Review
In the Beginning, Lipoproteins Cross the Endothelial Barrier
2024 Volume 31 Issue 6 Pages 854-860
Details
2024 Volume 31 Issue 6 Pages 854-860
Atherosclerosis begins with the infiltration of cholesterol-containing lipoproteins into the arterial wall. White blood cell (WBC)-associated inflammation follows. Despite decades of research using genetic and pharmacologic methods to alter WBC function, in humans, the most effective method to prevent the initiation and progression of disease remains low-density lipoprotein (LDL) reduction. However, additional approaches to reducing cardiovascular disease would be useful as residual risk of events continues even with currently effective LDL-reducing treatments. Some of this residual risk may be due to vascular toxicity of triglyceride-rich lipoproteins (TRLs). Another option is that LDL transcytosis continues, albeit at reduced rates due to lower circulating levels of this lipoprotein. This review will address these two topics. The evidence that TRLs promote atherosclerosis and the processes that allow LDL and TRLs to be taken up by endothelial cells leading to their accumulation with the subendothelial space.
Cholesterol accumulation within the arterial wall and the presence of lipid-filled macrophages, foam cells, is the pathological signature of atherosclerosis. This differentiates atherosclerosis from vasculitis. In his Connor lecture in 1954, Irvine Page reviewed atherogenesis1). He noted that the primary event, cholesterol accumulation, preceded and then activated processes leading to white blood cell (WBC) infiltration into the arterial wall. Clinical data confirming that cholesterol, primarily low-density lipoprotein (LDL), reduction prevented cardiovascular events re-enforced our understanding of vascular lipid infiltration as the primary cause of atherosclerosis. But efforts to understand this process, the movement of lipoproteins from the circulation into the arterial wall stalled.
Inflammation, a detour. The rapid advances in the science of inflammation and WBC biology led to decades of focus on the secondary processes that occur in lipid-rich arterial lesions. Widely quoted reviews by Ross promoted atherosclerosis as an inflammation disease2). This led to a focus on WBC biology, especially macrophages, and studies in this area have dominated atherosclerosis research for several decades, see reviews3, 4). Several WBC subsets have been shown to modulate vascular inflammation and changes in many WBC genes reduce atherosclerosis progression in mouse models. Efforts to reduce cardiovascular events by targeting WBCs and inflammation in general are on-going and may in the future become a more accepted approach to prevent vascular occlusive events, primarily in patients with documented disease.
Endothelial transport of fatty acids. All endothelial cells contain receptors for the uptake of lipids and lipoproteins allowing their movement across these cells. In capillaries, the major source of fatty acids especially during the postprandial period is the release of non-esterified fatty acids (NEFAs) due to lipoprotein lipase (LpL) hydrolysis of triglycerides (TGs). Most TG is within the core of the lipoprotein due to its hydrophobicity, but some migrates to the lipoprotein surface where it is LpL-accessible. Aside from NEFAs, this reaction leads to the transfer of lipids and apoproteins to HDL. For this reason, LpL deficiency, even in animals with no cholesteryl ester transfer protein, leads to low HDL levels5, 6). How the released NEFAs traverse the endothelial cell barrier is still unclear but could be due to changes in the endothelial cell barrier function allowing paracellular transcytosis, as seen with lipolysis in isolated vessels7), or NEFA transcytosis could involve receptor or non-receptor mediated transcytosis.
We have studied the movement of non-lipoprotein derived NEFAs, which are increased due to adipocyte intracellular lipolysis during fasting. These NEFAs are increased during fasting when low insulin levels lead to activation of intracellular lipases. CD36, a member of the scavenger receptor gene family, associates with NEFA via a specific molecular interaction and regulates movement of NEFA into heart, skeletal muscle, and adipose8). Copious data have implicated CD36 in NEFA uptake by adipocytes and cardiomyocytes, reviewed in9), but much of the actions of this receptor are due to its expression in endothelial cells. In the heart, CD36 is robustly expressed in endothelial cells for this reason; after floxing this gene, we explored how endothelial cell-specific CD36 knockout affected NEFA uptake into heart, muscle, and adipose tissue10). At the levels of NEFA that occur with fasting, endothelial cell CD36 deletion led to reduced NEFA uptake and hearts reverted to greater glucose utilization. Others have implicated endothelial CD36 in uptake of lipid emulsions by brown adipose tissue11). In contrast to uptake of albumin-associated NEFAs during fasting, the exorbitant local NEFA concentrations that occur during capillary chylomicron lipolysis are likely to saturate this transport system, a reason that uptake of chylomicron-derived fatty acids does not appear to involve CD36 12). The pathway involved in movement of these NEFAs across capillary endothelial cells is unclear, some thoughts on how this might occur are mentioned below.
Lipoprotein uptake by endothelial cells. The question of how endothelial cells interact with, take up, and then transcytose LDL has been studied for more than four decades. In 1979, Fielding and colleagues reported that proliferating cells internalized LDL and degraded the LDL in lysosomes, leading to changes in cellular cholesterol metabolism13). This is due to expression of LDL receptors, which are upregulated in proliferating cells. In contrast, with contact inhibition, this regulation of intracellular cholesterol metabolism was no longer modulated by LDL, which the investigators interpreted as a response to maintain an endothelial barrier to circulating LDL. Cationized LDL appeared to be internalized by a different receptor, likely a scavenger receptor. This group also reported that a separate receptor was responsible for endothelial uptake of TRLs14).
A series of studies in humans implicated changes in endothelial cell biology with increased circulating TRL levels, e.g. those that occur during the postprandial period. For example, Vogel et al. reported that a high fat meal impaired vascular dilatation15). Others reported that Intravenous lipid emulsions also alter endothelial function and affect blood pressure16). These changes in vascular function might be explained by the observation that arterial endothelial cells acquire lipid droplets during the postprandial period17). More recently, Kim et al. showed that accumulation of lipid droplets in endothelial cells due to adipose triglyceride lipase (ATGL) knockout alters eNOS activity18). The molecular events that mediate postprandial endothelial lipid droplet accumulation are described below.
Endothelial cell lipoprotein receptors and atherosclerosis: Can the primary event, infiltration of lipoproteins into the arterial wall, be blocked? Although some hypothesized that this process was due to endothelial cell damage19), the majority of atherosclerotic plaque lesions contain an intact endothelium20). Thus, the processes that allows transendothelial movement of lipoproteins must involve lipoprotein movement through or between endothelial cells. The movement of large proteins and certainly macromolecules like lipoproteins between endothelial cells is unlikely. Local conditions could make the endothelial barrier leaky, e.g. lipolysis leading to extremely high local concentrations of fatty acids and lysolipids might lead to defective barrier function21). However, LpL and its requisite binding site, GPIHBP1, primarily reside on capillaries22), making lipolysis induced arterial dysfunction unlikely. Multiple studies tracking labeled lipoproteins have also failed to demonstrate paracellular lipoprotein transport (reviewed in23). Thus, uptake of TRLs by endothelial cells is a likely cause of postprandial endothelial cell dysfunction.
Lipoproteins could cross the endothelial cell barrier via non-specific uptake into transcellular vesicles, as well as via a receptor-mediated saturable process24). Endothelial cells have multiple vesicles that form transcellular channels into which lipoproteins could passively move across the cells along with albumin and other plasma proteins. The introduction by Lee and his colleagues of the use of total internal reflection fluorescence microscopy (TIRF), allowed the tracking of lipoprotein exocytosis from the basolateral side of living cells23). Their studies in cultured cells established that specific receptors mediate lipoprotein transcytosis.
Scavenger receptor-BI (SCARB1, SR-BI), is a well-known lipoprotein receptor best associated with selective uptake of lipid from HDL. Knockout of SR-BI in apoE deficient mice leads to marked hyperlipidemia and atherosclerosis associated with a doubling of plasma cholesterol due to increased VLDL and large HDL25). Mice with reduced expression of both SR-BI and the LDL receptor knockout had greater circulating LDL and more disease26). SR-BI deficiency also increased atherosclerosis in humans, despite an increase in HDL and, unlike in mice, apoB lipoprotein levels were not affected27).
SR-BI mediates TRL metabolism. Van Berkel and colleagues had shown that SR-BI was a receptor for chylomicron/remnants28), indicating that it could bind to apoB-containing lipoproteins. SR-BI deficiency increases postprandial TG levels 4-fold in mice28) and SR-BI variants in humans increase TG29-31). Using isolated aortas, Armstrong et al.32) showed accumulation of LDL, but not dextran, in the arterial wall. They then used TIRF microscopy to show that LDL transcytosis was blocked by HDL and knockdown of SR-BI. Could eliminating endothelial cell SR-BI reduce atherosclerosis? Yes! Endothelial cell specific knockout of SR-BI protected hyperlipidemic mice from atherosclerosis33). This result suggests that the inhibition of LDL transcytosis more than compensates for the possible anti-atherogenic actions of SR-BI to mediate beneficial actions of HDL. The dichotomy in which both overexpression and knockout of endothelial SR-BI reduces atherosclerosis, might be explained by a need for many copies of the receptor to multimerize with HDL that has many molecules of apoAI, whereas LDL has but one copy of apoB to bind to SR-BI.
SR-BI and eNOS: The greater disease with whole body genetic loss of SR-BI could have resulted from a defect in reverse cholesterol transport or some non-lipoprotein mediated effects downstream of SR-BI. Shaul and colleagues had been studying how SR-BI interaction with HDL affected endothelial function. SR-BI functions as a signaling molecule; its activation in endothelial cells leads to greater eNOS activity, reduced expression of monocyte adhesion molecules, and other anti-atherogenic changes34). For this reason, it was not surprising that endothelial cell overexpression of SR-BI reduced atherosclerosis35).
Shear stress also regulates endothelial cell function through a SR-BI-eNOS signaling pathway to prevent atherosclerosis36). SR-B1 plays an important role in HDL-induced COX-2 expression and PGI2 release in endothelial cells through upregulating PI3K–Akt–eNOS signaling37). As noted above, generation of high levels of free fatty acids in humans leads to reduced levels of nitric oxide and decreased endothelium-dependent vasorelaxation, indicating a strong relation between lipids and disrupted eNOS activity. Kim et al 18) and Boutagy et al.38) both studied the effects of endothelial cell specific adipose TG lipase (ATGL) knockout on endothelial cell function and lipid droplet accumulation. ATGL knockout increased lipid droplet formation and reduced eNOS activation. Pharmacological prevention of lipid droplet formation reversed suppression of NO production in cell culture and in vivo and blunted blood pressure elevation in response to a high-fat high-salt diet18). In the second of these studies, atherosclerosis was increased along with the reduction in eNOS activity.
Triglyceride rich lipoproteins. In addition to LDL, TRLs track with atherosclerosis risk. This relationship is seen in cross sectional studies of humans39), genetic analysis of TG-modulating genes40), and studies of postprandial lipemia41). In addition, one interpretation of the recent PROMINENT study is that fibrate conversion of VLDL to LDL confirmed that these two lipoproteins were equally atherogenic42).
More than 50 years ago, Silversmit proposed that postprandial lipoproteins were as a cause of atherosclerosis. His 1970 Duff lecture was entitled “Atherogenesis: a postprandial phenomenon”43). His idea was merged with the observation of advanced disease in mice and humans with defects or deletion of apoE leading to accumulation of cholesterol-rich particles thought to be non-catabolized chylomicron remnants. The remnant hypothesis suggested that lipolysis along the arterial surface was atherogenic either because it led to the creation of smaller remnant lipoproteins that were more likely to cross the endothelial cell barrier, or to toxic lipolysis products44). Neither of these hypotheses is likely to be correct as lipolysis occurs in capillaries but not large arteries.
Endothelial uptake of chylomicrons. How do TRLs affect the arterial wall? In the mid-20th Century before the development of genetically modified mice, a common model used for studies of atherosclerosis was the cholesterol fed rabbit. In an effort to create a model to study why diabetes leads to more atherosclerosis in humans, Duff and McMillan used alloxan to destroy islet cells and create diabetes in rabbits. We imagine that they were surprised when these rabbits developed extraordinarily increased cholesterol levels, but less atherosclerosis45). Thirty years elapsed before an explanation of this seemingly contradictory observation in rabbits was reported. Nordestgaard and colleagues published several papers re-examining the development of atherosclerosis in diabetic rabbits. They showed that the enormous lipoproteins obtained from diabetic high cholesterol-fed rabbits with circulating triglyceride and cholesterol levels in excess of 4000 mg/dL failed to track with arterial accumulation of lipids. They concluded, likely correctly, that the particles were unable to cross the endothelial cell barrier46). This concept has been applied to suggest that human chylomicrons, very different particles, are also not-atherogenic. Prior to the studies by Nordestgaard, others had shown that diabetic rabbit lipoproteins do not cause macrophages to become foam cells, i.e. develop lipid droplets47), suggesting that they are unable to bind to cell surface lipoprotein receptors. Unlike particles from diabetic cholesterol-fed rabbits, chylomicrons bind to receptors.
While a number of studies had shown defects in vascular function related to postprandial lipemia, evidence that chylomicrons entered and then affected endothelial cells was lacking. Using confocal microscopy, Kuo et al. reported that postprandial lipemia led to lipid droplet accumulation within arterial endothelial cells17). Because as noted above, local lipolysis of chylomicrons was unlikely to occur along the arterial surface, we explored whether mice with LpL deficiency would also develop postprandial endothelial cell lipid droplets. They did. And, LpL knockout mice had fasting lipid droplets, likely a reflection of their fasting chylomicronemia48). Chylomicron uptake was also responsible for the development of lipid-filled skin macrophages. SR-BI, the same receptor responsible for LDL transcytosis, mediated chylomicron uptake and led to lysosomal targeting. Therefore loss of endothelial cell SR-BI not only reduced LDL transcytosis, but it likely also prevented any toxic effects of postprandial lipemia.
The interaction of chylomicrons with SR-BI differs from that of LDL, Using a RNA inhibition screening assay, Kraehling et al.49) discovered that activin-like kinase 1 (ACVLR1, ALK1) mediated LDL uptake and transcytosis of LDL. Unlike LDL, chylomicrons are not internalized by ALK1. Thus, either the size of the lipoprotein or the length of apoB confers some specificity for lipoprotein binding. Studies of proteoglycan binding showed that apoB48 and apoB100 particles bound to heparin-Sepharose with equal affinity50). In this regard, studies by Flood et al. suggested that the C-terminal region of apoB altered apoB confirmation and exposure of a proteoglycan binding region in the N-terminal non-lipid containing portion of the molecular51). Thus, apoB structure is likely affected by the length of the protein, the size of its carrier lipoprotein, and other associated proteins (apolipoproteins, such as apoC3 52)).
Lipid droplets and inflammation. If lipid uptake inflames endothelial cells as recently suggested53), then the biology of how TRLs relate to atherosclerosis requires re-examination. Lipid droplets are intracellular organelles comprised of a neutral lipid core (mostly TGs and cholesteryl esters) surrounded by a phospholipid monolayer with which a series of lipid droplet-associated proteins interact54). Lipid droplet biogenesis following the uptake of external lipid serves the double purpose of preventing lipotoxicity and storing lipids for catabolic or anabolic use upon cellular need. In addition, lipid droplet biogenesis can be triggered by a variety of stressors including nutrient deprivation, hypoxia, or inflammation. Stress-triggered lipid droplets promote cell survival by helping maintain redox and energy homeostasis, as well as by sequestering toxic lipids55). In addition, stress-induced lipid droplets constitute a depot of bioactive lipids involved in the regulation of inflammation and immunity (see56) for a comprehensive review). Although little is known about the role of endothelial lipid droplets in the immune and inflammatory responses that characterize atherosclerosis, as noted above, recent studies have linked their presence to endothelial dysfunction18).
Clinical implications of TRL-EC interactions. The clinical approach to reducing atherosclerosis has rightly focused on the use of medications that increase LDL receptors and the liver clearance of LDL. Although this approach has been remarkably successful, residual risk of cardiac events remain. Moreover, lipoprotein a (Lpa), a more potent risk factor than LDL57), is minimally affected by most LDL reducing medications. Rather than lowering the levels of atherogenic lipoproteins, is it possible that knowledge of the required receptors needed for their transcytosis could lead to novel therapies blocking cholesterol entry and accumulation in the artery wall?
Genetic analysis implicates defects in the lipolysis pathway and hence increased postprandial lipemia in atherosclerosis. Furthermore, it is well established that greater CVD occurs in populations ingesting a high saturated fat diet. While the most obvious way to reduce postprandial lipemia is dietary, there may be a role for targeting those patients most likely to benefit from this approach, likely those with low HDL levels. An alternative might be to block pathways, such as SR-BI, involved in chylomicron uptake into endothelial cells.
Although hyperlipidemia is a major treatable risk factor for CVD, not every patient with high circulating levels of LDL develops disease. Could it be that some people who have genetics or environmental factors that prevent lipoprotein endothelial cell interaction are protected from the harmful effects of circulating LDL and postprandial lipemia. If identified, such patients would be spared the life-long exposure to statins and restricted diets prescribed for patients with hyperlipidemias.
Decades of research into atherogenesis have illustrated the importance of hyperlipidemia as the cause of most atherosclerosis. Several lines of research have led to dead ends, awaiting new methods. One such area is the understanding of the interactions of lipoproteins with vascular endothelial cells. More recent research has uncovered the basic biology of how LDL crosses the endothelial cell barrier, the process that initiates lipid accumulation in the vascular wall. Parallel studies have explored how fatty acids and TRLs affect the biology of endothelial cells and the underlying vasculature tissue.
Using inducible LpL knockout mice, created by Hiro Yagyu58) we are studying how chylomicrons increase vascular inflammation and atherosclerosis development. In addition, we will define why LDL binds to both SR-BI and Alk1, while only SR-BI serves as a receptor for endothelial chylomicron uptake. Perhaps in the future, atherosclerosis prevention will focus on chylomicron as well as LDL reduction and methods to block endothelial uptake and transcytosis of apoB-containing lipoproteins.
Studies of chylomicron endothelial cell interactions are funded by grants HL151328, HL160470-01, HL045095, and HL164949 from the National Heart Lung and Blood Institute (USA).
None.
