Contrasting in vitro vs. in vivo effects of a cell membrane-specific CC-chemokine binding protein on macrophage chemotaxis
- 1.1k Downloads
Chemokines (CK) provide directional cues that mediate the recruitment of leukocytes to sites of inflammation. Broad-spectrum blockade of the CC-CK family, using the vaccinia virus 35K protein, has been shown to cause a potent reduction of systemic inflammation in models of atherosclerosis, vein graft disease and arthritis. We have used a cell membrane-targeted form of 35K, Mem35K, to probe whether cell-associated blockade of chemokine response is sufficient to reduce cell recruitment in inflammation. In Tie2cre mice, activation of a flox-stop Mem35K transgene directed conditional expression of Mem35K in leukocytes and endothelial cells, confirmed by Western blotting, flow cytometry and immunofluorescence microscopy. This conditional Mem35K expression was sufficient to increase cell surface CCL5 binding and reduce chemotaxis in vitro to CCL5, CCL2 and CCL3 but not to non-CC-CK chemoattractants, LTB4, C5a or chemerin. However, in vivo monocyte recruitment into the peritoneum driven by zymosan or CC-chemokine injection, which was demonstrated to be CC-CK dependent using CCR2−/− mice, was not reduced by Mem35K expression, despite the expression of functional Mem35K protein. These findings highlight differing requirements for cell-associated anti-inflammatory activity in in vitro and in vivo models.
Mem35K is a cell-associated CC-chemokine binding protein.
Conditional Mem35K transgenic mice show expression Mem35K in leukocytes.
Mem35K blocks in vitro primary macrophage chemotaxis specifically towards CC-chemokines.
Mem35K expression is not sufficient to reduce inflammation in vivo.
The requirements for anti-inflammatory activity in vitro and in vivo are different.
KeywordsChemokine Macrophage Therapy Inflammation Transgenic
CC-Chemokines (CC-CK) are a large class of low molecular weight 8–12 kDa chemoattractant cytokines that predominantly mediate the migration of mononuclear cells . CC-CK ligate GPCR receptors, activating downstream signalling and leading to cellular activation, cytoskeletal reorganisation and migration . Studies in mice deficient in CC-CK or their receptors have demonstrated a pivotal role for CC-CK signalling in mediating cell recruitment in multiple immune and inflammatory models including atherosclerosis, experimental autoimmune encephalomyelitis (EAE), asthma and graft versus host disease [3, 4, 5, 6].
Viruses have evolved a portfolio of immune defence molecules that can block or modulate the mammalian chemokine system. Pox family viruses express a soluble 35-kDa protein, ‘35K’, that binds and inactivates almost all members of the CC-chemokine family . Analysis of the purified protein shows it binds all human and mouse CC-CK with high affinity . The crystal structure of 35K bound to CCL4 reveals that when the chemokine is bound, the residues required for receptor binding are obscured . Systemic treatment of animals with 35K via hepatic gene transfer or with a soluble Fc-fusion protein form of this molecule is able to reduce disease in models of atherosclerosis, fibrosis, arthritis and allergic airway inflammation [7, 9, 10, 11]. The potency of this chemokine-binding molecule makes it an attractive tool for determining the role of CC-CK in disease models and for probing the therapeutic potential of broad-spectrum CC-CK blockade in experimental disease models. In cardiovascular disease in particular, there is evidence that multiple CC-CK and their receptors act in concert to promote atherosclerosis, since CCR2−/−, CCR5−/− and CCL2−/− mice have all shown reduced atherosclerosis in hyperlipidemic models [4, 12, 13]. Indeed, the combined inhibition of both CCL2 and CCL5 activity led to additive inhibition of atherosclerosis . It is not yet clear whether the potency of multiple CC-CK blockade is indicative of functional redundancy within the CC-CK system or is a result of a complex pathology manifesting through multiple parallel highly cell-specific roles for individual CC-CKs . Indeed, small-molecule CC-CK inhibitors are already being tested in clinical trials as therapies for inflammatory diseases, but there is uncertainty as to whether local, cell-specific or systemic inhibition of CC-CK action will be necessary for clinical efficacy, whilst avoiding potentially harmful effects on systemic immune function. Thus, there remains a pressing need to understand how cell-specific effects of CC-CK inhibition will impact on inflammatory cell recruitment.
To test the potential of CC-CK inhibition as an anti-inflammatory strategy targeting individual cells, we have engineered a novel, cell-localised form of the 35K chemokine binding protein, Mem35K, that has a cell-restricted expression pattern through a FasL transmembrane domain, facilitating localization to membranes, and a green fluorescent protein (GFP) moiety for cellular identification . Cell transfection studies and adenoviral delivery to the liver demonstrated that high-level expression of this molecule is sufficient to reduce the chemotaxis of Mem35k-expressing cells and can reduce hepatic inflammation in the concavalin A model of hepatitis . However, adenoviral delivery of a cell-localised molecule is a fundamentally limited strategy, since the principal target of gene transfer techniques such as adenoviral delivery is the liver. Whilst this is ideal for the production of secreted molecules, or effects mediated by high-level hepatocyte transgene expression, it does not allow the targeting of specific cell types such as leukocytes. Furthermore, the important question of whether CC-CK inhibition can be achieved at the level of individual inflammatory cell populations, as distinct from whole-organ or systemic effects, remains unknown.
To address these questions, we now report the development of a system for cell- or tissue-restricted expression of Mem35K, using a conditional transgenic mouse for expression of Mem35K under control of a CMV flox-stop promoter. To evaluate the transgenic cell-localised expression of Mem35K, we crossed these mice with transgenic mice expressing cre under control of the Tie2 promoter, in order to direct transgenic Mem35 expression to leukocytes.
Materials and methods
Additional methods can be found in the Supplementary Material.
Generation of the Mem35K transgenic mouse
The Mem35K mouse was commercially produced by genOway (Lyon, France). The Mem35K cDNA was introduced into the Hprt locus using the ‘Quick Knock-in’ targeting vector containing the CCAG promoter and a validated floxed STOP cassette  and the human HPRT allele to reconstitute the Hprt gene. The targeting cassette was linearised, isolated and purified prior to electroporation into E14Tg2a ES cells derived from 129P2/Ola mice. Positive selection was achieved by identification of HAT-resistant clones. Southern blotting identified 9 ES cell clones that were correctly targeted. The recombined ES cells were injected into blastocysts from pseudopregnant C57bl/6J mice. Chimeric male offspring with 80–100 % chimerism were selected for breeding to confirm germline transmission. Two founder 80 % chimeric males demonstrated germline transmission and produced 8 female Mem35K heterozygous mice.
Mem35K elicits GFP fluorescence, membrane-localised 35K protein and reduces CC-chemokine receptor-mediated chemotaxis
Transgenic expression of Mem35K
Functional transgenic expression of Mem35K detected in vitro
To further investigate the expression and localization of Mem35K protein in cells from Mem35Kflox Tie2cre mice, we performed fluorescence confocal microscopy on primary macrophages ex vivo, which had demonstrated Mem35K protein expression by Western blotting. Endogenous GFP fluorescence was detected in Biogel-elicited peritoneal macrophages from mice expressing Mem35K, by both significantly increased green fluorescence by flow cytometry (Fig. 2e) and confocal microscopy (2F), although expression was not sufficiently high to show clear localization of the transgene within the cell. The localization of mem35K to the cell surface was confirmed by detection of the HA epitope on the cell surface by flow cytometry (2F). The presence of active mem35K on the cell surface was confirmed by detection of significantly increased biotinylated CCL5 binding to the cell surface (Fig. 2f). Mem35Kflox Tie2cre mice yielded macrophages with a greater binding capacity for CCL5. Furthermore, we confirmed that this effect was not due to alterations in cell surface chemokine receptor expression by detecting no significant difference in CCR5 expression on the cell surface using an anti-CCR5 antibody (Fig. 2f). In primary endothelial cells from Mem35Kflox Tie2cre mice, identified using anti-CD31 immunohistochemistry, expression of the Mem35K molecule was revealed by anti-GFP immunohistochemistry, but only in Mem35Kflox Tie2cre mice that carried both the Mem35K and Tie2cre transgenes (Fig. 2f). Endothelial expression of Mem35K was also demonstrated using anti-GFP immunohistochemistry in the aorta from Mem35Kflox Tie2cre mice, compared to non-expressing controls (Fig. 2f).
Functional expression of Mem35K in vivo
Leukocyte counts from blood, spleen and bone marrow in Mem35Kflox and Mem35Kflox/Tie2cre animals were obtained by flow cytometry
Blood (cells × 106/ml)
Spleen (cells × 106)
Bone marrow (cells × 106)
Because 100 μg zymosan generates a potent sterile inflammation that takes 96 h to resolve, we next investigated a less potent model of peritonitis in order to determine whether the lack of effect of Mem35K was due to the production of inflammatory cytokines overwhelming the inhibitory capacity of cell-associated Mem35K. Mice were injected with 10 μg of zymosan for 4 h to evaluate at early leukocyte recruitment (Fig. 4d). Injection of zymosan caused recruitment of neutrophils and monocytes after 4 h in both Mem35K-expressing and non-expressing controls. However, no significant effect on the number of neutrophils or monocytes recruited was observed. Again the CC-CK dependence of this model was confirmed using CCR2−/− mice, which demonstrated a significant decrease in monocyte recruitment in this model (Fig. 4e and Supplementary Figure 1).
To directly assess whether Mem35K can inhibit specific CC-chemokines in vivo, we injected CCL2 i.p., which has previously been shown to cause monocyte recruitment to the peritoneal cavity . Injection of 2 μg CCL2 caused a significant recruitment of both monocytes and neutrophils after 4 h, but no effect of Mem35K expression was seen on the recruitment of either population (Fig. 4f). Although CCL2 would be assumed to cause recruitment via the receptor CCR2, we confirmed no confounding effect of the injection procedure on monocyte recruitment and confirmed the nature of the recombinant protein injected by performing a parallel experiment in CCR2−/− mice. We demonstrated no recruitment of monocytes to CCL2 in CCR2−/− mice, confirming this model is working via CC-CK as expected; however, the migration of neutrophils was unaffected by the loss of CCR2 indicating they are migrating in response to other substances in the recombinant protein preparation (Fig. 1g and Supplementary Figure 1).
These studies provide new and important insights to the application of CC-CK inhibition as a cell-specific therapeutic strategy in inflammatory diseases. These new insights were gained through cell-localised expression of a broad-spectrum CC-CK inhibitor by leukocytes and endothelial cells in transgenic mice, using the flox-stop conditional expression system. Transgenic expression of the Mem35K CC-CK inhibitor protein was readily detectable and sufficient to significantly inhibit leukocyte recruitment in in vitro models of chemotaxis. However, this cell-localised expression of the CC-CK inhibitor protein did not reduce leukocyte chemotaxis in an in vivo model of inflammation. These observations suggest that cell-restricted CC-CK inhibitors will not be sufficient to limit chemotaxis of inflammatory cells in vivo and that more systemic approaches to limit CC-CK action will be required.
Previous studies with a soluble, secreted form of 35K have shown potent effects on inflammatory cell recruitment [9, 22, 23, 24]. Furthermore, high-level viral-mediated expression of Mem35K in the liver is able to reduce hepatic inflammation induced by Concavalin A , demonstrating that the cell-restricted expression of Mem35K can also suppress the migration of leukocytes . However, whether this effect of Mem35K was achieved through competition for CC-CK binding at the surface of individual cells or through local sequestration of CC-CK in tissue could not be determined, as in vivo studies were limited to un-targeted local expression through adenoviral gene delivery, rather than through cell-targeted expression in the responder population of leukocytes. We have now achieved cell-targeted expression of Mem35K in leukocytes rather than hepatocytes that are the predominant target of adenoviral gene delivery. In transgenic leukocytes ex vivo, we again demonstrated that expression of the Mem35K is sufficient to reduce chemotaxis. In contrast, this reduced sensitivity to chemokine was not sufficient to alter leukocyte recruitment in a model of peritonitis in vivo. This finding suggests that the ability of a cell-targeted CC-CK inhibitor to exert an anti-inflammatory effect in vivo is more reliant on the ability of the molecule to sequester chemokines in the tissue, rather than by exerting a selective and cell-specific inhibition of chemokine receptor signalling at the surface of individual leukocytes. Such a cell-selective inhibition of chemokine signalling by Mem35K may require higher expression—it is of note that the migration of Mem35K-transfected cells towards CC-CK is almost completely ablated in our previous study, but is only blunted in our transgenic leukocytes . Alternatively, cells expressing Mem35K in vitro may show a more potent inhibition of CC-CK as they are ‘produced’ in a system that lacks CC-CK, therefore enabling the full complement of Mem35K to bind exogenously applied CC-CK, in contrast to cells in vivo, where cumulative saturation of Mem35K by CC-CK may ultimately diminish any functional impact on inflammatory cell recruitment.
Analysis of the structure of 35K bound to CCL4 reveals that 35K binds the residues responsible for CC-CK receptor binding. As a result, chemokines bound to 35K are unable to activate downstream signalling. We have previously shown that soluble 35K protein present in plasma binds chemokine and renders it inactive by demonstration that whilst chemokine levels in plasma are not decreased, or even are increased, in 35K expressing animals, that the chemokine mediated biological activity of those fluids is reduced [9, 24].
Clear evidence that the expression of a competitor for CC-CK binding is sufficient to regulate inflammation is provided by the D6 decoy chemokine receptor that regulates the resolution of inflammation by scavenging chemokines. Whilst D6−/− mice do not show any obvious phenotype in the absence of an inflammatory challenge, these mice show a significant increase in concentration of its ligands in plasma and inflamed tissue and the development of a more severe inflammatory response [25, 26]. This action of D6 demonstrates that expression of a cell-surface CC-CK binding molecule is a mechanism by which inflammation can be regulated in vivo. Given this is the case, then it may follow that Mem35K, in addition to the level of expression that can be achieved by a single copy transgene, may lack other features of its regulation that are required to act in this fashion. D6 undergoes rapid recycling between the plasma membrane to the endosome; however, in contrast to chemokine receptors that also internalise following ligand binding, D6 surface expression is not downregulated by prolonged exposure to its ligand . Indeed, there is little evidence that D6 activity is regulated by expression, rather that changes in the rate of cell surface recycling allows increased activity during inflammation . Mem35K lacks an intracellular signalling domain, so is unable to undergo the controlled endocytic process used by D6. Instead, membrane expression can only be controlled by gene expression, using a CMV-driven promoter that is not highly regulated by inflammation . The successful expression by adenovirus of this protein may be in part due to the high level of expression that is present acutely following infection. The fate of CC-CK once bound to 35K or Mem35K is unknown, so whether the binding capacity of each individual molecule is rapidly exhausted is key to understanding the utility of this molecule in therapy. If each molecule can only bind a single CC-CK molecule (as suggested by the co-crystal structure with CCL4), then a key aspect of this therapy is in sustaining a continuous high level of ‘free’ Mem35K. The binding capacity of Mem35K is not exhausted during the development and circulation of the transgenic leukocytes as the primary Biogel-elicited cells still demonstrate an in vitro phenotype. The Mem35Kflox/Tie2cre mouse also exhibits expression in endothelial cells, and despite this additional expression, which might be expected to exert a bystander effect on active chemokine availability, no in vivo reduction of inflammation is seen. Although we have not probed the functionality of the endothelial cell-expressed Mem35K in detail, we would hypothesise that the limitations of the level and non-regulated manner of expression would apply as they do to the leukocyte-expressed molecule.
Therapies that act by competition can be an effective strategy, since anti-TNF targets the soluble component to compete with TNF action at a systemic level . However, it is likely that cell-targeted therapies must be potent and regulated by inflammation, if they are a competitive inhibitor, or block the cell-autonomous part of the pathway such as the receptor. Furthermore, such a strategy is likely to require a sustained, high level of unbound inhibitor in order to exert a functional effect (as exemplified by viral delivery systems), but as a purified protein may be expensive and require frequent dosing.
In conclusion, expression of the cell-associated CC-CK inhibitor, Mem35K, by leukocytes effectively inhibits CC-CK action in vitro, but this cell-associated inhibition of CC-CK is not effective in blockade of leukocyte recruitment in vivo. Strategies to reduce inflammation by CC-CK binding will require high-level CC-CK binding to reduce CC-CK action in the whole tissue or systemically, making cell-targeted CC-CK inhibition strategies less likely to be effective as a therapeutic strategy.
This work was supported by the British Heart Foundation (RG/10/15/28578 and RG/12/5/29576), Wellcome Trust (091504/Z/10/Z and 090532/Z/09/Z) and the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre. We would like to the thank Dr A. Vincent for the access to CCR2−/− mice.
Experimental concept and design were done by EMcN, KMC and DRG; EMcN, AJI, JP, GEW and DR-K collected and interpreted the data; EMcN and KMC prepared the manuscript.
Conflict of interest
The authors have no conflict of interest to declare.
- 13.Combadiere C, Potteaux S, Rodero M, Simon T, Pezard A, Esposito B, Merval R, Proudfoot A, Tedgui A, Mallat Z (2008) Combined inhibition of CCL2, CX3CR1, and CCR5 abrogates Ly6C(hi) and Ly6C(lo) monocytosis and almost abolishes atherosclerosis in hypercholesterolemic mice. Circulation 117:1649–1657PubMedCrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.