HDL biogenesis revisited: how desmocollin-1 could sabotage reverse cholesterol transport in the arterial wall

This editorial refers to ‘Desmocollin 1 is abundantly expressed in atherosclerosis and impairs high-density lipoprotein biogenesis’†, by H.Y. Choi et al., on page 1194. High-density lipoproteins (HDLs) are highly heterogeneous complexes of hundreds of lipids and ∼100 proteins that can be classified according to their composition, intrinsic density, size, and functions.1 They are often defined as lipoproteins containing apolipoprotein (apo) A-I, the most abundant protein in HDL. HDL from healthy people exerts multiple antiatherogenic effects including cellular cholesterol efflux, anti-inflammatory, antithrombotic, and vasculoprotective activities. These normal functions of HDL may explain the established inverse relationship of HDL-cholesterol levels and incident cardiovascular disease.2 However, simply raising HDL-cholesterol levels is not a successful strategy to prevent cardiovascular events: disappointing results of phase III studies conducted with niacin or cholesterylester transfer protein (CETP) inhibitors in patients with high cardiovascular risk3–5 suggest that enhancing HDL functionality is rather the path to explore. Indeed, beneficial effects of HDL could not exist without HDL biogenesis, the molecular events that leads to lipidation of apo A-I with cellular membrane lipids by the key transporter ABCA1 and the subsequent esterification of cholesterol by lecithin:cholesterol acyltransferase (LCAT).6 In the arterial wall, the majority of apo A-I can be found in the lipid-poor state rather than in mature particles.7 In atherosclerotic plaques, lipid-poor apo A-I molecules undergo oxidative modifications, lose the capability to interact with the ATP-binding cassette A1 (ABCA1) transporter, and accumulate through mechanisms that remain to be defined.8 These observations support the concept that HDL biogenesis is impaired in the context of atherosclerotic plaques. In this issue of the journal, Choi and colleagues9 provide a seminal contribution to the understanding of how HDL biogenesis is prevented by the interaction of apo A-I with a subtype of cholesterol- and sphingomyelin-rich membrane microdomain, enriched in desmocollin-1 (Dsc1) and other desmosomal cadherins such as desmoglein 1 and 3.10 They isolated and characterized these microdomains from human fibroblasts with a new method involving: (i) detergent-free isolation of membranes on sucrose gradients in the presence of apo A-I; (ii) immunoprecipitation of apo A-I-associated microdomains after sonication in the absence of cross-linking agents; and (iii) lipidomic and proteomic analyses of the composition of microdomains. Co-immunoprecipitation of apo A-I with Dsc1 was also obtained in the absence of ABCA1 (in fibroblasts from Tangier disease patients) and was confirmed in human embryonic kidney (HEK) cells overexpressing human Dsc1. In both cell types, apo A-I co-localized with Dsc1 as shown by confocal microscopy. Using transient overexpression of Dsc1 or ABCA1 in HEK cells, the authors concluded that both proteins are able to increase apo A-I binding independently, but that cholesterol efflux was uniquely stimulated through ABCA1 expression and not by Dsc1 overexpression. While Dsc1 overexpression per se did not reduce ABCA1-mediated cholesterol efflux to apo A-I, interference with endogenous Dsc1 expression by stable transfection of short hairpin RNA or CRISPR/Cas9 constructs increased ABCA1-mediated cholesterol efflux. This was paralleled by an increase in ABCA1 protein, which was apparently stabilized by the increased availability of apo A-I for binding to ABCA1 or increased availability of membrane cholesterol following reduced Dsc1 expression. Both mechanisms have been shown to extend the half-life of this rapid turnover protein.11 , 12 The concept that Dsc1-containing microdomains sequester apo A-I and/or membrane lipids efficiently and impact ACBA1-mediated biogenesis is novel (Take home figure). As the authors justly point out, the balance between Dsc1 and ACBA1 expression at the cell surface could regulate the intensity of HDL biogenesis. One limitation of this study is that HDL biogenesis was described essentially by cellular cholesterol efflux assays, but the process of apo A-I lipidation also involves phospholipid efflux and the generation of nascent apo A-I-containing particles of different sizes, which could be detailed in future studies.13 The determinants of apo A-I binding to Dsc1 should also be investigated because several important questions remain unanswered. Which apo A-I residues are involved in binding to Dsc1? Is the affinity of apo A-I for Dsc1 higher than for the ABCA1 transporter? Does oxidized apo A-I bind with increased affinity to Dsc1? Do other desmocollins bind apo A-I as well as Dsc1? The interplay between Dsc1 and ABCA1 for apo A-I binding deserves detailed studies in relevant arterial wall cell types. Proposed role of desmocollin-1 (Dsc1) in impaired HDL biogenesis from apo A-I in the atherosclerotic plaques. Upper panel: in the normal arterial wall, lipid-poor apo A-I is able to complete HDL biogenesis, the first step in the antiatherosclerotic reverse cholesterol transport process. This occurs through binding to the ATP-binding cassette type A1 (ABCA1) transporter of macrophages and cellular cholesterol efflux via interactions with sphingolipid- and cholesterol-rich membrane microdomains generated by ABCA1 (pleated membrane in the figure). Discoidal particles of various sizes can be formed by ABCA1-mediated cellular efflux, and are optimal substrates for esterification of free cholesterol by lecithin:cholesterol acyltransferase (LCAT) and maturation of nascent HDL into larger spherical particles and their remodelling by phospholipid transfer protein (PLTP) which can occur in the circulation. Efficient HDL biogenesis and reverse cholesterol transport are key features of atherosclerosis prevention. Lower panel: in the atherosclerotic plaque, increased levels of Dsc1 have been found co-localized with apo A-I by Choi et al. 9 The cause of increased Dsc1 expression is uncertain. Dsc1 is present in sphingolipid- and cholesterol-rich microdomains in vitro. As Dsc1 was shown as a bona fide binding site for apo A-I binding in vitro and consequently reduces ABCA1-mediated cholesterol efflux, this interaction may contribute to reduced HDL biogenesis in the arterial wall. Increased levels of Dsc1 in atheroma could retain lipid-poor apo A-I and reduce its interaction with ABCA1 microdomains. As oxidation of apo A-I in the arterial wall impairs its cholesterol efflux capacity, prior oxidation of apo A-I may favour binding to Dsc1 or, conversely, binding to Dsc1 may favour its oxidation to prevent HDL biogenesis. As such, Dsc1–apo A-I interactions may be a key mechanism by which HDL functionality is lost in the context of atherosclerotic plaques, and interrupting this deleterious interaction could provide a new therapeutic strategy for enhancing reverse cholesterol transport and dampen atherosclerotic plaque progression. Intriguingly, the authors have shown that the human macrophage THP-1 cell line expresses Dsc1. This is unexpected for a cell type that normally does not assemble desmosomes. It is crucial to establish if Dsc1 expression is regulated by cholesterol levels and/or inflammatory signals in macrophages and if a similar effect of its knock-down on ABCA1 expression and cellular cholesterol efflux can be observed in macrophage-derived foam cells. All these questions will require further investigations as they could support a deleterious effect of Dsc1 expressed in atherosclerotic plaques. Indeed, Choi et al. demonstrated abundant expression of Dsc1 and its co-localization with CD68 in human atherosclerotic plaques. While the identity of the CD68-positive cells is debatable, as a significant fraction of those may derive from smooth muscle cells and not macrophages in human arteries,14 the finding is of importance. First, Dsc1 expression increased with the progression of atherosclerosis, from initial lesions to complex lesions with a necrotic core and calcifications. Secondly, the increase was found in coronary and carotid specimens. Thirdly, apo A-I was co-localized with Dsc1 in intermediate and advanced lesions. As the expression of Dsc1 is very high in the skin but very low in other tissues as reported by expression databases,15 Dsc1 may represent a selective marker for atherosclerosis burden or progression, especially if it is released from the vascular wall into the circulation. Indeed, increased Dsc1 expression was shown previously in various cancer cell types and is a prognostic marker in some of them.16 The impact of increased Dsc1 in atheroma is a matter of speculation as this study does not provide experimental evidence of Dsc1–apo A-I binding in the arterial wall. As lipid-poor apo A-I is abundant in atheroma and because Dsc1 and apo A-I are co-localized, it could be that Dsc1 expression is a driver of apo A-I retention, in which case targeting Dsc1 would be a valid approach to free apo A-I from its Dsc1-binding site. Nevertheless, other possibilities do exist. Lipid-free apo A-I in the vascular wall easily gets oxidized by the myeloperoxidase system. Is apo A-I oxidized first and in turn becomes a—perhaps better—ligand for binding to Dsc1? Dsc1 could capture already ‘exhausted’ oxidized apo A-I or native apo A-I to favour its oxidation in the arterial wall by preventing its lipidation by ABCA1 (Take home figure). The question is not trivial as prior oxidation of apo A-I makes it inefficient for HDL biogenesis.8 As the expression of Dsc1 is greatly increased in the arterial wall, it appears logical to target its interaction with apo A-I to enhance apo A-I availability for HDL biogenesis by ABCA1. A proof of concept was provided by the authors in an antibody-blocking experiment with human fibroblasts. Increased cholesterol efflux to apo A-I was shown in the presence of an anti-Dsc1 antibody directed against the extracellular domain of Dsc1, in a region that includes part of the extracellular cadherin (EC) repeat 4 and all EC5. These results are in agreement with apo A-I binding studies performed with EC deletion constructs pointing to a role for EC2 and EC5. A similar blocking antibody experiment conducted in macrophages or smooth muscle cell-derived foam cells would provide further support to this strategy. Another limitation of this study is the absence of an assessment of Dsc1 expression in mouse models of atherosclerosis. While Dsc1 is dispensable for desmosome formation and does not lead to embryonic lethality, it appears important for barrier function of the epidermis, as shown by the multiple skin defects in Dsc1 knockout mice, including chronic dermatitis and ventral hair loss in almost all animals between 6 weeks and 6 months.17 Nevertheless, Dsc1– /– mice are fertile, can be weaned successfully, and survive until adulthood, allowing for diet-induced atherosclerosis studies. A myeloid promoter-specific Dsc1 knock-out model with reduced macrophage Dsc1 would be useful to avoid skin problems already associated with atherogenic diets in addition to the severe phenotype of Dsc1 mice. Additionally, the phenotype of heterozygous Dsc1 mice was not reported, but could provide an intermediate phenotype for breeding on hyperlipidaemic mouse backgrounds. It will be important to ascertain whether it is possible to block Dsc1–apo A-I interaction in the arterial wall without affecting its normal role in the skin, i.e. leaving interactions with other desmosomal cadherins intact. It is noteworthy that human IgG autoantibodies against the desmoglein-3 protein cause its clathrin-independent endocytosis and disassembly of desmosomes in the skin, causing pemphigus vulgaris.18 Anti-Dsc1 antibodies could lead to raft-dependent endocytosis of the protein, freeing apo A-I from this ‘sticky patch’ of Dsc1, with tentatively beneficial effects in the plaque, but unpredictable effects in the epidermis. What is the normal physiological role of apo A-I binding to Dsc1, if any? Desmosome dynamics is undoubtedly influenced by the integrity of membrane microdomain (lipid rafts). Membrane cholesterol levels have been shown to be important for desmosome assembly and stability, as cholesterol-sequestering drugs and extraction of membrane cholesterol can prevent their formation or cause their disassembly.19 , 20 The ability of interstitial fluid apo A-I to interact with cholesterol-rich membrane microdomains and extract membrane lipids could be deleterious to the stability of specialized epithelial tissues where desmosomes are crucial to mechanical and barrier properties, such as the myocardium, intestinal epithelium or epidermis. Thus, it is tempting to speculate that Dsc1 has evolved as a protein that prevents unwanted destabilization of desmosomes by apo A-I and that this applies to other members of the desmocollin family. While the reductionist approach of raising blood HDL-cholesterol has been disappointing in the clinical setting, a better understanding of HDL biology in the vascular wall and especially how their functions are regulated by local factors such as desmocollin-1 may provide new therapeutic targets to prevent their dysfunction or unlock their therapeutic potential. As desmocollin-1 is increased in atherosclerotic plaques, it may be considered a new cause of impaired HDL biogenesis in the arterial wall. Hence, the work by Choi et al. is a good example of how integration of classical biochemical methods (cell fractionation), modern analytical methods (lipidomics and proteomics), cell-based assays, and clinical samples can open the treasure chest of new therapeutic targets for atherosclerotic cardiovascular diseases. Conflict of interest: none declared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.