Back To Schedule
Tuesday, July 28 • 17:10 - 17:15
Network perspective on metabolic diversity among mononuclear phagocytes

Log in to save this to your schedule, view media, leave feedback and see who's attending!

The diversity of myeloid cells across different tissues is truly astonishing, both in function and in their developmental trajectory. Additional dimension of this diversity is manifested by the metabolic characteristics of individual phagocytes which can vary significantly based on the cell type and its location. At present, direct metabolomics profiling of tissue residing subpopulations is not feasible, as the process of ex vivo sorting can be lengthy and cause significant metabolic perturbations. However, RNA levels are significantly more stable to the sorting process and can serve as a reasonably reliable proxy to activities of metabolic pathways. In this work we focus on understanding metabolic variability across phagocytic subpopulations through integrated examination of several large-scale datasets that transcriptionally profiled subsets of myeloid cells.
Specifically, we have assembled compendium of three datasets, including first public release of the new dataset generated by Mononuclear Phagocytes Open Source ImmGen project. This dataset totals 337 samples and provides a unique source of information about individual cell subpopulations. It extends previous ImmGen effort that included 202 samples of various mononuclear phagocytes, also analysed in this study. Furthermore, we have leveraged recently released single-cell RNA-seq profiling of the multiple murine organs and reanalysed those data by focusing only on the mononuclear phagocytic populations, comprising 36,480 cells across 18 tissues.
Using these transcriptional data, we sought to identify major metabolic features characteristic of different populations of phagocytic cells, and define how these features vary across cell types and locations. This is computational task that has not been address previously for the datasets of such scale. Indeed, a previously described computational approach, called GAM (PMID: 27098040) uses metabolic networks as the backbone for analysis of transcriptional data and provides a verifiable and systematic description of the metabolic differences. However, datasets in question contain hundreds of individual profiles, while GAM approach is designed to analyse comparison between two conditions. Therefore, in this work we have developed novel computational approach, GAM-clustering, which performs unbiased search of a collection of metabolic subnetworks that jointly define metabolic variability across large datasets. By doing so, GAM-clustering reveals metabolically similar subpopulations in a manner that does not require explicit annotation or pair-wise comparison of individual samples. Our analysis revealed major metabolic features associated with different cell subpopulations and highlighted a number of metabolic modules that are specific to individual cell types, tissues of residence, or developmental stages. As an example, GAM-clustering analysis revealed that cholesterol pathway might play an important role in the context of migratory dendritic cells (DC), which we validated using in vivo pharmacological inhibition of this pathway followed by tracking of DC migration. Consistent with the analysis, statins have demonstrated inhibitory effect on DC migratory ability, finding that has not been reported previously.
Taken together, our work provides both (1) unique data and analysis resource in terms of studying variability of phagocytes, as well as (2) validated computational approach that can unbiasedly analyse both single-cell RNA-seq data as well as multi-sample bulk RNA-seq datasets in terms of underlying metabolic features.

avatar for Anastasiia Gainullina

Anastasiia Gainullina

PhD student, ITMO University
Gene Expression Analysis, Biological Networks (Metabolic, etc), Teaching

Tuesday July 28, 2020 17:10 - 17:15 MSK
Zoom Conference https://zoom.us/j/94321101353?pwd=QlJBb09uM0NVVnVyK0FkbTJ3Nkcrdz09