Osteocalcin does not influence acute or chronic inflammation in human vascular cells

Abstract Some human observational studies have suggested an anti‐inflammatory role of osteocalcin (OCN). An inflammatory protocol using interferon‐γ and tumor necrosis factor‐α (10 ng/ml) was employed to examine the acute (24 hr) and chronic (144 hr) effects of uncarboxylated OCN (ucOCN) in commercial, primary, subcultured human aortic endothelial cells (HAEC), and human smooth muscle cells (HASMCs). The inflammatory protocol increased phosphorylation of intracellular signaling proteins (CREB, JNK, p38, ERK, AKT, STAT3, STAT5) and increased secretion of adhesion markers (vascular cell adhesion molecule‐1, intracellular adhesion molecule‐1, monocyte chemoattractant protein‐1) and proinflammatory cytokines (interleukin‐6 [IL‐6], IL‐8). After acute inflammation, there were no additive or reductive effects of ucOCN in either cell type. Following chronic inflammation, ucOCN did not affect cell responses, nor did it appear to have any pro‐ or anti‐inflammatory effects when administered acutely or chronically on its own in either cell type. Additionally, ucOCN did not affect lipopolysaccharide (LPS)‐induced acute inflammation in HAECs or HASMCs. The findings of this study do not support a causal role for OCN within the models of vascular inflammation chosen. Further confirmatory studies are warranted.

carboxylated to varying degrees, allowing for undercarboxylated forms to be present in the circulation (one or two carboxylated residues), and uncarboxylated OCN (ucOCN; no carboxylated residues). ucOCN has previously been regarded as the "active form," and it has less affinity to bind to hydroxyapatite crystals in bone than cOCN.
The potential role for OCN in the development of atherosclerosis has been recently extensively reviewed (Tacey et al., 2018). The question remains however whether OCN is a pathological bystander or in fact mediator. To our knowledge, no in vitro studies to date have investigated the inflammatory role of OCN in human vascular cells. The aim of the current experiments was to investigate the distinct role of ucOCN in HAECs and human aortic smooth muscle cells (HASMCs) after acute and chronic administration, with and without stimulated inflammation. We hypothesized that ucOCN may reduce proinflammatory markers when administered on its own and may reduce the stimulated inflammatory responses of HAECs and HASMCs.

| Cell culture
HAECs and HASMCs were purchased from PromoCell (UK) and maintained at 37°C in a humidified incubator supplemented with 5% CO 2. Cells were cultured in endothelial cell growth media and smooth muscle cell growth media, respectively, containing 1% penicillin-streptomycin (Sigma-Aldrich, UK) and supplemental mix (PromoCell, UK). In all experiments, cells were used between passages 3 and 5. After experimental treatments, cell media was collected and cells were washed once with phosphate-buffered saline (PBS; pH 7.4; Gibco™). Radioimmunoprecipitation buffer (Sigma-Aldrich) with protease and phosphatase inhibitors (A32959; Thermo Fisher Scientific) was added to lyse the cells and the plates were shaken at 4°C for an hour. The cells were then collected and centrifuged at 14,000g for 5 min at 4°C and cell supernatants were frozen at −80°C.

| 3415
To test whether ucOCN could effect the production of endogenously induced inflammation, in a subset of experiments, lipopolysaccharide (LPS; 10 ng/ml) was added to HAECs and HASMCs for 24 hr with and without ucOCN (10 ng/ml). Media was collected after this time and stored at −80°C until analyzed.

| Total protein content
A bicinchoninic acid protein assay was performed to quantify the total protein content in the cell lysates collected at the end of the experiments (Smith et al., 1985).

| Lactate dehydrogenase activity assay
Lactate dehydrogenase activity (LDH) Colorimetric Assay kit (category no. ab102526; Abcam) was performed on cell media from the chronic inflammation experiment as per manufacturer's instructions.

| Haematoxylin and eosin staining
After 48 hr of the chronic inflammation protocol, a representative selection of HAECs and HASMCs was washed with PBS and fixed with ice cold methanol/acetone (50:50) for 10 min at room temperature. Fixed cells were then washed with PBS and stained with 0.1% Mayer's haematoxylin and counterstained with 1% eosin Y solution to allow visualization of the nuclei and cytoplasm.

| Statistical analysis
One-way analysis of variances (ANOVAs) were performed to assess differences in protein secretions and protein phosphorylation after 24 hr. Data were mostly normally distributed and nonparametric t tests (Mann-Whitney) were performed in a few cases were the data were not normalized. Results were normalized to protein content. Twoway ANOVAs were performed to detect differences in IL-6 secretion and LDH activity using time and treatment as factors for the chronic inflammation experiment. Multiple comparisons were adjusted for by Dunnett's statistical hypothesis test. All statistical analyses were performed using Prism 7 for Windows (Version 7.01; GraphPad Software Inc.). p Values were considered significant at p < .05.

| Human aortic endothelial cell responses to chronic OCN
In the chronic experiments, IL-6 secretion increased over time but was not affected by OCN (10 ng/ml) after 48, 96, or 144 hr incubation ( Figure 3a). Similarly, no effects of ucOCN were observed on LDH activity, IL-10 secretion, or total protein content (Figure 3c-f).

| Human aortic endothelial cell responses to chronic inflammation
Because we found no anti-inflammatory effects of ucOCN in the acute inflammatory protocol, we next established whether any effects become apparent after prolonged inflammation and ucOCN treatment.
IL-6 secretion increased over time with the inflammatory protocol, with a peak observed at 48 hr, but ucOCN demonstrated no additional affects (Figure 3b). Nicotinamide adenine dinucleotide H concentration was measured as an indicator of LDH activity. LDH activity increased sharply following 48 hr of inflammation in HAECs, following a similar trajectory as with IL-6 production ( Figure 4d). There were no differences detected between cells treated with ucOCN alongside inflammation and those treated with inflammation alone (Figure 3d). IL-10 secretion at 144 hr did not differ between treatment groups F I G U R E 1 Protein secretion responses in HAECs after acute inflammation. Mean % change relative to vehicle and SEM of secreted cell proteins VCAM-1, ICAM-1, IL-6, endothelin, IL-10, IL-8, and MCP-1 when treated with vehicle, ucOCN (10 ng/mL), an inflammatory protocol (IFN-γ and TNF-α, both 10 ng/ml, 8 hr followed by 16 hr, respectively), or inflammatory protocol and ucOCN. Data were normalized to total protein content. Data were analyzed by one-way ANOVA. **** indicates p < .001 compared with vehicle. HAEC, human aortic endothelial cell; ICAM-1, intracellular adhesion molecule-1; IL, interleukin; inflm, inflammatory protocol; MCP-1, monocyte chemoattractant protein-1; ucOCN, uncarboxylated osteocalcin; SEM, standard error of mean; VCAM-1, vascular cell adhesion molecule-1 ( Figure 3e). Total protein content (as an indicator of cell death) was decreased with the inflammatory protocol (p < .05) but no additional effect of ucOCN was observed (Figure 3f).
After 48 hr of inflammatory protocol, HAECs adopted an activated, spindle-shaped morphology as opposed to the characteristic cobblestone appearance of noninflamed cells ( Figure S2).
Phosphorylation of CREB, p38, ERK, and STAT3 was increased with inflammation but no additional effect of ucOCN was observed (Figure 5a-i).

| HASMC cell responses to chronic OCN
In the chronic experiments, IL-6 secretion increased over time but

| HASMC responses to chronic inflammation
Following chronic exposure to inflammation, IL-6 secretion increased in HASMCs until 96 hr and remained elevated (Figure 6b). Addition of ucOCN did not affect IL-6 levels. Similarly, LDH activity was increased with inflammation but was not affected by the coadministration of ucOCN (Figure 6d). IL-10 and total protein content were not altered by the inflammation protocol nor ucOCN (Figure 6e,f).

| OCN secretion is not detected following chronic inflammation
Finally, total OCN was measured in cell media after chronic inflammation (144 hr) to test whether inflammation induced OCN F I G U R E 3 IL-6 secretion, LDH activity, IL-10 secretion, and total protein content in HAECs after chronic inflammation with or without OCN. The effects of ucOCN on secretion of IL-6 without inflammation (a) and with inflammation (b) after 0, 48, 96, and 144 hr. The effect of ucOCN on LDH activity was measured by NADH concentration without inflammation (c) and with inflammation (d) after 0, 48, 96, and 144 hr. IL-10 secretion was measured after 144 hr (e). Total protein content was measured by BCA assay at the end of the experiment. ucOCN (10 ng/ml) was added every 48 hr. Total n = 9 from three experimental repeats. Data are given as means with error bars representing SEM. *denotes a significant difference compared with vehicle, analyzed by one-way ANOVA (**p < .01, ***p < .005). ANOVA, analysis of variance; HAEC, human aortic endothelial cell; IL, interleukin; IFN-γ, interferon-γ; inflm, inflammatory protocol (8 hr of IFN-γ followed by addition of TNF-α; both 5 ng/ml); LDH, lactate dehydrogenase; NADH, nicotinamide adenine dinucleotide H; SEM, standard error of mean; TNF-α, tumor necrosis factor-α; ucOCN, uncarboxylated osteocalcin F I G U R E 4 Protein secretions in HASMCs after acute inflammation. Mean % change relative to vehicle and SEM of secreted cell proteins ICAM-1, MCP-1, endothelin, MMP-3, IL-17, and IL-8 when treated with vehicle, ucOCN (10 ng/ml), an inflammatory protocol (IFN-γ and TNF-α, both 10 ng/ml, 8 hr followed by 16 hr, respectively), or inflammatory protocol and ucOCN. Data were normalized to total protein content. Data were analyzed by one-way ANOVA. ****denotes a significant difference compared with vehicle (p < .005). ANOVA, analysis of variance; HASMC, human aortic smooth muscle cell; ICAM-1, intracellular adhesion molecule-1; IL, interleukin; IFN-γ, interferon-γ; MCP-1, monocyte chemoattractant protein-1; MMP-3, matrix metalloproteinase; SEM, standard error of mean; TNF-α, tumor necrosis factor-α; ucOCN, uncarboxylated osteocalcin; inflm, inflammatory protocol secretion. OCN was not detectable in media from HAECs nor HASMCs at the end of the experiments (data not shown).

| OCN does not prevent LPS-induced inflammation
As ucOCN did not affect inflammation induced by exogenously added cytokines, we then tested whether LPS-induced inflammation elicited a different response. Acute LPS treatment (24 hr) significantly increased secretion of VCAM-1 and ICAM-1 in HAECs and HASMCs, respectively, compared with control (Figure 1c,d). In both HAECs and HASMCs, there were no differences in VCAM-1 or ICAM-1 secretion respectively between LPS treated and LPS + ucOCN-treated cells (Figure 1c,d).
Here, we demonstrate for the first time using two human vascular cell types that ucOCN does not regulate or alter the inflammatory responses of HAECs or HASMCs during acute inflammation. Our observational results presented here do not reflect an anti-inflammatory role of OCN suggested by some cross-sectional, human studies, at least within the experimental models used. ucOCN did not affect inflammatory cytokine production nor inflammatory signaling pathways in HASMCs and HAECs. This may be representative of the nonbinary nature of OCN which cannot be limited to having a positive or negative influence overall in vascular pathology and physiology in humans. It may transpire that OCN plays a protective role in the later stages of atherosclerosis and have more of an impact on the process of calcification.
As no overall anti-inflammatory effects of OCN were reported in the acute inflammatory protocol, we next established whether any effects of OCN become apparent after prolonged inflammation and OCN treatment. Pro-and anti-inflammatory cytokine production as well as LDH activity were not affected by chronic ucOCN treatment, alone or in combination with the inflammatory protocol. Lastly, we further showed that in another model of inflammation using LPS, F I G U R E 6 IL-6 secretion, LDH activity, IL-10 secretion, and total protein content in HASMCs after chronic inflammation with or without OCN. The effects of ucOCN on secretion of IL-6 without inflammation (a) and with inflammation (b) after 0, 48, 96, and 144 hr. The effect of ucOCN on LDH activity was measured by NADH concentration without inflammation (c) and with inflammation (d) after 0, 48, 96, and 144 hr. IL-10 secretion was measured after 144 hr (e). Total protein content was measured by BCA assay at the end of the experiment. ucOCN (10 ng/ml) was added every 48 hr. Total n = 9 from three experimental repeats. BCA, bicinchoninic acid; HASMC, human aortic smooth muscle cell; LDH, lactate dehydrogenase; IFN-γ, interferon-γ; IL, interleukin; inflm, inflammatory protocol (8 hr of IFN-γ followed by addition of TNF-α; both 5 ng/ml); NADH, nicotinamide adenine dinucleotide H; TNF-α, tumor necrosis factor-α; ucOCN, uncarboxylated osteocalcin ucOCN did not affect VCAM-1 or ICAM-1 secretion in HAECs and HASMCs respectively. An interesting observation from this study was the transition of HAECs after 48 hr of chronic inflammation to an activated morphology that is characteristic of endothelial cells under sheer stress and inflammation, representative of the early stages of atherosclerosis (Hunt & Jurd, 1998). This corresponds to the upregulation of activated endothelium markers VCAM-1 and ICAM-1 reported at 24 hr. Additionally we demonstrated that inflammation does not induce secretion of OCN itself.
The role of OCN in cardiovascular disease has had conflicting results in humans. Some longitudinal studies have reported U-shaped associations between OCN and cardiovascular-related or all-cause mortality (Kanazawa, Yamaguchi, & Sugimoto, 2011;Yang et al., 2017;Yeap et al., 2012). However, other studies have found no association between OCN and cardiovascular disease risk (Holvik et al., 2014;Hwang et al., 2015). A systematic review and meta-analysis on the association between OCN and markers of atherosclerosis or calcification in humans found no conclusive relationship (Millar, Patel, Anderson, England, & O'sullivan, 2017). Although a causal role of ucOCN in vascular inflammation was not identified in this study, further work should explore whether OCN may influence systemic inflammation via directly acting on immune cells.
Additionally, the exploration of the possible biological activity of other forms of circulating OCN should not be disregarded. We recognize that phenotypic drift occurs with the isolation and subculturing of cells and we acknowledge this as a limitation of the current study. Further work should explore both the effect of OCN on intact vessels and on cells from different vascular beds including resistance and conduit vessels.
In conclusion, in our models of vascular inflammation, OCN did not display a role in the direct inflammatory responses of primary subcultured HAECs or smooth muscle cells in either an acute or chronic setting.

This work was supported by the Biotechnology and Biological
Sciences Research Council (grant number: BB/M008770/1).

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.