Viral T-cell epitopes - Identification, characterization and clinical application.

T-cell immunity, mediated by CD4+ and CD8+ T cells, represents a cornerstone in the control of viral infections. Virus-derived T-cell epitopes are represented by human leukocyte antigen (HLA)-presented viral peptides on the surface of virus-infected cells. They are the prerequisite for the recognition of infected cells by T cells. Knowledge of viral T-cell epitopes provides on the one hand a diagnostic tool to decipher protective T-cell immune responses in the human population and on the other hand various prophylactic and therapeutic options including vaccination approaches and the transfer of virus-specific T cells. Such approaches have already been proven to be effective against various viral infections, particularly in immunocompromised patients lacking sufficient humoral, antibody-based immune response. This review provides an overview on the state of the art as well as current studies regarding the identification and characterization of viral T-cell epitopes and approaches of clinical application. In the first chapter in silico prediction tools and direct, mass spectrometry-based identification of viral T-cell epitopes is compared. The second chapter provides an overview of commonly used assays for further characterization of T-cell responses and phenotypes. The final chapter presents an overview of clinical application of viral T-cell epitopes with a focus on human immunodeficiency virus (HIV), human cytomegalovirus (HCMV) and severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2), being representatives of relevant viruses.

Spectrum of splicing variants in disease genes and the ability of RNA analysis to reduce uncertainty in clinical interpretation.

The complexities of gene expression pose challenges for the clinical interpretation of splicing variants. To better understand splicing variants and their contribution to hereditary disease, we evaluated their prevalence, clinical classifications, and associations with diseases, inheritance, and functional characteristics in a 689,321-person clinical cohort and two large public datasets. In the clinical cohort, splicing variants represented 13% of all variants classified as pathogenic (P), likely pathogenic (LP), or variants of uncertain significance (VUSs). Most splicing variants were outside essential splice sites and were classified as VUSs. Among all individuals tested, 5.4% had a splicing VUS. If RNA analysis were to contribute supporting evidence to variant interpretation, we estimated that splicing VUSs would be reclassified in 1.7% of individuals in our cohort. This would result in a clinically significant result (i.e., P/LP) in 0.1% of individuals overall because most reclassifications would change VUSs to likely benign. In ClinVar, splicing VUSs were 4.8% of reported variants and could benefit from RNA analysis. In the Genome Aggregation Database (gnomAD), splicing variants comprised 9.4% of variants in protein-coding genes; most were rare, precluding unambiguous classification as benign. Splicing variants were depleted in genes associated with dominant inheritance and haploinsufficiency, although some genes had rare variants at essential splice sites or had common splicing variants that were most likely compatible with normal gene function. Overall, we describe the contribution of splicing variants to hereditary disease, the potential utility of RNA analysis for reclassifying splicing VUSs, and how natural variation may confound clinical interpretation of splicing variants.

Clinical relevance of genetic variants in the von Willebrand factor according to in-silico methods.

Clinical interpretation of genetic variants in the context of the patient's phenotype is a time-consuming and costly process. In-silico analysis using in-silico prediction tools, and molecular modeling have been developed to predict the influence of genetic variants on the quality and/or quantity of the resulting translated protein, and in this way, to alert clinicians of disease likelihood in the absence of previous evidence. Our objectives were to evaluate the success rate of the in-silico analysis in predicting the disease-causing variants as pathogenic and the single-nucleotide variants as neutral, and to establish the reliability of in-silico analysis for determining pathogenicity or neutrality of von Willebrand factor gene-associated genetic variants. Using in-silico analysis, we studied pathogenicity in 31 disease-causing variants, and neutrality in 61 single-nucleotide variants from patients previously diagnosed as type 2 von Willebrand disease. Disease-causing variants and non-synonymous single-nucleotide variants were explored by in-silico tools that analyze the amino acidic sequence. Intronic and synonymous single-nucleotide variants were analyzed by in-silico methods that evaluate the nucleotidic sequence. We found a consistent agreement between predictions achieved by in-silico prediction tools and molecular modeling, both for defining the pathogenicity of disease-causing variants and the neutrality of single-nucleotide variants. Based on our results, the in-silico analysis would help to define the pathogenicity or neutrality in novel genetic variants observed in patients with clinical and laboratory phenotypes suggestive of von Willebrand disease.

Prediction of Human Clearance Using In Silico Models with Reduced Bias.

Predicting human clearance with high accuracy from in silico-derived parameters alone is highly desirable, as it is fast, saves in vitro resources, and is animal-sparing. We derived random forest (RF) models from 1340 compounds with human intravenous pharmacokinetic (PK) data, the largest data set publicly available today. To assess the general applicability of the RF models, we systematically removed structural-therapeutic class analogues and other compounds with structural similarity from the training sets. For a quasi-prospective test set of 343 compounds, we show that RF models devoid of structurally similar compounds in the training set predict human clearance with a geometric mean fold error (GMFE) of 3.3. While the observed GMFE illustrates how difficult it is to generate a useful model that is broadly applicable, we posit that our RF models yield a more realistic assessment of how well human clearance can be predicted prospectively. We deployed the conformal prediction formalism to assess the model applicability and to determine the prediction confidence intervals for each prediction. We observed that clearance can be predicted better for renally cleared compounds than for other clearance mechanisms. We show that applying a classification model for predicting renal clearance identifies a subset of compounds for which clearance can be predicted with higher accuracy, yielding a GMFE of 2.3. In addition, our in silico RF human clearance models compared well to models derived from scaling human hepatocytes or preclinical in vivo data.

Evaluation of machine learning models for cytochrome P450 3A4, 2D6, and 2C9 inhibition.

Cytochrome P450 (CYP) enzymes are involved in the metabolism of approximately 75% of marketed drugs. Inhibition of the major drug-metabolizing P450s could alter drug metabolism and lead to undesirable drug-drug interactions. Therefore, it is of great significance to explore the inhibition of P450s in drug discovery. Currently, machine learning including deep learning algorithms has been widely used for constructing in silico models for the prediction of P450 inhibition. These models exhibited varying predictive performance depending on the use of machine learning algorithms and molecular representations. This leads to the difficulty in the selection of appropriate models for practical use. In this study, we systematically evaluated the conventional machine learning and deep learning models for three major P450 enzymes, CYP3A4, CYP2D6, and CYP2C9 from several perspectives, such as algorithms, molecular representation, and data partitioning strategies. Our results showed that the XGBoost and CatBoost algorithms coupled with the combined fingerprint/physicochemical descriptor features exhibited the best performance with Area Under Curve (AUC)  of 0.92, while the deep learning models were generally inferior to the conventional machine learning models (average AUC reached 0.89) on the same test sets. We also found that data volume and sampling strategy had a minor effect on model performance. We anticipate that these results are helpful for the selection of molecular representations and machine learning/deep learning algorithms in the P450 model construction and the future model development of P450 inhibition.

In Silico Prediction of N-Nitrosamine Formation Pathways of Pharmaceutical Products.

The recent discovery of N-nitrosodimethylamine (NDMA), a mutagenic N-nitrosamine, in pharmaceuticals has adversely impacted the global supply of relevant pharmaceutical products. Contamination by N-nitrosamines diverts resources and time from research and development or pharmaceutical production, representing a bottleneck in drug development. Therefore, predicting the risk of N-nitrosamine contamination is an important step in preventing pharmaceutical contamination by DNA-reactive impurities for the production of high-quality pharmaceuticals. In this study, we first predicted the degradation pathways and impurities of model pharmaceuticals, namely gliclazide and indapamide, in silico using an expert-knowledge software. Second, we verified the prediction results with a demonstration test, which confirmed that N-nitrosamines formed from the degradation of gliclazide and indapamide in the presence of hydrogen peroxide, especially under alkaline conditions. Furthermore, the pathways by which degradation products formed were determined using ranitidine, a compound previously demonstrated to generate NDMA. The prediction indicated that a ranitidine-related compound served as a potential source of nitroso groups for NDMA formation. In silico software is expected to be useful for developing methods to assess the risk of N-nitrosamine formation from pharmaceuticals.

Identification, trace level quantification, and in silico assessment of potential genotoxic impurity in Famotidine.

Potential genotoxic impurities in medications are an increasing concern in the pharmaceutical industry and regulatory bodies because of the risk of human carcinogenesis. To prevent the emergence of these impurities, it is crucial to carefully examine not only the final product but also the intermediates and key starting material (KSM) used in drug synthesis. During the related substances analysis of KSM of Famotidine, an unknown impurity in the range of 0.5-1.0% was found prompting the need for isolation and characterization due to the possibility of its to infiltrate into the final product. In this study, the impurity was isolated and characterized as 5-(2-chloroethyl)-3,3-dimethyl-3,4-dihydro-2H-1,2,4,6-thiatriazine 1,1-dioxide using multiple instrumental analysis, uncovering a structural alert that raises concern. Considering the potential impact of impurity on human health, an in silico genotoxicity assessment was established using Derek and Sarah tool in accordance with ICH M7 guideline. Furthermore, molecular docking and molecular dynamics simulation were performed to evaluate the specific interaction of the impurity with DNA. The findings reveal consistent interaction of the impurity with the dG-rich region of the DNA duplex and binding at the minor groove. Both in silico prediction and molecular dynamic study confirmed the genotoxic character of the impurity. The newly discovered impurity in famotidine has not been reported previously, and there is currently no analytical method available for its identification and control. A highly sensitive HPLC-UV method was developed and validated in accordance with ICH requirements, enabling quantification of the impurity at trace level in famotidine ensuring its safe release.

Applications of artificial intelligence in clinical laboratory genomics.

The transition from analog to digital technologies in clinical laboratory genomics is ushering in an era of "big data" in ways that will exceed human capacity to rapidly and reproducibly analyze those data using conventional approaches. Accurately evaluating complex molecular data to facilitate timely diagnosis and management of genomic disorders will require supportive artificial intelligence methods. These are already being introduced into clinical laboratory genomics to identify variants in DNA sequencing data, predict the effects of DNA variants on protein structure and function to inform clinical interpretation of pathogenicity, link phenotype ontologies to genetic variants identified through exome or genome sequencing to help clinicians reach diagnostic answers faster, correlate genomic data with tumor staging and treatment approaches, utilize natural language processing to identify critical published medical literature during analysis of genomic data, and use interactive chatbots to identify individuals who qualify for genetic testing or to provide pre-test and post-test education. With careful and ethical development and validation of artificial intelligence for clinical laboratory genomics, these advances are expected to significantly enhance the abilities of geneticists to translate complex data into clearly synthesized information for clinicians to use in managing the care of their patients at scale.

Prediction of the Contribution Ratio of a Target Metabolic Enzyme to Clearance from Chemical Structure Information.

The contribution ratio of metabolic enzymes such as cytochrome P450 to in vivo clearance (fraction metabolized: fm) is a pharmacokinetic index that is particularly important for the quantitative evaluation of drug-drug interactions. Since obtaining experimental in vivo fm values is challenging, those derived from in vitro experiments have often been used alternatively. This study aimed to explore the possibility of constructing machine learning models for predicting in vivo fm using chemical structure information alone. We collected in vivo fm values and chemical structures of 319 compounds from a public database with careful manual curation and constructed predictive models using several machine learning methods. The results showed that in vivo fm values can be obtained from structural information alone with a performance comparable to that based on in vitro experimental values and that the prediction accuracy for the compounds involved in CYP induction or inhibition is significantly higher than that by using in vitro values. Our new approach to predicting in vivo fm values in the early stages of drug discovery should help improve the efficiency of the drug optimization process.

Functional attributes of bioactive peptides of bovine milk origin and application of in silico approaches for peptide prediction and functional annotations.

Bovine milk peptides are the protein fragments with diverse bioactive properties having antioxidant, anticarcinogenic, other therapeutic and nutraceutical potentials. These peptides are formed in milk by enzymatic hydrolysis, gastrointestinal digestion and fermentation processes. They have significant health impact with high potency and low toxicity making them a suitable natural alternative for preventing and managing diseases. Antibiotic resistance has increased the quest for better peptide candidates with antimicrobial effects. This article presents a comprehensive review on well documented antimicrobial, immunological, opioid, and anti-hypertensive activities of bovine milk peptides. It also covers the usage of computational biology tools and databases for prediction and analysis of the food-derived bioactive peptides. In silico analysis of amino acid sequences of Bos taurus milk proteins have been predicted to generate peptides with dipeptidyl peptidase IV inhibitory and ACE inhibitory properties, making them favorable candidates for developing blood sugar lowering drugs and anti-hypertensives. In addition to the prediction of new bioactive peptides, application of bioinformatics tools to predict novel functions of already known peptides is also discussed. Overall, this review focuses on the reported as well as predicted biologically active peptide of casein and whey proteins of bovine milk that can be utilized to develop therapeutic agents.

Utility of in silico prediction of target suppression for antibodies against soluble targets: static versus dynamic models.

PURPOSE: Antibodies that bind soluble targets such as cytokines belong to an important class of immunotherapies. Target levels can significantly accumulate after antibody administration due to formation of antibody-target complex, accompanied with suppression in free target which is often difficult to measure. Being a surrogate for pharmacodynamic activity, free target suppression is often predicted using in silico tools. The objective of this work is to illustrate the utility of modelling and to compare static versus dynamic models in the prediction of free target suppression. METHODS: Using binding principles, we have derived a static equation to predict free target suppression at steady state (FTSS). This equation operates with five input parameters and accounts for target accumulation over time. Its predictivity was compared to a dynamic model and to other existing metrics in literature via simulations and assumptions were illustrated. RESULTS: We demonstrated the utility of in silico tools in prediction of free target suppression using static and dynamic models and clarified the assumptions in key input parameters and their limitations. Predicted values using the FTSS equation correlate very well with those from the dynamic model at level > 20% target suppression, relevant for antagonistic antibodies. CONCLUSION: In silico tools are needed to predict target suppression by antibody drugs. Static or dynamic models can be used dependant on the scope, available data and undertaken assumptions. These tools can be used to guide discovery and development of antibodies and has the potential to reduce clinical failure.

Towards Alzheimer's disease-related targets: One-pot Cu(I)- mediated synthesis of new nitroindazolyltriazoles.

In this work, we have investigated the one pot strategy for the Cu(I)-mediated synthesis of new triazoles bearing nitroindazole moieties using different copper catalysts. The biological activity of newly synthesized nitroindazolyltriazoles towards Alzheimer's disease-related targets, namely cholinesterases, monoamine oxidases, and amyloid aggregation, were investigated. Predictions of target affinity, physicochemical parameters, gastrointestinal absorption and brain penetration were achieved by means of in silico tools.

Outside the limit: questioning the distance restrictions for cooperative miRNA binding sites.

Among the concepts in biology that are widely taken granted is a potentiated cooperative effect of multiple miRNAs on the same target. This strong hypothesis contrasts insufficient experimental evidence. The quantity as well as the quality of required side constraints of cooperative binding remain largely hidden. For miR-21-5p and miR-155-5p, two commonly investigated regulators across diseases, we selected 15 joint target genes. These were chosen to represent various neighboring 3'UTR binding site constellations, partially exceeding the distance rules that have been established for over a decade. We identified different cooperative scenarios with the binding of one miRNA enhancing the binding effects of the other miRNA and vice versa. Using both, reporter assays and whole proteome analyses, we observed these cooperative miRNA effects for genes that bear 3'UTR binding sites at distances greater than the previously defined limits. Astonishingly, the experiments provide even stronger evidence for cooperative miRNA effects than originally postulated. In the light of these findings the definition of targetomes specified for single miRNAs need to be refined by a concept that acknowledges the cooperative effects of miRNAs.

Computational modeling (in silico) methods combined with ecotoxicological experiments (in vivo) to predict the environmental risks of an antihistamine drug (loratadine).

This study employed computational modeling (in silico) methods, combined with ecotoxicological experiments (in vivo) to predict the persistence/biodegradability, bioaccumulation, mobility, and ecological risks of an antihistamine drug (Loratadine: LOR) in the aquatic compartment. To achieve these goals, four endpoints of the LOR were obtained from different open-source computational tools, namely: (i) "STP total removal"; (ii) Predicted ready biodegradability; (iii) Octanol-water partition coefficient (KOW); and (iv) Soil organic adsorption coefficient (KOC). Moreover, acute and chronic, ecotoxicological assays using non-target freshwater organisms of different trophic levels (namely, algae Pseudokirchneriella subcapitata; microcrustaceans Daphnia similis and Ceriodaphnia dubia; and fish Danio rerio), were used to predict the ecological risks of LOR. The main results showed that LOR: (i) is considered persistent (after a weight-of-evidence assessment) and highly resistant to biodegradation; (ii) is hydrophobic (LOG KOW = 5.20), immobile (LOG KOC = 5.63), and thus, it can potentially bioaccumulate and/or can cause numerous deleterious effects in aquatic species; and (iii) after ecotoxicological evaluation is considered "toxic" and/or "highly toxic" to the three trophic levels tested. Moreover, both the ecotoxicological assays and risk assessment (RQ), showed that LOR is more harmful for the crustaceans (RQcrustaceans = moderate to high risks) than for algae and fish. Ultimately, this study reinforces the ecological concern due to the indiscriminate disposal of this antihistamine drug in worldwide aquatic ecosystems.

In Silico Prediction of Eye Irritation Using Hansen Solubility Parameters and Predicted pKa Values.

An in silico method has been developed that permits the binary differentiation between pure liquids causing serious eye damage or eye irritation, and pure liquids with no need for such classification, according to the UN GHS system. The method is based on the finding that the Hansen Solubility Parameters (HSP) of a liquid are collectively important predictors for eye irritation. Thus, by applying a two-tier approach in which in silico-predicted pKa values (firstly) and a trained model based solely on in silico-predicted HSP data (secondly) were used, we have developed, and validated, a fully in silico approach for predicting the outcome of a Draize test (in terms of UN GHS Cat. 1/Cat. 2A/Cat. 2B or UN GHS No Cat.) with high validation set performance (sensitivity = 0.846, specificity = 0.818, balanced accuracy = 0.832) using SMILES only. The method is applicable to pure non-ionic liquids with molecular weight below 500 g/mol, fewer than six hydrogen bond donors (e.g. nitrogen-hydrogen or oxygen-hydrogen bonds) and fewer than eleven hydrogen bond acceptors (e.g. nitrogen or oxygen atoms). Due to its fully in silico characteristics, this method can be applied to pure liquids that are still at the desktop design stage and not yet in production.

CLC-Pred 2.0: A Freely Available Web Application for In Silico Prediction of Human Cell Line Cytotoxicity and Molecular Mechanisms of Action for Druglike Compounds.

In vitro cell-line cytotoxicity is widely used in the experimental studies of potential antineoplastic agents and evaluation of safety in drug discovery. In silico estimation of cytotoxicity against hundreds of tumor cell lines and dozens of normal cell lines considerably reduces the time and costs of drug development and the assessment of new pharmaceutical agent perspectives. In 2018, we developed the first freely available web application (CLC-Pred) for the qualitative prediction of cytotoxicity against 278 tumor and 27 normal cell lines based on structural formulas of 59,882 compounds. Here, we present a new version of this web application: CLC-Pred 2.0. It also employs the PASS (Prediction of Activity Spectra for Substance) approach based on substructural atom centric MNA descriptors and a Bayesian algorithm. CLC-Pred 2.0 provides three types of qualitative prediction: (1) cytotoxicity against 391 tumor and 47 normal human cell lines based on ChEMBL and PubChem data (128,545 structures) with a mean accuracy of prediction (AUC), calculated by the leave-one-out (LOO CV) and the 20-fold cross-validation (20F CV) procedures, of 0.925 and 0.923, respectively; (2) cytotoxicity against an NCI60 tumor cell-line panel based on the Developmental Therapeutics Program's NCI60 data (22,726 structures) with different thresholds of IG50 data (100, 10 and 1 nM) and a mean accuracy of prediction from 0.870 to 0.945 (LOO CV) and from 0.869 to 0.942 (20F CV), respectively; (3) 2170 molecular mechanisms of actions based on ChEMBL and PubChem data (656,011 structures) with a mean accuracy of prediction 0.979 (LOO CV) and 0.978 (20F CV). Therefore, CLC-Pred 2.0 is a significant extension of the capabilities of the initial web application.

Challenges in Brugada Syndrome Stratification: Investigating SCN5A Mutation Localization and Clinical Phenotypes.

Brugada Syndrome (BrS) is a genetic heart condition linked to sudden cardiac death. Though the SCN5A gene is primarily associated with BrS, there is a lack of comprehensive studies exploring the connection between SCN5A mutation locations and the clinical presentations of the syndrome. This study aimed to address this gap and gain further understanding of the syndrome. The investigation classified 36 high-risk BrS patients based on SCN5A mutations within the transmembrane/structured (TD) and intra-domain loops (IDLs) lacking a 3D structure. We characterized the intrinsically disordered regions (IDRs) abundant in IDLs, using bioinformatics tools to predict IDRs and post-translational modifications (PTMs) in NaV1.5. Interestingly, it was found that current predictive tools often underestimate the impacts of mutations in IDLs and disordered regions. Moreover, patients with SCN5A mutations confined to IDL regions-previously deemed 'benign'-displayed clinical symptoms similar to those carrying 'damaging' variants. Our research illuminates the difficulty in stratifying patients based on SCN5A mutation locations, emphasizing the vital role of IDLs in the NaV1.5 channel's functioning and protein interactions. We advocate for caution when using predictive tools for mutation evaluation in these regions and call for the development of improved strategies in accurately assessing BrS risk.

Integrated Approach for Synthetic Cathinone Drug Prioritization and Risk Assessment: In Silico Approach and Sub-Chronic Studies in Daphnia magna and Tetrahymena thermophila.

Synthetic cathinones (SC) are drugs of abuse that have been reported in wastewaters and rivers raising concern about potential hazards to non-target organisms. In this work, 44 SC were selected for in silico studies, and a group of five emerging SC was prioritized for further in vivo ecotoxicity studies: buphedrone (BPD), 3,4-dimethylmethcathinone (3,4-DMMC), butylone (BTL), 3-methylmethcathinone (3-MMC), and 3,4-methylenedioxypyrovalerone (MDPV). In vivo short-term exposures were performed with the protozoan Tetrahymena thermophila (28 h growth inhibition assay) and the microcrustacean Daphnia magna by checking different indicators of toxicity across life stage (8 days sublethal assay at 10.00 µg L-1). The in silico approaches predicted a higher toxic potential of MDPV and lower toxicity of BTL to the model organisms (green algae, protozoan, daphnia, and fish), regarding the selected SC for the in vivo experiments. The in vivo assays showed protozoan growth inhibition with MDPV > BPD > 3,4-DMMC, whereas no effects were observed for BTL and stimulation of growth was observed for 3-MMC. For daphnia, the responses were dependent on the substance and life stage. Briefly, all five SC interfered with the morphophysiological parameters of juveniles and/or adults. Changes in swimming behavior were observed for BPD and 3,4-DMMC, and reproductive parameters were affected by MDPV. Oxidative stress and changes in enzymatic activities were noted except for 3-MMC. Overall, the in silico data agreed with the in vivo protozoan experiments except for 3-MMC, whereas daphnia in vivo experiments showed that at sublethal concentrations, all selected SC interfered with different endpoints. This study shows the importance to assess SC ecotoxicity as it can distress aquatic species and interfere with food web ecology and ecosystem balance.

Optimization of in silico tools for predicting genetic variants: individualizing for genes with molecular sub-regional stratification.

Genes are unique in functional role and differ in their sensitivities to genetic defects, but with difficulties in pathogenicity prediction. This study attempted to improve the performance of existing in silico algorithms and find a common solution based on individualization strategy. We initiated the individualization with the epilepsy-related SCN1A variants by sub-regional stratification. SCN1A missense variants related to epilepsy were retrieved from mutation databases, and benign missense variants were collected from ExAC database. Predictions were performed by using 10 traditional tools with stepwise optimizations. Model predictive ability was evaluated using the five-fold cross-validations on variants of SCN1A, SCN2A, and KCNQ2. Additional validation was performed in SCN1A variants of damage-confirmed/familial epilepsy. The performance of commonly used predictors was less satisfactory for SCN1A with accuracy less than 80% and varied dramatically by functional domains of Nav1.1. Multistep individualized optimizations, including cutoff resetting, domain-based stratification, and combination of predicting algorithms, significantly increased predictive performance. Similar improvements were obtained for variants in SCN2A and KCNQ2. The predictive performance of the recently developed ensemble tools, such as Mendelian clinically applicable pathogenicity, combined annotation-dependent depletion and Eigen, was also improved dramatically by application of the strategy with molecular sub-regional stratification. The prediction scores of SCN1A variants showed linear correlations with the degree of functional defects and the severity of clinical phenotypes. This study highlights the need of individualized optimization with molecular sub-regional stratification for each gene in practice.

Thrombosis-related circulating miR-16-5p is associated with disease severity in patients hospitalised for COVID-19.

SARS-CoV-2 tropism for the ACE2 receptor, along with the multifaceted inflammatory reaction, is likely to drive the generalized hypercoagulable and thrombotic state seen in patients with COVID-19. Using the original bioinformatic workflow and network medicine approaches we reanalysed four coronavirus-related expression datasets and performed co-expression analysis focused on thrombosis and ACE2 related genes. We identified microRNAs (miRNAs) which play role in ACE2-related thrombosis in coronavirus infection and further, we validated the expressions of precisely selected miRNAs-related to thrombosis (miR-16-5p, miR-27a-3p, let-7b-5p and miR-155-5p) in 79 hospitalized COVID-19 patients and 32 healthy volunteers by qRT-PCR. Consequently, we aimed to unravel whether bioinformatic prioritization could guide selection of miRNAs with a potential of diagnostic and prognostic biomarkers associated with disease severity in patients hospitalized for COVID-19. In bioinformatic analysis, we identified EGFR, HSP90AA1, APP, TP53, PTEN, UBC, FN1, ELAVL1 and CALM1 as regulatory genes which could play a pivotal role in COVID-19 related thrombosis. We also found miR-16-5p, miR-27a-3p, let-7b-5p and miR-155-5p as regulators in the coagulation and thrombosis process. In silico predictions were further confirmed in patients hospitalized for COVID-19. The expression levels of miR-16-5p and let-7b in COVID-19 patients were lower at baseline, 7-days and 21-day after admission compared to the healthy controls (p < 0.0001 for all time points for both miRNAs). The expression levels of miR-27a-3p and miR-155-5p in COVID-19 patients were higher at day 21 compared to the healthy controls (p = 0.007 and p < 0.001, respectively). A low baseline miR-16-5p expression presents predictive utility in assessment of the hospital length of stay or death in follow-up as a composite endpoint (AUC:0.810, 95% CI, 0.71-0.91, p < 0.0001) and low baseline expression of miR-16-5p and diabetes mellitus are independent predictors of increased length of stay or death according to a multivariate analysis (OR: 9.417; 95% CI, 2.647-33.506; p = 0.0005 and OR: 6.257; 95% CI, 1.049-37.316; p = 0.044, respectively). This study enabled us to better characterize changes in gene expression and signalling pathways related to hypercoagulable and thrombotic conditions in COVID-19. In this study we identified and validated miRNAs which could serve as novel, thrombosis-related predictive biomarkers of the COVID-19 complications, and can be used for early stratification of patients and prediction of severity of infection development in an individual.Abbreviations: ACE2, angiotensin-converting enzyme 2AF, atrial fibrillationAPP, Amyloid Beta Precursor ProteinaPTT, activated partial thromboplastin timeAUC, Area under the curveAß, amyloid betaBMI, body mass indexCAD, coronary artery diseaseCALM1, Calmodulin 1 geneCaM, calmodulinCCND1, Cyclin D1CI, confidence intervalCOPD, chronic obstructive pulmonary diseaseCOVID-19, Coronavirus disease 2019CRP, C-reactive proteinCV, CardiovascularCVDs, cardiovascular diseasesDE, differentially expressedDM, diabetes mellitusEGFR, Epithelial growth factor receptorELAVL1, ELAV Like RNA Binding Protein 1FLNA, Filamin AFN1, Fibronectin 1GEO, Gene Expression OmnibushiPSC-CMs, Human induced pluripotent stem cell-derived cardiomyocytesHSP90AA1, Heat Shock Protein 90 Alpha Family Class A Member 1Hsp90α, heat shock protein 90αICU, intensive care unitIL, interleukinIQR, interquartile rangelncRNAs, long non-coding RNAsMI, myocardial infarctionMiRNA, MiR, microRNAmRNA, messenger RNAncRNA, non-coding RNANERI, network-medicine based integrative approachNF-kB, nuclear factor kappa-light-chain-enhancer of activated B cellsNPV, negative predictive valueNXF, nuclear export factorPBMCs, Peripheral blood mononuclear cellsPCT, procalcitoninPPI, Protein-protein interactionsPPV, positive predictive valuePTEN, phosphatase and tensin homologqPCR, quantitative polymerase chain reactionROC, receiver operating characteristicSARS-CoV-2, severe acute respiratory syndrome coronavirus 2SD, standard deviationTLR4, Toll-like receptor 4TM, thrombomodulinTP53, Tumour protein P53UBC, Ubiquitin CWBC, white blood cells.