Pathways and Molecular Mechanisms Governing LDL Receptor Regulation

Abstract

Clearance of circulating plasma low density lipoprotein (LDL) cholesterol by the liver requires hepatic low density lipoprotein receptor (LDLR). Complete absence of functional LDLR manifests in severe hypercholesterolemia and premature atherosclerotic cardiovascular disease. Since the discovery of the LDLR fifty years ago by Brown and Goldstein, all approved lipid-lowering medications have been aimed at increasing abundance and availability of LDLR on the surface of hepatocytes to promote removal of LDL particles from the circulation. As such a critical regulator of circulating and cellular cholesterol, it is not surprising that LDLR activity is tightly regulated. Despite over half a century’s worth of study, there are still many facets of LDLR biology that remain unexplored. This review will focus on pathways that regulate the LDLR and emerging concepts of LDLR biology.

Introduction

Cholesterol is the primary sterol in most higher animals, with essential roles and function in membrane structure and cellular signaling. Cholesterol can be obtained through de novo biosynthesis or via dietary absorption. As a hydrophobic, water-insoluble molecule, cholesterol is transported with other lipids in lipoprotein particles. Dysregulation of cholesterol balance is implicated in the pathogenesis of multiple diseases, such as familial hypercholesterolemia (FH), Alzheimer’s disease, and most notably cardiovascular disease (CVD). The low-density lipoprotein receptor (LDLR) plays a crucial role in cholesterol metabolism and homeostasis through its ability to clear LDL particles from plasma, which carries approximately 70% of all plasma cholesterol.

High levels of plasma LDL cholesterol are the major risk factor for coronary heart disease (CHD) and atherosclerotic cardiovascular disease (ASCVD). Complete absence or defective LDLRs due to mutations in the LDLR gene cause FH, characterized by an autosomal dominant inheritance of hypercholesterolemia and premature ASCVD (a list of mutations can be found in Table 1)^1–4. Of these mutations, all result in FH, and most result in loss of LDLR function, either through complete loss of protein or production of a modified protein which is nonfunctional. The exception to this are Class V mutations where LDLR can still bind cargo for internalization, but binds with weak affinity or cannot release the cargo once internalized, prompting protein turnover¹. As such, some LDL is internalized but far less than occurs with typical LDLR recycling. This may account for potentially uncharacterized mild cases of hypercholesterolemia across the human population.

Table 1.

Class ID	LDLR Phenotype	Commonality	Regulatory Target/Defects
Class I	No detectable protein	First identified, easily diagnosed	Premature stop codon/instability of peptide
Class II	Transport Defective: Protein produced, but accumulates and does not leave ER or Golgi	Common	Glycosylation and protein folding defects
Class III	Binding Defective: LDLR makes it to the cell surface but fails to bind ApoB	Uncommon	ApoB or ApoE mutations
Class IV	Internalization Defective: Does not localize to clathrin pits and is not endocytosed	Uncommon	Mutations in Cytoplasmic Domain
Class V	Recycling Defective: LDLR binds cargo and is internalized, but may not release cargo and is not recycled to cell surface	Common	Mutations in Ligand Binding Domain and Cytoplasmic Domain
Class VI	New proposed class for variations of LDLR that do not thread in the PM correctly.	Rare	Mutations in TM domain and other structural mutations

List of Class grouped mutations in the LDLR protein, the resulting LDLR phenotype, the frequency of incidence of the class of mutation, and the regulatory pathway or mechanism that is specifically affected.

Current treatments for hypercholesterolemia target three principal mechanisms of cholesterol homeostasis: (1) the Statin class of drugs inhibit 3-hydroxy-3-methylglutaryl (HMG) CoA reductase (HMGCR), the rate-limiting enzyme of de novo cholesterol synthesis⁵, (2) Ezetimibe blocks the intestinal cholesterol transporter Niemann Pick C1-Like 1 (NPC1L1) that mediates cholesterol absorption⁶, and (3) monoclonal antibodies directed against Proprotein convertase subtilisin/kexin type 9 (PCSK9) that promotes internalization and degradation of LDLR⁷. Despite their large clinical success, these treatments do not always restore plasma cholesterol levels to within a normal physiological range⁸ and can increase the risk of developing other diseases such as type 2 diabetes⁹. For example, LDLR mutation Classes II and V are the most responsive to statin treatment. However, Class II mutations typically result in more severe FH phenotypes than Class V mutations¹⁰, suggesting there may be additional factors contributing to the observed efficacy of such treatments in different patients with different LDLR mutations.

The amount of LDLR on the surface of cells (particularly on hepatocytes which express over 75% of the body’s LDL receptors) is the primary determinant of plasma LDL cholesterol (LDL-C). LDLR is a type I membrane protein, meaning that it possesses an extracellular/luminal N-terminus, followed by a single transmembrane helix and a cytosolic C-terminus, such that this integral membrane protein spans the entirety of the cell membrane^11–13. The synthesis and intracellular trafficking of LDLR have been well documented, but there is a growing body of evidence that additional post-translational mechanisms regulate and fine tune the surface availability of LDLR, thereby modulating its ability to bind and internalize LDL-C. Gaining a more thorough understanding of the mechanisms regulating LDLR will not only aid in dissecting the complex biological mechanisms that control LDL-cholesterol homeostasis, but may also open up new avenues for development of next-generation pharmacologic strategies to treat hypercholesterolemia. This review will focus on transcriptional, post-transcriptional, and post-translational pathways that regulate the LDLR and its availability at the plasma membrane (Fig. 1).

Figure 1. The LDLR is tightly regulated at the transcriptional, post-transcriptional, and post-translational levels.

High-level summary of the complex pathways that regulate the LDLR. Each mechanism of regulation is described in further detail in the cited subsequent figures. The LDLR is regulated transcriptionally by several transcription factors including SREBP-2 which stimulate the transcription of LDLR mRNA (Fig. 2). The stability of LDLR mRNA is regulated post-transcriptionally by miRNAs and RNA binding proteins influencing the rate of translation (Fig. 3). Sensing of cholesterol content in the ER controls a negative feedback loop, allowing SREBP-2 to exit the ER and enter the nucleus to promote transcription (Fig. 4). The LDLR can be post-translationally modified (i.e. glycosylated) to license its transport from the Golgi to the plasma membrane (Fig. 5). After internalization of LDLR-LDL complexes, the LDLR can be recycled back to the plasma membrane (Fig. 6). The LDLR can be cleaved, releasing a soluble form that can bind LDL extracellularly (Fig. 7). There are several pathways that can promote degradation of the LDLR at the cell surface or following internalization (Fig. 8).

Life cycle of LDLR

LDLR is synthesized at the endoplasmic reticulum (ER) and is processed to its mature form by glycosylation in the Golgi apparatus (discussed separately below). Approximately 45 minutes after synthesis, the LDLR is inserted into the cell membrane¹⁴ where it is concentrated into clathrin-coated pits to internalize bound LDL by receptor-mediated endocytosis (discussed separately below)². Once internalized into the sorting endosome, the acidic pH causes the LDLR to undergo a conformational change that releases the bound LDL particle from LDLR¹⁵. The released LDL is confined to the vesicular part of the sorting endosome that eventually matures into the late endosome, while the LDLR enters the tubules of the sorting endosome for recycling back to the cell membrane, with each cycle taking approximately 10-15 minutes¹⁶. With the average life-span of the LDLR being approximately 24 hr, an individual LDLR molecule makes roughly 100 of these cycles before it is degraded in the lysosome¹⁷.

Transcriptional Regulation

The human LDLR gene, located on chromosome 19, spans approximately 45kb consisting of 18 exons and 17 introns¹⁸. The LDLR gene promoter is located on the 5’-flanking region where most cis-acting DNA regulatory elements are found between −58 and −234 bp. The promoter region contains three imperfect direct 16 bp repeats, two TATA-like sequences, and multiple transcription initiation sites, which are all required for gene expression and regulation¹⁹ (Fig. 2). Repeats 1 and 3 contain sequences that are recognized by the general transcription factor specificity protein 1 (Sp1) and function to maintain basal LDLR gene transcription, regardless of the presence or absence of sterols^19,20. However, this regulation is not sufficient for high-level expression of LDLR in the absence of sterols. This requires repeat 2, which contains a 10 bp sterol regulatory element (SRE), as we discuss in detail below.

Figure 2. Transcriptional regulation of LDLR occurs within four key binding sites in the promoter region.

The LDLR promotor region contains two 16 bp repeat regions (R1 and R3), an SRE domain, and a SIRE domain. The transcription factor Sp1 binds the R1 and R3 domains, regardless of the presence of sterols. The main transcriptional regulator of LDLR is SREBP2, which binds in the SRE domain, and mediates LDLR transcription in response to cholesterol content. Other transcription factors can bind the SRE domain independently of SREBP2 (MAFF) and EGR1, ATF3, CREB, and cEBPβ can also individually bind the SIRE domain.

Transcription of LDLR mRNA is regulated by several transcription factors including co-activation by EGR1 and cEBPβ²¹, activating transcription factor 3 (ATF3)²², MAFF²³, and CREB²⁴. However, the main transcriptional regulator of LDLR is sterol response element-binding protein 2 (SREBP-2). In addition to LDLR, SREBP2 is also a transcriptional regulator of many other genes involved in the cholesterol synthesis pathway^25,26. When activated by physiologic signals such as changes to cholesterol content in the endoplasmic reticulum (ER) membrane (see section below on feedback inhibition), SREBP-2 binds to LDLR mRNA²⁵. SREBP-2 contains a NH₂-terminal domain with a basic-helix-loop-helix-leucine zipper (bHLH-Zip)²⁶ that interacts with the SRE in the LDLR gene promotor region²⁷, leading to increases in transcription and translation of LDLR protein. Nuclear active SREBP-2 is not long-lived and is polyubiquitinated and rapidly degraded by the proteasome with an estimated half-life of 3 hr²⁸. SREBP-2 also interacts with transcriptional co-activators, such as CBP and p300 that acetylate SREBP-2 to stabilize its expression²⁹, and Sp1 which binds on either side of the SREBP-2 binding site for synergistic LDLR transcriptional activation^30,31. Similarly, the SIRE domain can be bound by factors including ATF3 and cEBP to activate transcription^24,32.

Regulation of LDLR transcription is tightly controlled; therefore, gene mutations affecting these regulatory pathways can have a significant impact on protein expression. Thus far, ~3,000 variants have been identified in the human LDLR gene³³ that could have functional consequences on LDLR expression and ultimately, circulating LDL levels. The LDLR variant database (generated by Leigh et. al.), identified 37 variants occurring in the promotor region³³. A second study by Muller et. al. screened exons 5-10 of LDLR from 70 hypercholesteremic patients³⁴ and identified several variations, with one being an amino acid exchange in the bHLH-Zip domain identified in three of the pateints³⁴. However, whether this mutation alters the ability of SREBP-2 to bind the LDLR gene was not studied. Numerous mutations have been described in the promotor region which could impact the ability of SREBP-2 to bind LDLR and activate transcription of the gene leading to elevated cholesterol levels^35–38.

Post-transcriptional Regulation

Post-transcriptional regulation is another crucial mechanism of control of LDLR expression (Fig. 3). The LDLR mRNA has a relatively short half-life of approximately 45 min in cultured liver cells³⁹. Recent studies have identified mRNA splicing factors that regulate LDLR expression⁴⁰ that may contribute to LDLR mRNA half-life. The stability of a given mRNA is largely determined by the structure of its 3’ untranslated region (UTR)⁴¹. There are many different regulators that can bind to UTRs and affect mRNA stability, including microRNAs (Fig. 3A) and RNA binding proteins (RBPs; Fig. 3B), which we discuss in more detail below.

Figure 3. The stability of LDLR mRNA expression is mediated by microRNAs and RNA binding proteins.

(A) Numerous microRNAs (miRNAs) have been identified to interact with and influence the stability of LDLR mRNA. These miRNAs typically bind in the 3’UTR, ultimately leading to degradation of the mRNA. (B) RNA Binding Proteins (RBPs) also typically bind within the 3’UTR of their mRNA targets. RBPs can stabilize or degrade their gene targets. HuR is one RBP that binds AU-rich regions in the 3’UTR of LDLR. When bound it stabilizes the mRNA, increasing mRNA expression and leading to greater translation of LDLR protein.

MicroRNA regulation of LDLR

Hundreds of miRNAs have been identified since the discovery of the first miRNA in C. elegans in 1993^42–44. miRNAs are small non-coding RNAs, approximately 22 nucleotides in length, that typically interact with the 3’UTR of their target mRNAs, although interaction with the 5’UTR and coding regions of targets has also been reported⁴⁵. Base-pairing interactions between the miRNA and mRNA target, allow for targeting of the mRNA for degradation or repression of translation^42,46. Many different miRNAs have been described as regulating a variety of enzymes and transporters involved in cholesterol and lipid metabolism⁴⁷. More specifically, several miRNAs are known to directly interact with the 3’UTR of LDLR, such as miR-148a, miR-130b, miR-301b, miR-128-1, miR-185, and miR-27a/b^47–49.

A genome wide association study (GWAS) from people of European ancestry identified miR-148a as being associated with triglyceride, total cholesterol and LDL cholesterol levels⁵⁰. miR-148a is primarily expressed in the liver and expression is induced in mice and rhesus macaques fed a high fat diet⁵¹. There are two predicted miR-148a binding sites (ACGUGA) in the human 3’UTR of LDLR at positions 872-878 and 1971-1978, the first being highly conserved across mammalian species⁵². Another GWAS identified miR-130b and miR-301b, two related miRNAs derived from the same seed sequence, miR-128-1, as well as miR-148a, as being highly associated with metabolic control in cardiometabolic disorders⁵¹. miR-128-1 is located within intron 18 of R3HDM1 and its expression is dependent on the expression of its host gene⁵¹. MiR-148a, miR-130b, miR-301b, and miR-128-1 were shown to directly interact with the 3’UTR of LDLR to decrease expression and functionally, overexpression of these miRNAs decreased LDL-C uptake in HepG2 cells⁵¹.

Two separate studies identified miR-185 as an indirect⁵³ and direct⁵⁴ post-transcriptional regulator of LDLR mRNA. There are two conserved miR-185 binding sites (UCUCUCCA) located in the 3’UTR of LDLR, at positions 249-256 and 271-278⁵⁴. Overexpression of miR-185 in HepG2 cells resulted in decreased expression of LDLR mRNA and protein and decreased LDL uptake⁵⁴. Additionally, luciferase reporter assays involving mutations of the miR-185 binding sites in the LDLR 3’UTR stabilized the luciferase expression, suggesting miR-185 regulation of LDLR at these binding sites⁵⁴. Finally, in vivo overexpression of a lentivirus containing miR-185 antisense oligonucleotides led to increased levels of hepatic LDLR expression and decreased plasma cholesterol and atherosclerosis burden in Apoe^–/– mice⁵⁴.

The miR-27 family (miR-27a and miR-27b) are two highly similar miRNAs originating from the same seed sequence with a difference of a single nucleotide⁵⁵. miR-27a and miR-27b have been shown to decrease LDLR mRNA expression^55–57. Choi et. al. demonstrated that miR-27a regulates LDLR and decreases the infectivity of hepatitis C virus⁵⁶. However, their study did not determine whether this regulation was direct or indirect. LDLR mRNA contains one well conserved miR-27a binding site⁵⁵. Alvarez et. al, demonstrated that miR-27a interacts directly with the predicted LDLR binding site, and that mutations in this site prevented interaction and regulation by the miR-27a in HepG2 cells⁵⁵. Conversely, overexpression of miR-27a decreased LDLR expression by 50% in HepG2 cells⁵⁵. The 3’UTR of LDLR contains one predicted miR-27b binding site at position 2472-2478 (ACUGUGA)⁵⁷. Goedeke et. al. demonstrated that miR-27b regulates LDLR expression using luciferase reporter assays in Cos7 cells, where addition of miR-27b decreased luciferase activity, but mutation of the predicted binding sites reversed this effect⁵⁷. Additionally, C57BL/6 mice treated with an AAV expressing the precursor of miR-27b had lower hepatic LDLR protein expression⁵⁷. Further investigation of these miRNAs is required to fully understand their role in LDLR regulation and functional consequences for cardiometabolic diseases in murine models and humans.

RNA Binding Proteins

RNA binding proteins (RBPs) are post-transcriptional regulators of protein expression that bind RNA recognition motifs within the 5’UTR, coding region (intron or exon), or 3’UTR of target genes to facilitate a variety of functions including alternative splicing, RNA export from the nucleus, translation, and stabilization or degradation of RNA transcripts^58,59. A summary of multiple RNA interactome capture studies identified 1,914 and 1,393 RBPs in mice and humans, respectively⁵⁸, making them much more abundant than transcription factors. RBPs interact with target RNA through a variety of molecular mechanisms within their RNA binding domain⁵⁸. Several RBPs have been described as regulators of LDLR mRNA.

LDLR mRNA contains three adenine-uridine rich elements (AREs) within the 3’UTR^60,61. Berberine, a natural compound used as a cholesterol lowering drug, was found to have similar results as statins, but function through an alternative mechanism involving the 3’UTR and independent of changes in LDLR transcriptional activity⁶². However, there has been no description of specific RBPs regulated by berberine treatment. These AU rich regions of the 3’UTR are particularly important as they interact with RBPs to promote degradation or stability of the mRNA⁶¹. Li et. al. identified heterogeneous nuclear ribonucleoprotein D (hnRNP D), hnRNP I, and KH-type splicing regulatory protein (KSRP) as important negative regulators of LDLR mRNA stability⁶¹. These RBPs bind AREs in the 3’UTR, demonstrating that multiple RBPs can influence the stability of LDLR mRNA⁶¹. More recently, it has been shown that high cholesterol diet increases hepatic expression of hnRNP D in murine models which in turn leads to degradation of LDLR mRNA, whereas adenovirus mediated depletion of hnRNP D decreased circulating LDL levels⁶³. Another family of RBPs that bind AREs in the 3’UTR of target genes, including LDLR, is the ZFP36 family of proteins⁶⁴. While these RBPs destabilize their mRNA targets leading to degradation of LDLR, human antigen R (HuR), also known as ELAV-like RNA binding protein 1 (ELAVL1), binds the same AREs to stabilize LDLR mRNA and increase LDLR proteins in hepatocytes^65,66.

While the 3’UTR of the mRNA is not translated into protein, this region is highly important for gene regulation by miRNAs and RBPs, as discussed above. Genetic variants in this region can alter binding of miRNAs and RBPs, influencing the protein expression and function. Typically, loss of function mutations in the LDLR coding region have been described. Recently, Bjornsson et. al. identified an Icelandic family with a gain of function mutation in LDLR that correlated with dramatically lower levels of circulating LDL levels⁶⁷. In screening 43,202 Icelanders for mutations in LDLR that resulted in improved LDL levels, a rare 2.5 kilobase deletion at the end of the 3’UTR was identified across several generations of the same family⁶⁷. Carriers of this variant had LDL levels in the first percentile and did not contain other known mutations in cholesterol handling genes known to cause FH⁶⁷. This truncated 3’UTR lacks binding sites for both miR128-1 and miR-148^52,53,67, while the miR-185 binding site is still present^54,67. This 2.5 kb region also contains several AU rich regions that can impact LDLR mRNA stability⁶⁷. Observed changes in LDLR expression and circulating LDL were not attributed to loss of specific binding sites for a particular miRNA or RBP⁶⁷. Therefore, it is currently unclear which regulatory element or combination of regulatory elements is causal for the improved LDL levels.

Feedback Inhibition (Cholesterol/SCAP/INSIG/SREBP cleavage)

SREBP-2 mediated regulation of LDLR expression, is controlled by a complex and tightly regulated negative feedback loop (Fig. 4). SREBP-2 is synthesized as an inactive form and anchored to the ER membrane by SREBP-cleavage activating protein (SCAP), in a 4:4 oligomer ratio in yeast⁶⁸. SCAP acts as a cholesterol sensor, to regulate the release of inactive SREBP-2 from the ER membrane to the nucleus⁶⁹. When the cholesterol content of the ER membrane drops below 5mol%, SCAP is converted from its closed to open state allowing it to interact with COPII-coated vesicles and promote SREBP-2 activation (Fig. 4A). SCAP and SREBP-2 are then incorporated into these COPII vesicles that bud off of the ER and can be transported to the Golgi apparaturs⁷⁰. When SREBP-2 reaches the Golgi, two proteases, site 1 protease (S1P) and site 2 protease (S2P), sequentially modify SREBP-2 to convert it to its active form, designated nuclear SREBP-2 (nSREBP-2)^71–73. nSREBP-2 can then be translocated to the nucleus where it can interact with SRE domains to turn on transcription of target genes, including LDLR as described above²⁷. In contrast, when cholesterol levels are above 5mol% of ER membrane lipid composition, SCAP interacts with cholesterol and insulin-induced gene (INSIG), to maintain inactive SREBP-2 in the ER membrane (Fig. 4B)²⁵.

Figure 4. Cholesterol sensing-mediated translocation of SREBP-2.

SREBP-2 is synthesized in an inactive form and tethered to the ER by SCAP, a cholesterol sensing protein. (A) When cholesterol levels are low, SCAP undergoes a conformational change to its open state allowing it to interact with COPII vesicles, which transport the SCAP:SREBP-2 complex to the Golgi. SREBP-2 is further modified by two proteases, S1P and S2P, which convert SREBP-2 to its active, nuclear form (nSREBP-2). nSREBP-2 is then transported to the nucleus where it can activate transcription of LDLR. (B) When cholesterol content is high, cholesterol interacts with SCAP to keep it in its closed conformation, maintaining SCAP-INSIG interaction which restricts SREBP-2 to the ER and preventing its translocation to the nucleus.

Post-translational modifications

In previous sections we have reviewed mechanisms by which expression of the LDLR gene is regulated, yet at the protein level additional modifications and interactions are required for its stability and function (Fig. 5). The LDLR peptide has an N-terminal ligand binding domain which has high affinity for Apolipoprotein E^74,75, as well as Apolipoprotein B (ApoB)⁷⁶, the main structural components of VLDL/LDL. Recently, two separate reports in Nature detailed the crystal structures of ApoB⁷⁷ and the LDLR ligand binding domain bound to ApoB⁷⁸. Though it was previously known that the ligand binding domain of LDLR bound ApoB, these new studies detailing the structure of ApoB as well as high resolution analysis of the LDLR-ApoB interaction have significant potential for future FH-related therapies. In detailing the LDLR ligand binding domain bound to a β-belt in ApoB, the impact of specific Class II mutations associated with FH can now be modeled and are in fact highlighted in structural diagrams reported in one of the studies⁷⁸. Class II mutations are the result of changes to ligand binding affinity either via mutations in ApoB or LDLR. These groundbreaking new structural assessments may allow for targeted design of therapeutics for mutations impacting LDLR-ApoB interactions.

Figure 5. Post-translational modification of LDLR protein license transport of LDLR to the plasma membrane.

The LDLR protein is comprised of several domains: ligand binding domain, EGF-like domains and YWTD repeats, O-glycosylation region, transmembrane domain, and a cytoplasmic domain. The main post-translational modification of LDLR is glycosylation which is required for proper trafficking of LDLR in vesicles to the plasma membrane.

Like other LDLR-related family members⁷⁹, following the ligand binding domain, the remaining extracellular portion of the protein includes repeated epidermal-growth factor (EGF) like domains, and a series of tyrosine-tryptophan-threonine-asparagine repeats (YWTD repeats). These repeats are thought to both allow proper positioning of the ligand binding domain and have been shown to promote the release of cargo in the endosome in a pH-dependent manner⁸⁰. Between this repetitive structural region and the plasma membrane is a serine-threonine-rich region where heavy O-glycosylation occurs. LDLR spans the plasma membrane with a single simple transmembrane domain, lacking a GPI anchor¹³. The intracellular cytoplasmic domain of LDLR comprises only a small part of the protein and contains an asparagine-proline-(any AA)-tyrosine (NPxY) motif. The intracellular domain is involved in interaction with intracellular regulatory factors such as E3 ligases, which are discussed later, and receptor-mediated endocytosis, which is discussed in this section.

Glycosylation is the major post-translational modification made to the LDL Receptor. The LDLR protein is manufactured in the rough ER, where the N-terminal sequence of LDLR targets it into the ER membrane. It is then translocated to the Golgi network where it is glycosylated and released via vesicles to the plasma membrane, where is it recruited by Dab2⁸¹ to reside within clathrin-rich pits^82–84. While many mutations that cause FH are receptor-null Class I mutations that result in no LDLR protein, Class II mutations result in production of the LDLR protein, but it is not properly processed^85–87. These mutations can result from misfolding in the ER or from point mutations at key sites for glycosylation, such as has been recently observed in human N-glycosylation mutants⁸⁸. Both misfolding and glycosylation defects result in failure of the LDLR protein to reach the Golgi apparatus and therefore lack of protein glycosylation. To model these Class II mutations ENU mutagenesis generated a mouse model, the Wicked High Cholesterol (WHC) mouse⁸⁹, that developed hypercholesterolemia at levels found in LDLR knockout animals. This mutation occurs at residue 699 and results in a cytosine to tyrosine amino acid conversion just upstream of a heavily glycosylated region of the extracellular domain of LDLR protein near the beta propeller. This modification was found to trap LDLR in the ER, preventing glycosylation in the Golgi apparatus, and resulting in protein degradation rather than trafficking to the cell surface^89–91. In addition to the glycosylated region near the transmembrane domain, the ligand binding domain of LDLR has also been shown to be O-glycosylated^79,92. Modification by O-glycosylation enhances binding of apolipoprotein ligands to LDLR, likely largely mediated by the GalNAc-transferase GalNaC-T11^79,92. Defects in glycosylation in the ligand binding domain of LDLR may contribute to Class III mutations where the LDLR protein is properly folded and trafficked to the cell surface but fails to effectively bind its cargo. These glycosylation sites not only impact LDLR stability and function, but may also affect degradation rates of the LDLR, specifically by ASGR1 (discussed in more detail below), an E3 ligase which recognizes a glycosylated region of LDLR⁷⁵. Lastly, there is a new proposed group of polymorphisms, Class VI mutations, that include variations of the LDLR that do not properly thread into the plasma membrane³, and may also affect degradation rates of the LDLR, specifically by PCSK9, IDOL, RNF130, and ASGR1, that can all act on LDLR at the cell surface (discussed further below).

Recycling (receptor-mediated endocytosis)

The concept that LDL was brought into the cell in a receptor-mediated manner (Fig. 6) was confirmed in the 1970s when Brown and Goldstein demonstrated that fibroblasts from healthy individuals can bind and internalize radiolabeled LDL particles, while fibroblasts from patients with familial hypercholesterolemia cannot⁷⁴.

Figure 6. After binding lipoprotein particles at the plasma membrane, LDLR can be internalized and recycled.

(A) Following interaction between LDLR and LDL, ARH binds to the cytoplasmic tail of LDLR. This allows for binding of AP-2, triggering clathrin polymerization and ultimately endocytosis of the LDL-LDLR complex. (B) Once internalized, the LDL-LDLR complex is trafficked to the lysosome, where the low pH causes dissociation of the interaction between LDL and LDLR. (C) LDLR can then either be degraded or recycled back to the cell surface, aided by the WASH and CCC complex to traffic to the plasma membrane via actin filaments.

The ligand binding region of LDLR contains seven cysteine-rich repeats which fold such that a cluster of negatively charged amino acids are presented for ligand binding⁸⁵ allowing for the positively-charged region of Apolipoproteins to bind LDLR. Following lipoprotein binding, the LDLR adaptor protein autosomal recessive hypercholesterolemia protein^93,94 (ARH) binds to the intracellular NPxY motif in the cytoplasmic tail of LDLR (Fig. 6A). This interaction allows binding of the AP-2 protein⁸¹, which recruits additional clathrin leading to clathrin polymerization and invagination of the plasma membrane to initiate endocytosis. After internalization, lipoprotein-bound LDLR is trafficked to the lysosome, where the acidic environment of the lysosome reduces affinity of the lipoprotein for LDLR leading to dissociation of the ApoB- or ApoE-containing lipoprotein from LDLR (Fig. 6B). Once dissociated, LDLR is either degraded or recycled back to the cell surface where it can bind new cargo and repeat this process^95,96. The low pH in the lysosome results in a conformational change in the extracellular EGF domains (Fig. 5, left panel) into a closed hairpin structure, aiding in recycling back to the plasma membrane^95,96. Efficient endocytic recycling of LDLR requires the WASH (Wiskott-Aldrich syndrome protein and SCAR homolog) and CCC (COMMD/CCDC22/CCDC93) complex (Fig. 6C)⁹⁷. Loss of WASH is associated with decreased surface LDLR and increased plasma cholesterol⁹⁸. The COMMD family of proteins are essential for stabilizing complexes for endocytosis and LDLR recycling. Hepatic loss of proteins in the CCC has been shown to result in increased plasma lipid levels due to reductions in LDLR⁹⁹. Importantly, a coding variant in CCDC93 increases CCDC93 protein stability, which is associated with lower LDL-C levels and reduced CVD mortality¹⁰⁰, suggesting that in addition to current treatment strategies, targeting endosomal trafficking and/or recycling could offer therapeutic benefit. Indeed, LDLR can also clear triglyceride-rich lipoprotein particles and hypertriglyceridemia is an independent risk factor for CVD. Inhibition of Apo-CIII using anti-sense oligonucleotides reduced plasma triglycerides in a process dependent on LDLR and its related family member LRP-1¹⁰¹, and is a promising therapeutic for combined familial hyperlipidemias.

Cleavage

i. γ-secretase

γ-secretase is a protease complex that consists of four transmembrane protein units (nicastrin, presenilin-1/2, PEN2, and Aph-1) and is most commonly associated with cleavage of amyloid precursor protein (APP) to produce amyloid-β peptide¹⁰². The nicastrin subunit binds to LDLR and recruits the remainder of the subunits, resulting in cleavage of LDLR (Fig. 7A). γ-secretase typically cleaves substrates within the transmembrane domain, however the exact cleavage site within LDLR has not yet been identified. Normally, the γ-secretase enzyme complex is autoinhibited, and this autoinhibition is removed when substrates are cleaved extracellularly, exposing the transmembrane domain for cleavage by γ-secretase¹⁰³. However, γ-secretase has been shown to cleave the full-length LDLR, although extracellular cleavage may still occur prior to γ-secretase cleavage¹⁰⁴. Though this enzyme directly cleaves the LDLR, inhibition of γ-secretase primarily improves plasma triglycerides as opposed to plasma cholesterol¹⁰⁵.

Figure 7. The LDLR can be cleaved at multiple sites resulting in degradation, or release of the full or fragmented LDLR protein.

(A) γ-secretase cleaves the LDLR within the transmembrane domain, which can lead to lysosomal or proteasomal degradation of the LDLR. (B) Cleavage of the LDLR by the transmembrane protease MT1-MMP releases the entire intact ligand binding domain from the cell membrane, resulting in the release of the soluble form into the extracellular space, where this cleaved form of LDLR is still capable of binding lipoproteins. (C) BMP is a secreted protease that cleaves the LDLR in the ligand binding domain, releasing a small fragment of the LDLR and limiting its ability to interact with LDL.

ii. MT1-MMP

MT1-MMP is a transmembrane protease that has well-known functions in extracellular matrix remodeling¹⁰⁶. LDLR protein levels were shown to be increased with MT1-MMP liver-specific deletion, and MT1-MMP overexpression resulted in decreased LDLR expression, increased plasma cholesterol levels, and increased atherosclerotic lesions¹⁰⁷. This cleavage event results in the release of the entire intact LDLR ectodomain into the extracellular space as a soluble form of the receptor (sLDLR), thereby providing a two-fold regulation (Fig. 7B). First, cleavage reduces the availability of full length LDLR at the cell surface and second, since sLDLR retains the complete ligand binding domain and can still bind lipoproteins, it produces a form of the receptor that can sequester LDL particles in the extracellular space. Interestingly, MT1-MMP is also highly expressed in multiple cancer cell types promoting metastasis and invasion^106,108, thus inhibiting MT1-MMP holds both promise as a therapeutic target to reduce the risk of cancer metastasis and lower plasma LDL cholesterol.

iii. BMP1

BMP1 is a metalloproteinase that plays a central role in remodeling extracellular matrix proteins¹⁰⁹. However, unlike MT1-MMP, BMP is a secreted protease¹⁰⁹. In contrast to MT1-MMP, BMP cleaves LDLR within its ligand binding domain, releasing a small N-terminal fragment, and retaining a 120kDa membrane-bound C-terminal fragment (Fig. 7C)^110,111. The disruption of the ligand binding domain reduces affinity of the receptor for LDL, thereby reducing the ability to bind and internalize LDL¹¹⁰. Importantly, the BMP1 cleavage site in human LDLR is between Gly192 and Asp193, with the aspartate residue being critical for cleavage to occur¹¹⁰, although in mice, Asp193 is replaced with a valine residue and mouse LDLR is not cleaved by BMP1¹¹⁰. The cleaved N-terminal fragment of LDLR is detected in human urine, providing evidence that LDLR is cleaved by BMP1 in humans^110,111. Indeed, synthetic peptides and antibodies against the linker region containing the BMP1 cleavage site, prevented cleavage of LDLR by BMP1, providing important evidence that blocking BMP1 action on LDLR could be a possible novel avenue to target reducing plasma LDL-C levels¹¹².

Degradation

i. PCKS9

Proprotein convertase subtilsin-kexin type 9 (PCSK9) is predominantly produced in hepatocytes, with some in intestinal and kidney cells^113,114. PCSK9 is a dual function regulator of LDLR, acting both intracellularly as a chaperone that guides LDLR towards degradation, as well as a secreted factor that promotes LDLR internalization (Fig. 8A). PCSK9 has a signal peptide sequence, a prodomain, a catalytic domain, and a C terminal histidine-rich domain. PCSK9 is synthesized as a proprotein zymogen in the ER that undergoes intramolecular autocatalytic cleavage to release the mature protein for transport out of the ER and for secretion^115–117. Unlike other proprotein convertases, PCSK9 does not undergo a second cleavage to release the prodomain and is secreted as an enzymatically inactive protein¹¹⁸. Therefore, the regulation of LDLR by PCSK9 is independent of its enzymatic activity.

Figure 8. Degradation of LDLR is mediated by multiple proteins.

(A) PCSK9 binding to the EGF-like domain of LDLR preferentially targets it for degradation when LDL-bound LDLR is internalized. (B) Binding affinity of PCSK9 is amplified under lysosomal conditions, preventing receptor recycling and promoting degradation. (C) IDOL binds the cytoplasmic tail of LDLR and catalyzes LDLR ubiquitination and promotes proteasomal degradation. (D) Binding of RNF130 to the cytoplasmic domain both sequesters LDLR to prevent recycling to the cell surface and allows for ubiquitination by RNF130 for proteasomal degradation. (E) The glycosylated domain of LDLR is bound by ASGR1 to promote internalization and lysosomal degradation of LDLR.

At the cell surface, the catalytic domain of PCSK9 binds to the EGF-A repeat of the LDLR extracellular domain and the LDLR:PCSK9 complex is internalized through clathrin-coated pits^119,120. Whereas LDL particles are released from the LDLR in the acidic environment of the sorting endosome, the affinity of PCSK9 for LDLR increases 120–150-fold at low pH^121,122. This enhanced binding forces the LDLR to remain in an open conformation^123,124, preventing its recycling and promoting degradation of LDLR. LDLR proteins lacking the cytoplasmic domain¹²⁵ still undergo PCSK9-mediated degradation, suggesting that modifications of the cytoplasmic domain or interactions with ubiquitin or adaptor proteins are not involved in routing LDLR:PCSK9 complexes for intracellular degradation. It is also likely that PCSK9-mediated degradation of LDLR does not involve commonly used pathways for lysosomal or proteasomal degradation since PCSK9 is still able to promote LDLR degradation when ubiquitination, ESCRT machinery, autophagocytic pathways, or proteasomal activity are inhibited¹²⁰.

Although the ligand binding domain of LDLR is not required for PCSK9 binding, it appears to be important for PCSK9 internalization and plays a critical role in routing LDLR for intracellular degradation. A series of deletion studies demonstrate that an LDLR lacking the ligand binding domain is able to internalize PCSK9 but is not degraded and instead recycles back to the plasma membrane^126,127. This recycling indicates that internalized PCSK9 is released from LDLR lacking the ligand binding domain in sorting endosomes, suggesting that the ligand binding domain may be required for PCSK9 to remain bound to LDLR at the low pH of the sorting endosome. The C terminal domain of PCSK9 is also not required for PCSK9 to bind and be internalized by LDLR, it appears to be required for PCSK9 to disrupt recycling of LDLR back to the cell surface^126,128. The PCSK9 C terminal domain has been shown to bind the extracellular ligand binding domain of LDLR at low pH (pH < 5.4). This is likely due to the large number of histidines in the C terminal domain of PCSK9 which gain a net increased positive charge at low pH. This allows them to have increased affinity for the negatively charged ligand binding domain of LDLR^122,128,129. Loss of function PCSK9 is associated with reduced CVD due to LDLR stabilization^130,131. In fact, PCSK9 has emerged as a key target for the treatment and prevention of FH with successful preclinical studies in macaques¹³² and human clinical trials for Evolocumab¹³³.

ii. IDOL

Inducible Degrader Of the LDLR (IDOL), also known as MYLIP, is an E3 ubiquitin ligase that mediates the ubiquitination and degradation of the LDLR (Fig. 8B). IDOL is a unique protein being the only protein in the human genome to possess both a RING (Really Interesting New Gene) E3 ligase domain and FERM (4.1 band, ezrin, radixin and moesin) homology domain. The discovery of IDOL was made by the laboratory of Dr. Peter Tontonoz, who observed that in addition to regulating cholesterol efflux from peripheral tissues and cells, the Liver X Receptor (LXR) signaling pathway also exerts a profound effect on the cellular uptake of cholesterol through rapid elimination of LDLR protein from the cell surface^134,135. The ability of IDOL to mediate LDLR protein degradation was demonstrated through a series of well-executed in vitro and in vivo studies¹³⁴. IDOL is a cytosolic protein and the cytoplasmic tail of the LDLR is the only portion of the LDLR protein accessible to IDOL in the cell. Biochemical mutational analysis demonstrated that an intact K830 or C839 residue in the LDLR cytoplasmic tail is required for anchoring of ubiquitin and IDOL-mediated degradation¹³⁴. As an E3 ubiquitin ligase, IDOL also possesses the ability to ubiquitinate itself and therefore undergoes autodegradation. Interestingly, degradation of the LDLR and IDOL autodegradation are carried out by separate and distinct pathways, with LDLR degradation being dependent on the lysosome and IDOL autodegradation on the proteasome¹³⁴. In addition to targeting itself and the LDLR, it was later revealed by sequence homology and LXR activation experiments that IDOL is also able to ubiquitinate and promote degradation of the related lipoprotein receptor family members, Very Low-density Lipoprotein Receptor (VLDL) and ApoE receptor 2 (ApoER2)¹³⁶.

The structural basis and mechanism for substrate recognition by IDOL were elucidated through elegant studies that utilized scanning mutagenesis and in vitro reconstitution assays^137–139. These studies demonstrated that although both the FERM and RING domains are required for IDOL’s activity, they have functionally independent roles. The FERM domain was shown to be responsible for IDOL target recognition by binding to a recognition sequence (WxxKNxxSI/MxF) in the cytoplasmic tails of LDLR, VLDLR, and ApoER2 and does not require the E3 ligase activity or RING domain¹³⁷. Interestingly, IDOL-mediated degradation of LDLR requires the cellular membrane. In chimeric reconstitution experiments, it was shown that IDOL cannot degrade fusion proteins containing the soluble LDLR cytoplasmic tail, yet IDOL can promote degradation of heterologous membrane proteins fused to the LDLR cytoplasmic tail¹³⁶. This membrane dependency was shown to be mediated by the IDOL FERM domain, which mediates the association of IDOL with negatively charged phospholipids¹³⁷.

There have been no complete loss-of-function or dominant negative alleles of IDOL described to date. However, several genome wide studies have identified polymorphisms in the IDOL genomic region that are linked to total and LDL cholesterol levels^140–142, yet the physiologic and/or pharmacologic relevance of the IDOL pathway for disease pathogenesis or treatment remains to be elucidated.

iii. RNF130

RNF130, also known as GOLIATH, is another RING domain E3 ubiquitin ligase and is the most recently reported regulator of LDLR (Fig. 8C)¹⁴³. GOLIATH is the mammalian homolog of the Drosophila protease-associated (PA) domain-containing E3 ligase Goliath (dGoliath). Despite there being over 600 RING E3 ligases¹⁴⁴, most are cytosolic and only a few (that includes GOLIATH) possess a transmembrane domain (TMD)¹⁴⁵. GOLIATH exhibits this distinct domain architecture of a signal peptide, PA domain, TMD, and RING domain¹⁴³.

Similar to IDOL, we recently identified that GOLIATH polyubiquitinates the LDLR¹⁴³, however the functional consequences are quite different to those catalyzed by IDOL. Although GOLIATH ubiquitination promotes degradation of LDLR and decreases total LDLR levels, the more significant effect of GOLIATH E3 ubiquitin ligase activity, is the subcellular redistribution of the LDLR¹⁴³. GOLIATH primarily functions to redistribute LDLR away from the cell surface, thereby reducing the ability of the receptor to bind and internalize LDL. Targeting of GOLIATH in mice resulted in increased LDLR abundance and decreased circulating cholesterol levels¹⁴³, indicating that the GOLAITH-LDLR axis could be manipulated to enhance clearance of LDL particles.

iv. ASGR1

Asialoglycoprotein receptor 1 (ASGR1) promotes the lysosomal degradation of LDLR (Fig. 8D). ASGR1 is the major subunit of the transmembrane asialoglycoprotein receptor that is primarily expressed on the cell surface of hepatocytes. ASGR1 binds to glycoproteins and promotes their endocytosis and eventual lysosomal degradation. LDLR is targeted by ASGR1 in hepatocytes. Knockdown of ASGR1 increased total LDLR levels and LDL particle uptake, suggesting that ASGR1 functions to negatively regulate LDLR surface availability and LDL uptake^146,147. Indeed, loss of function mutations in ASGR1 are associated with decreased non-HDL cholesterol levels and decreased risk of CAD¹⁴⁸.

Conclusions, Future Perspectives, and Remaining Questions

Maintenance of normal cholesterol homeostasis is crucial for human health, with elevated plasma cholesterol levels implicated in a wide range of major diseases. As such, effective treatment and management of hypercholesterolemia has been the focus of targeted therapies and has shown the most efficacy in reducing the incidence of CVD, dementia, and stroke. Despite past and current success, there is a pressing need to provide increased control of plasma cholesterol levels. The identification of alternative therapeutic targets will allow for a more nuanced approach to providing enhanced cholesterol lowering and managing plasma cholesterol levels in patients that are refractory to current treatments.

As the major determinant of plasma cholesterol levels, the LDLR is the inflection point of cholesterol homeostasis. Since Brown and Goldstein discovered the LDLR as the causative protein in hypercholesterolemia it has been the subject of intense investigation. Therapies like statin drugs and PCSK9 monoclonal antibodies have helped lower plasma LDL cholesterol levels in hypercholesterolemic individuals and those with FH. The discovery of LDLR-independent pathways regulating plasma LDL cholesterol, such as ANGPTL3, ApoC-III. Sortilin, and GPR146, offer additional modalities to improve cholesterol lowering. As we have discussed in this review, the extensive myriad of regulatory pathways that dictate abundance, availability, and activity of the LDLR provide multiple potential nodes for pharmacologic intervention. Exploring how these novel regulators of LDLR may be exploited for therapeutic advantage, not just in hypercholesterolemia, remains a significant question. Targeting the LDLR to improve its efficiency or function at multiple levels of regulation, such as with the combination of statin therapy and PCSK9 inhibitors, may allow for enhanced clearance of LDL from the circulation.

Despite having ubiquitous expression, extra-hepatic functions of LDLR remain poorly understudied. How LDLR functions in non-hepatic tissues in normo- and hyper-cholesterolemic individuals and how this contributes to CVD warrants further investigation. Continuing to make strides in understanding LDLR function and the factors which regulate it will improve our ability to develop new treatments to prevent CVD. Of note, this review highlights new LDLR-ApoB structural data, new directionalities for LDLR stabilization by inhibition of PCSK9, new regulators of LDLR protein stability and availability RNF130 and ASGR1. We expect the recently published structural data to improve the design and seed the discovery of new drugs which can improve LDLR cargo binding, import, and receptor recycling. Additionally, RNF130 and ASGR1 represent potential new targets which would allow stabilization and increased availability of the LDLR. Together, these provide new avenues to design and discover new treatments to treat hypercholesterolemia and FH, combined hyperlipidemias, and lower the incidence of CVD.

Non-standard abbreviations and acronyms:

ARE

AU-rich element

ASCVD

Atherosclerotic cardiovascular disease

ASGR1

Asialoglycoprotein receptor 1

ATF3

Activating transcription factor 3

bHLHzip

Basic helix loop helix zipper

cEBP

CCAAT/enhancer binding protein

CHD

Coronary heart disease

CREB

cAMP response element binding protein

CVD

Cardiovascular disease

EGR1

Early growth response 1

ELAVL1

ELAV-like RNA binding protein 1

ER

Endoplasmic reticulum

FH

Familial hypercholesterolemia

GWAS

Genome-wide association study

HMCGR

3-hydroxy-3-methylglutaryl (HMG) CoA reductase

hnRNP

heterogeneous nuclear ribonucleoprotein

HuR

Human antigen R

IDOL

Inducible degrader of the LDLR

KSRP

KH-type splicing regulatory protein

LDL

Low-density lipoprotein

LDLR

LDL receptor

MAFF

V-Maf avian musculoaponeurotic fibrosarcoma oncogene homolog F

NPC1L1

Niemann Pick C1-Like 1

PCSK9

Proprotein convertase subtilisin/kexin type 9

RING

Really interesting new gene

RNF130

RING finger protein 130

SCAP

Sterol cleavage activating protein

SRE

Sterol regulatory element

SREBP-2

Sterol regulatory element binding protein

Sp1

Specificity protein 1

UTR

Untranslated region