NFκB Inhibition Mitigates Serum Amyloid A-Induced Pro-Atherogenic Responses in Endothelial Cells and Leukocyte Adhesion and Adverse Changes to Endothelium Function in Isolated Aorta

Abstract

The acute phase protein serum amyloid A (SAA) is associated with endothelial dysfunction and early-stage atherogenesis. Stimulation of vascular cells with SAA increases gene expression of pro-inflammation cytokines and tissue factor (TF). Activation of the transcription factor, nuclear factor kappa-B (NFκB), may be central to SAA-mediated endothelial cell inflammation, dysfunction and pro-thrombotic responses, while targeting NFκB with a pharmacologic inhibitor, BAY11-7082, may mitigate SAA activity. Human carotid artery endothelial cells (HCtAEC) were pre-incubated (1.5 h) with 10 μM BAY11-7082 or vehicle (control) followed by SAA (10 μg/mL; 4.5 h). Under these conditions gene expression for TF and Tumor Necrosis Factor (TNF) increased in SAA-treated HCtAEC and pre-treatment with BAY11-7082 significantly (TNF) and marginally (TF) reduced mRNA expression. Intracellular TNF and interleukin 6 (IL-6) protein also increased in HCtAEC supplemented with SAA and this expression was inhibited by BAY11-7082. Supplemented BAY11-7082 also significantly decreased SAA-mediated leukocyte adhesion to apolipoprotein E-deficient mouse aorta in ex vivo vascular flow studies. In vascular function studies, isolated aortic rings pre-treated with BAY11-7082 prior to incubation with SAA showed improved endothelium-dependent vasorelaxation and increased vascular cyclic guanosine monophosphate (cGMP) content. Together these data suggest that inhibition of NFκB activation may protect endothelial function by inhibiting the pro-inflammatory and pro-thrombotic activities of SAA.

Keywords: aorta; atherosclerosis; endothelium; nuclear; serum amyloid A; transcription.

Figure 1

Secretion of VEGF from HCtAEC in response to SAA in the absence and presence of the NFκB inhibitor BAY11-7082. HCtAEC were pre-incubated (1.5 h) with 0, 1, 10 or 100 μM BAY11-7082 or vehicle (control) then treated with 10 μg/mL SAA. After 4.5 h the levels of secretory VEGF in cell supernatants was assessed by commercial ELISA assays. Data were expressed as a fold-change compared to control cells. Data represents the mean ± SD (n = 4 individual experiments, each performed in duplicate; therefore, data represent 4 technical replicates). * p < 0.05 compared to control cells. All data was normalised to the corresponding level of cell confluence (expressed as a percentage %) determined using an IncuCyte system immediately prior to cell harvest for ELISA. This approach to normalizing data was necessary as there was some level of toxicity for BAY11-7082 particularly at the higher doses of inhibitor tested.

Figure 2

Relative gene expression in HCtAEC in response to SAA in the absence and presence BAY11-7082 under normoglycaemic conditions. HCtAEC were incubated with 10 μg/mL SAA or vehicle (control) with or without 10 μM BAY11-7082 pre-treatment under normoglycaemic conditions (5 mM glucose). Total RNA was isolated after 4.5 h and gene regulation was assessed. Representative gels and semi-quantitative analysis are shown for (A) TNF (B) TF. Gene expression was normalised against β-actin and results represent fold change compared to control cells. Data shows the mean ± SD (n = 3 individual experiments each completed in duplicate; therefore, data represent 3 technical replicates); * p < 0.05 compared to control, where possible, target gene products were confirmed by sequence analysis (refer to supplementary data Figures S1–S3).

Figure 3

BAY11-7082 reduces secretory IL-6 induced by SAA. HCtAEC were treated with SAA and BAY11-7082 as defined in methods. A commercial ELISA kit was used to quantify levels of secretory IL-6 in the cell supernatant. All data was normalised to the corresponding level of cell confluence (expressed as a percentage %) determined using an IncuCyte system immediately prior to cell harvest for ELISA (See Supplemental Table S1 and Figure S4). Data represent the mean ± SD (n = 6 samples, each completed in duplicate; therefore, data represent 3 technical replicates). * Different to the control; p < 0.0001. # Different to SAA-treated cells; p < 0.0001.

Figure 4

Changes in intracellular IL-6 and TNF in response to SAA and BAY-11-7082. HCtAEC were treated with SAA and with and without BAY11-7082 pre-incubation. Next, cellular TNF and IL-6 content and distribution was visualised by immunocytochemistry. For all images shown, nuclei were stained with DAPI (blue) or IL-6⁺ (A–C) or TNF⁺ (D–F) immune-fluorescence using an appropriate secondary antibody (red) under 20× of fluorescence microscope. Representative images are shown for cells treated with DMSO (vehicle control) (A,D), 10 μg/mL SAA (B,E) or pre-incubated with 10 μM BAY11-7082 (C,F) prior to addition of 10 μg/mL SAA. Data shown are representative of 5 fields and (n = 2 control, n = 2 SAA or n = 3 BAY11-7082 + SAA independent repeat studies). Insets to panels A–F show higher magnification images of representative single cells from the different treatment groups.

Figure 5

BAY11-7082 reduces ex vivo leukocyte adhesion induced by SAA in mice aorta. Mice were administered BAY11-7082 (5 mg/kg) or PBS (control) over a 14day period in parallel with administration of SAA (120 μg/kg; PBS as control) as described in the Methods Section. Aorta were subsequently isolated and subjected to ex vivo leukocyte adhesion studies using total labelled leukocytes. Arrows indicate leukocytes adhered to the aortic surface of mice treated with (A) SAA alone and (B) SAA + BAY11-7082. (C) Adherent leukocytes were quantified per field of view (FOV) by counting the number of dye-stained leukocytes that ceased motion during a 30 s recording at various time intervals under 160x magnification. Data represent mean ± SD. Independent aortic samples n = 4 SAA (time = 0; 2.5 min); n = 3 SAA (time = 5 min); n = 2 SAA (time = 7.5; 10 min); n = 6 SAA + BAY11-7082 (time = 0; 2.5; 5 min); n = 5 SAA + BAY11-7082 (time = 0 = 7.5; 10 min). * Different to SAA-only group ** p < 0.05 and *** p < 0.001. Data represent 3 technical and 3 biological replicates.

Figure 6

ERK1/2 levels in aortae homogenates. Total phosphorylated ERK1/2 (p-ERK1/2) protein was analysed in homogenates prepared from isolated mouse aortae (refer to Methods section) using the commercial ERK1/2 (pT202/Y204) SimpleStep ELISA™ Kit (Abcam, Cambridge, MA, USA) (Catalogue number: ab176640) as per manufacturers guidelines. Absorbance readings were recorded at 450 nm using an Infinite^® M200 PRO Plate reader (Tecan, Männedorf, Germany) and analysed using Microsoft Excel (2013, v7). Standard curves generated on the same plate were employed to express protein units in ug/mL. Homogenate samples tested were from control (n = 4), SAA (n = 2) and SAA + BAY11 (n = 4), each measured in duplicate. *** p < 0.0001 and ### p < 0.0001.

Figure 7

SAA endothelium-dependent vaso-relaxation is partially reversed by BAY11-7082. Rat aorta were isolated, cut into ring segments and mounted onto a force transducer. Rings were incubated with 10 μg SAA protein/mL (■), pre-treated with BAY11 (▲), or control (●) and then pre-constricted with phenylephrine as described in the Methods section. (A) Vascular relaxation was then assessed in response to the endothelium-dependent (acetylcholine [ACh]) or endothelium-independent (sodium nitroprusside [SNP], positive control (♦) stimuli. * Different to the rings treated with SAA in the absence of BAY11-7082; p < 0.01. (B) Levels of cGMP were determined with a commercial ELISA assay. Data represent mean ± SD (SAA: n = 5; BAY11: n = 5; Control: n = 4; SNP: n = 4) independent aortic samples. Different to the control endothelium-dependent relaxation; # p < 0.05. Different to rat aortic rings treated with SAA in the absence of Bay 11-7082; *** p < 0.001.