How Glucagon-Like Peptide-1 Receptor Agonists May Reduce Artery Plaque: Direct or Indirect Actions?

The beneficial effects of GLP‐1 RAs on CV risk have been suggested to be mediated, at least in part, by the reduction of hyperglycemia in patients with T2D [20]. Hyperglycemia promotes the formation of advanced glycation end‐products (AGEs), which accumulate in tissues and contribute to endothelial activation and inflammatory responses [21]. The interaction of AGEs with their receptor (RAGE) activates intracellular signaling pathways, including nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, leading to increased generation of reactive oxygen species (ROS) and the development of oxidative stress [22]. Oxidative stress, defined as an imbalance between the production of ROS and antioxidant defense mechanisms, plays a key role in endothelial dysfunction [23]. However, many glucose‐lowering agents have demonstrated significant effects on glycemic control without producing a comparable reduction in CV risk to that observed with GLP‐1 RAs.

Furthermore, obesity and T2D are often accompanied by atherogenic dyslipidemia, characterized by an excess of modified low‐density lipoprotein (LDL) particles, such as small dense LDL (sdLDL), very low‐density lipoproteins (VLDL), triglycerides (TG), and decreased levels of high‐density lipoproteins (HDL). These abnormalities contribute to intracellular lipid accumulation and the development of atherosclerotic plaques [24].

Numerous clinical studies have demonstrated the lipid‐lowering effects of GLP‐1 RAs [25–28]. Short‐term treatment with GLP‐1 RAs has been shown to improve postprandial lipid profiles, particularly by reducing TG and VLDL levels both in patients with T2D and in individuals with obesity but without diabetes [29]. Importantly, similar improvements in fasting lipid parameters, additionally with a decrease in LDL concentrations, have been observed after several months of therapy [30]. Although these effects are partly attributed to reduced caloric intake, some clinical studies involving T2D patients treated with liraglutide have suggested additional mechanisms, such as a reduction in apolipoprotein B48 levels and enhanced lipid catabolism mediated by apolipoprotein B100‐containing lipoproteins [31, 32].

Treatment with liraglutide in patients with T2D has also been shown to reduce hepatic and adipose tissue expression of the PCSK9 gene both in vivo and in vitro [31]. At the same time, an upregulation of hepatic low‐density lipoprotein receptor (LDL‐R) expression is observed [31]. Since PCSK9 is a key regulator of LDL‐R degradation, its downregulation may indirectly enhance LDL‐R availability on hepatocyte surfaces, thereby promoting increased clearance of circulating LDL.

Insulin resistance (IR), defined as an impaired biological response to insulin that affects the regulation of carbohydrate, lipid, and protein metabolism, as well as endothelial function, is a key factor in the pathophysiology of T2D. Numerous studies have demonstrated that insulin possesses anti‐inflammatory, antiapoptotic, and vasodilatory properties [33, 34]. IR contributes to the development of atherosclerosis by promoting chronic low‐grade inflammation, hyperinsulinemia, atherogenic dyslipidemia, and decreased nitric oxide (NO) synthesis [35]. One of the effects of GLP‐1 RAs is the improvement of insulin sensitivity, which has been demonstrated in clinical trials using the hyperinsulinemic–euglycemic clamp technique—the gold standard for assessing insulin sensitivity [36]. This effect appears to be mediated primarily through weight loss; however, emerging evidence suggests the existence of direct pathways that may independently contribute to these mechanisms [37].

Dysfunctional adipose tissue plays a significant role in the development of IR [38]. These dysfunctions include impaired lipid storage, disrupted adipokine secretion, and enhanced inflammatory activity. GLP‐1 RAs influence cardiovascular risk in part through their effects on adipose tissue. Agents in this class reduce the volume of total body fat, including visceral, subcutaneous, epicardial, and hepatic fat. However, the most clinically relevant reductions are observed in visceral and epicardial fat depots, which are strongly associated with cardiometabolic risk [39–41]. The reduction of these fat depots is associated with improvements in body weight, glycemic control, and lipid profile, ultimately leading to a decrease in CV risk.

GLP‐1 RAs have been shown to induce a phenotypic shift in adipose tissue toward a less inflammatory state. This effect is mediated, at least in part, through the suppression of nuclear factor kappa B (NF‐κB) signaling and modulation of the Jun N‐terminal kinase (JNK) pathway in adipocytes, leading to decreased production of proinflammatory cytokines such as interleukin‐6 (IL‐6) and tumor necrosis factor‐α (TNF‐α) [42, 43]. By mitigating these inflammatory processes, GLP‐1 RAs may attenuate the chronic low‐grade inflammation that represents a hallmark of CVD. These indirect adipose tissue–related effects are regarded as a key mechanism underlying the observed reduction in MACE in large CVOTs involving GLP‐1 RAs [44–46].

Another important component of metabolic syndrome is arterial hypertension, which contributes to mechanical damage of the vascular endothelium and increased lipid accumulation within atherosclerotic plaques. Meta‐analyses involving both patients with and without T2D have demonstrated the antihypertensive effects of GLP‐1 RAs, particularly in reducing systolic blood pressure [47, 48]. This effect has been observed in patients with baseline hypertension as well as in normotensive individuals. However, the exact mechanism underlying the antihypertensive action of GLP‐1 RAs remains unclear.

Although the blood pressure–lowering effect of weight loss is well established, there is insufficient evidence to confirm a direct hypotensive effect of GLP‐1 RAs. In vitro experimental studies have shown no relaxation of preconstricted mouse aortic rings following liraglutide exposure [49]. However, the same study demonstrated an indirect vasodilatory effect of liraglutide mediated by atrial natriuretic peptide (ANP) secretion, which is dependent on GLP‐1R activation in cardiac muscle.

Another study in normotensive rats treated with exenatide revealed the induction of diuresis and natriuresis through inhibition of sodium reabsorption in the proximal renal tubules and an increase in estimated glomerular filtration rate (eGFR) [50]. Elucidating the precise mechanisms of the antihypertensive effects of GLP‐1 RAs remains a challenge due to difficulties in extrapolating animal model findings to the human population and the widespread use of oral antihypertensive agents, which may confound the results of human studies.

The American Heart Association (AHA) emphasizes that chronic kidney disease (CKD), particularly in the presence of albuminuria or reduced eGFR, is associated with a progressively increased incidence of major cardiovascular events and cardiovascular mortality [51]. This phenomenon is attributed to the proinflammatory and prothrombotic milieu that accompanies CKD [52]. Additionally, uremic toxins promote endothelial dysfunction and the formation of unstable atherosclerotic plaques [52].

GLP‐1 RAs reduce cardiovascular risk also through their beneficial effects on renal function. Large meta‐analyses and randomized controlled trials have shown that GLP‐1 RAs reduce the risk of composite kidney endpoints, including kidney failure, sustained decline in eGFR, or death due to kidney disease by approximately 17%–18%, while also reducing MACE incidence by 13%–14% and all‐cause mortality by 12% in patients with T2D [53–58]. These agents reduce proteinuria and attenuate the decline in eGFR, thereby delaying the development of diabetic kidney disease, an important risk factor for cardiovascular morbidity and mortality [53]. The AHA highlights that GLP‐1 RAs such as liraglutide and semaglutide reduce the risk of onset and progression of CKD, primarily by lowering the incidence of new‐onset macroalbuminuria [59]. These nephroprotective effects contribute to improved cardiovascular outcomes in high‐risk populations.

Although the renoprotective effect of GLP‐1 RAs is less pronounced compared with sodium‐glucose co‐transporter 2 (SGLT2) inhibitors, their favorable impact on cardiovascular outcomes remains clinically relevant across various patient subgroups, including those with established CKD [60]. Proposed mechanisms underlying this renoprotective action include modulation of CKD risk factors such as hypertension, hyperglycemia, and obesity [61]. Additionally, direct nephroprotective effects have been suggested, including attenuation of inflammation, oxidative stress, and renal fibrosis [61].

Already in the second half of the 20th century, a link between CVD and inflammation was recognized [69, 70]. Initial studies in animal models, including monkeys, pigs, and rabbits, demonstrated the antiatherogenic effects of glucocorticoids, immunosuppressive agents, and antihistamines. Patients with chronic inflammatory diseases such as psoriasis and rheumatoid arthritis have been shown to be at increased risk of developing CVD [71, 72]. Specifically, in patients with systemic lupus erythematosus, the elevated CV risk is closely linked to subclinical atherosclerosis, a process potentially mediated by increased IR [73]. A clear association also exists between T2D, obesity, IR, and chronic low‐grade inflammation, which contributes to increased CV risk.

It has been suggested that GLP‐1 RAs may inhibit atherogenesis in part by modulating specific steps of the inflammatory process within the atherosclerotic plaque. As early as 2010, exendin‐4 was shown to suppress monocyte adhesion to the endothelium of the thoracic aorta and attenuate atherosclerotic lesions in ApoE−/− mice [74]. Using positron emission tomography (PET), Jensen et al. demonstrated that semaglutide reduced vascular uptake of tracers targeting activated macrophages and cellular metabolism in an atherosclerotic rabbit model [75]. However, researchers point out that the observed effects may be, at least in part, dependent on the reduction in weight gain induced by GLP‐1 RAs.

Current clinical and experimental studies indicate that GLP‐1RAs may exert a direct effect on oxidative stress. Evidence derived from a meta‐analysis of randomized controlled trials suggests that GLP‐1RAs significantly reduce levels of oxidative stress biomarkers, such as malondialdehyde (MDA), in patients with T2D compared with placebo or standard therapy [76]. The underlying mechanisms may involve both the direct inhibition of ROS production and the activation of antioxidant pathways, for example through the translocation of the nuclear transcription factor erythroid 2p45‐related factor (Nrf2) and increased activity of glutathione‐related enzymes observed in pancreatic β‐cells [77, 78]. GLP‐1R activation has also been shown to attenuate oxidative stress via the AMPK/sterol regulatory element‐binding protein 1 (SREBP1) signaling pathway—an important mediator of vascular inflammation in atherosclerosis [79].

In another study involving patients with T2D, a reduction in serum monocyte chemoattractant protein‐1 (MCP‐1) levels was observed after 26 weeks of treatment with either dulaglutide or semaglutide, independent of changes in body mass index (BMI), body weight, or lipid profile [80]. MCP‐1 is a cytokine strongly involved in macrophage recruitment to the atherosclerotic plaque, and elevated MCP‐1 levels are associated with acute coronary syndromes and CVD‐related mortality [81]. However, the study did not rule out confounding effects from hyperglycemia or the concurrent use of anti‐inflammatory drugs such as statins and fibrates, and thus the results should be interpreted with caution.

In a study using murine macrophages, Hou et al. showed that both liraglutide and dulaglutide promoted macrophage polarization from the proinflammatory M1 phenotype to the anti‐inflammatory M2 phenotype [82]. Moreover, both agents reduced the expression of interleukin‐1 beta (IL‐1β) and TNF‐α in cocultured human umbilical vein endothelial cells (HUVECs), although dulaglutide required a longer exposure time to elicit these effects.

Similar results were observed in a study involving ApoE−/−Irs2+/− mice (a model of atherosclerosis and IR), where lixisenatide enhanced the activation of signal transducer and activator of transcription 3 (STAT3), a key factor in promoting M2 macrophage differentiation [83]. However, it should be noted that in CVOT evaluating lixisenatide did not demonstrate a reduction in cardiovascular risk [84]. In a Danish study using transgenic LDLR−/− mice treated with liraglutide for 4 weeks, expression of genes involved in leukocyte recruitment to the vascular wall—including C‐C chemokine receptor Type 5 (CCR5), fractalkine, and MCP‐1—was reduced [85]. However, this study did not observe a reduction in atherosclerotic plaque burden, possibly due to the relatively short intervention period.

In a similar study conducted in an animal model, Rakipovski et al. demonstrated that long‐term treatment (12–17 weeks) with liraglutide and semaglutide reduced plaque progression independently of weight loss or lipid lowering [86]. Furthermore, in a model of acute aortic wall inflammation induced by lipopolysaccharide (LPS), semaglutide reduced TNF‐α, IL‐6, interleukin‐1 receptor antagonist (IL‐1RN), and MCP‐1 levels.

Several studies have further explained the anti‐inflammatory effects of GLP‐1 RAs through modulation of signaling pathways such as NF‐κB, AMPK‐Sirt1, and RMRP/miR‐128‐1‐5P/Gadd45g [87, 88]. In contrast to these experimental findings, a clinical study involving patients with T2D and low to intermediate CV risk found that liraglutide did not reduce vascular inflammation, as assessed by 18F‐fluorodeoxyglucose PET imaging after 26 weeks of therapy [89]. This result may suggest that the anti‐inflammatory effects of GLP‐1 RAs are more pronounced in patients with advanced or high cardiovascular risk.

Endothelial activation followed by dysfunction is an integral component of atherogenesis, allowing inflammatory cells and plaque‐derived debris to penetrate into the intimal layers of the vessel wall. Under the influence of agents such as tobacco smoke constituents, proinflammatory cytokines, ROS, AGEs, and oxidized low‐density lipoprotein (ox‐LDL), the endothelium becomes more permeable to leukocytes and acquires prothrombotic properties—initiating the formation of atherosclerotic plaques [24, 90, 91]. Activated endothelial cells increase the surface expression of selectins and adhesion molecules, promoting leukocyte and platelet rolling, activation, and further propagation of inflammation.

Oxidized LDL, via its receptor lectin‐like oxidized LDL receptor‐1 (LOX‐1), has become a potential therapeutic target. A clinical trial with an anti–LOX‐1 monoclonal antibody is currently underway [92]. In vitro studies using human aortic endothelial cells (HAECs) exposed to ox‐LDL demonstrated that liraglutide may exert a protective effect on the transcription factor Krüppel‐like factor 2 (KLF2) via a mitogen‐activated protein kinase/extracellular signal–regulated kinase 5 (MAPK‐ERK5)–dependent pathway [93]. KLF2 is a key regulator of vascular endothelial biology, modulating endothelial nitric oxide synthase (eNOS), occludin expression, and NF‐κB pathway regulation [94]. Studies on HUVECs, a standard experimental model of the endothelium, have also shown that liraglutide reduces LOX‐1 expression in cells exposed to ox‐LDL [95]. These findings suggest that GLP‐1 RAs may inhibit LOX‐1–mediated endothelial activation. Additional mechanisms have been proposed, including inhibition of PTEN‐induced kinase 1 (PINK1)/Parkin‐dependent mitophagy [96]. HUVECs cultured under hyperglycemic conditions exhibited mitochondrial membrane potential preservation and increased eNOS phosphorylation when treated with liraglutide. A synergistic effect of liraglutide and metformin was also studied in HUVECs, where an antiproliferative effect on endothelial cells was observed, suggesting possible beneficial roles of GLP‐1 RAs in oncology [97].

Endothelial adhesion molecule expression, a contributor to atherogenesis, has also become a focus of interest in the context of GLP‐1 therapy. In addition to molecular and immunoassay‐based evaluations of adhesion molecules, contrast‐enhanced ultrasound molecular imaging (CEUMI) has been used. In one such in vivo study in ApoE−/− mice, contrast microbubbles targeted to vascular cell adhesion molecule‐1 (VCAM‐1) showed a threefold signal increase in the control group, whereas liraglutide treatment prevented this response, maintaining stable VCAM‐1 levels throughout the study [98]. This study also reported reductions in TNF‐α, IL‐1β, and MCP‐1 concentrations following treatment, independently of glucose levels; however, no analyses were performed to assess whether these effects correlated with weight loss.

Similar results were found in an earlier 2011 study using immunohistochemistry in an atherosclerotic mouse model, where liraglutide reduced the expression of intercellular adhesion molecule‐1 (ICAM‐1), whereas no significant changes in blood pressure or body weight were observed in the treatment groups [99]. These effects were hypothesized to be mediated by inhibition of TNF‐α through GLP‐1R activation. To further elucidate mechanisms, Krasner et al. demonstrated in HAEC cultures that liraglutide downregulated E‐selectin and VCAM‐1 expression via increased intracellular Ca²⁺ levels, activation of Ca²⁺/calmodulin‐dependent protein kinase kinase‐β (CaMKKβ), and subsequent AMPK activation—leading to suppression of the inflammatory response [100].

Another clinical study assessed the effects of liraglutide in obese women with and without polycystic ovary syndrome (PCOS) [101]. After 6 months of treatment, a similar reduction in soluble P‐selectin, ICAM‐1, and VCAM‐1 levels was observed in both groups, despite only moderate weight loss of 3%–4%. However, the absence of a placebo‐controlled group with similar weight loss made it difficult to determine whether these effects were drug‐specific or weight loss–related [101].

Conflicting data came from a study involving patients with T2D at high CV risk, where a 26‐week treatment with semaglutide or dulaglutide resulted in increased serum ICAM‐1 levels but decreased concentrations of L‐selectin and MCP‐1 [80]. No correlation with changes in body weight was observed; however, this study did not adequately account for the influence of hyperglycemia or concurrent medications, limiting the interpretability of the findings.

The enzymes eNOS and iNOS (inducible nitric oxide synthase) play a crucial role in regulating endothelial function and act as key mediators of vascular tone and homeostasis through the generation of NO [102]. Numerous animal studies have provided evidence of the impact of GLP‐1 RAs on this aspect of vascular endothelial function.

In the previously mentioned 2011 study, immunohistochemical analysis of aortic rings from ApoE−/− mice demonstrated a significant increase in eNOS expression in the endothelium following liraglutide treatment—a response that was abolished when coadministered with the GLP‐1R antagonist exendin‐9 [99]. Han et al. [103] showed in a model of obese mice that liraglutide improved endothelial function via activation of the AMPK/eNOS signaling pathway, a mechanism further confirmed in experimental studies using HAECs [100].

Other in vitro models have suggested that GLP‐1 RAs act via the Janus Kinase 2/signal transducer and activator of transcription 3 (JAK2/STAT3) signaling pathway, resulting in increased eNOS expression and angiogenic factor production—mechanisms that may contribute to their anti‐ischemic effects [104]. Liraglutide has also been shown to enhance eNOS phosphorylation and increase NO production, an effect at least partially dependent on activation of the mechanistic target of rapamycin (mTOR) signaling pathway [105]. This pathway is one of the core regulators of cellular metabolism. Additionally, studies have shown a protective effect of GLP‐1 RAs on pancreatic beta cells, cardiomyocytes, and neurons via the mTOR pathway, suggesting that mTOR signaling may be a central mechanism of action for this drug class [106–108].

In contrast to eNOS, GLP‐1 RAs appear to exert an inhibitory effect on iNOS. Reduced iNOS expression within atherosclerotic plaques is associated with diminished vascular inflammation and increased plaque stability [109, 110]. In murine models of atherosclerosis, intervention with lixisenatide led to decreased iNOS expression and increased Arginase I content within the plaques, indicating attenuated proinflammatory signaling and reduced oxidative stress [83]. As mentioned above, these results should be interpreted cautiously, as lixisenatide did not demonstrate cardiovascular risk reduction in CVOTs [84]. The positive effect of liraglutide on iNOS‐dependent nitro‐oxidative stress was also demonstrated in another animal model, suggesting that this effect is partly mediated through the AMPK pathway [111].

MMPs are a family of zinc‐dependent endopeptidases responsible for tissue remodeling and the degradation of extracellular matrix proteins. MMPs can modulate various cellular and signaling pathways involved in atherosclerosis, particularly those associated with plaque progression and rupture [112]. Their secretion is regulated by proinflammatory cytokines, as well as by specific endogenous tissue inhibitors (tissue inhibitors of metalloproteinases—TIMPs) [113].

The role of MMP‐9 in the development of atherosclerosis and endothelial dysfunction has been confirmed in multiple studies [114–116]. Several ex vivo investigations using animal vessels have demonstrated that GLP‐1 RAs reduce MMP‐9 levels [117–119]. In human coronary artery smooth muscle cells (hCASMCs) exposed to inflammatory stimuli, exenatide was shown to reduce the expression of MMP‐1, MMP‐2, and MMP‐9, likely through inhibition of the Akt signaling pathway [120]. Liraglutide has also been shown to upregulate the expression of TIMP‐2, which may represent an additional mechanism of action for this drug class [85].

In a murine model of atherosclerosis, GLP‐1 RAs reduced MMP‐9 expression, limited macrophage infiltration, and increased collagen content and fibrous cap thickness within plaques—changes associated with enhanced plaque stability [121]. However, the molecular mechanisms through which GLP‐1 RAs modulate MMP expression remain poorly understood, and there is a lack of clinical data directly linking GLP‐1 RAs with MMP modulation. Given the close association between MMPs and inflammatory signaling pathways, it is likely that the observed effects are mediated indirectly via the anti‐inflammatory properties of GLP‐1 RAs.

A characteristic feature of advanced atherosclerotic lesions is an enhanced prothrombotic state and remodeling of deeper vascular wall layers, including VSMCs [122, 123]. Although these processes are primarily associated with later stages of atherogenesis, increased VSMCs proliferation and migration—secondary to MMP activity—have also been demonstrated during earlier stages [124, 125].

A key aspect of pathological VSMCs transformation is the phenotypic switch from a contractile to a so‐called synthetic phenotype [122]. Studies using cultured VSMCs have shown that liraglutide inhibits VSMCs proliferation, migration, and phenotypic switching by suppressing the expression of proteins involved in lipid metabolism, such as PCSK9 and the LDL receptor [126]. Another study, using VSMC cultures under hyperglycemic conditions, reported a similar effect mediated through inhibition of the ERK1/2 and PI3K/Akt signaling pathways [127].

Liraglutide has also been shown to suppress AGE‐induced phenotypic switching of VSMCs. GLP‐1 RAs increased the expression of contractile VSMC markers while reducing collagen production, likely through downregulation of myocardin, inhibition of NF‐κB signaling, and activation of the PKA signaling pathway [128]. In another VSMC culture model, exendin‐4 was found to act via AMPK/Sirtuin 1/Forkhead Box O3a (AMPK/SIRT1/FOXO3a) signaling, leading to increased expression of contractile VSMC markers such as calponin and SM22α [129].

In contrast to these findings, Bjørnholm et al. reported a reduction in the expression of two contractile phenotype markers—Cnn1 and Cald1—in atherosclerotic vascular walls of animals treated with liraglutide for 4 weeks [85]. Since the contractile phenotype of VSMCs is considered physiological, these results may suggest a potentially adverse effect of GLP‐1 RAs. However, the role of VSMCs hypertrophy and phenotypic switching remains debated, as this transformation may actually promote fibrous cap formation and plaque stabilization, thereby reducing cardiovascular risk [122]. This hypothesis is supported by a study conducted in Watanabe heritable hyperlipidemic rabbits, where 12 weeks of liraglutide therapy led to increased fibrosis of VSMCs, as demonstrated by histopathological analysis of atherosclerotic arterial walls—suggesting enhanced plaque stability [130].

The association between plasminogen activator inhibitor‐1 (PAI‐1) and atherothrombosis is well established. Elevated levels of PAI‐1 are a recognized risk factor for thrombosis and vascular complications in diabetes [131]. Increased PAI‐1 concentrations have also been observed in patients with subclinical atherosclerotic changes [132]. This molecule is a key inhibitor of fibrinolysis, leading to pathological fibrin deposition and thrombus formation [133].

Initial studies on the effects of GLP‐1 RAs on PAI‐1 expression, conducted in HUVECs and dating back to 2009, demonstrated inhibition of PAI‐1 induction—suggesting a mechanism involving TNF‐α modulation via the Akt signaling pathway and the nuclear Receptor Nur77 [134, 135]. However, subsequent studies in human populations suggest that the inhibition of PAI‐1 expression by liraglutide may be indirect, resulting from improvements in hyperglycemia or inflammation, and is likely not driven by weight loss or improved IR [136].

A meta‐analysis of GLP‐1–based therapies in patients with T2D showed a minor reduction in serum PAI‐1 levels, with a mean difference of approximately −12.9% (p = 0.05), indicating a modest but consistent lowering effect in this population [137]. Further studies are warranted, particularly those conducted in larger patient cohorts under conditions that eliminate the confounding effects (e.g. hyperglycemia), to better elucidate the mechanisms underlying PAI‐1 downregulation by GLP‐1 RAs.