Role of Hic-5 in the formation of microvilli-like structures and the monocyte–endothelial interaction that accelerates atherosclerosis

Abstract

Aims

The adhesion of circulating monocytes to endothelial cells (ECs) is an early and critical event in the formation of atherosclerotic plaques. Hydrogen peroxide-inducible clone 5 (Hic-5) serves as an adaptor molecule in cell adhesion complexes. However, the role of endothelial Hic-5 in monocyte–EC interaction and atherogenesis remains unclear. We examined the roles of endothelial Hic-5 in monocyte–EC interaction and atherogenesis using mouse models of atherosclerosis and cultured human umbilical vein endothelial cells (HUVECs).

Methods and results

Hic-5 was expressed in ECs, but not in monocytes/macrophages. An ex vivo monocyte adhesion assay revealed that adhesion of THP-1 monocytes to aortas isolated from Apoe^−/− and LDLR^−/− mice stimulated by TNF-α or oxidized LDL was suppressed by Hic-5 deficiency. Scanning electron microscopic observations of aortas harvested from Apoe^−/− mice revealed that TNF-α- or oxidized LDL-induced microvilli-like structures were markedly suppressed by Hic-5 deficiency. Relative Hic-5 deficiency suppressed 60% of the atherosclerotic lesions in aortas from Apoe^−/− and LDLR^−/− mice. In contrast, overexpression of Hic-5 in HUVECs promoted induction of microvilli-like structures and adherence of THP-1 cells in an adhesion receptor such as intercellular adhesion molecule-1- and vascular cell adhesion molecule-1-dependent manner.

Conclusion

Hic-5 in ECs plays an important role in the formation of microvilli-like structures and in the interaction between ECs and monocytes, leading to monocyte recruitment and subsequent development of atherosclerosis.

Atherosclerosis, Microvilli-like structures, Endothelial cells

1. Introduction

The recruitment of monocytes to inflamed endothelium is a key event in the initiation of atherosclerosis.^1,2 The adhesion of monocytes to endothelial cells (ECs) involves a multistep process that is mediated by adhesion molecules and begins with rolling, which in turn is mediated by short-lived bonds between E-selectin expressed on the ECs and its ligands, such as P-selectin glycoprotein ligand-1, and monocytes.³ Monocytes are continuously arrested via interactions between their activated β1 and β2 integrins and vascular cell adhesion molecule-1 (VCAM-1) or intercellular adhesion molecule-1 (ICAM-1) on ECs.⁴ The arrested monocytes subsequently undergo transmigration through dynamic morphological changes and interact with other adhesion molecules.⁵ As described above, the interaction between adhesion molecules on monocytes and ECs may play a crucial role in the recruitment of monocytes to inflamed endothelium. However, studies using animal models of atherosclerosis reported that deficiency or mutation in adhesion molecules resulted in only a limited suppression of lesion formation.^6,7 These findings suggest a missing component required for the elucidation of the mechanism underlying monocyte–EC interaction.

Hydrogen peroxide-inducible clone 5 (Hic-5), a member of the paxillin family, was previously isolated as a gene induced by TGF-β or hydrogen peroxide.⁸ Hic-5 serves as the focal adhesion protein and adaptor for the recruitment of many signalling molecules, including focal adhesion kinase, talin, vinculin, C-terminal Src kinase, G protein-coupled receptor kinase interacting ArfGAP1, and protein tyrosine phosphatase PEST.^9,10 We previously demonstrated that Hic-5 was expressed in both ECs and vascular smooth muscle cells, but not in monocytes/macrophages, which are important cellular components involved in various types of vascular disorders.^10,11 In the present study, we investigated the effects of Hic-5 gene deletion on atherosclerotic lesion formation in mouse models of atherosclerosis. Our results demonstrate that Hic-5 in ECs plays a key role in the development of atherosclerosis through the formation of microvilli-like structures. These are tiny extensions of plasma membrane formed on both leucocytes and ECs, and are implicated in leucocyte recruitment by causing cellular adhesion and mechanotransduction.^12,13 Although microvilli-like structures on the leucocyte surface have been investigated extensively, few studies have focused on the role of microvilli-like structures on the EC surface in atherosclerosis.^14–16 This is the first study to demonstrate the involvement of microvilli-like structures on the EC surface in the development of atherosclerosis.

2. Methods

2.1 Animals

ApoE-deficient mice (Apoe^−/−: C57BL6) were obtained from Charles River Laboratories (Wilmington, MA, USA). LDL receptor-deficient (LDLR^−/−) mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). All mice were bred at the Center for Laboratory Animal Science at Showa University. We generated double-knockout mice by crossing Apoe^−/− or LDLR^−/− mice with previously generated Hic-5^−/− mice. Male mice were weaned at 12 weeks to a high-cholesterol diet containing 16.5% fat, 1.25% cholesterol, and 0.5% sodium cholate (F2HFD1; Oriental Yeast Co., Ltd, Tokyo, Japan), and were maintained on this diet for 10 weeks. Wild-type (WT) and Hic-5^−/− mice were maintained on a normal chow diet. WT C57BL/6J mice (Charles River Laboratories) were used as controls. All procedures involving experimental animals were approved by the Institutional Animal Care and Use Committee of Showa University, and complied with the Guide for the Care and Use of Laboratory Animals (7th and 8th edition, ILAR-NRC).

2.2 Scanning EM

Aortas were harvested from the mice, opened longitudinally, and pinned onto sterile agar. They were immediately incubated for 4 h with 1 × 10⁶ THP-1 monocytes in 1% fetal bovine serum (FBS)-DMEM with 10 ng/mL of TNF-α or 100 µg/mL of oxidized LDL, followed by immersion fixation with 2.5% glutaraldehyde for 24 h at 4°C. After washing in PBS, the specimens were stained with 1% buffered osmium tetroxide. They were then dehydrated by passing through a graded ethanol series, freeze-dried, coated with platinum, and observed with an S-4700 scanning electron microscope (SEM; Hitachi, Tokyo, Japan) at 25 kV. THP-1 cells on aortic ECs were photographed at magnifications of ×5000, ×10 000, and ×25 000. The numbers of microvilli-like structures per unit were based on their count at ×25 000 magnification.

2.3 Western blot analysis

Proteins were separated by sodium dodecyl sulfate–PAGE and blotted onto polyvinylidene difluoride membranes (GE Healthcare, Munich, Germany) blocked with 5% skim milk in PBS containing 0.1% Tween 20. Membranes were incubated with each primary antibody against VCAM-1 (Santa Cruz, Heidelberg, Germany), ICAM-1 (Santa Cruz), Hic-5 (BD Transduction Laboratories), or GAPDH (Millipore). Positive bands were visualized with a horseradish peroxidase-conjugated second antibody and Western Lightning chemiluminescence reagent (PerkinElmer, Boston, MA, USA), followed by exposure to X-ray film (RX-U; Fuji Film Co., Tokyo, Japan). The densities of the bands were measured using the Densitograph software (AE-6962FC, CS Analyzer ver2.0; ATTO, Tokyo, Japan).

2.4 Human aorta

Atherosclerotic lesions in the human aortic samples were collected from 10 autopsy cases (6 males and 4 females; age range, 37–85 years; see Supplementary material online, Table S2) and were graded by a pathologist. Expression of Hic-5 in the aortic segments was detected by conventional immunohistochemistry using an anti-Hic-5 antibody (Sigma). The study was approved by the Ethics Committee of the Kumamoto University School of Medicine.

2.5 Measurements of reactive oxygen species in HUVECs

Intracellular reactive oxygen species (ROS) generation was measured by fluorometric examination with dichlorofluorescein diacetate (H2DCFDA, Sigma). HUVECs were incubated with 10 ng/mL of TNF-α or 100 µg/mL of oxidized LDL, and then stained with 5 µM H2DCFDA for 30 min at 37°C. Images were captured using the image analysis software (Lumina Vision; Mitani Co., Fukui, Japan).

2.6 Isolation of mouse aortas and ex vivo monocyte adhesion assay

Mice were euthanized by cervical dislocation. Aortas were harvested from mice, opened longitudinally, and pinned onto sterile agar. Aortas were immediately incubated for 4 h with 1 × 10⁶ fluorescently labelled (using Calcein AM according to the manufacturer's instructions) THP-1 monocytes in 1% FBS-DMEM with 10 ng/mL of TNF-α or 100 µg/mL of oxidized LDL. THP-1 cells were maintained in DMEM containing 10% heat-inactivated FBS until use. After incubation, unbound THP-1 cells were rinsed away, and the number of THP-1 cells firmly bound to the aortas was counted in five fields using fluorescent microscopy.

2.7 Preparation of LDL

Human LDL (1.019–1.063 g/mL) was isolated from normal healthy human plasma (obtained from the Japanese Red Cross Society) by sequential ultracentrifugation. This part of the study was approved by the Ethics Committee of the Japanese Red Cross Society (approval ref.⁶). The isolation procedure conformed to the principles of the Declaration of Helsinki for the use of human tissues. Oxidized LDL was prepared by the incubation of isolated LDL at 0.1 mg/mL with 5 µmol/L of CuSO₄ for 20 h at 37°C.

2.8 siRNA transfection and adenovirus infection

Human umbilical vein endothelial cells (HUVECs) were bought from Lonza (USA) and cultured in six-well plates to 80% confluence. To down-regulate the expression of Hic-5, HUVECs were transfected for 24 h with siRNA (forward: 5′-GGA UCA UCU AUA CAG CAC-3′; reverse: 5′-CUC CUG CAA UAA ACC UAU A-3′ 10 nmol/L) or control siRNA using Lipofectamine 2000 (Invitrogen, CA, USA). To up-regulate the expression of Hic-5, HUVECs were infected for 24 h with a recombinant adenovirus encoding Hic-5 or β-galactosidase as a control.

2.9 Cell adhesion assay in HUVECs

HUVECs were incubated for 4 h with 1 × 10⁶ fluorescently labelled (using Calcein AM, according to the manufacturer's instructions) THP-1 monocytes in 1% FBS-DMEM with either TNF-α (10 ng/mL) or oxidized LDL (100 µg/mL). THP-1 cells were maintained in DMEM containing 10% heat-inactivated FBS until use. After incubation, unbound THP-1 cells were rinsed away, and the number of THP-1 cells that firmly bound to HUVECs was counted in five fields using fluorescent microscopy. In parallel experiments, HUVECs were treated with the following neutralizing antibodies: anti-ICAM-1 (R&D Systems, Minneapolis, MS, USA) and anti-VCAM-1 (Abcam).

2.10 Analysis of atherosclerotic lesions

Mice were euthanized by cervical dislocation. Aortas were opened longitudinally and fixed with 10% neutral buffered formalin for 24 h, followed by staining with Oil Red O. The Oil Red O-stained area relative to the whole surface area was calculated using the image analysis software (Lumina Vision; Mitani Co.). Hearts were excised and fixed in 4% formalin before being mounted in OCT medium and frozen at −20°C. Aortic sinus cross sections (10 µm) were stained with Oil Red O.

2.11 Immunohistochemistry

Macrophages were stained with MOMA-2, a rat anti-mouse macrophages/monocytes mAb (Serotec, Oxford, UK). Non-immune rat IgG was used as a negative control (Santa Cruz). A goat anti-rat IgG–HRP was used as the second antibody (Santa Cruz). Sections were developed with diaminobenzidine and counterstained with Mayer's haematoxylin. Images were captured using the image analysis software (Lumina Vision; Mitani Co.).

2.12 Statistical analysis

Data are presented as mean ± SEM. Statistical tests, including the t-test (for comparisons of parameters between the two groups) and a two-way ANOVA using the Bonferroni post hoc test (for comparisons of different parameters between two genotypes), were performed using the GraphPad Prism (version 5.0 for Mac) software (GraphPad Software, San Diego CA, USA). Values of P < 0.05 were considered statistically significant.

3. Results

3.1 Decreased monocyte–endothelial interaction in aortas harvested from Hic-5-deficient mice

To determine whether Hic-5 in ECs is involved in the formation of atherosclerotic lesions, we examined monocyte–EC interactions using aortas harvested from Apoe^−/−, LDLR^−/−, Apoe^−/−/Hic-5^−/−, and LDLR^−/−/Hic-5^−/− mice by SEM (Figure 1A) and an ex vivo monocyte adhesion assay (Figure 1C and D). SEM studies revealed that a number of monocytes with membrane ruffling, caused by adhesive stimuli, were attached to the surface of the aortas harvested from Apoe^−/− mice (Figure 1A, upper panels). In contrast, a limited number of monocytes without membrane ruffling were attached to the surface of the aortas harvested from Apoe^−/−/Hic-5^−/− mice (Figure 1A, lower panel). We previously reported that Hic-5 was expressed in mouse aortic ECs, but not in monocytes/macrophages.^9,10 Accordingly, the expression of Hic-5 was confirmed by western blotting in ECs, but not in THP-1 cells, a human monocytic cell line (Figure 1B). The numbers of fluorescein-labelled THP-1 cells that adhered to the surface of aortas harvested from Apoe^−/−/Hic-5^−/− mice (Figure 1C) and LDLR^−/−/Hic-5^−/− mice (Figure 1D) with ex vivo TNF-α stimulation were significantly lower than those from Apoe^−/− and LDLR^−/− mice. Similar results were obtained with oxidized LDL-stimulated aortas (Figure 1C and D), indicating that Hic-5 in ECs plays a key role in the monocyte–EC interaction.

Figure 1

Role of Hic-5 in monocyte–endothelial interaction in vivo and ex vivo. (A) Scanning electron microscopic images of aortas harvested from Apoe^−/− (upper panels) and Apoe^−/−/Hic-5^−/− (lower panels) mice (n = 7 in each group). (B) Western blotting showed the expression of Hic-5 in ECs but not in THP-1 monocytes. (C and D) Adherence of fluorescein-labelled THP-1 cells to harvested aortas from (C) Apoe^−/− and Apoe^−/−/Hic-5^−/− mice and (D) LDLR^−/− and LDLR^−/−/Hic-5^−/− mice. Isolated aortas were incubated with labelled THP-1 cells and stimulated ex vivo by TNF-α or oxidized LDL (n = 7 in each group). The number of adherent THP-1 cells was counted using fluorescence microscopy. All results are expressed as means ± SD of three independent experiments (*P < 0.01 and **P < 0.05). (E) Expression of Hic-5 on ECs in atherosclerotic lesions of the human aorta. Haematoxylin and eosin and immunohistochemical staining of Hic-5 in atherosclerosis-free (left) and severe atherosclerosis (right) patients. Scale bars indicate 200 µm. Arrowheads indicate Hic-5-positive ECs.

We then confirmed Hic-5 expression in the aortic sections derived from 10 patients with various grades of atherosclerotic lesions (see Supplementary material online, Table S2). Immunohistochemical staining revealed prominent Hic-5 expression in aortic ECs in the perilesional area of patients with advanced atherosclerosis (Figure 1E). Of note, Hic-5 expression was not observed in ECs located in the mature atherosclerotic lesion area.

3.2 Levels of Hic-5 expression parallel the numbers of microvilli-like structures on the EC surface

Next, we examined the mechanism responsible for the reduction in monocyte adhesion to ECs associated with Hic-5 deficiency. Expression of ICAM-1 and VCAM-1 in aortic ECs was enhanced by a high-cholesterol diet in Apoe^−/− mice; however, this process was not affected by Hic-5 deficiency (Figure 2A–C). In addition, Hic-5 knockdown in HUVECs did not affect TNF-α-induced monocyte chemoattractant protein-1 (MCP-1) expression (Figure 2D). Similarly, the secretion of MCP-1, IL-6, and VEGF from HUVECs after siRNA knockdown or adenovirus-mediated overexpression of Hic-5 did not differ from the secretion of those from control HUVECs (see Supplementary material online, Figure 1). We also investigated the role of Hic-5 in ROS production in ECs, as TNF-α is known to enhance ROS production and accelerate vascular inflammation. Both Hic-5 overexpression and suppression in HUVECs did not affect TNF-α-induced ROS production, and similar results were obtained with oxidized LDL-stimulated HUVECs (Figure 3A and B). On the other hand, SEM observation of the aortic endothelium revealed a striking morphological change; numerous microvilli-like structures emerged on the surface of aortas isolated from Apoe^−/− mice with ex vivo stimulation by TNF-α or oxidized LDL (Figure 4A and B). In contrast, the TNF-α- or oxidized LDL-induced formation of microvilli-like structures on the EC surface of aortas isolated from Apoe^−/−/Hic-5^−/− mice was dramatically suppressed when compared with the formation of the microvilli-like structures from Apoe^−/− mice (Figure 4A and B, P < 0.01). To determine whether Hic-5 is involved in the formation of microvilli-like structures on ECs, we stimulated HUVECs by TNF-α (Figure 4C and E) or oxidized LDL (Figure 4D and F), and examined the morphological changes after siRNA knockdown or adenovirus-mediated overexpression of Hic-5. The number of microvilli-like structures on the surface of HUVECs was significantly decreased by Hic-5 knockdown, but was markedly increased by Hic-5 overexpression (Figure 4C–F). Taken together, these results indicate that Hic-5 plays a key role in the formation of microvilli-like structures on the surface of ECs.

Figure 2

Effects of Hic-5 deletion or repression on the expression of ICAM-1, VCAM-1, or MCP-1 in ECs. (A and B) Western blotting for ICAM-1 and VCAM-1 in mouse ECs from Apoe^−/− and Apoe^−/−/Hic-5^−/− mice [HCD (high-cholesterol diet); n = 7 in each group]. (C) Representative ICAM-1 and VCAM-1 staining of aortic ECs from Apoe^−/− and Apoe^−/−/Hic-5^−/− mice. Scale bars indicate 50 µm. The graphs on the right show a densitometric analysis of ICAM-1 and VCAM-1 expression in ECs from Apoe^−/− and Apoe^−/−/Hic-5^−/− mice. (D) MCP-1 was induced by TNF-α in cultured HUVECs, but this process was not affected by Hic-5 knockdown. All results are expressed as means ± SD of three independent experiments.

Figure 3

Effect of Hic-5 on ROS production induced by TNF-α or oxidized LDL. HUVECs were transfected with siRNA or infected adenovirus and then stimulated with TNF-α (10 ng/mL) or oxidized LDL (100 µg/mL) for 4 h. (A) Images of representative fields viewed by a fluorescence microscope (n = 18 in each group). (B) Quantitation of intracellular ROS was determined by a fluorescence spectrophotometer. Scale bars indicate 50 µm. All results are expressed as means ± SD of three independent experiments.

Figure 4

Levels of Hic-5 expression parallel the numbers of microvilli-like structures on the EC surface. (A) Representative SEM images show microvilli-like structures on the aortic surface. Aortas were harvested from Apoe^−/− and Apoe^−/−/Hic-5^−/− mice and stimulated ex vivo by TNF-α or oxidized LDL (n = 7 in each group). (B) The numbers of microvilli-like structures on the surface of aortic ECs. Data are representative of three independent experiments (*P < 0.01). (C and D) Hic-5 in cultured HUVECs was knocked down by siRNA or overexpressed by adenovirus-mediated gene transfer. HUVECs were then stimulated by (C) TNF-α or (D) oxidized LDL for SEM observations. Scale bars indicate 5 µm. (E and F) The number of microvilli-like structures on the surface of HUVECs was counted after stimulation by (E) TNF-α or (F) oxidized LDL (n = 7 in each group). All results are expressed as means ± SD of three independent experiments (*P < 0.001 and **P < 0.05).

We also investigated whether the Hic-5-mediated formation of microvilli-like structures on ECs is essential for the monocyte–EC interaction. The monocyte adhesion assay using THP-1 cells and HUVECs stimulated by TNF-α or oxidized LDL revealed that the adherence of THP-1 cells to activated HUVECs was suppressed by Hic-5 knockdown and enhanced by adenovirus-mediated overexpression of Hic-5 in HUVECs (Figure 5A and B). Moreover, we determined whether Hic-5-mediated monocyte adhesion to the endothelium functionally depends on ICAM-1 and VCAM-1 membrane expression. HUVECs were pretreated with control IgG, anti-ICAM-1, or anti-VCAM-1 antibodies, followed by incubation with THP-1 cells. The preincubation of HUVECs with anti-ICAM-1 or anti-VCAM-1 antibodies resulted in a significant decrease in the monocyte–EC interaction on stimulation by TNF-α or oxidized LDL, despite the increase in microvilli-like structures following Hic-5 overexpression (Figure 5C and D).

Figure 5

Role of Hic-5 in the interaction between THP-1 cells and HUVECs. (A and B) HUVECs were incubated with fluorescein-labelled THP-1 cells and stimulated by TNF-α or oxidized LDL (n = 18 in each group). The numbers of adhered THP-1 cells were counted in five fields for each condition. Scale bars indicate 100 µm. (C and D) Effect of ICAM-1 and VCAM-1 neutralization on the adhesion of THP-1 cells to HUVECs. HUVECs were stimulated by (C) TNF-α or (D) oxidized LDL. All results are expressed as means ± SD of three independent experiments. *P < 0.01 and **P < 0.05, when compared with control IgG.

3.3 Hic-5 deficiency reduces the formation of atherosclerotic lesions and infiltration of macrophages

To evaluate the contribution of Hic-5 to atherogenesis, Apoe^−/− and Apoe^−/−/Hic-5^−/− mice were fed a high-cholesterol diet for 10 weeks, and their aortas were stained with Oil Red O. The surface areas of aortic lesions in Apoe^−/−/Hic-5^−/− mice were 60% smaller than those in Apoe^−/− mice. Moreover, aortic lesions in LDLR^−/−/Hic-5^−/− mice were also 60% smaller than those in LDLR^−/− mice (Figure 6A and B, P < 0.01).

Figure 6

Role of Hic-5 in the development of atherosclerosis in Apoe^−/− and LDLR^−/− mice. (A) Representative images depicting the en face analysis of Oil Red O staining. From left, representative results of WT (n = 12), Hic-5^−/− (n = 14), Apoe^−/− (n = 7), Apoe^−/−/Hic-5^−/− (n = 7), LDLR^−/− (n = 7), and LDLR^−/−/Hic-5^−/− (n = 7) mice after 10 weeks of a high-cholesterol diet. Scale bars indicate 100 µm. (B) Atherosclerotic lesions were evaluated by calculating the ratio of the total lesion area to the total surface area. (C) Representative results of aortic sinus plaque lesions in WT, Hic-5^−/−, Apoe^−/−, Apoe^−/−/Hic-5^−/−, LDLR^−/−, and LDLR^−/−/Hic-5^−/− mice after 10 weeks of a high-cholesterol diet. Plaques were stained with Oil Red O (n = 6 in each group). Scale bars indicate 100 µm. (D) Aortic sinus lesion areas were calculated as the ratio of the total plaque area to the total sinus area. (E) Representative images depicting sections of aortic sinus plaques from Apoe^−/− (n = 7) and Apoe^−/−/Hic-5^−/− (n = 7) mice, and (G) LDLR^−/− (n = 7) and LDLR^−/−/Hic-5^−/− (n = 7) mice after 10 weeks of a high-cholesterol diet. The plaques were stained with MOMA-2. Scale bars indicate 100 µm. (F and H) Macrophage-positive areas in the sections were calculated as the ratio to the total lesion area. All results are expressed as means ± SD of three independent experiments (*P < 0.01). (I) Schematic drawing of the possible roles of endothelial Hic-5 in the formation of microvilli-like structures on ECs, with subsequent EC–monocyte adhesion and membrane ruffling in monocytes.

We determined the aortic sinus lesion areas by calculating the ratio of the total plaque area to the total sinus area. Similar to the surface areas, sinus lesions were smaller in Apoe^−/−/Hic-5^−/− and LDLR^−/−/Hic-5^−/− mice compared with those in Apoe^−/− and LDLR^−/− mice (P < 0.05, Figure 6C and D). In addition, a quantitative analysis of lesions immunostained with MOMA-2 revealed that macrophage-positive areas were smaller in Apoe^−/−/Hic-5^−/− and LDLR^−/−/Hic-5^−/− mice compared with those in Apoe^−/− and LDLR^−/− mice (Figure 6E–H).

4. Discussion

Endothelial dysfunction is a key step in the initiation of atherosclerosis.¹⁶ Here, we describe endothelial Hic-5 as a novel player in atherosclerosis. Hic-5 plays a key role in the formation of microvilli-like structures on the surface of ECs (Figure 4), accelerating the monocyte–EC interaction (Figures 1 and 5), and serving as a novel regulator of atherosclerosis (Figure 6). As a result, Hic-5 deficiency suppresses atherosclerotic lesions in Apoe^−/− and LDLR^−/− mice (Figure 6) by suppressing the formation of microvilli-like structures on ECs (Figure 4) and the subsequent monocyte–EC interaction (Figure 1).

Monocytes transiently associated with ECs are easily detached from the ECs by the pressure from circulating blood flow. Thus, in addition to membrane protein–protein interactions, monocyte–EC interactions are also modulated by mechanical or haemodynamic factors. To overcome the detaching force, a three-dimensional docking structure on ECs in the form of microvilli-like structures increases the contact area between ECs and monocytes, thereby tightening monocyte capture by ECs even under blood flow. Nakashima et al.¹⁷ first reported the predominance of microvilli-like structures on ECs in sites prone to atherosclerosis. Immunoelectron microscopic observations revealed the homogeneous distribution of ICAM-1 over the surface of ECs, including microvilli-like structures.¹⁷ In the present study, neutralizing antibodies against ICAM-1 and VCAM-1 repressed monocyte–EC interactions, even though the number of microvilli-like structures on the surface of HUVECs was markedly increased by Hic-5 overexpression (Figure 5). These findings indicate that both microvilli-like structures and adhesion molecules such ICAM-1 and VCAM-1 on ECs are critical for monocyte–EC adhesion, and Hic-5-mediated monocyte adhesion to ECs seems to depend on adhesion molecules. Interestingly, very late Ag-4 (VLA-4; α4β1), which interacts with VCAM-1, has been reported to concentrate on microvilli in leucocytes.¹⁸ Similarly, interactions between lymphocyte function associated antigen-1 on leucocytes and ICAM-1 on EC are reportedly mediated by microvilli-like membrane projections.¹³ Therefore, microvilli on ECs and leucocytes may act as predominant sites of contact between each other under blood flow.

Furthermore, previous SEM studies have visualized microvillous structural alterations in the endothelium covering atherosclerotic plaques in human carotid arteries.¹⁶ To date, the molecular mechanism of the formation of microvilli-like structures on ECs has been unknown. However, the present study identified Hic-5 as the first known molecule involved in the formation of microvilli-like structures and atherogenicity; it mediates atherogenicity through morphological changes on the EC surface. Our study is also the first to demonstrate that the formation of microvilli-like structures on ECs is essential for the monocyte–EC interaction and the subsequent development of atherosclerosis. Considering our results regarding Hic-5 expression in the aortic sections derived from patients, Hic-5 may contribute to the initial stage of the development of atherosclerotic lesions. Hic-5 expression was induced in ECs located in the perilesional area, where macrophage accumulation was frequently observed, but not in ECs present in the mature atherosclerotic lesion area. Thus, future studies testing the repression of Hic-5 function in appropriate models may represent new therapeutic targets to suppress monocyte recruitment and the consequent progression of atherosclerosis. The present data derived from patients suggest that Hic-5 is highly expressed in ECs under pathological conditions (perilesional area of severe atherosclerotic patients), but weakly expressed in the cells under physiological conditions (atherosclerosis-free patients). The critical role of Hic-5 in atherosclerosis can be attributed to the enhanced expression of Hic-5 in ECs under pathological conditions.

In the current study involving ex vivo (Figure 1C and D) and in vitro models (Figure 5A and B), we first identified oxidized LDL, apart from TNF-α, as an inducer of microvilli-like structures on ECs.¹⁹ It is interesting to note that a major phospholipid component of oxidized LDL, lysophosphatidylcholine, is known to induce ICAM-1 and VCAM-1 in ECs.⁴ Therefore, oxidized LDL is likely to accelerate the monocyte–EC interaction synergistically by induction of microvilli-like structures and adhesion molecules.

As previously reported, one fundamental function of Hic-5 is to act as a scaffold to connect actin- and integrin-based adhesion complexes.²⁰ In addition, Hic-5 has been shown to promote the formation of invadopodia, which are actin-rich protrusive structures on the surface of cancer cells.²¹ Invadopodia look similar to microvilli-like structures, but are generally larger. A major role of invadopodia is cancer invasion, which is not the case for microvilli-like structures on ECs.²² We recently reported that Hic-5 is also involved in the formation of filopodia, which are actin-rich protrusive structures in platelets and essential for platelet aggregation. In addition, we found that the Rho inhibitor C3 transferase repressed the formation of Hic-5-mediated microvilli-like structures on the surface of HUVECs, and the number of actin stress fibre-positive HUVECs was significantly decreased by Hic-5 knockdown (see Supplementary material online, Figure 2). Considering that the forced expression of Hic-5 reportedly promotes the formation of ROCK-dependent actin stress fibres, whereas Hic-5 knockdown leads to the suppression of RhoA activation,²³ it is reasonable to assume that Hic-5 might control the organization of the cortical cytoskeleton of microvilli-like structures by regulating actin remodelling via Rho protein activation.

The results of the present study provide insights into three-dimensional structures located on the surface of ECs as a novel regulatory factor of the monocyte–EC interaction and the development of atherosclerosis, in addition to the expression levels of adhesion molecules in ECs. We are currently screening a compound that inhibits Hic-5 functions as a potential anti-inflammatory,¹¹ anti-platelet,²⁰ and anti-atherogenic agent.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Funding

This work was supported by Grants-in-Aid for Young Scientists (B) (26860588 to X.-F.L.) and Scientific Research (C) (26461149 to J.-r.K.-K. and 26461368 to A.M.) from the Japan Society for the Promotion of Science, Showa University Medical Foundation, Japan Heart Foundation, the Dr Hiroshi Irisawa and Aya Irisawa Memorial Research Grant, and a research grant from the Takeda Science Foundation. This work was also supported in part by the MEXT (Ministry of Education, Culture, Sports, Science and Technology)-Supported Program for the Strategic Research Foundation at Private Universities, 2012–16.

Acknowledgements

We sincerely thank Keitatsu Kou and Takenobu Nakagawa for their assistance with electron microscopy and immunohistochemical analysis, respectively.

Conflict of interest: none declared.