Targeting cell adhesion molecules with nanoparticles using in vivo and flow-based in vitro models of atherosclerosis

Abstract

Atherosclerosis is a leading cause of death worldwide; in addition to lipid dysfunction, chronic arterial wall inflammation is a key component of atherosclerosis. Techniques that target cell adhesion molecules, which are overexpressed during inflammation, are effective methods to detect and treat atherosclerosis. Specifically, research groups have identified vascular cell adhesion molecule-1, intercellular adhesion molecule-1, platelet endothelial cell adhesion molecule, and selectins (E-selectin and P-selectin) as correlated to atherogenesis. In this review, we discuss recent strategies both in vivo and in vitro that target cell adhesion molecules. First, we discuss peptide-based and antibody (Ab)-based nanoparticles utilized in vivo for diagnostic, therapeutic, and theranostic applications. Second, we discuss flow-based in vitro models that serve to reduce the traditional disadvantages of in vivo studies such as variability, time to develop the disease, and ethical burden, but preserve physiological relevance. The knowledge gained from these targeting studies can be translated into clinical solutions for improved detection, prevention, and treatment of atherosclerosis.

Impact statement

As atherosclerosis remains the leading cause of death, there is an urgent need to develop better tools for treatment of the disease. The ability to improve current treatments relies on enhancing the accuracy of in vitro and in vivo atherosclerotic models. While in vivo models provide all the relevant testing parameters, variability between animals and among models used is a barrier to reproducible results and comparability of NP efficacy. In vitro cultures isolate cells into microenvironments that fail to take into account flow separation and shear stress, which are characteristics of atherosclerotic lesions. Flow-based in vitro models provide more physiologically relevant platforms, bridging the gap between in vivo and 2D in vitro models. This is the first review that presents recent advances regarding endothelial cell-targeting using adhesion molecules in light of in vivo and flow-based in vitro models, providing insights for future development of optimal strategies against atherosclerosis.

Introduction

Cardiovascular disease (CVD) is the leading cause of death in the USA, with approximately 801,000 deaths each year.¹ The most common type of CVD is atherosclerosis, which accounts for 68% of CVD deaths.² Atherosclerosis is a degenerative disease that is characterized by the irregularity of lipid metabolism, endothelial dysfunction, and the buildup of rupture-prone plaques, resulting in acute coronary blockage syndromes and/or sudden cardiac arrest.^3,4 A prominent feature of atherosclerosis is inflammation of the endothelium, which instigates the influx of monocytes to the plaque, driving plaque thickening and rupture.^5,6 Thus, targeting inflamed endothelia has emerged as a therapeutic strategy against atherosclerosis.

The onset of atherosclerotic plaques begins with the accumulation of low-density lipoproteins (LDLs) in the subendothelium.^7,8 Endothelial cells (ECs) are activated to express leukocyte adhesion molecules that result in monocyte recruitment, attachment, and differentiation into macrophages. These macrophages eventually produce enzymes that degrade the extracellular matrix and cause plaque rupture.^9,10 The ECs in these regions respond through mechanosensors, which transduce the local flow shear stress into changes in levels of gene expression.¹¹ For instance, activated ECs begin to overexpress cell adhesion molecules (CAMs), such as vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), platelet endothelial cell adhesion molecule (PECAM-1), and selectins (E-selectin and P-selectin) on their surfaces.

There are two major weaknesses in current treatments that prevent conventional small molecule drugs from optimal effectiveness against atherosclerosis: (1) their short blood circulation time and (2) their inability to select for diseased tissue. Nanoparticles (NPs) offer numerous advantages over traditional molecules, including high payloads, tunable sizes, tailorable surface properties, controllable drug kinetics, and improved pharmacokinetics.^12–18 Many NP-based platforms with the ability to target plaques, the hallmark of atherosclerosis, have thus been developed as a novel means to enhance diagnostic sensitivity and therapeutic efficacy.^19–22

Interactions between NPs and their local environments are dependent on both size and surface properties. Because of their small size (10–200 nm in diameter), NPs can penetrate through “leaky” regions of inflamed vasculature, which is known to be part of the enhanced permeability and retention (EPR) effect.²³ When the charge of NP surface is highly positive or negative, clearance by the mononuclear phagocytic system (MPS) macrophages is significantly enhanced. In order to stabilize particles against rapid decomposition in circulation, research groups have encapsulated them within a lipid layer containing water-soluble polymers such as polyethylene glycol (PEG).²⁴ This coating shields the NP from the immune system, prevents degradation, and increases circulation time to improve drug accumulation and controlled drug release. Furthermore, small molecules, dyes, and targeting moieties can be encapsulated inside or conjugated to the surface of NPs.²⁵ Targeting moieties, which include peptides, antibodies (Abs), integrin, and endothelial-specific ligands, can drive NPs to localize at the desired site and bind to molecules or receptors expressed on the cell surface.²⁶ The high surface-to-volume ratio of NPs makes them optimal for conjugating therapeutics or targeting ligands onto their surfaces for enhanced efficacy and avidity. Nanoparticles can also carry drug payloads in their cores, which has been shown to promote therapeutic effectiveness and reduce side effects by improving pharmacokinetics.^17,27–29

Prior reviews have provided a biological understanding of how CAMs play a critical role in the development of atherosclerosis and plaque rupture, but only a few have considered NPs targeting CAMs and their testing in both in vivo and in vitro atherosclerotic models.^30–33 The treatment of atherosclerosis is often limited by the lack of understanding regarding the interplay between the NP and the endothelium. In vivo models provide a platform that includes all the parameters in a physiologically functional system, which usually suggests clinical relevance. However, the ethical issues and high costs associated with the use of animals and the inherent variable nature between individual specimens are barriers to repeatable results and comparability of NP efficacy. At the other end of the spectrum are highly controlled 2D cell cultures, which force cells into isolation and do not provide a physiologically relevant microenvironment. In vitro flow models can close the gap between 2D cell culture and animal experiments, providing additional parameters such as shear stress, 3D architecture, and co-culture conditions. In this review, current targeted strategies using NPs are reviewed, focusing on targeting moieties that enable NP localization to activated ECs expressing VCAM-1 as well as other major CAMs. We first discuss recent developments of diagnostic and therapeutic NPs targeting ECs using peptides and Abs in in vivo models (Table 1). The second half of this review focuses on the in vitro flow models that have been specifically developed for evaluating the targeting efficiency of particles to the endothelium using CAMs, as well as models investigating CAM expression and leukocyte recruitment in response to disturbed flow conditions.

Table 1.

Nanoparticles targeting cell adhesion cells for in vivo diagnostic, therapeutic, and theranostic applications in atherosclerotic-related diseases

Targets	Targeting moieties	Treatments	Particle types	References
VCAM-1	VHSPNKK Peptide	Diagnostics	Magnetofluorescent	38
	VHPKQHR Peptide	Diagnostics	Magnetofluorescent	39
	VHPKQHR Peptide	Diagnostics	Micelle	40
	VLTTGLPALISWIKRKRQQ Peptide	Diagnostics	Perfluorocarbon Cored Nanocarrier	41
	NNSKSHT Peptide	Diagnostics	SPIO	42,43
	VHPKQHRAEEAK Peptide	Diagnostics	Tobacco Mosaic Virus	44
	C*NNSKSHTC*C Peptide	Diagnostics	Micelle	45
	VHPK Peptide	Therapeutics	Cationic Lipoparticle	46
	VHPKQHRGGSKGC Peptide	Therapeutics	Liposome	47
	VHSPNKK	Theranostics	Simian Virus 40	49
	Ab(M/K-2.7)	Diagnostics	Iron Oxide	54
	Ab(429)	Diagnostics	Gold Nanoshell	55
	Nanobodies	Diagnostics	Nanobody	56
	Ab	Therapeutics	Liposome	57
PECAM-1	Ab	Therapeutics	Polymer Nanocarrier	58
ICAM-1	LFA-1 Integrin	Imaging	SPIO	59
	Ab(R6.5)	Therapeutics	Antibody carrier	60

In vivo models and CAMs expression

Peptide-based nanomaterials for targeting VCAM in vivo

VCAM-1 is an adhesion molecule that is overexpressed on the surfaces of inflamed ECs in atherosclerosis.^34,35 VCAM-1 acts as a mediator in the recruitment of monocytes to the plaque.³¹ It plays a critical role in the inflammatory process and its expression is often correlated with the progression of atherosclerotic lesions. For these reasons, VCAM-1 expression is a reliable target to consider in the development of several in vivo imaging tools and therapies against atherosclerosis. One strategy to incorporate a biomarker for cell-specific binding and localization is by modifying the surface of NPs with peptides. Peptide-based nanomaterials provide greater selectivity than free drugs, therefore limiting the potential off-target side effects generally associated with small molecule targeting.³⁶ Due to their ability to form secondary structures, such as helices and coils, peptides can be presented on the exterior of the NP for active targeting.³⁷ In addition, their small size offers enhanced penetration into tissues over whole proteins.³⁶

Recent efforts have been directed toward enhancing in vivo diagnostic methods to detect vulnerable, atherosclerotic plaques prone to rupturing, which can allow for earlier intervention and may ultimately reduce the numbers of heart attacks and strokes. A number of imaging modalities exist for vulnerable plaque detection, from optical imaging to magnetic resonance imaging (MRI). Kelly et al. used phage display to derive the peptide sequence VHSPNKK that could target and be internalized by VCAM-1.³⁸ This sequence is homologous to the alpha-chain of late antigen (VLA-4), which is expressed by the monocytes and lymphocytes that bind to VCAM-1 on activated EC surfaces. Kelly et al. synthesized VHSPNKK-modified magnetofluorescent NPs with superparamagnetic and fluorescent properties to enable evaluation of VCAM-1 expression via fluorescence imaging or MRI. The peptide-conjugated NP successfully targeted VCAM-1-expressing ECs and accumulated in the vessel wall of apolipoprotein E-knockout (ApoE^−/−) mice. Under fluorescence imaging, the peptide-conjugated NPs exhibited enhanced target-to-background ratios of about 12-fold higher compared to those functionalized with VCAM-1 monoclonal Abs. Moreover, it was demonstrated for the first time that endothelial targets could be detected in vivo by MRI without the use of VCAM-1 Abs. Nahrendorf et al. developed a second generation of these VCAM-1-targeting magnetofluorescent NPs for the detection of early-stage inflammation in cholesterol-fed ApoE^−/− mice. More specifically, the NPs were conjugated to a different linear VCAM-1-targeting peptide, VHPKQHR (VHP), which was shown to exhibit higher cellular internalization for enhanced MRI resolution and optical imaging.³⁹ Injecting these particles into ApoE^−/− mice led to a 77% increased contrast-to-noise ratio between the aortic wall and adjacent blood pool by MRI imaging. Fluorescent images confirmed these results, as the aortic root had a 350% higher plaque target-to-background ratio compared to saline-injected ApoE^−/− mice. Mlinar et al. also reported conjugating the same peptide discovered by Nahrendorf et al. to the surfaces of spherical, self-assembled peptide amphiphile micelles (PAMs). Apart from having targeting capabilities, the PAMs were labeled with the fluorochrome Cy7 to enable particle tracking and fluorescence imaging in vivo.⁴⁰ After 24 h post injection, these VCAM-1-targeting PAMs were able to differentiate between early- and mid-stage atherosclerotic plaque in ApoE^−/− mice (Figure 1). Others have reported using alternative peptides to target VCAM-1 for MRI. Pan et al. developed multifunctional perfluorocarbon (PFC)-cored nanocarrier functionalized with the VCAM-1-targeting linker peptide, VLTTGLPALISWIKRKRQQ, for imaging in ApoE^−/− mice.⁴¹ The VCAM-1-targeted lipidic PFC NP showed a four-fold increase in accumulation in the aortas of ApoE^−/− mice, resulting in enhanced ¹⁹F MR contrast compared to the control (Figure 2).

Figure 1.

Cy7-labelled PAMs are detectable by near-IR in vivo imaging in ApoE^−/− mice. (a) to (d) 24-h post injection, control (not functionalized with a targeting moiety) PAMs mostly show a strong signal in the bladder and liver, but not in the aorta. (e) to (g) In contrast, VCAM-1-targeting PAMs localize in the cardiovascular system (denoted by arrow), primarily in the aorta. Adapted from Mlinar et al.⁴⁰ with permission from Elsevier. (A color version of this figure is available in the online journal.)

Figure 2.

Multifunctional perfluorocarbon (PFC)-cored nanocarrier carrying VCAM-1-targeting linker peptide. (a) The VCAM-1-targeted NPs localized in the aortas of ApoE^−/− mice, as signified by the higher ¹⁹F signals compared to the control mice receiving targeted or non-targeted NPs and ApoE^−/− mice receiving non-targeted nanoparticles. *P < 0.05. Histological analyses of in vivo targeting show that (b) only autofluorescence signals were detected in the aorta of control mice. Scale bar: 500 µm. (c) VCAM-1 staining (green) was strongly detected in the aorta of ApoE^−/− mice (FITC-labeled secondary antibody). Blue is staining of cell nuclei by DAPI. Scale bar: 200 µm. Reprinted and adapted from Pan et al.⁴¹ with permission from Federation of American Societies for Experimental Biology. (A color version of this figure is available in the online journal.)

Because MRI has shown promise in its superior capability of characterizing atherosclerotic plaques with high accuracy and reproducibility, efforts have been directed towards optimizing the delivery of contrast agents. For instance, Burtea et al. synthesized ultrasmall superparamagnetic iron oxide particles (USPIO), which were revealed to be strong MRI contrast agents, and functionalized with VCAM-1-binding cyclic heptapeptides (NNSKSHT motif) for effective VCAM-1 targeting in aorta plaques.⁴² The VCAM-1-targeting USPIO NPs were able to reach vulnerable plaques as quickly as 32 min post-injection compared to 221 min using the control, and at low doses of administration. This was further supported in a study by Michalska et al. in which the same cyclic heptapeptide was functionalized onto USPIO NPs to detect early and advanced atherosclerotic lesions.⁴³ By combining ultra-high-field MRI with surface-enhanced anti-Stokes Raman scattering microscopy, these USPIO NPs showed enhanced contrast, higher spatial resolution, and precise localizations in atherosclerotic lesions and cells expressing VCAM-1 within the vessel wall of ApoE^−/− mice.

Regarding MRI, several NP-based platforms modified with chelated gadolinium (Gd), a contrast-enhancing agent, have been developed to target ECs and image atherosclerosis. Bruckman et al. synthesized VCAM-1 targeting, tobacco mosaic virus (TMV) NPs modified with sulfo-Cy5-azide dyes and Gd ions for dual optical-MR imaging capabilities in vivo.⁴⁴ Immunofluorescence imaging of aortas in ApoE^−/− mice showed that the VCAM-1 targeting TMV NPs selectively localized to the activated ECs on atherosclerotic plaques compared to non-targeting control NPs. In addition, there was no accumulation of VCAM-1 targeting TMV NPs in healthy C57B1/6 mice. The authors hypothesized that the elongated shape of the NPs contributed to improved plaque localization, as more copies of the ligand could be presented to the flat vessel wall of atherosclerotic plaques compared to spherical NPs. Moreover, chelating Gd to TMV enabled plaque detection by MRI at an injected dose 400 times lower (0.00025 mmol/kg Gd ion) compared to the typical clinical dose of chelated Gd molecules (0.1 mmol/kg Gd ions). In a different study, Pagoto et al. developed phospholipid-based micelles containing the amphiphilic Gd-DOTAMA(C18)₂ and functionalized with the VCAM-1 receptor-targeting, cyclic peptide C*NNSKSHTC*C.⁴⁵ These Gd-based micelles demonstrated enhanced MRI T₁ signal (2–3-fold) in the inflamed region of lipopolysaccharide-induced mice compared to healthy and diseased mice administered free Gd, confirming successful targeting and increased contrast for inflammation detection.

In addition to diagnostics, advances have been made in the application of peptide-targeting NPs for therapeutic purposes in atherosclerosis. Kheirolomoom et al. developed a coated, cationic lipoparticle (CCL) nanocarrier carrying anti-miR-712 and VHP peptide (VHPK-CCL).⁴⁶ VHPK-CCLs selectively delivered anti-miR-712 to the pro-atherogenic, inflamed endothelium expressing VCAM-1 in disturbed-flow regions in vivo, while silencing miR-712 expression. As a result, VHPK-CCL-anti-miR-712 significantly reduced atherosclerotic lesions, ultimately inhibiting further disease progression. Furthermore, VHPK-CCL-anti-miR-712 could be injected at a dose 80% lower than that of naked anti-miR-712 to successfully prevent atheroma formation in ApoE^−/− mice. Another peptide-therapeutic strategy by Calin et al. demonstrated the ability of PEGylated target-sensitive liposomes (TSL) carrying Teijin compound 1, an antagonist of the chemokine CCR2, to interfere with chemokine/receptor interaction at the surface of activated ECs.⁴⁷ The surface of the TSL was conjugated with VCAM-1 targeted peptide (VHPKQHRGGSKGC motif) to specifically bind to the developing atherosclerotic plaque in the aorta of ApoE^−/− mice and deliver Teijin 1 compound to reduce monocyte infiltration and block inflammation. The effects of peptide-targeting were confirmed by fluorescent imaging, which revealed that non-targeted TSL had a 1.3-fold lower radiant efficiency compared to VCAM-1-targeting TSL.

Nanoparticles with the ability to co-deliver diagnostic and therapeutic agents on a single platform are termed “theranostic” NPs. These NPs offer benefits over NPs dedicated solely to diagnostics or therapeutics, as they can simultaneously treat the disease and monitor particle biodistribution in the body over time.⁴⁸ In this pursuit, Sun et al. developed self-assembling trifunctional Simian virus 40 (SV40) NPs encapsulated with quantum dots and the anticoagulant drug Hirulog and functionalized with a VCAM-1 targeting peptide (VHSPNKK motif) for targeting.⁴⁹ These multifunctional NPs were able to target, image, and deliver drugs to early, developmental, and late-stage atherosclerotic plaques in ApoE^−/− mice. Selectively targeting VCAM-1 allowed a higher concentration of Hirulog to the sites of interest, confirming the ability of VHSPNKK to target VCAM-1. To overcome challenges of optical imaging such as tissue absorbance and scattering, Sun et al. incorporated near-infrared quantum dots (NIR QDs) for their high detection sensitivity in deep-tissue imaging for early- and late-stage atherosclerosis in ApoE^−/− mice. They demonstrated that the NIR QDs functionalized with the VCAM-1-targeting VHSPNKK peptide resulted in a 6.5-fold increase in accumulation to atherosclerotic plaques compared to QDs without the targeting moiety.

Ab-based nanomaterials for targeting VCAM in vivo

Functionalizing NPs with Abs is another common strategy used to enhance the diagnostic and therapeutic index at the diseased site while reducing off-target side effects. Abs, which are nanometer sized biological proteins part of the body’s specific immune system, offer many advantages for in vivo imaging, targeting, and drug delivery in atherosclerosis.^50,51 For example, Abs allow NPs to bind to the target with enhanced affinity and increased cell penetration.⁵² Like peptides, the small structure of antibodies allows for higher loading on the surfaces of NPs with limited steric constraints, substantially increasing the number of NPs that can be bound to the target.⁵³ Moreover, Abs enable NPs to be internalized via receptor-mediated endocytosis. These improvements in cellular uptake and intracellular stability allow for a higher intracellular concentration of drugs to be delivered at the intended site at or within the atherosclerotic plaque.

Tsourkas et al. developed magnetooptical NPs conjugated to anti-VCAM-1 Abs (VCAM-NP) to target and noninvasively image early-stage inflammation of the endothelium.⁵⁴ These VCAM-NPs were cross-linked with dextran-coated iron oxide, which has superparamagnetic properties, and Cy5.5 near-infrared fluorescent markers to enable detection by MRI and optical imaging. Following intravenous administration of VCAM-NPs to C57B1/6 mice, the fluorescent signal from NPs localizing to the activated EC surface signified selective targeting by the anti-VCAM-1 Abs.

Photoacoustic tomography (PAT) is another attractive imaging modality due to its nanomolar sensitivity to molecular contrast agents and sub-millimeter spatial resolution compared to MRI and fluorescence imaging. This is extremely important in the detection and characterization of atherosclerotic plaques and plaque progression, which require high spatial resolution and high molecular contrast to get an accurate prognosis and for subsequent therapy development. Gold nanoshells (AuNS) have strong absorption abilities and have been previously used for imaging applications in oncology, making these NPs an attractive molecular probe for imaging. Rouleau et al. synthesized gold nanoshell VCAM-1-targeted photoacoustic probes, known as immunonanoshells.⁵⁵ These immunonanoshells accumulated in VCAM-1 expressing atherosclerotic lesions of ApoE^−/− mice by PAT. Ex vivo optical projection tomography of excised aorta confirmed this result, showing enhanced MR contrast.

Apart from optical imaging, MRI, and PAT, positron emission tomography-computed tomography (PET-CT) has been shown to have extremely high spatial resolution (3–5 nm), accurate quantification, and heightened sensitivity, making it a promising clinical imaging modality for atherosclerotic plaques detection. Bala et al. incorporated Fluorine-18 (¹⁸F) on VCAM-1 nanobodies ([18F]FB-cABCAM-1-5), which are small antigen-binding fragments (12-15 kDa) from heavy-chain portion of Abs in camelids.⁵⁶ In vivo PET/CT imaging of ApoE^−/− mice showed that cABCAM-1-5 were highly expressed in atherosclerotic lesions. Ex vivo analysis showed an increased uptake of targeted nanobodies in aortas 2.7 times higher than in control mice and 4.3 times higher compared to non-targeting nanobodies.

Modification of NPs with Abs to achieve therapeutic efficiency has also been accomplished. The anti-inflammatory properties of antiproliferative cyclopentenone prostaglandins (CP-PGs) make them optimal to incorporate in therapeutics against atherosclerosis. Antiproliferative cyclopentenone prostaglandins (CP-PGs) exhibit potent anti-inflammatory properties. Homen de Bittencourt et al. developed a negatively charged, liposome-based particle that was conjugated to anti-VCAM-1 Abs (LipoCardium) and was able to selectively deliver PGA2 to the injured arterial wall cells of adult LDL receptor knockout (ldlr^−/−) atherosclerotic mice.⁵⁷ At the site of inflammation, LipoCardium accumulated at the sites of adhesion molecules expressed on the inflamed EC surfaces and released PGA2 (Figure 3). It inhibited plaque progression through its tetravalent effects (anti-inflammatory, anti-inflammatory, anti-proliferative, anti-cholesterogenic, and cytoprotective), reversed vascular lesions, and reduced myocardium infractions in ldlr^−/− mice compared to control non-treated ldlr^−/− mice.

Figure 3.

Endothelial dysfunction triggers inflammation, leading to an overexpression of VCAM-1 receptors on the surface of the activated ECs. (1) Following injection, the LipoCardium travel through the bloodstream. (2) The anti-VCAM-1 Abs lead the LipoCardium to VCAM-1 receptors, where it is taken up by the endothelial cell. (3) Once internalized by the cell, the lysosome disassembles and releases its contents, PGA2 molecules. Reprinted from Homem de Bittencourt et al.⁵⁷ with permission from Elsevier. (A color version of this figure is available in the online journal.)

Targeting other CAMs in vivo

PECAM-1 and ICAM-1 are responsible for the transmigration of leukocytes from the blood vessel into the endothelium and intima.³⁰ PECAM-1 - and ICAM-1-targeting platforms have also been developed to interfere with the inflammation of integrins and leukocytes responsible for endothelial monolayer integrity in various diseases. Although PECAM-1 and ICAM-1 have not been used directly in any in vivo NP applications for atherosclerosis, many researchers have used both of these targets to better understand their mechanistic and immunogenic properties. For example, Dziubla et al. developed a PECAM-targeted polymer nano-carrier (PNC) loaded with catalase enzyme to protect against vascular oxidative stress.⁵⁸ These PECAM-targeted nanocarriers demonstrated sufficient accumulation in the pulmonary vasculature, leading to reduced vascular oxidative stress. The authors theorized that this application could also be applied for atherosclerosis treatment. In another example, leukocyte-mimetic SPIO-based micelles (LMN) incorporated with LFA-1 integrin for ICAM-1 targeting demonstrated a rapid and non-invasive MR imaging technique to identify organ-specific inflammation in mice treated with lipopolysaccharides (LPS).⁵⁹ Due to the overexpression of ICAM-1 from LPS inflammation, there was a two-fold greater LMN accumulation in the liver of mice treated with LPS than that of non-ICAM-1 control mice. Serrano et al. demonstrated that the binding between ICAM-1 and ICAM-1 targeted carriers could induce CAM-mediated endocytosis.⁶⁰ The binding of these anti-ICAM carriers to ICAM-1 in the endothelium led to intracellular transport of these carriers by CAM-mediated endocytosis. The engagement of ICAM-1 in lipid domains, complexing with NHE1 (a linker between CAM-mediated endocytosis and cytoskeleton), acid shingomyelinase drives the hydrolysis of sphingomyelin into ceramide. The resulting actin polymerization and cytoskeleteon remodeling stabilizes the drug delivery platform and suggests that ECs can internalize relatively large drug carriers that are targeted to ICAM-1.

Flow-based in vitro models and CAMs expression

According to Zheng et al., the interplay between the arterial microenvironment and atherogenesis remains unclear, partially due to the gap between cell culture and animal experiments.⁶¹ In vitro models can span the gap between 2D culture and in vivo testing, thus reducing the cost, time, and ethical burden of current approaches.⁶² Furthermore, the lack of plaque rupture and thrombosis in animal models, gives advantage to in vitro systems for mimicking events that are characteristic of the late stages of the disease.⁶³ Studying endothelial mechanobiology requires engineered tools that can maintain EC in a controlled in vitro culture environment while exposing them to mechanical stimuli.⁶⁴

Flow chamber-based models have been used to study both early- and late-stage pathogenic processes in atherosclerosis.⁶³ Various macro- and micro-scale cell culture flow (CCF) systems have been reviewed by Young and Simmons for studying adhesion, migration, flow-induced mechano-trasnduction, and permeability of ECs.⁶⁴ Cone-and-plate and parallel plate flow chamber (PPFC) are two most popular macro-scale CCF in vitro devices for controlled shear stress studies.⁶⁵ The family of CAMs are used for targeted delivery of drugs and imaging probes to the endothelium.^66,67 In this next section, we briefly review the in vitro flow models and their implementation with micro- and nanoparticles that have been specifically developed for: (1) studying the effect of disturbed flow on CAM expression and leukocyte recruitment and (2) evaluating the targeting efficiency to endothelium by using CAMs.

CAM expression, leukocyte recruitment, and disturbed flow

Leukocytes, including neutrophils and monocytes, adhere to the site of inflammation.^68,69 Tethering and rolling of leukocytes on ECs are controlled by selectins (E and P), adhesion is regulated by integrins, and transmigration is guided through the gap junctions by PECAM-1. Consequently, evaluating CAM expression in flow for ECs is central to understanding the nature of interaction of leukocytes with ECs. Initiation of atherosclerosis involves endothelial dysfunction in response to oxidative cholesterol particles as well as altered local hemodynamics.⁶³ Disturbed and non-laminar flow below healthy physiological levels of shear stress (10–70 dyn/cm²) can cause EC phenotype changes resulting in pro-inflammatory endothelium.^70–72 Here, we review in vitro models developed for determining the role of disturbed flow patterns on CAM expression and leukocyte recruitment, and conclude with the testing of targeted particles against CAMs in such models.

Cicha et al. studied adhesion of human monocyte cell-line (THP-1) at 5 dyn/cm² to HUVECs that have been perfused at 10 dyn/cm² for 18 h in a bifurcating flow-through cell (y-shaped µ-Slide).⁷³ Cells in the straight segment of the channel exposed to uniform shear stress had polygonal shape and elongated and aligned with the direction of the flow. On the contrary, cells exposed to non-uniform stress in bifurcations had irregular shapes. The authors observed an induction of E/P-selectins, VCAM-1, and ICAM-1 for the cells exposed to non-uniform shear stress compared to the cells in the straight segment. A moderate increase in the cells’ ability to recruit monocytes was observed upon exposure to non-uniform shear stress.

Adhesion of neutrophils to human abdominal aortic endothelial cells (HAAEC) has been studied by Rouleau et al. in an eccentric stenosis model (divided into five sections: inlet, proximal, peak, recirculation and distal zones) with 50% area reduction (Figure 4(a)).⁷⁴ Acute promyelocytic leukemia cell line (NB4) activated with all-trans-retionic acid (ATRA) appeared to be significantly attached to the recirculation zone of the model for both the non-stimulated and TNF-α stimulated cells under stress conditions of 1.25 and 6.25 dyn/cm². Higher levels of VCAM-1 were detected near the stenosis peak of 1.25 dyn/cm². The adhesion was also dependent on time (1 vs. 6 h) and shear magnitude. Pre-shearing the cells at 1.25 dyn/cm² for 24 h significantly decreased the adhesion of NB4 cells (lasting 1 h) to both stimulated and non-stimulated ECs. Pre-shearing caused cells to express higher levels of ICAM-1 and VCAM-1 in the recirculation zone.

Figure 4.

Various in vitro models for studying disturbed flow and CAM expression. (a) Asymmetric stenosis model. Model flow regions, including the inlet, proximal, peak, recirculation and distal regions. Adapted and reprinted from Rouleau et al.⁷⁴ with permission from the American Society of Hermatology. (b) Schematic diagram of the flow channel and test section. Flow separation occurs in the region distal to the step, forming four specific flow areas. Adapted and reprinted from Chen et al.⁷⁵ with permission from IOP Publishing. (c) PDMS mold of the linear shear flow chamber and surrounding vacuum network that seals the chamber to the HAEC monolayer on a cover slip. Adapted and reprinted from Tsou et al.⁷⁷ with permission from the American Physiological Society. (d) Representative bright field images of multiple vascular geometries including a straight channel, stenosis, aneurysm, and bifurcation. Adapted and reprinted from Mannino et al.⁸⁰ with permission from Nature Publishing Group. (e) Modular, sub-millimeter-sized cylinders are made from collagen with embedded functional cells and surface-attached endothelial cells. Adapted and reprinted from Khan and Sefton⁸¹ with permission from Elsevier. (f) The image on the top shows the view of laminar flow, and the image at the bottom shows the disturbed flow and regions selected for the study DS1, DS2, and DS3 based on the WSS measurement. Adapted and reprinted from Balaguru et al.⁸² with permission from Nature Publishing Group. (A color version of this figure is available in the online journal.)

Chen et al. used a vertical step flow system (divided into four sections: stagnant flow area, recirculation eddy, reattachment area, and developed laminar flow) that exposed the cells to shear stresses (Figure 4(b)) in the range of 0–7 dyn/cm² and investigated the adhesion/transmigration of leukocytes in ECs co-cultured with smooth muscle cells (SMCs).⁷⁵ There was significant adhesion of neutrophils, lymphocytes, and monocytes to IL-1β activated co-culture system consisting of ECs seed on SMCs in the reattachment zone of this model. Co-culturing ECs with SMCs significantly amplified the adhesion of leukocytes relative to ECs cultured alone. Compared to neutrophils and lymphocytes, monocytes showed very low mobility underneath the ECs, suggesting that they possess the greatest likelihood to be deposited in the vessel wall following transmigration into the sub-endothelial space.

Conway et al. developed a PPFC reversing flow system to recreate the physiological form of the reversing shear stress (time-average: 1 dyn/cm², maximum 11 dyn/cm² and minimum of −11 dyn/cm²) of the carotid sinus wall.⁷⁶ HUVECs exposed to high steady shear stress (15 dyn/cm²) for 24 h aligned to the direction of flow, while cells exposed to low steady shear stress of 1 dyn/cm² or to the reversing flow were randomly oriented. Upon stopping the perfusion of THP-1 monocytes after 24 h, significant increase for adhesion was observed in the cells exposed to the reversing flow compared to static, low and high stress. Increasing values for ICAM-1 expression was observed for reversing flow compared to the static condition, while no increase for VCAM-1 expression could be detected.

In another study, Tsou et al. studied monocyte recruitment by TNF-α activated HAECs in a PDMS microfluidic flow chamber exposing the cells to a linear gradient of shear stress that increased from 0 to 16 dyn/cm² (Figure 4(c)).⁷⁷ Spatial variations of ICAM-1, VCAM-1, and E-selectin were determined and it was concluded that VCAM-1 expression is regulated with the highest spatial acuity in response to small fluctuations in shear stress. Furthermore, spatial CAM regulation is dependent on the magnitude of the applied stress and not the gradient. A specially designed configuration consisting of three independent rectangular flow channels allowed the authors to assess leukocyte recruitment at a constant shear rate (0, 2, 6, and 12 dyn/cm²). Moreover, in separate experiments, the authors assessed monocyte rolling, arrest, and transmigration in the direction parallel to that of shear preconditioning. At a critical steady shear stress of 7 dyn/cm², CAM expression was altered and monocyte recruitment and arrest was significantly increased.

Reduced wall shear stress (WSS) that occurs following implementation of coronary stents favors restenosis due to an increase in neointimal proliferation.⁷⁸ Punchard et al.⁷⁹ evaluated the effect of stent deployment on HUVECs in a tubular flexible silicon model of artery fabricated by dip-coating a non-adherent mandrel. Stent deployment affected the local flow pattern and significantly reduced cell alignment. Alteration of expression of inflammation-associated genes was also measured. Upregulation of inflammatory gene expressions (E-selectin, ICAM-1, and VCAM-1) was found 24 h post stent deployment. Mannino et al. prepared different geometries (stenosis, aneurysm, and bifurcation) with PDMS channels of circular cross section (Figure 4(d)).⁸⁰ WSSs in different locations of multiple variations of these geometries were calculated using computational fluid dynamics (CFD). By varying the percent stenosis, aneurysm radius, and bifurcation angle in their “do-it-yourself” models, the wall shear stress could be varied at different locations in their model. The authors could demonstrate a correlation between WSS and VCAM-1 expression in these models.

Khan et al. created a modular tissue-engineered construct by randomly assembling HUVEC seeded collagen made modules in a PDMS microfluidic chamber to study the effect of disturbed flow on cell activation (Figure 4(e)).⁸¹ Based on VCAM-1 and ICAM-1 expressions after 1 and 24 h of flow, no statistically significant up/downregulation was observed. Khan et al. also replicated the disturbed flow pattern of a modular tissue engineering construct in a microfluidic device exposing HUVEC cells to low (2.8 dyn/cm²) and high (15.6 dyn/cm²) average shear stress, where flow is in transitional state.⁷⁰ The PDMS microfluidic chamber was constructed by producing 3D images of packed modules using microcomputed tomography and then extracting a single plane image for photomask production. Perfusion of device with THP-1 cells confirmed that areas where multiple channels converge (node region of their model) are most prone to THP-1 attachment. Strong correlation between ICAM-1 and VCAM-1 expression was observed and high shear stress caused downregulation of activation markers.

Estrada et al. fabricated a microfluidic device capable of generating normal (average shear stress = 11 dyn/cm²) and disturbed flow (average shear stress = 1.3 dyn/cm²) patterns of atherosclerosis susceptible regions of abdominal aorta.⁷² By exposing HAECs to these types of flow, the authors concluded that in the absence of pro-inflammatory stimulation, PECAM-1, VCAM-1, and ICAM-1 were not significantly upregulated by flow condition. Cells cultured under disturbed flow exhibited round shape and random orientation.

Balaguru and co-workers created disturbed flow by placing a circular block at the center of a PPFC seeded with human endothelial line (EAhy926) to create three regions (DS1, DS2, and DS3) with different flow characteristics around this block (Figure 4(f)).⁸² Based on CFD analysis, lowest shear stress (5 dyn/cm²) was experienced behind the block (DS2) where retrograde flow was present and cells maintained polygonal shape similar to that of static control. For this region, the authors observed a different pattern of actin compared to the other regions where actin fibers were seen at the periphery of the cells opposed to the centralized dispersive pattern. ICAM-1 and PECAM-1 gene expression levels were quantified in each region and were compared to normal shear stress setting where no block was present. Replacing the circular block with a different shape (e.g. square, triangle, and irregular) is an interesting approach for studying real plaque mimicking conditions.

Targeting efficacy of particles to endothelium using CAMs

As stated by Kusunose et al., having an effective model for quantifying particle binding under near in vivo conditions is crucial for development and optimization of targeted nanocarriers.⁸³ According to Muzykantov et al., endothelial adhesion molecules are ideal markers for detection and drug delivery targets to treat vascular inflammation, thrombosis and oxidative stress.⁶⁶ We next turn our attention to the in vitro determination of binding and internalization of VCAM-1 -, ICAM-1 -, PECAM-1 - and selectin-targeting particulate carriers under the action of a flow field.

VCAM-1

Yang et al. prepared VCAM-1 targeted core-shell Fe₃O₄@SiO₂ NPs (355 ± 37 nm) by reverse micro-emulsion.⁸⁴ They investigated adhesion of these particles to a monolayer of HUVECs in PPFS (shear stresses of 0, 1.1, 5.15, and 9.94 dyn/cm²) upon stimulation with lipopolysaccharide (LPS) for 5 h for induction of inflammation. Adhesion of VCAM-1 targeted Fe₃O₄@SiO₂ particles was significantly higher when the cells were treated with LPS. The observed degree of adhesion decreased with increasing the shear stress and duration of exposure (0, 1, 5, and 10 min). Treatment of activated cells with anti-VCAM-1 Abs slightly decreased the binding of NPs which verified the VCAM-1 mediated adhesion of these particles. Kusunose et al. fabricated two different PDMS microfluidic chambers seeded with HUVEC to expose these cells to uniform (4.4 dyn/cm²) and gradient shear stresses (2.4–8.6 dyn/cm²).⁸³ Binding avidities of VHP peptide coated liposome were compared in static and dynamic conditions to TNF-α-activated HUVEC cells. Greatest accumulation for both 1% and 2% conjugated VHP liposomes occurred at the lowest shear rate and decreased steadily due to increased shear stress in the gradient shear chamber. In an uniform shear chamber, higher binding was observed by increasing the VHP peptide concentration in the liposomes while increasing the polyethylene-glycol (PEG) brush layer decreased the accumulation in both static and dynamic (4.4 dyn/cm²) cultures (Figure 5(a)). Relative to static culture, liposome binding was significantly amplified upon the application of the uniform shear stress for 1% and 2% VHP-conjugated liposomes. Gosk et al. prepared immunoliposomes conjugated to VCAM-1 Abs (83.1 ± 14.4 nm).⁸⁵ Binding of these liposomes to murine brain endothlioma cells (bEnd.3) was assessed in a doublet flow chamber activated with TNF-α (4 h) at shear rates of about 200 s⁻¹ (20 h). The authors determined that approximately 8% of the bound liposomes were internalized after 2 h. To confirm specificity, binding of VCAM-1 targeted liposomes was microscopically compared with human IgG conjugated liposomes of irrelevant specificity but similar size (91.3 ± 13.2 nm) over a period of 20 h.

Figure 5.

Effect of targeting ligand concentration on binding, internalization, and detachment of CAM targeted particles in flow. (a) VHP-conjugated liposome binding under static and shear conditions. Particle binding consistently increased with peptide concentration (p < 0.01), and also increased under shear (p < 0.01 except for 1%, up to 13-fold) (n ≥ 3). Reprinted from Kusunose et al.⁸³ with permission from Biomedical Engineering Society. (b) Shear stress-induced detachment of anti-ICAM carriers from HUVEC. Data represent mean ± standard errors (n = 30 carriers). Reprinted from Calderon et al.⁸⁶ with permission from IOS press. Acute exposure to flow stimulates endocytosis of anti-PECAM/NCs in EC. Effect of flow (1 dyn/cm², 30 min) on endothelial internalization (c) and binding (d) of Abs/NC carrying 50, 100, and 200 Ab molecules per NC. Adapted from Han et al.⁹³ with permission from the American Chemical Society. (NC: nanocarrier)

ICAM-1

Calderon et al. investigated the dynamics of adhesion and detachment of 1 µm poly(styrene) (PS) beads coated with Abs against ICAM-1 to TNF-α activated human umbilical vein endothelial cells (HUVEC) using PPFC.⁸⁶ Fixing cells with paraformaldehyde allowed exclusive investigation of particle binding of without the interference of carrier internalization. Four distinct patterns for the carriers were observed: (1) no interaction, (2) continuous rolling, (3) first rolling and then binding, and (4) first rolling and then detachment. Increasing ICAM-1 surface density (370, 1100, 4100 molecules per µm²) on the carrier resulted in a reduction in rolling velocity while increasing shear stress (0.1 and 0.5 Pa) amplifies this velocity. For the lowest Ab density tested (370 molecules per µm²), the detachment of the particle was steep when increasing the shear stress (Figure 5(b)). However, higher Ab densities (1100 and 4100 molecules per µm²) showed a markedly different behavior consisting of a minimal initial detachment and formation of a stable bond for the range of shear stresses explored (0–13 Pa). This study provided insight into monocyte attachment mechanisms for further investigation.

CAM-mediated endocytosis of NP in PPFC was explored by Bhowmick et al.⁸⁷ They determined the internalization of anti-ICAM Ab-coated PS particles (180 nm) in TNF-α activated flow-adapted (4 dyn/cm²) HUVEC cells cultured in PPFC. Cells that adapted a 24-h flow internalized the particles at a slower rate compared to none flow-adapted cells. The effect of treatment of flow-adapted ECs with inhibitors of endocytic pathways was also investigated by these authors. Among the inhibitors tested, amiloride (which inhibits CAM endocytosis) showed the strongest inhibitory effect. Targeting efficiency of a multimodal N-succinimidyl S-acetylthioacetate (SATA) ICAM-1-conjugated liposomal contrast agent (163 ± 2 nm) to TNF-α activated ECs in a unidirectional flow system (µ-Slide) has been studied by Paulis et al.⁸⁸ Although higher levels of stress resulted in higher ICAM-1 expression, application of flow reduced the ability of the targeted liposomes to adhere to mouse endothelioma cells (bEnd.5) within the range of stresses studied (0, 0.25, and 0.5 Pa).

Rosano et al. fabricated a physiologically realistic polydimethylsiloxane (PDMS)-based synthetic microvascular network (SMN) from the mapped microvascular network of a hamster cremaster muscle using a modified Geographic Information Systems (GIS) approach.⁸⁹ Compared to control IgG conjugated spheres, a significantly larger number of mouse anti-ICAM-1-conjugated particles (2 µm PS spheres) adhered to the ECs using SMN seeded with bovine aortic endothelial cells (BAEC) and activated with TNF-α for 4 and 24 h. The effect of shape on targeting Ab-coated NPs to endothelium was investigated by Kolhar et al.⁹⁰PS nanospheres (200 ± 0.01 nm in diameter) and nanorods (501 ± 43.6 nm × 123.6 ± 13.3 nm) of equal volume were coated with anti-ICAM-1 Abs for evaluating their endothelial targeting in a PDMS-based SMN seeded with rat brain endothelial cells (RBE4) activated with TNF-α. Greater attachment and internalization for nanorods was observed compared to anti-ICAM-1-coated nanospheres and IgG-coated spheres/rods flown at the shear rate of 240 s⁻¹. Muro et al.⁹¹ conjugated anti-ICAM-1 molecules to PS spheres (7000 molecules per µm²) and evaluated their targeting efficacy. PPFC seeded with HUVEC and activated with TNF-α at the shear stress of 9 dyn/cm² was used to determine the kinetics of NP binding under both static and flow conditions, which were both very rapid and indistinguishable from one another. The effect of exposure of HUVECs plated in PPFC to microparticles from human atherosclerotic plaques has been investigated by Rautou et al.⁹² The authors measured the adhesion of human monocytic cells (U937) in low (1 dyn/cm²) and high (10 dyn/cm²) shear stress and concluded that the exposure of ECs to plaque MP favors monocytic cell adhesion by transferring ICAM-1 to ECs.

PECAM-1 and selectins

Han et al. studied binding and internalization of PS NPs coated with anti-PECAM-1 Abs (180–200 nm) to endothelium following exposure to chronic and acute shear stress.⁹³ A six-channel µ-Slide was used to subject HUVEC cells to laminar shear stresses. These authors demonstrated that following a 16-h exposure to chronic steady flow at 5 dyn/cm², endocytosis of the particles over a period of 30 min is diminished due to phenotype changes associated with flow adaptation, compared to non-flow-adapted cells. On the contrary, acute shear stress exposure (1 dyn/cm² for 30 min) increased the internalization of these NPs containing 50, 100, and 200 Ab molecule per particle (Figure 5(c)). When exposed to acute stress, binding of low avidity NPs (50 Ab molecules per particle) to endothelium was decreased, while for high avidity particles (200 Ab molecules per particle) the binding was increased compared to the static culture (Figure 5(d)). Maximal internalization for low avidity NPs was observed at shear stresses in the range of 2–4 dyn/cm² for the values explored at 0–8 dyn/cm². Charoenphol et al.⁹⁴ studied the targeting efficiency of PS particles coated (avidin-biotin) with silaly Lewis^a (sLe^a), selectins-specific ligand, to HUVECs using PPFC. For 10 µm particles, the highest binding was observed in the presence of red blood cells (RBC) in plasma flow at the WSR of 200 s⁻¹ upon activation with IL-1β. At a fixed volume concentration, particle adhesion increased with an increase in size within the ranges (0.1, 0.5, 2, 5, and 10 µm) explored. These researches observed a critical WSR beyond which disruptive hemodynamic forces interfere with particle adhesion. The effect of channel height was also investigated by varying the height (127, 254, 508, 762 µm) at the WSR of 200 s⁻¹. For 2, 5, and 10 µm particles, an increase in channel height results in higher particle binding. No significant difference was observed for adhesion levels when comparing inverted channel to upright. A significant difference in binding was only seen with 5 and 10 µm particles at low channel heights and WSR in the inverted chamber.

The effect of particle shape on adhesion of sLe^a coated PS particles was investigated by Thompson et al.⁹⁵ HUVECs in PPFC were activated with IL-1β and targeting efficacy was evaluated in reconstituted blood using steady and pulsatile flow profiles. Three sets of particles with equivalent spherical diameters (ESD) of 2.07, 1.01, and 0.52 µm were used where the aspect ratio (AR) varied in each set. Increasing WSR did not result in a significant difference in the adhesion of any particles for the smallest ESD. For particle shapes with ESD of 2 µm, increasing WSR resulted in a decrease in normalized binding efficiency of particles with ARs of 1, 2, and 4, while for rods with AR of 9, high binding efficiency was observed for all the shear rates tested (200, 500, and 1000 s⁻¹). In pulsatile flow, highest binding was observed for 2 µm particles with ARs of 9 to 11. Moreover, the adhesion of these particles in various locations of the step channel geometry was measured. Far downstream from the stagnation point, the adhesion of particles with AR of 9 and ESD of 2 µm was highest at the WSR of 200 and 500 s⁻¹ (Figure 6(a) and (b) correspondingly) Interestingly, for WSR of 200 s⁻¹, at the stagnation point of the step channel, particle adhesion did not show any shape or AR dependency. For ESD of 0.5 µm, no improvement in adhesion due to particle shape or WSR was detected in the step channel and adhesion was minimal (Figure 6c and Figure 6d). Lin et al. prepared PS NPs (100 nm) conjugated to glycocalicin (the extracellular segment of platelet glycoprotein Ibα that specifically binds to P-selectin) using avidin-biotin complex.⁹⁶ The authors investigated the effect of shear stress on uptake of these NPs by histamine-activated human aortic endothelial cells (HAECs) exposed to 0, 5, and 15 dyn/cm² of shear stress in a PPFC. Increasing shear stress drastically decreased the uptake of non-conjugated NPs, but not for glycocalicin conjugated particles.

Figure 6.

Role of aspect ratio (AR) on binding of targeted particles in various locations of the step channel geometry. Adhesion of spheres and rods with 2 µm ESD [(a) 200 s^-1, (b) 500 s^-1] and 500 nm ESD [(c) 200 s^-1, (d) 500 s^-1] to activated ECs from 40% RBC in buffer at a fixed particle concentration of 5 × 10⁵ #/mL. Reprinted from Thompson et al.⁹⁵ with permission from Elsevier

Conclusion

A variety of peptides and Abs have been utilized in vivo to target the characteristic overexpressed CAMs in atherosclerosis. This gives researchers a wide selection of potential designs when developing diagnostic or therapeutic strategies. For instance, the development of molecular, contrast agents for MRI has been a large focus of current imaging research using nanomedicine that can translate into the clinic.⁴⁵ Therapies that seek out sites of CAM do not rely on traditional symptoms of atherosclerosis such as unnatural levels of lipoproteins, high blood pressure, and body weight, but are able to directly deliver drugs to the plaque site and even reverse vascular lesions based on molecular and cellular markers of the disease.⁵⁸ Many CAM-targeting particles (Table 2) have been tested in vitro in microfluidic devices and parallel-plate flow chambers that are able to elucidate atherogenic physical stimuli. As such, patterns of disturbed flow⁷⁰ and wall shear stresses⁸² and their relation to CAM expression can set the stage for yet undiscovered in vitro devices. As CVD is expected to remain a priority health issue, the rapid pace of advances in NP theranostics is sure to continue. As in vitro models move to those of greater complexity and physiological relevance with the use of sophisticated and advanced technologies including 3D printing and complex co-culture conditions, the multifaceted parameters of atherosclerosis can be further replicated. Current and upcoming in vivo and in vitro studies will allow the testing of novel therapeutics in a timely and cost-effective manner for the atherosclerosis.

Table 2.

Various synthetic CAM targeting particles tested in vitro by different researchers using flow-based models

Targets	Targeting moieties	Particle types	In vitro models	References
VCAM-1	VHPKQHR Peptide	Liposome	PDMS Microfluidic	83
	Phycoerythrin-labeled Ab	Core-Shell Fe3O4@SiO2	PPFC	84
	Ab(MK271)	Liposome	Doublet Flow Chamber	85
ICAM-1	Ab(R6.5)	PS Latex Particles	PPFC	86
	Ab(R6.5)	PS Spheres	PPFC	87
	Ab(YN1/1.7.4)	Liposome	µ-Slide (Ibidi, Germany)	88
	Ab	PS Spheres	PDMS Microvascular Networks	89
	Ab	PS Nanospheres/Nanorods	PDMS Microfluidic Networks	90
	Ab(R6.5)	PS Spheres	PPFC	91
PECAM-1	Ab(Ab62)	PS Spheres	µ-Slide (Ibidi, Germany)	93
Selectin-E	sLea	PS Spheres	PPFC	94
	sLea	PS Spheres/Rods	PPFC	95
Selectin-P	Glycocalicin	Carboxylated PS Spheres	PPFC	96

Note: The corresponding Ab clones are indicated in the parentheses.

Impact statement

As atherosclerosis remains the leading cause of death, there is an urgent need to develop better tools for treatment of the disease. The ability to improve current treatments relies on enhancing the accuracy of in vitro and in vivo atherosclerotic models. While in vivo models provide all the relevant testing parameters, variability between animals and among models used is a barrier to reproducible results and comparability of NP efficacy. In vitro cultures isolate cells into microenvironments that fail to take into account flow separation and shear stress, which are characteristics of atherosclerotic lesions. Flow-based in vitro models provide more physiologically relevant platforms, bridging the gap between in vivo and 2D in vitro models. This is the first review that presents recent advances regarding endothelial cell-targeting using adhesion molecules in light of in vivo and flow-based in vitro models, providing insights for future development of optimal strategies against atherosclerosis.

Authors’ contribution

KK, JJML and CP contributed equally. KK, CP, and EJC conceived the concept. KK, JJML, CP, JW, and EJC wrote the review.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.