Apolipoprotein F - A natural inhibitor of CETP and key regulator of lipoprotein metabolism

Abstract

Purpose of review

To highlight recent studies that have advanced our understanding of apolipoprotein F (ApoF) and its role in lipid metabolism.

Recent findings

Previous studies showed that ApoF hepatic mRNA levels are suppressed by fat-enriched diets. Recent studies show this down-regulation is mediated by agonist-induced binding of LXR and PPARalpha to a regulatory element in the ApoF promoter. First-of-kind in vivo studies show ApoF lowers low density lipoprotein levels and enhances reverse cholesterol transport in fat-fed hamsters.

Summary

Diverse studies collectively provide compelling evidence that cholesteryl ester transfer protein (CETP) plays an important role in regulating lipid metabolism. Inhibiting CETP raises high density lipoprotein cholesterol. However, considering the recent failures of pharmacological inhibitors of CETP in clinical trials, it does not seem likely that global inhibition of CETP will be beneficial. ApoF is a minor apolipoprotein that functions as a natural inhibitor of CETP. However, ApoF is not a general inhibitor of CETP, but rather it preferentially inhibits CETP activity with LDL. Therefore, ApoF tailors CETP activity so that less tissue-derived cholesterol traffics from HDL into the LDL compartment. Lower LDL cholesterol levels have recognized clinical benefit for reduced cardiovascular disease.

Introduction

Elevated low density lipoprotein (LDL) and reduced high density lipoprotein (HDL) cholesterol are risk factors for cardiovascular disease. A major beneficial function of HDL is to serve as a platform for the transport of peripheral tissue cholesterol to the liver for excretion. Cholesteryl ester transfer protein (CETP) impacts this process by moving HDL-associated cholesterol (as cholesteryl ester (CE)) to other lipoproteins, creating alternative pathways for peripheral tissue cholesterol to be delivered to the liver or diverted to other tissues. As a result, CETP activity directly alters LDL and HDL cholesterol levels. Our lab is interested in understanding the mechanism of CETP, its impact on lipid and lipoprotein metabolism, and the factors that regulate its activity. This review focuses on apolipoprotein F (ApoF), a plasma protein that regulates CETP activity in a lipoprotein-specific fashion. Past studies and recent advances are reviewed, and several areas in need of further investigation are identified.

Cholesteryl ester transfer protein (CETP)

A review of ApoF literature necessarily requires an introduction to CETP since much of what is understood about the function of ApoF is in the context of this transfer protein. CETP is a plasma protein that facilitates the movement of cholesteryl ester (CE) and triglyceride (TG) between lipoproteins. Through its capacity to exchange CE in one lipoprotein for TG in another (1), CETP modifies lipoprotein composition and controls lipoprotein metabolism (2–4). Its role in modulating lipoprotein metabolism and altering HDL levels in humans has been extensively studied (5, 6). CETP activity is also critical to whole body cholesterol clearance mechanisms in humans since 70% of tissue-derived cholesterol (as CE) in HDL is transferred to VLDL and LDL by CETP prior to its removal by the liver for eventual excretion (7, 8). Therefore, factors controlling CETP activity have the potential to modify whole body cholesterol metabolism.

An inhibitor of CETP

From the earliest efforts to purify CETP from human plasma it became apparent that there is a factor, or factors, in plasma that reduce CETP activity. We identified a novel, acidic protein called lipid transfer inhibitor protein (LTIP) (9) that equally inhibits the transfer of both CE and TG by CETP. Importantly, subsequent in vitro studies with isolated lipoproteins, or human plasma supplemented with partially purified LTIP, show that while LTIP reduces CETP transfer activity between all plasma lipoproteins, it preferentially suppresses lipid transfers involving LDL while having the least effect on CETP-mediated transfers between VLDL and HDL (10, 11). While much remains to be determined about the mechanism of LTIP’s action, the data suggest LTIP blocks CETP activity by disrupting and/or preventing the binding of CETP to the lipoprotein surface (12, 13). The binding of CETP to lipoproteins is an essential step in the lipid transfer process (12, 14, 15). CETP interacts with phospholipids residing in the lipoprotein surface (12, 16, 17). The molecular organization of these lipids has a significant impact on CETP binding (14). We hypothesize that LTIP binds to the surface of lipoproteins and modifies this molecular organization.

The preferential inhibition of CETP lipid transfer events involving LDL by LTIP explains an apparent contradiction between lipid transfer studies with isolated lipoproteins and those measured in human plasma. In plasma, CETP-mediated CE transfer from HDL exceeds that from LDL by up to 2-fold relative to their individual contents of CE (18). But with isolated lipoproteins, there is no apparent preference of CETP for LDL or HDL as a substrate. However, the preference for HDL as a transfer protein substrate can be restored by the addition of LTIP to the isolated lipoprotein assay (19). This observation explains previous findings in plasma from continuous ambulatory peritoneal dialysis patients. In these plasmas, CETP shows little preference for LDL versus HDL as a lipid transfer substrate, and these plasmas also have very low LTIP activity (20).

Further purification and eventual molecular cloning of LTIP revealed its identity Apolipoprotein F (21).

What is Apolipoprotein F?

ApoF was first identified as a minor apolipoprotein in human plasma in 1978 (22). ApoF primarily associates with lipoproteins involved in cholesterol transportation and esterification (23). The gene coding ApoF resides on human chromosome 12 (24). APOF mRNA encodes a 308 amino acid protein containing a signal peptide and a 286 amino acid proprotein (Figure 1). ApoF protein is the 162 amino acid, C-terminal fragment of proApoF. PCSK7, a type-I membrane-bound protease highly enriched in liver, appears to mediate this cleavage following secretion of proApoF (25). The apparent molecular weight of mature ApoF is ~30kDa, not the expected 17.4kDa, due to its content of N-linked and O-linked sugars (26–28).

Figure 1 – Schematic of the primary structure of preproApoF.

ApoF is derived from a much larger precursor protein that is cleaved upon secretion by a membrane bound protease. The amino acid sequence of this cleave site is shown. Abbreviations: MW, molecular weight; pI, isoelectric point; pcsk7, proprotein convertase subtilisin/kexin type 7.

ProApoF has a unique structure. At physiologic pH, the N-terminal peptide is strongly positively charged, whereas the C-terminal portion (ApoF) is negatively charged (Figure 1). In our experience, the N-terminal portion of proApoF is essential for secretion of ApoF, suggesting that these two oppositely charged sequences might work together to prevent ectopic binding of proApoF to intracellular membranes, or to promote proper folding of ApoF. It cannot be excluded that the N-terminal portion of proApoF also has functions extracellularly.

Regulation of APOF gene expression

APOF gene expression is primarily restricted to the liver (24,28,29). Several recent studies have begun to provide insight into transcriptional factors involved in the regulation of APOF mRNA expression. Agonists of the farnesoid X receptor (FXR) increased APOF mRNA levels several-fold in liver cell lines and in C57/BL6 mouse liver. This regulation depends on binding of FXR to the FXR element ER1 (−2904 to −2892 nt) in the APOF promoter (30). Also, multiple binding sites for ETS-1/2 and C/EBP transcription factors were identified in the −198 to −2 nt region of the APOF promoter. ETS-1 and C/EBPalpha physically associate and act synergistically to activate APOF transcription (31). In contrast, we recently reported that liver X receptor (LXR) and peroxisome proliferator-activated receptor alpha (PPARa) negatively regulate APOF expression through a complex regulatory element about 1900 base pairs 5’ to the APOF proximal promoter (32). These reductions in APOF mRNA caused parallel declines in ApoF protein secretion. This negative regulation entails agonist-mediated recruitment of these nuclear receptors to the promoter element. Regulation through this element likely explains the marked reduction in hepatic APOF mRNA levels that occurs in both hamsters and rabbits consuming diets enriched in cholesterol or fatty acids [29]. Lipid metabolites derived from these diets are known agonists for LXR and PPARa receptors. Moreover, a similar mechanism may also exist in mice. Two studies reported that APOF expression is significantly downregulated in LDLR-deficient mice when fed high fat compared with chow-fed mice (33,34). The negative regulation of APOF gene expression by metabolites of both fatty acid and sterol pathways shows that APOF gene expression is closely integrated with hepatic lipid metabolism.

Plasma ApoF levels in hyperlipidemia

The concentration of ApoF in normolipidemic plasma is approximately 84 μg/ml or 5 μM (26), although this value differs between quantitation methods. In normolipidemic plasma, ApoF levels are 30% higher in males than in females. ApoF is positively associated with HDL cholesterol in normolipidemic males but not in females (26). ApoF levels are significantly elevated in hypercholesterolemic but reduced in hypertriglyceridemic subjects (35). However, the response of ApoF to plasma TG levels is gender-specific. In male hypertriglyceridemic subjects, ApoF levels are ~50% of normolipidemic plasma, whereas in females ApoF levels trend upward. Plasma ApoF levels are also increased in hamster and rabbit models of hypercholesterolemia, where they correlate positively with plasma cholesterol and LDL levels (29). In a small proteomics study of men with recently diagnosed coronary artery disease, ApoAI levels in isolated HDL₃ were unchanged but ApoF levels were reduced 40% compared to healthy controls. One year of niacin + statin therapy normalized these ApoF levels (36).

The increased plasma ApoF levels observed in hypercholesterolemia are inconsistent with the down regulation of APOF gene expression and reduced ApoF secretion that occurs in this setting (see above). As described below, ApoF can reside on different plasma lipoproteins and this distribution is influenced in hypercholesterolemia. We hypothesize that the plasma residence time of ApoF is variable dependent on the lipoproteins with which it associates. Increased residence time in hypercholesterolemia could explain the apparent disconnect between ApoF synthesis and plasma levels. This hypothesis remains to be tested.

Regulation of ApoF activity

An indication that ApoF activity might be regulated came from the observation that ApoF activity, as measured by inhibition of CETP, is primarily associated with plasma LDL (10) but most ApoF protein is present in the HDL fraction (37). Further analysis of the distribution of ApoF protein in normolipidemic plasma resolved this contradiction. ApoF exists in two states, one associated with LDL, and the other bound in a large, 470 kDa complex (38). Although ApoF was not detected on VLDL in that study, a recent lipoproteomic analysis showed that ApoF in VLDL is 5-fold lower that than associated with LDL (39). The 470kDa complex is an HDL particle with a size slightly larger than that of HDL₂ but with a density in the HDL₃ range. Most importantly, the ApoF associated with LDL is active in regulating CETP, whereas ApoF in the 470kDa complex is inactive (38). In normolipidemic plasma, more than 85% of ApoF is contained in the inactive complex (40).

The distribution of ApoF between active and inactive pools is under metabolic control. In vitro incubation of plasma at 37°C caused ApoF to redistribute from the inactive pool to the active pool (40). This redistribution could be driven by the activities of CETP and/or lecithin:cholesterol acyltransferase contained in plasma. These in vitro incubations increased LDL-associated ApoF from 15% to ~50% of total plasma ApoF. The acquisition of ApoF by LDL depended on the extent of LDL modification and correlated positively with changes in its ratio of CE + TG -to- phospholipid + cholesterol. The dependence of ApoF binding to LDL on LDL’s lipid composition suggests that the distribution of ApoF between active and inactive pools may be different in hyperlipidemia since altered LDL composition is commonly encountered in this setting. This was confirmed in a small group of hypercholesterolemic individuals, where ~50% of plasma ApoF was associated with LDL (38). This suggests that hyperlipidemia will alter ApoF activity independent of ApoF concentration.

Impact of ApoF on lipoprotein metabolism

To further define the role of ApoF, we recently investigated the consequence of ApoF knockdown on lipid metabolism in hamsters (41). Hamsters naturally express CETP and ApoF, unlike mice that lack the CETP gene. siRNA-mediated ApoF knockdown reduced plasma ApoF levels by more than 85%. ApoF knockdown in chow-fed hamsters had no effect on lipoproteins over the 6-day course of the study, but the ApoF-deficient lipoproteins from these animals did support 50–100% higher CETP activity with LDL in vitro. In fat-fed animals, ApoF knockdown created an atherogenic profile. Endogenous CETP-mediated transfer of HDL CE to LDL increased up to 2-fold, plasma LDL cholesterol increased, HDL cholesterol declined, and both LDL and HDL lipid compositions were altered. Notably, in reverse cholesterol transport assays, ApoF knockdown impaired the excretion of HDL-associated CE into feces in fat-fed hamsters. Conversely, ApoF knockdown in chow-fed animals increased reverse cholesterol transport. These findings are illustrated in Figure 2. These in vivo data validate in vitro findings that ApoF regulates lipid transfer to LDL and demonstrate that the impact of ApoF on lipoproteins and sterol excretion depends on the underlying lipid status. We conclude from these in vivo studies that ApoF minimizes the transfer of HDL-derived CE to LDL, which helps control LDL cholesterol levels when LDL clearance mechanisms are limiting.

Figure 2 – Effect of ApoF on the fate of HDL CE.

CE in HDL can be transferred to VLDL and LDL by CETP. LDL and VLDL are cleared by separate hepatic receptor mechanisms. ApoF preferentially blocks CETP activity on LDL, thus reducing the flow of HDL derived CE along this pathway. For both panels, arrow size reflects relative CETP activity between the indicated lipoprotein pairs. Panel A - In chow-fed animals, depletion of ApoF increases CE transfer to LDL. Since hepatic LDL receptor activity is high, the increased CE in LDL is efficiently delivered to the liver, thus increasing the overall delivery of HDL CE to the liver and preventing LDL cholesterol from rising. Panel B - In fat-fed animals, LDL-associated ApoF is increased by the diet, which further reduces the transfer of HDL CE to LDL compared to chow-fed animals. When ApoF is depleted, overall CE efflux from HDL is increased, but due to its initial greater repression by ApoF, CE transfer to LDL is markedly increased. However, since hepatic LDL receptor levels are low in fat-fed animals, HDL-derived CE accumulates in LDL. This results in an overall decrease in HDL CE delivery to the liver. Direct HDL CE clearance by SR-BI is not shown in these models. Abbreviations: LDLR, low density lipoprotein receptor; HSPG, heparan sulfate proteoglycans.

ApoF is present in the plasma of animals that express CETP and those that do not. This suggests that ApoF may have functions independent of CETP. For example, since most plasma ApoF is bound to HDL, it may have a role in HDL stability and/or turnover. Insights into this can be gained by mouse studies, which express ApoF but lack CETP. Overexpression of either murine or human ApoF in mice by adeno-associated viruses decreased total and HDL cholesterol levels, and accelerated the clearance of HDL from plasma (27). However, an effect on HDL was not evident in ApoF deficient mice (28). While other functions of ApoF seem likely given its presence in species not expressing CETP, these remain to be determined.

Insight into possible non-lipoprotein-related functions of ApoF

A proteomics study identified ApoF as a novel biomarker for hepatic fibrosis in hepatitis C patients. Plasma ApoF levels correlated negatively with the extent of fibrosis (42). Patients with severe periodontitis also have reduced HDL ApoF levels compared to age-matched controls without disease (43). And in both columnaris infected channel catfish and Vibrio harveyi infected sea bass, hepatic APOF mRNA expression is dramatically induced (44, 45).

In addition to an inflammation link, APOF expression is down-regulated in hepatocellular carcinoma and associated with low recurrence-free survival rate. APOF overexpression inhibited proliferation and migration in hepatoma cell lines, and also effectively repressed tumor growth in a xenograft nude mouse model (46).

Conclusion

Current data support the hypothesis that ApoF provides a mechanism for fine-tuning where CE goes once removed from HDL by CETP. ApoF preferentially inhibits CETP-mediated lipid transfer specifically involving LDL. This likely occurs because ApoF is most active in inhibiting CETP when it is bound to LDL. As a result, HDL-derived CE traffics to VLDL, whose remnants are rapidly cleared by the liver, which prevents the accumulation of cholesterol in LDL when hepatic LDL receptors are down regulated. In this way, ApoF may reduce the risk for atherosclerosis. Demonstrating this anti-atherogenic role for ApoF will require long-term diet studies in an animal model where CETP and ApoF are naturally expressed.

Key Points:

• ApoF selectively inhibits CETP lipid transfer events involving LDL.

• APOF gene expression is regulated in liver by a series of lipid-regulated transcription factors including FXR, LXR and PPARalpha.

• LDL composition regulates the distribution of ApoF between its active form on LDL and its inactive form in HDL.

• Plasma ApoF levels are increased in hypercholesterolemia and a larger percentage is associated with LDL.

• In fat-fed hamsters, ApoF expression is associated with lower LDL cholesterol levels and higher reverse cholesterol transport, suggesting it may have anti-atherogenic properties.