SPaRTAN, a computational framework for linking cell-surface receptors to transcriptional regulators

Our plain-English summary of SPaRTAN, a computational framework for linking cell-surface receptors to transcriptional regulators, by X. Ma, A. Somasundaram, Z. Qi, et al (Nucleic Acids Research, 2021).

The study

• Authors: X. Ma, A. Somasundaram, Z. Qi, D.J. Hartman, H. Singh, H.U. Osmanbeyoglu
• Journal: Nucleic Acids Research
• Published: 2021;49(17):9633-9647
• Cited by: 8 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

The identity and functions of specialized cell types are dependent on the complex interplay between signaling and transcriptional networks. Recently single-cell technologies have been developed that enable simultaneous quantitative analysis of cell-surface receptor expression with transcriptional states. To date, these datasets have not been used to systematically develop cell-context-specific maps of the interface between signaling and transcriptional regulators orchestrating cellular identity and function. We present SPaRTAN (Single-cell Proteomic and RNA based Transcription factor Activity Network), a computational method to link cell-surface receptors to transcription factors (TFs) by exploiting cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq) datasets with cis-regulatory information. SPaRTAN is applied to immune cell types in the blood to predict the coupling of signaling receptors with cell context-specific TFs. Selected predictions are validated by prior knowledge and flow cytometry analyses. SPaRTAN is then used to predict the signaling coupled TF states of tumor infiltrating CD8+ T cells in malignant peritoneal and pleural mesotheliomas. SPaRTAN enhances the utility of CITE-seq datasets to uncover TF and cell-surface receptor relationships in diverse cellular states.

© The Author(s) 2021. Published by Oxford University Press on behalf of Nucleic Acids Research.

In plain English

This article described SPaRTAN, a computer method the authors developed to link proteins on the surface of single cells (receptors) to the predicted activity of transcription factors, the proteins that switch genes on and off. It used CITE-seq data, which measures surface-protein levels and messenger RNA (gene expression) levels in the same cells. The authors first tested it on blood immune cells from a healthy donor. There it predicted gene expression significantly better than a simpler comparison method (P less than 0.001). They checked selected predictions by flow cytometry on blood from 9 healthy donors. They then applied the method to CD8+ T cells, a type of immune cell, taken from malignant peritoneal and pleural mesothelioma tumors. In both tumor types, a group of 'exhausted' CD8+ T cells showed increased predicted activity of the factor TCF7. In peritoneal mesothelioma, predicted BCL3 activity was correlated with the receptor PD-1. Staining of tumor specimens showed the two proteins together in CD8+ T cells. The authors called their analysis a proof-of-concept. The study did not examine asbestos exposure.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Computational methods study: development and testing of a computer model, with laboratory checks of selected predictions by flow cytometry and immunohistochemistry
• Who was studied: Single cells from a public blood-cell (PBMC) CITE-seq dataset from a healthy donor. An in-house CITE-seq dataset of human malignant peritoneal and pleural mesothelioma tumors from the University of Pittsburgh (number of patients not stated in the text). Blood from 9 healthy donors for flow cytometry. Peritoneal mesothelioma specimens from two patients for immunohistochemistry.
• Exposure: Not applicable: the article does not study asbestos or any other exposure. It analyzed surface-protein and mRNA measurements in single cells.
• Main results:
  • In blood immune cells, the cell-type specific SPaRTAN models predicted gene expression significantly better than a nearest-neighbour approach (P less than 0.001, one-sided Wilcoxon signed-rank test)
  • Flow cytometry checks used B, CD8+ and CD4+ memory T cells from the blood of 9 healthy donors
  • Consistent with the predictions, flow cytometry showed increased FOSL2 and decreased NFKB in CD27+ CD8+ T and CD4+ memory T cells compared with CD27-negative cells
  • In malignant peritoneal mesothelioma CD8+ T cells, one cluster had high PD-1, TIM3 and TIGIT (characteristic of exhausted cells). It showed increased inferred activity of TCF7, STAT6, BCL3 and FOS (FDR-corrected P less than 10 to the power of minus 50).
  • In malignant pleural mesothelioma CD8+ T cells, the exhausted-type cluster (Cluster 4) also showed increased inferred TCF7 activity (FDR-corrected P less than 10 to the power of minus 50)
  • Predicted HIF1A activity was correlated with PD-1 protein expression in CD8+ T cells from blood, peritoneal mesothelioma (printed as 'MePM') and pleural mesothelioma
  • Predicted BCL3 activity was correlated with PD-1 in peritoneal mesothelioma CD8+ T cells but not in pleural mesothelioma or blood CD8+ T cells. Immunohistochemistry on peritoneal mesothelioma specimens found PD-1 and BCL3 co-expressed in CD8+ T cells.
• The authors' conclusion: The authors concluded that SPaRTAN generates hypotheses about signaling-regulated transcription factors that can be tested in the laboratory or clinic. They described their analysis as a proof-of-concept.
• Limitations the authors state:
  • The curated transcription factor and target-gene interactions used by the model are noisy, incomplete and not context-specific
  • The model does not represent whether a transcription factor activates or represses a gene
  • More complex combinations of transcription factor binding, and crosstalk between receptor signals, are not modelled
  • The analysis is limited to about 200 surface proteins for which barcoded antibodies are commercially available
  • The analysis is a proof-of-concept and is not large or diverse enough to serve as a reference set

Topics

Medical terms in this study, explained in our Lexicon: Pleural Mesothelioma.

Related studies

• Expression of growth factor and growth factor receptor RNA in rat pleural mesothelial cells in culture — E. Bermudez, J. Everitt, et al. (1990)
• The plexin-A1 receptor activates vascular endothelial growth factor-receptor 2 and nuclear factor-kappaB to mediate survival and anchorage-independent growth of malignant mesothelioma cells — A. Catalano, R. Lazzarini, et al. (2009)
• Human mononuclear phagocytes express the insulin-like growth factor-II/mannose-6-phosphate receptor — W.N. Rom (1991)
• Urokinase receptor in human malignant mesothelioma cells: role in tumor cell mitogenesis and proteolysis — S. Shetty, A. Kumar, et al. (1995)
• Expression and activity of urokinase and its receptor in endothelial and pulmonary epithelial cells exposed to asbestos — A. Barchowsky, R.R. Roussel, et al. (1998)
• Asbestos upregulates expression of the urokinase-type plasminogen activator receptor on mesothelial cells — R.C. Perkins, V.C. Broaddus, et al. (1999)

Browse return to medical research.