov model. Modified surveillance consisted of FGFR3 mutation analysis of voided urine samples every 3 months, and cystoscopy at 3, 12 and 24 months. Standard surveillance was defined as cystoscopy every 3 months and minimal surveillance was defined as cystoscopy at 3, 12 and 24 months. Analysis was stratified for 3 risk profiles, including surveillance after 1) the primary tumor, 2) the first to third recurrence and 3) the fourth recurrence or more. Sensitivity analysis was performed to evaluate the impact of variations in cost, sensitivity and specificity.Results: The probability of n...

Abstract 2259: Identifying causative genes and prognostic markers for bladder cancer

Proceedings: AACR 102nd Annual Meeting 2011‐‐ Apr 2‐6, 2011; Orlando, FL Upon presentation 70% of the bladder tumors are confined to the epithelial (stage pTa) and stromal layers (stage pT1) of the bladder. These non-muscle invasive tumors (NMI-BC) frequently recur and 10-15% will eventually invade the detrusor muscle and may cause metastases. The management of these tumors is characaterized by the difficult choice between transurethral resection (TURB) or total bladder removal by cystectomy. Our first objective was to identify DNA copy number changes as biomarkers predicting tumor progression in high grade NMI-BC to aid decision making. A set of 18 DNA samples isolated from fresh frozen high grade NMI-BC was hybridized on Illumina SNP arrays. Analysis of alterations was done with Nexus software. Hierarchical clustering of significantly differing SNPs showed that tumors that progressed and those that did not clustered separately. A comparison of the differential alterations between tumor groups resulted in many chromosomal locations encompassing chromosome 1p, 2-6, 8-10, 12p, 13q, 17q, 18p and 22q that were fit for inclusion in a validation assay. A custom Agilent 15K array was designed to include probes for these locations. We collected a retrospective series of 96 high grade NMI-BC paraffin embedded tumors with a median follow-up of 5 years. Half of these samples were from patients who developed a muscle-invasive tumor at a later time. Preliminary results show an interesting region on chromosome 8 with the best independent prognostic value. This region and the genes therein will be studied in more detail. This and other validated regions will be used to design a definitive prognostic assay. Complex losses on chromosome 9 characterize bladder tumors regardless of stage. We hypothesized that the complex rearrangements on this chromosome in BC are the results of attempts of the tumor cell to loose tumor suppressor genes whilst struggling to retain other (onco)genes. Therefore, the second aim was to identify pivotal tumor suppressor and/or oncogenes on chromosome 9 for development of bladder cancer, with the aim to exploit these genes or the pathways in which they function as potential targets for therapy for tumors from all stage categories. A custom Agilent 44K array was designed, containing probes for all genes on chromosome 9 (and 8p). DNA from a series of 144 tumors from different stages was hybridized to the arrays. Analysis of alterations was done with Nexus software. We found many regions of loss as expected but also a region that appeared amplified in most tumors. Several candidate regions were validated of their copy number status with specific MLPA assays. Two potential candidate genes from chromosome 9 were selected for mutational and functional analysis. In one of those genes we found several mutations affecting protein function, suggesting that this gene is a new tumor suppressor gene for bladder cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2259. doi:10.1158/1538-7445.AM2011-2259

Abstract 3012: Genome-wide analysis of CpG island methylation in bladder cancer identifies novel biomarkers for diagnosis, prediction of progression and survival

Proceedings: AACR 102nd Annual Meeting 2011‐‐ Apr 2‐6, 2011; Orlando, FL Cancers of the urinary bladder (BC) present as muscle-invasive (MI) or non-muscle invasive (NMI). Major problems with NMI-BC are that 70% of the tumors will recur and 10-20% will eventually progress to MI-BC. These patients require life-long surveillance by cystoscopy. Over 50% of patients with primary or secondary MI-BC die of their disease. Epigenetic modifications have been shown to contribute to the pathogenesis of various cancers including bladder cancer. Hypermethylation is often found in CGIs in the 5’ regions of genes overlapping promoter regions. Methylation of these CGIs negatively affects gene expression and consequently DNA modifications may serve as useful biomarkers, both for diagnostic and prognostic purposes. We aimed to identify DNA methylation markers as a prognostic tool to predict progression, survival and to enable detection of recurrent tumors in urine. In this study, we performed a genome wide screening for DNA methylation in different subtypes of bladder cancer with the aid of differential methylation hybridization (DMH) coupled with Agilent 244k human CpG island microarrays. We have found 731 significant probes to be more methylated in bladder tumors than in blood, which represents 392 unique CGIs. The adjacent CpG dinucleotides within a CGI were co-methylated in most of the significantly methylated CGIs. In contrast, CGIs neighboring a methylated island were usually not methylated. Extensive methylation indicative of a CGI methylator phenotype was observed in FGFR3 wild-type NMI-BC. Most de novo methylated genes in bladder cancer are known targets of repression by polycomb group proteins in embryonic stem cells. CGI markers for the detection of recurrences and prediction of progression and survival were validated in an independent set of 90 FFPE tumors on a 384-plex custom Illumina Golden Gate Methylation Assay (GGMA). We identified 110 CGIs that are differentially methylated between tumor cells and control urine and 20 of these were further validated for the detection of recurrent and primary tumors in voided urine. Methylation of 34 CGIs was significantly associated with progression, with 22/34 prediction no progression and 13 predicting progression. Methylation of 7/13 CGIs was inversely associated with survival, while 8 were positively correlated with survival. These results imply that therapies targeting DNA methylation in bladder tumors should not be used without further studies. In summary, DNA methylation markers were identified that will help to predict disease progression and help to stratify patients for personalized follow-up. In addition, a vast number of markers were selected to employ in urine-based tests in order to reduce cystoscopy frequency. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3012. doi:10.1158/1538-7445.AM2011-3012

Abstract 4897: Genome-wide analysis of CGIs in bladder cancer identifies novel epigenetic biomarkers for diagnosis and follow-up

Epigenetic modifications have been shown to contribute to the pathogenesis of various cancers including bladder cancer. Hypermethylation in cancer is often found in CGIs in the 5′ regions of genes. These DNA modifications may serve as useful biomarkers, both for diagnostic and prognostic purposes. Although the list of aberrantly methylated genes in cancer is expanding, only a few genes so far were investigated for their usefulness as tumor biomarkers for early diagnosis and risk assessment of bladder cancer. Thus, a large scale (genome wide) screening of aberrant methylation of CGIs is needed to identify bladder-specific epigenetic markers. In this study, we performed a genome wide screening for DNA methylation in different subtypes of bladder cancer using differential methylation hybridization (DMH) coupled with Agilent 244k human CpG island microarrays. We next validated the genome-wide data on a 384-plex custom Illumina Golden Gate Methylation Assay (GGMA), which is based on bisulfite conversion. We found and validated a large number of possible novel epigenetic biomarkers for early detection and prognosis of bladder cancer. We also observed that subgroups of bladder tumors cluster separately due to their different methylation patterns suggesting possible differences in pathogenesis. Our data clearly shown that the adjacent CpG loci within a CGI were co-methylated. In contrast, CGIs neighboring a methylated island were usually not methylated. The data showed no bias toward promoter methylation as the relative proportion of methylated islands in genes was similar to those encompassing promoter regions. We observed that 70% of the methylated genes are downregulated in bladder cancer. Moreover, gene body methylation also seems to be negatively correlated with gene expression. Interestingly, we found an overrepresentation of methylated genes in those genes that are the target of polycomb group protein complexes in embryonic stem cells. Further validation of selected CGIs on a separate group of bladder cancers on a custom Illumina Golden Gate Methylation assay (GGMA) allowed us to confirm 70 % of CGIs differentiating bladder cancers from normal cells, besides validating CGIs discriminating subgroups of bladder tumors. In conclusion, this is the first genome-wide study on bladder cancer which identified potential methylation target genes that provides putative biomarkers that may be correlated with disease course as well as insight into the pathogenesis of bladder cancer subtypes. Note: This abstract was not presented at the AACR 101st Annual Meeting 2010 because the presenter was unable to attend. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4897.