Background and motivation
Fibrosis is a process that can arise from injury and inflammation. Under normal conditions, an injury remits by wound healing and scarring. However, chronic injury or inflammation generates an excess of connective tissue and proliferating fibroblasts which build up an excess extracellular matrix (ECM) and may affect the correct performance of the involved organ or tissue. Both excess of fibrotic cells and aberrant remodelling of ECM empowers further inflammatory signals that lead to chronic fibrosis, damages of organ, tissue architecture and, ultimately to loss of function. Regardless of the different conditions leading to fibrosis and the diverse organs that can be afflicted by it, the main mechanism involves various cell types mainly (myo)fibroblasts, which produce excessive amounts of ECM components. Fbrosis-related diseases, mainly in the form of chronic kidney disease (CKD), liver cirrhosis or lung fibrosis, are highly prevalent especially in the developed world. Hence, understanding the molecular mechanisms orchestrating them alongside with the identification of early biomarkers and development of efficient therapies is essential (Friedman et al., Sci Transl Med., 2013).
Research with systems biology approaches is relatively scarce when compared with other diseases, in particular in the context of CKD (Saez-Rodriguez et al. 2019). We work on kidney and heart fibrosis in collaboration with the Division of Nephrology and Clinical Immunology at the University Hospital Aachen, in particular with Rafael Kramann. We also work on fibrosis within the Molecular Medicine Partnership Unit (MMPU) on CKD with EMBL, together with Matthias Gaida, Rainer Pepperkok and Christoph Merten. We also work in liver fibrosis in the context of the LiSyM network.
Data and approaches
In order to gain deeper insight into the mechanisms driving fibrosis and to find possible markers and treatments, we consider several types of data and a variety of modeling approaches. In the case of transcriptomic data, we try to reduce the vast amount of the transcriptome into a functional context by applying our functional genomics footprint methods (PROGENy, DoRothEA). Using single cell RNA-seq data, we are identifying cell-type-based molecular signatures. We then attempt to find cell-cell interactions. Furthermore, we integrate different layers of omics data and approaches in order to unveil the signaling mechanisms and metabolism driving fibrotic processes.
Our research focuses on liver and kidney fibrosis. From these, we obtain different types of data as well as curated prior knowledge on already well-studied mechanisms involved from our tool Omnipath (Türei et al., Nat Methods, 2016) and other resources. We integrate them using available or in house tools to gain further insight into the underlying processes driving fibrosis. We hope to leverage the obtained knowledge to find potential new markers and therapeutic targets (Saez-Rodriguez et al. 2019).
In a preliminary study, we focused on the meta-analysis of public transcriptomics data from human CKD samples with different functional genomic techniques to characterize the differences across CKD subtypes (Tajti et al. 2020). We aim to relate molecular data with clinical parameters such as the filtering function of the kidney (glomerular filtration rate, GFR) using machine learning-driven approaches capable of exploiting several feature selection and dimensionality reduction methods. Simultaneously, we are integrating these techniques with multiple omics data in order to create a comprehensive model of kidney fibrosis. Furthermore, we are trying to identify disease-causing changes in tissue microenvironment inferred from single-cell RNA-seq. We are currently working on the characterization of cell populations involved in kidney fibrosis and homeostasis with Rafael Kramann from RWTH Aachen University. Among other approaches, we use our functional genomics tools (PROGENy, DoRothEA and CARNIVAL) as well as biological information of drug-response and methods for deconvolution of gene expression. In parallel, we are integrating sequencing data from bulk and single-cell measurements to deconvolute heterogeneity across CKD samples.
Finally, we use microfluidics to do high-throughput screenings of drug combinations and characterize responses of specific cell-types. This is within the Molecular Medicine Partnership Unit on CKD, together with the groups of Christoph Merten and Rainer Pepperkok in EMBL and Susanne Delecluse and Martin Zeier in the Kidney Center in Heidelberg. Our long term aim is to develop multi-cellular models of the kidney tissue to unveil the mechanisms driving fibrosis in multiple scales, namely from the intracellular point of view to the tissue level.
Chronic Liver Disease
In the context of chronic liver disease, fibrosis frequently leads to cirrhosis. The number of global cirrhosis deaths is steadily growing, resulting in over 1 million deaths worldwide (Mokdat et al., BMC Med, 2014). Our contribution to understand this condition is to investigate the dynamic profile of chronic liver disease progression at RNA level in mouse models provided by Marie-Luise Berres from University Hospital Aachen and Ahmed Ghallab from Technical University Dortmund. This is part of the Lisym (Liver Systems Medicine) network. Functional analysis of the transcriptome will decipher disease-stage associated molecular changes which can help us identify critical parameters for liver fibrosis progression.
In a collaboration with Jan Hengstler also from Technical University Dortmund we investigated the influence of liver fibrosis on lobular zonation with the outcome that during fibrosis pericentral regions adopts periportal features (Ghallab et al. 2019).
For future work, we plan to integrate also single-cell RNA-seq datasets.