A draft network of ligand-receptor-mediated multicellular signalling in human.
Ramilowski JA, Goldberg T, Harshbarger J, Kloppmann E, Lizio M, Satagopam VP, Satagopam VP, Itoh M, Kawaji H, Carninci P, Rost B, Forrest AR. · 2015
Read A draft network of ligand-receptor-mediated multicellular signalling in human. by Ramilowski JA, Goldberg T, Harshbarger J, Kloppmann E, Lizio M, Satagopam VP, Satagopam VP, Itoh M, Kawaji H, Carninci P, Rost B, Forrest AR. (2015) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4525178.txt.
Introduction
Jordan A. Ramilowski leads an eleven-author team drafting the wiring diagram of human cell talk, in Nature Communications in 2015. Cells coordinate through ligands, messenger molecules one cell releases, and receptors, the locks they fit on another cell's surface. A full map of such pairs across cell types had been missing.
Drawing on gene data from 144 primary human cell types, the team charted which cells send and which receive each signal. Most cells express tens to hundreds of ligands and receptors, weaving a dense network. About two-thirds of pairs may also act on the same cell that sent them. Surface and secreted proteins proved the most cell-specific, and receptors mostly evolved before their ligands. An online tool lets others query the map.
Catalog record
| CREATOR | Ramilowski JA, Goldberg T, Harshbarger J, Kloppmann E, Lizio M, Satagopam VP, Satagopam VP, Itoh M, Kawaji H, Carninci P, Rost B, Forrest AR. |
|---|---|
| YEAR | 2015 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A draft network of ligand–receptor-mediated multicellular signalling in human By Jordan A. Ramilowski, Tatyana Goldberg, Jayson Harshbarger, Edda Kloppman, Marina Lizio, Venkata P. Satagopam, Masayoshi Itoh, Hideya Kawaji, Piero Carninci, Burkhard Rost, Alistair R. R. Forrest Published in: Nature Communications DOI: 10.1038/ncomms8866 ABSTRACT Cell-to-cell communication across multiple cell types and tissues strictly governs proper functioning of metazoans and extensively relies on interactions between secreted ligands and cell-surface receptors. Herein, we present the first large-scale map of cell-to-cell communication between 144 human primary cell types. We reveal that most cells express tens to hundreds of ligands and receptors to create a highly connected signalling network through multiple ligand–receptor paths. We also observe extensive autocrine signalling with approximately two-thirds of partners possibly interacting on the same cell type. We find that plasma membrane and secreted proteins have the highest cell-type specificity, they are evolutionarily younger than intracellular proteins, and that most receptors had evolved before their ligands. We provide an online tool to interactively query and visualize our networks and demonstrate how this tool can reveal novel cell-to-cell interactions with the prediction that mast cells signal to monoblastic lineages via the CSF1–CSF1R interacting pair. Development of multicellular organisms from unicellular ancestors is one of the most profound evolutionary events in the history of life on Earth1. In this transition, cells of multicellular organisms had to acquire various modes of cell-to-cell (intercellular) communication to develop and then control their coordinate functioning2. This process is critical during early embryonic development where the cell's differentiation and ultimate fate are controlled by communication with neighbouring cells345. In the developed organism, intercellular communication coordinates the activities of multiple cell types required for complex organismal processes such as immune response6, growth7 and homeostasis8. Defects in cell-to-cell communication, including dysregulation of autocrine signalling, are also medically important in cancer9, autoimmune10 and metabolic diseases11. Despite its importance, studies of intercellular communication across specialized cells of higher metazoa have generally focused on communication between only a few cell types and via limited numbers of ligand–receptor pairs. Currently there are no reports of systematic studies trying to elucidate and quantify the repertoire of signalling routes between different cell types. To address this, we have systematically reviewed the expression profiles of 642 ligands and their 589 cognate receptors in our 1,894 literature-supported interacting pairs across a panel of 144 human primary cell types12. In particular, we used known interacting ligand–receptor pairs and public protein–protein interaction (PPI) information to generate the first large-scale draft map of primary cell-to-cell interactions. Highlighting their important role in the evolution of higher order metazoans, we show that receptors and ligands have more cell-type-specific expression profiles and are evolutionarily younger as a class than nuclear and cytoplasmic proteins. Applying a 10 tags per million (TPM; ∼3 transcripts per cell) detection threshold to our data, we find that primary cells express on average less than one-third of all ligands and receptors (roughly 140 ligands and 140 receptors). We also find that messages between any two given cell types are carried in a rather specific manner despite the hundreds of possible connecting paths and have significant potential for autocrine signalling. We also discuss in more detail the level of communication between different cell lineages. Finally, to benefit the research community, we provide an interactive visualization and query tool for ligand–receptor networks in humans (available at http://fantom.gsc.riken.jp/5/suppl/Ramilowski_et_al_2015/). This work is part of the FANTOM5 project. Data download, genomic tools and co-published manuscripts have been summarized at http://fantom.gsc.riken.jp/5/. RESULTS PM and secreted proteins are young and cell-type specific Recently the FANTOM5 consortium used Cap Analysis of Gene Expression (CAGE) to generate a promoter level expression atlas12. Based on CAGE measurements across a collection of 975 human samples (primary cells, cell lines and tissues), gene expression profiles were classified as non-ubiquitous (cell-type restricted), ubiquitous-non-uniform and ubiquitous-uniform (housekeeping)12. Gene Ontology (GO)13 analysis of genes with cell-type-restricted expression showed their enrichment for proteins annotated with the terms receptor activity, plasma membrane (PM) and multicellular organismal process. This suggested that proteins involved in intercellular communication were more likely to have cell-type-restricted expression profiles. To explore this more systematically, we used protein experimental localization information1415 and computational predictions1617 (Methods) to classify human protein-coding genes (HGNC18 release 03 April 2014; http://www.genenames.org/cgi-bin/hgnc_downloads) based on the subcellular localization of the proteins they encode into: PM, secreted, cytosolic, nuclear, multiple and ‘other' proteins (Supplementary Data 1). Comparing the cell-type specificity of each class, we find that secreted and PM proteins are significantly more cell-type specific (Fig. 1) than proteins that localize to other cellular compartments (Mann–Whitney U-test, each adjusted P value<0.001). We also confirmed this trend using whole cell proteome data available for five haematopoietic primary cell types19 (Mann–Whitney U-test, each adjusted P value<0.001; Supplementary Fig. 1). As cell-type-specific proteins are likely to appear with the emergence of new cell types and increased organismal complexity, we next examined the predicted ages of proteins from each subcellular localization using Protein Historian20 (pre-computed estimates based on Wagner parsimony21 and P-POD's22 OrthoMCL23 clustering of proteins in the PANTHER24 database were used). Evolutionary profiles of proteins from the different subcellular localizations show that secreted proteins (average age 412.2 mya) and PM proteins (average age 517.2 mya) are younger (Mann–Whitney U-test, each adjusted P values<0.001) than proteins that localize to the nucleus (average age 663.1 mya), cytoplasm (average age 855.1 mya) (Supplementary Data 1; Fig. 1c,d) or to other compartments. Using additional protein age estimates2526, also confirmed the trend that PM and secreted proteins are generally the youngest proteins (Supplementary Fig. 2). Identification of putative ligand–receptor pairs We next sought to examine in more detail PM and secreted proteins specifically involved in cell-to-cell communication. Building on previous efforts to curate lists of ligand–receptor pairs, we merged the lists from Database of Ligand−Receptor Partners (DLRP)27, IUPHAR28 and Human Plasma Membrane Receptome (HPMR)29 databases to generate a non-redundant set of 1,179 known interacting ligand–receptor pairs. Given that these resources originated many years ago, and are not extensively updated, we found many genuine ligand–receptor pairs were missing, for example GDF2->ACVR1 (ref. 30) and CCL4->CCR3 (ref. 31). To extend this set, we first expanded the lists of candidate ligands and receptors by incorporating proteins predicted to be secreted and localized to the PM, respectively. We then searched for PPIs between all putative ligands and putative receptors (Supplementary Fig. 3a) as described in the Methods section. From this analysis, we inferred 2,117 experimentally supported interactions in the HPRD15 and STRING32 databases, which included 1,288 ligand–receptor pairs absent from our known collection of DLRP, IUHPAR and HPMR interactions. From the above, we compiled a unique list of 2,467 known and inferred interactions. We next aimed to curate these interactions with a primary citation (PubMed ID), either by extracting the reference from the primary data sources (IUHPAR, HPMR and HPRD) or by manually searching the literature. Through the manual curation, we excluded 135 pairs, as the partners were not a genuine ligand or receptor, and found an additional 90 pairs. This resulted in a final curated set of 2,422 interactions: 1,894 interactions with primary literature support which we refer to as ‘reference' and use in our subsequent analysis, and the remaining set of 528 curated interactions without primary literature support we refer to as