Magdalena Krochmal

The application of multi-omics and systems biology to identify therapeutic targets in chronic kidney disease

The quest for the ideal therapeutic target in chronic kidney disease (CKD) has been riddled with many obstacles stemming from the molecular complexity of the disease and its co-morbidities. Recent advances in omics technologies and the resulting amount of available data encompassing genomics, proteomics, peptidomics, transcriptomics and metabolomics has created an opportunity for integrating omics datasets to build a comprehensive and dynamic model of the molecular changes in CKD for the purpose of biomarker and drug discovery. This article reviews relevant concepts in omics data integration using systems biology, a mathematical modelling method that globally describes a biological system on the basis of its modules and the functional connections that govern their behaviour. The review describes key databases and bioinformatics tools, as well as the challenges and limitations of the current state of the art, along with practical application to CKD therapeutic target discovery. Moreover, it describes how systems biology and visualization tools can be used to generate clinically relevant molecular models with the capability to identify specific disease pathways, recognize key events in disease development and track disease progression.

Omics databases on kidney disease: where they can be found and how to benefit from them

In the recent decades, the evolution of omics technologies has led to advances in all biological fields, creating a demand for effective storage, management and exchange of rapidly generated data and research discoveries. To address this need, the development of databases of experimental outputs has become a common part of scientific practice in order to serve as knowledge sources and data-sharing platforms, providing information about genes, transcripts, proteins or metabolites. In this review, we present omics databases available currently, with a special focus on their application in kidney research and possibly in clinical practice. Databases are divided into two categories: general databases with a broad information scope and kidney-specific databases distinctively concentrated on kidney pathologies. In research, databases can be used as a rich source of information about pathophysiological mechanisms and molecular targets. In the future, databases will support clinicians with their decisions, providing better and faster diagnoses and setting the direction towards more preventive, personalized medicine. We also provide a test case demonstrating the potential of biological databases in comparing multi-omics datasets and generating new hypotheses to answer a critical and common diagnostic problem in nephrology practice. In the future, employment of databases combined with data integration and data mining should provide powerful insights into unlocking the mysteries of kidney disease, leading to a potential impact on pharmacological intervention and therapeutic disease management.

PeptiCKDdb—peptide- and protein-centric database for the investigation of genesis and progression of chronic kidney disease

The peptiCKDdb is a publicly available database platform dedicated to support research in the field of chronic kidney disease (CKD) through identification of novel biomarkers and molecular features of this complex pathology. PeptiCKDdb collects peptidomics and proteomics datasets manually extracted from published studies related to CKD. Datasets from peptidomics or proteomics, human case/control studies on CKD and kidney or urine profiling were included. Data from 114 publications (studies of body fluids and kidney tissue: 26 peptidomics and 76 proteomics manuscripts on human CKD, and 12 focusing on healthy proteome profiling) are currently deposited and the content is quarterly updated. Extracted datasets include information about the experimental setup, clinical study design, discovery-validation sample sizes and list of differentially expressed proteins (P-value < 0.05). A dedicated interactive web interface, equipped with multiparametric search engine, data export and visualization tools, enables easy browsing of the data and comprehensive analysis. In conclusion, this repository might serve as a source of data for integrative analysis or a knowledgebase for scientists seeking confirmation of their findings and as such, is expected to facilitate the modeling of molecular mechanisms underlying CKD and identification of biologically relevant biomarkers. Database URL: www.peptickddb.com

Urinary peptidomics in kidney disease and drug research

ABSTRACT Introduction: Due to its close connection with the renal system, urine is considered a valuable source of information in kidney disease research. Peptidomics methods focus on the discovery of endogenous peptides, given their wide range of biological functions and diagnostic and therapeutic potential. Representing a non-invasive and sensitive method, technological prospects of urinary peptidomics should be evaluated in the context of drug discovery and research. Areas covered: This review describes urinary peptidomics with focus on its application in drug research in the field of kidney diseases. The authors provide an overview of current achievements and potential future applications. Expert opinion: The urinary peptidome is a dynamically changing source of information, able to reflect sudden and long-term changes affecting the renal system. Studies utilizing urinary peptidomics techniques have demonstrated their value in patient stratification and detection of early pathological changes in kidney disease. Serving as a liquid biopsy, urinary peptides are an invaluable tool for drug response monitoring. Nevertheless, peptidomics is largely underexplored in drug research in general, as evidenced by the scarce number of scientific publications on this topic. Further progress will be driven by the successful validation of current discoveries and continued efforts to improve the translation of results into therapeutic applications.

Urinary peptidomics analysis reveals proteases involved in diabetic nephropathy

Mechanisms underlying the onset and progression of nephropathy in diabetic patients are not fully elucidated. Deregulation of proteolytic systems is a known path leading to disease manifestation, therefore we hypothesized that proteases aberrantly expressed in diabetic nephropathy (DN) may be involved in the generation of DN-associated peptides in urine. We compared urinary peptide profiles of DN patients (macroalbuminuric, n = 121) to diabetic patients with no evidence of DN (normoalbuminuric, n = 118). 302 sequenced, differentially expressed peptides (adjusted p-value < 0.05) were analysed with the Proteasix tool predicting proteases potentially involved in their generation. Activity change was estimated based on the change in abundance of the investigated peptides. Predictions were correlated with transcriptomics (Nephroseq) and relevant protein expression data from the literature. This analysis yielded seventeen proteases, including multiple forms of MMPs, cathepsin D and K, kallikrein 4 and proprotein convertases. The activity of MMP-2 and MMP-9, predicted to be decreased in DN, was investigated using zymography in a DN mouse model confirming the predictions. Collectively, this proof-of-concept study links urine peptidomics to molecular changes at the tissue level, building hypotheses for further investigation in DN and providing a workflow with potential applications to other diseases.

Identification of novel molecular signatures of IgA nephropathy through an integrative -omics analysis

IgA nephropathy (IgAN) is the most prevalent among primary glomerular diseases worldwide. Although our understanding of IgAN has advanced significantly, its underlying biology and potential drug targets are still unexplored. We investigated a combinatorial approach for the analysis of IgAN-relevant -omics data, aiming at identification of novel molecular signatures of the disease. Nine published urinary proteomics datasets were collected and the reported differentially expressed proteins in IgAN vs. healthy controls were integrated into known biological pathways. Proteins participating in these pathways were subjected to multi-step assessment, including investigation of IgAN transcriptomics datasets (Nephroseq database), their reported protein-protein interactions (STRING database), kidney tissue expression (Human Protein Atlas) and literature mining. Through this process, from an initial dataset of 232 proteins significantly associated with IgAN, 20 pathways were predicted, yielding 657 proteins for further analysis. Step-wise evaluation highlighted 20 proteins of possibly high relevance to IgAN and/or kidney disease. Experimental validation of 3 predicted relevant proteins, adenylyl cyclase-associated protein 1 (CAP1), SHC-transforming protein 1 (SHC1) and prolylcarboxypeptidase (PRCP) was performed by immunostaining of human kidney sections. Collectively, this study presents an integrative procedure for -omics data exploitation, giving rise to biologically relevant results.

Comparison of Urine and Plasma Peptidome Indicates Selectivity in Renal Peptide Handling

Urine is considered to be produced predominantly as a result of plasma filtration in the kidney. However, the origin of the native peptides present in urine has never been investigated in detail. Therefore, the authors aimed to obtain a first insight into the origin of urinary peptides based on a side‐by‐side comprehensive analysis of the plasma and urine peptidome.

Urinary Glycopeptide Analysis for the Investigation of Novel Biomarkers

Urine is a rich source of potential biomarkers, including glycoproteins. Glycoproteomic analysis remains difficult due to the high heterogeneity of glycans. Nevertheless, recent advances in glycoproteomics software solutions facilitate glycopeptide identification and characterization. The aim is to investigate intact glycopeptides in the urinary peptide profiles of normal subjects using a novel PTM‐centric software—Byonic.

Deep Learning and Association Rule Mining for Predicting Drug Response in Cancer. A Personalised Medicine Approach.

A major challenge in cancer treatment is predicting the clinical response to anticancer drugs for each individual patient. For complex diseases, such as cancer, characterized by high inter-patient variance, the implementation of precision medicine approaches is dependent upon understanding the disease process at the molecular level. While the ‘omics’ era provides unique opportunities to dissect the molecular features of diseases, the ability to apply it to targeted therapeutic efforts is hindered by both the massive size and diverse nature of the ‘omic’ data. Recent advances with Deep Learning Neural Networks (DLNN), suggests that DLNN could be trained on large data sets to efficiently predict therapeutic responses. We present the application of Association Rule Mining (Market Basket Analysis) in combination with Deep Learning to integrate and extract knowledge in the form of easily interpretable rules from the molecular profiles of 689 cancer cell lines and predict pharmacological responses to 139 anti-cancer drugs. The proposed algorithm achieved superior classification and outperformed Random Forests which currently represents the state-of-the-art classification process. Finally, the in silico pipeline presented introduces a novel strategy for identifying drug combinations with high therapeutic potential. For the first time, we demonstrate that DLNN trained on a large pharmacogenomic data set can effectively predict the therapeutic response of specific drugs in specific cancer types, from a large panel of both drugs and cancer cell lines. These findings serve as a proof of concept for the application of DLNN to predict therapeutic responsiveness, a milestone in precision medicine.