Abstract

The incidence of CHD is still increasing, which underscores the need for new preventive and therapeutic approaches to decrease CHD risk. In this respect, increasing apoA-I concentrations may be a promising approach, especially through increasing apoA-I synthesis. This review first provides insight into current knowledge on apoA-I production, clearance, and degradation, followed by a systematic review of dietary and novel pharmacological approaches to target apoA-I metabolism. For this, a systematic search was performed to identify randomized controlled intervention studies that examined effects of whole foods and (non)nutrients on apoA-I metabolism. In addition, novel pharmacological approaches were searched for, which were specifically developed to target apoA-I metabolism. We conclude that both dietary components and pharmacological approaches can be used to increase apoA-I concentrations or functionality. For the dietary components in particular, more knowledge about the underlying mechanisms is necessary, as increasing apoA-I per se does not necessarily translate into a reduced CHD risk.

1. Introduction

The global incidence of coronary heart diseases (CHD) is still increasing, which underscores the need for novel and alternative approaches to prevent the initiation and progression of this disease already at an early stage. Since elevated serum low-density lipoprotein cholesterol (LDL-C) concentrations are causally related to CHD, most dietary life style interventions and pharmaceutical treatments to prevent CHD so far are focused on lowering serum LDL-C concentrations. Despite successful intervention possibilities, there is still a substantial residual cardiovascular risk. Therefore, a possibility of further lowering CHD risk is to target multiple metabolic pathways simultaneously [1, 2]. For example, statin treatment, to lower serum LDL-C concentrations, can be combined with other pharmaceutical agents, such as proprotein convertase subtilisin/kexin type 9 inhibitors, which substantially further lower serum LDL-C concentrations [3]. Also, the Niemann-Pick Like Intracellular Cholesterol Transporter 1 inhibitor ezetimibe can be used, which has been shown to further lower the number of myocardial infarctions with 13%, strokes with 14%, and ischemic strokes with 21% [4]. Besides combined interventions to further increase the LDL-C lowering potential, it can be considered to target at the same time other CHD risk parameters including serum high density lipoprotein (HDL) cholesterol (HDL-C), apolipoprotein A-I (apoA-I), triacylglycerol or lipoprotein(a) concentrations, and/or blood pressure [5]. These parameters may be interrelated. An inverse relationship exists, for example, between serum triacylglycerol and HDL-C concentrations. Thus, interventions that change triacylglycerol may therefore also affect HDL metabolism [6]. In this review we will however focus on possibilities to further reduce CHD risk via novel and alternative dietary and pharmacological interventions targeting apoA-I metabolism.

1.1. Increasing HDL Functionality by Increasing ApoA-I

So far, interventions specifically targeting to increase serum HDL-C concentrations did not report any protective cardiovascular effect, which has clearly negatively influenced the interest to develop novel interventions to elevate serum HDL-C. However, recent evidence suggests that the focus should be on optimizing HDL functionality instead of elevating circulating serum HDL-C concentrations [7]. By increasing their functionality, HDL particles are able to take up more cholesterol from peripheral tissues, that is, the so-called cholesterol efflux. In addition, a more functional HDL particle will be more antioxidative—in particular by inhibiting LDL oxidation—and more antithrombotic and will have a higher anti-inflammatory and antiapoptotic activity [8]. A wealth of evidence from epidemiological, in vitro, and in vivo studies suggests that higher apoA-I concentrations protect against CHD development [9]. By increasing apoA-I concentrations, the resulting newly produced small HDL particles (i.e., prebeta HDL) will be highly functional, thereby enhancing cholesterol efflux [8]. Indeed, it has been found that apoA-I concentration is the strongest predictor for cholesterol efflux capacity [10]. ApoA-I is the major protein of HDL particles [11] contributing to approximately 33% of the total HDL particle mass and up to 60% of the HDL protein mass [12]. The most likely mechanism explaining the beneficial effects of elevated serum apoA-I concentrations origins from the fact that apoA-I is the ligand for ATP-binding cassette transporter A1 (ABCA1), as such mediating cholesterol efflux from lipid-laden macrophages [8]. Based on this information, Smits et al. wrote a clear plead for strategies to increase serum apoA-I concentrations as the most promising target for enhancing HDL functionality, thereby decreasing cardiovascular disease (CVD) risk [13]. However, lowering CHD risk by increasing endogenous apoA-I production, by decreasing apoA-I degradation, or by providing exogenous apoA-I has for unknown reasons not yet been investigated into great detail. Therefore, the question remains whether specifically targeting apoA-I metabolism is a suitable target to reduce CHD risk.

In this review we will first briefly provide insight into the current knowledge of apoA-I synthesis, clearance, and degradation, followed by a detailed overview of dietary and novel experimental pharmaceutical developments targeting circulating apoA-I concentrations.

2. ApoA-I

2.1. ApoA-I Synthesis

ApoA-I mRNA is expressed in cells of the liver and small intestine [14], where it is translated into a pre-pro-apoA-I protein. The presegment needs cotranslational cleavage [15], which takes place during translocation of the protein into the endoplasmatic reticulum by a signal peptidase [16, 17]. This results in a stable intracellular, pro-apoA-I protein [15], which is secreted into blood and lymph. Directly after secretion of pro-apoA-I, the proprotein is cleaved of by Bone Morphogenetic Protein-1 (BMP-1) and Procollagen C-proteinase Enhancer-2 Protein (PCPE2) (Figure 1) [18, 19]. It is evident that the cleavage of the prosegment is essential for the secretion of newly formed intracellular apoA-I. Deletion of the coding sequence of the prosegment causes accumulation of apoA-I in the cell [20], decreases the efficiency of apoA-I mRNA expression [17], and impairs the secretion of apoA-I into blood and lymph [17, 20]. The cleavage of the proprotein occurs relatively rapid, while the residence time for pro-apoA-I in plasma is only 5.5 hours, in contrast to the residence time for mature apoA-I of 6.5 days [21]. About 4–8% of the circulating apoA-I pool is pro-apoA-I [15, 22, 23]. After cleavage of the prosegment, apoA-I accepts cholesterol and phospholipids from ABCA1 [24] to form a pre-β HDL particle (Figure 1). In other words, apoA-I is the starting point for the synthesis of a functional HDL particle and therefore essential for the formation and maturation of novel HDL particles [16]. In the circulation, lecithin-cholesterol acyltransferase esterifies the free cholesterol in these pre-β HDL particles, thereby forming HDL3 and finally HDL2 [25]. The ATP-binding cassette G1 transporter and scavenger receptor class B type 1 (SR-B1) contribute to the cholesterol efflux from peripheral tissues and macrophages to these mature HDL particles. After binding of HDL2 to SR-B1 on the liver, cholesterol esters are taken up and lipid-depleted apoA-I is returned to the circulation. These apoA-I-rich lipid-depleted HDL particles can again acquire cholesterol and phospholipids—forming an pre-β HDL particle—or can be cleared from the circulation [26].

2.2. ApoA-I Clearance

Several organs are involved in apoA-I clearance and degradation [26]. Calculations in rabbits have indicated that renal apoA-I clearance accounts for approximately 68–70% of total apoA-I catabolism. Also in humans, the kidney is the major site for apoA-I clearance [26, 27]. In the kidneys, the uptake of HDL particles is limited, because the intact lipoprotein particles are too large to pass the glomerular filtration barrier. However, newly formed or recycled lipid free apoA-I can pass this barrier. In the proximal tubule of the glomerulus, apoA-I binds the receptors cubilin and megalin [28], which mediate endocytosis and delivery of the protein to the lysosomes [29, 30], resulting in complete degradation of the apoA-I protein. The amino acids can be reused for de novo protein synthesis [31]. While the kidney plays a major role in apoA-I degradation, the liver accounts for 26% of the apoA-I clearance, at least in rats. It is not known how the hepatocytes take up the apoA-I particles, but the apoA-I catabolic products are excreted from the liver via the bile into the gut. In the gut, they are further digested and absorbed or excreted from the body. Other tissues, besides kidney and liver, which are to a lesser extent involved in the degradation of apoA-I, are ovaries, adrenals, and spleen, which secrete apoA-I catabolic products into the urine (Figure 1) [26].

Increasing apoA-I concentrations via reducing apoA-I clearance is for unknown reasons not a subject of investigation. Consequently, it is also not known whether inhibiting apoA-I clearance affects HDL functionality. Therefore, decreasing apoA-I clearance is currently not a target for interventions, whereas elevating de novo apoA-I production certainly is [32].

3. Dietary Interventions Affecting ApoA-I Metabolism

It has been clearly shown that dietary components can change serum apoA-I concentrations. We here provide an overview of randomized controlled dietary intervention studies that have examined the effects of whole foods and (non)nutrients on apoA-I concentrations or apoA-I metabolism. Only crossover and parallel studies were included. Potentially relevant studies published before January 2017 were identified by a systematic search of the database PubMed (https://www.ncbi.nlm.nih.gov). The following search terms were used to search in titles and abstracts: (((Clinical Trial[Publication Type]) OR randomized controlled trial[Publication Type])) AND apolipoprotein [MeSH Terms]. The selection was performed in two steps. First, titles and abstracts were screened. Studies were selected if they met the following inclusion criteria: human intervention study with adults, dietary intervention study, and measurement of apoA-I concentrations. In the second step, full-texts of the selected articles were read to extract fasting or postprandial apoA-I values. Then, a search was performed to find meta-analysis of each food or (non)nutrient group. When a meta-analysis was found, it is included in this review together with the articles identified by us, which were not part of the meta-analysis. Changes in apoA-I concentrations were expressed as percentages, if possible. When percentages were not reported, they were calculated from the mean values as reported in the articles. Furthermore, the list of articles was screened for studies that investigated the effects on cholesterol efflux, apoA-I production rate (PR), or fractional catabolic rate (FCR).

3.1. Alcohol

Based on a meta-analysis including 16 studies with in total 374 subjects, Brien et al. concluded that alcohol consumption (women: >15 g alcohol/day; men: >30 g alcohol/day) increased fasting plasma apoA-I concentrations with 10.1 mg/dL (95% CI 7.3–12.9 mg/dL) [33]. A later study, not included in this meta-analysis, also showed a higher fasting apoA-I concentration after alcohol consumption as compared with no alcohol consumption [34]. Moreover, postprandial apoA-I concentrations also increased after alcohol consumption [35]. These effects did not depend on the source (red wine, beer, and Dutch gin) of alcohol [36]. Lavy et al. however reported that red wine increased apoA-I as compared with white wine consumption [37]. Also, Gepner et al. observed that red wine increased apoA-I concentrations as compared with water consumption, but white wine did not significantly change apoA-I concentrations as compared with water or red wine [38]. Furthermore, alcohol consumption not only elevated circulating apoA-I concentrations but also improved HDL functionality as shown by an increased cholesterol efflux capacity [36, 39, 40]. In one study, the kinetics of apoA-I have been examined. It was reported that apoA-I PR increased and apoA-I FCR decreased after alcohol consumption (Table 1) [41].

3.2. Boiled and Filtered Coffee, Caffeine, and Tea

In six studies, the effects of boiled or filtered coffee, caffeine, and tea on fasting apoA-I concentrations have been compared. In none of the studies, significant differences in apoA-I concentrations were observed (Table 2) [4247].

3.3. Fatty Acids

In a recent meta-analysis including 104 diets from forty-two well-controlled intervention studies the effects of the various fatty acids on fasting serum apoA-I concentrations were estimated. Effects of fish-fatty acids were not included in that meta-analysis and will be discussed in the next paragraph. A significant increase in serum apoA-I concentrations was found when 1 energy% of carbohydrates was replaced by saturated fatty acids (SFA; 8.4 mg/dL, 95% CI 6.4–10.5), cis-monounsaturated fatty acids (cis-MUFA; 5.5 mg/dL, 95% CI 3.7–7.3), and cis-polyunsaturated fatty acids (cis-PUFA; 2.3 mg/dL, 95% CI 0.1–4.6). cis-MUFA mainly referred to oleic acid and cis-PUFA to linoleic acid plus some α-linolenic acid. This meta-analysis further showed that fasting apoA-I concentrations were significantly increased by replacement of 1 energy% from carbohydrates with lauric acid (C12:0; 19.2 mg/dL, 95% CI 14.6–12.7), myristic acid (C14:0; 8.8 mg/dL, 95% CI 0.5–13.1), and palmitic acid (C16:0; 6.5 mg/dL, 95% CI 3.8–9.3), while replacement with stearic acid (C18:0) did not change apoA-I concentrations [48]. For these latter analyses, 88 diets from 34 studies were included. In another meta-analysis based on 17 diets from 10 studies, Brouwer (2016) described the effects of trans-fatty acids (TFA) on circulating fasting apoA-I concentrations. It was reported that replacement of 1 energy% of carbohydrates for total TFA increased apoA-I concentrations (3.3 mg/dL, 95% CI 4.7–1.9). When a difference was made between industrial and ruminant TFA, it was found that replacement with industrial TFA significantly increased fasting apoA-I concentrations (3.3 mg/dL, 95% CI 4.8–1.8), while ruminant TFA did not (4.6 mg/dL, 95% CI: −22.0–12.9). This may be due to a lack of power, since only two studies investigated ruminant TFA. Furthermore, this meta-analysis also showed that replacement of 1 energy% from TFA with SFA increased fasting apoA-I concentrations (2.6 mg/dL, 95% CI 1.4–3.9), while replacement with MUFA did not change apoA-I concentrations and replacement with cis-PUFA decreased fasting apoA-I concentrations (−1.7 mg/dL, 95% CI −2.8–−0.6) (Table 3) [49].

Several studies have examined the effects of the various fatty acids on serum apoA-I metabolism. A TFA diet increased apoA-I FCR as compared with SFA, but the FCR after cis-PUFA consumption did not differ from the TFA or SFA diets. ApoA-I PR was not different between the various diets [52]. Moreover, a cis-PUFA diet did not affect apoA-I FCR [50] and both FCR and PR decreased after low fat consumption compared with high cis-MUFA consumption [51]. In contrast, Labonté et al. have reported that replacing 13 energy% of carbohydrates with cis-MUFA decreased apoA-I FCR with no change in apoA-I PR (Table 3) [53]. The different results between these two studies [51, 53] may have been due to the significant weight loss in the study of Desroches et al., which may have confounded to some extent the effect of MUFA on apoA-I kinetics.

3.4. Fish and Fish-Fatty Acids

Most studies investigating the effects of omega-3 fatty acids from fatty fish, mainly eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), did not observe any differences in fasting and postprandial apoA-I concentrations [54, 55, 5766, 6871, 7379]. However, in two studies, all in healthy men, fasting apoA-I concentrations decreased after fish oil supplementation. The first study showed lower apoA-I concentrations after pollock oil (rich in EPA) and salmon oil (rich in DHA), but not after tuna oil (rich in DHA) consumption as compared with butter [56]. The second study found lower apoA-I concentrations after EPA oil supplementation compared with DHA oil supplementation [67]. On the other hand, one study found an increase in fasting apoA-I concentrations after a diet high in fish-fatty acids compared with a diet low in fish-fatty acids. In this study, the diets were matched for total fat (Table 4) [72].

Five studies have investigated the effects of fish on fasting apoA-I concentrations. In one study, fatty fish (salmon, rainbow trout, Baltic herring, whitefish, vendace, and tuna) consumption increased apoA-I concentrations compared with lean fish (pike, pike-perch, perch, saithe, and cod) consumption. However, it did not change apoA-I concentrations as compared with lean meat (beef and pork) consumption [83]. The other three studies did not find differences in apoA-I concentrations after fish consumption, of which two compared fatty fish with lean meat [82, 96] and one compared prawns with crab [84]. A limitation of the study of Lindqvist et al. is that participants consumed in total 35 energy% of fat in the herring period and only 10 energy% of fat in the meat period [82], which may have affected apoA-I concentrations. Comparisons between fish and meat consumption are probably not confounded by differences in the intake of the source of protein, as suggested by Gascon et al. In that study, the effects of proteins in lean fish (cod, sole, pollack, and haddock) were compared with those of animal protein (lean beef, pork, veal, eggs, skimmed milk, and milk products). No differences on fasting apoA-I concentrations were found (Table 4) [81].

3.5. Fibers

Studies comparing the effects of oat germ, low in fiber, with those of wheat germ, high in fiber, consumption did not find any differences in fasting apoA-I concentrations [80, 8589]. In four of these studies, it was explicitly reported that the macronutrient composition of the experimental diets was comparable [80, 85, 86, 88]. Mekki et al. observed that a high-fiber diet did not change fasting apoA-I concentrations as compared with a low fiber diet [90]. On the other hand, decreased fasting apoA-I concentrations were found after a high β-glucan and psyllium diet as compared with a low fat, low-cholesterol control diet [91]. The water-soluble fiber arabinoxylan also decreased fasting apoA-I concentrations as compared with the control diet, which had a similar macronutrient composition [92]. Furthermore, no differences in fasting apoA-I concentrations were observed between the soluble and insoluble forms of P. ovate [93]. The water-soluble fiber β-glucan did not affect fasting apoA-I concentrations [94]. Furthermore, wheat germ consumption increased fasting apoA-I concentrations compared with flaxseed consumption (Table 5) [95].

3.6. Nuts

In one short-term study, walnut consumption significantly increased fasting serum apoA-I concentrations [97], but these effects were not found in two longer-term studies [98, 99]. Almond consumption did also not affect fasting apoA-I concentrations [100, 101]. Likewise, hazelnuts [102, 103] and pistachio nuts did not change fasting apoA-I concentrations [104]. A limitation of some of the studies is that not all experimental diets were matched for differences in fat and fatty acid composition. In some of these studies, the diets containing nuts provided more energy from fat than the control diets [99102]. Furthermore, the nut diets were sometimes also lower in SFA and higher in PUFA than the control diets [99, 101]. Although these differences in nutrient intakes are inherent to consuming more nuts, it is not likely that the effects observed are due to minor component in nuts, since fatty acids increase apoA-I concentrations as compared with carbohydrates [48]. However, most other studies that used a control diet with similar fat and fatty acid composition did also not find any effects of the consumption of nuts on apoA-I concentrations (Table 6) [98, 103, 104].

3.7. Plant Sterols and Stanols

Most studies examining the effects of plant sterols on serum lipids did not demonstrate an effect of plant sterols on fasting apoA-I concentrations [106, 107, 109114]. In one study, comparing olive oil and olive oil with plant sterol esters and sunflower oil with plant sterol esters, fasting apoA-I concentrations increased when plant sterol esters were consumed together with olive oil, but apoA-I concentrations were comparable during the other two interventions [108]. Furthermore, one study showed an increase in fasting apoA-I concentrations comparing 3 months of prudent diet consumption (National Cholesterol Education Program) with added plant sterols, with prudent diet consumption alone [111]. One study examined the effects of plant stanols on fasting apoA-I concentrations and found increased apoA-I concentrations comparing 6 weeks of sitostanol consumption with no sitostanol consumption [105]. Finally, no changes in apoA-I PR and FCR were found after plant sterol or stanol consumption (Table 7) [105, 110].

3.8. Soy Proteins or Isoflavones Isolated from Soy

Studies investigating the effects of soy protein on fasting apoA-I concentrations showed inconsistent outcomes. Eight studies using different amounts of soy protein for 3 weeks till 3 months did not find changes in fasting apoA-I concentrations [115, 116, 118121, 123125]. On the other hand, in one study products containing soy protein increased fasting apoA-I concentrations as compared with products containing casein [117], while in another study products with soy protein decreased fasting apoA-I as compared with products containing casein [122]. Furthermore, two studies found different effects of various soy products on fasting apoA-I concentrations [126, 127]. Soy-milk increased apoA-I concentrations as compared with soy nuts and soy flour, but no differences were found as compared with animal protein [126]. Soy nut and soy protein consumption increased apoA-I concentrations as compared with the control group without soy [127]. Two studies have investigated the effects of isoflavones isolated from soy on apoA-I concentrations and showed no effect on fasting [128, 129] and postprandial apoA-I concentrations (Table 8) [129].

3.9. Others

Many other products and food components have been studied for their effects on apoA-I concentrations. In most of these studies, which included eggs [130], dried garlic [131, 134], beta-carotene [132], phytochemicals with cytochrome P-450-inducing activity [133], magnesium [135], eggplant [136], dry beans [137], kiwifruits [138], and polyphenols [139], no effects on fasting apoA-I concentrations were observed. In addition, sphingolipids did not change postprandial apoA-I concentrations [140]. On the other hand, red grape juice [141], a mixture of citrus flavonoids and tocotrienols [142], vitamin D supplementation [143, 144], vitamin D plus calcium supplementation [144], theobromine [145], orange juice [146], and a high dose of grape pomace and omija fruit [147] all increased fasting apoA-I concentrations (Table 9).

4. Pharmacological Approaches Targeting ApoA-I Metabolism

Although not always specifically developed for this purpose, several well-known drugs like statins [7, 167] and CETP inhibitors [168171] affect serum apoA-I concentrations. However, since this review focuses on novel strategies to increase serum apoA-I concentrations, we here describe only approaches that are currently in development and are specifically designed to target a change in apoA-I metabolism. Potentially relevant studies published before January 2017 were identified by a systematic search of the database PubMed (https://www.ncbi.nlm.nih.gov). The following search terms were used to search in titles and abstracts: (Pharmacological AND approaches AND apoA-I). First, all abstracts were screened and the pharmacological approaches were divided into three categories: apoA-I mimetics, apoA-I infusions, and others. Second, a new search was performed with the search terms: (apoA-I mimetics AND apoA-I infusions AND RVX-208 AND LCAT infusion AND clinical trial) to select all studies published before January 2017 that investigate apoA-I mimetics, apoA-I infusions, and RVX-208 in humans.

4.1. ApoA-I Mimetics

ApoA-I mimetics are small amphipathic peptides that resemble apoA-I in biological function and structure [172]. These mimetics are not the intact apoA-I protein, but small fragments of the protein with certain biological functions. These small peptides can be given orally or can be infused [15, 173]. To prevent digestion in the gastrointestinal tract, mimetics are made from D-amino acids, which are resistant to human gastrointestinal proteolytic enzymes [14]. Over the years, several mimetics have been produced, but none of them has all the antiatherogenic functions of apoA-I. However, combining several mimetics can be a theoretical approach to mimic all antiatherosclerotic properties of apoA-I [174]. The only mimetic that has been tested in humans is D-4F. When 50 patients with coronary artery disease received a single oral dose (30, 100, 300, and 500 mg) of this mimetic, the two highest doses increased the anti-inflammatory activity of the HDL fraction. However, no changes in lipids or lipoprotein concentrations were seen. D-4F was shown to be safe and well tolerated (Table 10) [148]. Unfortunately, the effects of D-4F on cholesterol efflux in humans have not yet been investigated.

4.2. ApoA-I Infusions

Besides apoA-I mimetics, apoA-I itself, either by using delipidated HDL or by using delipidated HDL combined with phospholipids, can be infused directly into the circulation. The theoretical advantage of using apoA-I or apoA-I-phospholipid complexes instead of using apoA-I mimetics is that the apoA-I protein is completely intact and still possesses all its biological functions and might therefore have a larger atheroprotective effect. So far, three different forms of apoA-I have been tested, that is, apoA-I Milano (MDCO-216), CSL-111/CSL112, and CER-001.

4.2.1. ApoA-I Milano

In a randomized human controlled trial, 47 patients with acute coronary syndromes received for 5 weeks one infusion of placebo or recombinant apoA-I Milano/phospholipid complex (ETC-216) at 15 or 45 mg/kg per week. At the end of the study a significant reduction in atheroma volume was found in the high dose group [149]. This reduction in atheroma volume was accompanied by a reduction in external elastic membrane volume of the artery, but not with a change in lumen volume [175]. Recently, in a randomized controlled study, patients with stable coronary artery disease received 5 doses of 10, 20, 30, and 40 mg/kg MDCO-216 infusion. This resulted in a dose-dependent increase in apoA-I concentrations and a dose-dependent shift from small- to large-sized HDL particles [150]. Moreover, a profound increase in ABCA1-mediated cholesterol efflux was observed [151]. However, very recently the MILANO-PILOT study failed to slow down the regression of coronary atherosclerosis with 5 weekly infusions of 20 mg/kg MDCO-216 in 120 patients with acute coronary syndromes. In fact, significant reductions in HDL-C and apoA-I concentrations were observed, while there were no effects found on percent atheroma volume and total atheroma volume (Table 10) [152].

4.2.2. CSL-111/CSL112

In one clinical study, 40 and 80 mg/kg CSL-111 were infused once a week for one month in 183 patients elected for coronary angiography. Treatment of the high dose group (80 mg/kg) was discontinued early, because some of the patients exceeded the upper level of alanine aminotransferase by 100-fold. The low dose group (40 mg/kg) showed a significant reduction in atheroma volume. However, this reduction was not significantly different from the decrease in the placebo group [176]. After this, the further development program of CSL-111 was discontinued because of the unfavorable hepatic abnormalities. As a follow-up, one phase I study has been performed using CSL112, which is a similar compound, but postulated without effects on liver function. In this study, a single dose (5, 15, 40, 70, 105, or 135 mg/kg) or multiple doses for 4 weeks (3.4 or 6.7 g once a week or 3.4 g twice a week) of CSL112 was administrated intravenously to healthy volunteers. Both the single and multiple doses of CSL112 dose dependently increased serum apoA-I and serum HDL-C concentrations. Moreover, also pre-β HDL particle concentrations and cholesterol efflux capacity were increased. In the single dose study, dose-dependent effects were found on HDL-C [153, 154]. Recently, two studies showed that CSL112 was indeed safe for human consumption, with no effects on liver function parameters [155, 156]. In the first study, patients with atherosclerosis were given infusions of 1.7, 3.4, and 6.8 g CSL112 or placebo. The CSL112 infusions resulted in a dose-dependent increase in apoA-I and total cholesterol efflux [155]. In the second study patients with myocardial infarction received infusions of 2 or 6 g CSL112 or placebo for 4 weeks. Here also a dose-dependent increased in HDL-C, apoA-I, and cholesterol efflux was shown (Table 10) [156].

4.2.3. CER-001

In one clinical study, 417 patients with acute coronary syndromes were randomized for 6 weekly infusions of 3, 6, and 12 mg/kg CER-001 or placebo. No changes in atheroma volumes were found. It was speculated that a higher dose or a different patient group would have shown more positive results [157]. In a recent human study, 9 infusions of 8 mg/kg CER-001 were given twice weekly for 28 days to 7 patients with familial hypoalphalipoproteinemia, who were severely deficient in HDL. In this patient group, CER-001 significantly increased serum apoA-I and HDL-C concentrations and reduced atherosclerotic lesion size, measured using Magnetic Resonance Imaging. Moreover, an increase in cholesterol efflux from macrophages and a higher fecal neutral sterol excretion was seen, which may indicate improved RCT [158]. Additionally, 12 biweekly infusions with 8 mg/kg CER-001 showed increased apoA-I concentrations, a decrease in vessel wall area, and a trend toward a reduction in vessel wall thickness [159]. Recently, a study evaluated the effects of 3 mg/kg CER-001, in patients with atherosclerotic carotid artery disease, and showed increased apoA-I concentrations, with a simultaneously increased cholesterol efflux capacity [160]. Unfortunately, preliminary data of a recent clinical trial in patients with coronary atherosclerosis did not show beneficial effects of CER-001 on atheroma volume and LDL-C [161] (Table 10).

4.3. Others
4.3.1. RVX-208

The first class of compounds affecting apoA-I metabolism refers to the apoA-I transcriptional upregulator RVX-208. RVX-208 is an oral, small synthetic quinazoline molecule, which binds bromo- and extra terminal (BET) proteins and upregulates apoA-I gene transcription via an epigenetic mechanism.

In the first human clinical trial, 18 healthy subjects received varying and multiple doses (1 to 20 mg/kg per day) of RVX-208 or placebo for 7 days. Plasma apoA-I concentrations were increased, and more importantly, an increase in pre-β1-HDL concentrations and a higher ABCA1-mediated cholesterol efflux was demonstrated [162]. The outcome of the recent phase 2 randomized placebo-controlled clinical ASSERT trial, evaluating the effect of RVX-208 on serum apoA-I concentrations and CHD risk in human, was less positive. In that study, 299 patients with stable coronary artery disease received placebo or RVX-208 at three different dosages (50, 100, and 150 mg) twice daily for 12 weeks. Only a nonsignificant increase in serum apoA-I concentrations was found. Unfortunately, HDL functionality and cholesterol efflux capacity were not studied [163]. A second study using RVX-208 is the phase 2b clinical trial SUSTAIN. In this trial, 172 statin-treated patients (Rosuvastatin or Atorvastatin) with low serum HDL-C concentrations were treated with 200 mg/day RVX-208 for 24 weeks. Both serum apoA-I concentrations and HDL particle numbers increased significantly. Furthermore, RVX-208 was found to be safe for oral use [164]. In another phase 2 clinical trial, the ASSURE study, 323 statin (Rosuvastatin or Atorvastatin) treated patients with coronary artery disease and low serum HDL-C concentrations received 100 mg RVX-208 twice daily for 26 weeks. However, no significant reductions in atheroma volume or increases in HDL-C and apoA-I concentrations were seen [32]. Finally, a recent study in subjects with prediabetes showed that 100 mg RVX-208 for 29–33 days did not increase HDL-C and apoA-I concentrations, while it increased the concentration of medium-sized HDL and decreased the concentration of small-sized HDL particles. Furthermore, RVX-208 delayed and reduced oral glucose absorption and endogenous glucose production (Table 10) [165].

4.3.2. LCAT Infusion

The first human study investigating the effects of lecithin-cholesterol acyltransferase (LCAT) infusion investigated only one patient with familial LCAT deficiency. Recombinant human LCAT was infused 3 times for 1 hour in a dose optimization phase (0.3, 3.0, and 9.0 mg/kg) and after this 1 to 2 weekly infusions were given of 3.0 or 9.0 mg/kg for 7 months. LCAT infusion improved renal function, increased apoA-I, HDL-C, and to a lesser extent LDL-C. Furthermore, after infusion, postprandial triacylglycerol concentrations decreased [166]. These results are promising; however, before drawing conclusions about LCAT infusion clinical trials including more patients should be done.

5. Conclusion

Alcohol consumption increases fasting apoA-I concentrations and may improve cholesterol efflux, possibly via increasing apoA-I PR and decreasing FCR. Further, replacement of carbohydrates for SFA, cis-MUFA, cis-PUFA, and TFA increases fasting apoA-I concentrations. The effects of the various SFA are different, since lauric, palmitic, and myristic acids increase apoA-I concentrations, while stearic acid does not. The different fatty acids affect apoA-I metabolism differently, but results are conflicting. Therefore more studies are needed to better understand the effects of the various macronutrients on apoA-I kinetics.

Coffee, caffeine, tea, omega 3 fatty acid, fish, nuts, plant sterol and stanol, different soy proteins, and isoflavones isolated from soy do not change fasting apoA-I concentrations. Moreover, the effects of the various types of fibers may be different; the consumption of diets rich in wheat germ did not modify apoA-I concentrations, while the consumption of diets rich in psyllium, arabinoxylan, and flaxseed may decrease fasting apoA-I concentrations. However, these types of fibers have only been examined in a limited number of studies. Therefore, we conclude that fiber consumption does not have a profound impact on fasting apoA-I concentrations.

Finally, five other food components showed a promising increase in fasting apoA-I concentrations: citrus, vitamin D, theobromine, orange juice, and a high dose of grape pomace and omija fruit. However, these findings need to be confirmed in future studies. Additional research is also needed to examine the effects of these products or food components not only on apoA-I kinetics, but also on HDL functionality.

Overall, all three categories of pharmacological approaches showed that targeting apoA-I concentrations and/or HDL functionality by a pharmacologic approach can increase apoA-I functionality and might improve CHD risk markers, including vessel wall characteristics and inflammation. The mimetic D-4F is promising, but clinical studies are required to investigate the effects on HDL functionality. The CSL112 and LCAT infusions are the most promising of the infusion therapies, but studies are needed to investigate the effects of CSL112 on CHD risk markers, including vessel wall characteristics and inflammation, and LCAT infusions need to be investigated in clinical trials with more patients. Unfortunately, recent clinical studies showed no improvement in CHD risk markers after apoA-I Milano, RVX-208, or CER-001 therapy.

Although we cannot exclude that we have missed studies during the systematic searches and studies with positive results are overrepresented, we conclude that both dietary components and pharmacological approaches can be used to increase apoA-I concentrations. For the dietary components in particular, more knowledge about underlying mechanisms is necessary, as increasing apoA-I per se does not necessarily translate into a reduced CHD risk.

Conflicts of Interest

The authors have no conflicts of interest.

Acknowledgments

This research was supported by the Dutch Technology Foundation (STW), which is part of the Netherlands Organisation for Scientific Research (NWO) and which is partly funded by the Ministry of Economic Affairs.