StellarPath: Hierarchical-vertical multi-omics classifier synergizes stable markers and interpretable similarity networks for patient profiling.

The Patient Similarity Network paradigm implies modeling the similarity between patients based on specific data. The similarity can summarize patients' relationships from high-dimensional data, such as biological omics. The end PSN can undergo un/supervised learning tasks while being strongly interpretable, tailored for precision medicine, and ready to be analyzed with graph-theory methods. However, these benefits are not guaranteed and depend on the granularity of the summarized data, the clarity of the similarity measure, the complexity of the network's topology, and the implemented methods for analysis. To date, no patient classifier fully leverages the paradigm's inherent benefits. PSNs remain complex, unexploited, and meaningless. We present StellarPath, a hierarchical-vertical patient classifier that leverages pathway analysis and patient similarity concepts to find meaningful features for both classes and individuals. StellarPath processes omics data, hierarchically integrates them into pathways, and uses a novel similarity to measure how patients' pathway activity is alike. It selects biologically relevant molecules, pathways, and networks, considering molecule stability and topology. A graph convolutional neural network then predicts unknown patients based on known cases. StellarPath excels in classification performances and computational resources across sixteen datasets. It demonstrates proficiency in inferring the class of new patients described in external independent studies, following its initial training and testing phases on a local dataset. It advances the PSN paradigm and provides new markers, insights, and tools for in-depth patient profiling.

ciRS-7 and miR-7 regulate ischemia-induced neuronal death via glutamatergic signaling.

Brain functionality relies on finely tuned regulation of gene expression by networks of non-coding RNAs (ncRNAs) such as the one composed by the circular RNA ciRS-7 (also known as CDR1as), the microRNA miR-7, and the long ncRNA Cyrano. We describe ischemia-induced alterations in the ncRNA network both in vitro and in vivo and in transgenic mice lacking ciRS-7 or miR-7. Our data show that cortical neurons downregulate ciRS-7 and Cyrano and upregulate miR-7 expression during ischemia. Mice lacking ciRS-7 exhibit reduced lesion size and motor impairment, while the absence of miR-7 alone results in increased ischemia-induced neuronal death. Moreover, miR-7 levels in pyramidal excitatory neurons regulate neurite morphology and glutamatergic signaling, suggesting a potential molecular link to the in vivo phenotype. Our data reveal the role of ciRS-7 and miR-7 in modulating ischemic stroke outcome, shedding light on the pathophysiological function of intracellular ncRNA networks in the brain.

Alzheimer's disease-induced phagocytic microglia express a specific profile of coding and non-coding RNAs.

INTRODUCTION: Alzheimer's disease (AD) is a neurodegenerative disease and the main cause of dementia in the elderly. AD pathology is characterized by accumulation of microglia around the beta-amyloid (Aß) plaques which assumes disease-specific transcriptional signatures, as for the disease-associated microglia (DAM). However, the regulators of microglial phagocytosis are still unknown. METHODS: We isolated Aß-laden microglia from the brain of 5xFAD mice for RNA sequencing to characterize the transcriptional signature in phagocytic microglia and to identify the key non-coding RNAs capable of regulating microglial phagocytosis. Through spatial sequencing, we show the transcriptional changes of microglia in the AD mouse brain in relation to Aß proximity. RESULTS: Finally, we show that phagocytic messenger RNAs are regulated by miR-7a-5p, miR-29a-3p and miR-146a-5p microRNAs and segregate the DAM population into phagocytic and non-phagocytic states. DISCUSSION: Our study pinpoints key regulators of microglial Aß clearing capacity suggesting new targets for future therapeutic approaches.

Human iPSC-derived microglia carrying the LRRK2-G2019S mutation show a Parkinson's disease related transcriptional profile and function.

LRRK2-G2019S is one of the most common Parkinson's disease (PD)-associated mutations and has been shown to alter microglial functionality. However, the impact of LRRK2-G2019S on transcriptional profile of human induced pluripotent stem cell-derived microglia-like cells (iMGLs) and how it corresponds to microglia in idiopathic PD brain is not known. Here we demonstrate that LRRK2-G2019S carrying iMGL recapitulate aspects of the transcriptional signature of human idiopathic PD midbrain microglia. LRRK2-G2019S induced subtle and donor-dependent alterations in iMGL mitochondrial respiration, phagocytosis and cytokine secretion. Investigation of microglial transcriptional state in the midbrains of PD patients revealed a subset of microglia with a transcriptional overlap between the in vitro PD-iMGL and human midbrain PD microglia. We conclude that LRRK2-G2019S iMGL serve as a model to study PD-related effects in human microglia.

Esearch3D: propagating gene expression in chromatin networks to illuminate active enhancers.

Most cell type-specific genes are regulated by the interaction of enhancers with their promoters. The identification of enhancers is not trivial as enhancers are diverse in their characteristics and dynamic in their interaction partners. We present Esearch3D, a new method that exploits network theory approaches to identify active enhancers. Our work is based on the fact that enhancers act as a source of regulatory information to increase the rate of transcription of their target genes and that the flow of this information is mediated by the folding of chromatin in the three-dimensional (3D) nuclear space between the enhancer and the target gene promoter. Esearch3D reverse engineers this flow of information to calculate the likelihood of enhancer activity in intergenic regions by propagating the transcription levels of genes across 3D genome networks. Regions predicted to have high enhancer activity are shown to be enriched in annotations indicative of enhancer activity. These include: enhancer-associated histone marks, bidirectional CAGE-seq, STARR-seq, P300, RNA polymerase II and expression quantitative trait loci (eQTLs). Esearch3D leverages the relationship between chromatin architecture and transcription, allowing the prediction of active enhancers and an understanding of the complex underpinnings of regulatory networks. The method is available at: https://github.com/InfOmics/Esearch3D and https://doi.org/10.5281/zenodo.7737123.

Dynamic release of neuronal extracellular vesicles containing miR-21a-5p is induced by hypoxia.

Hypoxia induces changes in the secretion of extracellular vesicles (EVs) in several non-neuronal cells and pathological conditions. EVs are packed with biomolecules, such as microRNA(miR)-21-5p, which respond to hypoxia. However, the true EV association of miR-21-5p, and its functional or biomarker relevance, are inadequately characterised. Neurons are extremely sensitive cells, and it is not known whether the secretion of neuronal EVs and miR-21-5p are altered upon hypoxia. Here, we characterised the temporal EV secretion profile and cell viability of neurons under hypoxia. Hypoxia induced a rapid increase of miR-21a-5p secretion in the EVs, which preceded the elevation of hypoxia-induced tissue or cellular miR-21a-5p. Prolonged hypoxia induced cell death and the release of morphologically distinct EVs. The EVs protected miR-21a-5p from enzymatic degradation but a remarkable fraction of miR-21a-5p remained fragile and non-EV associated. The increase in miR-21a-5p secretion may have biomarker potential, as high blood levels of miR-21-5p in stroke patients were associated with significant disability at hospital discharge. Our data provides an understanding of the dynamic regulation of EV secretion from neurons under hypoxia and provides a candidate for the prediction of recovery from ischemic stroke.

Hypocalcemia is associated with adverse outcomes in patients hospitalized with COVID-19.

PURPOSE: Calcium ions are involved in the regulation of several cellular processes and may also influence viral replication. Hypocalcemia has been frequently reported during infectious diseases and in critically ill patients, including also COVID-19 patients, significantly related with the pro-inflammatory state and mortality. The aim of this study is to investigate the prevalence of hypocalcemia at admission in patients hospitalized for COVID-19 (Coronavirus disease 2019) and to evaluate association of hypocalcemia with in-hospital COVID-19 outcomes. METHODS: Retrospective analysis on 118 consecutive patients, hospitalized for COVID-19 between March and May 2020. Clinical characteristics, inflammation markers, biochemical routine and mineral metabolism parameters at admission were collected. Hypocalcemia was defined as total serum calcium <2.2 mmol/L. Population was stratified by tertiles of total serum calcium. Primary outcome was the composite of in-hospital death or admission to intensive care unit (ICU). Secondary outcomes included in-hospital death, admission to ICU and need for non-invasive ventilation as separate events. Associations were tested by logistic regression and Cox-regression analysis with survival curves. RESULTS: Overall prevalence of hypocalcemia was 76.6%, with just 6.7% of patients reporting levels of 25-(OH)-vitamin D > 30 ng/ml. Total serum calcium was inversely related with selected inflammatory biomarkers (p < 0.05) and poorer outcome of COVID-19 during hospitalization. Lower tertile of total calcium (≤2.02 mmol/L) had increased risk of in-hospital mortality (HR 2.77; 1.28-6.03, p = 0.01) compared with other groups. CONCLUSION: Total serum calcium detected on admission is inversely related with proinflammatory biomarkers of severe COVID-19 and is useful to better define risk stratification for adverse in-hospital outcome.

Combined Large Cell Neuroendocrine Carcinomas of the Lung: Integrative Molecular Analysis Identifies Subtypes with Potential Therapeutic Implications.

BACKGROUND: Combined large cell neuroendocrine carcinoma (CoLCNEC) is given by the association of LCNEC with adeno or squamous or any non-neuroendocrine carcinoma. Molecular bases of CoLCNEC pathogenesis are scant and no standardized therapies are defined. METHODS: 44 CoLCNECs: 26 with adenocarcinoma (CoADC), 7 with squamous cell carcinoma (CoSQC), 3 with small cell carcinoma (CoSCLC), 4 with atypical carcinoid (CoAC) and 4 napsin-A positive LCNEC (NapA+), were assessed for alterations in 409 genes and transcriptomic profiling of 20,815 genes. RESULTS: Genes altered included TP53 (n = 30), RB1 (n = 14) and KRAS (n = 13). Targetable alterations included six KRAS G12C mutations and ALK-EML4 fusion gene. Comparison of CoLCNEC transcriptomes with 86 lung cancers of pure histology (8 AC, 19 ADC, 19 LCNEC, 11 SCLC and 29 SQC) identified CoLCNEC as a separate entity of neuroendocrine tumours with three different molecular profiles, two of which showed a non-neuroendocrine lineage. Hypomethylation, activation of MAPK signalling and association to immunotherapy signature specifically characterized each of three CoLCNEC molecular clusters. Prognostic stratification was also provided. CONCLUSIONS: CoLCNECs are an independent histologic category. Our findings support the extension of routine evaluation of KRAS mutations, fusion genes and immune-related markers to offer new perspectives in the therapeutic management of CoLCNEC.

Microglial amyloid beta clearance is driven by PIEZO1 channels.

BACKGROUND: Microglia are the endogenous immune cells of the brain and act as sensors of pathology to maintain brain homeostasis and eliminate potential threats. In Alzheimer's disease (AD), toxic amyloid beta (Aß) accumulates in the brain and forms stiff plaques. In late-onset AD accounting for 95% of all cases, this is thought to be due to reduced clearance of Aß. Human genome-wide association studies and animal models suggest that reduced clearance results from aberrant function of microglia. While the impact of neurochemical pathways on microglia had been broadly studied, mechanical receptors regulating microglial functions remain largely unexplored. METHODS: Here we showed that a mechanotransduction ion channel, PIEZO1, is expressed and functional in human and mouse microglia. We used a small molecule agonist, Yoda1, to study how activation of PIEZO1 affects AD-related functions in human induced pluripotent stem cell (iPSC)-derived microglia-like cells (iMGL) under controlled laboratory experiments. Cell survival, metabolism, phagocytosis and lysosomal activity were assessed using real-time functional assays. To evaluate the effect of activation of PIEZO1 in vivo, 5-month-old 5xFAD male mice were infused daily with Yoda1 for two weeks through intracranial cannulas. Microglial Iba1 expression and Aß pathology were quantified with immunohistochemistry and confocal microscopy. Published human and mouse AD datasets were used for in-depth analysis of PIEZO1 gene expression and related pathways in microglial subpopulations. RESULTS: We show that PIEZO1 orchestrates Aß clearance by enhancing microglial survival, phagocytosis, and lysosomal activity. Aß inhibited PIEZO1-mediated calcium transients, whereas activation of PIEZO1 with a selective agonist, Yoda1, improved microglial phagocytosis resulting in Aß clearance both in human and mouse models of AD. Moreover, PIEZO1 expression was associated with a unique microglial transcriptional phenotype in AD as indicated by assessment of cellular metabolism, and human and mouse single-cell datasets. CONCLUSION: These results indicate that the compromised function of microglia in AD could be improved by controlled activation of PIEZO1 channels resulting in alleviated Aß burden. Pharmacological regulation of these mechanoreceptors in microglia could represent a novel therapeutic paradigm for AD.

Atherogenic Dyslipidemia on Admission Is Associated With Poorer Outcome in People With and Without Diabetes Hospitalized for COVID-19.

OBJECTIVE: Identifying metabolic factors associated with critical disease can help to improve management of patients hospitalized for coronavirus disease 2019 (COVID-19). High triglycerides and low HDL levels characterize the atherogenic dyslipidemia closely related to insulin resistance and diabetes. We examined associations of atherogenic dyslipidemia detected on admission with outcome of COVID-19 during hospitalization. RESEARCH DESIGN AND METHODS: We retrospectively analyzed clinical reports of 118 consecutive patients hospitalized for COVID-19 in Rome, Italy, between March and May 2020. Clinical characteristics, inflammation markers, and glucose and lipid metabolism parameters at admission were collected. Critical disease was defined as in-hospital death or need for endotracheal intubation. Associations were tested using logistic regression analysis. RESULTS: Patients with critical COVID-19 (n = 43) were significantly older than those with noncritical disease (n = 75) and presented higher levels of fasting glucose, triglycerides, C-reactive protein, interleukin-6, procalcitonin, and d-dimer (P < 0.01 for all), whereas HDL levels were lower (P = 0.003). Atherogenic dyslipidemia was more frequent in patients with critical COVID-19 (46 vs. 24%, P = 0.011), as well as diabetes (37 vs. 19%, P = 0.026), and significantly associated with death or intubation (odds ratio 2.53 [95% CI 1.16-6.32], P = 0.018). Triglycerides were significantly associated with selected inflammatory biomarkers (P < 0.05 for all) and poorer outcome of COVID-19 during hospitalization in both the overall population and the subgroup with atherogenic dyslipidemia. CONCLUSIONS: Atherogenic dyslipidemia detected on admission can be associated with critical in-hospital course of COVID-19. Further investigations are needed to elucidate the hypothetical role of insulin resistance and related lipid abnormalities in severe acute respiratory syndrome coronavirus 2 pathogenesis. Assessment of lipid profile should be encouraged in patients hospitalized for COVID-19.

"Time window" effect of Yoda1-evoked Piezo1 channel activity during mouse skeletal muscle differentiation.

AIM: Mechanosensitive Piezo1 ion channels emerged recently as important contributors to various vital functions including modulation of the blood supply to skeletal muscles. The specific Piezo1 channel agonist Yoda1 was shown to regulate the tone of blood vessels similarly to physical exercise. However, the direct role of Piezo1 channels in muscle function has been little studied so far. We therefore investigated the action of Yoda1 on the functional state of skeletal muscle precursors (satellite cells and myotubes) and on adult muscle fibres. METHODS: Immunostaining, electrophysiological intracellular recordings and Ca2+ imaging experiments were performed to localize and assess the effect of the chemical activation of Piezo1 channels with Yoda1, on myogenic precursors, adult myofibres and at the adult neuromuscular junction. RESULTS: Piezo1 channels were detected by immunostaining in satellite cells (SCs) and myotubes as well as in adult myofibres. In the skeletal muscle precursors, Yoda1 treatment stimulated the differentiation and cell fusion rather than the proliferation of SCs. Moreover, in myotubes, Yoda1 induced significant [Ca2+ ]i transients, without detectable [Ca2+ ]i response in adult myofibres. Furthermore, although expression of Piezo1 channels was detected around the muscle endplate region, Yoda1 application did not alter either the nerve-evoked or spontaneous synaptic activity or muscle contractions in adult myofibres. CONCLUSION: Our data indicate that the chemical activation of Piezo1 channels specifically enhances the differentiation of skeletal muscle precursors, suggesting a possible new strategy to promote muscle regeneration.

Intracerebral overexpression of miR-669c is protective in mouse ischemic stroke model by targeting MyD88 and inducing alternative microglial/macrophage activation.

BACKGROUND: Ischemic stroke is a devastating disease without a cure. The available treatments for ischemic stroke, thrombolysis by tissue plasminogen activator, and thrombectomy are suitable only to a fraction of patients and thus novel therapeutic approaches are urgently needed. The neuroinflammatory responses elicited secondary to the ischemic attack further aggravate the stroke-induced neuronal damage. It has been demonstrated that these responses are regulated at the level of non-coding RNAs, especially miRNAs. METHODS: We utilized lentiviral vectors to overexpress miR-669c in BV2 microglial cells in order to modulate their polarization. To detect whether the modulation of microglial activation by miR-669c provides protection in a mouse model of transient focal ischemic stroke, miR-669c overexpression was driven by a lentiviral vector injected into the striatum prior to induction of ischemic stroke. RESULTS: Here, we demonstrate that miR-669c-3p, a member of chromosome 2 miRNA cluster (C2MC), is induced upon hypoxic and excitotoxic conditions in vitro and in two different in vivo models of stroke. Rather than directly regulating the neuronal survival in vitro, miR-669c is capable of attenuating the microglial proinflammatory activation in vitro and inducing the expression of microglial alternative activation markers arginase 1 (Arg1), chitinase-like 3 (Ym1), and peroxisome proliferator-activated receptor gamma (PPAR-γ). Intracerebral overexpression of miR-669c significantly decreased the ischemia-induced cell death and ameliorated the stroke-induced neurological deficits both at 1 and 3 days post injury (dpi). Albeit miR-669c overexpression failed to alter the overall Iba1 protein immunoreactivity, it significantly elevated Arg1 levels in the ischemic brain and increased colocalization of Arg1 and Iba1. Moreover, miR-669c overexpression under cerebral ischemia influenced several morphological characteristics of Iba1 positive cells. We further demonstrate the myeloid differentiation primary response gene 88 (MyD88) transcript as a direct target for miR-669c-3p in vitro and show reduced levels of MyD88 in miR-669c overexpressing ischemic brains in vivo. CONCLUSIONS: Collectively, our data provide the evidence that miR-669c-3p is protective in a mouse model of ischemic stroke through enhancement of the alternative microglial/macrophage activation and inhibition of MyD88 signaling. Our results accentuate the importance of controlling miRNA-regulated responses for the therapeutic benefit in conditions of stroke and neuroinflammation.

Integrated analysis of transcriptome, methylome and copy number aberrations data of marginal zone lymphoma and follicular lymphoma in dog.

Marginal zone lymphoma (MZL) and follicular lymphoma (FL) are classified as indolent B-cell lymphomas in dogs. Aside from the clinical and histopathological similarities with the human counterpart, the molecular pathogenesis remains unclear. We integrated transcriptome, genome-wide DNA methylation and copy number aberration analysis to provide insights on the pathogenesis of canine MZL (n = 5) and FL (n = 7), also comparing them with diffuse large B-cell lymphoma (DLBCL). Transcriptome profiling highlighted the presence of similar biological processes affecting both histotypes, including BCR and TLR signalling pathways. However, FLs showed an enrichment of E2F targets, whereas MZLs were characterized by MYC-driven transcriptional activation signatures. FLs showed a distinctive loss on chr1 containing CEACAM23 and 24, conversely MZLs presented multiple recurrent gains on chr13, where MYC is located. The distribution of methylation peaks was similar between the two histotypes. Integrating data from the three omics, FLs resulted clearly separated from MZLs and DLBCL dataset. MZLs showed the enrichment of FoxM1 network and TLR associated TICAM1-dependent IRFs activation pathway. However, no specific signatures differentiated MZLs from DLBCLs. In conclusion, our study presents the first comprehensive analysis of molecular and epigenetic pathogenesis of canine FL and MZL.

netDx: Software for building interpretable patient classifiers by multi-'omic data integration using patient similarity networks.

Patient classification based on clinical and genomic data will further the goal of precision medicine. Interpretability is of particular relevance for models based on genomic data, where sample sizes are relatively small (in the hundreds), increasing overfitting risk netDx is a machine learning method to integrate multi-modal patient data and build a patient classifier. Patient data are converted into networks of patient similarity, which is intuitive to clinicians who also use patient similarity for medical diagnosis. Features passing selection are integrated, and new patients are assigned to the class with the greatest profile similarity. netDx has excellent performance, outperforming most machine-learning methods in binary cancer survival prediction. It handles missing data - a common problem in real-world data - without requiring imputation. netDx also has excellent interpretability, with native support to group genes into pathways for mechanistic insight into predictive features. The netDx Bioconductor package provides multiple workflows for users to build custom patient classifiers. It provides turnkey functions for one-step predictor generation from multi-modal data, including feature selection over multiple train/test data splits. Workflows offer versatility with custom feature design, choice of similarity metric; speed is improved by parallel execution. Built-in functions and examples allow users to compute model performance metrics such as AUROC, AUPR, and accuracy. netDx uses RCy3 to visualize top-scoring pathways and the final integrated patient network in Cytoscape. Advanced users can build more complex predictor designs with functional building blocks used in the default design. Finally, the netDx Bioconductor package provides a novel workflow for pathway-based patient classification from sparse genetic data.

Long Non-Coding RNAs as Molecular Signatures for Canine B-Cell Lymphoma Characterization.

BACKGROUND: Diffuse large B-cell lymphoma (DLBCL), marginal zone lymphoma (MZL) and follicular lymphoma (FL) are the most common B-cell lymphomas (BCL) in dogs. Recent investigations have demonstrated overlaps of these histotypes with the human counterparts, including clinical presentation, biologic behavior, tumor genetics, and treatment response. The molecular mechanisms that underlie canine BCL are still unknown and new studies to improve diagnosis, therapy, and the utilization of canine species as spontaneous animal tumor models are undeniably needed. Recent work using human DLBCL transcriptomes has suggested that long non-coding RNAs (lncRNAs) play a key role in lymphoma pathogenesis and pinpointed a restricted number of lncRNAs as potential targets for further studies. RESULTS: To expand the knowledge of non-coding molecules involved in canine BCL, we used transcriptomes obtained from a cohort of 62 dogs with newly-diagnosed multicentric DLBCL, MZL and FL that had undergone complete staging work-up and were treated with chemotherapy or chemo-immunotherapy. We developed a customized R pipeline performing a transcriptome assembly by multiple algorithms to uncover novel lncRNAs, and delineate genome-wide expression of unannotated and annotated lncRNAs. Our pipeline also included a new package for high performance system biology analysis, which detects high-scoring network biological neighborhoods to identify functional modules. Moreover, our customized pipeline quantified the expression of novel and annotated lncRNAs, allowing us to subtype DLBCLs into two main groups. The DLBCL subtypes showed statistically different survivals, indicating the potential use of lncRNAs as prognostic biomarkers in future studies. CONCLUSIONS: In this manuscript, we describe the methodology used to identify lncRNAs that differentiate B-cell lymphoma subtypes and we interpreted the biological and clinical values of the results. We inferred the potential functions of lncRNAs to obtain a comprehensive and integrative insight that highlights their impact in this neoplasm.