Integrin-Mediated Leukocyte Recruitment in Atherosclerosis: A Mechanistic Review of β2 Integrins and VLA-4 Signaling Pathways

Abstract

Atherosclerosis is a lipid-driven, chronic immunoinflammatory disease in which site-selective leukocyte recruitment converts endothelial dysfunction into sustained arterial-wall inflammation. Among the integrin superfamily, β2 integrins merit focused consideration because they are predominantly expressed on leukocytes and couple chemokine sensing to arrest, adhesion strengthening, crawling, transendothelial migration, phagocytic responses, and outside-in signaling. This mechanistic review examines LFA-1 (αLβ2), Mac-1 (αMβ2), CR4 (αXβ2/CD11c/CD18), and αDβ2 (CD11d/CD18), while positioning VLA-4 (α4β1) as a functionally complementary β1 integrin. We emphasize that leukocyte recruitment is not a rigid receptor-by-receptor sequence: β2 integrins and VLA-4 act cooperatively but are not fully interchangeable, and their contribution varies with leukocyte subset, vascular bed, ligand availability, and inflammatory stage. Disturbed-flow signaling through endothelial fibronectin-binding integrins, particularly α5β1 and αvβ3, provides an upstream mechanotransduction context that increases endothelial adhesiveness and thereby conditions β2/VLA-4-dependent recruitment. Within circulating and lesional myeloid cells, integrin signaling extends beyond adhesion to cytoskeletal organization, phagocytosis, inflammatory gene regulation, lipid-associated phenotypic transitions, and tissue retention. Platelet-leukocyte coupling, especially Mac-1 interaction with platelet GPIbα after selectin-dependent tethering, further links vascular inflammation to thrombosis. Recent single-cell and spatial studies underscore marked plaque-cell heterogeneity, while also cautioning that transcript abundance does not report integrin affinity or avidity state. Clinical translation remains difficult: broad CD18 blockade did not improve outcomes in acute myocardial infarction trials, whereas successful α4-directed therapies in other inflammatory diseases demonstrate druggability but also expose the safety costs of systemic leukocyte-trafficking inhibition. Future strategies should therefore prioritize context-, ligand-, conformation-, and cell-selective modulation rather than indiscriminate integrin suppression.

Introduction

Atherosclerosis is not simply a disorder of lipid storage. It is a chronic disease of the arterial wall in which retention and modification of apoB-containing lipoproteins, endothelial dysfunction, maladaptive immune responses, and extracellular-matrix remodeling interact over years to produce clinically heterogeneous plaques. Lesion progression can generate a large necrotic core, loss or disorganization of matrix, fibrous-cap thinning, intraplaque hemorrhage, or surface disruption; the immediate clinical consequences include myocardial infarction, ischemic stroke, and peripheral ischemic events. Importantly, lesion size, inflammatory burden, and propensity to rupture are related but non-equivalent biological outcomes.¹

The focal distribution of plaques at arterial branch points and curvatures illustrates why local vascular biology matters. Atheroprone hemodynamics promote endothelial activation and a permissive adhesion landscape, while retained lipoproteins and inflammatory mediators establish chemokine gradients that recruit circulating monocytes and other leukocytes. Recruitment is therefore a mechanistic bridge between a systemic metabolic disturbance and a spatially restricted arterial inflammatory lesion. The canonical adhesion cascade comprises tethering and rolling, chemokine-triggered activation, firm arrest, intraluminal crawling, and transendothelial migration, but the relative molecular contribution of each step varies by leukocyte subset and vascular context.^2,3 These interconnected biochemical and cellular events from lipid-driven endothelial activation to leukocyte recruitment, foam-cell formation, and inflammatory amplification are summarized schematically in Figure 1.

Figure 1.

Integrin-mediated cellular and biochemical processes during atherogenesis. (A) The biochemical phase of atherogenesis begins with subendothelial retention and oxidative modification of apoB-containing lipoproteins, accompanied by increased reactive oxygen species (ROS), endothelial dysfunction and activation, and enhanced expression of pro-inflammatory mediators and adhesion molecules. These changes establish a vascular microenvironment that promotes leukocyte recruitment through selectin-dependent tethering and rolling, chemokine-induced integrin activation, firm adhesion, intraluminal crawling, and transendothelial migration. (B) The cellular phase illustrates monocyte recruitment into the arterial intima, differentiation into macrophages, uptake of modified lipoproteins, foam-cell formation, and subsequent amplification of vascular inflammation and plaque progression. Platelet–leukocyte interactions further contribute to thromboinflammation: platelet P-selectin interacts with leukocyte PSGL-1 during initial tethering, whereas Mac-1 (αMβ2; CD11b/CD18) can stabilize platelet–leukocyte adhesion through platelet GPIbα and, in appropriate contexts, through fibrinogen-mediated bridging involving platelet αIIbβ3. LFA-1 (αLβ2; CD11a/CD18) may additionally interact with platelet ICAM-2. The schematic also depicts oxidative stress, inflammatory mediator release, neutrophil extracellular trap formation, and progressive remodeling of the atherosclerotic plaque. The pathways shown are interconnected and context dependent and should not be interpreted as a strictly linear sequence or as implying that plaque size alone determines plaque instability. The symbol ↑ denotes increased expression, activity, or biological process intensity, as applicable.

Abbreviations: GPIbα, glycoprotein Ib alpha; ICAM-2, intercellular adhesion molecule-2; LFA-1, lymphocyte function-associated antigen-1; Mac-1, macrophage-1 antigen; NETs, neutrophil extracellular traps; oxLDL, oxidized low-density lipoprotein; PSGL-1, P-selectin glycoprotein ligand-1; ROS, reactive oxygen species.

Integrins occupy a distinctive position in this cascade because they are simultaneously adhesion receptors and bidirectional signaling machines. Ligand binding is regulated from inside the cell through conformational activation and clustering, while ligand engagement feeds mechanical and biochemical information back into the cytoskeleton and signaling networks.⁴ For leukocytes, the β2 family - LFA-1 (αLβ2; CD11a/CD18), Mac-1 (αMβ2; CD11b/CD18), CR4 (αXβ2; CD11c/CD18), and αDβ2 (CD11d/CD18) - is especially relevant because these receptors are concentrated in hematopoietic cells and directly govern leukocyte arrest, adhesive strengthening, migration, and effector responses. VLA-4 (α4β1), although a β1 integrin, is included because its interaction with VCAM-1 is a major and often non-redundant route for monocyte recruitment to atheroprone endothelium.^2,3

The rationale for focusing on β2 integrins is therefore not that other integrins are unimportant. Rather, β2 receptors form a leukocyte-centered signaling network that links vascular recruitment to cell activation, phagocytic behavior, macrophage retention, and platelet-leukocyte thromboinflammation. Recent work also makes it increasingly difficult to interpret these receptors in isolation: endothelial α5β1 and αvβ3 shape the adhesive microenvironment under disturbed flow, while CR4 and VLA-4 can cooperate during monocyte arrest. This review consequently places β2 integrins within the wider vascular integrin network, distinguishes direct atherosclerosis evidence from mechanistic extrapolation, and critically considers whether integrin biology can be therapeutically exploited without compromising essential immune surveillance.

Integrins in Atherosclerosis: Positioning β2 Integrins Within the Vascular Adhesion Network

Integrins are heterodimeric α/β receptors whose ligand repertoire and signaling output depend on subunit composition, cellular context, extracellular matrix, and activation state.⁴ In atherosclerosis, this diversity means that “integrin function” cannot be reduced to leukocyte adhesion alone. Leukocyte β2 integrins and α4β1 mediate recruitment, endothelial α5β1 and αvβ3 participate in matrix sensing and inflammatory mechanotransduction, platelet αIIbβ3 supports aggregation, and several β1-family integrins influence smooth-muscle and matrix biology. The biological question is therefore not whether one integrin family is dominant across the whole lesion, but which receptor-ligand modules are rate-limiting in a specific cell at a specific stage.

Endothelial α5β1 is particularly relevant to early atherogenesis. On a fibronectin-rich matrix, α5β1 amplifies oxidized LDL-induced focal adhesion kinase (FAK) signaling, NF-κB activation, VCAM-1 expression, and monocyte adhesion; endothelial α5 deletion or disruption of α5 signaling reduces experimental atherosclerotic inflammation.⁵ α5β1 also drives endothelial deposition of cellular fibronectin, creating a feed-forward matrix environment that favors inflammatory signaling.⁶ These findings provide a mechanistic context for leukocyte recruitment: endothelial integrins do not replace leukocyte β2 integrins or VLA-4, but can determine how strongly the endothelium expresses and organizes the ligands encountered by them.

αvβ3 has a related but non-identical role. In disturbed-flow models, αvβ3 inhibition attenuates shear-induced NF-κB signaling and proinflammatory gene expression, with corresponding reductions in early atherogenic inflammation.⁷ The distinction between α5β1 and αvβ3 is important because both bind fibronectin-associated matrices yet can transmit different mechanical and inflammatory signals. This heterodimer specificity argues against treating “β1” or “αv” signaling as a single pathway.

The present review therefore uses a leukocyte-centered scope while explicitly recognizing a multicellular integrin network. Endothelial and smooth-muscle integrins are discussed where they condition leukocyte recruitment or plaque phenotype; platelet integrins are considered where they intersect with leukocyte receptors. This scope is deliberate: it permits mechanistic depth on β2 integrins and VLA-4 without implying that leukocyte receptors operate independently of the vascular wall. Representative mechanistic and translational evidence supporting the roles of the principal integrins discussed in this review is summarized in Table 1.

Table 1.

Representative Mechanistic and Translational Evidence for Integrin Pathways Relevant to Atherosclerosis

Integrin/Pathway	Biological Context	Model/Evidence Type	Principal Mechanistic Finding	Translational Interpretation	Reference
Endothelial α5β1	oxLDL on fibronectin	Human aortic endothelial cells; ApoE−/− mice	α5β1–FAK signaling amplified NF-κB/VCAM-1 activation and monocyte adhesion; α5-directed inhibition reduced early lesion inflammation.	Supports an upstream endothelial target; preclinical evidence does not establish clinical efficacy.	[5]
Endothelial α5β1/fibronectin	Matrix remodeling	Endothelial-specific α5 deletion; ApoE−/− mice	α5β1 promoted cellular fibronectin deposition and early atherogenic inflammation.	Supports a matrix-dependent feed-forward mechanism that conditions leukocyte recruitment.	[6]
Endothelial αvβ3	Disturbed-flow signaling	Endothelial cells; disturbed-flow and diet-induced mouse models	αvβ3 supported flow-induced NF-κB activation, pro-inflammatory gene expression, and early atherogenic inflammation.	Provides mechanotransduction context upstream of leukocyte recruitment.	[7]
Endothelial α5β1/annexin A2	Oscillatory shear stress	Endothelial cells; ApoE−/− mice	Piezo1/PTP1B/annexin A2-dependent α5β1 trafficking and activation promoted endothelial inflammation and plaque development.	Provides a mechanistic explanation for site-selective endothelial activation.	[8]
α5/αv fibronectin-binding integrins	Disturbed flow and oxLDL	Endothelial cells; endothelial genetic perturbation	Fibronectin-integrin signaling promoted endoplasmic-reticulum stress and JNK–c-Jun inflammatory output.	Extends endothelial integrin signaling beyond a single FAK/NF-κB pathway.	[9]
LFA-1 (αLβ2)	Arterial leukocyte adhesion	Intravital microscopy; LFA-1-deficient mice	LFA-1 deficiency markedly reduced leukocyte adhesion in the aorta and femoral artery.	Strong functional evidence for arterial arrest and adhesion.	[10]
VLA-4 (α4β1)	Early-lesion recruitment	Perfused carotid arteries from ApoE−/− mice	VLA-4 blockade strongly reduced monocyte adhesion; VCAM-1 and the fibronectin connecting segment-1 contributed as ligands.	Supports a major VCAM-1-dependent route for monocyte recruitment.	[11]
β2-integrin activation/α4β1 compensation	Myeloid talin1 signaling	ApoE−/− mice with myeloid talin1 deletion	Talin1 deletion abolished β2-integrin-mediated adhesion but spared α4β1-dependent adhesion and unexpectedly increased atherosclerosis.	Demonstrates non-redundancy, compensatory trafficking, and signaling complexity.	[12]
CR4 (αXβ2; CD11c/CD18)	Hypercholesterolemic monocyte recruitment	CD11c−/−/ApoE−/− mice	CD11c deficiency reduced firm arrest, lesional monocyte/macrophage accumulation, and atherosclerosis.	Causal preclinical evidence linking CR4 to recruitment.	[13]
CR4 (αXβ2; CD11c/CD18)	Postprandial hypertriglyceridemia	Human feeding study; adhesion under flow	Monocyte CD11c increased after a high-fat meal and was associated with enhanced arrest on VCAM-1 under shear.	Human translational evidence linking metabolic state to adhesive phenotype.	[14]
CR4 → VLA-4 cross-talk	Foamy-monocyte arrest	Human/mouse monocytes; flow assays	High-affinity CD11c signaling activated VLA-4 through Syk/paxillin-associated signaling and promoted VCAM-1-dependent capture.	Direct example of cooperative, non-redundant integrin signaling.	[15]
CD11c-associated signaling	Macrophage migration and inflammation	Experimental atherosclerosis	GM-CSFR–SYK–JNK–FOXO1 signaling increased CD11c and promoted macrophage migration and atherosclerosis.	Connects CD11c expression to intracellular inflammatory signaling.	[16]
αDβ2 (CD11d/CD18)	Macrophage retention	CD11d−/−/ApoE−/− mice; adoptive transfer	CD11d deficiency reduced macrophage accumulation after lesion entry and reduced atherosclerosis.	Supports a retention-focused mechanism; direct human validation is lacking.	[17]
Mac-1 (αMβ2)–platelet GPIbα	Platelet–leukocyte thromboinflammation	Biochemical mapping and vascular-injury models	Mac-1 directly binds platelet GPIbα; disruption of this interface attenuated thrombosis in mice.	Ligand-interface targeting may preserve more immune function than pan-integrin blockade.	[18, 19]
β2 integrins	Macrophage phagocytic mechanics	Macrophage biophysics and phosphoproteomics	β2 integrins, talin1, and vinculin matched cargo rigidity to distinct phagocytic architectures.	Evidence for β2-integrin signaling beyond adhesion; not yet plaque-specific.	[20]
Human plaque cellular states	Plaque heterogeneity	Single-cell and multimodal human carotid studies	Multiple macrophage, lymphocyte, endothelial, and smooth-muscle cell states coexist within plaques.	Transcript abundance identifies cellular context but does not establish integrin activation state.	[21–23]

Notes: This table prioritizes studies that directly inform the mechanistic scope of the review rather than functioning as a catalogue of loosely related inflammatory or cardiovascular conditions. Findings from animal or cell models should not be interpreted as established clinical efficacy. Reference numbers correspond to the revised manuscript.

Abbreviations: FAK, focal adhesion kinase; GM-CSFR, granulocyte-macrophage colony-stimulating factor receptor; GPIbα, glycoprotein Ib alpha; NF-κB, nuclear factor kappa B; oxLDL, oxidized low-density lipoprotein; VCAM-1, vascular cell adhesion molecule-1; VLA-4, very late antigen-4.

Molecular Organization, Conformational Control, and Bidirectional Signaling of β2 Integrins

All four β2 integrins share the β2 chain (CD18) paired with a distinct α subunit. The extracellular headpiece of each α chain contains an inserted αI domain with a metal-ion-dependent adhesion site that directly contributes to ligand recognition. Structural differences around this domain account for substantial differences in ligand repertoire: LFA-1 is comparatively specialized for ICAM binding, whereas Mac-1 and CR4 recognize a broader set of protein and complement-associated ligands.^4,24

Resting integrins are maintained predominantly in low-affinity conformations. Chemokine and other G-protein-coupled receptor signals activate intracellular pathways that recruit talin and kindlin to the β-chain cytoplasmic tail, destabilizing the α/β tail association and favoring ectodomain extension, headpiece opening, and clustering. Talin and kindlin are not simply structural adaptors; their coordinated engagement converts a rapidly reversible receptor into a mechanically competent adhesion system.²⁵ For leukocytes exposed to blood flow, affinity regulation and lateral organization are both important: a cell may express abundant surface integrin yet remain poorly adhesive if the receptor is not in an appropriate conformational or nanoscale state.

Outside-in signaling begins once ligand-bound integrins bear force and assemble signaling complexes. β2-integrin engagement can activate Src-family kinases and Syk-associated pathways, PI3K-AKT signaling, MAPKs, and Rho-family GTPases, with downstream effects on actin polymerization, cell polarity, spreading, migration, phagocytic architecture, reactive oxygen species, and transcriptional programs.²⁴ The exact pathway is strongly stimulus- and cell-dependent, which is why generic statements that an integrin “activates FAK/PI3K/Rho” can be misleading. In leukocytes, integrin signaling is best viewed as an input integrated with chemokine, Fc-receptor, complement, Toll-like-receptor, and cytokine signals.

This bidirectionality also creates a conceptual distinction that is central to interpreting atherosclerosis studies: expression is not activation. Flow cytometry or transcriptomics can identify receptor abundance, but neither approach by itself establishes high-affinity conformation, ligand occupancy, mechanical load, or signaling competence. Studies that use conformation-sensitive probes, functional adhesion under flow, receptor-specific blockade, or genetic perturbation therefore provide different levels of mechanistic evidence than studies based only on expression.

Integrin-Mediated Mechanotransduction in Atheroprone Arteries

The arterial endothelium is continuously exposed to shear stress, cyclic stretch, and matrix-dependent mechanical cues. Sustained unidirectional laminar flow generally promotes an atheroprotective endothelial program, whereas low, oscillatory, or otherwise disturbed flow at branch points and curvatures favors inflammatory activation. Integrins are embedded in this mechanosensory system because they connect extracellular matrix to focal adhesions and the actin cytoskeleton, allowing force and matrix composition to modify signaling output.^5–8

Fibronectin is particularly important in this setting. Compared with basement-membrane proteins, a fibronectin-rich subendothelial matrix sensitizes endothelial cells to oxidized LDL and disturbed-flow signals. α5β1 activation can recruit FAK-dependent pathways that augment NF-κB activation and VCAM-1 expression, while α5-dependent fibronectin fibrillogenesis further reinforces the inflammatory matrix.^5,6 In parallel, αvβ3 is a major mediator of flow-induced inflammatory signaling in experimental endothelial systems, including NF-κB-dependent ICAM-1 and VCAM-1 expression.⁷ These pathways are overlapping but not identical, and their relative contribution depends on the stimulus and matrix environment.

More recent work has refined this model. Oscillatory shear can drive α5β1 trafficking to membrane microdomains through a pathway involving Piezo1-dependent calcium influx, PTP1B, and annexin A2; perturbing this axis reduces endothelial activation and plaque development at atheroprone sites in ApoE-deficient mice.⁸ In 2025, fibronectin-binding α5 and αv integrins were also linked to endoplasmic-reticulum stress induced by disturbed flow or oxidized LDL, with a JNK-c-Jun inflammatory output that was separable from canonical NF-κB activation.⁹ These observations expand integrin mechanotransduction beyond a single FAK-NF-κB pathway.

Mechanotransduction by endothelial integrins should be distinguished from force-dependent regulation of leukocyte integrins. The former determines whether the endothelial surface becomes adhesive and chemokine-rich; the latter determines whether a circulating leukocyte can convert transient contacts into shear-resistant arrest, crawling, and transmigration. The systems meet at the vessel wall: disturbed-flow signaling increases endothelial ligands such as VCAM-1 and ICAM-1, while chemokines and mechanical loading activate VLA-4 and β2 integrins on leukocytes. This coupling helps explain why leukocyte recruitment is spatially concentrated at atheroprone sites even when systemic inflammatory and lipid signals are widespread.

Integrin-Dependent Leukocyte Recruitment: A Cooperative Rather Than Linear Cascade

Leukocyte recruitment is often illustrated as a linear cascade, but the underlying receptor biology is better described as a network of overlapping checkpoints. Selectins mediate much of the initial tethering and rolling, while chemokines presented on activated endothelium trigger rapid integrin activation. Activated α4β1 can support both rolling-related interactions and firm monocyte adhesion to VCAM-1, whereas LFA-1 and Mac-1 contribute to arrest, adhesion strengthening, crawling, and transmigration.^2,3,11 The relative order of these events is therefore not fixed across leukocyte subsets or vascular beds. The overlapping steps of the leukocyte adhesion cascade, together with the principal ligand interactions and functional specialization of β2 integrins and VLA-4, are illustrated in Figure 2. The context-dependent contributions of these integrins across leukocyte recruitment, plaque development, inflammatory-cell retention, and thromboinflammatory progression are summarized in Table 2.

Figure 2.

Integrin architecture, ligand recognition, and functional specialization during leukocyte recruitment in atherosclerosis. (A) Schematic representation of the multistep leukocyte adhesion cascade. Initial capture and rolling are predominantly mediated by selectin–ligand interactions, whereas endothelial chemokines trigger inside-out activation of leukocyte integrins, promoting increased affinity and avidity for endothelial ligands. The symbol ↑ denotes an increase or upregulation in the indicated molecular expression, signaling activity, or biological process, as applicable. Activated LFA-1 (αLβ2; CD11a/CD18), Mac-1 (αMβ2; CD11b/CD18), and VLA-4 (α4β1) cooperate during firm arrest, adhesion strengthening, intraluminal crawling, and transendothelial migration. (B) Major leukocyte integrins and representative ligands and functions relevant to atherosclerosis. LFA-1 primarily recognizes ICAM family members and contributes to firm adhesion, transmigration, and immune-cell interactions; Mac-1 binds multiple endothelial, platelet, complement, fibrinogen, and extracellular-matrix ligands and links leukocyte adhesion to phagocytic and thromboinflammatory signaling; CR4 (αXβ2; CD11c/CD18, historically termed p150,95) participates in complement-dependent recognition, phagocytosis, antigen-presenting-cell functions, and inflammatory-cell retention; αDβ2 (CD11d/CD18) is associated particularly with retention of activated monocytes/macrophages in chronically inflamed tissues; and VLA-4 binds VCAM-1 and fibronectin and supports leukocyte arrest and transmigration under inflammatory flow conditions. Chemokine-induced inside-out signaling through Rap1, RIAM, talin, and kindlin promotes integrin extension, affinity maturation, and clustering, whereas ligand engagement initiates outside-in signaling through pathways including Src-family kinases/FAK, PI3K–AKT, MAPKs, and Rho-family GTPases, thereby regulating cytoskeletal remodeling, migration, survival, cytokine production, and phagocytic responses. The depicted functions are overlapping and context dependent and should not be interpreted as a rigid sequential hierarchy.

Abbreviations: CR4, complement receptor 4; ECM, extracellular matrix; FAK, focal adhesion kinase; GPIbα, glycoprotein Ib alpha; ICAM, intercellular adhesion molecule; LFA-1, lymphocyte function-associated antigen-1; Mac-1, macrophage-1 antigen; MAPK, mitogen-activated protein kinase; PI3K, phosphoinositide 3-kinase; PSGL-1, P-selectin glycoprotein ligand-1; RIAM, Rap1-GTP-interacting adaptor molecule; VCAM-1, vascular cell adhesion molecule-1; VLA-4, very late antigen-4.

Table 2.

Context-Dependent Functions of Integrins During Atherogenesis

Atherosclerotic Context	Integrin(s)	Representative Ligand/Partner	Primary Function	Likely Disease Consequence	Evidence/Caveat
Atheroprone endothelial activation	Endothelial α5β1, αvβ3	Fibronectin; focal-adhesion signaling	Mechanosensing and inflammatory signaling	Increases VCAM-1/ICAM-1 expression and creates a permissive recruitment niche	Strong experimental evidence; these are endothelial rather than leukocyte integrins. Refs5–9
Tethering and rolling	VLA-4 (α4β1) can contribute; β2 integrins are activated downstream	VCAM-1; selectin/chemokine systems act in parallel	Slowing and transition toward arrest	Increases the probability of stable monocyte capture	Integrins should not be portrayed as the only initial tethering system. Refs3,11
Firm arrest/adhesion strengthening	LFA-1, VLA-4; Mac-1 contributes contextually	ICAM-1/2; VCAM-1; additional Mac-1 ligands	Shear-resistant adhesion and adhesion strengthening	Permits recruitment at activated arterial sites	Functions overlap and vary by leukocyte subset, ligand density, and vascular bed. Refs10–12
CR4-dependent monocyte recruitment	CR4 (αXβ2) + VLA-4 (α4β1)	VCAM-1-centered cooperative signaling	CR4-to-VLA-4 cross-talk and firm capture	Preferential recruitment of lipid-associated/foamy monocytes in experimental hyperlipidemia	CR4 can amplify VLA-4 function but is not simply interchangeable with it. Refs13–15,26
Intraluminal crawling/transmigration	LFA-1, Mac-1, VLA-4	ICAM-family ligands; VCAM-1; junctional partners	Cytoskeletal polarization, crawling, and endothelial crossing	Facilitates intimal entry	Relative contributions depend on leukocyte subset and vascular context; avoid a rigid sequence. Refs2,3,24
Plaque macrophage persistence	αDβ2; Mac-1 and matrix receptors may contribute	Fibrin and extracellular-matrix ligands	Adhesion and retention within inflamed tissue	Sustains macrophage burden and chronic inflammation	Retention evidence for αDβ2 is strongest in experimental mouse models. Ref17
Phagocytic function	β2 integrins, including complement receptors	iC3b/opsonized cargo; cytoskeletal machinery	Mechanical control of engulfment and phagocytic architecture	May influence plaque macrophage function	Direct independent roles in plaque efferocytosis remain unresolved. Refs20,27
Platelet–leukocyte thromboinflammation	Mac-1 with selectin-dependent tethering	Platelet GPIbα; P-selectin–PSGL-1 acts upstream/parallel	Stable platelet–leukocyte adhesion and signaling	Links inflammation to thrombosis at disrupted vascular surfaces	P-selectin is not a direct Mac-1 ligand in this interaction. Refs18,19
Advanced plaque phenotype	Multiple integrin networks	Endothelial, matrix, leukocyte, and platelet partners	Recruitment, retention, signaling, and thrombosis	May influence necrotic-core burden, inflammation, cap remodeling, and thrombogenicity	Lesion area should not be equated with plaque instability. Refs27,28

Notes: Leukocyte recruitment is multistep but not rigidly sequential. LFA-1, Mac-1, CR4, and VLA-4 can act concurrently or partially compensate for one another depending on leukocyte subset, vascular bed, ligand density, and inflammatory stage. αDβ2 is presented primarily as a retention-associated receptor.

Abbreviations: CR4, complement receptor 4; GPIbα, glycoprotein Ib alpha; ICAM, intercellular adhesion molecule; PSGL-1, P-selectin glycoprotein ligand-1; VCAM-1, vascular cell adhesion molecule-1; VLA-4, very late antigen-4.

Direct arterial imaging supports an important role for LFA-1 in shear-resistant leukocyte adhesion: LFA-1 deficiency markedly reduces adhesion in mouse aorta and femoral artery.¹⁰ In early atherosclerotic carotid lesions, VLA-4 blockade strongly reduces monocyte adhesion, and both VCAM-1 and the connecting segment-1 region of fibronectin can function as α4β1 ligands.¹¹ These classic experiments remain mechanistically informative because they measure function under flow rather than receptor expression alone.

Recent genetic evidence further argues against a simple hierarchy. In ApoE-deficient mice, myeloid talin1 deletion abolished β2-integrin-mediated monocyte adhesion in flow assays yet left α4β1-dependent adhesion intact; unexpectedly, the deletion increased rather than decreased atherosclerotic lesion formation and macrophage deposition.¹² This result illustrates two principles. First, α4β1 can preserve trafficking when β2 activation is impaired in some settings. Second, disrupting a core integrin activator may alter inflammatory cell state in ways that are not predictable from adhesion alone. Therapeutic models based solely on “less adhesion equals less atherosclerosis” are therefore incomplete.

LFA-1 (αLβ2): Arrest, Adhesion Strengthening, and Immune Signaling

LFA-1 is the most specialized β2 integrin for ICAM-dependent adhesion and is widely expressed across leukocyte populations. During recruitment, rapid inside-out activation converts LFA-1 from a low-affinity receptor into a high-affinity, force-bearing adhesion system capable of stabilizing leukocytes against arterial shear. Its interaction with endothelial ICAM-1 is particularly important during the transition from rolling to arrest and during subsequent locomotion on the endothelial surface.^3,10

The atherosclerosis-specific evidence base for LFA-1 is strongest for recruitment rather than for a unique plaque-resident signaling program. This distinction matters. Increased LFA-1 expression on circulating cells or lesional leukocytes may accompany inflammation, but the causal evidence is more persuasive when LFA-1 is genetically deleted or functionally blocked and arterial adhesion is directly measured.¹⁰ The receptor should therefore be described as a key contributor to leukocyte arrest and trafficking, not as an isolated master regulator of plaque instability.

LFA-1 also has functions beyond vascular anchoring. In lymphocytes it contributes to immune-synapse organization, and in myeloid cells outside-in signaling reorganizes actin and cell polarity.²⁴ These functions are biologically relevant to atherosclerosis, which contains activated T cells and diverse myeloid states, but direct demonstration that LFA-1 signaling independently determines those states inside human plaques remains limited. The most defensible interpretation is that LFA-1 couples recruitment to subsequent signaling competence, while the direction and magnitude of that signaling depend on the broader receptor environment.

Mac-1 (αMβ2): Adhesion, Outside-in Signaling, and Platelet-Leukocyte Thromboinflammation

Mac-1 is a highly promiscuous β2 integrin expressed prominently on neutrophils and monocytes/macrophages. Its ligand repertoire includes ICAMs, fibrinogen and other matrix or soluble proteins, complement-derived ligands, and platelet GPIbα. This breadth allows Mac-1 to reinforce endothelial adhesion and crawling while also connecting leukocytes to complement, extracellular matrix, and platelets.^18,24

The platelet interaction requires particular precision because two adhesive modules are often conflated. Initial platelet-leukocyte tethering is classically supported by platelet P-selectin binding leukocyte PSGL-1; firm interaction can then be strengthened by Mac-1 binding platelet GPIbα, with fibrinogen-dependent bridging possible in some settings. Ehlers et al mapped a GPIbα-binding region within the Mac-1 αM I-domain, providing direct molecular evidence for this counter-receptor pair.¹⁸ Thus, P-selectin should not be described as a direct Mac-1 ligand in this context.

The Mac-1-GPIbα axis also has functional consequences for thrombosis. In mouse vascular-injury models, genetic disruption of Mac-1 or its GPIbα-binding interface delayed thrombosis, whereas targeted interruption of the interaction reduced thrombus formation without the same hemostatic phenotype expected from direct platelet αIIbβ3 blockade.¹⁹ These experiments support a mechanistic link between leukocyte adhesion and thrombosis, but they do not by themselves prove that chronic Mac-1 inhibition will stabilize human atherosclerotic plaques.

Within plaques, Mac-1 signaling can affect migration, phagocytic responses, and inflammatory activation. The receptor’s broad ligand repertoire is therefore both an opportunity and a therapeutic problem: inhibiting one ligand-specific interaction may preserve more immune function than blocking the entire αMβ2 receptor. This provides a strong rationale for interface-selective or conformation-selective strategies rather than pan-Mac-1 antagonism.

CR4/CD11c (αXβ2): Hyperlipidemic Monocyte Arrest, Integrin Cross-Talk, and Phagocytic Biology

CR4 (αXβ2; CD11c/CD18) is expressed by dendritic cells and subsets of monocytes, macrophages, and lymphocytes. In atherosclerosis, its most compelling mechanistic evidence comes from hyperlipidemia-associated monocyte phenotypes. CD11c deficiency in ApoE-deficient mice reduced firm monocyte arrest, lesional monocyte/macrophage accumulation, and atherosclerosis, linking the receptor to recruitment rather than merely using CD11c as a lineage marker.¹³

Human translational observations are directionally consistent. After a standardized high-fat meal, circulating monocytes displayed increased CD11c surface expression and greater arrest on VCAM-1 under flow, with these changes tracking the transient rise in triglycerides.¹⁴ In severe experimental hypercholesterolemia, lipid-containing circulating monocytes are enriched for CD11c and contribute to nascent lesion formation.²⁶ These data support a model in which metabolic cues alter the adhesive phenotype of circulating monocytes before they enter the arterial wall.

A particularly important contribution of CR4 is its cross-talk with VLA-4. In foamy monocytes, high-affinity CD11c signaling can promote VLA-4 activation through Syk- and paxillin-associated signaling, thereby increasing shear-resistant capture on VCAM-1.¹⁵ This is a mechanistic example of integrin cooperation that is non-redundant: CR4 can prime or amplify α4β1 function rather than simply substitute for it. More recent work has identified a GM-CSF receptor-SYK-JNK-FOXO1-CD11c pathway that promotes macrophage migration and experimental atherosclerosis, further connecting CD11c expression to intracellular inflammatory signaling.¹⁶

CR4 is also a complement receptor and can participate in iC3b-dependent recognition and phagocytosis. However, this should not be equated automatically with efferocytosis of apoptotic cells in advanced plaque. β2 integrins can regulate phagocytic mechanics in macrophages,²⁰ but the best-established causal pathways for defective plaque efferocytosis involve dedicated corpse-recognition and signaling systems such as MerTK.²⁷ The current evidence therefore supports a role for CR4 in phagocytic biology while leaving its specific, independent contribution to atherosclerotic efferocytosis unresolved.

αDβ2/CD11d: Macrophage Retention and Chronic Plaque Inflammation

αDβ2 (CD11d/CD18) differs from LFA-1 and Mac-1 in that its atherosclerosis phenotype is most clearly linked to persistence of macrophages within inflamed tissue rather than to the initial capture of circulating leukocytes. In ApoE-deficient mice, CD11d was strongly upregulated on inflammatory macrophages; deletion of CD11d reduced aortic lipid deposition and lesional macrophage accumulation without an equivalent defect in initial monocyte recruitment. Adoptive-transfer experiments supported impaired retention or accumulation after entry into the tissue.¹⁷

These findings make αDβ2 mechanistically attractive because chronic macrophage residence can sustain cytokine production, protease activity, oxidative stress, and lipid handling even after the initial recruitment event has occurred. Yet the available evidence remains predominantly preclinical, and older “M1/M2” descriptions of CD11d biology should be interpreted cautiously. Contemporary single-cell studies demonstrate that plaque macrophages occupy multiple overlapping inflammatory, lipid-associated, proliferative, and remodeling states that are not captured by a binary polarization framework.^21–23

The therapeutic implication is therefore stage-specific. A receptor that primarily regulates retention could, in principle, be modulated after plaque initiation, whereas blockade of an arrest receptor would alter cell entry from the circulation. Whether αDβ2 inhibition would promote safe egress or beneficial inflammatory resolution in human plaques is unknown and requires direct testing.

VLA-4 (α4β1): VCAM-1-Dependent Recruitment and Non-Redundant Cross-Talk with β2 Integrins

VLA-4 (α4β1) is a β1 integrin with a central role in monocyte interactions with VCAM-1-rich endothelium. In early atherosclerotic lesions, VLA-4 blockade markedly reduced monocyte adhesion under flow, and VCAM-1 accounted for a substantial component of this interaction.¹¹ VLA-4 can also engage fibronectin connecting segment-1, giving it both endothelial and matrix-associated ligands. This makes α4β1 particularly well suited to regions where endothelial activation and provisional-matrix remodeling coexist.

It is nevertheless inaccurate to describe VLA-4 as the sole or universally dominant determinant of firm arrest. LFA-1, Mac-1, and VLA-4 can act concurrently, with dominance changing according to leukocyte subset, chemokine input, ligand density, and vascular bed. CR4-to-VLA-4 signaling in foamy monocytes provides one example of cooperation,¹⁵ whereas the 2024 myeloid talin1 study illustrates partial functional independence: α4β1-dependent adhesion persisted despite loss of β2-integrin activation.¹²

This combination of cooperation and non-redundancy is a central theme of the review. Cross-talk does not mean interchangeability. An intervention directed against one receptor may be compensated by another at the level of trafficking, while simultaneously altering outside-in signaling, cell survival, or inflammatory state. The net effect on atherosclerosis may therefore diverge from the immediate effect observed in an adhesion assay.

Integrins Beyond Recruitment: Immune-Cell State, Phagocytosis, Efferocytosis, and Plaque Heterogeneity

Once leukocytes enter the arterial wall, integrins continue to shape cell behavior through mechanical coupling to the extracellular environment and through outside-in signaling. β2-integrin engagement can influence actin architecture, migration, reactive oxygen species, phagocytosis, and transcriptional responses.^20,24 In 2024, Settle et al showed that macrophage β2 integrins function as a mechanical checkpoint during phagocytosis, matching cargo rigidity to distinct engulfment architectures.²⁰ Although that study was not an atherosclerosis experiment, it provides contemporary evidence that β2 integrins can regulate immune-cell function independently of vascular arrest.

Efferocytosis deserves separate treatment because defective clearance of apoptotic cells is a major driver of advanced plaque necrosis. Cleavage-resistant MerTK improves efferocytosis, reduces necrotic-core formation, and increases fibrous-cap thickness in experimental atherosclerosis; human carotid data also link MerTK cleavage with necrosis and symptoms.²⁷ These data establish the importance of efferocytosis for plaque phenotype, but they do not demonstrate that CR4 is the dominant efferocytosis receptor in plaques. The revised mechanistic model therefore treats CR4-mediated complement phagocytosis and plaque efferocytosis as related but distinct processes.

High-dimensional human plaque studies have further changed how integrin expression should be interpreted. Single-cell immune profiling identified distinct activated T-cell and macrophage states in symptomatic versus asymptomatic carotid plaques, while single-cell transcriptomic and epigenomic mapping revealed substantial heterogeneity among endothelial, smooth-muscle, lymphoid, and myeloid populations.^21,22 More recent multimodal profiling resolved 25 cellular populations, including multiple inflammatory and lipid-associated macrophage states.²³ These datasets are valuable for identifying which cells express integrin subunits and ligands, but transcript abundance alone cannot establish receptor affinity, clustering, ligand occupancy, or mechanical force. Functional integrin mapping will require pairing single-cell/spatial data with conformation-sensitive or ligand-engagement measurements.

From Plaque Growth to Plaque Instability and Atherothrombosis

A recurrent problem in the integrin literature is the tendency to infer “plaque instability” from a change in lesion area. These are not equivalent outcomes. Plaque vulnerability is determined by composition and structure - including necrotic-core burden, fibrous-cap integrity, collagen and smooth-muscle content, inflammatory protease activity, endothelial disruption, and thrombogenic surface exposure - rather than size alone.^1,27 Integrin studies should therefore report and interpret the specific plaque phenotype they measure.

β2-integrin pathways can influence plaque composition through several routes. Continued recruitment and retention increase the number of inflammatory myeloid cells; outside-in signaling can modify their effector programs; defective clearance of dying cells can enlarge the necrotic core; and platelet-leukocyte interactions can amplify thromboinflammatory signaling at disrupted surfaces. Mac-1-GPIbα is especially relevant at the inflammation-thrombosis interface because it stabilizes leukocyte-platelet interactions and contributes to thrombosis in experimental injury models.^18,19

Human spatial transcriptomics reinforces the importance of geography and mechanics. In carotid plaques, rupture-prone proximal and stenotic regions show distinct transcriptional programs and structural features compared with distal regions, linking local biomechanics to plaque vulnerability.²⁸ This finding does not prove that leukocyte integrins cause rupture, but it strengthens the rationale for studying integrin activation in spatially defined plaque niches rather than treating the lesion as a uniform tissue.

Human Genetics and Clinical Relevance

Human genetics offers two different kinds of evidence relevant to integrin targeting. Rare loss-of-function mutations can reveal what is biologically indispensable, whereas common variants may indicate whether modest lifelong perturbation changes cardiovascular risk. Severe ITGB2 mutations cause leukocyte adhesion deficiency type 1, characterized by profound defects in leukocyte adhesion and host defense.²⁹ This does not establish atherosclerosis protection, but it is an important safety signal: broad, durable suppression of β2-integrin function is unlikely to be immunologically neutral.

Evidence that common leukocyte-integrin variants materially alter atherosclerotic cardiovascular risk is much weaker. A small Turkish case-control study reported associations involving an LFA-1 variant and coronary heart disease,³⁰ but candidate-gene studies of this size are vulnerable to population stratification, multiple testing, and non-replication. At present, it would be premature to cite such observations as robust human genetic validation of LFA-1 or ITGAM as cardiovascular drug targets. Large-scale, ancestry-diverse genetic analyses linked to functional integrin phenotypes remain an important unmet need.

Therapeutic Targeting: Lessons from Clinical Integrin Inhibitors

The attractiveness of integrin targeting is obvious: these receptors sit at a physical and signaling checkpoint between circulating leukocytes and inflamed tissue. The clinical record, however, argues strongly against assuming that successful blockade in experimental models will translate directly to cardiovascular disease. Two randomized trials of anti-CD18/CD11-CD18 antibodies in acute myocardial infarction - LIMIT-AMI and HALT-MI - failed to reduce infarct size or improve the measured cardiac end points despite a strong preclinical rationale.^31,32 These trials addressed acute reperfusion injury rather than chronic atherosclerosis, but they remain an instructive example of the translational gap between leukocyte-adhesion biology and clinical efficacy.

Conversely, α4-directed therapy demonstrates that integrins are druggable. Natalizumab, an α4-integrin antagonist, substantially reduces inflammatory disease activity in relapsing multiple sclerosis,³³ yet systemic interference with α4-mediated trafficking carries clinically important infectious risk, including progressive multifocal leukoencephalopathy observed in natalizumab-treated patients.³⁴ Vedolizumab provides a different lesson: targeting α4β7 produces gut-selective anti-trafficking activity in ulcerative colitis, demonstrating that tissue-selective integrin blockade can preserve efficacy while narrowing the anatomical footprint of immune interference.³⁵ Neither drug is evidence for treating atherosclerosis with α4 blockade; rather, together they define the efficacy-safety problem that cardiovascular strategies must solve.

For atherosclerosis, the most plausible next generation of approaches will therefore be more selective than pan-integrin inhibition. Options include blocking a disease-relevant ligand interface, stabilizing a non-adhesive conformation, interfering with a context-specific intracellular signaling node, transient inhibition during high-risk inflammatory windows, or vascular/plaque-directed delivery. The Mac-1-GPIbα interface is an example of a ligand-specific concept with preclinical antithrombotic activity.¹⁹ Likewise, endothelial α5β1 signaling offers the possibility of modulating the atheroprone vascular niche upstream of leukocyte arrest rather than chronically disabling leukocyte β2 function.^5,6,8 These strategies remain preclinical and require careful host-defense and vascular-safety testing.

Controversies, Knowledge Gaps, and Future Directions

The first unresolved issue is redundancy versus specialization. LFA-1, Mac-1, CR4, αDβ2, and VLA-4 share portions of the recruitment cascade, yet experiments increasingly show that they are not interchangeable. VLA-4 can maintain monocyte adhesion when β2 activation is disabled in some settings, while CR4 can signal VLA-4 activation.^12,15 A therapeutic intervention may therefore shift traffic through a parallel pathway rather than simply reduce total recruitment.

Second, integrin effects are stage- and cell-dependent. Blocking an arrest receptor during lesion initiation is biologically different from inhibiting a retention receptor in an established plaque, and the same integrin can signal differently in a neutrophil, monocyte, macrophage, dendritic cell, or T cell. Studies should define both the disease stage and the cellular compartment being targeted. The unexpected acceleration of atherosclerosis after myeloid talin1 deletion is a reminder that perturbing a shared activation machinery can alter inflammatory biology beyond adhesion.¹²

Third, mechanistic plausibility is not equivalent to causal validation. A recent study found that vascular smooth-muscle- and myeloid-derived α9β1 was not directly required for atherosclerosis development in mice, despite reasonable biological arguments for involvement.³⁶ Negative studies of this type are valuable because they test whether an attractive adhesion pathway is actually rate-limiting in disease.

Fourth, current human datasets measure integrin state incompletely. Single-cell and spatial technologies can localize ITGAL, ITGAM, ITGAX, ITGAD, ITGA4, ITGB1, and ligand transcripts across plaque niches, but integrin function is heavily post-translational.^21–23,28 Future studies should combine spatial transcriptomics with conformation-sensitive antibodies, ligand-mimetic probes, imaging of receptor clustering, or force-sensitive reporters to map where integrins are functionally engaged in vivo.

Finally, translation requires biomarkers. An integrin-directed therapy is unlikely to be appropriate for all patients with atherosclerosis. Candidate biomarkers could include imaging of vascular inflammation, circulating platelet-leukocyte aggregates, conformation-specific integrin epitopes, soluble ligand profiles, or molecular signatures of plaque macrophage recruitment and retention. These markers must be validated prospectively and linked to a drug mechanism, not merely correlated with disease severity.

A practical research agenda follows from these gaps: define the spatiotemporal activation state of each integrin in human lesions; separate adhesion-dependent from signaling-dependent effects using ligand- and pathway-specific perturbations; test interventions in established as well as early disease; quantify host-defense consequences; and develop cell- or vascular-selective delivery systems. Such work would move the field from descriptive expression maps toward clinically actionable integrin biology.

Conclusion

β2 integrins and VLA-4 form a cooperative but non-redundant control system for leukocyte recruitment in atherosclerosis. LFA-1 provides rapid ICAM-dependent arrest and adhesion strengthening; Mac-1 expands the adhesive network to matrix, complement, and platelet GPIbα; CR4 links hyperlipidemic monocyte phenotypes to VLA-4 activation and recruitment; αDβ2 is strongly implicated in macrophage retention; and VLA-4 provides a major VCAM-1-dependent route for monocyte entry into atheroprone regions. These pathways operate within an endothelial environment whose adhesiveness is itself shaped by matrix-sensitive integrin mechanotransduction.

The therapeutic implication is more nuanced than simply blocking leukocyte adhesion. β2 integrins also regulate phagocytic mechanics and intracellular signaling, α4 blockade can impair protective immune trafficking, and inhibition of shared activation machinery can generate unexpected inflammatory phenotypes. Moreover, human plaque heterogeneity means that the relevant integrin pathway may differ across disease stage and anatomical niche. The failure of broad anti-CD18 approaches in myocardial-infarction trials, contrasted with the efficacy and safety lessons from α4-directed therapies in other diseases, argues for selectivity rather than stronger global inhibition.

Future progress will depend on context-specific modulators, direct measurement of integrin activation in human plaques, spatially resolved mapping of receptor-ligand interactions, and biomarkers that identify patients in whom a defined integrin pathway is genuinely rate-limiting. Integrins remain compelling targets not because they are universal switches for atherosclerosis, but because they sit at a tractable interface between mechanics, immunity, and thrombosis - an interface that can be exploited only if its cell- and stage-specific biology is respected.

Scope and Limitations

This is a mechanistic narrative review centered on leukocyte β2 integrins and VLA-4. Endothelial, platelet, and vascular smooth-muscle integrins are discussed when they alter the leukocyte adhesion environment, plaque composition, or thromboinflammatory signaling. The review does not attempt a systematic quantitative synthesis and does not infer human clinical efficacy from preclinical blockade studies. A further limitation is that much of the causal evidence for individual leukocyte integrins derives from mouse genetics, antibody blockade, or reductionist flow systems; direct measurements of integrin activation state in human atherosclerotic plaques remain sparse.

Funding Statement

The authors report that this work received no specific financial support.

AI Declaration

During the revision of this manuscript, the authors used OpenAI ChatGPT (GPT-5.6 Sol; OpenAI, San Francisco, CA, USA) solely for English-language refinement, including improvement of grammar, sentence structure, readability, and linguistic clarity. The tool was used only to improve the presentation of text that had already been scientifically developed and finalized by the authors. It was not used to generate scientific content, formulate or interpret scientific arguments, select or evaluate references, analyze data, develop conclusions, or prepare the figures and tables.

All AI-assisted language edits were critically reviewed and approved by the authors. The authors independently verified the scientific content, terminology, references, tables, figures, and interpretations presented in the manuscript. The authors take full responsibility for the originality, accuracy, validity, and integrity of the entire content of the paper.

Data Sharing Statement

No new datasets were generated or analyzed as part of this narrative review. Data reported in the cited studies are available from the original publications and their associated repositories, where applicable.

Author Contributions

Hojat Shahraki: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Davood Bashash: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Mohammad Hossein Mohammadi: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Hamideh Kouhpeikar: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Mohammad Bahloli: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Saeedeh Bagheri: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Omolbanin Sargazi Aval: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Masoud Kargar: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Alireza Khiabani: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Javad Yasbolaghi Sharahi: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Mohammad Esmail Gheydari: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

Mohsen Hamidpour: Conceptualization, Investigation, Writing – original draft, Writing – review & editing.

All authors contributed equally to the work. All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare no competing interests in this study.