DISP1 deficiency: Monoallelic and biallelic variants cause a spectrum of midline craniofacial malformations.

PURPOSE: DISP1 encodes a transmembrane protein that regulates the secretion of the morphogen, Sonic hedgehog, a deficiency of which is a major cause of holoprosencephaly (HPE). This disorder covers a spectrum of brain and midline craniofacial malformations. The objective of the present study was to better delineate the clinical phenotypes associated with division transporter dispatched-1 (DISP1) variants. METHODS: This study was based on the identification of at least 1 pathogenic variant of the DISP1 gene in individuals for whom detailed clinical data were available. RESULTS: A total of 23 DISP1 variants were identified in heterozygous, compound heterozygous or homozygous states in 25 individuals with midline craniofacial defects. Most cases were minor forms of HPE, with craniofacial features such as orofacial cleft, solitary median maxillary central incisor, and congenital nasal pyriform aperture stenosis. These individuals had either monoallelic loss-of-function variants or biallelic missense variants in DISP1. In individuals with severe HPE, the DISP1 variants were commonly found associated with a variant in another HPE-linked gene (ie, oligogenic inheritance). CONCLUSION: The genetic findings we have acquired demonstrate a significant involvement of DISP1 variants in the phenotypic spectrum of midline defects. This underlines its importance as a crucial element in the efficient secretion of Sonic hedgehog. We also demonstrated that the very rare solitary median maxillary central incisor and congenital nasal pyriform aperture stenosis combination is part of the DISP1-related phenotype. The present study highlights the clinical risks to be flagged up during genetic counseling after the discovery of a pathogenic DISP1 variant.

Deep Learning for Detecting BRCA Mutations in High-Grade Ovarian Cancer Based on an Innovative Tumor Segmentation Method From Whole Slide Images.

BRCA1 and BRCA2 genes play a crucial role in repairing DNA double-strand breaks through homologous recombination. Their mutations represent a significant proportion of homologous recombination deficiency and are a reliable effective predictor of sensitivity of high-grade ovarian cancer (HGOC) to poly(ADP-ribose) polymerase inhibitors. However, their testing by next-generation sequencing is costly and time-consuming and can be affected by various preanalytical factors. In this study, we present a deep learning classifier for BRCA mutational status prediction from hematoxylin-eosin-safran-stained whole slide images (WSI) of HGOC. We constituted the OvarIA cohort composed of 867 patients with HGOC with known BRCA somatic mutational status from 2 different pathology departments. We first developed a tumor segmentation model according to dynamic sampling and then trained a visual representation encoder with momentum contrastive learning on the predicted tumor tiles. We finally trained a BRCA classifier on more than a million tumor tiles in multiple instance learning with an attention-based mechanism. The tumor segmentation model trained on 8 WSI obtained a dice score of 0.915 and an intersection-over-union score of 0.847 on a test set of 50 WSI, while the BRCA classifier achieved the state-of-the-art area under the receiver operating characteristic curve of 0.739 in 5-fold cross-validation and 0.681 on the testing set. An additional multiscale approach indicates that the relevant information for predicting BRCA mutations is located more in the tumor context than in the cell morphology. Our results suggest that BRCA somatic mutations have a discernible phenotypic effect that could be detected by deep learning and could be used as a prescreening tool in the future.

Synonymous variants in holoprosencephaly alter codon usage and impact the Sonic Hedgehog protein.

Synonymous single nucleotide variants (sSNVs) have been implicated in various genetic disorders through alterations of pre-mRNA splicing, mRNA structure and miRNA regulation. However, their impact on synonymous codon usage and protein translation remains to be elucidated in clinical context. Here, we explore the functional impact of sSNVs in the Sonic Hedgehog (SHH) gene, identified in patients affected by holoprosencephaly, a congenital brain defect resulting from incomplete forebrain cleavage. We identified eight sSNVs in SHH, selectively enriched in holoprosencephaly patients as compared to healthy individuals, and systematically assessed their effect at both transcriptional and translational levels using a series of in silico and in vitro approaches. Although no evidence of impact of these sSNVs on splicing, mRNA structure or miRNA regulation was found, five sSNVs introduced significant changes in codon usage and were predicted to impact protein translation. Cell assays demonstrated that these five sSNVs are associated with a significantly reduced amount of the resulting protein, ranging from 5% to 23%. Inhibition of the proteasome rescued the protein levels for four out of five sSNVs, confirming their impact on protein stability and folding. Remarkably, we found a significant correlation between experimental values of protein reduction and computational measures of codon usage, indicating the relevance of in silico models in predicting the impact of sSNVs on translation. Considering the critical role of SHH in brain development, our findings highlight the clinical relevance of sSNVs in holoprosencephaly and underline the importance of investigating their impact on translation in human pathologies.

A framework for validating AI in precision medicine: considerations from the European ITFoC consortium.

BACKGROUND: Artificial intelligence (AI) has the potential to transform our healthcare systems significantly. New AI technologies based on machine learning approaches should play a key role in clinical decision-making in the future. However, their implementation in health care settings remains limited, mostly due to a lack of robust validation procedures. There is a need to develop reliable assessment frameworks for the clinical validation of AI. We present here an approach for assessing AI for predicting treatment response in triple-negative breast cancer (TNBC), using real-world data and molecular -omics data from clinical data warehouses and biobanks. METHODS: The European "ITFoC (Information Technology for the Future Of Cancer)" consortium designed a framework for the clinical validation of AI technologies for predicting treatment response in oncology. RESULTS: This framework is based on seven key steps specifying: (1) the intended use of AI, (2) the target population, (3) the timing of AI evaluation, (4) the datasets used for evaluation, (5) the procedures used for ensuring data safety (including data quality, privacy and security), (6) the metrics used for measuring performance, and (7) the procedures used to ensure that the AI is explainable. This framework forms the basis of a validation platform that we are building for the "ITFoC Challenge". This community-wide competition will make it possible to assess and compare AI algorithms for predicting the response to TNBC treatments with external real-world datasets. CONCLUSIONS: The predictive performance and safety of AI technologies must be assessed in a robust, unbiased and transparent manner before their implementation in healthcare settings. We believe that the consideration of the ITFoC consortium will contribute to the safe transfer and implementation of AI in clinical settings, in the context of precision oncology and personalized care.

Integrated clinical and omics approach to rare diseases: novel genes and oligogenic inheritance in holoprosencephaly.

Holoprosencephaly is a pathology of forebrain development characterized by high phenotypic heterogeneity. The disease presents with various clinical manifestations at the cerebral or facial levels. Several genes have been implicated in holoprosencephaly but its genetic basis remains unclear: different transmission patterns have been described including autosomal dominant, recessive and digenic inheritance. Conventional molecular testing approaches result in a very low diagnostic yield and most cases remain unsolved. In our study, we address the possibility that genetically unsolved cases of holoprosencephaly present an oligogenic origin and result from combined inherited mutations in several genes. Twenty-six unrelated families, for whom no genetic cause of holoprosencephaly could be identified in clinical settings [whole exome sequencing and comparative genomic hybridization (CGH)-array analyses], were reanalysed under the hypothesis of oligogenic inheritance. Standard variant analysis was improved with a gene prioritization strategy based on clinical ontologies and gene co-expression networks. Clinical phenotyping and exploration of cross-species similarities were further performed on a family-by-family basis. Statistical validation was performed on 248 ancestrally similar control trios provided by the Genome of the Netherlands project and on 574 ancestrally matched controls provided by the French Exome Project. Variants of clinical interest were identified in 180 genes significantly associated with key pathways of forebrain development including sonic hedgehog (SHH) and primary cilia. Oligogenic events were observed in 10 families and involved both known and novel holoprosencephaly genes including recurrently mutated FAT1, NDST1, COL2A1 and SCUBE2. The incidence of oligogenic combinations was significantly higher in holoprosencephaly patients compared to two control populations (P < 10-9). We also show that depending on the affected genes, patients present with particular clinical features. This study reports novel disease genes and supports oligogenicity as clinically relevant model in holoprosencephaly. It also highlights key roles of SHH signalling and primary cilia in forebrain development. We hypothesize that distinction between different clinical manifestations of holoprosencephaly lies in the degree of overall functional impact on SHH signalling. Finally, we underline that integrating clinical phenotyping in genetic studies is a powerful tool to specify the clinical relevance of certain mutations.

Implementation of a molecular tumor board at a regional level to improve access to targeted therapy.

BACKGROUND: With the development of precision oncology, Molecular Tumor Boards (MTB) are developing in many institutions. However, the implementation of MTB in routine clinical practice has still not been thoroughly studied. MATERIAL AND METHODS: Since the first drugs approved for targeted therapies, patient tumor samples were centralized to genomic testing platforms. In our institution, all tumor samples have been analyzed since 2014 by Next Generation Sequencing (NGS). In 2015, we established a regional MTB to discuss patient cases with 1 or more alterations identified by NGS, in genes different from those related to drug approval. We conducted a retrospective comparative analysis to study whether our MTB increased the prescriptions of Molecular Targeted Therapies (MTT) and the inclusions of patients in clinical trials with MTT, in comparison with patients with available NGS data but no MTB discussion. RESULTS: In 2014, 86 patients had UGA, but the results were not available to clinicians and not discussed in MTB. During the years 2015 and 2016, 113 patients with an UGA (unreferenced genomic alteration) were discussed in MTB. No patients with an UGA were included in 2014 in a clinical trial, versus 2 (2%) in 2015-2016. 13 patients with an UGA (12%) were treated in 2015-2016 with a MTT whereas in 2014, no patient (p = 0.001). CONCLUSIONS: In this retrospective analysis, we showed that the association of large-scale genomic testing and MTB was feasible, and could increase the prescription of MTT. However, in routine clinical practice, the majority of patients with UGA still do not have access to MTT.

Integration of transcriptome and proteome profiles in glioblastoma: looking for the missing link.

BACKGROUND: Glioblastoma (GB) is the most common and aggressive tumor of the brain. Genotype-based approaches and independent analyses of the transcriptome or the proteome have led to progress in understanding the underlying biology of GB. Joint transcriptome and proteome profiling may reveal new biological insights, and identify pathogenic mechanisms or therapeutic targets for GB therapy. We present a comparison of transcriptome and proteome data from five GB biopsies (TZ) vs their corresponding peritumoral brain zone (PBZ). Omic analyses were performed using RNA microarray chips and the isotope-coded protein label method (ICPL). RESULTS: As described in other cancers, we found a poor correlation between transcriptome and proteome data in GB. We observed only two commonly deregulated mRNAs/proteins (neurofilament light polypeptide and synapsin 1) and 12 altered biological processes; they are related to cell communication, synaptic transmission and nervous system processes. This poor correlation may be a consequence of the techniques used to produce the omic profiles, the intrinsic properties of mRNA and proteins and/or of cancer- or GB-specific phenomena. Of interest, the analysis of the transcription factor binding sites present upstream from the open reading frames of all altered proteins identified by ICPL method shows a common binding site for the topoisomerase I and p53-binding protein TOPORS. Its expression was observed in 7/11 TZ samples and not in PBZ. Some findings suggest that TOPORS may function as a tumor suppressor; its implication in gliomagenesis should be examined in future studies. CONCLUSIONS: In this study, we showed a low correlation between transcriptome and proteome data for GB samples as described in other cancer tissues. We observed that NEFL, SYN1 and 12 biological processes were deregulated in both the transcriptome and proteome data. It will be important to analyze more specifically these processes and these two proteins to allow the identification of new theranostic markers or potential therapeutic targets for GB.

Recent advances in understanding inheritance of holoprosencephaly.

Holoprosencephaly (HPE) is a complex genetic disorder of the developing forebrain characterized by high phenotypic and genetic heterogeneity. HPE was initially defined as an autosomal dominant disease, but recent research has shown that its mode of transmission is more complex. The past decade has witnessed rapid development of novel genetic technologies and significant progresses in clinical studies of HPE. In this review, we recapitulate genetic epidemiological studies of the largest European HPE cohort and summarize the novel genetic discoveries of HPE based on recently developed diagnostic methods. Our main purpose is to present different inheritance patterns that exist for HPE with a particular emphasis on oligogenic inheritance and its implications in genetic counseling.

Mutational Spectrum in Holoprosencephaly Shows That FGF is a New Major Signaling Pathway.

Holoprosencephaly (HPE) is the most common congenital cerebral malformation in humans, characterized by impaired forebrain cleavage and midline facial anomalies. It presents a high heterogeneity, both in clinics and genetics. We have developed a novel targeted next-generation sequencing (NGS) assay and screened a cohort of 257 HPE patients. Mutations with high confidence in their deleterious effect were identified in approximately 24% of the cases and were held for diagnosis, whereas variants of uncertain significance were identified in 10% of cases. This study provides a new classification of genes that are involved in HPE. SHH, ZIC2, and SIX3 remain the top genes in term of frequency with GLI2, and are followed by FGF8 and FGFR1. The three minor HPE genes identified by our study are DLL1, DISP1, and SUFU. Here, we demonstrate that fibroblast growth factor signaling must now be considered a major pathway involved in HPE. Interestingly, several cases of double mutations were found and argue for a polygenic inheritance of HPE. Altogether, it supports that the implementation of NGS in HPE diagnosis is required to improve genetic counseling.

Genome-wide association study identifies TF as a significant modifier gene of iron metabolism in HFE hemochromatosis.

BACKGROUND & AIMS: Hereditary hemochromatosis (HH) is the most common form of genetic iron loading disease. It is mainly related to the homozygous C282Y/C282Y mutation in the HFE gene that is, however, a necessary but not a sufficient condition to develop clinical and even biochemical HH. This suggests that modifier genes are likely involved in the expressivity of the disease. Our aim was to identify such modifier genes. METHODS: We performed a genome-wide association study (GWAS) using DNA collected from 474 unrelated C282Y homozygotes. Associations were examined for both quantitative iron burden indices and clinical outcomes with 534,213 single nucleotide polymorphisms (SNP) genotypes, with replication analyses in an independent sample of 748 C282Y homozygotes from four different European centres. RESULTS: One SNP met genome-wide statistical significance for association with transferrin concentration (rs3811647, GWAS p value of 7×10(-9) and replication p value of 5×10(-13)). This SNP, located within intron 11 of the TF gene, had a pleiotropic effect on serum iron (GWAS p value of 4.9×10(-6) and replication p value of 3.2×10(-6)). Both serum transferrin and iron levels were associated with serum ferritin levels, amount of iron removed and global clinical stage (p<0.01). Serum iron levels were also associated with fibrosis stage (p<0.0001). CONCLUSIONS: This GWAS, the largest one performed so far in unselected HFE-associated HH (HFE-HH) patients, identified the rs3811647 polymorphism in the TF gene as the only SNP significantly associated with iron metabolism through serum transferrin and iron levels. Because these two outcomes were clearly associated with the biochemical and clinical expression of the disease, an indirect link between the rs3811647 polymorphism and the phenotypic presentation of HFE-HH is likely.

Characterizing the peritumoral brain zone in glioblastoma: a multidisciplinary analysis.

Glioblastoma (GB) is the most frequent and aggressive type of primary brain tumor. Recurrences are mostly located at the margin of the resection cavity in the peritumoral brain zone (PBZ). Although it is widely believed that infiltrative tumor cells in this zone are responsible for GB recurrence, few studies have examined this zone. In this study, we analyzed PBZ left after surgery with a variety of techniques including radiology, histopathology, flow cytometry, genomic, transcriptomic, proteomic, and primary cell cultures. The resulting PBZ profiles were compared with those of the GB tumor zone and normal brain samples to identify characteristics specific to the PBZ. We found that tumor cell infiltration detected by standard histological analysis was present in almost one third of PBZ taken from an area that was considered normal both on standard MRI and by the neurosurgeon under an operating microscope. The panel of techniques used in this study show that the PBZ, similar to the tumor zone itself, is characterized by substantial inter-patient heterogeneity, which makes it difficult to identify representative markers. Nevertheless, we identified specific alterations in the PBZ such as the presence of selected tumor clones and stromal cells with tumorigenic and angiogenic properties. The study of GB-PBZ is a growing field of interest and this region needs to be characterized further. This will facilitate the development of new, targeted therapies for patients with GB and the development of approaches to refine the per-operative evaluation of the PBZ to optimize the surgical resection of the tumor.

Mouse genetic background impacts both on iron and non-iron metals parameters and on their relationships.

Iron is reported to interact with other metals. In addition, it has been shown that genetic background may impact iron metabolism. Our objective was to characterize, in mice of three genetic backgrounds, the links between iron and several non-iron metals. Thirty normal mice (C57BL/6, Balb/c and DBA/2; n = 10 for each group), fed with the same diet, were studied. Quantification of iron, zinc, cobalt, copper, manganese, magnesium and rubidium was performed by ICP/MS in plasma, erythrocytes, liver and spleen. Transferrin saturation was determined. Hepatic hepcidin1 mRNA level was evaluated by quantitative RT-PCR. As previously reported, iron parameters were modulated by genetic background with significantly higher values for plasma iron parameters and liver iron concentration in DBA/2 and Balb/c strains. Hepatic hepcidin1 mRNA level was lower in DBA/2 mice. No iron parameter was correlated with hepcidin1 mRNA levels. Principal component analysis of the data obtained for non-iron metals indicated that metals parameters stratified the mice according to their genetic background. Plasma and tissue metals parameters that are dependent or independent of genetic background were identified. Moreover, relationships were found between plasma and tissue content of iron and some other metals parameters. Our data: (i) confirms the impact of the genetic background on iron parameters, (ii) shows that genetic background may also play a role in the metabolism of non-iron metals, (iii) identifies links between iron and other metals parameters which may have implications in the understanding and, potentially, the modulation of iron metabolism.

Impact of Sonic Hedgehog-dependent sphenoid bone defect on craniofacial growth.

OBJECTIVES: The main objective of this study was to evaluate how an apparently minor anomaly of the sphenoid bone, observed in a haploinsufficient mouse model for Sonic Hedgehog (Shh), affects the growth of the adult craniofacial region. This study aims to provide valuable information to orthodontists when making decisions regarding individuals carrying SHH mutation. MATERIALS AND METHODS: The skulls of embryonic, juvenile and adult mice of two genotypes (Shh heterozygous and wild type) were examined and measured using landmark-based linear dimensions. Additionally, we analysed the clinical characteristics of a group of patients and their relatives with SHH gene mutations. RESULTS: In the viable Shh+/ - mouse model, bred on a C57BL/6J background, we noted the presence of a persistent foramen at the midline of the basisphenoid bone. This particular anomaly was attributed to the existence of an ectopic pituitary gland. We discovered that this anomaly led to premature closure of the intrasphenoidal synchondrosis and contributed to craniofacial deformities in adult mice, including a longitudinally shortened skull base. This developmental anomaly is reminiscent of that commonly observed in human holoprosencephaly, a disorder resulting from a deficiency in SHH activity. However, sphenoid morphogenesis is not currently monitored in individuals carrying SHH mutations. CONCLUSION: Haploinsufficiency of Shh leads to isolated craniofacial skeletal hypoplasia in adult mouse. This finding highlights the importance of radiographic monitoring of the skull base in all individuals with SHH gene mutations.

Tailored Digital PCR Follow-Up of Rare Fusion Transcripts after Initial Detection through RNA Sequencing in Hematological Malignancies.

Minimal residual disease (MRD) monitoring plays a pivotal role in the management of hematologic malignancies. Well-established molecular targets, such as PML::RARA, CBFB::MYH11, or RUNX1::RUNX1T1, are conventionally tracked by quantitative RT-PCR. Recently, a broader landscape of fusion transcripts has been unveiled through transcriptomic analysis. These newly discovered fusion transcripts may emerge as novel molecular markers for MRD quantification. In this study, we compared a targeted RNA-sequencing (RNA-seq) approach (FusionPlex) with a whole-transcriptomic strategy (Advanta RNA-Seq XT) for fusion detection in a training set of 21 samples. We evidenced a concordance of 100% for the detection of known fusions, and showed a good correlation for gene expression quantification between the two techniques (Spearman r = 0.77). Additionally, we prospectively evaluated the identification of fusions by targeted RNA-seq in a real-life series of 126 patients with hematological malignancy. At least one fusion transcript was detected for 60 patients (48%). We designed tailored digital PCR assays for 11 rare fusions, and validated this technique for MRD quantification with a limit of detection of <0.01%. The combination of RNA-seq and tailored digital PCR may become a new standard for MRD evaluation in patients lacking conventional molecular targets.

MiR-31-3p do not predict anti-EGFR efficacy in first-line therapy of RAS wild-type metastatic right-sided colon cancer.

BACKGROUND: Low miR-31-3p expression was identified as predictive of anti-EGFR efficacy in RAS-wt mCRC. Primary tumor side was also proposed as a predictive factor of anti-EGFR benefit. This retrospective multicentric study evaluated the predictive role of miR-31-3p in right-sided RAS-wt mCRC patients treated with first-line CT+anti-EGFR or CT+bevacizumab (Beva). METHODS: Seventy-two right-sided RAS-wt mCRC patients treated in first-line with CT+anti-EGFR (n = 43) or Beva (n = 29) were included. Overall survival (OS), progression-free survival (PFS) and response rate (RR) were analyzed and stratified according to tumor miR-31-3p expression level and targeted therapy (TT). RESULTS: BRAF V600E mutation was more frequent in high vs low miR-31-3p expressers (60.6% vs 15.4%, P < 0.001). PFS was significantly longer with CT+Beva than with CT+anti-EGFR (13 vs 7 months; P = 0.024). Among low miR-31-3p expressers, PFS, OS and RR were not significantly different between the two groups, while in high miR-31-3p expressers, only PFS was longer in the CT+Beva group (11 vs 6 months; P = 0.03). In patients treated with CT+anti-EGFR, low miR-31-3p expressers had a significantly longer OS (20 vs 13 months; P = 0.02) than high miR-31-3p expressers. ORR was not significantly different between the two groups of treatment, in both low and high miR-31-3p expressers. MiR-31-3p expression status was statistically correlated between primary tumors and corresponding metastases. CONCLUSION: In this study, miR-31-3p couldn't identify a subgroup of patients with right-sided RAS-wt mCRC who might benefit from anti-EGFR and suggest that Beva is the TT of choice in first-line treatment of these patients.

MITF reprograms the extracellular matrix and focal adhesion in melanoma.

The microphthalmia-associated transcription factor (MITF) is a critical regulator of melanocyte development and differentiation. It also plays an important role in melanoma where it has been described as a molecular rheostat that, depending on activity levels, allows reversible switching between different cellular states. Here, we show that MITF directly represses the expression of genes associated with the extracellular matrix (ECM) and focal adhesion pathways in human melanoma cells as well as of regulators of epithelial-to-mesenchymal transition (EMT) such as CDH2, thus affecting cell morphology and cell-matrix interactions. Importantly, we show that these effects of MITF are reversible, as expected from the rheostat model. The number of focal adhesion points increased upon MITF knockdown, a feature observed in drug-resistant melanomas. Cells lacking MITF are similar to the cells of minimal residual disease observed in both human and zebrafish melanomas. Our results suggest that MITF plays a critical role as a repressor of gene expression and is actively involved in shaping the microenvironment of melanoma cells in a cell-autonomous manner.

DNA methylation in glioblastoma: impact on gene expression and clinical outcome.

BACKGROUND: Changes in promoter DNA methylation pattern of genes involved in key biological pathways have been reported in glioblastoma. Genome-wide assessments of DNA methylation levels are now required to decipher the epigenetic events involved in the aggressive phenotype of glioblastoma, and to guide new treatment strategies. RESULTS: We performed a whole-genome integrative analysis of methylation and gene expression profiles in 40 newly diagnosed glioblastoma patients. We also screened for associations between the level of methylation of CpG sites and overall survival in a cohort of 50 patients uniformly treated by surgery, radiotherapy and chemotherapy with concomitant and adjuvant temozolomide (STUPP protocol). The methylation analysis identified 616 CpG sites differentially methylated between glioblastoma and control brain, a quarter of which was differentially expressed in a concordant way. Thirteen of the genes with concordant CpG sites displayed an inverse correlation between promoter methylation and expression level in glioblastomas: B3GNT5, FABP7, ZNF217, BST2, OAS1, SLC13A5, GSTM5, ME1, UBXD3, TSPYL5, FAAH, C7orf13, and C3orf14. Survival analysis identified six CpG sites associated with overall survival. SOX10 promoter methylation status (two CpG sites) stratified patients similarly to MGMT status, but with a higher Area Under the Curve (0.78 vs. 0.71, p-value < 5e-04). The methylation status of the FNDC3B, TBX3, DGKI, and FSD1 promoters identified patients with MGMT-methylated tumors that did not respond to STUPP treatment (p-value < 1e-04). CONCLUSIONS: This study provides the first genome-wide integrative analysis of DNA methylation and gene expression profiles obtained from the same GBM cohort. We also present a methylome-based survival analysis for one of the largest uniformly treated GBM cohort ever studied, for more than 27,000 CpG sites. We have identified genes whose expression may be tightly regulated by epigenetic mechanisms and markers that may guide treatment decisions.

Integrative genome-wide analysis reveals a robust genomic glioblastoma signature associated with copy number driving changes in gene expression.

Glioblastoma multiforme shows multiple chromosomal aberrations, the impact of which on gene expression remains unclear. To investigate this relationship and to identify putative initiating genomic events, we integrated a paired copy number and gene expression survey in glioblastoma using whole human genome arrays. Loci of recurrent copy number alterations were combined with gene expression profiles obtained on the same tumor samples. We identified a set of 406 "cis-acting DNA targeted genes" corresponding to genomic aberrations with direct copy-number-driving changes in gene expression, defined as genes with either significantly concordant or correlated changes in DNA copy number and expression. Functional annotation revealed that these genes participate in key processes of cancer cell biology, providing insights into the genetic mechanisms driving glioblastoma. The robustness of the gene selection was validated on an external microarray data set including 81 glioblastomas and 23 non-neoplastic brain samples. The integration of array CGH and gene expression data highlights a robust cis-acting DNA targeted genes signature that may be critical for glioblastoma progression, with two tumor suppressor genes PCDH9 and STARD13 that could be involved in tumor invasiveness and resistance to etoposide.