Effect of cobalt-mediated Toll-like receptor 4 activation on inflammatory responses in endothelial cells

Cobalt-containing metal-on-metal hip replacements are associated with adverse reactions to metal debris (ARMD), including inflammatory pseudotumours, osteolysis, and aseptic implant loosening. The exact cellular and molecular mechanisms leading to these responses are unknown. Cobaltions (Co2+) activate human Toll-like receptor 4 (TLR4), an innate immune receptor responsible for inflammatory responses to Gram negative bacterial lipopolysaccharide (LPS). We investigated the effect of Co2+-mediated TLR4 activation on human microvascular endothelial cells (HMEC-1), focusing on the secretion of key inflammatory cytokines and expression of adhesion molecules. We also studied the role of TLR4 in Co2+-mediated adhesion molecule expression in MonoMac 6 macrophages. We show that Co2+ increases secretion of inflammatory cytokines, including IL-6 and IL-8, in HMEC-1. The effects are TLR4-dependent as they can be prevented with a small molecule TLR4 antagonist. Increased TLR4-dependent expression of intercellular adhesion molecule 1 (ICAM1) was also observed in endothelial cells and macrophages. Furthermore, we demonstrate for the first time that Co2+ activation of TLR4 upregulates secretion of a soluble adhesion molecule, sICAM-1, in both endothelial cells and macrophages. Although sICAM-1 can be generated through activity of matrix metalloproteinase-9 (MMP-9), we did not find any changes in MMP9 expression following Co2+ stimulation. In summary we show that Co2+ can induce endothelial inflammation via activation of TLR4. We also identify a role for TLR4 in Co2+-mediated changes in adhesion molecule expression. Finally, sICAM-1 is a novel target for further investigation in ARMD studies.


INTRODUCTION
Metal-on-metal (MoM) hip replacements are associated with the development of adverse reactions to metal debris (ARMD), which includes inflammatory pseudotumours, soft tissue necrosis, osteolysis and resulting aseptic implant loosening. Peri-implant tissues are often infiltrated by monocytes, macrophages and lymphocytes (referred to as aseptic lymphocyte-dominated vasculitis-associated lesion, ALVAL) which is indicative of an inflammatory response. However the cellular and molecular mechanisms that underlie ARMD are not wellunderstood.
Co 2+ from MoM implants activates human Tolllike receptor 4 (TLR4) [1][2][3], an innate immune receptor expressed on immune cells as well as endothelial and epithelial cells. The major ligand for TLR4 is lipopolysaccharide from Gram negative bacteria, and

Research Paper: Immunology
Oncotarget 76472 www.impactjournals.com/oncotarget receptor activation causes adaptor protein recruitment and an intracellular signalling cascade that upregulates the activity of transcription factors including NFκB [3].
We have previously shown that activation of TLR4 by Co 2+ increases the secretion of inflammatory cytokines, including interleukin-8 (IL-8) and chemokine (C-X-C motif) ligand 10 (CXCL10), in MonoMac 6 macrophages [4]. Previous studies investigating the inflammatory effects of Co 2+ in endothelial cells have primarily focused on endothelial cells transfected with TLR4 and its coreceptor MD2 [3,5], but few studies have investigated the effect of Co 2+ on endogenous TLR4-expressing endothelial cell lines. Endothelial cells are exposed to Co 2+ present in the blood of MoM hip replacement patients [6] and therefore understanding the cellular response is important in defining the causes of ARMD and identifying potential therapeutic targets for ARMD prevention.
In the present study we assessed the immune response of endothelial cells to Co 2+ , with a focus on the role of TLR4. We also investigated the effect of Co 2+ on adhesion molecule expression by endothelial cells and macrophages because of their critical role in inflammatory process such as leukocyte binding and extravasation.
To assess the role of TLR4 in the observed cytokine secretion, HMEC-1 were pre-incubated with 1μg/ml CLI-095 (a small molecule TLR4 antagonist) for 6h followed by stimulation with 0.75mM Co 2+ or 100ng/ ml LPS for 24h. IL-8 and IL-6 secretion were measured by ELISA. Pre-treatment with CLI-095 significantly HMEC-1 were pre-treated with 1μg/ ml CLI-095 followed by 24h stimulation with 0.75mM Co 2+ or 100ng/ml LPS. C. IL-8 and D. IL-6 secretion was quantified by ELISA. All data is representative of three independent experiments.
Oncotarget 76473 www.impactjournals.com/oncotarget decreased secretion of both cytokines in response to Co 2+ (p < 0.001), showing that their release is TLR4dependent. The cytokine release was not a result of Co 2+mediated cytotoxicity as trypan blue staining revealed no change in HMEC-1 viability following cobalt stimulation (Supplementary Material, Figure 6).

Co 2+ -mediated TLR4 activation increases ICAM1 expression in endothelial cells and macrophages
Endothelial cells are known to express adhesion molecules, including intercellular adhesion molecule-1  1μg/ml CLI-095 for 6h prior to 24h stimulation with 0.75mM Co 2+ or 100ng/ml LPS. RNA was isolated and cDNA synthesised by reverse transcription. ICAM1 expression was quantified by qRT-PCR. Data is representative of three independent experiments.
Co 2+ induced a small but significant 3-fold upregulation in ICAM1 expression by HMEC-1 (p = 0.013) ( Figure 2A) and a larger 35-fold upregulation in MonoMac 6 cells (p < 0.001) ( Figure 2B). In both cell lines the response was found to be TLR4-dependent because it was inhibited by the TLR4 antagonist CLI-095 (both p < 0.001).
HMEC-1 exhibited a significant 16-fold increase in MMP9 expression following stimulation with 100ng/ ml LPS (p < 0.001) ( Figure 4A). This was inhibited by CLI-095, showing that it is a TLR4-dependent effect (p < 0.001). In contrast there was no change in MMP9 expression in response to Co 2+ (p = 0.999) ( Figure 4A). A similar pattern was observed in MonoMac 6 cells; following LPS stimulation there was a 7-fold increase in MMP9 expression by (p < 0.001) ( Figure 4B). CLI-095 inhibited this upregulated expression, showing that it is TLR4-dependent. However there was no increase in MMP9 expression in response to Co 2+ , although CLI-095 decreased its expression further.

DISCUSSION
In the present study we describe a TLR4-dependent inflammatory response to Co 2+ in human endothelial cells Oncotarget 76475 www.impactjournals.com/oncotarget and macrophages. HMEC-1 exhibited significant increases in secretion of inflammatory cytokines IL-8 and IL-6 when stimulated with Co 2+ . This was inhibited by the TLR4 antagonist CLI-095, showing that the receptor is central to the responses. Previous studies have shown that Co 2+ upregulates adhesion molecule expression [8][9][10], but have not demonstrated the exact signalling pathways involved. The data obtained in this study supports the findings of these studies and also indicates a previously unidentified role for TLR4 in Co 2+ -mediated ICAM1 expression in both endothelial cells and macrophages. Furthermore, for the first time a soluble adhesion molecule, sICAM-1, was detected in conditioned media from Co 2+ and LPSstimulated HMEC-1 and MonoMac 6 cells. CLI-095 inhibited sICAM-1 changes and consequently they are TLR4-dependent.
We investigated the effect of Co 2+ on MMP9 expression because MMP-9 can cleave membranebound ICAM-1 resulting in the release of its soluble form, sICAM-1. In addition, MMP-9 can be regulated by LPS activation of TLR4 [11] and therefore it is possible that Co 2+ -mediated TLR4 activation results in MMP-9 activity and sICAM-1 generation. However, although LPS increased MMP9 expression in a TLR4-dependent manner, there was no change in expression in response to Co 2+ . The absence of any effect was consistent between HMEC-1 and MonoMac 6 cells. The lack of change in MMP9 expression following Co 2+ stimulation suggests that the enzyme is not responsible for the changes in sICAM-1 secretion observed in response to Co 2+ . Other proteolytic enzymes potentially involved in sICAM-1 cleavage include serine proteases [12], neutrophil elastase [13], and cathepsin G [14]. However the effect of Co 2+ on these factors remains to be elucidated.
sICAM-1 has previously been proposed as a marker of inflammation [15] that is cleaved to regulate inflammatory responses but studies are now reporting a broader role for sICAM-1, including promotion of angiogenesis and neovascularisation [16]. This is of particular interest to the present study because blood vessel Oncotarget 76476 www.impactjournals.com/oncotarget formation is required for pseudotumour development, which is a major factor in ARMD. Soft tissue necrosis is also a common feature of ARMD and can result from vascular inflammation restricting oxygen supply to the tissues. The ability of Co 2+ to cause an inflammatory response, including pro-inflammatory cytokine release, in endothelial cells may indicate that similar effects occur in vitro, which could result in ischaemia and subsequent tissue death.
A limitation of the present study is the high Co 2+ concentrations that we have used to stimulate the cells. Even the concentrations at the lower end of the range are considerably higher than those detected in the serum and synovial fluid of patients with failed MoM implants [17][18][19]. However, the Co 2+ concentrations used in our study are in line with those of similar in vitro studies of the inflammatory effects of metal ions [3,10,20,21]. Hence, they are appropriate and relevant for this study.
A working model of the possible mechanisms indicated by our results is shown in Figure 5. In summary, we have shown that Co 2+ has the potential to induce an inflammatory response in the endothelium through activation of TLR4. It also shows for the first time that Co 2+ increases sICAM-1 secretion in a TLR4-dependent manner. Although the exact mechanism of its release remains unclear, sICAM-1 is an interesting target for further investigation in ARMD because of its previously described roles in angiogenesis, neovascularisation and tumour formation [16].
MonoMac 6 cells are a human TLR4-expressing cell line derived from acute monocytic leukaemia. Cells were cultured as previously described [22].

Cell stimulation
Cells were stimulated with cobalt chloride hexahydrate (referred to as Co 2+ in this study) in complete culture medium appropriate for each cell line. Complete culture medium was used as a negative control while 100ng/ml TLR4-specific LPS (Alexis Biochemicals, San Diego, USA) provided a positive control.

qRT-PCR
Gene expression changes were assessed by quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) using TaqMan primers and probes (ThermoFisher Scientific, Massachusetts, USA). RNA was isolated using a Qiagen RNeasy Mini kit (Qiagen, Venlo, Netherlands) and cDNA synthesised using Superscript III reverse transcriptase (ThermoFisher Scientific). Each qRT-PCR reaction contained 5μl TaqMan Gene Expression Mastermix (ThermoFisher Scientific), 2μl diluted cDNA template, 2.5μl nuclease-free H 2 O and 0.5μl TaqMan Gene Expression Assay (ThermoFisher Scientific). No template controls with nuclease-free H 2 O instead of cDNA were included. All reactions were performed in triplicate and target gene expression was normalised to GAPDH expression.

CLI-095
Inhibition of TLR4 was performed by pre-incubating cells for 6h with 1μg/ml CLI-095. CLI-095 (Invivogen, UK) is a small molecule TLR4 antagonist that binds to the intracellular domain of the receptor and prevents recruitment of downstream adaptor proteins.

Cytotoxicity assay
Cytotoxicity was assessed by trypan blue staining. Stimulated cells were resuspended in a small volume of supernatant and 10μl cell suspension was mixed with 10μl trypan blue dye. Staining was visualised on a Luna II automated cell counter (Logos Biosystems, Virginia, USA)

Statistical analysis
Statistical significance was calculated using a oneway analysis of variance (ANOVA). When samples were compared to an untreated control ( Figures 1A, 1B, 3A, and 3B), Dunnett's test for multiple comparisons was performed. When comparing all samples to each other, Tukey's test for multiple comparisons was performed.