Abstract

Mononuclear phagocytes have intrinsic functions in development and tissue remodeling as well as extrinsic functions in host defense, antigen presentation and innate immunity. To obtain an unbiased view of genes that are characteristic of the macrophage, we compare gene expression in the 15 macrophage gene expression profiling studies (185 samples) with data from 544 platform matched control studies (10,160 samples) using a nonparametric rank based metric. 769 genes show consistently higher expression in both human and mouse macrophage samples defining an unbiased macrophage gene set. Automated natural language processing defines a second group of macrophage related genes. 545 genes are mentioned frequently in sentences containing the word “macrophage” in a PubMed abstract. We refer to these as the literature gene set. 87 genes occurred in both the unbiased and the literature gene sets. Annotations of the unbiased and literature gene sets are remarkably consistent with a 0.93 correlation in the normalized frequency of Gene Ontology (GO) term use in the two sets relative to GO term use for all genes in the genome. The most highly over represented GO terms for macrophage specific genes include: chemotaxis, chemokines and chemokine receptors, inflammation, cytokine and chemokine signaling, JAK-STAT cascade, peroxidase activity. To better define the genes involved in host defense functions, we considered evolutionary behavior. Genes involved in intrinsic host functions show purifying selection and are constrained in their rate of evolutionary change. In contrast, host defense genes (HDG) may be subject to diversifying selection and rapid rates of evolutionary divergence. To control for different rates of mutation at different loci, analysis was based on the rate of nonsynonymous to synonymous mutations (dN/dS). As a group, macrophage genes exhibit higher rates of change (median dN/dS=0.17) compared to all genes in the genome (median dN/dS=0.14, P<1e-6). Cell surface receptors (FCGR3a&b, IL15Ra, CD33, C-type lectin, CD1e and CD54) and secreted ligands (CGA, IL3, osteopontin, CCL1, CXCL13 and CXCL2) dominate the list of most rapidly evolving macrophage genes for both the unbiased and literature gene sets. Variation in dN/dS is seen, even within a single gene with secreted and extracellular domains exhibiting consistently higher rates of change compared to intracellular domains. The GO terms with the highest dN/dS ratios are interleukin receptors, pattern recognition receptors, IgG binding, MHC class I and II and T cell activation. We found many macrophage genes that show high dN/dS ratios including IL15RA, CD33, CSN3, CGA, SPP1, ubiquitin D, CD300a, CXCL13, CD1E, CXCL2, ICAM1, CD34 and APOE. The accelerated rates of evolution of these genes, their expression in the macrophage and often their extracellular location, suggest that these may be novel HDG. Overall, 1064 of the macrophage genes are found in the MiMI database of molecular interactions. 2,493 interactions are found among 683 genes in the macrophage gene set. Thus macrophage specific genes operate as a highly connected network. In summary, we present an unbiased systems biology approach to define genes characteristic of mononuclear phagocytes and macrophage host defense responses, and we identify numerous and novel candidate host defense genes.

Supported by NIH grants R01 LM008106, R01 CA85368, U54 DA021519.

Author notes

Disclosure: No relevant conflicts of interest to declare.