Since one of the first reports in 1951 , a plethora of prospective epidemiological studies have shown that plasma HDL-C levels and the risk for cardiovascular disease (CVD) are inversely related . The magnitude of the association is quantified in a widely cited metaanalysis showing that a 1 mg/dl increase of HDL-C level is associated with a 2–3% decrease in CVD risk . A number of mechanisms by which HDL-C may influence the process of atherosclerosis have been described (reviewed in Kontush and Chapman ). Fielding and Fielding  demonstrated in 1982 that HDL can act as an acceptor of cellular cholesterol, which is proposed to constitute the first step in a hypothetical pathway that is known as reverse cholesterol transport (RCT). RCT is defined as the uptake of cholesterol from peripheral cells by lipid-poor apolipoprotein A-I (apoA-I) and HDL, and the subsequent delivery to the liver for excretion into the feces. The function ofHDL as a atheroprotective particle extends well beyond this role in cholesterol transport; it has also been found to have antiapoptotic , antiinflammatory , and antithrombotic  capacities.
Cardiovascular disease remains a major cause of morbidity and mortality, despite interventions to decrease the number or severity of known risk factors. As a consequence, HDL-C increasing therapy is considered a suitable target for CVD prevention. Indeed, infusion of exogenous HDL has been shown to result in improvements in hallmarks of atherosclerosis, such as endothelial function , coronary atheroma volume , and plaque morphology . These studies, however, were small scaled and should be regarded as a proof-of-concept that an increase in HDL-C does change biomarkers of atherosclerosis.
Repetitive infusions of HDL are not considered a feasible therapy for the wide scaled burden of CVD; it would require large resources to produce considerable quantities of this lipoprotein. Consequently, large efforts have been put in the strive for orally administered HDL-C increasing medication. Several of these are available (i.e. fibrates or nicotinic acid derivates).
Most of these drugs, however, do not exclusively raise HDL, but have beneficial effects on other risk factors as well, and as such, the contribution of the HDL-C increase per se has been impossible to delineate.
However, efforts in the field of HDL therapeutics have resulted in the recent development of cholesteryl ester transfer protein (CETP) inhibitors. This class of drugs induces significant HDL-C increases. The ILLUMINATE trial, in which 15 000 individuals were randomized to either statin or statin combined with the CETP inihibitor torcetrapib, did show increased mortality among the torcetrapib-treated patients, despite an impressive (>100%) increase in HDL-C [12–14]. The result is counterintuitive, and the underlying mechanism has not been fully elucidated, but off target effects of the torcetrapib molecule are the most likely explanationto date.
After ILLUMINATE, a subtle shift in the HDLatherosclerosis paradigm has occurred. Whereas in the past the main focus was to increase HDL-C levels, the concept has changed to increase HDL functionality [15,16].
With this new paradigm kept in mind, apoA-I should be considered an attractive target. apoA-I is the structural protein of the HDL particle and it plays a crucial role in many of the beneficial effects of HDL. Studies in rodents have established that apoA-I does play a role in atherosclerosis, although the magnitude of the effect differed among the studies, which might be in part attributable to the genetic background of the animals and additional manipulations used in these models (reviewed in Vergeer et al. ). The fact that apoA-I is a crucial protein for HDLgeneration is exemplified by the finding that apoA-I KO mice have 75% lower HDL-C levels. These apoA-I knockout mice, however, do not show increased susceptibility to atherosclerosis . In an atherogenic background, such as an LDL receptor null mice fed with an atherogenic diet, apoA-I deficiency results in markedly increased atherosclerosis .
A combined knockout apoA-I and human apoB transgenic mice model also showed increased atherosclerosis when fed a western diet . Overexpression of hApoA-I in atherosclerosis prone animal models do show uniform results. Hepatic overexpression of human apoA-I has been shown to reduce the extent of atherosclerosis in mice on proatherogenic diets  and these findings have been reproduced in different animal models [22,23]. Clearly, these data point towards apoA-I as a very attractive interesting target and the finding that human carriers of mutations in the apoA-I gene do suffer from increased atherosclerosis strongly supports this view . The development of apolipoprotein A-I mimetic peptides apoA-I is a relatively large protein, comprising 243 amino acids in 10 amphipathic alpha helices, of which most are separated by a proline residue. Due to its size, apoA-I requires parenteral administration. As a consequence, it was compulsory to design smaller mimetic peptides, without losing the lipid binding and antiatherogenic capacities of these molecules. In 1985, Anantharamaiah et al.  produced an 18 amino acid containing peptide folded into an alpha-helix. The peptide did not have direct sequence homology to either one of the helices of the apoA-I protein, but the secondary structure of apoA-I was replicated. This so called 18A peptide was found to have similar properties to the apoA-I helices in terms of charge distribution and lipid-binding capacity. Cellular
efflux-inducing capacity of 18A was shown to be similar to apoA-I in an assay where cultured mouse fibroblasts were used . In addition, like apoA-I, 18A was shown to activate lecithin : cholesterol acyltransferase (LCAT), a pivotal enzyme in HDL maturation. As such, 18A was thus shown to have quite similar properties in lipid metabolism as apoA-I.
In an attempt to optimize the peptide, stability was improved by replacing the existing nonpolar amino acids on 18A with phenylalanine (F) residues. The increasing number of F residues (resulting in 2F, 3F, 4F, 5F, 6F, and 7F, reflecting the number of F residues) did result in increased lipid affinity and hydrophobicity; 6F and 7F were the most hydrophobic, but these peptides were shown to lose the affinity to bind phospholipids, and in subsequent experiments the main focus has been on the 4F and 5F peptides.
Peptides can either be synthesized from L-amino acids (i.e., L-4F) or D-amino acids (D-4F). In-vitro properties have been shown to be similar for both isomers [27,28], but the L-isomer is significantly more prone to proteolysis compared to the D-isomer , as mammalian enzymes recognize peptides made from L-amino acids, but not from D-amino acids.
In addition to lipid binding and hydrophobicity, peptides have been tested for their anti-inflammatory and antioxidant characteristics. A commonly described method is the one described by Navab et al. , who have used a model in which LDL-C induced activation of monocyte chemotactic activity was tested in an in-vitro arterial wall co-culture model In these models, mimetic peptides are compared with each other and with apoA-I. Attempts to mimick apoA-I in all its antiatherogenic effects have thus far not been successful, but mimetics have been shown to be superior to apoA-I when it comes to specific characteristics. 4F, for example, was shown to bind oxidized lipids with orders of magnitude higher affinity compared to apoA-I . This finding might explain why these peptides have biological effects despite low plasma concentrations. The large difference between the effects of native and mimetic apoA-I suggests that ‘mimetics’ might be considered not to be the right nomenclature.
Tandem helical mimetic peptides
As said above, apoA-I contains multiple helices linked by a proline residue. Single-helix peptides have been shown to be efficient antiatherogeneic molecules, but tandem helical mimetic peptides have been created in an attempt to further increase efficacy by more closely mimicking the apoA-I peptide. In a direct comparison, Wool et al.  showed that symmetrical peptides, composed of two 4F peptides linked by a proline or alanine residue, were more efficient in a cholesterol efflux from lipid-loaded murine macrophages compared to the single 4F peptide. Coppermediated oxidation of purified mouse LDL was inhibited by the 4F peptide, but the tandem peptides increased oxidation, clearly emphasizing the variability of effects between the single and tandem peptides in in-vitro assays. Recently, 22 bihelical apoA-I mimetic peptides were investigated in vitro for their capacity and specificity of cholesterol efflux, and their inhibiting effects on inflammation and LDL-oxidation . In this comprehensive first systematic analysis of multiple structural modifications, none of the peptides tested were found to be equally effective in all antiatherogenic functions. Moreover, the anti-inflammatory, antioxidant, and efflux capacities were found to be related to differential structural features of the peptides. For efflux efficiency and specificity, for example, mean hydrophobicity, charge, size, and angle of the link between two helices were crucial, whereas for antioxidant properties the presence of cysteine and histidine residues was important. The latter has also been proposed as a potential explanation for the increased antioxidant property of apoA-I Milano, a genetic variant known to be associated with reduced risk for CVD .
The fact that none of the peptides tested could outcompete the beneficial functions of apoA-I in all facets of antiatherogeneity suggests that different portions of the apoA-I peptide might be involved. As a consequence mimicking one part of apoA-I does not necessarily result in an overall beneficial property of the designed apoA-I mimetic peptide. A combination of different mimetic
peptides, harboring various beneficial effects, might therefore be a prerequisite to mimick the full spectre of antiatherogenic characteristics of apoA-I. One could also envision that tailored therapy can be achieved by the combination of different peptides.
Apolipoprotein A-I mimetics in animal studies
The initial in-vivo studies addressing the effect of apoA-I mimetics were performed by Garber et al. . In this study, C57Bl/6J mice were put on an atherogenic diet and 5F was administered daily by intraperitoneal injection at a dose of 20mg/day for 16 weeks. Mice treated with phosphate-buffered saline (PBS) or murine apoA-I (50mg/day) were used as controls. Lipids and lipoproteins were not significantly altered upon 5F injections and administration of 5F was found to be nontoxic. Importantly, no antibodies to the injected materials were observed. The aortic atherosclerotic lesion area was significantly less (44% reduction) in 5F-treated mice compared to mice receiving placebo or murine apoA-I. HDL isolated from 5F-injected mice was shown to be superior to HDL derived from PBS and apoA-I-treated mice in terms of reduction of monocyte chemotaxis and lipid hydroxyperoxidase formation . This, in combination with the finding that lipid levels were not changed by 5F injection, shows that the functionality of the HDL pool was enhanced by 5F. In subsequent in-vivo studies, the main focus has been on 4F. This is in part due to the fact that D-4F can be administered orally in drinking water or by gavage. In apoE knockout mice as well as in LDL-r KO mice, D-4F was shown to decrease the aortic root lesions by more than 70% , despite very low bioavailability and low plasma concentrations. The marked reduction in atherosclerotic lesions occurred independent of changes in total plasma or HDL-C. However, administration of D-4F in the drinking water resulted in a substantial change of HDL from a proinflammatory to an anti-inflammatory particle, emphasizing that D-4F therapy induced a qualitative rather than a quantitative effect on HDL. In a subsequent study, this suggestion was confirmed by showing that administration of D-4F caused a change in HDL distribution towards prebeta HDL (a fraction known for its efflux efficacy), and that it increased HDLassociated paraoxonase activity .
Low doses of D-4F and pravastatin work synergistically on HDL-C level, HDL function, and atherosclerosis prevention in an apoE null mice model. This might be relevant for future clinical perspectives . Bielicki et al.  published the beneficial effect of a new HDL mimetic single-helix peptide (ATI-5261) in LDL-r KO mice and apoE null mice on a 13–18 week during high-fat western diet. Daily intraperitoneal injections of ATI-5261 (30 mg/kg) for 6 weeks reduced atherosclerosis, as judged by lesion area covering the aorta, by 30% and 45% inLDL-R/ and apoE/ mice, respectively. Interestingly, one single intraperitoneal injection of ATI-5261was found to increase reverse cholesterol transport from macrophage foam cells to feces over 24–48 h.
Emerging perspectives for apolipoprotein A-I mimetics
The effects of apoA-I mimetics is not restricted to dyslipidemic mice models. In a recent study, Vaziri et al.  showed the benefit of L-4F (5 mg/kg s.c. three times weekly for 4 weeks) on inflammation and oxidative stress in a chronic kidney disease (CKD) rat model. L-4F attenuated a large number of the markers for inflammation and oxidation (such as NAD(P)H oxidase subunits, COX-2, 12-lipoxygenase, MCP-1, PAI-1, myeloperoxidase) in the thoracic aorta without altering plasma lipids.
Compared with WT mice, ApoE-deficient mice have been shown to suffer significantly more from signs of renal dysfunction such as proteinuria, tubulo-interstitial inflammation, and mesangial expansion. Oral D-4F administration was recently found to have favorable effects on all of these characteristics . Further studies in CKD patients are awaited to address whether apoA-I mimetics hold promise for atherosclerosis regression and preservation of kidney function. Another patient population that might benefit from apoAI mimetics are septic patients. Apolipoproteins and HDL have been shown to beneficially change the inflammatory response to lipopolysacccharide (LPS) [41–43], and a recent study  showed that intraperitoneal injection of 4F significantly blunted the hypotensive response to LPS (LPS-treated rats: 34% decrease, LPS and 4F treated rats: 17% decrease in systolic blood pressure). In order to unravel the underlying mechanism, aortic ring segments from LPS-treated rats were isolated. These exvivo studies showed a reduced contractile response to phenylephrine in aortae of LPS-treated rats compared to controls, and this reduced contractility was reverted by 4F via nitric oxide synthase 2 (NOS2) downregulation.
The concentration of circulating endotoxin was significantly reduced in 4F-treated rats, and interestingly, HDL-C levels increased in the LPSþ4F-treated animals (38–45 mg/dl). Similar to other studies, an HDL-C reduction was found in LPSþvehicle-treated rats (38–28 mg/dl). The findings of this study are in line with previous studies showing that the apoA-I mimetic 4F significantly inhibits the induction of proinflammatory mediators by LPS in cultured endothelial cells . Interestingly, 4F was shown to reduce 24 h mortality in LPS-treated rats (60% mortality in LPS and 10% in LPSþ4F, total 40 animals).
In addition, transplant-associated vasculopathy is reverted upon intraperitoneal D-4F administration in a mouse heart transplant model . The authors hypothesize that this is partly mediated through D-4F induced hemeoxygenase-1 upregulation .
In a mouse lupus model , L-4F was not only shown to reduce atherosclerosis progression, but it also resulted in a reduction of IgG anti-dsDNA, proteinuria, and glomerulonephritis, suggestive of protective effect on lupuslike disease. Beneficial vascular effects have also been described in a systemic sclerosis model .
Administration of L-4F (2 mg/kg/day) to ob/ob mice reduced adiposity, improved insulin sensitivity, and improved glucose tolerance , which might be related to apoA-I mimetic-induced increase in uncoupling protein 1 (UCP1) mRNA and protein levels as well as the stimulation of AMPK phosphorylation in brown adipocytes in culture .
Other favorable effects of mimetics have shown to be a reduction of platelet aggregation , an increase of cognitive function in an Alzheimer’s mice model  and prevention of fibrosis after onset of steatohepatitis .
Apolipoprotein A-I mimetics in human studies
On the basis of all intriguing beneficial findings in animal studies, apoA-I mimetics hold promise for human therapy. Apart from its anticipated efficacy, apoA-I mimetics have more positive characteristics: they are safe, well tolerated, and relatively inexpensive compared to rHDL . Moreover, the size of the peptide does implicate that administration might be possible in an oral formulation, which is a prerequisite for a long-term treatment in large numbers of patients. Modifications for optimal oral delivery are in development .
The first clinical trial of oral D-4F was performed in CAD and high-risk patients who received a single dose of 30, 100, 300, or 500mg of D-4F (n¼8 for each dose) or placebo (n¼8) under fasting conditions. Ten additional patients received 500mg (n¼8) or placebo (n¼2) with a low-fat meal. The Tmax was shown to be 30 min and D-4F was detectable in plasma at all dosages. The single dose was well tolerated and shown to be safe . No effect on lipids and lipoproteins was found, but the anti-inflammatory index, assessed by comparing the ability to inhibit LDL-induced monocyte chemotactic activity in cultures of human aortic endothelial cells, increased in the highest doses compared to placebo.
Additional studies focussing on these vascular effects are expected.
Other apolipoprotein A-I-based therapies
Apart from apoA-I mimetics, several other apoA-I-based therapies have emerged and these have reached different stages of research. Full-length apoA-I, recombinant HDL, apoA-I enhancers and active delipidation are among these therapies. Results from studies with these are eagerly awaited.
Recent studies have increased our understanding of the effects of apoA-I mimetics. All peptides studied have been shown to induce beneficial effects on oxidation, inflammation, and cholesterol efflux and combining different peptides might be a prerequisite to establish the full spectrum of possible beneficial effects. Whereas the focus for apoA-I mimetics has traditionally
been directed towards atherosclerosis, recent studies have shown effects in several other disease states. These provocative findings do require further investigations in carefully designed clinical trials in humans.