Reverse Cholesterol Transport and Atherosclerosis: Heterogeneous Observations by Vascular Bed and Plaque Architecture

Despite improvements in atherosclerotic cardiovascular disease (ASCVD) by targeting reduction of apolipoprotein B-containing lipoproteins, ASCVD remains the leading cause of death. In addition, participants randomized to the intervention arms of contemporary lipid-lowering clinical trials continue to experience ASCVD events at alarming and unacceptable rates.¹ Recent efforts have begun to focus on the active removal of cholesterol from the arterial wall, termed reverse cholesterol transport (RCT). The RCT process is an endogenous mechanism that allows cells to export cholesterol and maintain homeostasis. It has become increasing clear from animal and human studies that impaired RCT can lead to accelerated atherosclerosis and, vice versa, enhanced RCT can limit or prevent atherosclerosis.²

The most studied aspect of RCT has been the initial step of cholesterol movement out of a cell to an acceptor, typically high-density lipoprotein (HDL) particles. Experimental studies in mice have demonstrated that macrophage-specific cholesterol efflux capacity (CEC) is causally related to atherosclerosis.³ Techniques to measure CEC ex vivo have allowed assessment of the capacity of plasma or serum to accept cholesterol from standardized cell lines as a readout of CEC,⁴ a novel marker of HDL function that may relate to ASCVD risk. Indeed, multiple observational studies in large human cohorts have demonstrated an inverse relationship between baseline CEC and incident ASCVD events^{5, 6} and even CV mortality in several high-risk cohorts.^{7, 8} Novel infusion agents that directly manipulate CEC are being tested in human trials to assess whether this strategy of targeting RCT can improve CV outcomes. However, atherosclerosis development is not uniform throughout the vascular tree. Furthermore, arterial plaques are characterized by a spectrum of morphologic features tracking with developmental stage and propensity for rupture and thrombosis. How does RCT impact atherosclerosis in different vascular beds and plaques with architectural diversity? These questions are extremely important if CEC and other measures of RCT will gain clinical relevance as predictors of ASCVD and as targets of therapy.

The study by Shea and colleagues in this issue of ATVB directly addresses this knowledge gap using an observational nested case-control study design in the Multi-Ethnic Study of Atherosclerosis (MESA), a large, middle-aged population-based cohort free of CVD at baseline.[REF] In an analysis comprising 465 matched incident CVD cases, they found a robust inverse association between baseline cholesterol efflux and incident CVD. When outcomes were stratified by vascular territory, the inverse association was driven exclusively by a 28% reduced odds of incident coronary events per standard deviation increase in CEC, while there was no apparent association with incident stroke. The robust inverse association with coronary events is similar to the findings from prior observational studies in low risk populations in the Dallas Heart Study⁵ and EPIC-Norfolk⁶ and is consistent with cross-sectional studies reporting inverse associations with severity of prevalent coronary atherosclerosis.⁹ Though this link between impaired RCT and coronary atherosclerosis has become quite clear and is relatively consistent, only recently have studies begun to hint at how RCT influences coronary plaque architecture. Lipid plaque burden, calcification, and thin cap fibroatheroma all reflect overall plaque burden; however, thin cap fibroatheromas and non-calcified plaque most reliably predict risk of plaque rupture and acute coronary syndrome.¹⁰ In this regard, several studies in humans with increased CV risk or undergoing clinically indicated coronary angiography have revealed that impaired CEC is associated with the prevalence of both thin cap fibroatheroma and non-calcified plaque;^{11, 12} whereas large and small cohorts have revealed a relative lack of association between CEC and calcified coronary plaque.^{5, 13}

The more intriguing finding from this study was the lack of association with cerebrovascular disease. Baseline efflux not only was unrelated to incident stroke but also was essentially unrelated with progression of carotid plaque as determined by B-mode ultrasound. Although cerebrovascular and coronary artery disease share many common risk factors, there are several important differences in the atherobiology of these vascular beds. The phenotype of cerebrovascular disease is more heterogeneous, comprising 1) ischemic plaque-mediated brain infarction, 2) ischemic embolic-mediated brain infarction, or 3) hemorrhagic brain infarction from arterial wall (not plaque) rupture. Processes related to the latter two outcomes are clearly distinct from the classical plaque-mediated process; yet many clinical studies assessing cerebrovascular disease do not disentangle these disparate stroke events and reduce the ability to discern factors specifically linked to cerebrovascular plaque development. Furthermore, carotid atherosclerosis in human studies is typically assessed by carotid intima media thickness or presence of carotid plaque by ultrasound or magnetic resonance imaging. In multiple large cohorts, carotid intima media thickness is actually a weak predictor of stroke,¹⁴ whereas the presence of carotid plaque is a robust predictor of ischemic stroke events.¹⁵ Not surprisingly, findings from human studies assessing associations between cholesterol efflux with cerebrovascular disease have been heterogeneous. Prior cross-sectional reports of convenience samples have revealed inverse associations with both prevalent carotid intima media thickness⁹ as well as with prevalent carotid artery stenosis.¹⁶ In contrast to these inverse associations in small and restricted study designs, the current study in this issue of ATVB now reports no meaningful association with incident stroke or with carotid plaque progression in a large, population-based cohort followed longitudinally over many years. Importantly, the investigators were able to parse out the different stroke subtypes (of ~200 incident strokes, ~120 were not due to hemorrhage or embolic events) and convincingly show a lack of association with non-embolic ischemic stroke. In the Dallas Heart Study, which was a younger population with fewer events, efflux capacity was inversely associated with incident stroke in secondary analyses; however, the number of stroke events was small (N~30) and not parsed by stroke subtype and the efflux assays used in that study versus this study differed modestly.⁵ Thus, the current study in MESA is the largest and most epidemiologically rigorous to date with regard to assessing the contribution of RCT to cerebrovascular plaque development.

In summary, there is likely no significant link between RCT, as reflected by macrophage-specific cholesterol efflux capacity, and cerebrovascular plaque and ischemic stroke. Epidemiologic studies to date demonstrate consistent associations between impaired cholesterol efflux and coronary atherosclerosis, specifically non-calcified plaques or thin-cap fibroatheromas. Further studies in large unselected cohorts such as the current analysis in MESA are warranted to elucidate the role of RCT in atherosclerosis in other vascular beds such as aorto-iliac disease, lower extremity peripheral arterial disease, and transplant vasculopathy.

Figure 1.