Search

Abstract Background Specific microglia responses are thought to contribute to the development and progression of neurodegenerative diseases, including Parkinson’s disease (PD). However, the phenotypic acquisition of microglial cells and their role during the underlying neuroinflammatory processes remain largely elusive. Here, according to the multiple-hit hypothesis, which stipulates that PD etiology is determined by a combination of genetics and various environmental risk factors, we investigate microglial transcriptional programs and morphological adaptations under PARK7/DJ-1 deficiency, a genetic cause of PD, during lipopolysaccharide (LPS)-induced inflammation. Methods Using a combination of single-cell RNA-sequencing, bulk RNA-sequencing, multicolor flow cytometry and immunofluorescence analyses, we comprehensively compared microglial cell phenotypic characteristics in PARK7/DJ-1 knock-out (KO) with wildtype littermate mice following 6- or 24-h intraperitoneal injection with LPS. For translational perspectives, we conducted corresponding analyses in human PARK7/DJ-1 mutant induced pluripotent stem cell (iPSC)-derived microglia and murine bone marrow-derived macrophages (BMDMs). Results By excluding the contribution of other immune brain resident and peripheral cells, we show that microglia acutely isolated from PARK7/DJ-1 KO mice display a distinct phenotype, specially related to type II interferon and DNA damage response signaling, when compared with wildtype microglia, in response to LPS. We also detected discrete signatures in human PARK7/DJ-1 mutant iPSC-derived microglia and BMDMs from PARK7/DJ-1 KO mice. These specific transcriptional signatures were reflected at the morphological level, with microglia in LPS-treated PARK7/DJ-1 KO mice showing a less amoeboid cell shape compared to wildtype mice, both at 6 and 24 h after acute inflammation, as also observed in BMDMs. Conclusions Taken together, our results show that, under inflammatory conditions, PARK7/DJ-1 deficiency skews microglia towards a distinct phenotype characterized by downregulation of genes involved in type II interferon signaling and a less prominent amoeboid morphology compared to wildtype microglia. These findings suggest that the underlying oxidative stress associated with the lack of PARK7/DJ-1 affects microglia neuroinflammatory responses, which may play a causative role in PD onset and progression.

Background: Epidemiological studies suggest that metformin reduces the risk of developing several types of cancer, including gliomas, and improves the overall survival in cancer patients. Nevertheless, while the effect of metformin on cancer cells has been extensively studied, its impact on other components of the tumour microenvironment, such as macrophages, is less understood. Results: Metformin-treated mouse bone marrow cells differentiate into spindle-shaped macrophages exhibiting increased phagocytic activity and tumour cell cytotoxicity coupled with modulated expression of co-stimulatory molecules displaying reduced sensitivity to inflammatory cues compared with untreated cells. Transcriptional analyses of metformin-treated mouse bone marrow-derived macrophages show decreased expression levels of pro-tumour genes, including Tgfbi and Il1β, related to enhanced mTOR/HIF1α signalling and metabolic rewiring towards glycolysis. Significance: Our study provides novel insights into the immunomodulatory properties of metformin in macrophages and its potential application in preventing tumour onset and in cancer immunotherapy.

Abstract Background A major contributing factor to glioblastoma (GBM) development and progression is its ability to evade the immune system by creating an immune-suppressive environment, where GBM-associated myeloid cells, including resident microglia and peripheral monocyte-derived macrophages, play critical pro-tumoral roles. However, it is unclear whether recruited myeloid cells are phenotypically and functionally identical in GBM patients and whether this heterogeneity is recapitulated in patient-derived orthotopic xenografts (PDOXs). A thorough understanding of the GBM ecosystem and its recapitulation in preclinical models is currently missing, leading to inaccurate results and failures of clinical trials. Methods Here, we report systematic characterization of the tumor microenvironment (TME) in GBM PDOXs and patient tumors at the single-cell and spatial levels. We applied single-cell RNA sequencing, spatial transcriptomics, multicolor flow cytometry, immunohistochemistry, and functional studies to examine the heterogeneous TME instructed by GBM cells. GBM PDOXs representing different tumor phenotypes were compared to glioma mouse GL261 syngeneic model and patient tumors. Results We show that GBM tumor cells reciprocally interact with host cells to create a GBM patient-specific TME in PDOXs. We detected the most prominent transcriptomic adaptations in myeloid cells, with brain-resident microglia representing the main population in the cellular tumor, while peripheral-derived myeloid cells infiltrated the brain at sites of blood–brain barrier disruption. More specifically, we show that GBM-educated microglia undergo transition to diverse phenotypic states across distinct GBM landscapes and tumor niches. GBM-educated microglia subsets display phagocytic and dendritic cell-like gene expression programs. Additionally, we found novel microglial states expressing cell cycle programs, astrocytic or endothelial markers. Lastly, we show that temozolomide treatment leads to transcriptomic plasticity and altered crosstalk between GBM tumor cells and adjacent TME components. Conclusions Our data provide novel insights into the phenotypic adaptation of the heterogeneous TME instructed by GBM tumors. We show the key role of microglial phenotypic states in supporting GBM tumor growth and response to treatment. Our data place PDOXs as relevant models to assess the functionality of the TME and changes in the GBM ecosystem upon treatment. Graphical Abstract

Abstract Early-life adversity covers a range of physical, social and environmental stressors. Acute viral infections in early life are a major source of such adversity and have been associated with a broad spectrum of later-life effects outside the immune system or “off-target”. These include an altered hypothalamus–pituitary–adrenal (HPA) axis and metabolic reactions. Here, we used a murine post-natal day 14 (PND 14) Influenza A (H1N1) infection model and applied a semi-holistic approach including phenotypic measurements, gene expression arrays and diffusion neuroimaging techniques to investigate HPA axis dysregulation, energy metabolism and brain connectivity. By PND 56 the H1N1 infection had been resolved, and there was no residual gene expression signature of immune cell infiltration into the liver, adrenal gland or brain tissues examined nor of immune-related signalling. A resolved early-life H1N1 infection had sex-specific effects. We observed retarded growth of males and altered pre-stress (baseline) blood glucose and corticosterone levels at PND42 after the infection was resolved. Cerebral MRI scans identified reduced connectivity in the cortex, midbrain and cerebellum that were accompanied by tissue-specific gene expression signatures. Gene set enrichment analysis confirmed that these were tissue-specific changes with few common pathways. Early-life infection independently affected each of the systems and this was independent of HPA axis or immune perturbations.

We provide in this paper a comprehensive comparison of various transfer learning strategies and deep learning architectures for computer-aided classification of adult-type diffuse gliomas. We evaluate the generalizability of out-of-domain ImageNet representations for a target domain of histopathological images, and study the impact of in-domain adaptation using self-supervised and multi-task learning approaches for pretraining the models using the medium-to-large scale datasets of histopathological images. A semi-supervised learning approach is furthermore proposed, where the fine-tuned models are utilized to predict the labels of unannotated regions of the whole slide images (WSI). The models are subsequently retrained using the ground-truth labels and weak labels determined in the previous step, providing superior performance in comparison to standard in-domain transfer learning with balanced accuracy of 96.91% and F1-score 97.07%, and minimizing the pathologist's efforts for annotation. Finally, we provide a visualization tool working at WSI level which generates heatmaps that highlight tumor areas; thus, providing insights to pathologists concerning the most informative parts of the WSI.

This article aims to investigate sufficient conditions for the stability of the trivial solution of stochastic differential equations with a random structure, particularly in contexts involving the presence of concentration points. The proof of asymptotic stability leverages the use of Lyapunov functions, supplemented by additional constraints on the magnitudes of jumps and jump times, as well as the Markov property of the system solutions. The findings are elucidated with an example, demonstrating both stable and unstable conditions of the system. The novelty of this work is in the consideration of jump concentration points, which are not considered in classical works. The assumption of the existence of concentration points leads to additional constraints on jumps, jump times and relations between them.

Abstract Melanoma has the highest propensity of all cancers to metastasize to the brain with a large percentage of late‐stage patients developing metastases in the central nervous system (CNS). It is well known that metastasis establishment, cell survival, and progression are affected by tumour‐host cell interactions where changes in the host cellular compartments likely play an important role. In this context, miRNAs transferred by tumour derived extracellular vesicles (EVs) have previously been shown to create a favourable tumour microenvironment. Here, we show that miR‐146a‐5p is highly expressed in human melanoma brain metastasis (MBM) EVs, both in MBM cell lines as well as in biopsies, thereby modulating the brain metastatic niche. Mechanistically, miR‐146a‐5p was transferred to astrocytes via EV delivery and inhibited NUMB in the Notch signalling pathway. This resulted in activation of tumour‐promoting cytokines (IL‐6, IL‐8, MCP‐1 and CXCL1). Brain metastases were significantly reduced following miR‐146a‐5p knockdown. Corroborating these findings, miR‐146a‐5p inhibition led to a reduction of IL‐6, IL‐8, MCP‐1 and CXCL1 in astrocytes. Following molecular docking analysis, deserpidine was identified as a functional miR‐146a‐5p inhibitor, both in vitro and in vivo. Our results highlight the pro‐metastatic function of miR‐146a‐5p in EVs and identifies deserpidine for targeted adjuvant treatment.

Summary: Treatment options for patients with NRAS-mutant melanoma are limited and lack an efficient targeted drug combination that significantly increases overall and progression-free survival. In addition, targeted therapy success is hampered by the inevitable emergence of drug resistance. A thorough understanding of the molecular processes driving cancer cells’ escape mechanisms is crucial to tailor more efficient follow-up therapies. We performed single-cell RNA sequencing of NRAS-mutant melanoma treated with MEK1/2 plus CDK4/6 inhibitors to decipher transcriptional transitions during the development of drug resistance. Cell lines resuming full proliferation (FACs [fast-adapting cells]) and cells that became senescent (SACs [slow-adapting cells]) over prolonged treatment were identified. The early drug response was characterized by transitional states involving increased ion signaling, driven by upregulation of the ATP-gated ion channel P2RX7. P2RX7 activation was associated with improved therapy responses and, in combination with targeted drugs, could contribute to the delayed onset of acquired resistance in NRAS-mutant melanoma.

People with COVID-19 can experience impairing symptoms that require enhanced surveillance. Our objective was to train an artificial intelligence-based model to predict the presence of COVID-19 symptoms and derive a digital vocal biomarker for easily and quantitatively monitoring symptom resolution. We used data from 272 participants in the prospective Predi-COVID cohort study recruited between May 2020 and May 2021. A total of 6473 voice features were derived from recordings of participants reading a standardized pre-specified text. Models were trained separately for Android devices and iOS devices. A binary outcome (symptomatic versus asymptomatic) was considered, based on a list of 14 frequent COVID-19 related symptoms. A total of 1775 audio recordings were analyzed (6.5 recordings per participant on average), including 1049 corresponding to symptomatic cases and 726 to asymptomatic ones. The best performances were obtained from Support Vector Machine models for both audio formats. We observed an elevated predictive capacity for both Android (AUC = 0.92, balanced accuracy = 0.83) and iOS (AUC = 0.85, balanced accuracy = 0.77) as well as low Brier scores (0.11 and 0.16 respectively for Android and iOS when assessing calibration. The vocal biomarker derived from the predictive models accurately discriminated asymptomatic from symptomatic individuals with COVID-19 (t-test P-values

Advanced sequencing technologies such as RNASeq provide the means for production of massive amounts of data, including transcriptome-wide expression levels of coding RNAs (mRNAs) and non-coding RNAs such as miRNAs, lncRNAs, piRNAs and many other RNA species. In silico analysis of datasets, representing only one RNA species is well established and a variety of tools and pipelines are available. However, attaining a more systematic view of how different players come together to regulate the expression of a gene or a group of genes requires a more intricate approach to data analysis. To fully understand complex transcriptional networks, datasets representing different RNA species need to be integrated. In this review, we will focus on miRNAs as key post-transcriptional regulators summarizing current computational approaches for miRNA:target gene prediction as well as new data-driven methods to tackle the problem of comprehensively and accurately dissecting miRNome-targetome interactions.

Glioblastomas (GBMs) are highly invasive brain tumours, but the underlying mechanisms of GBM invasion are unclear. Here, the authors perform an RNA interference screen and identify AN1-Type Zinc Finger protein 3 (ZFAND3) as a regulator of GBM invasion, and find that it acts through the transcriptional regulation of invasion-related genes.

An optimal control for a dynamical system optimizes a certain objective function. Here, we consider the construction of an optimal control for a stochastic dynamical system with a random structure, Poisson perturbations and random jumps, which makes the system stable in probability. Sufficient conditions of the stability in probability are obtained, using the second Lyapunov method, in which the construction of the corresponding functions plays an important role. Here, we provide a solution to the problem of optimal stabilization in a general case. For a linear system with a quadratic quality function, we give a method of synthesis of optimal control based on the solution of Riccati equations. Finally, in an autonomous case, a system of differential equations was constructed to obtain unknown matrices that are used for the construction of an optimal control. The method using a small parameter is justified for the algorithmic search of an optimal control. This approach brings a novel solution to the problem of optimal stabilization for a stochastic dynamical system with a random structure, Markov switches and Poisson perturbations.

Abstract Background The amount of publicly available cancer-related “omics” data is constantly growing and can potentially be used to gain insights into the tumour biology of new cancer patients, their diagnosis and suitable treatment options. However, the integration of different datasets is not straightforward and requires specialized approaches to deal with heterogeneity at technical and biological levels. Methods Here we present a method that can overcome technical biases, predict clinically relevant outcomes and identify tumour-related biological processes in patients using previously collected large discovery datasets. The approach is based on independent component analysis (ICA) – an unsupervised method of signal deconvolution. We developed parallel consensus ICA that robustly decomposes transcriptomics datasets into expression profiles with minimal mutual dependency. Results By applying the method to a small cohort of primary melanoma and control samples combined with a large discovery melanoma dataset, we demonstrate that our method distinguishes cell-type specific signals from technical biases and allows to predict clinically relevant patient characteristics. We showed the potential of the method to predict cancer subtypes and estimate the activity of key tumour-related processes such as immune response, angiogenesis and cell proliferation. ICA-based risk score was proposed and its connection to patient survival was validated with an independent cohort of patients. Additionally, through integration of components identified for mRNA and miRNA data, the proposed method helped deducing biological functions of miRNAs, which would otherwise not be possible. Conclusions We present a method that can be used to map new transcriptomic data from cancer patient samples onto large discovery datasets. The method corrects technical biases, helps characterizing activity of biological processes or cell types in the new samples and provides the prognosis of patient survival.

Cancer stem cells (CSCs) comprise a putative population that can drive growth and resistance. Here, in glioblastoma models the authors show that rather than being a distinct clonal entity, the CSC population represents a plastic state adoptable by most cancer cells via reversible state transitions induced by the microenvironment.

Summary: C/EBPα represents a paradigm intrinsically disordered transcription factor containing short linear motifs and post-translational modifications (PTM). Unraveling C/EBPα protein interaction networks is a prerequisite for understanding the multi-modal functions of C/EBPα in hematopoiesis and leukemia. Here, we combined arrayed peptide matrix screening (PRISMA) with BioID to generate an in vivo validated and isoform specific interaction map of C/EBPα. The myeloid C/EBPα interactome comprises promiscuous and PTM-regulated interactions with protein machineries involved in gene expression, epigenetics, genome organization, DNA replication, RNA processing, and nuclear transport. C/EBPα interaction hotspots coincide with homologous conserved regions of the C/EBP family that also score as molecular recognition features. PTMs alter the interaction spectrum of C/EBP-motifs to configure a multi-valent transcription factor hub that interacts with multiple co-regulatory components, including BAF/SWI-SNF or Mediator complexes. Combining PRISMA and BioID is a powerful strategy to systematically explore the PTM-regulated interactomes of intrinsically disordered transcription factors.

Clinical studies in glioblastoma and pancreatic carcinoma patients strongly support the further development of H-1 protoparvovirus (H-1PV)-based anticancer therapies. The identification of cellular factors involved in the H-1PV life cycle may provide the knowledge to improve H-1PV anticancer potential. Recently, we showed that sialylated laminins mediate H-1PV attachment at the cell membrane. In this study, we revealed that H-1PV also interacts at the cell surface with galectin-1 and uses this glycoprotein to enter cancer cells. Indeed, knockdown/out of LGALS1, the gene encoding galectin-1, strongly decreases the ability of H-1PV to infect and kill cancer cells. This ability is rescued by the re-introduction of LGALS1 into cancer cells. Pre-treatment with lactose, which is able to bind to galectins and modulate their cellular functions, decreased H-1PV infectivity in a dose dependent manner. In silico analysis reveals that LGALS1 is overexpressed in various tumours including glioblastoma and pancreatic carcinoma. We show by immunohistochemistry analysis of 122 glioblastoma biopsies that galectin-1 protein levels vary between tumours, with levels in recurrent glioblastoma higher than those in primary tumours or normal tissues. We also find a direct correlation between LGALS1 transcript levels and H-1PV oncolytic activity in 53 cancer cell lines from different tumour origins. Strikingly, the addition of purified galectin-1 sensitises poorly susceptible GBM cell lines to H-1PV killing activity by rescuing cell entry. Together, these findings demonstrate that galectin-1 is a crucial determinant of the H-1PV life cycle.

Abstract Background One of the main current challenges in computational biology is to make sense of the huge amounts of multidimensional experimental data that are being produced. For instance, large cohorts of patients are often screened using different high-throughput technologies, effectively producing multiple patient-specific molecular profiles for hundreds or thousands of patients. Results We propose and implement a network-based method that integrates such patient omics data into Patient Similarity Networks. Topological features derived from these networks were then used to predict relevant clinical features. As part of the 2017 CAMDA challenge, we have successfully applied this strategy to a neuroblastoma dataset, consisting of genomic and transcriptomic data. In particular, we observe that models built on our network-based approach perform at least as well as state of the art models. We furthermore explore the effectiveness of various topological features and observe, for instance, that redundant centrality metrics can be combined to build more powerful models. Conclusion We demonstrate that the networks inferred from omics data contain clinically relevant information and that patient clinical outcomes can be predicted using only network topological data. Reviewers This article was reviewed by Yang-Yu Liu, Tomislav Smuc and Isabel Nepomuceno.

Abstract Background RNA sequencing (RNA-seq) and microarrays are two transcriptomics techniques aimed at the quantification of transcribed genes and their isoforms. Here we compare the latest Affymetrix HTA 2.0 microarray with Illumina 2000 RNA-seq for the analysis of patient samples - normal lung epithelium tissue and squamous cell carcinoma lung tumours. Protein coding mRNAs and long non-coding RNAs (lncRNAs) were included in the study. Results Both platforms performed equally well for protein-coding RNAs, however the stochastic variability was higher for the sequencing data than for microarrays. This reduced the number of differentially expressed genes and genes with predictive potential for RNA-seq compared to microarray data. Analysis of this variability revealed a lack of reads for short and low abundant genes; lncRNAs, being shorter and less abundant RNAs, were found especially susceptible to this issue. A major difference between the two platforms was uncovered by analysis of alternatively spliced genes. Investigation of differential exon abundance showed insufficient reads for many exons and exon junctions in RNA-seq while the detection on the array platform was more stable. Nevertheless, we identified 207 genes which undergo alternative splicing and were consistently detected by both techniques. Conclusions Despite the fact that the results of gene expression analysis were highly consistent between Human Transcriptome Arrays and RNA-seq platforms, the analysis of alternative splicing produced discordant results. We concluded that modern microarrays can still outperform sequencing for standard analysis of gene expression in terms of reproducibility and cost.

Background: The topological analysis of networks extracted from different types of “omics” data is a useful strategy for characterizing biologically meaningful properties of the complex systems underlying these networks. In particular, the biological significance of highly connected genes in diverse molecular networks has been previously determined using data from several model organisms and phenotypes. Despite such insights, the predictive potential of candidate hubs in gene co-expression networks in the specific context of cancer-related drug experiments remains to be deeply investigated. The examination of such associations may offer opportunities for the accurate prediction of anticancer drug responses. Methods: Here, we address this problem by: a) analyzing a co-expression network obtained from thousands of cancer cell lines, b) detecting significant network hubs, and c) assessing their capacity to predict drug sensitivity using data from thousands of drug experiments. We investigated the prediction capability of those genes using a multiple linear regression model, independent datasets, comparisons with other models and our own in vitro experiments. Results: These analyses led to the identification of 47 hub genes, which are implicated in a diverse range of cancer-relevant processes and pathways. Overall, encouraging agreements between predicted and observed drug sensitivities were observed in public datasets, as well as in our in vitro validations for four glioblastoma cell lines and four drugs. To facilitate further research, we share our hub-based drug sensitivity prediction model as an online tool. Conclusions: Our research shows that co-expression network hubs are biologically interesting and exhibit potential for predicting drug responses in vitro. These findings motivate further investigations about the relevance and application of our unbiased discovery approach in pre-clinical, translationally-oriented research.

The present data are related to the article entitled “Insights into ligand stimulation effects on gastro-intestinal stromal tumors signaling” (C. Bahlawane, M. Schmitz, E. Letellier, K. Arumugam, N. Nicot, P.V. Nazarov, S. Haan, 2016) [1]. Constitutive and ligand-derived signaling pathways mediated by KIT and PDGFRA mutated proteins found in gastrointestinal stromal tumors (GIST) were compared. Expression of mutant proteins was induced by doxycycline in an isogenic background (Hek293 cells). Kit was identified by FACS at the cell surface and found to be quickly degraded or internalized upon SCF stimulation for both Kit Wild type and Kit mutant counterparts. Investigation of the main activated pathways in GIST unraveled a new feature specific for oncogenic KIT mutants, namely their ability to be further activated by Kit ligand, the stem cell factor (scf). We were also able to identify the MAPK pathway as the most prominent target for a common inhibition of PDGFRA and KIT oncogenic signaling. Western blotting and micro-array analysis were applied to analyze the capacities of the mutant to induce an effective STATs response. Among all Kit mutants, only Kit Ex11 deletion mutant was able to elicit an effective STATs response whereas all PDGFRA were able to do so.

Abstract The zebrafish has the capacity to regenerate its heart after severe injury. While the function of a few genes during this process has been studied, we are far from fully understanding how genes interact to coordinate heart regeneration. To enable systematic insights into this phenomenon, we generated and integrated a dynamic co-expression network of heart regeneration in the zebrafish and linked systems-level properties to the underlying molecular events. Across multiple post-injury time points, the network displays topological attributes of biological relevance. We show that regeneration steps are mediated by modules of transcriptionally coordinated genes, and by genes acting as network hubs. We also established direct associations between hubs and validated drivers of heart regeneration with murine and human orthologs. The resulting models and interactive analysis tools are available at http://infused.vital-it.ch . Using a worked example, we demonstrate the usefulness of this unique open resource for hypothesis generation and in silico screening for genes involved in heart regeneration.

Long non-coding RNAs (lncRNAs) have recently emerged as a novel group of non-coding RNAs able to regulate gene expression. While their role in cardiac disease is only starting to be understood, their involvement in cardiac hypertrophy is poorly known. We studied the association between lncRNAs and left ventricular hypertrophy using whole transcriptome microarrays. Wild-type mice and mice overexpressing the adenosine A2A receptor were subjected to transverse aortic constriction (TAC) to induce left ventricular hypertrophy. Expression profiles of lncRNAs in the heart were characterized using genome-wide microarrays. An analytical pipeline was specifically developed to extract lncRNA data from microarrays. We identified 2 lncRNAs up-regulated and 3 lncRNAs down-regulated in the hearts of A2A-receptor overexpressing-mice subjected to TAC compared to wild-type mice. Differential expression of these 2 lncRNAs was validated by quantitative PCR. Complete microarray dataset is available at Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE45423. Here, we describe in details the experimental design, microarray performance and analysis.

Background: The topological analysis of networks extracted from different types of “omics” data is a useful strategy for characterizing biologically meaningful properties of the complex systems underlying these networks. In particular, the biological significance of highly connected genes in diverse molecular networks has been previously determined using data from several model organisms and phenotypes. Despite such insights, the predictive potential of candidate hubs in gene co-expression networks in the specific context of cancer-related drug experiments remains to be deeply investigated. The examination of such associations may offer opportunities for the accurate prediction of anticancer drug responses. Methods: Here, we address this problem by: a) analyzing a co-expression network obtained from thousands of cancer cell lines, b) detecting significant network hubs, and c) assessing their capacity to predict drug sensitivity using data from thousands of drug experiments. We investigated the prediction capability of those genes using a multiple linear regression model, independent datasets, comparisons with other models and our own in vitro experiments. Results: These analyses led to the identification of 47 hub genes, which are implicated in a diverse range of cancer-relevant processes and pathways. Overall, encouraging agreements between predicted and observed drug sensitivities were observed in public datasets, as well as in our in vitro validations for four glioblastoma cell lines and four drugs. To facilitate further research, we share our hub-based drug sensitivity prediction model as an online tool. Conclusions: Our research shows that co-expression network hubs are biologically interesting and exhibit potential for predicting drug responses in vitro. These findings motivate further investigations about the relevance and application of our unbiased discovery approach in pre-clinical, translationally-oriented research.