CONNECTOR, fitting and clustering of longitudinal data to reveal a new risk stratification system.

MOTIVATION: The transition from evaluating a single time point to examining the entire dynamic evolution of a system is possible only in the presence of the proper framework. The strong variability of dynamic evolution makes the definition of an explanatory procedure for data fitting and clustering challenging. RESULTS: We developed CONNECTOR, a data-driven framework able to analyze and inspect longitudinal data in a straightforward and revealing way. When used to analyze tumor growth kinetics over time in 1599 patient-derived xenograft growth curves from ovarian and colorectal cancers, CONNECTOR allowed the aggregation of time-series data through an unsupervised approach in informative clusters. We give a new perspective of mechanism interpretation, specifically, we define novel model aggregations and we identify unanticipated molecular associations with response to clinically approved therapies. AVAILABILITY AND IMPLEMENTATION: CONNECTOR is freely available under GNU GPL license at https://qbioturin.github.io/connector and https://doi.org/10.17504/protocols.io.8epv56e74g1b/v1.

Cancer Cells Haploinsufficient for ATM Are Sensitized to PARP Inhibitors by MET Inhibition.

The MET oncogene encodes a tyrosine kinase (TK) receptor. Its activation protects cells from death but also stimulates DNA damage response by triggering excess replicative stress. Transcriptomic classification of cancer cell lines based on MET expression showed that response to the PARP inhibitor (PARPi) olaparib is poorer in MET overexpressing cell lines. Accordingly, a high MET expressing lung carcinoma cell line was sensitized to PARPi by MET TK inhibition. This was not linked solely to MET overexpression: other MET overexpressing cell lines were biochemically but not functionally responsive to combined inhibition. Moreover, exogenously induced MET overexpression was unable to induce resistance to PARPi. The MET overexpressing cell line, responsive to the combined PARP and MET inhibition, carried a heterozygous mutation of the ATM gene and showed an attenuated response of ATM to PARPi. Among the downstream targets of ATM activation, NuMA was phosphorylated only in response to the combined PARP and MET inhibition. Given the role played by NuMA in mitosis, data show that the latter is affected by MET and PARP inhibition in cells with haploinsufficient ATM. This is important as ATM heterozygous mutation is frequently found in human cancer and in lung carcinomas in particular.

Ovarian Cancer Cells in Ascites Form Aggregates That Display a Hybrid Epithelial-Mesenchymal Phenotype and Allows Survival and Proliferation of Metastasizing Cells.

Peritoneal metastases are the leading cause of morbidity and mortality in ovarian cancer. Cancer cells float in peritoneal fluid, named ascites, together with a definitely higher number of non neo-neoplastic cells, as single cells or multicellular aggregates. The aim of this work is to uncover the features that make these aggregates the metastasizing units. Immunofluorescence revealed that aggregates are made almost exclusively of ovarian cancer cells expressing the specific nuclear PAX8 protein. The same cells expressed epithelial and mesenchymal markers, such as EPCAM and αSMA, respectively. Expression of fibronectin further supported a hybrid epithelia-mesenchymal phenotype, that is maintained when aggregates are cultivated and proliferate. Hematopoietic cells as well as macrophages are negligible in the aggregates, while abundant in the ascitic fluid confirming their prominent role in establishing an eco-system necessary for the survival of ovarian cancer cells. Using ovarian cancer cell lines, we show that cells forming 3D structures neo-expressed thoroughly fibronectin and αSMA. Functional assays showed that αSMA and fibronectin are necessary for the compaction and survival of 3D structures. Altogether these data show that metastasizing units display a hybrid phenotype that allows maintenance of the 3D structures and the plasticity necessary for implant and seeding into peritoneal lining.

Different miRNA Profiles in Plasma Derived Small and Large Extracellular Vesicles from Patients with Neurodegenerative Diseases.

Identifying biomarkers is essential for early diagnosis of neurodegenerative diseases (NDs). Large (LEVs) and small extracellular vesicles (SEVs) are extracellular vesicles (EVs) of different sizes and biological functions transported in blood and they may be valid biomarkers for NDs. The aim of our study was to investigate common and different miRNA signatures in plasma derived LEVs and SEVs of Alzheimer's disease (AD), Parkinson's disease (PD), Amyotrophic Lateral Sclerosis (ALS) and Fronto-Temporal Dementia (FTD) patients. LEVs and SEVs were isolated from plasma of patients and healthy volunteers (CTR) by filtration and differential centrifugation and RNA was extracted. Small RNAs libraries were carried out by Next Generation Sequencing (NGS). MiRNAs discriminate all NDs diseases from CTRs and they can provide a signature for each NDs. Common enriched pathways for SEVs were instead linked to ubiquitin mediated proteolysis and Toll-like receptor signaling pathways and for LEVs to neurotrophin signaling and Glycosphingolipid biosynthesis pathway. LEVs and SEVs are involved in different pathways and this might give a specificity to their role in the spreading of the disease. The study of common and different miRNAs transported by LEVs and SEVs can be of great interest for biomarker discovery and for pathogenesis studies in neurodegeneration.

Reproducible bioinformatics project: a community for reproducible bioinformatics analysis pipelines.

BACKGROUND: Reproducibility of a research is a key element in the modern science and it is mandatory for any industrial application. It represents the ability of replicating an experiment independently by the location and the operator. Therefore, a study can be considered reproducible only if all used data are available and the exploited computational analysis workflow is clearly described. However, today for reproducing a complex bioinformatics analysis, the raw data and the list of tools used in the workflow could be not enough to guarantee the reproducibility of the results obtained. Indeed, different releases of the same tools and/or of the system libraries (exploited by such tools) might lead to sneaky reproducibility issues. RESULTS: To address this challenge, we established the Reproducible Bioinformatics Project (RBP), which is a non-profit and open-source project, whose aim is to provide a schema and an infrastructure, based on docker images and R package, to provide reproducible results in Bioinformatics. One or more Docker images are then defined for a workflow (typically one for each task), while the workflow implementation is handled via R-functions embedded in a package available at github repository. Thus, a bioinformatician participating to the project has firstly to integrate her/his workflow modules into Docker image(s) exploiting an Ubuntu docker image developed ad hoc by RPB to make easier this task. Secondly, the workflow implementation must be realized in R according to an R-skeleton function made available by RPB to guarantee homogeneity and reusability among different RPB functions. Moreover she/he has to provide the R vignette explaining the package functionality together with an example dataset which can be used to improve the user confidence in the workflow utilization. CONCLUSIONS: Reproducible Bioinformatics Project provides a general schema and an infrastructure to distribute robust and reproducible workflows. Thus, it guarantees to final users the ability to repeat consistently any analysis independently by the used UNIX-like architecture.

Heat-shock protein 27 (HSP27, HSPB1) is up-regulated by MET kinase inhibitors and confers resistance to MET-targeted therapy.

The tyrosine kinase encoded by the MET oncogene is activated by gene mutation or amplification in tumors, which in most instances maintain addiction, i.e., dependency, to MET activation. This makes MET an attractive candidate for targeted therapies. Here we show that, in 3/3 MET-addicted human gastric cancer cell lines, MET kinase inhibition resulted in a 3- to 4-fold increased expression of the antiapoptotic small heat-shock protein of 27 kDa (HSP27, HSPB1). HSP27 increase depended on the inhibition of the MEK/ERK pathway and on heat-shock factor 1 (HSF1) and hypoxia-inducible factor-1α (HIF-1α) regulation. Importantly, HSP27-silenced MET-addicted cells underwent 2- and 3-fold more apoptosis following MET inhibition in vitro and in vivo, respectively. Likewise, in human cancer cells susceptible to epidermal growth factor receptor (EGFR) inhibition, EGFR inhibitors induced HSP27 expression and were strengthened by HSP27 suppression. In control cell lines that were not affected by drugs targeting MET or EGFR, these drugs did not induce HSP27 increase. Therefore, in cancer therapies targeting the MET pathway, the induction of HSP27 might limit the efficacy of anti-MET agents. As HSP27 increase also impairs the effectiveness of EGFR inhibitors and is known to protect cells from chemotherapeutics, the induction of HSP27 by targeted agents might strongly affect the success of combination treatments.

PIM2 kinase is induced by cisplatin in ovarian cancer cells and limits drug efficacy.

Platinum-based chemotherapy is widely used to treat various cancers, but many patients ultimately relapse due to drug resistance. We employed phosphoproteomic analysis and functional assays of the response of SK-OV-3 ovarian cancer cells to cisplatin as a strategy to identify kinases as candidate druggable targets to sensitize cells to platinum. A SILAC-based approach combined with TiO2-based phosphopeptide enrichment allowed the direct identification of ERK1/2, p90RSK, and ERBB2 as kinases whose phosphorylation is regulated by cisplatin. Bioinformatic analysis revealed enrichment in linear phosphorylation motifs predicted to be targets of p38MAPK, CDK2, and PIM2. All three PIM kinases were found expressed in a panel of 10 ovarian cancer cell lines, with the oncogenic PIM2 being the most commonly induced by cisplatin. Targeting PIM2 kinase by either biochemical inhibitors or RNA interference impaired cell growth, decreased cisplatin-triggered BAD phosphorylation, and sensitized ovarian cancer cells to drug-induced apoptosis. Overexpression of PIM2 triggered anchorage-independent growth and resulted in increased BAD phosphorylation and cell resistance to DNA damaging agents. The data show that the PIM2 kinase plays a role in the response of ovarian cancer cells to platinum drugs and suggest that PIM inhibitors may find clinical application as an adjunct to platinum-based therapies.

HSP27 is required for invasion and metastasis triggered by hepatocyte growth factor.

The hepatocyte growth factor (HGF) also known as scatter factor activates cancer cell invasion and metastasis. We show that in ovarian cancer cells HGF induced the phosphorylation of the small heat shock protein of 27 kDa (HSP27) by activating the p38MAPK. HSP27 is increased in many cancers at advanced stage including ovarian cancer and associated with cancer resistance to therapy and poor patients' survival. The phosphorylation of HSP27 regulates both its chaperone activity and its control of cytoskeletal stability. We show that HSP27 was necessary for the remodeling of actin filaments induced by HGF and that motility in vitro depended on the p38MAPK-MK2 axis. In vivo, HSP27 silencing impaired the ability of the highly metastatic, HGF-secreting ovarian cancer cells to give rise to spontaneous metastases. This was due to defective motility across the vessel wall and reduced growth. Indeed, HSP27 silencing impaired the ability of circulating ovarian cancer cells to home to the lungs and to form experimental hematogenous metastases and the capability of cancer cells to grow as subcutaneous xenografts. Moreover, HSP27 suppression resulted in the sensitization of xenografts to low doses of the chemotherapeutic paclitaxel, likely because HSP27 protected microtubules from bundling caused by the drug. Altogether, these data show that the HSP27 is required for the proinvasive and prometastatic activity of HGF and suggest that HSP27 might be not only a marker of progression of ovarian cancer, but also a suitable target for therapy.

AKT activation drives the nuclear localization of CSE1L and a pro-oncogenic transcriptional activation in ovarian cancer cells.

The human homolog of the yeast cse1 gene (CSE1L) is over-expressed in ovarian cancer. CSE1L forms complex with Ran and importin-α and has roles in nucleocytoplasmic traffic and gene expression. CSE1L accumulated in the nucleus of ovarian cancer cell lines, while it was localized also in the cytoplasm of other cancer cell lines. Nuclear localization depended on AKT, which was constitutively active in ovarian cancer cells, as the CSE1L protein translocated to the cytoplasm when AKT was inactivated. Moreover, the expression of a constitutively active AKT forced the translocation of CSE1L from the cytoplasm to the nucleus in other cancer cells. Nuclear accrual of CSE1L was associated to the nuclear accumulation of the phosphorylated Ran Binding protein 3 (RanBP3), which depended on AKT as well. Also in samples of human ovarian cancer, AKT activation was associated to nuclear accumulation of CSE1L and phosphorylation of RanBP3. Expression profiling of ovarian cancer cells after CSE1L silencing showed that CSE1L was required for the expression of genes promoting invasion and metastasis. In agreement, CSE1L silencing impaired motility and invasiveness of ovarian cancer cells. Altogether these data show that in ovarian cancer cells activated AKT by affecting RanBP3 phosphorylation determines the nuclear accumulation of CSE1L and likely the nuclear concentration of transcription factors conveying pro-oncogenic signals.

Additional copies of 1q negatively impact the outcome of multiple myeloma patients and induce transcriptomic deregulation in malignant plasma cells.

Additional copies of chromosome 1 long arm (1q) are frequently found in multiple myeloma (MM) and predict high-risk disease. Available data suggest a different outcome and biology of patients with amplification (Amp1q, ≥4 copies of 1q) vs. gain (Gain1q, 3 copies of 1q) of 1q. We evaluated the impact of Amp1q/Gain1q on the outcome of newly diagnosed MM patients enrolled in the FORTE trial (NCT02203643). Among 400 patients with available 1q data, 52 (13%) had Amp1q and 129 (32%) Gain1q. After a median follow-up of 62 months, median progression-free survival (PFS) was 21.2 months in the Amp1q group, 54.9 months in Gain1q, and not reached (NR) in Normal 1q. PFS was significantly hampered by the presence of Amp1q (HR 3.34 vs. Normal 1q, P < 0.0001; HR 1.99 vs. Gain1q, P = 0.0008). Patients with Gain1q had also a significantly shorter PFS compared with Normal 1q (HR 1.68, P = 0.0031). Concomitant poor prognostic factors or the failure to achieve MRD negativity predicted a median PFS < 12 months in Amp1q patients. Carfilzomib-lenalidomide-dexamethasone plus autologous stem cell transplantation treatment improved the adverse effect of Gain1q but not Amp1q. Transcriptomic data showed that additional 1q copies were associated with deregulation in apoptosis signaling, p38 MAPK signaling, and Myc-related genes.

The cellular apoptosis susceptibility CAS/CSE1L gene protects ovarian cancer cells from death by suppressing RASSF1C.

The cellular apoptosis susceptibility gene CAS/CSE1L is overexpressed in cancer, although it was originally identified as a gene that renders cells vulnerable to apoptotic stimuli. CAS/CSE1L has roles in the nucleocytoplasmic recycling of importin-α and in the regulation of gene expression, cell migration, and secretion. We identified CAS/CSE1L as a survival factor for ovarian cancer cells in vitro and in vivo. In 3/3 ovarian cancer cell lines, CAS/CSE1L was down-modulated by the unorthodox proapoptotic signaling of the MET receptor. CAS/CSE1L knockdown with RNA interference committed the ovarian cancer cells to death, but not immortalized normal cells and breast and colon cancer cells. In 70 and 95% of these latter cells, respectively, CAS/CSE1L was localized in the cytoplasm, while it accumulated in the nucleus in >90% of ovarian cancer cells. Nuclear localization depended on AKT, which was constitutively active in ovarian cancer cells. In the nucleus, CAS/CSE1L regulated the expression of the proapoptotic Ras-association domain family 1 gene products RASSF1C and RASSF1A, which mediated death signals evoked by depletion of CAS/CSE1L. Our data show that CAS/CSE1L protects ovarian cancer cells from death through transcriptional suppression of a proapoptotic gene and suggest that the localization of CAS/CSE1L dictates its function.

Single-Cell RNAseq Data QC and Preprocessing.

The first step in single-cell RNAseq data analysis is the evaluation of the overall quality of the cell transcriptome and the preparation of the single-cell transcription data for clustering. In this chapter, we describe one of the possible approaches to perform single-cell data preprocessing for 3' end single-cell RNAseq transcriptomics data.

Targeting CCR7-PI3Kγ overcomes resistance to tyrosine kinase inhibitors in ALK-rearranged lymphoma.

Anaplastic lymphoma kinase (ALK) tyrosine kinase inhibitors (TKIs) show potent efficacy in several ALK-driven tumors, but the development of resistance limits their long-term clinical impact. Although resistance mechanisms have been studied extensively in ALK-driven non-small cell lung cancer, they are poorly understood in ALK-driven anaplastic large cell lymphoma (ALCL). Here, we identify a survival pathway supported by the tumor microenvironment that activates phosphatidylinositol 3-kinase γ (PI3K-γ) signaling through the C-C motif chemokine receptor 7 (CCR7). We found increased PI3K signaling in patients and ALCL cell lines resistant to ALK TKIs. PI3Kγ expression was predictive of a lack of response to ALK TKI in patients with ALCL. Expression of CCR7, PI3Kγ, and PI3Kδ were up-regulated during ALK or STAT3 inhibition or degradation and a constitutively active PI3Kγ isoform cooperated with oncogenic ALK to accelerate lymphomagenesis in mice. In a three-dimensional microfluidic chip, endothelial cells that produce the CCR7 ligands CCL19/CCL21 protected ALCL cells from apoptosis induced by crizotinib. The PI3Kγ/δ inhibitor duvelisib potentiated crizotinib activity against ALCL lines and patient-derived xenografts. Furthermore, genetic deletion of CCR7 blocked the central nervous system dissemination and perivascular growth of ALCL in mice treated with crizotinib. Thus, blockade of PI3Kγ or CCR7 signaling together with ALK TKI treatment reduces primary resistance and the survival of persister lymphoma cells in ALCL.

Factor VIII as a potential player in cancer pathophysiology.

BACKGROUND: Trousseau sign was the first demonstration of a close relationship between cancer and thrombosis. Currently, venous thromboembolism (VTE) is five to six times more likely to occur in cancer patients, whereas there is a greater risk of cancer diagnoses following thromboses. In considering novel players, factor VIII (FVIII), an essential coagulation cofactor with emerging extracoagulative functions, has been identified as an independent VTE risk factor in cancer; however, the basis of this increase is unknown. OBJECTIVE: To investigate the possible direct expression and secretion of FVIII by cancer cells. METHODS: Bladder cancer, with a high VTE risk, and normal bladder tissue and epithelium, were used to investigate FVIII. Factor VIII protein and secretion were examined in bladder cancer cell lines. Expanding to other cancers, the Cancer Cell line Encyclopedia database was used to analyze FVIII, tissue factor, FV, FVII, FIX, FX, and von Willebrand factor (VWF) mRNA in 811 cell lines subdivided according to origin. Factor VIII protein synthesis, secretion, and bioactivity were investigated in a profile of cancer cell lines of differing origins. RESULTS AND CONCLUSIONS: Although expressed in the normal bladder epithelium, FVIII mRNA and protein were higher in matched bladder neoplasms, with synthesis and secretion of bioactive FVIII evident in bladder cancer cells. This can be extended to other cancer cell lines, with a pattern reflecting the tumor origin, and that is independent of VWF and other relevant players in the coagulation cascade. Here, evidence is provided of a possible independent role for FVIII in cancer-related pathophysiology.

Circulating Extracellular Vesicles Contain Liver-Derived RNA Species as Indicators of Severe Cholestasis-Induced Early Liver Fibrosis in Mice.

Aims: Biliary diseases represent around 10% of all chronic liver diseases and affect both adults and children. Currently available biochemical tests detect cholestasis but not early liver fibrosis. Circulating extracellular vesicles (EVs) provide a noninvasive, real-time molecular snapshot of the injured organ. We thus aimed at searching for a panel of EV-based biomarkers for cholestasis-induced early liver fibrosis using mouse models. Results: Progressive and detectable histological evidence of collagen deposition and liver fibrosis was observed from day 8 after bile duct ligation (BDL) in mice. Whole transcriptome and small RNA sequencing analyses of circulating EVs revealed differentially enriched RNA species after BDL versus sham controls. Unsupervised hierarchical clustering identified a signature that allowed for discrimination between BDL and controls. In particular, 151 microRNAs (miRNAs) enriched in BDL-derived EVs were identified, of which 66 were conserved in humans. The liver was an important source of circulating EVs in BDL animals as evidenced by the enrichment of several hepatic mRNAs, such as Albumin and Haptoglobin. Interestingly, among experimentally validated miRNAs, miR192-5p, miR194-5p, miR22-3p, and miR29a-3p showed similar enrichment patterns also in EVs derived from 3,5-diethoxycarboncyl-1,4-dihydrocollidine-treated (drug-induced severe cholestasis) but not in mice with mild phenotype or non-cholestatic liver fibrosis. Innovation: A panel of mRNAs and miRNAs contained in circulating EVs, when combined, indicates hepatic damage and fibrosis in mice and represents promising biomarkers for human severe cholestasis-induced liver fibrosis. Conclusion: Analysis of EV-based miRNAs, in combination with hepatic injury RNA markers, can detect early cholestatic liver injury and fibrosis in mice. Antioxid. Redox Signal. 36, 480-504.

Regulation of CD45 phosphatase by oncogenic ALK in anaplastic large cell lymphoma.

Anaplastic Large Cell Lymphoma (ALCL) is a subtype of non-Hodgkin lymphoma frequently driven by the chimeric tyrosine kinase NPM-ALK, generated by the t (2,5)(p23;q35) translocation. While ALK+ ALCL belongs to mature T cell lymphomas, loss of T cell identity is observed in the majority of ALCL secondary to a transcriptional and epigenetic repressive program induced by oncogenic NPM-ALK. While inhibiting the expression of T cell molecules, NPM-ALK activates surrogate TCR signaling by directly inducing pathways downstream the TCR. CD45 is a tyrosine phosphatase that plays a central role in T cell activation by controlling the TCR signaling and regulating the cytokine responses through the JAK/STAT pathway and exists in different isoforms depending on the stage of T-cell maturation, activation and differentiation. ALK+ ALCL cells mainly express the isoform CD45RO in keeping with their mature/memory T cell phenotype. Because of its regulatory effect on the JAK/STAT pathway that is essential for ALK+ ALCL, we investigated whether CD45 expression was affected by oncogenic ALK. We found that most ALK+ ALCL cell lines express the CD45RO isoform with modest CD45RA expression and that NPM-ALK regulated the expression of these CD45 isoforms. Regulation of CD45 expression was dependent on ALK kinase activity as CD45RO expression was increased when NPM-ALK kinase activity was inhibited by treatment with ALK tyrosine kinase inhibitors (TKIs). Silencing ALK expression through shRNA or degradation of ALK by the PROTAC TL13-112 caused upregulation of CD45RO both at mRNA and protein levels with minimal changes on CD45RA, overall indicating that oncogenic ALK downregulates the expression of CD45. CD45 repression was mediated by STAT3 as demonstrated by ChIP-seq data on ALCL cells treated with the ALK-TKI crizotinib or cells treated with a STAT3 degrader. Next, we found that knocking-out CD45 with the CRISPR/Cas9 system resulted in increased resistance to ALK TKI treatment and CD45 was down-regulated in ALCL cells that developed resistance in vitro to ALK TKIs. Overall, these data suggest that CD45 expression is regulated by ALK via STAT3 and acts as a rheostat of ALK oncogenic signaling and resistance to TKI treatment in ALCL.

Fumarase tumor suppressor gene and MET oncogene cooperate in upholding transformation and tumorigenesis.

Loss of the fumarate hydratase (FH) tumor suppressor gene results in the development of benign tumors that rarely, but regrettably, progress to very aggressive cancers. Using mouse embryo fibroblasts (MEFs) to model transformation, we found that fh knockdown results in increased expression of the met oncogene-encoded tyrosine kinase receptor through hypoxia-inducible factor (hif) stabilization. MET-increased expression was alone able to stabilize hif, thus establishing a feed forward loop that might enforce tumor progression. The fh-defective MEFs showed increased motility and protection from apoptosis. Motility, but not survival, relied on hif-1alpha and was greatly enhanced by MET ligand hepatocyte growth factor. Met cooperated with a weakly oncogenic ras in making MEFs transformed and tumorigenic, as shown by in vitro and in vivo assays. Loss of fh was not equally effective by itself but enhanced the transformed and tumorigenic phenotype induced by ras and MET. Consistently, the rescue of fumarase expression abrogated the motogenic and transformed phenotype of fh-defective MEFs. In conclusion, the data suggest that the progression of tumors where FH is lost might be boosted by activation of the MET oncogene, which is able to drive cell-autonomous tumor progression and is a strong candidate for targeted therapy.

HGF/MET Axis Induces Tumor Secretion of Tenascin-C and Promotes Stromal Rewiring in Pancreatic Cancer.

Pancreatic ductal adenocarcinoma is an aggressive tumor characterized by the presence of an abundant stromal compartment contributing significantly to the malignant phenotype. Pancreatic stellate cells are peculiar fibroblasts present in the stroma and represent the predominant source of extracellular matrix proteins, pro-inflammatory cytokines, and growth factors, including hepatocyte growth factor (HGF). Exploiting a co-culture system of human pancreatic stellate cells and cancer cells, we demonstrated that fibroblast activation was reduced upon HGF/MET axis inhibition. To unveil the signaling pathways sustaining stroma modulation orchestrated by MET activation in the tumor, we analyzed the gene expression profile in pancreatic cancer cells stimulated with HGF and treated with HGF/MET inhibitors. Transcriptome analysis showed that, among all the genes modulated by HGF, a subset of 125 genes was restored to the basal level following treatment with the inhibitors. By examining these genes via ingenuity pathway analysis, tenascin C emerged as a promising candidate linking MET signaling and tumor microenvironment. MET-dependent tenascin C modulation in pancreatic cancer cells was validated at RNA and protein levels both in vitro and in vivo. In conclusion, this work identifies tenascin C as a gene modulated by MET activation, suggesting a role in MET-mediated tumor-stroma interplay occurring during pancreatic tumor progression.

Sparsely-connected autoencoder (SCA) for single cell RNAseq data mining.

Single-cell RNA sequencing (scRNAseq) is an essential tool to investigate cellular heterogeneity. Thus, it would be of great interest being able to disclose biological information belonging to cell subpopulations, which can be defined by clustering analysis of scRNAseq data. In this manuscript, we report a tool that we developed for the functional mining of single cell clusters based on Sparsely-Connected Autoencoder (SCA). This tool allows uncovering hidden features associated with scRNAseq data. We implemented two new metrics, QCC (Quality Control of Cluster) and QCM (Quality Control of Model), which allow quantifying the ability of SCA to reconstruct valuable cell clusters and to evaluate the quality of the neural network achievements, respectively. Our data indicate that SCA encoded space, derived by different experimentally validated data (TF targets, miRNA targets, Kinase targets, and cancer-related immune signatures), can be used to grasp single cell cluster-specific functional features. In our implementation, SCA efficacy comes from its ability to reconstruct only specific clusters, thus indicating only those clusters where the SCA encoding space is a key element for cells aggregation. SCA analysis is implemented as module in rCASC framework and it is supported by a GUI to simplify it usage for biologists and medical personnel.

Evolution of HER2-positive mammary carcinoma: HER2 loss reveals claudin-low traits in cancer progression.

HER2-positive breast cancers may lose HER2 expression in recurrences and metastases. In this work, we studied cell lines derived from two transgenic mammary tumors driven by human HER2 that showed different dynamics of HER2 status. MamBo89HER2stable cell line displayed high and stable HER2 expression, which was maintained upon in vivo passages, whereas MamBo43HER2labile cell line gave rise to HER2-negative tumors from which MamBo38HER2loss cell line was derived. Both low-density seeding and in vitro trastuzumab treatment of MamBo43HER2labile cells induced the loss of HER2 expression. MamBo38HER2loss cells showed a spindle-like morphology, high stemness and acquired in vivo malignancy. A comprehensive molecular profile confirmed the loss of addiction to HER2 signaling and acquisition of an EMT signature, together with increased angiogenesis and migration ability. We identified PDGFR-B among the newly expressed determinants of MamBo38HER2loss cell tumorigenic ability. Sunitinib inhibited MamBo38HER2loss tumor growth in vivo and reduced stemness and IL6 production in vitro. In conclusion, HER2-positive mammary tumors can evolve into tumors that display distinctive traits of claudin-low tumors. Our dynamic model of HER2 status can lead to the identification of new druggable targets, such as PDGFR-B, in order to counteract the resistance to HER2-targeted therapy that is caused by HER2 loss.