Epigenetic profiling demarcates molecular subtypes of muscle-invasive bladder cancer.

Muscle-invasive bladder cancer (MIBC) is a heterogeneous disease that often recurs despite aggressive treatment with neoadjuvant chemotherapy and (radical) cystectomy. Basal and luminal molecular subtypes have been identified that are linked to clinical characteristics and have differential sensitivities to chemotherapy. While it has been suggested that epigenetic mechanisms play a role in defining these subtypes, a thorough understanding of the biological mechanisms is lacking. This report details the first genome-wide analysis of histone methylation patterns of human primary bladder tumours by chromatin immunoprecipitations and next-generation sequencing (ChIP-seq). We profiled multiple histone marks: H3K27me3, a marker for repressed genes, and H3K4me1 and H3K4me3, which are indicators of active enhancers and active promoters. Integrated analysis of ChIP-seq data and RNA sequencing revealed that H3K4 mono-methylation demarcates MIBC subtypes, while no association was found for the other two histone modifications in relation to basal and luminal subtypes. Additionally, we identified differentially methylated H3K4me1 peaks in basal and luminal tumour samples, suggesting that active enhancers play a role in defining subtypes. Our study is the first analysis of histone modifications in primary bladder cancer tissue and provides an important resource for the bladder cancer community.

Mouse models in the era of large human tumour sequencing studies.

Cancer is a complex disease in which cells progressively accumulate mutations disrupting their cellular processes. A fraction of these mutations drive tumourigenesis by affecting oncogenes or tumour suppressor genes, but many mutations are passengers with no clear contribution to tumour development. The advancement of DNA and RNA sequencing technologies has enabled in-depth analysis of thousands of human tumours from various tissues to perform systematic characterization of their (epi)genomes and transcriptomes in order to identify (epi)genetic changes associated with cancer. Combined with considerable progress in algorithmic development, this expansion in scale has resulted in the identification of many cancer-associated mutations, genes and pathways that are considered to be potential drivers of tumour development. However, it remains challenging to systematically identify drivers affected by complex genomic rearrangements and drivers residing in non-coding regions of the genome or in complex amplicons or deletions of copy-number driven tumours. Furthermore, functional characterization is challenging in the human context due to the lack of genetically tractable experimental model systems in which the effects of mutations can be studied in the context of their tumour microenvironment. In this respect, mouse models of human cancer provide unique opportunities for pinpointing novel driver genes and their detailed characterization. In this review, we provide an overview of approaches for complementing human studies with data from mouse models. We also discuss state-of-the-art technological developments for cancer gene discovery and validation in mice.

Endogenous androgen receptor proteomic profiling reveals genomic subcomplex involved in prostate tumorigenesis.

Androgen receptor (AR) is a key player in prostate cancer development and progression. Here we applied immunoprecipitation mass spectrometry of endogenous AR in LNCaP cells to identify components of the AR transcriptional complex. In total, 66 known and novel AR interactors were identified in the presence of synthetic androgen, most of which were critical for AR-driven prostate cancer cell proliferation. A subset of AR interactors required for LNCaP proliferation were profiled using chromatin immunoprecipitation assays followed by sequencing, identifying distinct genomic subcomplexes of AR interaction partners. Interestingly, three major subgroups of genomic subcomplexes were identified, where selective gain of function for AR genomic action in tumorigenesis was found, dictated by FOXA1 and HOXB13. In summary, by combining proteomic and genomic approaches we reveal subclasses of AR transcriptional complexes, differentiating normal AR behavior from the oncogenic state. In this process, the expression of AR interactors has key roles by reprogramming the AR cistrome and interactome in a genomic location-specific manner.

Towards a global cancer knowledge network: dissecting the current international cancer genomic sequencing landscape.

BACKGROUND: While next generation sequencing has enhanced our understanding of the biological basis of malignancy, current knowledge on global practices for sequencing cancer samples is limited. To address this deficiency, we developed a survey to provide a snapshot of current sequencing activities globally, identify barriers to data sharing and use this information to develop sustainable solutions for the cancer research community. METHODS: A multi-item survey was conducted assessing demographics, clinical data collection, genomic platforms, privacy/ethics concerns, funding sources and data sharing barriers for sequencing initiatives globally. Additionally, respondents were asked as to provide the primary intent of their initiative (clinical diagnostic, research or combination). RESULTS: Of 107 initiatives invited to participate, 59 responded (response rate = 55%). Whole exome sequencing (P = 0.03) and whole genome sequencing (P = 0.01) were utilized less frequently in clinical diagnostic than in research initiatives. Procedures to identify cancer-specific variants were heterogeneous, with bioinformatics pipelines employing different mutation calling/variant annotation algorithms. Measurement of treatment efficacy varied amongst initiatives, with time on treatment (57%) and RECIST (53%) being the most common; however, other parameters were also employed. Whilst 72% of initiatives indicated data sharing, its scope varied, with a number of restrictions in place (e.g. transfer of raw data). The largest perceived barriers to data harmonization were the lack of financial support (P < 0.01) and bioinformatics concerns (e.g. lack of interoperability) (P = 0.02). Capturing clinical data was more likely to be perceived as a barrier to data sharing by larger initiatives than by smaller initiatives (P = 0.01). CONCLUSIONS: These results identify the main barriers, as perceived by the cancer sequencing community, to effective sharing of cancer genomic and clinical data. They highlight the need for greater harmonization of technical, ethical and data capture processes in cancer sample sequencing worldwide, in order to support effective and responsible data sharing for the benefit of patients.

Targeting the RB-E2F pathway in breast cancer.

Mutations of the retinoblastoma tumor-suppressor gene (RB1) or components regulating the CDK-RB-E2F pathway have been identified in nearly every human malignancy. Re-establishing cell cycle control through cyclin-dependent kinase (CDK) inhibition has therefore emerged as an attractive option in the development of targeted cancer therapy. The most successful example of this today is the use of the CDK4/6 inhibitor palbociclib combined with aromatase inhibitors for the treatment of estrogen receptor-positive breast cancers. Multiple studies have demonstrated that the CDK-RB-E2F pathway is critical for the control of cell proliferation. More recently, studies have highlighted additional roles of this pathway, especially E2F transcription factors themselves, in tumor progression, angiogenesis and metastasis. Specific E2Fs also have prognostic value in breast cancer, independent of clinical parameters. We discuss here recent advances in understanding of the RB-E2F pathway in breast cancer. We also discuss the application of genome-wide genetic screening efforts to gain insight into synthetic lethal interactions of CDK4/6 inhibitors in breast cancer for the development of more effective combination therapies.

Breast cancer subtyping by immunohistochemistry and histological grade outperforms breast cancer intrinsic subtypes in predicting neoadjuvant chemotherapy response.

Intrinsic subtypes are widely accepted for the classification of breast cancer. Lacking gene expression data, surrogate classifications based on immunohistochemistry (IHC) have been proposed. A recent St. Gallen consensus meeting recommends to use this "surrogate intrinsic subtypes" for predicting adjuvant chemotherapy resistance, implying that "Surrogate Luminal A" breast cancers should only receive endocrine therapy. In this study we assessed both gene expression based intrinsic subtypes as well as surrogate intrinsic subtypes regarding their power to predict neoadjuvant chemotherapy benefit. Single institution data of 560 breast cancer patients were reviewed. Gene expression data was available for 247 patients. Subtypes were determined on the basis of IHC, Ki67, histological grade, endocrine responsiveness, and gene expression, and were correlated with chemotherapy response and recurrence-free survival. In ER+/HER2- tumors, a high histological grade was the best predictor for chemotherapy benefit, both in terms of pCR (p = 0.004) and recurrence-free survival (p = 0.002). The gene expression based and surrogate intrinsic subtype based on Ki67 had no predictive or prognostic value in ER+/HER2- tumors. Histological grade, ER, PR, and HER2 were the best predictive factors for chemotherapy response in breast cancer. We propose to continue the conventional use of these markers.

Prognostication and prediction using gene expression profiling in oesophageal cancer.

AIM: To evaluate current literature on gene expression profiling in oesophageal cancer. METHODS: We performed a review of the literature (2000-2010) on prognostication and prediction using gene expression analysis in oesophageal cancer. RESULTS: Seventeen papers comprising 638 patients were included. Gene expression profiles studied in relation to survival, lymph node metastasis and response to neoadjuvant therapy. Most studies included a limited number of patients. Several prognostic and predictive gene signatures were identified with different accuracies. In only one study, the gene signature was validated in a large, independent patient cohort. CONCLUSION: Gene expression profiling has potential clinical applications in oesophageal cancer. Especially a signature which is predictive for response to neoadjuvant treatment could be of great clinical value. To date, most published studies suffer from an underpowered training cohort or lack adequate validation. Clinicians should put effort in the collection of high quality tissue samples and should participate in biobank initiatives, considering the increasing availability and possibilities of sequencing technology.

A non-BRCA1/2 hereditary breast cancer sub-group defined by aCGH profiling of genetically related patients.

Germline mutations in BRCA1 and BRCA2 explain approximately 25% of all familial breast cancers. Despite intense efforts to find additional high-risk breast cancer genes (BRCAx) using linkage analysis, none have been reported thus far. Here we explore the hypothesis that BRCAx breast tumors from genetically related patients share a somatic genetic etiology that might be revealed by array comparative genomic hybridization (aCGH) profiling. As BRCA1 and BRCA2 tumors can be identified on the basis of specific genomic profiles, the same may be true for a subset of BRCAx families. Analyses used aCGH to compare 58 non-BRCA1/2 familial breast tumors (designated BRCAx) to sporadic (non-familiar) controls, BRCA1 and BRCA2 tumors. The selection criteria for BRCAx families included at least three cases of breast cancer diagnosed before the age of 60 in the family, and the absence of ovarian or male breast cancer. Hierarchical cluster analysis was performed to determine sub-groups within the BRCAx tumor class and family heterogeneity. Analysis of aCGH profiles of BRCAx tumors indicated that they constitute a heterogeneous class, but are distinct from both sporadic and BRCA1/2 tumors. The BRCAx class could be divided into sub-groups. One subgroup was characterized by a gain of chromosome 22. Tumors from family members were classified within the same sub-group in agreement with the hypothesis that tumors from the same family would harbor a similar genetic background. This approach provides a method to target a sub-group of BRCAx families for further linkage analysis studies.

Exome sequencing identifies frequent mutation of the SWI/SNF complex gene PBRM1 in renal carcinoma.

The genetics of renal cancer is dominated by inactivation of the VHL tumour suppressor gene in clear cell carcinoma (ccRCC), the commonest histological subtype. A recent large-scale screen of ∼3,500 genes by PCR-based exon re-sequencing identified several new cancer genes in ccRCC including UTX (also known as KDM6A), JARID1C (also known as KDM5C) and SETD2 (ref. 2). These genes encode enzymes that demethylate (UTX, JARID1C) or methylate (SETD2) key lysine residues of histone H3. Modification of the methylation state of these lysine residues of histone H3 regulates chromatin structure and is implicated in transcriptional control. However, together these mutations are present in fewer than 15% of ccRCC, suggesting the existence of additional, currently unidentified cancer genes. Here, we have sequenced the protein coding exome in a series of primary ccRCC and report the identification of the SWI/SNF chromatin remodelling complex gene PBRM1 (ref. 4) as a second major ccRCC cancer gene, with truncating mutations in 41% (92/227) of cases. These data further elucidate the somatic genetic architecture of ccRCC and emphasize the marked contribution of aberrant chromatin biology.

An aCGH classifier derived from BRCA1-mutated breast cancer and benefit of high-dose platinum-based chemotherapy in HER2-negative breast cancer patients.

BACKGROUND: Breast cancer cells deficient for BRCA1 are hypersensitive to agents inducing DNA double-strand breaks (DSB), such as bifunctional alkylators and platinum agents. Earlier, we had developed a comparative genomic hybridisation (CGH) classifier based on BRCA1-mutated breast cancers. We hypothesised that this BRCA1-like(CGH) classifier could also detect loss of function of BRCA1 due to other causes besides mutations and, consequently, might predict sensitivity to DSB-inducing agents. PATIENTS AND METHODS: We evaluated this classifier in stage III breast cancer patients, who had been randomly assigned between adjuvant high-dose platinum-based (HD-PB) chemotherapy, a DSB-inducing regimen, and conventional anthracycline-based chemotherapy. Additionally, we assessed BRCA1 loss through mutation or promoter methylation and immunohistochemical basal-like status in the triple-negative subgroup (TN subgroup). RESULTS: We observed greater benefit from HD-PB chemotherapy versus conventional chemotherapy among patients with BRCA1-like(CGH) tumours [41/230 = 18%, multivariate hazard ratio (HR) = 0.12, 95% confidence interval (CI) 0.04-0.43] compared with patients with non-BRCA1-like(CGH) tumours (189/230 = 82%, HR = 0.78, 95% CI 0.50-1.20), with a significant difference (test for interaction P = 0.006). Similar results were obtained for overall survival (P interaction = 0.04) and when analyses were restricted to the TN subgroup. Sixty-three percent (20/32) of assessable BRCA1-like(CGH) tumours harboured either a BRCA1 mutation (n = 8) or BRCA1 methylation (n = 12). CONCLUSION: BRCA1 loss as assessed by CGH analysis can identify patients with substantially improved outcome after adjuvant DSB-inducing chemotherapy when compared with standard anthracycline-based chemotherapy in our series.

Refinement of breast cancer classification by molecular characterization of histological special types.

Most invasive breast cancers are classified as invasive ductal carcinoma not otherwise specified (IDC NOS), whereas about 25% are defined as histological 'special types'. These special-type breast cancers are categorized into at least 17 discrete pathological entities; however, whether these also constitute discrete molecular entities remains to be determined. Current therapy decision-making is increasingly governed by the molecular classification of breast cancer (luminal, basal-like, HER2+). The molecular classification is derived from mainly IDC NOS and it is unknown whether this classification applies to all histological subtypes. We aimed to refine the breast cancer classification systems by analysing a series of 11 histological special types [invasive lobular carcinoma (ILC), tubular, mucinous A, mucinous B, neuroendocrine, apocrine, IDC with osteoclastic giant cells, micropapillary, adenoid cystic, metaplastic, and medullary carcinoma] using immunohistochemistry and genome-wide gene expression profiling. Hierarchical clustering analysis confirmed that some histological special types constitute discrete entities, such as micropapillary carcinoma, but also revealed that others, including tubular and lobular carcinoma, are very similar at the transcriptome level. When classified by expression profiling, IDC NOS and ILC contain all molecular breast cancer types (ie luminal, basal-like, HER2+), whereas histological special-type cancers, apart from apocrine carcinoma, are homogeneous and only belong to one molecular subtype. Our analysis also revealed that some special types associated with a good prognosis, such as medullary and adenoid cystic carcinomas, display a poor prognosis basal-like transcriptome, providing strong circumstantial evidence that basal-like cancers constitute a heterogeneous group. Taken together, our results imply that the correct classification of breast cancers of special histological type will allow a more accurate prognostication of breast cancer patients and facilitate the identification of optimal therapeutic strategies.

Combinatorial influence of environmental parameters on transcription factor activity.

MOTIVATION: Cells receive a wide variety of environmental signals, which are often processed combinatorially to generate specific genetic responses. Changes in transcript levels, as observed across different environmental conditions, can, to a large extent, be attributed to changes in the activity of transcription factors (TFs). However, in unraveling these transcription regulation networks, the actual environmental signals are often not incorporated into the model, simply because they have not been measured. The unquantified heterogeneity of the environmental parameters across microarray experiments frustrates regulatory network inference. RESULTS: We propose an inference algorithm that models the influence of environmental parameters on gene expression. The approach is based on a yeast microarray compendium of chemostat steady-state experiments. Chemostat cultivation enables the accurate control and measurement of many of the key cultivation parameters, such as nutrient concentrations, growth rate and temperature. The observed transcript levels are explained by inferring the activity of TFs in response to combinations of cultivation parameters. The interplay between activated enhancers and repressors that bind a gene promoter determine the possible up- or downregulation of the gene. The model is translated into a linear integer optimization problem. The resulting regulatory network identifies the combinatorial effects of environmental parameters on TF activity and gene expression. AVAILABILITY: The Matlab code is available from the authors upon request. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Creating gene set activity profiles with time-series expression data.

The use of predefined gene sets has become crucial in the interpretation of genomewide expression data. A limitation of the existing techniques that relate gene expression levels to gene sets is that they cannot readily be applied to time-course microarray data. The ability to attach statistical significance to the behaviour of biological processes over time would greatly contribute to understanding the complex regulatory mechanisms in the cell. We propose a statistical testing procedure based on the central limit theorem to assess the enrichment of a gene set. The technique is applied on time-course microarray data to generate gene-set specific 'activity profiles'.

No common denominator for breast cancer lymph node metastasis.

The axillary lymph node status is the most powerful prognostic factor for breast cancer patients to date. The molecular mechanisms that control lymph node metastasis, however, remain poorly understood. To define patterns of genes or gene regulatory pathways that drive breast cancer lymph node metastasis, we compared the gene expression profiles of 15 primary breast carcinomas and their matching lymph node metastases using microarrays. In general, primary breast carcinomas and lymph node metastases do not differ at the transcriptional level by a common subset of genes. No classifier or single gene discriminating the group of primary tumours from those of the lymph node metastases could be identified. Also, in a series of 295 breast tumours, no classifier predicting lymph node metastasis could be developed. However, subtle differences in the expression of genes involved in extracellular-matrix organisation and growth factor signalling are detected in individual pairs of matching primary and metastatic tumours. Surprisingly, however, different sets of these genes are either up- or downregulated in lymph node metastases. Our data suggest that breast carcinomas do not use a shared gene set to accomplish lymph node metastasis.

DNA microarrays for comparison of gene expression profiles between diagnosis and relapse in precursor-B acute lymphoblastic leukemia: choice of technique and purification influence the identification of potential diagnostic markers.

Microarrays for gene expression profiling are rapidly becoming important research tools for the identification of novel markers, for example, for novel classification of leukemias and lymphomas. Here, we review the considerations and infrastructure for microarray experiments. These considerations are illustrated via a microarray-based comparison of gene expression profiles of paired diagnosis-relapse samples from patients with precursor-B acute lymphoblastic leukemia (ALL), who relapsed during therapy or after completion of treatment. Initial experiments showed that several seemingly differentially expressed genes were actually derived from contaminating non-leukemic cells, particularly myeloid cells and T-lymphocytes. Therefore, we purified the ALL cells of the diagnosis and relapse samples if their frequency was lower than 95%. Furthermore, we observed in earlier studies that extra RNA amplification leads to skewing of particular gene transcripts. Sufficient (non-amplified) RNA of purified and paired diagnosis-relapse samples was obtained from only seven cases. The gene expression profiles were evaluated with Affymetrix U95A chips containing 12 600 human genes. These diagnosis-relapse comparisons revealed only a small number of genes (n=6) that differed significantly in expression: mostly signaling molecules and transcription factors involved in cell proliferation and cell survival were highly upregulated at relapse, but we did not observe any increase in drug-resistance markers. This finding fits with the observation that tumors with a high proliferation index have a poor prognosis. The genes that changed between diagnosis and relapse are currently not in use as diagnostic or disease progression markers, but represent potential new markers for such applications. Leukemia (2003) 17, 1324-1332. doi:10.1038/sj.leu.2402974

Genetic network modeling.

The inference of genetic interactions from measured expression data is one of the most challenging tasks of modern functional genomics. When successful, the learned network of regulatory interactions yields a wealth of useful information. An inferred genetic network contains information about the pathway to which a gene belongs and which genes it interacts with. Furthermore, it explains the function of the gene in terms of how it influences other genes and indicates which genes are pathway initiators and therefore potential drug targets. Obviously, such wealth comes at a price and that of genetic network modeling is that it is an extremely complex task. Therefore, it is necessary to develop sophisticated computational tools that are able to extract relevant information from a limited set of microarray measurements and integrate this with different information sources, to come up with reliable hypotheses of a genetic regulatory network. Thus far, a multitude of modeling approaches have been proposed for discovering genetic networks. However, it is unclear what the advantages and disadvantages of each of the different approaches are and how their results can be compared. In this review, genetic network models are put in a historical perspective that explains why certain models were introduced. Various modeling assumptions and their consequences are also highlighted. In addition, an overview of the principal differences and similarities between the approaches is given by considering the qualitative properties of the chosen models and their learning strategies.