Using Systems Biology to Determine Possible Modulators of Inflammation in Patients with COVID-19
This web page was produced as an assignment for an undergraduate course at Davidson College.
A single cell-based computational method was used to predict proteins that modulate the dysregulated inflammatory response and determine novel-therapeutic targets against COVID-19.
Inflammation is an important defense mechanism to pathogenic factors such as infection, chemical substances, or tissue injury. To regulate the level of inflammation, tissue cells release cytokines to communicate with each other and activate cell specific functions necessary to clear the infection (Bennett et al. 2018). The immune response is tightly controlled to minimize tissue injury and restore homeostasis. When the infection cannot be cleared a subtle inflammatory response leads towards a hyper-inflammatory condition called cytokine storm due to an excessive release of cytokines and an accumulation of immune cells in the tissue (Bennett et al. 2018). A novel highly contagious virus called COVID-19 has been identified as the main pathogen of an ongoing outbreak of viral pneumonia (He et al. 2020). Patients with severe symptoms of COVID-19 develop a hyper-inflammatory immune response. Although, studies have begun in assessing the immune status of patients, the molecular mechanisms of the hyper-inflammatory response in patients with severe symptoms are not known. The challenge today is how to mitigate the cytokine storm without impairing the clearance of the disease.
Current strategies for regulating the immune response relies on identifying biomarkers as therapeutic targets or in vitro screening of compounds. These methods are labor intensive and are mostly not effective which halts the design of new therapeutic strategies for modulating inflammation (Netea et al. 2017). Single cell technologies have enabled the analysis of multiple cell population tissues at an unprecedented resolution and allows the development of computational methods that could identify immunomodulatory therapeutic targets and compounds inhibiting them (Netea et al. 2017). There has been no computational method for predicting immunomodulatory target proteins that alleviates dysregulations of the physiological immune responses.
In this paper, the researchers present a single cell-based computational method for predicting protein targets to modulate the inflammatory response. The method infers the functional cell-cell communication networks of two tissue-specific single-cell profiles of a dysregulated immune response by integrating a collection of 1756 extracellular receptor-ligand interactions from more than 6000 protein-protein interactions with intracellular signaling and gene regulatory networks. Once the network has been inferred, to construct positive feedback loops, the method examines for those interactions causing the expression of ligands secreted by this cell population. Targeting these positive feedback loops is a plausible therapeutic strategy to modulate the immune response. To validate the method for predicting protein products, the researchers applied it to 12 diseases characterized by hyper-inflammatory or chronic inflammation. Finally, the scientists applied the method to a dataset of bronchoalveolar lavage fluid from patients with mild and severe symptoms of COVID-19.
To detect positive feedback loops established between immune and non-immune cells, 6000 protein-protein interactions between receptors and ligands were manually curated and identified 1756 extracellular interactions. Transcription factors were selected based on expression preserved across all cells in each population and identified the active receptors regulating them by using a chain model that assesses signal transduction possibilities. Once the functional cell-cell communication network has been predicted the method searches for interactions by determining the existence of a sustained regulatory path from the receptor to the ligand. Positive feedback loops are established by combining extracellular and intracellular ligand-receptor interactions.
To validate the method, scientists decided to apply it to a vast array of diseases that are characterized by a pathological immune response. They identified two target proteins having the highest scores. In the cases of liver cirrhosis and macular degeneration, they identified approximately 15 and 20 target proteins. Validation with previous literature revealed evidence for on average 90 percent of the predicted immunomodulatory proteins.
A case study was performed on single-cell RNA sequencing data of nine patients with COVID-19 with mild and severe symptoms and three healthy individuals. The authors aggregated data of different patients showing mild and severe symptoms along with healthy individuals into a single representative sample of each condition and clustered cells to identify cell types using known sets of markers. Differential expression analysis was performed for each cell type under both conditions. The researchers observed that mild and severe groups were significantly different in the level of expression of inflammatory molecules.
Although there are clinical trials ongoing that examines the efficacy of drugs in the treatment of patients with a hyperinflammatory immune response, no effective treatment has been found. The scientists tried to use their methodology to identify proteins that could be targets for modulating the immune response in patients with severe symptoms of COVID-19. They first identified the positive feedback loops underlying mild and severe cases and detected these groups. Three groups were identified. The first group represents the immune response common to patients with severe and mild symptoms and is characterized by feedback between pro and anti-inflammatory cytokines predominantly released by macrophages and dendritic cells. The second and third groups contain interactions unique to patients with mild and severe symptoms.
After determining the molecular differences and commonalities in patients with mild and severe COVID-19 symptoms, the investigators wanted to identify potential target genes for modulating the immune response. They simulated the effects of inhibiting ligands and receptors by manually removing each gene individually from the positive feedback loops. The simulation identified an extracellular matrix glycoprotein known as versican (VCAN) and toll-like receptor 2 as novel target genes whose inhibition disrupts around 75% of the interconnected feedback loops unique to severe cases while not interfering with the immune response mechanisms common to mild and severe cases.
In this study the scientists came up with a computational method for predicting immunomodulatory compounds and target proteins to treat severe symptoms in patients with COVID-19. The method detects and utilizes the amplifying feedback loops governing the dysregulated inflammatory response, providing a holistic view of the extracellular cell-cell communication networks underlying this disease. This is the first method incorporating molecular information about inflammatory processes. The proposed methodology relies on positive feedback loops which play a key role in the amplification of the immune response to diseases. Using the method, the researchers were able to identify VCAN and TLR2 as potential targets for immunomodulatory attempts which was further confirmed by analyzing two independent patients with severe symptoms. Over-expression of VCAN in severe disease cases results in the excessive accumulation of pro-inflammatory monocytes in the lungs which constitutes a plausible, novel target gene for modulating the hyper-inflammatory response in patients with severe symptoms of COVID-19. Despite its ability to predict regulatory proteins and compounds there are limitations. It requires single cell RNA-seq data of tissues displaying pathological and physiological immune responses which is not widely available. Other signaling mechanisms exist that are important for establishing a proper inflammatory response, such as thorough exchange of exosomes which could be an experiment to think about in the future. Overall, the researchers believe that their method is useful in characterizing the pathological immune response and in the design of novel therapeutic interventions for a wide range of diseases associated with persistent inflammation including COVID-19. The article does a great job talking about the contents of the new computational method they created and explains how they validated their method multiple times.
Nikhil Virani is currently enrolled at Davidson College. Contact him at nivirani@davidosn.edu
References
Bennett J. M., G. Reeves, G. E. Billman, and J. P. Sturmberg, 2018 Inflammation–Nature’s Way to Efficiently Respond to All Types of Challenges: Implications for Understanding and Managing “the Epidemic” of Chronic Diseases. Front. Med. 5: 316. https://doi.org/10.3389/fmed.2018.00316
He R., Z. Lu, L. Zhang, T. Fan, R. Xiong, et al., 2020 The clinical course and its correlated immune status in COVID-19 pneumonia. Journal of Clinical Virology 127: 104361. https://doi.org/10.1016/j.jcv.2020.104361
Netea M. G., F. Balkwill, M. Chonchol, F. Cominelli, M. Y. Donath, et al., 2017 A guiding map for inflammation. Nat Immunol 18: 826–831. https://doi.org/10.1038/ni.3790
© Copyright 2020 Department of Biology, Davidson College, Davidson, NC 28036.
