The Internal Mammary Artery and its Resilience to Atherogenesis: Shifting from Risk to Resistance to Address Unmet Needs

Abstract

Fuelled by the global surge in ageing, atherosclerotic cardiovascular disease reached pandemic dimensions putting affected individuals at enhanced risk of myocardial infarction, stroke and premature death. Atherosclerosis is a systemic disease driven by a wide spectrum of factors, including cholesterol, pressure and disturbed flow. Although all arterial beds encounter a similar atherogenic milieu, the development of atheromatous lesions occurs discontinuously across the vascular system. Indeed, the internal mammary artery (IMA) possesses unique biological properties that confer protection to intimal growth and atherosclerotic plaque formation, thus making it a conduit of choice for coronary artery bypass grafting. Its endothelium abundantly expresses nitric oxide (NO) synthase and shows accentuated NO release, whilst its vascular smooth muscle cells (VSMC) exhibit reduced tissue factor expression, high tissue-type plasminogen activator production and blunted migration and proliferation, which may collectively mitigate intimal thickening and ultimately the evolution of atheromatous plaques. We aim here to provide insights into the anatomy, physiology, cellular and molecular aspects of the IMA thereby elucidating its remarkable resistance to atherogenesis. We propose a change in perspective from risk to resilience to decipher mechanisms of atheroresistance and eventually identification of novel therapeutic targets presently not addressed by currently available remedies.

Introduction

Our familiarity with atherosclerosis obscures a fact that challenges our conventional understanding of disease initiation and progression: Most risk factors for atherosclerosis are systemic, yet the disease is focal in nature. For example, all vessels bath in the same metabolic milieu and are uniformly exposed to the untoward effects of hypertension. Yet, atherosclerosis involves arteries segmentally and affects different arterial beds unequally.¹ Specifically, atherogenesis tends to spare the internal mammary artery (IMA), while the neighbouring and similarly sized coronary arteries bear the brunt of atherosclerosis (FIG. 1).^2,3 Also, when used as a bypass graft, the IMA has superior durability compared to vein grafts,^4,5 although they are exposed to the same arterial pressure and lipoprotein levels.

Figure 1:

Representative angiography of the coronary artery tree (A) and the LIMA (B) of a 58-year-old patient with multi-vessel CAD. (A) Note that the coronary artery bed displays multiple stenotic lesions. In particular, there is a severely occlusive process (red triangle) present in the left main coronary artery that involves its trifurcation and coincides with multiple lesions distally. While the proximal part of the ramus intermedius artery is 80% occluded (green triangle), the middle part of the circumflex artery exhibits a moliminous stenotic process (yellow triangle). (B) Conversely, the LIMA of the very same patient exhibits no stenotic lesions throughout its course. Abbreviations: CAD, coronary artery disease; LIMA, left internal mammary artery.

What protects the IMA from atherosclerosis? Does this protection arise because of embryological differences of arterial endothelial and smooth muscle cells? Does this difference reflect local flow conditions or protective properties at cellular and molecular levels? This review considers these issues and provides insights in its unique biological properties and a perspective for future research into atherosclerosis, aiming to shift our focus to so far unknown protective factors that might provide opportunities for novel measures in patients at risk for atherosclerotic cardiovascular disease (ASCVD) and its sequelae.

According to the current concept, atherosclerosis occurs due to risk factors, such as age, hypertension, hypercholesterolemia, and diabetes.⁶ The risk factor concept was introduced by the Framingham Study,⁷ but cohorts worldwide have confirmed the association of cardiovascular (CV) risk factors with coronary artery disease (CAD), infarction and cardiac death.^8–12 More recently, inflammation has been shown to drive atherosclerosis, specifically via the NLRP3 inflammasome and its downstream effectors, such as interleukins-1β and −6. Clinical studies using C-reactive protein (CRP) as a readout confirmed that inflammation predicts the risk of major adverse CV events (MACE).^13,14 Furthermore, selected randomized controlled trials (RCTs) with anti-inflammatory agents largely reported favourable outcomes and a reduction of MACE in patients with a history of acute coronary syndromes (ACS).^15–18 Other emerging risk factors include triglyceride rich lipoproteins such as apolipoprotein C-III.¹⁹ Also, clonal haematopoiesis of indeterminant potential or CHIP, is another recently recognized CV risk factor.^20–22

Contemporary CV medicine possesses a powerful armamentarium of tools for modifying CV risk factors such as statins and inhibitors of intestinal cholesterol uptake and proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors^23–26 as well as effective anti-hypertensive medications.²⁷ Novel anti-diabetic drugs such as glucagon-like peptide-1 receptor agonists and sodium glucose co-transporter-2 (SGLT-2) inhibitors further improve clinical outcomes,^28,29 partly even in the absence of diabetes.³⁰ Finally, although the interleukin-1β antibody canakinumab will not become available for CV indications,^31,18,32 colchicine may assume a place as an anti-inflammatory remedy.^16,17,33,34

Despite this impressive armamentarium for ASCVD therapy, the unmet medical need remains high. Indeed, in the FOURIER trial, at 3 years the composite of CV death, myocardial infarction, stroke, hospitalization for unstable angina, or coronary revascularization still occurred in 12.6% with evolocumab on top of statin therapy despite a median low-density lipoprotein cholesterol (LDL-C) reduction of 30 mg/dl.²⁴ Similarly, in CANTOS the risk of MACE remained high in spite of an effective reduction of CRP by canakinumab therapy.²⁴ Hence, there is considerable room for improving outcomes in ASCVD patients by targeting pathways beyond those addressed by current treatment strategies. Moving from risk to resilience by deciphering the IMA’s resistance to atherosclerosis may provide hints for the development of novel therapies to address the unmet CV risk that remains despite all the contemporary therapies outlined above.

The ‘resilience’ of the IMA to atherogenesis

The term resilience (lat.: ‘re-silire’; ‘to bounce back’) refers to the ability of a system to adapt and return to equilibrium despite injurious insults, a multidisciplinary concept broadly established in the field of (psycho-)biology. This notion, as we postulate, is also applicable to the IMA with respect to its resistance to atherogenic effects of systemic CV risk factors, such as high LDL-C, diabetes and hypertension. Indeed, irrespective of their systemic nature, the IMA rarely develops atherosclerotic plaques (FIG. 2) with prevalence rates of histologically-proven lesions ranging between 3.1–4.2% in unselected individuals^3,35 and 0.7–7% in patients with multivessel CAD.^36,37

Figure 2:

Native LIMA (A, B), RIMA (C, D) and descending aorta (E, F) obtained from a 58-year-old patient during autopsy. While the LIMA (A, B) and RIMA (C, D) are histologically normal and free of atheromatous lesions, cross-sections of the descending aorta (E, F) sampled from the same patient display profound intimal hyperplasia and fibrotic changes (i.e., fibroelastosis), as indicated by black arrow heads. Black arrow heads in D denote the IEL. Note that the white arrow head in F points at the poorly defined IEL. In all sections either H.E. (left) or Elastica van Gieson staining (right) was used. A-F: Magnification 50x. Asterisks mark the adventitia. Abbreviations: IEL, internal elastic lamina; LIMA, left internal mammary artery; RIMA, right internal mammary artery.

Intimal thickening and the IMA:

Intimal thickening – a preclinical process closely associated with atherosclerosis - can be broadly categorized as being eccentric or diffuse, the former being observed in the left anterior descending coronary artery (LAD) of children as early as the first week of life.³⁸ Histological mapping of eccentric intimal thickening and advanced atherosclerotic lesions in humans revealed that the topographical distribution of these two processes is very similar in the coronary arteries,³⁹ the aorta⁴⁰ and the internal carotid artery,⁴¹ suggesting that eccentric intimal thickening may furnish an anatomical soil required for lipid accumulation and in turn atherosclerotic plaque formation. However, as the build-up of plaques is not necessarily confined to regions of the vasculature where eccentric intimal thickening occurs, the latter may mark regions with a high propensity to plaque formation when exposed to a pro-atherogenic milieu.⁴² The large-scale necropsy study by F.H. Sims⁴³ demonstrates beyond any doubt that the intima of the IMA enjoys remarkable protection from structural changes occurring with advancing age. Indeed, while the thickness of the coronary intima sharply increases after the third decade, particularly in males, the intimal thickness of the IMA remains surprisingly constant until the fifth decade and never exceeds 60% of the media’s width during a normal life-span.⁴³ These findings suggest that age-associated intimal alterations develop at a much slower pace in the IMA, a notion strengthened by several independent reports.^44–47

Susceptibility to atheroma compared to similarly-sized arteries:

In a subsequent autopsy series,³ comprising 112 male and 48 female individuals of high-risk ethnicity aged 20 to 94 years, above 97% of the IMAs studied had an intimal thickness below 25% of the luminal diameter. Importantly, the majority of macroscopic atherosclerotic lesions, if present, tended to be below 25% of the luminal surface and showed no correlation with a diagnosis of CAD.³ Remarkably, even the radial artery, a vessel also known to be far less susceptible to intimal hyperplasia and atheroma formation as compared to the LAD,⁴⁸ develops atheromatous lesions much more frequently and to a much higher degree than the IMA.^36,49,50 Indeed, in a paired analysis of radial and IMAs obtained from 150 patients undergoing coronary artery bypass grafting (CABG), histology-proven fibroatheromas were found in 5.3% of radial arteries, but only in 0.7% of the IMAs.³⁶ In concert, morphometric indices of intimal hyperplasia and atherosclerosis (i.e., percentage of luminal narrowing, intimal thickness index, and intima-to-media ratio) were much more pronounced in the IMA than the radial artery.³⁶ Also, in another autopsy study involving 153 patients aged 20–89 years, all IMAs studied were patent with only one patient having a plaque-like lesion in the IMA. Conversely, 44 of the coronary arteries were obstructed and 37 displayed atherosclerotic plaques.⁵¹ Some authors described incipient intimal lesions in the IMA associated with advanced age,^{35,36,44,49,52} hypertension,^3,35,44 diabetes,^35,49 smoking^36,44 or hypercholesterolemia.⁵³ Yet, the similar prevalence rates observed in population-based autopsy series^3,35 and cohort studies involving patients undergoing CABG^36,37 combined with the high variation observed within some of these studies,³⁶ indicate that only few if any intimal alterations are attributable to risk factor exposure. In fact, a non-invasive in vivo study using linear ultrasound probes showed that the relative intimal thickness of the IMA does not differ between patients with or without CAD (0.51±0.11 vs. 0.50±0.17 mm, P<0.05),⁵⁴ findings in accord with independent reports.^50,55 These observational findings, available from autopsy^3,35,43,44, ex vivo^36,49,56 and in vivo studies^50,57,58, strengthen the notion that early and advanced stages of atherogenesis occur to a much lesser extent and at a much lower pace in the IMA as compared to other similarly-sized segments of the arterial system such as the radial artery or LAD.

Indication for and benefits of IMA grafting:

Following the first successful aortocoronary bypass surgery using an autogenous saphenous vein graft (SVG) by Garrett and colleagues in 1964,⁵⁹ CABG advanced markedly, becoming the standard of care for many patients with multi-vessel CAD, specifically in those with intermediate-to-high anatomical complexity as assessed by the SYNTAX Score.⁶⁰ Over decades, the combined use of both the left IMA (LIMA) and SVGs has been the standard therapy for patients undergoing CABG, but SVGs tend to fail. Indeed, already in 1986 Loop et al. showed that patients receiving only SVGs are more likely to die within 10 years than those in who an IMA graft has been placed, a difference that widens over time.⁶¹ Independent studies consistently replicated these findings^62–64 showing that the use of IMA grafts is an even better predictor for survival than native CAD progression. While RIMA and LIMA grafts display 15-year patency rates >90%,⁶⁵ only 75% to 86% and 55% to 60% of SVGs remain patent over a period of 5 and 10 years, respectively.^4,5,66

Late-graft patency is limited by atherosclerosis-like processes:

Limitations to SVG patency >1 year after CABG result mainly from the development and progression of atherosclerosis,^67,68 a process that – similar to the native disease – can be hastened by traditional risk factors, such as hypercholesterolemia and smoking.⁶⁹ While accelerated atherosclerosis occurs frequently in SVGs, IMA grafts rarely develop atherosclerotic plaques. Indeed, in the paired post-mortem study by M.E. Shelton et al.⁶⁸ roughly 71% of SVGs showed atherosclerotic changes after a mean follow-up of 56 months, while only about 6% of IMA grafts had histologically-proven lesions.

Downstream effects of IMA grafts:

Evidence accumulated over the last three decades supports the notion that IMA conduits also blunt the progression of atheromatous lesions in the native downstream coronary bed.^70–75 In an observational series comprising 772 patients with a symptom-driven indication for coronary angiography, Kaplan-Meier estimated overall CAD progression in segments distal to the conduits were 8% in LIMA grafted territories at 10 years, but 43% in those receiving a venous conduit. Also, in patients with patent grafts, freedom from CAD progression 10 years upon CABG was much more common in territories grafted with the LIMA as compared to the radial artery (72% vs. 60%, P<0.05).⁷⁴

Benefits of bilateral IMA grafting:

Considering the vast survival benefit associated with the use of IMA conduits, bilateral IMA grafting gained centre-stage in CABG, with observational data providing convincing evidence that bilateral IMA grafting improves survival over short- and long-term as compared to the use of single IMA grafts.⁷⁶ The arterial revascularization trial (ART) was the first, and so far only, large-scale RCT designed to compare outcomes of bilateral versus single LIMA CABG at 10-years. While no difference in survival and event-free survival between single LIMA and bilateral IMA was noted in the intention-to-treat analysis, an as-treated analysis showed that in patients receiving multiple arterial grafts (i.e., LIMA and RA, and/or bilateral IMA) the incidence of MACE (i.e., composite measure of all-cause death, MI or stroke) was 20% lower at 10 years compared to those in whom a single arterial graft (i.e., LIMA) was placed.⁷⁷ As findings of ART were confounded by high crossover rates, results from the ROMA trial (Randomized Comparison of the Clinical Outcome of Single Versus Multiple Arterial Grafts)^78,79 are eagerly awaited, with primary results likely becoming available in 2025.

Anatomic, cellular, molecular, and physiological insights

Anatomy of the IMA:

Both, the RIMA and LIMA arise from the proximal part of the brachiocephalic trunk and subclavian artery, respectively. Proximally, the IMA runs forward but then courses behind the cartilages of the upper ribs. Along its way caudally it gives rise to multiple branches, including the anterior intercostal branches, before dividing into the musculophrenic and the superior epigastric arteries approximately at the level of the sixth intercostal space.⁸⁰ Of note, similar to the right IMA,⁸¹ the luminal diameter of the LIMA decreases marginally from proximal to distal, with 2.6±0.4 mm and 2.4±0.4 mm at the level of the 2^nd and 5^th rib, respectively.⁸² Microscopically, the IMA shows features of both, elastic type arteries (e.g., aorta) and vessels of muscular phenotype (e.g., LAD).⁸³ Indeed, its media exhibits circumferentially aligned vascular smooth muscle cells (VSMC) that are firmly belted by a plethora of longitudinally and circularly interlacing elastic fibers,^84,85 which might limit VSMC proliferation and migration.⁸⁶ Furthermore, its well-defined (i.e., non-fragmented) internal elastic lamina (IEL)^47,87 whose morphology is characterized by coarse folds⁸⁸ might form an additional barrier to the migration of these cells.⁸⁶ Conversely, the fragmentation of the coronary IEL may provoke VSMC proliferation and thus progressive thickening of the coronary intima.^47,87 Vasa vasorum foster blood and oxygen delivery, allow resident inflammatory and progenitor cell trafficking that – according to the non-traditional ‘outside-in’ hypothesis⁸⁹ –can contribute to vascular inflammation and neointima formation.⁹⁰ Therefore, it is tempting to speculate that the low density of vasa vasorum of the IMA, as assessed by micro-CT,^91,92 might limit immune-cell trafficking, neointima development and eventually – at least in part –decelerate the evolution of atherosclerotic plaques.⁹⁰ While multifaceted factors certainly contribute to both, atherosclerosis and adventitial remodelling, the precise mechanisms linking these entities and their contribution to the IMA’s resilience to atherosclerosis remain to be fully untangled.

Intimal extracellular matrix composition:

The thickened intima mainly comprises VSMCs and proteoglycans which are key for the retention of atherogenic lipoproteins and thus for the initiation of atherogenesis.⁹³ Indeed, negatively charged side-chains of proteoglycans such as biglycan, versican or perlecan interact with positively charged apolipoprotein B containing LDL particles thus playing a major role during disease initiation. Importantly, proteomic studies revealed that the proteoglycan lumican is abundantly expressed by the intima of the atherosclerosis-prone carotid artery, relative to the intima of the IMA.⁹⁴ Notably, while lumican can impact cell proliferation and apoptosis, it also shapes macrophage function.^95,96 These data indicate that the intima of the IMA possesses a distinct extracellular matrix composition compared to atherosclerosis-susceptible vessels, which may attenuate lipid retention, immune-cell trafficking and in turn atherosclerotic plaque formation.

Endothelial structure, function and lipid homing:

Endothelial cells of the IMA are densely stacked in the intimal layer, show low intercellular permeability and display enhanced expression of junctional molecules which might attenuate endothelial injury, lipid trafficking and eventually immune-cell trafficking.^67,97,98 Indeed, endothelial coverage is well preserved in the IMA and vein, while saphenous vein dissected from patients undergoing CABG shows areas of endothelial denudation.^99,100 Also, macrophages residing within the IMA accumulate largely at the medial/adventitial border suggesting that the migration of immune-cells must mainly occur abluminally, which underscores the resistance of the IMA’s endothelium.¹⁰¹ Besides endothelial structure, endothelial vasodilator function is also well-preserved in the IMA as compared to atherosclerosis-prone vessels.⁹⁹ Of note, endothelium-dependent vasorelaxation evoked by acetylcholine, adenosine diphosphate or thrombin is remarkably preserved in the IMA. Notably, while indomethacin-induced blockage of prostaglandin synthesis shows no effect on vasorelaxation, relaxation of the IMA is profoundly impaired upon methylene blue and hemoglobin exposure⁹⁹ as well as L-NMMA¹⁰² suggesting that this phenomenon is mainly mediated by nitric oxide (NO). This finding has particular interest, as endothelial cells of the IMA contain abundant in endothelial nitric oxide synthase (eNOS) and show accentuated NO release in comparison to the carotid artery,¹⁰³ a vessel highly susceptible to develop atheroma. Besides the importance of endothelial NO synthesis, transendothelial transport of cholesterol-rich lipoproteins is a process highly relevant to atherogenesis, specifically during disease initiation. Indeed, recent experimental evidence suggests that the scavenger receptor class B type 1 (SR-B1) actively drives the endothelial uptake of LDL particles and thus promoting atherogenesis.¹⁰⁴ Although it is tempting to postulate that the SR-B1 expression shows spatial differences across the vascular system, further studies are warranted to investigate whether its expression is indeed differentially regulated depending on vascular bed and disease-susceptibility, and if its (hypothetically) blunted expression contributes to atheroprotection of the IMA.

VSMC proliferation and migration:

VSMCs of the IMA are mainly located in the media and are highly resistant to migration and proliferation^105–108 – processes that play important roles through different stages of atherogenesis.¹⁰⁹ According to the current paradigm, atherogenic risk factors can provoke phenotypic modulation of VSMCs and stimulate their migration to eventually transdifferentiate into different lineages that can elicit harmful as well as protective effects. While their transition to macrophage-like cells may foster plaque rupture, VSMC-derived fibroblast-like cells (so-called ‘fibromyocytes’) may actually consolidate the fibrous cap.¹¹⁰ Indeed, at early disease stages VSMC proliferation can promote the evolution of (eccentric) intimal thickening through a complex interplay of lipid retention, inflammation, phenotype switching and cellular death.¹¹¹ At more advanced disease stages, however, VSMC-derived cells may actually stabilize atherosclerotic plaques making them less prone to rupture,¹¹⁰ as demonstrated by experimental evidence showing that VSMCs residing within the fibrous cap mainly derive from medial VSMCs as they undergo cell migration and proliferation.^112,113 Proliferative responses of VSMCs are orchestrated by different stimuli, including pulsatile stress and secreted growth factors (by e.g., activated endothelial cells and monocytes).^114,115 Interestingly, VSMCs of the IMA possess an intrinsic resistance to these stimuli, as shown by several studies (FIG. 3).^105–107 Specifically, VSMCs of the IMA exposed to pulsatile stretch (1 Hz) over 24 hours do not show enhanced ³H-thymidine incorporation and numbers of VSMCs obtained from the IMA remain stable during 6 days of stretch.¹⁰⁶ Likewise, VSMC proliferation upon exposure to platelet-derived growth factor (PDGF) BB¹⁰⁷ or thrombin¹⁰⁵ is much more pronounced in atherosclerosis-prone vessels compared to the IMA, although VSMCs of the IMA express functional PDGF and thrombin receptors.^105–108 Importantly, PDGF triggers mitogen-activated protein kinase (MAPK) and blunts downregulation of the cyclin-dependent kinase inhibitor p27^Kip1 (which blocks cell cycle progression at G1 and is thus required for the transition from quiescence to the proliferative state) in cultured VSMCs of the IMA.^105,116 VSMC migration and proliferation is also shaped by tissue factor,¹¹⁷ a key initiator of the coagulation cascade upon plaque rupture.¹¹⁸ Notably, Payeli et al. found attenuated expression of tissue factor in VSMCs isolated from the IMA. In concert, the supernatant of VSMCs isolated from the IMA displayed lower tissue factor activity, higher tissue factor pathway inhibitor and elevated tissue-type plasminogen activator and, consequently, prolonged clotting time when added to human plasma.¹¹⁹ These findings suggest that (i) VSMCs of the IMA lack an array of intrinsic stimuli for cellular growth and (ii) display antithrombotic properties which may not only blunt plaque progression following CABG but also mitigate (athero-)thrombotic events, as evidenced by its low rates of early and late graft attrition.¹²⁰

Figure 3:

Outgrowth of VSMCs. Vascular tissue specimen of the saphenous vein and IMA, respectively, obtained from the same patient show different outgrowth rates when cultured in 20% FBS DMEM over a 20-day period. Abbreviations: FBS, fetal bovine serum; IMA, internal mammary artery; VSMCs, vascular smooth muscle cells. Reprinted with permission from REF¹⁰⁷.

Molecular mediators - nitric oxide and beyond:

Endothelial dysfunction, arising from the disparity of endothelium-derived relaxing and contracting factors,¹²¹ plays a major role in atherogenesis and may precede the evolution of atheromatous lesions.¹²² In fact, established CV risk factors, such as smoking, hypercholesterolemia, hypertension and diabetes, enhance intracellular reactive-oxygen species (ROS) production and impair endothelial NO synthesis thus facilitating the development of atherosclerotic plaques.¹²³ Interestingly, in patients with CAD eNOS protein expression and NO release are considerably higher in the endothelial layer of the IMA than the carotid¹⁰³ (FIG. 4) and radial arteries,¹²⁴ respectively, which may blunt key processes of atherogenesis, such as lipoprotein oxidation, leukocyte trafficking and VSMC proliferation and migration due to the abundance of NO.^{123,125–127} Interestingly, the antioxidant paraoxonase 2 (PON2) is highly expressed in vascular tissue specimen of the IMA at both mRNA and protein levels, while its abundance declines in carotid arteries as atherosclerosis progresses.¹²⁸ Aggregating platelets promote endothelium-dependent NO-induced relaxation in the IMA (even in patients with severe CAD), but such aggregates foster constriction in the saphenous vein, highlighting platelet-derived adenine nucleotides and the subsequent activation of purinergic endothelial receptors as potent mediators that eventually promote protection against (athero-)thrombotic complications following CABG.¹²⁹ NO bioavailability declines substantially during the course of ageing, which coincides with altered eNOS function.^130–132 Indeed, endothelial cells undergoing replicative senescence display decreased eNOS protein levels and activity,⁴⁵ which can, at least in part, be restored by stable human telomerase reverse transcriptase expression.¹³³ Of note, the low rate of telomere loss in the intima of the IMA⁴⁶ combined with the lack of senescence-associated β-galactosidase activity⁴⁵ suggests that the IMA resists ageing-associated endothelial changes which would enhance the propensity to atherogenesis.

Figure 4:

The endothelium of the IMA displays high basal eNOS protein expression and NO release. (A) eNOS protein expression (dark purple color) is higher in endothelial cells of the IMA (left) as compared to the carotid artery (right). (B) Representative amperogram showing NO release upon calcium ionophore administration in the IMA and carotid artery, respectively (left). Note that initial rate of NO release and NO peak concentration are elevated in the IMA as compared to the carotid artery (right). Sections in A are haematoxylin counterstained. Values in B are mean±SEM with n=10. Abbreviations: IMA, internal mammary artery; NO, nitric oxide; LU, lumen. Adapted with permission from REF¹⁰³.

Hemodynamic flow in the IMA and its effect on vascular homeostasis:

Flowing blood generates mechanical forces in blood vessels that vary temporally and spatially according to vascular geometry, tissue composition and pulsatility. The flow of blood generates shear stress (frictional drag),¹³⁴ which is sensed by vascular cells and has profound effects on vascular structure, homeostatic mechanisms and disease. Regions of the arterial tree that encounter relatively uniform anterograde (forward direction) flow enjoy relative protection from atherosclerosis because endothelial cells respond to forward flow by inducing multiple cytoprotective mechanisms and suppressing inflammation.¹³⁵ By contrast, arching and bifurcating arteries encounter disturbed flow generating vortices with both retrograde and anterograde components. These structural attributes generate oscillatory shear stress resulting in endothelial dysfunction and inflammation.¹³⁶ Shear stress is also an important trigger of acute and chronic vascular remodelling. Indeed, acute increases in flow velocity lead to increased shear stress which in turn induces nitric oxide, prostacyclin and endothelin-1 to trigger vasodilation, thereby increasing the capacity for blood supply.¹³⁷ Sustained increases in flow can cause remodelling of arteries to increase diameter.¹³⁸ For example, the canine IMA exhibits dilation in response to chronic increases in flow velocity via an arteriovenous fistula, thereby preserving hemodynamic homeostasis.¹³⁹ Similarly, the diameter of the IMA grows in size after implantation into the coronary circulation. The native mechanical environment of the IMA may be a key factor determining its resilience to atherogenesis and thus its high rates of patency in CABG. Measurements of flow by transcutaneous duplex ultrasound in patients with CAD that were awaiting CABG revealed that flow in the IMA is almost exclusively anterograde with only minor retrograde components,¹⁴⁰ a condition that is known to induce cytoprotective and anti-inflammatory mechanisms. Consistent with this, endothelial-dependent increases in flow can be induced experimentally in IMAs of patients that were awaiting CABG.¹⁴¹ Moreover, van Son revealed using ultrasound-based imaging that flow characteristics in the IMA were similar in younger and older patients awaiting CABG, suggesting that IMA hemodynamic characteristics show relative age resistance.¹⁴² Other studies have focused on the responses of conduit vessels to the altered flow conditions that they experience after CABG. Saphenous veins grafted into the coronary artery circulation experience an acute increase in shear stress that induces inflammation and tissue damage.¹⁴³ By contrast, the arterial endothelium – as evidenced by experimental studies in aortic endothelial cells¹⁴³ - resists acute inflammation. Indeed, high shear stress generated by anterograde flow, as it is observed in the native environment of the IMA,¹⁴⁰ induces the expression of an anti-inflammatory enzyme called MKP-1 (MAPK phosphatase-1) that inhibits inflammatory MAPK signalling. IMA grafts also retain the ability to adapt to chronic changes in flow following grafting. Doppler recordings in bypass grafts showed that peak and time-averaged flow velocity is higher in early IMA grafts compared to saphenous veins.¹⁴⁴ However, flow velocity in the IMA is reduced at 1 year post-CABG due to remodelling, while the IMA increases flow capacity more efficiently in response to hyperaemia indicating an improvement in hemodynamic homeostasis.¹⁴⁴ Similarly, a study by quantitative angiography and frequency-domain optical tomography revealed that IMAs that were grafted into coronary arteries at least 10 years prior undergo structural changes, probably in response to altered hemodynamics, and that these long surviving IMA grafts also retain the capacity for endothelium-derived vasodilation.¹⁴⁵ IMA grafts in paediatric CABG secondary to Kawasaki disease were also observed to increase in diameter over time suggesting that they have potential to adapt to chronic changes in hemodynamic conditions.¹⁴⁶ In summary, IMAs encounter flow conditions in their native environment that elicit anti-inflammatory and cytoprotective responses which protect them from inflammatory and injurious insults. Upon grafting, the IMA responds to the coronary flow environment by increasing flow capacity eventually maintaining physiological homeostasis. Collectively, these hemodynamic factors may contribute to the high patency rates of IMA grafts.

A new perspective in atherosclerosis research:Targeting atheroprotective mechanisms

Despite the advances made in the prevention and therapy of atherosclerosis and its associated sequelae, the burden of residual CV risk remains high. Thus, to confront the global pandemic of ASCVD, beyond control of risk factors, protective molecular mechanisms merit consideration as novel therapeutic targets. In this regard, the IMA is a paradigmatic blood vessel that resists atherosclerotic plaque formation (FIG. 5), even in patients with multi-vessel CAD, and exerts atheroprotective effects on the distal coronary circulation when implanted as a bypass graft.^70–75 Ever since 1986, when Loop et al. showed that the use of IMA conduits improves longevity and lowers incidence of MACE,^2,61 endogenous atheroprotective mechanisms deserve focus. Indeed, an increased understanding of the IMA’s resilience to atherogenesis may inform the development of novel medical therapies by unravelling and harnessing pathways that confer atheroresistance on the IMA. To achieve this goal, we need to deepen our currently limited understanding of the molecular and cellular mechanisms involved (FIG. 6). Besides conducting hypothesis-driven research, the application of contemporary technologies such as multi-omics and systems biology approaches merits consideration to unveil previously unsuspected mediators and pathways, and shed light into the IMA’s resilience to atherogenesis. This change in perspective from attacking deleterious risk factors to enhancing endogenous protective mechanisms may ultimately offer novel approaches to combat residual CV risk.

Figure 5:

Atheroprotective features of the IMA. The intimal layer of the IMA is characterized by few fenestrations and low intercellular permeability. Interestingly, intimal hyperplasia is almost absent in the IMA even in patients with progressive atherosclerotic disease. Enhanced eNOS expression contributes to the accentuated NO release upon exposure to calcium-ionophore. Note that the age-dependent decrease in the telomeric length is lower in intimal tissue of the IMA compared to the iliac artery. Anatomically, the IMA is characterized by well-defined IEL which is rich in heparan-sulfate. The tunica media of the IMA is abundant in elastic fibers which VSMCs are embedded in. Interestingly, these cells show low TF and tPA expression and display a remarkable resistance to migration and proliferation, even if exposed to growth factors or pulsatile stress. In stark contrast to vessels with high propensity to develop atheroma, the tunica externa of the IMA is almost devoid of vasa vasorum. Abbreviations: eNOS, endothelial nitric oxide synthase; IEL, internal elastic lamina; IMA, internal mammary artery; NO, nitric oxide; TF, tissue factor; TFPI, tissue factor pathway inhibitor; tPA, tissue-type plasminogen activator.

Figure 6:

Cellular and molecular mechanisms that may underlie the IMA’s resilience to atheroma formation. The anatomy of the IMA favours anterograde flow which may induce MKP-1 expression and in turn limit endothelial cell activation. Also, endothelial cells of the IMA show high basal eNOS expression and pronounced NO release, which hampers ROS formation, which eventually blunts LDL oxidation, endothelial cell activation and VSMC proliferation. Furthermore, telomeric DNA stability may contribute to preserved endothelial function and enhanced eNOS protein expression. Of note, VSMCs lack an array of intrinsic stimuli for cell proliferation, while showing high tPA and TFPI, but low TF expression, which may not only mitigate VSMC proliferation, but may also protect from thrombotic events. Abbreviations: EEL, external elastic lamina; eNOS, endothelial nitric oxide synthase; IEL, internal elastic lamina; LDL, low-density lipoprotein; MKP-1, MAPK phosphatase-1; NO, nitric oxide; PON2, paraoxonase 2; ROS, reactive oxygen species; TF, tissue factor; TFPI, tissue factor pathway inhibitor; tPA, tissue-type plasminogen activator; VSMC, vascular smooth muscle cell.

Limitations

The majority of evidence presented here derives from observational work, and thus might be subjected to treatment allocation bias and/or selection bias. Also, the harvesting technique of SVGs greatly differs from the one used for LIMA grafts and, thus, beside biological factors, intraoperative endothelial injury could contribute to the propensity of vein grafts to degenerate and fail. Importantly, although risk factors of late graft failure resemble those of atherosclerosis,¹⁴⁷ vein graft disease and atherosclerosis represent two different entities with distinct pathophysiologic processes at play. Indeed, our understanding of atherosclerosis and vein graft disease has shifted considerably during the past decades, thus results obtained from former studies using SVGs as comparators to LIMA grafts should be interpreted with caution, since veins are clearly not well suited to be placed in the arterial system. Finally, patency rates not only depend on the biological properties of the grafts, but also on competitive flow and surgical techniques, factors that might be unevenly distributed between patients receiving SVGs and IMA grafts, respectively.

Conclusions

At present, findings from observational and randomized trials support the superiority of the IMA as a bypass graft in terms of patency rates and clinical outcomes compared to the saphenous vein. These observations translate into remarkably low prevalence rates of histology-proven atherosclerotic lesions. Interestingly, the grafted IMA also protects the native coronary artery bed by mitigating the development of downstream lesions, likely through accentuated endothelial NO release.¹⁴⁸ While laminar flow strongly regulates eNOS activity and thus endothelial NO bioavailability, eNOS protein expression and peak NO release is accentuated in the IMA as compared to arteries with similar flow patterns,^100,146 suggesting that endothelial NO biosynthesis is enhanced in the IMA independent from hemodynamic determinants. Indeed, severely hypercholesterolemic mice readily develop atherosclerotic plaques in the straight segment of the carotid artery if its flow is mechanically disturbed and thus retrograde components are present.¹⁵⁰ Yet, IMA graft patency in humans with ASCVD is only marginally affected by altered flow.¹⁵¹ Hence, we postulate that the resilience of the IMA to atherosclerosis results primarily from (1) its unique anatomical architecture which favours age-independent anterograde blood flow (thereby fuelling anti-inflammatory and cytoprotective mechanisms) and (2) the unique biology of its endothelium and VSMCs residing within its wall. A combination of structural and functional features likely contributes to the IMA’s atheroresistence. Yet, how these factors interact and whether they rely on each other to exert their atheroprotective effects requires further study. Despite of years of research into the unique vascular biology of the IMA, the key mechanisms that orchestrate its striking resilience to atherogenesis remain elusive, opening an exciting avenue of research into novel anti-atherosclerotic strategies. Deciphering the IMA’s atheroresistance might be the key to unveil novel molecular targets to eventually overcome the high residual CV risk in patients with ASCVD.

Highlights.

• Although most traditional and emerging risk factors are encountered homogenously in the arterial circulation, the prevalence of atherosclerotic lesions in the internal mammy artery (IMA) is among the lowest in the vasculature

• The anatomic architecture of the IMA favours age-independent anterograde blood flow which fuels anti-inflammatory and cytoprotective mechanisms

• Structural and functional senescence-associated changes of the intimal layer of the internal mammary artery occur at a much lower extent and pace as compared to atherosclerosis-susceptible vessels

• The intimal layer of the IMA displays high endothelial nitric oxide synthase (eNOS) protein expression and accentuated nitric oxide release, which coincides with decreased telomeric shortening during ageing

• Vascular smooth muscle cells (VSMCs) in the IMA resist migration/proliferation, even if exposed to growth factors or pulsatile stress

Non-standard acronyms and abbreviations

ACS

Acute coronary syndrome

ART

Arterial revascularization trial

ASCVD

Atherosclerotic cardiovascular disease

CAD

Coronary artery disease

CABG

Coronary artery bypass grafting

CRP

C-reactive protein

CV

Cardiovascular

eNOS

Endothelial nitric oxide synthase

IEL

Internal elastic lamina

IMA

Internal mammary artery

LAD

Left anterior descending coronary artery

LDL-C

Low-density lipoprotein cholesterol

LIMA

Left internal mammary artery

MACE

Major adverse CV events

MAPK

Mitogen-activated protein kinase

NO

Nitric oxide

PCSK9

Proprotein convertase subtilisin/kexin type 9

PDGF

Platelet-derived growth factor

PON2

Paraoxonase 2

RCT

Randomized controlled trial

ROS

Reactive-oxygen species

SGLT-2

Sodium glucose co-transporter-2

SVG

Saphenous vein graft

VSMC

Vascular smooth muscle cells