François Radvanyi

Analysis of array CGH data: from signal ratio to gain and loss of DNA regions

MOTIVATION Genomic DNA regions are frequently lost or gained during tumor progression. Array Comparative Genomic Hybridization (array CGH) technology makes it possible to assess these changes in DNA in cancers, by comparison with a normal reference. The identification of systematically deleted or amplified genomic regions in a set of tumors enables biologists to identify genes involved in cancer progression because tumor suppressor genes are thought to be located in lost genomic regions and oncogenes, in gained regions. Array CGH profiles should also improve the classification of tumors. The achievement of these goals requires a methodology for detecting the breakpoints delimiting altered regions in genomic patterns and assigning a status (normal, gained or lost) to each chromosomal region. RESULTS We have developed a methodology for the automatic detection of breakpoints from array CGH profile, and the assignment of a status to each chromosomal region. The breakpoint detection step is based on the Adaptive Weights Smoothing (AWS) procedure and provides highly convincing results: our algorithm detects 97, 100 and 94% of breakpoints in simulated data, karyotyping results and manually analyzed profiles, respectively. The percentage of correctly assigned statuses ranges from 98.9 to 99.8% for simulated data and is 100% for karyotyping results. Our algorithm also outperforms other solutions on a public reference dataset. AVAILABILITY The R package GLAD (Gain and Loss Analysis of DNA) is available upon request.

Frequent FGFR3 mutations in papillary non-invasive bladder (pTa) tumors.

We recently identified activating mutations of fibroblast growth factor receptor 3 (FGFR3) in bladder carcinoma. In this study we assessed the incidence of FGFR3 mutations in a series of 132 bladder carcinomas: 20 carcinoma in situ (CIS), 50 pTa, 19 pT1, and 43 pT2-4. All 48 mutations identified were identical to the germinal activating mutations that cause thanatophoric dysplasia, a lethal form of dwarfism. The S249C mutation, found in 33 of the 48 mutated tumors, was the most common. The frequency of mutations was higher in pTa tumors (37 of 50, 74%) than in CIS (0 of 20, 0%; P < 0.0001), pT1 (4 of 19, 21%; P < 0.0001) and pT2-4 tumors (7 of 43, 16%; P < 0.0001). FGFR3 mutations were detected in 27 of 32 (84%) G1, 16 of 29 (55%) G2, and 5 of 71 (7%) G3 tumors. This association between FGFR3 mutations and low grade was highly significant (P < 0.0001). FGFR3 is the first gene found to be mutated at a high frequency in pTa tumors. The absence of FGFR3 mutations in CIS and the low frequency of FGFR3 mutations in pT1 and pT2-4 tumors are consistent with the model of bladder tumor progression in which the most common precursor of pT1 and pT2-4 tumors is CIS.

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

PURPOSE Fibroblast growth factor receptor 3 (FGFR3) mutations were recently found at a high frequency in well-differentiated urothelial cell carcinoma (UCC). We investigated the relationship between FGFR3 status and three molecular markers (MIB-1, P53, and P27kip1) associated with worse prognosis and determined the reproducibility of pathologic grade and molecular variables. PATIENTS AND METHODS In this multicenter study, we included 286 patients with primary (first diagnosis) UCC. The histologic slides were reviewed. FGFR3 status was examined by polymerase chain reaction-single strand conformation polymorphism and sequencing. Expression levels of MIB-1, P53, and P27kip1 were determined by immunohistochemistry. Mean follow-up was 5.5 years (range, 0.4 to 18.4 years). RESULTS FGFR3 mutations were detected in 172 (60%) of 286 UCCs. Grade 1 tumors had an FGFR3 mutation in 88% of patient samples and grade 3 tumors in 16% of patient samples. Conversely, aberrant expression patterns of MIB-1, P53, and P27kip1 were seen in 5%, 2%, and 3% of grade 1 tumors and in 85%, 60%, and 56% of grade 3 tumors, respectively. In multivariate analysis with recurrence rate, progression, and disease-specific survival as end points, the combination of FGFR3 and MIB-1 proved independently significant in all three cases. By using these two molecular markers, three molecular grades (mGs) could be identified: mG1 (mutation; normal expression), favorable prognosis; mG2 (two remaining combinations), intermediate prognosis; and mG3 (no mutation; high expression), poor prognosis. The molecular variables were more reproducible than pathologic grade (85% to 100% v 47% to 61%). CONCLUSION The FGFR3 mutation represents the favorable molecular pathway of UCC. Molecular grading provides a new, simple, and highly reproducible tool for clinical decision making in UCC patients.

Regional copy number–independent deregulation of transcription in cancer

Genetic and epigenetic alterations have been identified that lead to transcriptional deregulation in cancers. Genetic mechanisms may affect single genes or regions containing several neighboring genes, as has been shown for DNA copy number changes. It was recently reported that epigenetic suppression of gene expression can also extend to a whole region; this is known as long-range epigenetic silencing. Various techniques are available for identifying regional genetic alterations, but no large-scale analysis has yet been carried out to obtain an overview of regional epigenetic alterations. We carried out an exhaustive search for regions susceptible to such mechanisms using a combination of transcriptome correlation map analysis and array CGH data for a series of bladder carcinomas. We validated one candidate region experimentally, demonstrating histone methylation leading to the loss of expression of neighboring genes without DNA methylation.

Gene Expression Signatures Predict Outcome in Non–Muscle-Invasive Bladder Carcinoma: A Multicenter Validation Study

Purpose: Clinically useful molecular markers predicting the clinical course of patients diagnosed with non–muscle-invasive bladder cancer are needed to improve treatment outcome. Here, we validated four previously reported gene expression signatures for molecular diagnosis of disease stage and carcinoma in situ (CIS) and for predicting disease recurrence and progression. Experimental Design: We analyzed tumors from 404 patients diagnosed with bladder cancer in hospitals in Denmark, Sweden, England, Spain, and France using custom microarrays. Molecular classifications were compared with pathologic diagnosis and clinical outcome. Results: Classification of disease stage using a 52-gene classifier was found to be highly significantly correlated with pathologic stage (P < 0.001). Furthermore, the classifier added information regarding disease progression of Ta or T1 tumors (P < 0.001). The molecular 88-gene progression classifier was highly significantly correlated with progression-free survival (P < 0.001) and cancer-specific survival (P = 0.001). Multivariate Cox regression analysis showed the progression classifier to be an independently significant variable associated with disease progression after adjustment for age, sex, stage, grade, and treatment (hazard ratio, 2.3; P = 0.007). The diagnosis of CIS using a 68-gene classifier showed a highly significant correlation with histopathologic CIS diagnosis (odds ratio, 5.8; P < 0.001) in multivariate logistic regression analysis. Conclusion: This multicenter validation study confirms in an independent series the clinical utility of molecular classifiers to predict the outcome of patients initially diagnosed with non–muscle-invasive bladder cancer. This information may be useful to better guide patient treatment.

Assessing HER2 gene amplification as a potential target for therapy in invasive urothelial bladder cancer with a standardized methodology: results in 1005 patients

Background: This study assessed the human epidermal growth factor receptor-2 (HER2) protein expression and its relationship with gene amplification in invasive bladder carcinoma, using the same criteria than for breast cancer. Patients and methods: In 1005 patients, paraffin-embedded tissues of transurethral resection or cystectomy were evaluated by immunohistochemistry (IHC), using antibodies against HER2. All samples with a 2+ or 3+ HER2 overexpression were evaluated by FISH. Results: HER2 overexpression was observed in 93 (9.2%) tumors (2+: 42 tumors and 3+: 51 tumors). Using FISH, all HER2 3+ tumors had a gene amplification, whereas no amplification was found in 2+ tumors. Intratumoral heterogeneity was observed in 35% of cases. These tumors showed the same heterogeneous pattern, with adjacent 3+ positive and negative areas by both IHC and FISH. Conclusions: This study showed that 5.1% of invasive bladder carcinomas had a HER2 gene amplification. These findings may have clinical implications for the management of patients with HER2-positive locally advanced or metastatic bladder cancer, as they could be potential candidates for targeted therapy.

Molecular Grade (FGFR3/MIB-1) and EORTC Risk Scores Are Predictive in Primary Non–Muscle-Invasive Bladder Cancer

BACKGROUND The European Organization for Research and Treatment of Cancer (EORTC) risk scores are not validated in an independent patient population. Molecular grade (mG) based on fibroblast growth factor receptor 3 (FGFR3) gene mutation status and MIB-1 expression was proposed as an alternative to pathologic grade in bladder cancer (BCa) [1]. OBJECTIVE To validate the EORTC risk score and to determine its relation to mG in a series with long-term follow-up as well as to determine reproducibility of pathologic grade and mG. DESIGN, SETTING, AND PARTICIPANTS In this multicenter study, we included 230 patients with primary non-muscle-invasive BCa (NMIBC). MEASUREMENTS Four uropathologists reviewed the slides. FGFR3 mutation status was examined by two assays. MIB-1 was assessed by immunohistochemistry. The EORTC risk scores for recurrence and progression were determined. Multivariable analyses were used to find prognostic factors. RESULTS AND LIMITATIONS Median follow-up was 8.62 yr (interquartile range: 6.6-11.8). FGFR3 mutations were significantly related to favorable disease parameters, whereas altered MIB-1 was frequently seen with pT1, high grade, and high EORTC risk scores. EORTC risk scores were significant in multivariable analyses for recurrence and progression. In multivariable analyses for progression and disease-specific survival, the mG had independent significance. The addition of mG to the multivariable model for progression increased the predictive accuracy from 74.9% to 81.7% (p<0.001; Mantel-Haenszel test). The mG (89%) was more reproducible than the pathologic grade (41-74%). CONCLUSIONS We validated the EORTC risk scores for primary NMIBC in a clinical and biomarker setting. Next to EORTC risk score, mG proved highly reproducible and predictive. Our long-term results justify an independent prospective analysis of mG and EORTC risk scores.

Novel fibroblast growth factor receptor 3 (FGFR3) mutations in bladder cancer previously identified in non-lethal skeletal disorders

Activating mutations in the fibroblast growth factor receptor 3 (FGFR3) gene are responsible for several autosomal dominant craniosynostosis syndromes and chondrodysplasias i.e. hypochondroplasia, achondroplasia, SADDAN and thanatophoric dysplasia – a neonatal lethal dwarfism syndrome. Recently, activating FGFR3 mutations have also been found to be present in cancer, i.e. at high frequency in carcinoma of the bladder and rarely in multiple myeloma and carcinoma of the cervix. Almost all reported mutations in carcinomas corresponded to the mutations identified in thanatophoric dysplasia. We here screened a series of 297 bladder tumours and found three FGFR3 somatic mutations (G380/382R; K650/652M and K650/652T) that were not previously identified in carcinomas or thanatophoric dysplasia. Another novel finding was the occurrence of two simultaneous FGFR3 mutations in four tumours. Two of the three new mutations in bladder cancer, the G380/382R and the K650/652M mutations, were previously reported in achondroplasia and SADDAN, respectively. These syndromes entail a longer life span than thanatophoric dysplasia. The K650/652T mutation has not previously been detected in patients with skeletal disorders, but affects a codon that has been shown to be affected in some cases of thanatophoric dysplasia, SADDAN and hypochondroplasia. From a clinical perspective, the patients with FGFR3-related, non-lethal skeletal disorders might be at a higher risk for development of bladder tumours than the general population.

Differences in steroid 5α-reductase iso-enzymes expression between normal and pathological human prostate tissue

We studied the expression level and cell-specific expression patterns of 5alpha-reductase (5alpha-R) types 1 and 2 iso-enzymes in human hyperplastic and malignant prostate tissue by semi-quantitative RT-PCR and in situ hybridisation analyses. In situ hybridisation established that 5alpha-R1 mRNA is preferentially expressed by epithelial cells and little expressed by stromal cells whereas 5alpha-R2 mRNA is expressed by both epithelium and stroma. Semi-quantitative RT-PCR has been performed on total RNA from different zones of normal prostate, BPH tissues and liver. We found that 5alpha-R1 and 5alpha-R2 mRNAs expression was near the same in all zones of normal prostate. In BPH tissue, 5alpha-R1 and 5alpha-R2 mRNAs expression was slightly but significantly increased, when it was compared to the levels recorded for normal prostate. In cancer samples, 5alpha-R1 mRNA expression was higher than in normal and hyperplastic prostate but the level of 5alpha-R2 mRNA was not statistically different from that observed in the different zones of normal prostate. In liver, 5alpha-R2 mRNA level was similar to that measured in BPH but 5alpha-R1 mRNA expression was ten times higher. The increase observed in 5alpha-R isoenzymes expression in BPH tissue could play an important role in the pathogenesis and/or maintenance of the disease.

Integrated Genomic and Transcriptomic Analysis of Ductal Carcinoma In situ of the Breast

Purpose: To gain insight into genomic and trancriptomic subtypes of ductal carcinomas in situ of the breast (DCIS). Experimental Design: We did a combined phenotypic and genomic analysis of a series of 57 DCIS integrated with gene expression profile analysis for 26 of the 57 cases. Results: Thirty-two DCIS exhibited a luminal phenotype; 21 were ERBB2 positive, and 4 were ERBB2/estrogen receptor (ER) negative with 1 harboring a bona fide basal-like phenotype. Based on a CGH analysis, genomic types were identified in this series of DCIS with the 1q gain/16q loss combination observed in 3 luminal DCIS, the mixed amplifier pattern including all ERBB2, 12 luminal and 2 ERBB2-/ER- DCIS, and the complex copy number alteration profile encompassing 14 luminal and 1 ERBB2-/ER- DCIS. Eight cases (8 of 57; 14%) presented a TP53 mutation, all being amplifiers. Unsupervised analysis of gene expression profiles of 26 of the 57 DCIS showed that luminal and ERBB2-amplified, ER-negative cases clustered separately. We further investigated the effect of high and low copy number changes on gene expression. Strikingly, amplicons but also low copy number changes especially on 1q, 8q, and 16q in DCIS regulated the expression of a subset of genes in a very similar way to that recently described in invasive ductal carcinomas. Conclusions: These combined approaches show that the molecular heterogeneity of breast ductal carcinomas exists already in in situ lesions and further indicate that DCIS and invasive ductal carcinomas share genomic alterations with a similar effect on gene expression profile.