Transporters as Drug Targets: Discovery and Development of NPC1L1 Inhibitors

Abstract

The potent cholesterol absorption inhibitor ezetimibe was developed as a first-in-class drug for treating hypercholesterolemia even before its molecular target, Niemann–Pick C1-like 1 (NPC1L1), had been identified. The NPC1L1 protein mediates sterol transport across the enterocyte brush border membrane and is essential for intestinal cholesterol absorption, a major pathway controlling whole-body cholesterol homeostasis. An elucidation of the mechanism underlying NPC1L1-dependent cholesterol absorption would greatly facilitate the discovery and development of new cholesterol-lowering agents for treating hypercholesterolemia and other cholesterol-related metabolic disorders.

DISCOVERY OF EZETIMIBE, THE FIRST NPC1L1 INHIBITOR

Hypercholesterolemia is a major risk factor for atherosclerotic coronary heart disease. Cholesterol homeostasis is a complex balance of de novo biosynthesis, intestinal absorption, and biliary excretion.¹ The current first-line therapy for hypercholesterolemia is statins, which reduce de novo cholesterol biosynthesis by inhibiting the rate-limiting enzyme, 3-hydroxyl-3-methylglutaryl-CoA reductase. Although statins significantly reduce blood cholesterol levels and mortality attributable to cardiovascular disease, statin monotherapy rarely lowers the level of blood low-density lipoprotein cholesterol to the ideal target of 70 mg/dl for high-risk populations. High-dose statins offer only a limited benefit and markedly increase the incidence of serious side effects (e.g., myotoxicity) and drug–drug interactions. Therefore, combination therapy and novel cholesterol-lowering drugs are needed. As one of the major pathways governing cholesterol homeostasis, intestinal cholesterol absorption represents an attractive drug target. With this knowledge, a potent cholesterol-absorption inhibitor, ezetimibe, was developed. Indeed, ezetimibe monotherapy efficiently lowers plasma total cholesterol and low-density lipoprotein cholesterol levels in humans.²

A member of the 2-azetidinone drug class, ezetimibe was identified in an acyl-CoA:cholesterol acyltransferase inhibitor discovery program that aimed to reduce atherosclerosis by preventing cholesterol esterification and deposition in the body.³ Although ezetimibe only weakly inhibited acyl-CoA:cholesterol acyltransferase, it was a potent inhibitor of intestinal cholesterol absorption in animals. After the discovery of statins,³ the discovery of ezetimibe, detailed in an excellent review,³ was a major breakthrough in the development of novel cholesterol-lowering drugs.

Ezetimibe’s potency in inhibiting cholesterol absorption is at least partially attributable to its unique metabolism in the body.⁴ After oral administration, ezetimibe exerts its initial action in the small intestine and is then absorbed and metabolized by the intestine and liver to its phenolic glucuronide. After its secretion into bile and delivery to the small intestine, the glucuronidated form of ezetimibe is more potent than the parent compound in inhibiting cholesterol absorption.⁵ This enterohepatic cycling results in a relatively long half-life for ezetimibe in human plasma (~22 h), and enhances its inhibitory activity by providing multiple opportunities for ezetimibe to encounter its target. Although it is very difficult to distinguish between the effects of ezetimibe and its glucuronidated metabolite in humans, evidence suggests that both forms may directly inhibit intestinal cholesterol absorption by interacting with the molecular target, Niemann–Pick C1-like 1 (NPC1L1). The drug–target interactions may occur on the cell surface and/or at an intracellular level.

REVERSE TRANSLATION: FROM DRUG (EZETIMIBE) TO TARGET (NPC1L1)

In the recent paradigm, drugs are developed for a well-documented molecular target. Such forward translational medicine proceeds from the basic discovery of “druggable targets” to the development of inhibitors or activators as therapeutic agents (“bench to bedside”). In this approach, molecular targets in a specific physiological pathway are identified before drugs are developed. For example, the ileal apical sodium-dependent bile acid transporter is essential for intestinal reabsorption of bile acids,⁶ the major catabolites of cholesterol. Inhibition of intestinal bile acid reabsorption lowers blood cholesterol by increasing the catabolism of cholesterol to bile acids as well as by reducing intestinal cholesterol absorption. With this molecular target and mechanism in mind, the apical sodium-dependent bile acid transporter inhibitor SC-435 was developed.⁷ In contrast, ezetimibe was discovered and developed without prior knowledge of its specific molecular target.³

The discovery of ezetimibe revolutionized our understanding of intestinal cholesterol absorption and demonstrated that cholesterol uptake at the enterocyte apical membrane is mediated by a specific protein rather than by passive diffusion (which was the historical hypothesis). Attempts to identify the molecular target of ezetimibe culminated in the discovery of NPC1L1 as a critical sterol transporter or receptor.⁸ Ezetimibe is therefore an example of reverse translational medicine, i.e., drug development followed by clarification of its molecular target. This highlights the importance of bidirectional flow of information in translational medicine.

STRUCTURE AND FUNCTION OF NPC1L1

Using a genomics–bioinformatics approach, Altmann et al. identified NPC1L1 as a target for ezetimibe.⁸ NPC1L1 is homologous to Niemann–Pick C1 (NPC1), a gene that is found to be defective in those with the genetic lipid-storage disorder Niemann–Pick disease type C1. Like its homologue NPC1, NPC1L1 is a polytopic membrane protein with 13 predicted transmembrane domains, 5 of which represent a typical sterol-sensing domain⁹ (Figure 1). NPC1L1 is postulated to mediate cholesterol trafficking because of its sterol-sensing domain.^8,9 Although a high-resolution crystal structure of NPC1L1 and its domains is unavailable, recent studies of the crystal structure of the cysteine-rich N-terminal domain of NPC1 identified a sterol-binding pocket.¹⁰ Given that the N terminus of NPC1L1 encompasses a similar cysteine-rich globular domain, it is likely that this region binds sterols as well.

Figure 1.

Amino acid sequences and predicted topological structure of human Niemann–Pick C1-like 1 (NPC1L1). Two potential YXXØ endocytic motifs are outlined in red squares. Residues in dark circles denote sterol-sensing domains. The N-terminal 21 amino acids are assumed to be the signal peptide and are not shown in this figure.

In rodents, NPC1L1 is almost exclusively expressed in the small intestines; in humans, however, it is expressed in multiple tissues.^8,9,11 Details relating to the transcriptional regulation of NPC1L1 expression remain unclear and controversial.¹² In the small intestine, the NPC1L1 protein resides at the apical surface of absorptive enterocytes.⁸ In mice, genetic ablation of NPC1L1 dramatically reduces intestinal cholesterol absorption to levels similar to those observed in ezetimibe-treated mice, and ezetimibe treatment in NPC1L1-null mice causes no further reduction in cholesterol absorption.^8,13 In cultured cells, overexpression of NPC1L1 facilitates cholesterol uptake, which can be inhibited by ezetimibe.^14–16 These findings clearly demonstrate that NPC1L1 mediates ezetimibe-sensitive sterol uptake.

INTERACTION BETWEEN NPC1L1 AND EZETIMIBE

Although the association between NPC1L1 and the ezetimibe-sensitive pathway for cholesterol absorption is well established, the issue of whether NPC1L1 is the direct molecular target has been a matter of debate for years. Using intestinal brush border membrane vesicles from wild-type and NPC1L1-deficient mice, Garcia-Calvo et al. performed in vitro drug-membrane vesicle binding studies.¹⁷ They showed that binding occurred between radiolabeled ezetimibe glucuronide and intestinal brush border membrane vesicles isolated from wild-type mice but not in those from NPC1L1-null mice. In addition, the binding affinity of the radiolabeled ezetimibe for intestinal brush border membranes from different species correlated well with its ability to inhibit intestinal cholesterol absorption. These data unequivocally reveal that ezetimibe glucuronide binds specifically to NPC1L1-containing membrane vesicles and support the concept that this interaction is important for ezetimibe’s mechanism of action. Recently, Weinglass et al. extended those findings by purifying the NPC1L1–ezetimibe complex and showing that NPC1L1 accounts for essentially all the ezetimibe binding to protein in those cells.¹⁸ This group further mapped the ezetimibe binding site to the extracellular loop-1 of NPC1L1 protein (Figure 1). Collectively, these functional and biochemical findings strongly support the concept that NPC1L1 is the molecular target of ezetimibe.

MOLECULAR BASIS FOR NPC1L1-DEPENDENT CHOLESTEROL UPTAKE

We have shown that, in cultured McArdle rat hepatoma cells, NPC1L1 resides mainly in the endocytic recycling compartment and the protein moves to and from the cell surface when cells are cholesterol depleted and enriched, respectively.¹⁴ We further demonstrated that the cholesterol-regulated NPC1L1 translocation to the cell surface facilitates ezetimibe-sensitive cholesterol uptake,¹⁴ which can be blocked by potassium depletion,¹⁹ a condition known to arrest clathrin-mediated endocytosis.²⁰ Caveolin-1, a structural protein of caveolae, is not required for intestinal cholesterol absorption,²¹ thereby suggesting that caveolin-mediated endocytosis, another important endocytic pathway, is unlikely to be the cellular basis for NPC1L1-dependent cholesterol uptake. On the basis of these observations, we hypothesized that clathrin-mediated endocytosis may underlie NPC1L1-dependent cholesterol uptake.^12,19 This hypothesis is consistent with the results from recent studies that used an identical McArdle cell model, showing that NPC1L1 co-immunoprecipitates with the μ₂ subunit of the adaptor protein AP2 and with the clathrin heavy chain,²² two critical proteins in the clathrin endocytic pathway. In this cell model, ezetimibe appears to inhibit the sterol-induced internalization of NPC1L1.²²

Many important questions remain to be answered before we can fully understand the molecular mechanisms governing NPC1L1-dependent and ezetimibe-sensitive cholesterol uptake. For example, does NPC1L1 directly bind cholesterol? If so, which region(s) of the protein are involved, and what role does cholesterol binding play in NPC1L1-mediated cholesterol uptake? Does an NPC1L1–cholesterol interaction serve as a signal for triggering the clathrin-mediated endocytosis of NPC1L1 or for recruiting free cholesterol to NPC1L1-containing membrane microdomains for subsequent internalization? Does NPC1L1 bind cholesterol that is intercalated within the plasma membrane, or does this interaction occur through cell surface interactions with cholesterol donors such as bile acid micelles in the gut lumen and hepatic bile canaliculus?²³ If NPC1L1 functions as a free-cholesterol receptor at the cell surface and is internalized through the clathrin-mediated endocytic pathway after cholesterol binding, what is the sorting signal in the protein? How does ezetimibe inhibit sterol-induced NPC1L1 internalization? The base of the microvilli of the small intestine is the only likely site for endocytic membrane trafficking in vivo, given the dense and rigid microvilli cytoskeleton. Under these conditions, how does the endocytosis of the microvillus-localized NPC1L1 and its cargo take place? Clearly, further work is needed to elucidate the molecular mechanisms underlying NPC1L1-dependent cholesterol uptake. Findings from this work will not only facilitate our understanding of how cells handle extracellular free cholesterol (and perhaps other lipophilic substrates) but also aid in the development of novel hypocholesterolemic agents.

FORWARD TRANSLATION: FROM TARGET (NPC1L1) TO DRUGS (NOVEL NPC1L1 INHIBITORS)

Ezetimibe is the only cholesterol absorption inhibitor available by prescription that is known to specifically target the NPC1L1 pathway. Large interindividual variations exist in responses to ezetimibe treatment.²⁴ The mechanisms responsible for the variability are not clear, but variability may be partly accounted for by the presence of inherited polymorphisms in NPC1L1. These observations highlight the need for additional inhibitors of NPC1L1 or cholesterol absorption. Indeed, a series of spiroimidazolidinone-based NPC1L1 inhibitors were recently identified by three-dimensional similarity–based virtual screening of the Merck corporate sample repository and in vitro NPC1L1 binding assay.²⁵

Although ezetimibe is clinically safe overall, drug toxicity (elevated levels of liver and muscle enzymes) did occur in some individuals. To reduce the possibility of systemic toxicity and drug–drug interactions, there is a potential to develop nonabsorbable NPC1L1 inhibitors that function exclusively in the small intestine. Several nonabsorbable cholesterol absorption inhibitors are in early clinical trials, including the compounds AZD4121 (AstraZeneca), MD-0727 (Microbia Pharmaceuticals), and AVE5530 (Sanofi-Aventis). Details regarding the clinical outcomes for these compounds are limited; however, the AVE5530 hypercholesterolemia clinical trial was halted because of insufficient efficacy. Despite this, therapy involving combinations of nonabsorbable NPC1L1 inhibitors with statins or other hypocholesterolemic agents is expected to be a viable approach.

COMBINATION OF NPC1L1 INHIBITORS WITH OTHER DRUGS

The liver attempts to compensate for decreases in intestinal cholesterol absorption by upregulating endogenous cholesterol synthesis; consequently, coadministration of ezetimibe with a statin has an additive effect in lowering plasma cholesterol levels.² A completed clinical trial (Extended Evaluation of Recombinant Human Activated Protein C; ENHANCE) examined the effects of ezetimibe plus simvastatin vs. simvastatin alone on atherosclerosis in the carotid artery in familial hypercholesterolemic patients, using the surrogate marker, carotid intima-media thickness, as the primary end point.²⁶ After 2 years of treatment, patients on combination therapy had the same carotid intima-media thickness as patients taking simvastatin alone. Unfortunately, the observed outcomes of the trial did not match expected outcomes, probably because of study design weaknesses related to patient selection and methodology. The effects of combined therapy on clinical end points such as myocardial infarction and stroke will be evident after the completion of the Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT) involving 18,000 hypercholesterolemic patients.²

An attractive and aggressive approach to reducing blood cholesterol may be the combination of three drugs targeting the three major pathways governing whole-body cholesterol homeostasis: a statin for inhibiting cholesterol biosynthesis, an intestinal cholesterol absorption inhibitor, and a small-molecule activator of the heterodimeric adenosine triphosphate–binding cassette transporters G5 and G8 (ABCG5/G8) that promotes cholesterol excretion. Mice that overexpress human ABCG5/G8 are hypersensitive to the hypocholesterolemic effect of a statin,²⁷ which suggests that increasing the ABCG5/G8 function represents a promising pharmacologic goal.

Agonists to the nuclear receptor liver X receptor (LXR) activate an array of cholesterol efflux genes, including ABCG5/G8,²⁸ thereby facilitating reverse cholesterol transport from peripheral tissues, including macrophages, to the feces. However, LXR agonists also drive hepatic lipogenesis, which substantially hampers their clinical use. We found that treatment of NPC1L1-deficient mice with an LXR agonist increases hepatic ABCG5/G8 expression without causing a dramatic elevation in liver triglycerides.²⁹ Therefore, coadministration of ezetimibe or other cholesterol absorption inhibitors with LXR agonists may reduce side effects and preserve the beneficial outcomes of LXR agonists.

POTENTIAL ROLES OF NPC1L1 INHIBITORS IN METABOLIC AND GALLSTONE DISEASES

There is exciting evidence that ezetimibe-treated wild-type mice and NPC1L1-deficient mice are resistant to diet-induced non-alcoholic fatty liver disease, weight gain, insulin resistance, and gallstone formation.^30–32 These studies suggest that an NPC1L1-mediated and/or ezetimibe-sensitive pathway has a significant physiological role beyond cholesterol uptake. It is clear that drug discovery and development pertaining to NPC1L1 and cholesterol absorption will be an exciting field with far-reaching consequences.

SUMMARY AND CONCLUSIONS

The development of the potent hypocholesterolemic agent ezetimibe has played a central role in the discovery of the NPC1L1 pathway for cellular uptake of cholesterol and is a powerful example of the bidirectional nature of translational medicine. Although inhibition of NPC1L1 by ezetimibe dramatically reduces blood cholesterol, the long-term outcomes of NPC1L1 blockade on clinical end points are still unknown. Given the critical role of NPC1L1 in sterol metabolism, the future for NPC1L1 inhibitors is bright, especially with respect to the combined use of NPC1L1 inhibitors and other cholesterol-lowering agents.