The transcriptome of human monocyte subsets begins to emerge
© BioMed Central Ltd 2009
Published: 23 December 2009
Human monocytes can be divided into subsets according to their expression or lack of the cell-surface antigen CD16. In papers published recently in the Journal of Proteome Research and in BMC Genomics, two groups publish independent transcriptome analyses of CD16+ and CD16-monocytes, with revealing results.
See research article http://www.biomedcentral.com/1471-2164/10/403
Monocytes are a heterogeneous group of cells constituting 5-10% of the total white blood cells in humans. They originate in the bone marrow, circulate in the bloodstream and enter tissues, where they differentiate into macrophages, either to replenish the stock of tissue macrophages or to contribute to an inflammatory response to infection. Monocytes can remain in the circulation for up to 72 hours, after which, if they have not been activated, they die and are removed.
The heterogeneity of monocytes was noticed soon after their definition, and it includes differences in density, production of reactive oxygen species, antigen-presenting capacity, maturation status, and phagocytic and adhesive properties. In 1989, Ziegler-Heitbrock and colleagues  noticed that human monocytes can be divided into three main populations according to their expression of the cell-surface antigens CD16 (Fcγ receptor III) and CD14 (a receptor for bacterial lipopolysaccharide (LPS)). The CD16 Fcγ receptor is a relatively low-affinity receptor for the Fc portion of IgG antibodies in complex with their antigens, and stimulates the monocyte to take up antibody-antigen complexes by phagocytosis and thus remove them from the circulation. CD14 is essential for the recognition of bacterial LPS present in Gram-negative bacteria, which include many common pathogens. CD14 acts as pattern recognition protein which accepts LPS from LPS-Binding protein. To elicit the endotoxin cellular response the CD14-LPS complex interacts with various Toll like receptors (TLR) including TLR4-MD2 (myeloid differentiation factor-2), TLR2/TLR6 and TLR2/TLR1.
The ratio of CD16+ to CD16- monocytes changes greatly in disease. Results accumulated over the past decade suggest that the CD16+ subset is expanded in a vast number of inflammatory diseases, irrespective of their etiology. A higher CD16+ to CD16- ratio (as compared with that typical of a healthy person) has been found in Crohn's disease , rheumatoid arthritis, asthma and sepsis, among other diseases . In hypoxic episodes, such as those caused by myocardial infarction and stroke, higher levels of CD16- monocytes (but not CD16+) correlate with disease severity and poor outcome. The increase in numbers of CD16+ monocytes are such a general feature of disease that, although it can indicate severity and outcome, it cannot suggest a specific diagnosis. Various explanations have been proposed for the increase in circulating CD16+ cells in inflammatory disease: the maturation of CD16- cells into CD16+ cells; the increased movement of CD16- monocytes out of the blood vessels into tissues; and even the stimulation of a putative CD16+ monocyte developmental pathway.
Investigating monocyte subset transcriptomes
In the past decade, we have learned a great deal about the functional and phenotypic differences between the monocyte subsets, with most investigations looking at well-established molecules or defined monocytic properties. In the late 1990s and early 2000s, with the sequencing of the human and other genomes, microarray technology for studying the transcriptomes of human cells became available, and in the past few years this methodology has been used to investigate the detailed differences between human monocyte subsets. Microarrays are now commercially available that allow investigation of the expression of all known human genes and as-yet-unidentified transcribed sequences.
At least three microarray studies focused on monocyte subsets have been published. In 2007, Mobley et al.  studied the differences in gene expression between monocyte subsets using the Affymetrix Human 133A 2.0 array, which analyzes the expression level of 18,400 transcripts and variants, including 14,500 well characterized human genes. In this study , human peripheral mononuclear cells (PBMCs) were isolated from total blood using a Ficoll gradient and were subsequently incubated with paramagnetic beads conjugated with specific antibodies to deplete natural killer (NK) cells, T lymphocytes and B lymphocytes. After separation the resultant fraction, highly enriched in monocytes, was then split to isolate CD16+ and CD16- monocytes. CD16+ monocytes were positively selected using an anti-CD16 antibody. To isolate CD16- monocytes the authors  used magnetic beads conjugated with antibodies against the cell adhesion molecule CD62L, also known as L-selectin. This molecule is found on most peripheral leukocytes and divides human monocytes into two fractions. The CD62L+ monocytes are mostly CD14++CD16-, whereas the CD62L- monocytes are CD14+CD16+. This procedure of isolation differs from that used in other studies [6, 7].
Mobley et al.  provided a selective list of genes differentially expressed in CD16+ and CD16- monocytes. The list includes 15 genes highly expressed in CD14++CD16- monocytes and 19 genes highly expressed in CD14+CD16+ monocytes. They found that CD16+ monocytes have higher mRNA levels of known subset biomarkers such as CD16 and the chemokine receptor CX3CR1, but also new markers such as the colony-stimulating factor 1 receptor (CSF1R), the receptor for macrophage colony-stimulating factor (MCSF) and the complement component factors C1QA, C1QB and C3. MCSF is a potent maturation signal for monocytes and a survival and proliferative factor for macrophages and their precursors, and it is required for the development of many types of tissue macrophage.
In CD16- monocytes Mobley et al.  found higher expression of CD14 and the chemokine receptor CCR2. Their data show for the first time higher expression in CD16- monocytes of the colony-stimulating factor 3 receptor (CSF3R). Colony-stimulating factor 3 (CSF3; also called GMCSF) is another important maturation factor for monocytes and is also a maturation factor for granulocytes. In this study , no details about the total number of differentially expressed genes or ways of accessing the dataset are given.
Recently, two other investigations have been published, by Zhao et al.  in the Journal of Proteome Research and Ancuta et al.  in BMC Genomics. Both of these comply with MIAME (Minimal Information about Microarray Experiments) requirements and provide their data. In a well executed study, Zhao et al.  used magnetic beads carrying anti-CD16 antibodies to isolate CD16+ monocytes from PBMCs that had been depleted of NK cells and neutrophils. CD16- monocytes were isolated from the CD16-negative fraction using anti-CD14 beads. CD16+ and CD16- monocytes were also isolated by fluorescence-activated cell sorting (FACS) from total monocytes purified by CD14-positive selection. The transcriptomes of the two subsets were then defined using the Illumina BeadArray HG-6v2. In this study , 521 genes were scored as differentially expressed between the subsets: 305 characterized the CD16- subset and 216 the CD16+ subset. The authors also investigated differences between subsets at the protein level. The proteomic approach showed that out of 1,006 proteins robustly expressed, 235 were differen tially expressed between the subsets: of these, 123 proteins characterized the CD16+ monocytes and 112 the CD16- monocytes.
The three most represented Gene Ontology (GO) categories for differentially expressed mRNAs were cellular growth and proliferation, cell death, and metabolism; for differentially expressed proteins they were cell death, metabolism, and cellular assembly . Known subset biomarkers such as CD16 and CD14, the chemokine receptors CX3CR1 and CCR2, the integrin alphaL (ITGAL) and alphaM (ITGAM) chains, and the leukocyte adhesion molecule CD62L were among the genes modulated at the mRNA level (Figure 1), of which CD16, ITGAL and ITGAM were also identified at the protein level.
In addition to CD16 itself, in the CD16+ subset the authors  found overexpression of genes that participate in FcγR-mediated phagocytosis. They confirmed higher mRNA levels for heme oxygenase 1 (HMOX1), villin 2 (VIL2), hematopoietic cell kinase (HCK) and the tyrosine protein kinase Lyn (LYN). At the protein level they confirmed higher expression of actin-related protein 2/3 complex (ARP2 and ARP3), HCK and LYN. In the CD16- subset they found instead overexpression of genes involved in antimicrobial functions. In this subset were confirmed higher mRNA levels of myeloperoxidase (MPO), lysozyme C (LYZ), Protein S100-A9 (S100A9), eosinophil cationic protein (RNASE3) and phospholipase B domain containing 1 (PLBD1, also FLJ22662). Higher protein levels were confirmed for cathepsin G (CTSG), MPO, LYZ and S100-A9. Among other interesting conclusions, this study  clearly showed that mRNA and protein levels do not always correlate in the subsets. Thus, mRNA levels seem to represent the potential, more than the actual, functional capacity of the monocytes.
Ancuta et al.  isolated total monocytes from PBMCs by negative selection using magnetic beads to remove the other cell types; the CD16+ monocytes were subsequently isolated by positive selection using anti-CD16 magnetic beads. The transcriptomes were defined using the Affymetrix HGU133 microarray. Applying rigorous and exemplary statistics, the authors defined a set of 361 genes that distinguish CD16+ from CD16- monocytes: these comprise 172 genes and unknown transcribed sequences that are highly expressed in CD16+ monocytes, and 189 genes and unknown transcribed sequences highly expressed by CD16- monocytes. Applying more stringent statistics they provide a shortlist of 61 genes, of which 30 transcripts are upregulated in CD16+ and 31 in CD16-monocytes.
Classifying their larger dataset of differentially expressed genes, Ancuta et al.  found over-representation in key GO categories, including immune response, inflammation, metabolism and stress response, cell cycle, proliferation and differentiation. They also found over-representation of more informative functional subcategories: for example, cytokines, chemokines and complement (both ligands and receptors); signaling and signal transduction; cytoskeleton; and transcription factors.
Ancuta et al. , like the other authors [5, 6], confirmed expected subset biomarkers. All three authors found at the mRNA level, and Ancuta et al.  at the protein level, that CD16+ monocytes express higher levels of CSFR1. In agreement with Zhao et al. , Ancuta et al.  found that CD16+ monocytes have higher mRNA levels of the IL-12 receptor 1 (IL12RB1). IL-12 is a cytokine produced by activated monocytes, macrophages and dendritic cells and is essential for resistance to bacterial and intracellular parasite infection. In addition they  found differential expression of the complement component C3 receptor 1 (C3AR1) on CD16+ monocytes. This receptor recognizes the chemotactic and inflammatory peptide anaphylatoxin C3a. C3a is one of the products of the proteolytic cleavage of complement component C3, which was found to be highly expressed by CD16+ monocytes by Mobley et al. .
This gene-expression pattern functionally contrasts with the higher mRNA levels in CD16- monocytes of the IL-13 receptor 1 (IL13RA1), found by Zhao et al.  and Ancuta et al. . IL13RA1 is a subunit of one of the receptors for IL-4 and IL-13. These cytokines induce the 'alternative' activation of monocytes and macrophages, enhancing macrophage capacity for fluid-phase pinocytosis and endocytosis, and inducing giant cell formation and specific gene signatures. Zhao et al.  and Ancuta et al.  found at the mRNA level, and Ancuta et al.  confirmed at the protein level, that CD16- monocytes express higher levels of CD93, the receptor for complement component C1q1 (also called C1QR1). This receptor is part of a larger receptor complex for C1q complement factor, mannose-binding lectin (MBL2) and pulmonary surfactant protein A (SPA), all proteins that enhance phagocytosis in monocytes.
All three groups [5–7] found that colony-stimulating factor 3 receptor (CSF3R) is highly expressed at the mRNA level in CD16- monocytes and Ancuta et al.  demonstrated it at the protein level. The contrasting expression of the colony-stimulating factor receptors CSF1R and CFS3R increases the repertoire of confirmed membrane markers that characterize human monocyte subsets. These findings reveal an unnoticed compensatory loop in the activation balance and perhaps even origin of the subset phenotypes. CD16+ pro-inflammatory monocytes express higher levels of MCSF receptor (CFS1R); MCSF induces macrophages with less pro-inflammatory capacity than GMCSF; and the GMCSF receptor (CFS3R) is in turn highly expressed in CD16- classical monocytes .
The discrepancy between the studies may be due to differences in cell isolation methodology and the purity of the cell populations isolated, the use of negative versus positive selection, and the microarray methodology, among other factors. In fact, although Barnes et al.  and others have demonstrated that Affymetrix and Illumina platforms yield highly comparable data, especially for genes predicted to be differentially expressed, the platforms use different amounts of total RNA for the hybridization and different probes to identify the genes and even distinct solid supports for the probes . Further transcriptomic and proteomic studies will clarify the discrepancies found so far and will shed light on this topic.
Comparison with macrophage maturation transcriptomes
It has been hypothesized that the difference between monocyte subsets is due to a difference in stage of maturation . This could be directly influenced (at least in part) by levels of MCSF in the environment and by the differential expression in the subsets of the receptor for MCSF. It is therefore of interest to compare the genes that distinguish the monocyte subsets [6, 7] with those involved in the maturation of total human monocytes induced by MCSF in vitro and in macrophage activation (approximately 3,530 genes in total) . Representatives of the latter category of cells are macrophages stimulated with a combination of interferon gamma (INF-γ) and LPS, which induces a classical pro-inflammatory and antimicrobial phenotype in macrophages, and IL-4, which as previously mentioned induces an alternative type of activation. This comparison shows that out of the 434 genes extracted as differentially expressed from , 190 are also contained in the maturation/activation gene set, and out of the 318 genes identified from , 180 overlap with the maturation/activation data.
Of the 145 differentially expressed genes found by both studies [6, 7], a total of 86 (59% of 145) were shared with the maturation/activation gene set (Figure 2b; Additional data file 2). Hierarchical clustering of the expression ratio of these 86 genes shows, however, that only a proportion of them are regulated by MCSF. Of the set of genes regulated by MCSF (data from [6–8]), clusters 1 and 2 contain genes whose levels correlate with the ratio in CD16+ versus CD16- monocytes (Figure 2b). These two clusters support the view that CD16+ monocytes are more responsive to MCSF. However, cluster 3 shows that not all genes induced by MCSF are higher in CD16+ monocytes than in CD16- monocytes. In addition, these genes seem to correlate with the pattern induced by interferon gamma (IFN-γ) and bacterial LPS when these cytokines are used to drive macrophages toward classical activation. The correlation between CD16+/CD16- profiles and those of inflammatory stimuli reinforces the suggestion that not only MCSF but also other factors may contribute to differences between monocyte subsets [2, 6, 7].
Many issues about monocyte subsets await clarification. The origin of the subsets and the basis and meaning of the fluctuations in their numbers in health and disease remain unexplained. The therapeutic potential of depleting specific subsets has been assessed in several studies, with the aim of reducing inflammation. Several therapies for autoinflammatory diseases decrease the numbers of CD16+ monocytes in the blood; for example, glucocorticoids reduce the number of CD16+ monocytes by 95% after 5 days. However their systemic effects are many, including potent immunosuppression, osteoporosis and hypertension . Monocyte depletion by apheresis using an ad sorptive Adacolumn seems more selective, and the results from early trials of this device in Crohn's disease are promising . Understanding the role and full potential of the monocyte subsets in the inflammatory response will be essential for creating novel and directed therapeutic approaches.
Additional data files
Additional data are provided with this article. Additional data file 1 lists the genes differentially expressed by CD16+ and CD16- monocytes that were detected in common by Zhao et al.  and Ancuta et al. Additional data file 2 lists genes differentially expressed by CD16+ and CD16- monocytes (from [6, 7]) in common with a dataset of genes expressed during monocyte maturation (3 days and 7 days after stimulation with MCSF) and activation by the classical (M1) and alternative (M2) pathway .
I thank Wong Siew Cheng and Petronela Ancuta for reviewing the manuscript and providing helpful suggestions, and Megan Kerr and Janet Digby for proofreading.
- Passlick B, Flieger D, Ziegler-Heitbrock HW: Identification and characterization of a novel monocyte subpopulation in human peripheral blood. Blood. 1989, 74: 2527-2534.PubMedGoogle Scholar
- Ziegler-Heitbrock L: The CD14+ CD16+ blood monocytes: their role in infection and inflammation. J Leukoc Biol. 2007, 81: 584-592. 10.1189/jlb.0806510.View ArticlePubMedGoogle Scholar
- Tacke F, Randolph GJ: Migratory fate and differentiation of blood monocyte subsets. Immunobiology. 2006, 211: 609-618. 10.1016/j.imbio.2006.05.025.View ArticlePubMedGoogle Scholar
- Hanai H, Iida T, Takeuchi K, Watanabe F, Yamada M, Kikuyama M, Maruyama Y, Iwaoka Y, Hirayama K, Nagata S, Takai K: Adsorptive depletion of elevated proinflammatory CD14+CD16+DR++ monocytes in patients with inflammatory bowel disease. Am J Gastroenterol. 2008, 103: 1210-1216. 10.1111/j.1572-0241.2007.01714.x.View ArticlePubMedGoogle Scholar
- Mobley JL, Leininger M, Madore S, Baginski TJ, Renkiewicz R: Genetic evidence of a functional monocyte dichotomy. Inflammation. 2007, 30: 189-197. 10.1007/s10753-007-9036-0.View ArticlePubMedGoogle Scholar
- Zhao C, Zhang H, Wong WC, Sem X, Han H, Ong SM, Tan YC, Yeap WH, Gan CS, Ng KQ, Koh MB, Kourilsky P, Sze SK, Wong SC: Identification of novel functional differences in monocyte subsets using proteomic and transcriptomic methods. J Proteome Res. 2009, 8: 4028-4038. 10.1021/pr900364p.View ArticlePubMedGoogle Scholar
- Ancuta P, Liu KY, Misra V, Wacleche VS, Gosselin A, Zhou X, Gabuzda D: Transcriptional profiling reveals developmental relationship and distinct biological functions of CD16+ and CD16- monocyte subsets. BMC Genomics. 2009, 10: 403-10.1186/1471-2164-10-403.PubMed CentralView ArticlePubMedGoogle Scholar
- Martinez FO, Gordon S, Locati M, Mantovani A: Transcriptional profiling of the human monocyte-to-macrophage differentiation and polarization: new molecules and patterns of gene expression. J Immunol. 2006, 177: 7303-7311.View ArticlePubMedGoogle Scholar
- Barnes M, Freudenberg J, Thompson S, Aronow B, Pavlidis P: Experimental comparison and cross-validation of the Affymetrix and Illumina gene expression analysis platforms. Nucleic Acids Res. 2005, 33: 5914-5923. 10.1093/nar/gki890.PubMed CentralView ArticlePubMedGoogle Scholar
- Dayyani F, Belge KU, Frankenberger M, Mack M, Berki T, Ziegler-Heitbrock L: Mechanism of glucocorticoid-induced depletion of human CD14+CD16+ monocytes. J Leukoc Biol. 2003, 74: 33-39. 10.1189/jlb.1202612.View ArticlePubMedGoogle Scholar