The CD44-HA Axis and Inflammation in Atherosclerosis: A Temporal Perspective

Abstract

Cardiovascular disease (CVD) due to atherosclerosis is a disease of chronic inflammation at both the systemic and the tissue level. CD44 has previously been implicated in atherosclerosis in both humans and mice. This multi-faceted receptor plays a critical part in the inflammatory response during the onset of CVD, though little is known of CD44’s role during the latter stages of the disease. This review focuses on the role of CD44-dependent HA-dependent effects on inflammatory cells in several key processes, from disease initiation throughout the progression of atherosclerosis. Understanding how CD44 and HA regulate inflammation in atherogenesis is key in determining the utility of the CD44-HA axis as a therapeutic target to halt disease and potentially promote disease regression.

Introduction

CVD due to atherosclerosis is the number one cause of death worldwide [1]. Lesion formation is triggered by endothelial dysfunction and subendothelial accumulation of oxidized low-density lipoprotein (oxLDL), which promotes the adhesion and extravasation of leukocytes and a sustained inflammatory response. Leukocytes may consume oxLDL to become foam cells, a cell type found in early lesions (fatty streaks) through the later stages of disease. High inflammatory cell content in atherosclerotic lesion correlates with an increase in susceptibility to rupture, which triggers thrombosis and myocardial infarction, whereas a paucity of inflammation combined with the - generation of lesional fibrotic caps is associated with lesion stabilization. It therefore follows that inflammation is a key process that should be targeted to halt the progression of atherosclerosis and prevent acute catastrophic events associated with this chronic disease [2]. A great body of evidence suggests that local tissue inflammation drives atherogenesis, however the molecular mechanisms that govern this are only partly understood [3]. It is crucial to gain a thorough understanding of the mechanisms that govern atherogenesis to support the rational design of novel therapies to reduce risk for a myocardial event [2]. It is important to consider signaling pathways that may be targeted to modulate inflammation in atherosclerosis, including but not limited to pathways involving NFκB, PPARγ, endothelial nitric oxide synthase (eNOS), and iNOS ; all of which are discussed in more detail later. These important signaling pathways may modulate the inflammatory milieu of lesion, and have previously been targeted to halt the progression of atherosclerosis [4].

While these are well-established pathways involved in the progression of atherosclerosis, there is continued interest in the discovery of novel therapeutic targets to combat CVD. In recent years CD44 has emerged as a potential therapeutic target that may provide insights into local inflammation at the site of lesion. CD44 is a type 1 transmembrane receptor and its standard form is widely expressed on multiple cell types [5]. Under inflammatory conditions, CD44 is upregulated and functionally activated on vascular endothelial, smooth muscle and inflammatory cells [5]. CD44 modulates leukocyte adhesion, migration, and functional phenotype. It is a multi-faceted receptor, which exists in multiple activation states, variant isoforms, as well as intracellular and soluble forms [6]. CD44 is regulated with respect to the affinity for its primary ECM ligand, hyaluronan (HA). This regulation occurs via multiple mechanisms (alternative splicing of variant exons, post-translational modifications including glycosylation, sulfation, phosphorylation and clustering) all of which can modulate CD44’s capacity to bind HA [7,8].

HA is a polymeric non-sulfated glycosaminoglycan comprised of repeating units of n-acetyl-glucosamine and d-glucuronic acid. HA is a key biophysical component of extracellular matrices due to its hydrating viscoelastic physical properties. Furthermore, HA exists over a broad range of different molecular weights (MW) that have contrasting effects on cell behavior. Generally speaking, low molecular weight HA (LMW HA) tends to be pro-inflammatory, whereas high molecular weight HA (HMW HA) tends to have anti-inflammatory properties [9]. Control of HA MW occurs at both its anabolic and catabolic stages of metabolism. Significant research has gone into understanding the metabolic, biophysical, and signaling roles that HA exerts during homeostasis [10–13]. HA synthases (HAS 1-3) are responsible for de novo synthesis of HA of different MW. These synthases have been implicated in a variety of inflammatory conditions including rheumatoid arthritis, asthma, atherosclerosis and murine neointimal hyperplasia [14,15]. HA is degraded by multiple different processes; enzymatic digestion by hyaluronidases (HYALs) or by non-enzymatic oxidative processes in response to reactive oxygen or nitrogen species (ROS or RNS) that generate a pool of HA fragments of various size. Interestingly, increased HYAL activity is often a hallmark of pro-inflammatory conditions [9].

Both CD44 and HA play critical roles in atherogenesis. CD44 has been implicated in both human and in murine models of atherosclerosis. Gene co-expression analysis has revealed that CD44 is a critical mediator of murine atherosclerosis susceptibility [16]. It is expressed on all key cells that contribute to the progression of disease and is both upregulated and functionally activated in atheroprone regions of the vasculature [17–20]. Furthermore, CD44 alters gene expression in these areas prior to lesion development and even in the absence of any difference in cholesterol levels. This suggests that CD44 exerts its effects in response to the local inflammatory environment, not at the systemic level [20]. Since vascular CD44 is upregulated in response to both altered shear stress and hyperlipidemia (two key risk factors for CVD) [21,22] it implies that CD44 might be an early indicator of risk for the development of atherosclerosis. Such findings led to a focus on a potential role for CD44 in the initiation of CVD and paved the way for studies examining the role of CD44 in lesion burden in vivo. Global CD44 deletion resulted in a significant decrease in lesion burden compared to wild-type control mice on an atherogenic background (apoE^−/− mice) [18]. In order to delve into the cell types involved and the mechanism by which CD44 controls atherogenesis, bone marrow chimeras were generated using WT and CD44-null donor bone marrow transplantation into apoE^−/− and apoE^−/− CD44^−/− mice. CD44 expression in both the vascular and the bone marrow-derived compartments contributed to lesion development in the apoE^−/− mouse model, implying that CD44 on both resident as well as recruited cells may be required for CD44 to exert its full pro-atherogenic affect in vivo [17]. Furthermore, CD44 deletion also promoted an increase in fibrotic content in apoE^−/−CD44^−/− compared to apoE^−/− mice demonstrating that CD44 also regulates the composition of lesion [23].

Interestingly, global CD44 deletion on an alternative atherogenic LDLR^−/− background did not result in decreased lesion burden compared to wild-type littermate controls [24]. This apparent discrepancy between the two studies may be due to the difference in diets used in the two models. Specifically, development of atherosclerosis in LDLR^−/− requires a high fat diet (HFD), whereas disease develops spontaneously in apoE^−/− mice even if maintained on a regular chow diet, as was used in the studies in the model described above. Although both these mouse models develop hypercholesterolemia, cholesterol levels are remarkably higher in HFD-fed LDLR^−/− mice than chow-fed apoE^−/− mice and plasma lipid profiles are also distinct between the two models [25,26]. In support of the concept that differences in diet and lipid profiles might underlie the differences in CD44 deletion between these two models, maintenance of apoE^−/− mice on a HFD overwhelmed the impact of CD44 deletion on lesion burden (Zhao and Puré, unpublished data). This implies that the overwhelming inflammatory environment associated with a HFD can overcome any differences between CD44 wild-type and CD44-null mice regardless of model.

HA is a modulator of atherosclerotic lesion stability. It is highly upregulated at atheroprone regions in mouse models of atherosclerosis and may regulate smooth muscle cell migration, proliferation, leukocyte activation and lipid accumulation within the plaque to alter atherosclerotic plaque stability [17,20,27]. In human atheroma HA is present in both early and late lesion and an increase in HA accumulation correlates with lesion progression [28,29]. In addition, increased hyaluronic acid metabolism correlates with increased plaque destabilization in human atheroma [30]. Furthermore, both HA and its receptor CD44 accumulate at sites of plaque erosion in human ruptured lesions [31,32]. Since LMW HA and HMW HA often exert contrasting effects on diverse cell types in lesion, it is difficult to determine the net result of total HA content on lesion progression and therefore represents an important area for future study. There is also evidence that both HASs and HYALs may contribute to the progression of atherosclerosis. There is an increase in HYAL1 expression in complex human lesions compared to fibrous plaque, and one report observed that overexpression of HAS2 in smooth muscle cells promotes the development of atherosclerosis, potentially by altering vascular stiffness [33,34]. Moreover, it has been suggested that plasma HYAL activity may serve as a biomarker for patients with atherosclerosis [35].

There is a body of evidence showing crucial roles for both CD44 and/or HA in CVD [16–18,20,24,36–47]. Several studies have separately examined the role of CD44 and HA on inflammation in other pathological settings [48–71]. However, the specific roles of the CD44-HA axis on inflammation in atherosclerosis have not yet been comprehensively elucidated. CD44 exerts a prominent role in the onset of CVD, regulates lesion composition and is itself regulated by the inflammatory environment. In addition, CD44 on multiple-interacting cells may be required to exert a pro-atherogenic effect. Generally, previous work has focused on the role of CD44 as an adhesion receptor during the initiation of disease. Interestingly, the role of CD44 in the processes that govern mid- and advanced-stage disease has been under-explored. Particularly intriguing is the potential role that CD44 may regulate the balance between inflammation and fibrosis in plaque composition, which ultimately dictates risk for a myocardial event. Herein, we examine the potential impact of CD44 throughout the progression of disease, which may provide insights into the mechanisms by which CD44 affects atherogenesis and whether targeting the CD44-HA axis may be utilized to achieve disease regression.

Initiation of Disease Endothelial Dysfunction

Local tissue inflammation promotes the initiation and progression of disease, which can lead to lesion rupture and myocardial infarction [3]. Under homeostatic conditions, laminar flow and low LDL levels maintain the HA-rich glycocalyx, a protective carbohydrate network that prevents leukocyte adhesion to the endothelium (Figure 1A). However under atherogenic conditions (Figure 1B), hyperlipidemia and disturbed blood flow at curved areas and branched points in the vasculature together promote increased endothelial permeability and activation to promote leukocyte adhesion [2]. oxLDL may compromise glycocalyx integrity by increasing leukocyte adhesion, endothelial permeability, and pro-inflammatory gene expression, all of which support the progression of disease. oxLDL may also promote leukocyte adhesion by upregulating key adhesion molecules via LOX-1-CD40 dependent mechanisms [72,73] or increase the permeability of the endothelial cell layer via LDLR and caveolin 1-dependent pathways [74]. oxLDL may also directly degrade the glycocalyx since it has been shown that administration of oxLDL reduces the thickness of the EC layer and promotes leukocyte adhesion. This phenotype is reversed in response to superoxide dismutase (SOD), implying that the damage to the EC layer may be dependent on the presence of reactive oxygen species [75]. Since the glycocalyx also directly modulates the inflammatory micro-environment by docking with key enzymes including protective SOD (which quenches oxygen radicals and promotes the availability of protective NO), it follows that NO is one of the key factors that maintains and prevents destruction of the glycocalyx [76]. In contrast, several factors lead to alterations in glycocalyx integrity, including low shear stress, oxidized LDL, pro-inflammatory mediators (IL1-β, TNFα, LPS), matrix metalloproteinases and elastase [76].

Figure 1. The CD44 / HA axis in the initiation of atherosclerosis.

1A: During homeostasis, an intact, HMW HA-rich EC glycocalyx prevents adhesion of circulating leukocytes (monocytes and T cells). CD44 is inducible, but inactive. VSMCs are quiescent and contractile. 1B: Under pro-atherogenic conditions, high circulating levels of oxLDL promote EC activation and increased endothelial permeability by degrading the protective glycocalyx. EC adhesion molecule expression, including CD44, is increased, supporting adhesion of blood-borne leukocytes. Incoming active leukocyte CD44 bridges with its ligand HA on EC CD44 to facilitate leukocyte adhesion. The release of pro-inflammatory cytokines and chemokines (TNFα and IL-1β) also promote monocyte and T cell CD44 expression, as well as cell adhesion and extravasation into the sub-endothelium. VSMCs adopt a synthetic phenotype characterized by production of pro-inflammatory stimuli and increased ECM production.

Disruption of the glycocalyx barrier allows more LDL to access the subendothelial space to initiate the inflammatory response. Both HA and CD44 are critical to sustain robust vascular endothelial integrity. In the LDLR^−/− model of atherosclerosis, administration of 4-methylumbellieferone, an inhibitor of HA synthesis, resulted in increased leukocyte recruitment to the carotid artery by intravital imaging, increased macrophage recruitment to lesion and increased atherosclerosis, highlighting the critical anti-atherogenic role of an intact HA-rich glycocalyx [77]. LMW HA promotes endothelial permeability by inducing HABP2 activity and subsequent protease activated receptor (PAR) activation on endothelial cells [78]. In contrast, HMW HA-CD44 binding protects murine pulmonary endothelium from LPS-induced permeability [79]. In addition, endothelial HAS2 appears to be anti-atherogenic and is upregulated under atheroprotective conditions of high shear stress [80]. However, pro-inflammatory, pro-atherogenic cytokines may also induce endothelial cell HAS2 synthesis and subsequent HMW HA production, promoting CD44-HA mediated adhesion of monocytes to human activated endothelium under static conditions [81]. These conflicting studies highlight the need for additional studies to better ascertain the net effect of endothelial HAS2 on the disease to better-determine whether this enzyme may serve as a potential therapeutic target to promote endothelial cell integrity at the initiation of atherosclerosis.

One of the critical signaling pathways that alters inflammation of ECs during the initiation of disease is the NFκB pathway. NFκB is known be critical for promotion of pro-inflammatory mediator production in immune cells - T cells, macrophages and vascular endothelial cells [82]. The NFκB signaling pathway is comprised of NFκB heterodimers or homodimers that contain the Rel homology domain containing polypeptides and their inhibitor proteins, the IκBs. NFκB activation requires proteasomal degradation of the IκBs. The IκB Kinase (IKK) complex phosphorylates and thus inactivates IκB [83]. In vascular endothelial cells NFκB signaling promotes pro-inflammatory cytokine production and upregulation of critical adhesion molecules, both of which promote atherogenesis [84]. In vivo, it has been shown that deletion of NEMO (IKKγ) in endothelial cells inhibits monocyte recruitment and lesion burden in apoE^−/− mice on a HFD [84].

Leukocyte Adhesion to Activated Endothelium

Concomitant with endothelial dysfunction, a pro-inflammatory environment promotes the production of cytokines/chemokines and the subsequent upregulation of key adhesion molecules ICAM-1, VCAM-1, P-selectin, E-selectin and CD44 to support leukocyte adhesion and extravasation into the subendothelial space [2]. The role of CD44 in lesion initiation is in part attributed to its ability to mediate the interaction between incoming leukocytes and activated endothelial cells. Specifically, HA associated with the luminal surface via association with endothelial CD44 serves as a bridge binding to CD44 expressed on activated leukocytes to facilitate leukocyte adhesion and transmigration (Figure 1B).

Progression of Disease

While the role of CD44 in the initiation of disease is well established, there have been relatively few studies examining CD44 during the later stages of atherosclerosis. Atherosclerosis progresses due to the actions of multiple cells types working inconcert to promote a sustained pro-inflammatory environment (Figure 2) [2]. eNOS is also a critical anti-atherogenic target in atherogenesis since NO regulates arterial tone and suppresses proliferation of vascular smooth muscle cells (VSMCs) [85]. It is important to understand how to maintain eNOS activity. Dysregulation of eNOS results in decreased NO synthesis/activity, which enhances the progression of atherosclerosis [86]. It has also been shown that administration of L-arginine, a precursor of nitric oxide, may ameliorate the severity of atherosclerosis [87].

Figure 2. Role of CD44 in the inflammatory response and impact on the progression of atherosclerosis.

The impact of CD44 on the progression of the atherosclerosis remains under-explored. CD44 may influence the balance between prolonged inflammation versus potential anti-inflammatory, disease-resolving processes through varying mechanisms on different cell types. Pro-inflammatory cytokines TNFα and IL-1β increase M1 macrophage CD44 bioactivity and LMW HA binding, leading to increased pro-inflammatory gene expression. M2 macrophages, localized predominantly in the adventitia, contribute to the resolution of inflammation via the production of anti-inflammatory cytokines, although the bioactivity of CD44, or the affect of HMW HA versus LMW HA binding in these cells is unknown. The expression, activity and role of CD44/HA on Mox macrophages (a separate anti-inflammatory macrophage subset in lesions) has not been investigated. TNFα and IL-1β rapidly induce expression and bioactivity on T cells, in turn prolonging the inflammatory response. Conversely, T-reg CD44 expression is critical for the production of the anti-inflammatory cytokines TGFβ and IL-10. CD44-LMW HA binding promotes foam cell formation and consequent CD36 expression and increased uptake of oxLDL. In turn, this facilitates foam cell release of pro-inflammatory cytokines TNFα and IL-1β. CD44 activity on dendritic cells is required for dendritic cell-induced T cell activation, although this has not yet been confirmed in the context of atherosclerosis.

Since NO production is a key factor in modulating the inflammatory environment, it is important to consider how CD44 interacts with these known targets to combat atherosclerosis. CD44 is known to associate with cholesterol-rich microdomains termed lipid rafts. It has been reported that HA binding to endothelial CD44v10 promotes engagement with ankyrin and IP3 receptor into lipid rafts, which promotes enhanced NO production which may be due to eNOS activity[88]. This therefore represents a potential avenue to modulate endothelial cell phenotype, and subsequent effects on VSMCs in the inflammatory micro-environment.

Regulating the Lesion Milieu: Pro-inflammatory versus pro-fibrotic functions of infiltrating leukocytes

Macrophages are key leukocytes that drive atherogenesis as they exert critical affects throughout the course of disease. An abundance of macrophages is associated with lesions containing thin fibrotic caps relative to more stable, complex fibrotic lesions [89]. After blood-borne monocytes firmly adhere and extravasate into the subendothelium, these cells differentiate into mature tissue macrophages in the presence of macrophage colony stimulated factor (MCSF) produced by the activated endothelium [90]. Macrophages express receptors for oxLDL that mediate signal transduction to regulate cell behavior and gene expression and function to internalize oxLDL leading to the formation of lipid-laden foam cells. Macrophages within lesion are heterogeneous, representing cells along a spectrum ranging from pro-inflammatory M1-like macrophages to anti-inflammatory M2-like macrophages. Notably, cells along this spectrum have been shown to exhibit plasticity in response to factors in the milieu which opens the intriguing possibility of therapeutically driving reprogramming of macrophages in the context of lesions [91,92]. The details of macrophage biology in atherosclerosis has been covered extensively elsewhere [93]. Briefly, M1 macrophages are characterized by the production of ROS and RNS, as well as pro-inflammatory cytokines including IL-1β, TNFα and IL-12 all of which have been shown to promote atherogenesis. Each of these has also been implicated in CD44 activation and/or downstream of CD44 to influence inflammatory gene expression, or both. IL-12 is of particular interest in this context as it has been shown that the CD44/HA axis is a potent inducer of IL-12 production that in turn drives the generation of type 1 inflammation and the production of interferon-γ – a potent macrophage activating and anti-fibrogenic factor whose function in part attributed to inhibition of VSMC proliferation and macrophage production of ECM degrading enzymes [94]. This may further amplify a pro-atherogenic inflammatory circuit that may potentially prevent the formation or cause the degradation of fibrotic caps that would tip the balance toward plaque destabilization [94]. M2-like macrophages are generated in response to IL-4, glucocorticoids, or IL-10 and TGFβ. M2 macrophages promote healing and contribute to the resolution of inflammation through production of anti-inflammatory cytokines [95]. Another anti-inflammatory macrophage subset (Mox) has been implicated in atherosclerosis and is generated in response to oxLDL administration in a Nrf2-dependent manner. These cells have distinct gene expression profiles compared to both the M1 and M2 macrophages and represent an intriguing avenue of future study [96]. Relatively-high M1 lesion content correlates with increased lesion vulnerability whereas relatively-high M2 lesion content correlates with lesion stability lesion in both human and mouse models of atherosclerosis [97]. In addition, different macrophage subsets exhibit distinct spatial localization within lesion - M1 macrophages populate the vulnerable shoulder region of lesion, M2s are found to be prominent in the adventitia, and Mox macrophages are found throughout lesion, as shown in Figure 2 ([98–100] and Monslow and Puré, unpublished data). Increasing the preponderance of M2 or Mox macrophages relative to M1s or decreasing the abundance of M1s at vulnerable shoulder regions of lesion may promote lesion stability. Thus, understanding how to regulate macrophage abundance, localization and functional phenotype by reprogramming macrophage subsets may serve as a key therapeutic opportunity to regulate disease progression.

NFκB signaling is critical for the development of the pro-inflammatory M1 macrophage phenotype [101]. Deletion of myeloid IκBα increases lesion burden in LDLR^−/− mice, while deletion of myeloid IκBβ reduces lesion burden in the LDLR^−/− mice fed a HFD. Thus, targeting the NFκB pathway may alter macrophage phenotype and regulate the inflammatory micro-environment [102,103].

As stated earlier, while eNOS is a potent anti-atherogenic target, other sources of NO may have pro-atherogenic effects. It has been shown that iNOS deletion decreases lesion burden in both male and female apoE^−/− iNOS^−/− mice compared to apoE^−/− mice[104]. iNOS is a well-established marker for M1 pro-inflammatory macrophages in both human and mouse lesion, a cell subset that as stated previously, is increased in unstable lesions [97,98,105]. This contrasting impact of NOS isoform on the progression of disease is complex and may reflect the low levels of basal NO produced by eNOS relative to the transient but markedly greater levels of NO produced by activation of iNOS. In any case, iNOS represents a pro-atherogenic target to potentially impede the progression of disease.

Peroxisome proliferator activated receptors (PPARs) also play a critical role in atherogenesis often by modulating macrophage phenotype. PPARs are lipid-derived nuclear hormone receptors that may regulate dyslipidemia and thus the course of atherosclerosis. Of particular note is PPARγ, which forms heterodimeric transcription factor with the retinoid X receptor (RXR), binding to distinct peroxisome proliferator elements (PPRE) in the promoter region of their target genes to alter lipid metabolism. It is expressed on multiple cell types - adipocytes, macrophages, hepatocytes, and skeletal muscle [106,107]. While scavenger receptor CD36 is one of the key target genes of PPARγ, PPARγ agonists may inhibit macrophage foam cell formation in vitro and in vivo [108–110]. PPARγ also inhibits monocyte production of key pro-inflammatory cytokines - IL-6, IL-1β, and TNFα, and thus may directly alter the lesional micro-environment [111,112]. In vivo the PPARγ agonist troglitazone has been shown to decrease carotid artery intimal thickness [113]. Thus, it is important to keep these signaling pathways in mind with regard to how they may specifically intersect with CD44.

As stated previously, ROS is a strong pro-atherogenic mediator. It degrades protective nitric oxide, leading to endothelial dysfunction and VSMC proliferation. It is thus critical to fully understand the role that ROS generation has on the CD44-HA axis. NADPH oxidase is a membrane-bound enzyme that produces ROS. It has been shown that CD44 immunostaining is significantly decreased in apoE^−/− mice with global NADPH oxidase deletion (apoE^−/−ph47^phox−/− mice) [47]. This decrease in CD44 staining was localized to macrophages in lesion. Coupled with the observation that apoE^−/− ph47^phox−/− mice had decreased lesion burden compared to apoE^−/− mice, this implies a potential regulatory role for NADPH on macrophage CD44 expression in an atherogenic setting, total macrophage abundance, and the progression of disease.

CD44 is critical for maximal macrophage recruitment and/or retention in atherosclerosis and other inflammatory conditions [17,18,114,115]. It may also regulate the pro-inflammatory state of monocytes and macrophages. Freshly isolated human monocytes required stimulation with either IFNy or LPS in short-term culture to increase both CD44 expression and to induce the high affinity state of CD44 [116]. Additional pro-atherogenic stimuli may also induce CD44 activity, including, IL-1α, IL-1β, IL-2, GMCSF, TNFα, and IL-4. These include a few of the key stimuli that govern macrophage differentiation and thus it will be interesting to determine whether different macrophage subsets have differential CD44 expression and/or activity under normal conditions as well as under pro-atherogenic conditions of atherosclerosis [117–119].

The balance between M1, M2 and Mox macrophages may influence disease progression, since the balance between inflammation and fibrosis is a critical determinant of plaque fate. The key M1 pro-inflammatory stimuli - IFNγ and TNFα as well as the M2 anti-inflammatory stimuli (IL-4, IL-10) alter CD44 activity of murine bone marrow-derived macrophages (levels of CD44 on Mox macrophages are currently unknown). CD44 activity can be further modified by post-translational modification of the receptor in a chondroitin sulfate-specific manner [119–121]. Thus, CD44 activity can be regulated on diverse macrophage subsets, and is thus a viable therapeutic target. This study paves the way for an in-depth examination of whether macrophage CD44 expression and regulation alters function of the subsets in the context of atherosclerosis.

Since HA promotes inflammatory cell retention and is also an important component of eroded atherosclerotic lesion, it is important to understand HA bioactivity and its impact on the key inflammatory cells in lesion. [32,54]. LMW HA promotes pro-inflammatory gene expression in murine macrophages in a CD44-dependent manner [122–125]. There is also evidence that suggests that this interaction is due to the signaling properties of LMW HA, and not solely due to its physical viscoelastic properties [123]. In contrast, HMW HA promotes anti-inflammatory gene expression profiles in both human and murine macrophages [125,126]. Thus, HA modulates macrophage inflammatory state and the status of macrophage activation can modulate CD44 activity. In addition, in macrophages NFκB signaling is downstream of CD44 binding and thus represents an important area for future study. CD44-LMW HA ligation leads to degradation of IκB and subsequent activation of NFκB and pro-inflammatory gene expression and iNOS activity [124,127–129]. Nevertheless, several questions remain regarding the impact of CD44 and HA on macrophages and how they may regulate lesion fate. Do CD44 and/or HA alter macrophage differentiation capacity in lesion? Can CD44 and/or HA be used to reprogram macrophage phenotype? Is CD44 expression or activity differentially regulated between macrophage phenotypes (and does this alter their functions in a subtype-specific manner)?

While macrophages are important inflammatory cells in lesion, T cells also modulate progression of disease. Macrophages recruit T cells to lesion by releasing TNFα. CD4+ T cells can then differentiate to a variety of subsets in response to different cytokines, found in both human and murine atherosclerosis. Th1 cells, driven in part by macrophage-derived IL-12, are pro-atherogenic; T-reg cells are atheroprotective, while Th2 cells (driven by the pro-fibrotic cytokines IL-4 and IL-13) may exert both pro-atherogenic and anti-atherogenic effects [130,131].

Cytokines and chemokines critical for atherogenesis can also rapidly induce increased expression of CD44 and its transition from a high-to-low affinity state for HA binding on T cells. This occurs under both static and shear stress conditions and may promote T cell adhesion and migration into inflamed sites in vivo [132–134]. In addition, CD44 on both bone marrow and non-bone marrow derived cells regulates T cell recruitment and retention in both the apoE^−/− and the LDLR^−/− models [23,24]. It has also been shown that deletion of HAS3 in apoE^−/− mice reduced Th1 cell polarization and atherosclerosis, supporting a pro-atherogenic role for HAS3 [135].

T cells can also adopt both pro-inflammatory and anti-inflammatory phenotypes and thus may influence the lesion microenvironment. One previous study reported that Th1 T cells produce a variety of pro-inflammatory cytokines, including IFNγ and TNFα, while T-regs produce anti-inflammatory, anti-atherogenic cytokines such as IL-10 and TGFβ [131]. T-regs exert a dominant role in atherogenesis relative to the other T cell subtypes, and CD44 is necessary for T-regs to produce TGFβ and IL-10 in wild-type mice [130,136]. In addition, CD44 deletion has been reported to decrease the production of IL-10 and IL-5 in a highly inflammatory milieu (LDLR^−/− mice on a HFD), suggesting that CD44 may exert an anti-atherogenic role on the dominant T cell population in lesion [24]. This contradictory anti-atherogenic role may in fact be dwarfed by the overwhelming pro-inflammatory role of CD44 on other key cells in lesion, resulting in a net pro-atherogenic role for CD44.

In contrast, VSMCs play seemingly paradoxical roles on lesion fate. Under basal conditions VSMCs exist in a contractile, low proliferative state in the media where they control the vascular tone of blood vessels [137]. During atherogenesis pro-inflammatory cytokines and oxLDL promote VSMC migration into the intima where they adopt a synthetic phenotype, as shown in Figure 1B [138,139]. Synthetic VSMCs produce large amounts of pro-inflammatory cytokines (MCP-1, IL-6) and increased extracellular matrix (ECM), notably collagen and fibronectin, which may help stabilize the plaque [70,140–143]. However, these ECM components and any derived matrix protein fragments may also exert pro-atherogenic effects. Macrophage adherence to type-I collagen induces macrophage phagocytosis of modified LDL; this interaction also increases MMP9 expression which degrades the protective fibrotic cap in lesion [144–146]. There is also some evidence that lesional VSMCs can phagocytose cholesterol to become foam cells in lesion [147]. Understanding the net role that these VSMC-derived foam cells play is complicated, since the increased ECM production and pro-inflammatory cytokine production from these cells may support both stabilization and weakening of the fibrotic cap, respectively.

Vascular CD44 gene expression is increased and CD44 is functionally activated in atheroprone regions of the vasculature in apoE^−/− mice relative to wild-type mice [20]. LMW HA regulates VSMC migration and phenotypic de-differentiation to the synthetic state in a CD44-dependent manner [17]. In addition, LMW HA-CD44 ligation enhances aortic smooth muscle migration in an ERK 1-2 dependent manner [148], and it has been shown that LMW HA-CD44 promotes thrombin-induced VSMC proliferation [46]. NADPH-deficient apoE^−/− mice demonstrated decreased CD44 and HA expression in both plasma and in atherosclerotic lesion compared to apoE^−/− mice, implying a role for NADPH at both the systemic and the local level. Administration of a Nox4-NADPH inhibitor reduced both ROS generation, atherosclerotic lesion burden, and CD44-HA expression in human aortic VSMCs [46]. Thus, it is critical to consider the role of NADPH as a potential therapeutic target which may impact the CD44-HA axis to ameliorate the progression of atherosclerosis.

Since synthetic VSMCs exhibit contradictory functions - pro-inflammatory cytokine production and increased MMP1 and MMP9 expression (which degrade type-I and type-IV collagen) versus ECM production and potential plaque stabilization, it is difficult to ascertain the net atherogenic effect of these cells on lesion stability. Although HAS2 is responsible for the majority of VSMC-derived HA production, both HAS1 and HAS3 contribute to VSMC HA production [65]. VSMC production of HA may promote either a pro- or anti-inflammatory environment depending on the size of HA (LMW or HMW) due to differential processing by HYALs and/or ROS, which influences its signaling capabilities, structure and interestingly, collagen architecture. HAS1-derived HA may form anti-atherogenic cables that trap leukocytes to prevent their active intralesional migration [149]. It has also been shown that oxLDL-LOX1 ligation on human aortic VSMCs may promote HAS2 and HAS3 overexpression, increased HA deposition, and subsequent enhanced monocyte adhesion in vitro [150]. Thus, it will be of interest to explore how different VSMC phenotypes regulate HA metabolism as well as the effect of HA on VSMC phenotype.

Foam Cell Formation:

One of the hallmarks of atherosclerotic lesion is the formation of lipid-laden foam cells, which may originate from macrophages, VSMCs and potentially dendritic cells [140,151–155]. Foam cells produce a variety of factors that promote necrotic cell death and the formation of a necrotic core. Of note, VSMC lineage tracing using Myh11-CreER^T2-ROSA-floxed-STOP-eYFP apoE^−/− mice has highlighted a novel insight that lesional VSMC-derived cells may adopt a macrophage-like phenotype [156]. In addition, these cells may be modulated in a KLF4-dependent manner, since VSMC-specific knock-out of KLF4 resulted in decreased lesion size and an increase in fibrotic cap area. There is evidence that CD44 and its pro-inflammatory ligand LMW HA both promote the formation of macrophage-derived foam cells by enhancing scavenger receptor CD36 expression and promoting the in vitro uptake of oxLDL in a PKC-dependent manner [157,158]. It would be interesting to determine whether CD44 alters the foam cell formation capacity of different foam cell precursors (VSMCs, dendritic cells, macrophages) using a similar mechanism in vivo. It would also be of interest to examine whether CD44 alters foam cell migration and subsequent egress from lesion back into the blood or via the lymphatic system.

Necrotic Cell Death and Efferocytosis:

Dead cell accumulation contributes to the formation of a necrotic core and the progression of atherosclerosis due to increased apoptosis or defective efferocytosis (clearance of apoptotic cells) [159]. These apoptotic cells are primarily macrophages/VSMCs/foam cells, but the apoptosis of any cells in lesion promotes a sustained pro-inflammatory milieu that is pro-atherogenic [160]. In vitro it has been shown that HMW HA prevents 4-MU-induced apoptosis in human VMSCs in a CD44/TLR4-dependent manner, highlighting the anti-inflammatory potential of HMW HA to reduce the sustained chronic inflammatory micro-environment of established lesion [161].

In both human and rabbit atherosclerosis lesional macrophages co-localize with classic apoptosis and oxidative stress markers - p53, cJun-API, and MnSOD. These highly-stressed cells also co-localize with TUNEL-positive nuclei, marking DNA strand breakage [162]. A follow up study revealed that CD44⁺ macrophages also co-localize with pro-inflammatory markers such as iNOS, TNFα and MHC class 1 and 2 implying that CD44 may contribute to pro-inflammatory M1 cell death in lesion [162,163]. If CD44 preferentially promotes M1 macrophage apoptosis in lesion, this again provides a pro-atherogenic role for the receptor. A pertinent extension of this study would be to examine the co-localization of macrophage subset specific markers such as heme oxygenase 1 for the Mox subset or CD206 for the M2 subset with apoptotic markers and CD44, which would provide a more complete picture of CD44’s role in apoptosis on different macrophage subsets.

CD44 regulates the general phagocytic capacity, as well as the efferocytic capacity of macrophages under inflammatory conditions. CD44 appears to be necessary for efficient phagocytosis of apoptotic leukocytes, specifically neutrophils [164–167]. Thus, a loss of CD44 on macrophages may in fact cause defective efferocytosis and a sustained pro-inflammatory environment. To answer this it would be pertinent to stain cells for markers of distinct scavenger receptors such as MERTK, which is the primary receptor responsible for efferocytosis in lesion [168].

Future Research Questions and Directions CD44 and Atherosclerosis Regression

The ideal therapeutic intervention for atherosclerosis would not only halt the progression of disease, but also enable lesion stabilization as defined as the alteration in plaque morphology that tilts the balance of lesion content to a decrease in lipid and inflammatory content and an increase in fibrotic content and/or regression [169]. Atherosclerosis regression has been studied in several different models - arguably the most direct approach so far involved the transplantation of lesion-filled aortas into wild-type mice which resulted in reduced lesion burden [170]. Other compelling evidence comes from studies using the Reversa transgenic mouse (LDLR^−/−Apob^100/100Mttp^fl/flMx1Cre^+/+) in which the severe hypercholesterolemia induced by HFD in LDLR^−/− mice can be normalized by conditionally knocking-out microsomal triglyceride transfer protein (Mttp) [169]. In the Reversa model, aged mice with advanced lesion and hypercholesterolemia were given polyinosinicpolyctidylic acid (pIpC) to induce the Mx1-Cre transgene to enable deletion of Mttp. This resulted in the normalization of plasma cholesterol levels in these mice. Analysis of plaque content revealed that mice with normalized lipid profiles had decreased total CD68⁺ macrophage content and lipid content. In addition, the group found that lipid normalization also increased both collagen content and classic M2 macrophage gene expression (Arg 1, CD163, Fizz-1, Mannose Receptor) in CD68+ laser captured cells from lesion relative to untreated mice. This implies that one of the key forces that promotes lesion regression is a decrease in total macrophage content combined with an increase in the relative amount of M2 macrophages [169].

One study has examined the role of CD44 in atherosclerosis regression. In the apo3Leiden atherosclerosis model, administration of an LXR agonist resulted in atherosclerosis regression. This regression was accompanied by decreased vascular CD44 expression in treated mice compared to untreated mice, as determined by mRNA and staining of aortic roots [171]. Co-staining with CD44 and cell-specific markers was not performed, so it is unknown whether this staining is localized to specific cell types, but it provides the basis for a more in-depth analysis of CD44’s role in regression. Co-staining aortic roots for CD44 and MERTK, the classic critical marker for efferocytosis in these models will provide a clearer understanding of the mechanism by which CD44 influences lesion regression.

Lesion regression may be influenced by many of the key processes that drive atherogenesis by directly influencing the inflammatory microenvironment of lesion. Since CD44 is necessary for efficient macrophage efferocytosis, the receptor may possess the capacity to promote lesion regression. However, given CD44’s pro-atherogenic role in the apoE^−/− mouse model as well as the previously mentioned regression study in the apo3Leiden model, any CD44 ‘pro-regression’ roles may prove complex to dissect. It is possible that an anti-atherogenic effect exerted by macrophage CD44 in efferocytosis is not sufficient to overcome the overwhelming pro-inflammatory milieu that CD44 promotes by influencing the other processes in lesion. It is also possible that CD44’s positive effect on regression is not sufficient to overcome CD44’s initial pro-inflammatory role by supporting leukocyte recruitment. Thus, it will be necessary to examine the impact of deleting or blocking the function of CD44 at multiple time points over the course of disease on atherogenesis in mouse models in order to fully delineate the temporal role that CD44 plays during key processes that govern regression. Co-staining for classic markers of apoptosis, efferocytosis, foam cell formation and cell-types will give a deeper understanding of how each of these cells works in concert to promote disease, and ultimately whether CD44 may be utilized to turn the tide of atherogenesis to favor disease regression.

Cellular Cross-Talk

Co-culture of T-regs with macrophages can decrease the capacity of macrophages to become foam cells in a IL-10 and TGFβ-dependent manner [172]. Conversely, CD11c⁺ antigen presenting cells (APCs) can activate Th1 cells to produce IFNγ and TNFα, which in turn promote macrophage-derived foam cell formation [173]. Thus, T cell subsets produce pro-inflammatory cytokines that may regulate monocyte CD44 activity and alter macrophage polarization, which may tip the balance in lesion to favor inflammation compared to fibrosis [119]. It is important to be cognizant of how CD44 expressed on one particular cell-type may indirectly regulate the actions of adjacent cells in lesion. This concept opens an additional avenue of study that may be illuminated by examining the effect of wild-type versus CD44^−/− cells in co-culture experiments (i.e. examining the role of wild-type versus CD44^−/− macrophages on wild-type or CD44^−/− VSMC and T cell function).

Role of CD44 on Other Myeloid cells in Lesion

This review has focused on key inflammatory players in lesion — endothelial cells, macrophages, T cells, and VSMCs. While these cells play key roles in atherosclerosis, it is essential to consider the roles of other inflammatory cells, such as dendritic cells (DCs) and neutrophils. DCs influence atherogenesis by activating T cells and phagocytosing oxLDL to form foam cells [173,174]. There is some evidence that CD44 is critical for DC-induced T cell activation but the function of CD44 in this regard has not been studied in the context of atherosclerosis [175–177]. Neutrophils contribute to both the initiation of disease by disrupting endothelial permeability and enhancing recruitment of blood-borne monocytes, but also weaken advanced fibrotic lesion by degrading ECM components, and increasing the expression of proteases MMP8 and MMP13 [178]. Although CD44 has been implicated in these processes in other contexts, the role that CD44 plays on neutrophil cell recruitment, phenotype, and survival in the context of atherosclerosis has not yet been elucidated. Since both DC and neutrophils can contribute to lesion morphology, understanding the impact of DC and neutrophil CD44 on atherogenesis remains an exciting avenue for the field either by tissue-specific deletion in vivo, or using in vitro models for the key processes mentioned above - inflammatory functional phenotype, foam cell formation, proliferation and efferocytosis.

CD44 Regulation: Post-Translational Modification and Variant Expression

While this review has focused on CD44 and HA in atherosclerosis there are, of course, several complexities that have not been discussed in detail. As already briefly mentioned, CD44 exists in several states (inducible, active, and inactive). Examining CD44 expression is not sufficient to fully-understand the role that CD44 exerts on atherogenesis. CD44’s activity may be regulated differently on multiple cell types (via alternative splicing of variant exons, post-translational modifications, glycosylation, sulfation, phosphorylation, clustering, receptor activation status) [7,8]. CD44 has multiple variants, which may be targets for regulation of the receptor’s activity. The standard form of CD44 and its variants are highly expressed in macrophage-rich advanced atheromatous lesion [179], which hints that macrophage-specific CD44 variant expression may correlate with lesion severity - another intriguing avenue of future study.

CD44-Dependent HA-Independent, vs. CD44-lndependent HA-Dependent Effects

This review has mainly discussed CD44-dependent HA-dependent effects on inflammation in atherosclerosis. However, it is important to briefly note that while CD44 is the principal receptor for HA, HA binds other target proteins including LYVE-1, TSG-6 and RHAMM, all of which are expressed in atherosclerotic lesions and may alter lesion burden [33,180,181]. In addition, CD44 may also have independent effects divorced from HA. CD44’s other potential ligands - osteopontin, fibronectin and collagen all contribute to atherogenesis [182–184]. Thus, it is imperative to take into account CD44-dependent HA-dependent effects, but also CD44-independent HA-dependent effects in order to fully dissect out the role that the CD44/HA axis has on atherosclerosis.

Conclusions

Lesion progression is governed by a constant tug-of-war between pro-inflammatory and disease-resolving processes, and CD44 regulates these pathways on multiple cell types to alter the progression of atherosclerosis. CD44 and HA are intriguing potential therapeutic targets for atherosclerosis as they play critical roles that influence the key processes that govern atherogenesis - maintenance of a pro-inflammatory milieu, foam cell formation, necrotic cell death, and efferocytosis. By analyzing the entire course of disease instead of just the end state of advanced lesion, it may be possible to identify critical checkpoints and dominant CD44-positive cell types that will allow us to alter the course of disease to halt the progression of disease and potentially enable lesion regression.

Highlights.

• Cardiovascular disease (CVD) due to atherosclerosis is a disease of chronic inflammation

• The CD44-HA axis has been implicated in CVD and thus is an attractive therapeutic target to halt the course of atherosclerosis

• While the role of CD44 and HA in the initiation of the inflammatory response in atherosclerosis is well-established, the role of CD44 and HA during the progression of the disease remains relatively unexplored

• We review the role of CD44 and HA in the key processes and multiple cell types throughout atherosclerosis to determine the possibilities of targeting the CD44-HA axis to halt disease progression and enable lesion regression

Abbreviations used

CVD

cardiovascular disease

VSMC

vascular smooth muscle cell

ECM

extracellular matrix

DC

dendritic cell

oxLDL

oxidized LDL

HA

hyaluronan

LMW HA

low molecular weight HA

HMW HA

high molecular weight HA

MERTK

proto-oncogene tyrosine-protein kinase MER

ROS

reactive oxygen species

RNS

reactive nitrogen species

T-reg

regulatory T cell

APC

antigen presenting cell

NO

nitric oxide