Collagen VI sustains cell stemness and chemotherapy resistance in glioblastoma.

Microenvironmental factors are known fundamental regulators of the phenotype and aggressiveness of glioblastoma (GBM), the most lethal brain tumor, characterized by fast progression and marked resistance to treatments. In this context, the extracellular matrix (ECM) is known to heavily influence the behavior of cancer cells from several origins, contributing to stem cell niches, influencing tumor invasiveness and response to chemotherapy, mediating survival signaling cascades, and modulating inflammatory cell recruitment. Here, we show that collagen VI (COL6), an ECM protein widely expressed in both normal and pathological tissues, has a distinctive distribution within the GBM mass, strongly correlated with the most aggressive and phenotypically immature cells. Our data demonstrate that COL6 sustains the stem-like properties of GBM cells and supports the maintenance of an aggressive transcriptional program promoting cancer cell proliferation and survival. In particular, we identified a specific subset of COL6-transcriptionally co-regulated genes, required for the response of cells to replicative stress and DNA damage, supporting the concept that COL6 is an essential stimulus for the activation of GBM cell response and resistance to chemotherapy, through the ATM/ATR axis. Altogether, these findings indicate that COL6 plays a pivotal role in GBM tumor biology, exerting a pleiotropic action across different GBM hallmarks, including phenotypic identity and gene transcription, as well as response to treatments, thus providing valuable information for the understanding of the complex microenvironmental cues underlying GBM malignancy.

Involvement of Mitochondria in the Selective Response to Microsecond Pulsed Electric Fields on Healthy and Cancer Stem Cells in the Brain.

In the last few years, pulsed electric fields have emerged as promising clinical tools for tumor treatments. This study highlights the distinct impact of a specific pulsed electric field protocol, PEF-5 (0.3 MV/m, 40 µs, 5 pulses), on astrocytes (NHA) and medulloblastoma (D283) and glioblastoma (U87 NS) cancer stem-like cells (CSCs). We pursued this goal by performing ultrastructural analyses corroborated by molecular/omics approaches to understand the vulnerability or resistance mechanisms triggered by PEF-5 exposure in the different cell types. Electron microscopic analyses showed that, independently of exposed cells, the main targets of PEF-5 were the cell membrane and the cytoskeleton, causing membrane filopodium-like protrusion disappearance on the cell surface, here observed for the first time, accompanied by rapid cell swelling. PEF-5 induced different modifications in cell mitochondria. A complete mitochondrial dysfunction was demonstrated in D283, while a mild or negligible perturbation was observed in mitochondria of U87 NS cells and NHAs, respectively, not sufficient to impair their cell functions. Altogether, these results suggest the possibility of using PEF-based technology as a novel strategy to target selectively mitochondria of brain CSCs, preserving healthy cells.

Effects of Ultra-Short Pulsed Electric Field Exposure on Glioblastoma Cells.

Glioblastoma multiforme (GBM) is the most common brain cancer in adults. GBM starts from a small fraction of poorly differentiated and aggressive cancer stem cells (CSCs) responsible for aberrant proliferation and invasion. Due to extreme tumor heterogeneity, actual therapies provide poor positive outcomes, and cancers usually recur. Therefore, alternative approaches, possibly targeting CSCs, are necessary against GBM. Among emerging therapies, high intensity ultra-short pulsed electric fields (PEFs) are considered extremely promising and our previous results demonstrated the ability of a specific electric pulse protocol to selectively affect medulloblastoma CSCs preserving normal cells. Here, we tested the same exposure protocol to investigate the response of U87 GBM cells and U87-derived neurospheres. By analyzing different in vitro biological endpoints and taking advantage of transcriptomic and bioinformatics analyses, we found that, independent of CSC content, PEF exposure affected cell proliferation and differentially regulated hypoxia, inflammation and P53/cell cycle checkpoints. PEF exposure also significantly reduced the ability to form new neurospheres and inhibited the invasion potential. Importantly, exclusively in U87 neurospheres, PEF exposure changed the expression of stem-ness/differentiation genes. Our results confirm this physical stimulus as a promising treatment to destabilize GBM, opening up the possibility of developing effective PEF-mediated therapies.

AKR1C enzymes sustain therapy resistance in paediatric T-ALL.

BACKGROUND: Despite chemotherapy intensification, a subgroup of high-risk paediatric T-cell acute lymphoblastic leukemia (T-ALL) patients still experience treatment failure. In this context, we hypothesised that therapy resistance in T-ALL might involve aldo-keto reductase 1C (AKR1C) enzymes as previously reported for solid tumors. METHODS: Expression of NRF2-AKR1C signaling components has been analysed in paediatric T-ALL samples endowed with different treatment outcomes as well as in patient-derived xenografts of T-ALL. The effects of AKR1C enzyme modulation has been investigated in T-ALL cell lines and primary cultures by combining AKR1C inhibition, overexpression, and gene silencing approaches. RESULTS: We show that T-ALL cells overexpress AKR1C1-3 enzymes in therapy-resistant patients. We report that AKR1C1-3 enzymes play a role in the response to vincristine (VCR) treatment, also ex vivo in patient-derived xenografts. Moreover, we demonstrate that the modulation of AKR1C1-3 levels is sufficient to sensitise T-ALL cells to VCR. Finally, we show that T-ALL chemotherapeutics induce overactivation of AKR1C enzymes independent of therapy resistance, thus establishing a potential resistance loop during T-ALL combination treatment. CONCLUSIONS: Here, we demonstrate that expression and activity of AKR1C enzymes correlate with response to chemotherapeutics in T-ALL, posing AKR1C1-3 as potential targets for combination treatments during T-ALL therapy.

Outcome of patients affected by newly diagnosed glioblastoma undergoing surgery assisted by 5-aminolevulinic acid guided resection followed by BCNU wafers implantation: a 3-year follow-up.

The purpose of the study was to evaluate the clinical outcome of the association of BCNU wafers implantation and 5-aminolevulinic acid (5-ALA) fluorescence in the treatment of patients with newly diagnosed glioblastoma (ndGBM). Clinical and surgical data from patients who underwent 5-ALA surgery followed by BCNU wafers implantation were retrospectively evaluated (20 patients, Group I) and compared with data of patients undergoing surgery with BCNU wafers alone (42 patients, Group II) and 5-ALA alone (59 patients, Group III). Patients undergoing 5-ALA assisted resection followed by BCNU wafers implantation (Group I) resulted long survivors (>3 years) in 15 % of cases and showed a median PFS and MS of 11 and 22 months, respectively. Patients treated with BCNU wafers presented a significantly higher survival when tumor was removed with the assistance of 5-ALA (22 months with vs 18 months without 5-ALA, p < 0.0001); these data could be partially explained by the significantly higher CRET achieved in patients operated with 5-ALA assistance (80 % with vs 47 %% without 5-ALA). Moreover, patients of Group I showed a significant increased survival compared with Group III (5-ALA without BCNU) (22 months with vs 21 months without BCNU wafers, p = 0.0025) even with a comparable CRET (80 % vs 76 %, respectively). The occurrence of adverse events related to wafers did not significantly increase with 5-ALA (20 % with and 19 % without 5-ALA) and did not impact in survival outcome. In conclusion, our experience shows that on selected ndGBM patients 5-ALA technology and BCNU wafers implantation show a synergic action on patients' outcome without increasing adverse events occurrence.

A synthetic BMP-2 mimicking peptide induces glioblastoma stem cell differentiation.

BACKGROUND: Glioblastoma (GBM) is the most aggressive type of primary brain tumor, characterized by the intrinsic resistance to chemotherapy due to the presence of a highly aggressive Cancer Stem Cell (CSC) sub-population. In this context, Bone Morphogenetic Proteins (BMPs) have been demonstrated to induce CSC differentiation and to sensitize GBM cells to treatments. METHODS: The BMP-2 mimicking peptide, named GBMP1a, was synthesized on solid-phase by Fmoc chemistry. Structural characterization and prediction of receptor binding were obtained by Circular Dicroism (CD) and NRM analyses. Activation of BMP signalling was evaluated by a luciferase reporter assay and western blot. Pro-differentiating effects of GBMP1a were verified by immunostaining and neurosphere assay in primary glioblastoma cultures. RESULTS: CD and NMR showed that GBMP1a correctly folds into expected tridimensional structures and predicted its binding to BMPR-IA to the same epitope as in the native complex. Reporter analysis disclosed that GBMP1a is able to activate BMP signalling in GBM cells. Moreover, BMP-signalling activation was specifically dependent on smad1/5/8 phosphorylation. Finally, we confirmed that GBMP1a treatment is sufficient to enhance osteogenic differentiation of Mesenchymal Stem Cells and to induce astroglial differentiation of glioma stem cells (GSCs) in vitro. CONCLUSIONS: GBMP1a was demonstrated to be a good inducer of GSC differentiation, thus being considered a potential anti-cancer tool to be further developed for GBM treatment. GENERAL SIGNIFICANCE: These data highlight the role of BMP-mimicking peptides as potential anti-cancer agents against GBM and stimulate the further development of GBMP1a-based structures in order to enhance its stability and activity.

Proteomic Alterations in Response to Hypoxia Inducible Factor 2α in Normoxic Neuroblastoma Cells.

Hypoxia inducible factor (HIF)-2α protein expression in solid tumors promotes stem-like phenotype in cancer stem cells and increases tumorigenic potential in nonstem cancer cells. Recently, we have shown that HIF-1/2α gene expression is correlated to neuroblastoma (NB) poor survival and to undifferentiated tumor state; HIF-2α protein was demonstrated to enhance aggressive features of the disease. In this study, we used proteomic experiments on NB cells to investigate HIF-2α downstream-regulated proteins or pathways with the aim of providing novel therapeutic targets or bad prognosis markers. We verified that pathways mostly altered by HIF-2α perturbation are involved in tumor progression. In particular, HIF-2α induces alteration of central metabolism and splicing control pathways. Simultaneously, WNT, RAS/MAPK, and PI3K/AKT activity or expression are affected and may impact the sensitivity and the intensity of HIF-2α-regulated pathways. Furthermore, genes coding the identified HIF-2α-related markers built a signature able to stratify NB patients with unfavorable outcome. Taken together, our findings underline the relevance of dissecting the downstream effects of a poor survival marker in developing targeted therapy and improving patient stratification. Future prospective studies are needed to translate the use of these data into the clinical practice.

Phenotypic and functional characterization of Glioblastoma cancer stem cells identified through 5-aminolevulinic acid-assisted surgery [corrected].

5-aminolevulinic acid (5-ALA) introduction in the surgical management of Glioblastoma (GBM) enables the intra-operatively identification of cancer cells in the mass by means of fluorescence. Here, we analyzed the phenotype of GBM cells isolated from distinct tumour areas determined by 5-ALA (tumour core, 5-ALA intense and vague layers) and the potency of 5-ALA labelling in identifying GBM cells and cancer stem cells (CSCs) in the mass. 5-ALA identified distinct layers in the mass, with less differentiated cells residing in the core of the tumour. 5-ALA was able to stain up to 68.5% of CD133(+) cells in the 5-ALA intense layer and, although 5-ALA(+) cells retrieved from different tumour areas contained a similar proportion of CD133(+) cells (range 27.5-35.6%), those from the vague layer displayed the lowest ability to self-renew. In conclusion, our data demonstrate that a substantial amount of GBM cells and CSCs in the mass are able to avoid 5-ALA labelling and support the presence of heterogenic CSC populations in the GBM mass.

Intra-operative 5-aminolevulinic acid (ALA)-induced fluorescence of medulloblastoma: phenotypic variability and CD133(+) expression according to different fluorescence patterns.

5-Aminolevulinic acid (5-ALA) fluorescence has been proved advantageous in glioma surgery. Conflicting results have been reported by few studies published in literature about intra-operative 5-ALA-induced fluorescence of medulloblastoma (MDB). The aim of this study is to verify if these conflicting results could be explained by intra-tumoral histological and phenotypic differences. In the present case of a 45-year-old patient affected by a cerebellar MDB, histological analysis of cell phenotype and 5-ALA and CD133 correlation were performed in multiple samples according to different fluorescence patterns. Intra-operatively, the tumor appeared unevenly fluorescent under blue-violet light. Histologically, 5-ALA-intense biopsies from inner areas were characterized by a significant amount of cancer cells, whereas 5-ALA faint regions from peripheral areas displayed normal cerebellar features, with MDB cells infiltrating healthy tissues. Presenting our findings, we show the correlation between different 5-ALA fluorescence patterns of medulloblastoma with specific histological and phenotypical features. Thus, we hypothesize that a distinct relationship between CD133 expression and fluorescence accumulation presented in our study could partially explain the divergent results published in literature.

A yeast based assay establishes the pathogenicity of novel missense ACTA2 variants associated with aortic aneurysms.

The ACTA2 gene codes for alpha-smooth muscle actin, a critical component of the contractile apparatus of the vascular smooth muscle cells. Autosomal dominant variants in the ACTA2 gene have been associated to familial non-syndromic thoracic aortic aneurysm/dissection (TAAD). They are thought to act through a dominant-negative mechanism. These variants display incomplete penetrance and variable expressivity, complicating the validation of ACTA2 variants pathogenicity by family segregation studies. In this study, we developed a yeast based assay to test putative TAAD-associated ACTA2 variants. We identified five new heterozygous ACTA2 missense variants in TAAD patients through next generation sequencing. We decided to test their pathogenicity in Saccharomyces cerevisiae, since yeast actin is very similar to human alpha-smooth muscle actin, and the residues at which the TAAD-associated variants occur in ACTA2 are well conserved. A wild type yeast strain was transformed with a vector expressing the different mutant alleles, to model the heterozygous condition of patients. Then, we evaluated yeast growth by spot test and cytoskeletal and mitochondrial morphology by fluorescence microscopy. We found that mutant yeast strains displayed only mild growth defects but a significant increase in the percentage of cells with abnormal mitochondrial distribution and abnormal organization of the actin cytoskeleton compared to controls. All variants appeared to interfere with the activity of wild type actin in yeast, suggesting a dominant-negative pathogenic mechanism. Our results demonstrate the utility of using the yeast actin model system to validate the pathogenicity of TAAD-associated ACTA2 variants.

On the effects of 30.5 GHz sinusoidal wave exposure on glioblastoma organoids.

Introduction: Glioblastoma (grade IV) is the most aggressive primary brain tumor in adults, representing one of the biggest therapeutic challenges due to its highly aggressive nature. In this study, we investigated the impact of millimeter waves on tridimensional glioblastoma organoids derived directly from patient tumors. Our goal was to explore novel therapeutic possibilities in the fight against this challenging disease. Methods: The exposure setup was meticulously developed in-house, and we employed a comprehensive dosimetry approach, combining numerical and experimental methods. Biological endpoints included a global transcriptional profiling analysis to highlight possible deregulated pathways, analysis of cell morphological changes, and cell phenotypic characterization which are all important players in the control of glioblastoma progression. Results and discussion: Our results revealed a significant effect of continuous millimeter waves at 30.5 GHz on cell proliferation and apoptosis, although without affecting the differentiation status of glioblastoma cells composing the organoids. Excitingly, when applying a power level of 0.1 W (Root Mean Square), we discovered a remarkable (statistically significant) therapeutic effect when combined with the chemotherapeutic agent Temozolomide, leading to increased glioblastoma cell death. These findings present a promising interventional window for treating glioblastoma cells, harnessing the potential therapeutic benefits of 30.5 GHz CW exposure. Temperature increase during treatments was carefully monitored and simulated with a good agreement, demonstrating a negligible involvement of the temperature elevation for the observed effects. By exploring this innovative approach, we pave the way for improved future treatments of glioblastoma that has remained exceptionally challenging until now.

TR-644 a novel potent tubulin binding agent induces impairment of endothelial cells function and inhibits angiogenesis.

TR-644 is a novel combretastatin A-4 (CA-4) analogue endowed with potent microtubule depolymerizing activity superior to that of the lead compound and it also has high affinity to colchicines binding site of tubulin. We tested TR-644 anti-angiogenic effects in human umbilical endothelial cells (HUVEC). It showed no significant effects on the growth of HUVEC cells at concentrations below 1,000 nM, but at much lower concentrations (10-100 nM) it induced inhibition of capillary tube formation, inhibition of endothelial cell migration and affected endothelial cell morphology as demonstrated by the disruption of the microtubule network. TR-644 also increased permeability of HUVEC cells in a time dependent manner. The molecular mechanism for the anti-vascular activity of TR-644 was investigated in detail. TR-644 caused G2/M arrest in endothelial cells and this effect correlated with downregulation of the expression of Cdc25C and Cdc2(Tyr15). Moreover TR-644 inhibited VEGF-induced phosphorylation of VE-cadherin but did not prevent the VEGF-induced phosphorylation of FAK. In chick chorioallantoic membrane in vivo assay, TR-644 (0.1-1.0 pmol/egg) efficiently counteracted the strong angiogenic response induced by FGF. Also CA-4, used as reference compound, caused an antagonistic effect, but in contrast, it induced per se, a remarkable angiogenic response probably due to an inflammatory reaction in the site of treatment. In a mice allogenic tumor model, immunohistochemical staining of tumors with anti-CD31 antibody showed that TR-644 significantly reduced the number of vessel, after 24 h from the administration of a single dose (30 mg/Kg).

Wnt Signaling in Brain Tumors: A Challenging Therapeutic Target.

The involvement of Wnt signaling in normal tissue homeostasis and disease has been widely demonstrated over the last 20 years. In particular, dysregulation of Wnt pathway components has been suggested as a relevant hallmark of several neoplastic malignancies, playing a role in cancer onset, progression, and response to treatments. In this review, we summarize the current knowledge on the instructions provided by Wnt signaling during organogenesis and, particularly, brain development. Moreover, we recapitulate the most relevant mechanisms through which aberrant Wnt pathway activation may impact on brain tumorigenesis and brain tumor aggressiveness, with a particular focus on the mutual interdependency existing between Wnt signaling components and the brain tumor microenvironment. Finally, the latest anti-cancer therapeutic approaches employing the specific targeting of Wnt signaling are extensively reviewed and discussed. In conclusion, here we provide evidence that Wnt signaling, due to its pleiotropic involvement in several brain tumor features, may represent a relevant target in this context, although additional efforts will be needed to: (i) demonstrate the real clinical impact of Wnt inhibition in these tumors; (ii) overcome some still unsolved concerns about the potential systemic effects of such approaches; (iii) achieve efficient brain penetration.

Surviving the hunger games: Metabolic reprogramming in medulloblastoma.

Medulloblastoma is a highly malignant pediatric brain tumor characterized by its aggressive nature and limited treatment options. Metabolic changes have recently emerged as key factors in the development, progression, and response to therapy in various types of cancer. Cancer cells exhibit remarkable adaptability by modulating glucose, lipids, amino acids, and nucleotide metabolism to survive in nutrient- and oxygen-deprived environments. Although medulloblastoma has been extensively studied from a genomic perspective, leading to the identification of four subgroups and their respective subcategories, the investigation of its metabolic phenotype has remained relatively understudied. This review focus on the available literature, aiming to summarize the current knowledge about the main metabolic pathways that are deregulated in medulloblastoma tumors, while emphasizing the controversial aspects and the progress that is yet to be made. Furthermore, we underscored the insights gained so far regarding the impact of metabolism on the development of drug resistance in medulloblastoma and the therapeutic strategies employed to target specific metabolic pathways.

Identification of Homoharringtonine as a potent inhibitor of glioblastoma cell proliferation and migration.

We previously demonstrated that Annexin A2 (ANXA2) is a pivotal mediator of the pro-oncogenic features displayed by glioblastoma (GBM) tumors, the deadliest adult brain malignancies, being involved in cell stemness, proliferation and invasion, thus negatively impacting patient prognosis. Based on these results, we hypothesized that compounds able to revert ANXA2-dependent transcriptional features could be exploited as reliable treatments to inhibit GBM cell aggressiveness by hampering their proliferative and migratory potential. Transcriptional signatures obtained by the modulation of ANXA2 activity/levels were functionally mapped through the QUADrATiC bioinformatic tool for compound identification. Selected compounds were screened by cell proliferation and migration assays in primary GBM cells, and we identified Homoharringtonine (HHT) as a potent inhibitor of GBM cell motility and proliferation, without affecting their viability. A further molecular characterization of the effects displayed by HHT, confirmed its ability to inhibit a transcriptional program involved in cell migration and invasion. Moreover, we demonstrated that the multiple antitumoral effects displayed by HHT are correlated to the inhibition of a platelet derived growth factor receptor α (PDGFRα)-dependent intracellular signaling through the impairment of Signal transducer and activator of transcription 3 (STAT3) and Ras homolog family member A (RhoA) axes. Our results demonstrate that HHT may act as a potent inhibitor of cancer cell proliferation and invasion in GBM, by hampering multiple PDGFRα-dependent oncogenic signals transduced through the STAT3 and RhoA intracellular components, finally suggesting its potential transferability for achieving an effective impairment of peculiar GBM hallmarks.

Molecular and functional profiling of chemotolerant cells unveils nucleoside metabolism-dependent vulnerabilities in medulloblastoma.

Chemotherapy resistance is considered one of the main causes of tumor relapse, still challenging researchers for the identification of the molecular mechanisms sustaining its emergence. Here, we setup and characterized chemotherapy-resistant models of Medulloblastoma (MB), one of the most lethal pediatric brain tumors, to uncover targetable vulnerabilities associated to their resistant phenotype. Integration of proteomic, transcriptomic and kinomic data revealed a significant deregulation of several pathways in resistant MB cells, converging to cell metabolism, RNA/protein homeostasis, and immune response, eventually impacting on patient outcome. Moreover, resistant MB cell response to a large library of compounds through a high-throughput screening (HTS), highlighted nucleoside metabolism as a relevant vulnerability of chemotolerant cells, with peculiar antimetabolites demonstrating increased efficacy against them and even synergism with conventional chemotherapeutics. Our results suggest that drug-resistant cells significantly rewire multiple cellular processes, allowing their adaptation to a chemotoxic environment, nevertheless exposing alternative actionable susceptibilities for their specific targeting.

High-Throughput Microscopy Analysis of Mitochondrial Membrane Potential in 2D and 3D Models.

Recent proteomic, metabolomic, and transcriptomic studies have highlighted a connection between changes in mitochondria physiology and cellular pathophysiological mechanisms. Secondary assays to assess the function of these organelles appear fundamental to validate these -omics findings. Although mitochondrial membrane potential is widely recognized as an indicator of mitochondrial activity, high-content imaging-based approaches coupled to multiparametric to measure it have not been established yet. In this paper, we describe a methodology for the unbiased high-throughput quantification of mitochondrial membrane potential in vitro, which is suitable for 2D to 3D models. We successfully used our method to analyze mitochondrial membrane potential in monolayers of human fibroblasts, neural stem cells, spheroids, and isolated muscle fibers. Moreover, by combining automated image analysis and machine learning, we were able to discriminate melanoma cells from macrophages in co-culture and to analyze the subpopulations separately. Our data demonstrated that our method is a widely applicable strategy for large-scale profiling of mitochondrial activity.

Long-term proliferation of immature hypoxia-dependent JMML cells supported by a 3D in vitro system.

Juvenile myelomonocytic leukemia (JMML) is a rare clonal stem cell disorder that occurs in early childhood and is characterized by the hyperactivation of the RAS pathway in 95% of the patients. JMML is characterized by a hyperproliferation of granulocytes and monocytes, and little is known about the heterogeneous nature of leukemia-initiating cells, as well as of the cellular hierarchy of the JMML bone marrow. In this study, we report the generation and characterization of a novel patient-derived three-dimensional (3D) in vitro JMML model, called patient-derived JMML Atypical Organoid (pd-JAO), sustaining the long-term proliferation of JMML cells with stem cell features and patient-specific hallmarks. JMML cells brewed in a 3D model under different microenvironmental conditions acquired proliferative and survival advantages when placed under low oxygen tension. Transcriptomic and microscopic analyses revealed the activation of specific metabolic energy pathways and the inactivation of processes leading to cell death. Furthermore, we demonstrated the pd-JAO-derived cells' migratory, propagation, and self-renewal capacities. Our study contributes to the development of a robust JMML 3D in vitro model for studying and defining the impact of microenvironmental stimuli on JMML disease and the molecular mechanisms that regulate JMML initiating and propagating cells. Pd-JAO may become a promising model for compound tests focusing on new therapeutic interventions aimed at eradicating JMML progenitors and controlling JMML disease.

MGMT expression and promoter methylation status may depend on the site of surgical sample collection within glioblastoma: a possible pitfall in stratification of patients?

We recently described a three-layer concentric model of a glioblastoma (GBM) related to a specific distribution of molecular and phenotypic characteristics driven by the intratumoral hypoxic gradient in which the cancer stem cells niche is located in the hypoxic necrotic core of the tumour. The purpose of this study was to investigate the relationship between O(6)-methylguanine-DNA methyltransferase (MGMT) promoter methylation status and MGMT expression in GBM samples collected according to the three-layer concentric model. Multiple tissue samples were obtained, by means of image-guided surgery, from the three concentric layers of newly diagnosed GBM. Two samples from each layer were collected from 12 patients (total 72 samples). Immunohistochemical analysis was performed on formalin-fixed paraffin-embedded tissue samples. The methylation status of the MGMT promoter was determined by methylation-specific polymerase-chain-reaction analysis. In all tumours, MGMT protein expression decreased progressively from the inner to the outer layer, and methylation of the MGMT promoter was unrelated to tumour layer. In particular, the MGMT promoter was unmethylated in all layers in 41.7% of tumours, methylated in all layers in 25%, and variably methylated in the three layers in 33.3%. We recorded concordance between MGMT expression and MGMT promoter methylation status within the GBM in only 58.8% of the samples collected. Our data suggest that both MGMT expression and promoter methylation data may be variable throughout GBM and that they may, consequently, depend on the site of surgical sample collection, even in the same patient. However, whereas MGMT expression is pre-operatively predictable when sampling is performed according to the three-layer concentric model, MGMT promoter methylation is not. These results must be considered when sample collection is performed for assessment of MGMT data.

Notch3 signalling promotes tumour growth in colorectal cancer.

Increased Notch1 activity has been observed in intestinal tumours, partially accomplished by ß-catenin-mediated up-regulation of the Notch ligand Jagged-1. Whether further mechanisms of Notch activation exist and other Notch receptors might be involved is unclear. Microarray data indicated that Notch3 transcript levels are significantly up-regulated in primary and metastatic CRC samples compared to normal mucosa. Moreover, Notch3 protein was expressed at strong/moderate levels by 19.7% of 158 CRC samples analysed, and at weak levels by 51.2% of the samples. Intrigued by these findings, we sought to investigate whether Notch3 modulates oncogenic features of CRC cells. By exploiting xenografts of CRC cells with different tumourigenic properties in mice, we found that the aggressive phenotype was associated with altered expression of components of the Notch pathway, including Notch3, Delta-like 4 (DLL4), and Jagged-1 ligands. Stimulation with immobilized recombinant DLL4 or transduction with DLL4-expressing vectors dramatically increased Notch3 expression in CRC cells, associated with accelerated tumour growth. Forced expression of an active form of Notch3 mirrored the effects of DLL4 stimulation and increased tumour formation. Conversely, attenuation of Notch3 levels by shRNA resulted in perturbation of the cell cycle followed by reduction in cell proliferation, clonogenic capacity, and inhibition of tumour growth. Altogether, these findings indicate that Notch3 can modulate the tumourigenic properties of CRC cells and contributes to sustained Notch activity in DLL4-expressing tumours.