scholarly journals GENE-55. GENE EXPRESSION SIGNATURE ASSOCIATED WITH AGGRESSIVE GLIOBLASTOMA GROWTH IS ENRICHED IN CHROMATIN MODIFICATION AND STEMNESS TRANSCRIPTIONAL REGULATION PROGRAMS

2019 ◽  
Vol 21 (Supplement_6) ◽  
pp. vi109-vi109
Author(s):  
Artem Berezovsky ◽  
Oluwademilade Nuga ◽  
Yuling Meng ◽  
Laila Poisson ◽  
Houtan Noushmehr ◽  
...  

Abstract Prognosis of patients diagnosed with IDHwt glioblastoma is influenced by known clinical and demographic factors, and likely by physiological characteristics. Our goal is to determine tumor-intrinsic gene expression signatures associated with aggressive tumor growth in patient-derived xenografts (PDXs). Cancer stem cell (CSC) lines established from 10 IDHwt glioblastoma tumors were implanted orthotopically in cohorts of 10 to 15 nude mice (3x10E5 viable cells/mouse), for development of PDX under uniform conditions. Mice were monitored and sacrificed when symptomatic. Five PDX lines, presenting median survival of 29 to 59 days were classified as short (S) survivors, and 5 lines with median survival between 111 and 134 days as long (L) survivors. RNA was isolated from terminal PDX tumors (n=3/line) and sequenced using Illumina HiSeq 2000. Differential gene expression analysis between tumors in S and L survival groups was conducted using the lmFit and eBayes, and genes were ranked by Benjamini-Hochberg adjusted P-values, set to adj.p< 0.05, resulting in 1663 genes upregulated and 1539 genes downregulated in the aggressive S group. Gene ontology analysis was performed using Metacore (Clarivate Analytics) and Metascape (http://metascape.org). Chromatin modification was significantly enriched in the aggressive tumor group (Metacore, p= 1x 10–12). Remarkably, 40% (654/1663) of the genes upregulated in the aggressive PDX tumors were co-expressed with an epigenetic master regulator in the TCGA glioblastoma RNAseq dataset (cBio Portal, q< 10E-15), and this subset was also highly enriched in chromatin modification, stemness transcriptional regulation and DNA repair (q=10E-45 to q=10E-13). These results indicate that novel host-independent prognostic gene expression signatures can be derived from the PDX models and underline the potential of epigenetic regulators as therapeutic target for aggressive glioblastomas. Our results further indicate that these models are suitable for testing a new generation of epigenetic drugs currently in pre-clinical and clinical development.

2019 ◽  
Author(s):  
Mario A. Inchiosa

AbstractPrevious clinical studies with the FDA-approved alpha-adrenergic antagonist, phenoxybenzamine, showed apparent efficacy to reverse the symptoms and disabilities of the neuropathic condition, Complex Regional Pain Syndrome; also, the anatomic spread and intensity of this syndrome has a proliferative character and it was proposed that phenoxybenzamine may have an anti-inflammatory, immunomodulatory mode of action. A previous study gave evidence that phenoxybenzamine had anti-proliferative activity in suppression of growth in several human tumor cell cultures. The same report demonstrated that the drug possessed significant histone deacetylase inhibitory activity. Utilizing the Harvard/Massachusetts Institute of Technology Broad Institute genomic database, CLUE, the present study suggests that the gene expression signature of phenoxybenzamine in malignant cell lines is consistent with anti-inflammatory/immunomodulatory activity and suppression of tumor expansion by several possible mechanisms of action. Of particular note, phenoxybenzamine demonstrated signatures that were highly similar to those with glucocorticoid agonist activity. Also, gene expression signatures of phenoxbenzamine were consistent with several agents in each case that were known to suppress tumor proliferation, notably, protein kinase C inhibitors, Heat Shock Protein inhibitors, epidermal growth factor receptor inhibitors, and glycogen synthase kinase inhibitors. Searches in CLUE also confirmed the earlier observations of strong similarities between gene expression signatures of phenoxybenzamine and several histone deacetylase inhibitors.


Author(s):  
Oscar Mendez-Lucio ◽  
Benoit Baillif ◽  
Djork-Arné Clevert ◽  
David Rouquié ◽  
Joerg Wichard

Finding new molecules with a desired biological activity is an extremely difficult task. In this context, artificial intelligence and generative models have been used for molecular <i>de novo</i> design and compound optimization. Herein, we report the first generative model that bridges systems biology and molecular design conditioning a generative adversarial network with transcriptomic data. By doing this we could generate molecules that have high probability to produce a desired biological effect at cellular level. We show that this model is able to design active-like molecules for desired targets without any previous target annotation of the training compounds as long as the gene expression signature of the desired state is provided. The molecules generated by this model are more similar to active compounds than the ones identified by similarity of gene expression signatures, which is the state-of-the-art method for navigating compound-induced gene expression data. Overall, this method represents a novel way to bridge chemistry and biology to advance in the long and difficult road of drug discovery.


2018 ◽  
Author(s):  
Oscar Mendez-Lucio ◽  
Benoit Baillif ◽  
Djork-Arné Clevert ◽  
David Rouquié ◽  
Joerg Wichard

Finding new molecules with a desired biological activity is an extremely difficult task. In this context, artificial intelligence and generative models have been used for molecular <i>de novo</i> design and compound optimization. Herein, we report the first generative model that bridges systems biology and molecular design conditioning a generative adversarial network with transcriptomic data. By doing this we could generate molecules that have high probability to produce a desired biological effect at cellular level. We show that this model is able to design active-like molecules for desired targets without any previous target annotation of the training compounds as long as the gene expression signature of the desired state is provided. The molecules generated by this model are more similar to active compounds than the ones identified by similarity of gene expression signatures, which is the state-of-the-art method for navigating compound-induced gene expression data. Overall, this method represents a novel way to bridge chemistry and biology to advance in the long and difficult road of drug discovery.


2020 ◽  
Author(s):  
Zeid M Rusan ◽  
Michael P Cary ◽  
Roland J Bainton

AbstractMulticellular organisms employ concurrent gene regulatory programs to control development and physiology of cells and tissues. The Drosophila melanogaster model system has a remarkable history of revealing the genes and mechanisms underlying fundamental biology yet much remains unclear. In particular, brain xenobiotic protection and endobiotic regulatory systems that require transcriptional coordination across different cell types, operating in parallel with the primary nervous system and metabolic functions of each cell type, are still poorly understood. Here we use the unsupervised machine learning method independent component analysis (ICA) on majority fresh-frozen, bulk tissue microarrays to define biologically pertinent gene expression signatures which are sparse, i.e. each involving only a fraction of all fly genes. We optimize the gene expression signature definitions partly through repeated application of a stochastic ICA algorithm to a compendium of 3,346 microarrays from 221 experiments provided by the Drosophila research community. Our optimized ICA model of pan fly gene expression consists of 850 modules of co-regulated genes that map to tissue developmental stages, disease states, cell-autonomous pathways and presumably novel processes. Importantly, we show biologically relevant gene modules expressed at varying amplitudes in whole brain and isolated adult blood-brain barrier cell levels. Thus, whole tissue derived ICA transcriptional signatures that transcend single cell type boundaries provide a window into the transcriptional states of difficult to isolate cell ensembles maintaining delicate brain physiologies. We believe the fly ICA gene expression signatures set, by virtue of the success of ICA at inferring robust often low amplitude patterns across large datasets and the quality of the input samples, to be an important asset for analyzing compendium and newly generated microarray or RNA-seq expression datasets.


2019 ◽  
Author(s):  
Joske Ubels ◽  
Pieter Sonneveld ◽  
Martin H. van Vliet ◽  
Jeroen de Ridder

AbstractMany cancer drugs only benefit a subset of the patients that receive them, but are often associated with serious side effects. Predictive classification methods that can identify which patients will benefit from a specific treatment are therefore of great clinical utility. We here introduce a novel machine learning method to identify predictive gene expression signatures, based on the idea that patients who received different treatments but exhibit similar expression profiles can be used to model response to the alternative treatment. We use this method to predict proteasome inhibitor benefit in Multiple Myeloma (MM). In a dataset of 910 MM patients we identify a 14-gene expression signature that can successfully predict benefit to the proteasome inhibitor bortezomib, with a hazard ratio of 0.47 (p = 0.04) in class ‘benefit’, while in class ‘no benefit’ the hazard ratio is 0.91 (p = 0.68). Importantly, we observe a similar classification performance (HR class benefit = 0.46, p = 0.04) in an independent patient cohort which was moreover measured on a different platform, demonstrating the robustness of the signature. Moreover, we find that the genes in the discovered signature are essential, as no equivalent signature can be found when they are excluded from the analysis. Multiple genes in the signature are linked to working mechanisms of proteasome inhibitors or MM disease progression. In conclusion, our method allows for identification of gene expression signatures that can aid in treatment decisions for MM patients and provide insight into the biological mechanism behind treatment benefit.


2019 ◽  
Author(s):  
J. Javier Díaz-Mejía ◽  
Elaine C. Meng ◽  
Alexander R. Pico ◽  
Sonya A. MacParland ◽  
Troy Ketela ◽  
...  

AbstractIdentification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated computational steps like data normalization, dimensionality reduction and cell clustering. However, assigning cell type labels to cell clusters is still conducted manually by most researchers, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. Two bottlenecks to automating this task are the scarcity of reference cell type gene expression signatures and that some dedicated methods are available only as web servers with limited cell type gene expression signatures. In this study, we benchmarked four methods (CIBERSORT, GSEA, GSVA, and ORA) for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used scRNA-seq datasets from liver, peripheral blood mononuclear cells and retinal neurons for which reference cell type gene expression signatures were available. Our results show that, in general, all four methods show a high performance in the task as evaluated by Receiver Operating Characteristic curve analysis (average AUC = 0.94, sd = 0.036), whereas Precision-Recall curve analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). CIBERSORT and GSVA were the top two performers. Additionally, GSVA was the fastest of the four methods and was more robust in cell type gene expression signature subsampling simulations. We provide an extensible framework to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.


Blood ◽  
2015 ◽  
Vol 126 (23) ◽  
pp. 3719-3719
Author(s):  
Marta Sanchez-Martin ◽  
Alberto Ambesi-Impiombato ◽  
Luyao Xu ◽  
Yue Qin ◽  
Daniel Herranz ◽  
...  

Abstract Oncogenic NOTCH signaling is a major driver of T-cell transformation in T-cell acute lymphoblastic leukemia (T-ALL). However, clinical studies testing the efficacy of NOTCH1 inactivation with γ-secretase inhibitors (GSIs) have shown limited antileukemic activity for these drugs as single agents. Here we used an expression-based virtual screening approach and network perturbation analyses to identify and functionally characterize new highly active antileukemic drugs synergistic with NOTCH1 inhibition in T-ALL. Gene expression profiling studies have shown a prominent gene expression signature dominated by genes involved in growth and metabolism downstream of NOTCH1 in T-ALL. Notably, loss of the PTEN tumor suppressor gene confers resistance to GSI therapy and effectively rescues the gene expression signature induced by NOTCH1 inhibition in T-ALL. We hypothesized that drugs inducing transcriptional programs overlapping with those driven by NOTCH1 inhibition and antagonizing those resulting from PTEN loss could have synergistic antileukemic effects with GSIs in PTEN wild type and PTEN null leukemia cells. To address this question we generated gene expression signatures from Pten conditional-inducible knockout NOTCH1-driven leukemias in basal condition, upon NOTCH1 inhibition by GSI treatment and upon deletion of Pten. Connectivity Map (cMAP) analysis in this series identified 17 high scoring compounds as candidate antileukemic drugs (p<0.01). Reassuringly these included two inhibitors of the mTOR/PI3K/AKT pathway (rapamycin, wortmannin), but also histone deacetylase inhibitors (vorinostat, trichostatin A and valproic acid), phenothiazine antipsychotic drugs (trifluoperazine and thioridazine), antimalarial agents (astemizole, mefloquine) and compounds with less characterized activities such as withaferin A, parthenolide and pyrvinium pamoate. Transcriptional profiling followed by pairwise gene set enrichment analysis of these compounds identified groups of drugs with highly interconnected transcriptional programs suggestive of an overlapping mechanism of action (e.g. mTOR/PI3K inhibitors, HDAC inhibitors and phenothiazines), as well as compounds with more unique expression signatures suggestive of a more distinct mode of action (e.g. withaferin A, astemizole and mefloquine). Detailed characterization of the antileukemic effects of these 17 cMAP hits alone and in combination with the GSI DBZ in a broad panel of human NOTCH1-mutated T-ALL cell lines, identified withaferin A, rapamycin, wortmannin, parthenolide and vorinostat as the most active (lethal dose 50 <0.5 µM) and GSI-synergistic (combination index <0.4) drugs in this series. Among these, withaferin A, stood out as the most cytotoxic and GSI-synergistic compound against both PTEN positive and PTEN null T-ALL cell lines. Moreover, withaferin A treatment of primary mouse NOTCH1-induced T-ALLs and primary human T-ALL xenografts demonstrated strong and GSI-synergistic antileukemic activity in vivo. To address the mechanisms mediating the antileukemic effects of withaferin A we performed a detailed analysis of the gene expression signatures induced by this drug in T-ALL lymphoblasts. These studies revealed a strong enrichment of downregulated genes involved in translation regulation in T-ALL cells upon treatment with withaferin A (p<0.001). Mechanistically, transcriptional network perturbation analysis identified the eIF2A translation initiation complex as a potential effector of the antileukemic effects of withaferin A, and withaferin A treatment induced strong dose dependent phosphorylation of eIF2S1 in position S51, a modification responsible for blocking the activity of the eIF2A complex. Consistently, polysome profiling and nascent-protein assays revealed decreased translation in T-ALL cells treated with withaferin A. In this context, expression a phosphomimetic mutant form of eIF2S1 (S51D) impaired leukemia cell viability. Moreover, expression of a non-phosphorylatable form of eIF2S1 (eIF2S1 S51A) in T-ALL cells abrogated the antileukemic effects of withaferin A.These results support a direct role of eIF2S1 phosphorylation and the inhibition of eIF2A-dependent translation as a critical mediators of the antileukemic effects of withaferin A in T-ALL and a role for the combination of GSIs and inhibitors of protein translation for the treatment of high risk T-ALL. Disclosures Califano: Therasis Inc: Employment; Cancer Genetics Inc: Consultancy; Ipsen pharmaceuticals: Consultancy; Thermo Fischer Scientific: Consultancy.


Author(s):  
Giuseppe Rizzo ◽  
Ehsan Vafadarnejad ◽  
Panagiota Arampatzi ◽  
Jean-Sébastien Silvestre ◽  
Alma Zernecke ◽  
...  

AbstractRationaleMonocytes and macrophages have a critical and dual role in post-ischemic cardiac repair, as they can foster both tissue healing and damage. To decipher how monocytes/macrophages acquire heterogeneous functional phenotypes in the ischemic myocardium, we profiled the gene expression dynamics at the single-cell level in circulating and cardiac monocytes/macrophages following experimental myocardial infarction (MI) in mice.Methods and resultsUsing time-series single-cell transcriptome and cell surface epitope analysis of blood and cardiac monocytes/macrophages, as well as the integration of publicly available and independently generated single-cell RNA-seq data, we tracked the transitions in circulating and cardiac monocyte/macrophage states from homeostatic conditions up to 11 days after MI in mice. We show that MI induces marked and rapid transitions in the cardiac mononuclear phagocyte population, with almost complete disappearance of tissue resident macrophages 1 day after ischemia, and rapid infiltration of monocytes that locally acquire discrete and time-dependent transcriptional states within 3 to 7 days. Ischemic injury induced a shift of circulating monocytes towards granulocyte-like transcriptional features (Chil3, Lcn2, Prtn3). Trajectory inference analysis indicated that while conversion to Ly6Clow monocytes appears as the default fate of Ly6Chi monocytes in the blood, infiltrated monocytes acquired diverse gene expression signatures in the injured heart, notably transitioning to two main MI-associated macrophage populations characterized by MHCIIhi and Trem2hiIgf1hi gene expression signatures. Minor ischemia-associated macrophage populations with discrete gene expression signature suggesting specialized functions in e.g. iron handling or lipid metabolism were also observed. We further identified putative transcriptional regulators and new cell surface markers of cardiac monocyte/macrophage states.ConclusionsAltogether, our work provides a comprehensive landscape of circulating and cardiac monocyte/macrophage states and their regulators after MI, and will help to further understand their contribution to post-myocardial infarction heart repair.


2020 ◽  
Vol 38 (15_suppl) ◽  
pp. 4539-4539
Author(s):  
Anita Lavery ◽  
Leanne Stevenson ◽  
Damian McManus ◽  
Gemma E. Logan ◽  
Steven M. Walker ◽  
...  

4539 Background: The Dual Erb B Inhibition in Oesophago-gastric Cancer (DEBIOC) trial reported an acceptable safety profile for neoadjuvant Xelox +/- AZD8931 but limited efficacy. We utilized EAC patient samples from DEBIOC to evaluate the impact of neoadjuvant Xelox +/-AZD8931 on biological pathways using a unique software driven solution. Methods: 24 pre-treatment FFPE EAC biopsies and 17 matched surgical resection specimens were transcriptionally profiled using the Almac Diagnostics Xcel Array. Gene expression data was analyzed using the Almac claraT total mRNA report V3.0.0, reporting on 92 gene expression signatures and 7337 single genes associated with 10 key biologies. Paired Wilcoxon tests (5% significance level) were used to evaluate changes in claraT scores pre- and post-treatment. EGFR and Her2 expression were assessed by IHC and FISH. Results: 15 patients received Xelox+AZD8931 and 9 Xelox alone. Hierarchical clustering of biopsies identified 4 major clusters: Inflammation active, Genomic Instability active, EGFR & MAPK active, and EMT & Angiogenesis active. Comparison of signature scores pre- and post- neoadjuvant treatment demonstrated a significant reduction in scores relating to DNA damage repair (DDR) deficiency (Almac DNA Damage assay, p< 0.0001; BRCAness Profile, p= 0.0025; HRD Gene Signature, p< 0.0001; BRCA1ness Signature, p= 0.0004) and a significant increase in angiogenesis signatures (Almac Angiogenesis Assay, p= 0.0002; Angio Predictive G model, p= 0.0228; Angiogenesis Signature A, p= 0.0034) and EMT signatures (EMT Signature, p= 0.0031, EMT Enrichment Score, p= 0.0013, Pan-Can EMT Signature B, p= 0.0001). Comparing pre- and post-treatment signature scores in patients treated with Xelox +/-AZD8931 revealed a significant reduction in EGFR Sensitivity Signature ( p= 0.0088), ERBB2-specific Gene Expression Signature ( p= 0.0127) and Hallmark PI3K-AKT-MTOR Signaling ( p= 0.0195) in those treated with Xelox + AZD8931 in keeping with the mechanism of action of AZD8931. Downregulation of AKT signaling was confirmed in AZD8931 treated and resistant cell lines. Conclusions: We report the use of a novel software tool to apply 92 gene expression signatures to EAC biopsy and resection specimens from the DEBIOC trial to provide insight into mechanisms of action. Neoadjuvant treatment was associated with a reduction in DDR deficiency and an increase in angiogenesis and EMT signatures whilst a reduction in EGFR, Her2 and AKT pathways was noted with AZD8931 treatment.


Sign in / Sign up

Export Citation Format

Share Document