ApoA‐I_Milano from structure to clinical application

Abstract

Apolipoprotein A‐I_Milano (apoA‐I_M) is characterized by a Cys for Arg substitution and formation of homo‐ and apoA‐II heterodimers. In the last years, some important insights on apoA‐I_M have been provided, i.e. that apoA‐I_M/apoA‐I_M has a very long permanence in blood, as well as a high capacity to remove tissue cholesterol, in addition to further attractive properties. These effects stimulated a number of attempts on alternative developments in high‐density lipoprotein (HDL) therapy, i.e. to the so‐called apoA‐I mimetics. A frequent criticism, i.e. the lack of a detailed comparative assessment of the capacity of apoA‐I versus apoA‐I_M in arterial cholesterol removal, has recently been overcome by a trial of gene therapy with apoA‐I_M versus wild‐type apoA‐I in hyperlipidemic mice. Structural studies on the dimer indicate that this very stable compound is present in small HDL, containing a single molecule of dimer, and apparently is responsible for the higher effectiveness of apoA‐I_M sera in mobilizing cholesterol from macrophages, when the ATP‐binding cassette, subfamily A, member 1 (ABCA1)‐driven efflux is stimulated. An additional protective trait of apoA‐I_M carriers is the peculiar susceptibility to proteolysis of the heterodimer, an interesting hypothetical mechanism for cardiovascular protection. The heterodimer would, in fact, act as a suicide substrate for arterial metalloproteinases, responsible for plaque rupture.

• Animal model
• apolipoprotein A‐I_Milano
• cardiovascular risk
• high‐density lipoprotein (HDL)
• HDL therapy
• IVUS
• liposome
• phospholipid
• synthetic HDL
• transgenic mouse

Background: epidemiological data on apoA‐I_M carriers

Apolipoprotein A‐I_Milano (apoA‐I_M) is currently the most widely studied mutant of human apolipoproteins, besides being the first described mutation of these proteins 1. ApoA‐I_Milano differs from wild‐type apoA‐I with an Arg¹⁷³→Cys substitution, leading to the formation of disulfide‐linked homodimers and heterodimers with apoA‐II 2, 3. All the 40 apoA‐I_M carriers identified up to now are heterozygous for the mutation and are characterized by severe hypoalphalipoproteinemia, normal to elevated plasma low‐density lipoprotein (LDL)‐cholesterol, and moderate hypertriglyceridemia 4. Contrary to expectations from their lipoprotein pattern, apoA‐I_M carriers appear to have a very low cardiovascular risk. In fact, during the last century only two carriers (males, age 54 and 57 years) have developed acute coronary syndromes. This extremely low incidence should be related to an approximately 3,500 life years' exposure, typically involving a 10–20‐fold higher risk 5. Furthermore, the carriers do not differ to any extent in carotid intima media thickness from their close relatives living in the same environment, with high‐density lipoprotein (HDL) levels in the normal/high range, as recently assessed by a highly sensitive B‐mode ultrasound method 6.

Molecular mechanism(s)

Considerable effort is being applied to the understanding of the mechanism(s) whereby the apoA‐I_M mutation may be linked to the apparent cardiovascular protection. The presence of three different protein forms of apoA‐I_M—monomeric with a very rapid turnover, dimeric with a very slow turnover 7 and recently selected for clinical development 8, and apoA‐I_M/apoA‐II heterodimer, frequently percentage‐wise the most represented protein form in the carriers—has led to a dispute on the effectiveness of this natural product versus e.g. wild‐type apoA‐I. This dispute has led to two different approaches, one investigating physicochemical properties of the three conformations of apoA‐I_M, with special interest for the heterodimer apoA‐I_M/apoA‐II, the second addressing the differences between apoA‐I_M and wild‐type apoA‐I in terms of potential therapeutic properties.

Studies in human carriers and transgenic mice have disclosed a higher capacity of apoA‐I_M sera to extract cholesterol from peripheral cells compared to control sera 9, thus implying a greater efficiency of apoA‐I_M in the first step of reverse cholesterol transport. Indeed, among the circulating forms of apoA‐I_M, the homodimer is characterized by a high efficiency in promoting cell cholesterol efflux 3. Very recently it was observed that particles containing the apoA‐I_M homodimer are optimal substrates for cell membrane ATP‐binding cassette, subfamily A, member 1 (ABCA1) 10, i.e. the membrane system allowing interaction with apoA‐I and mediating cell cholesterol efflux 11. Detailed studies on the mechanism of interaction of HDL from carriers with different cells (fibroblasts, macrophages) clearly indicate the unique presence of a lipoprotein component with electrophoretic mobility intermediate between the pre‐beta and the alpha positions 10. This component, a small HDL particle containing a single molecule of the apoA‐I_M dimer, appears to be responsible for the higher effectiveness of apoA‐I_M sera in mobilizing cholesterol from macrophages when the ABCA1‐driven efflux is stimulated. Sera from carriers thus have both pre‐beta HDL containing the wild‐type apoA‐I, as well as the newly identified small particle, specifically including the dimer, both effective in removing cell cholesterol via ABCA1 10. In contrast, in a very recent paper no difference was observed between apoA‐I_M and apoA‐I in promoting cholesterol efflux from macrophages using reconstituted particles and lipoproteins from transgenic mice 12. This discrepancy could be explained by the size/composition of the test particles: heterogeneity in reconstituted HDL (rHDL) diameter, HDL subclass distribution and lack of apoA‐I_M/apoA‐II heterodimer in mouse lipoproteins.

Key messages

• ApoA‐I_Milano results from a Cys for Arg substitution; in spite of a reduced high‐density lipoprotein (HDL) cholesterolemia, the carrier status appears to be a protective trait for the cardiovascular system.

• Carriers' HDL are characterized by a higher efficiency in mobilizing cholesterol via ABCA1 (ATP‐binding cassette, subfamily A, member 1) and by a higher susceptibility to metalloproteinase degradation.

• A trial of gene therapy with apoA‐I_Milano versus wild‐type apoA‐I in hyperlipidemic mice confirmed a better protection with the mutant phenotype; infusion of apoA‐I_Milano liposomes in a rabbit model of atheromatous lesions removed cholesterol from lipid‐rich plaques within few hours.

Potential additional mechanism for coronary protection in the apoA‐I_M carriers

Considerable uncertainty exists as to whether apoA‐I_M carriers may have further beneficial traits, besides the above‐listed properties affecting lipid metabolism 2, 13. Human apoA‐I is a substrate for a variety of proteolytic enzymes, both in plasma and tissues 14–16. While in plasma, the presence of inhibitors prevents significant degradation, allowing only small amounts of fragments for urinary excretion 17. ApoA‐I appears to be degraded to a larger extent in tissues. Tissues with a pronounced proteolytic activity include the arterial walls, where inflammatory factors, including cholesterol crystals 18, may lead to plaque instability and in turn to plaque rupture and acute ischemic events 19, 20.

While these observations have never been associated to any definite enzymatic process within the arterial wall, recent findings indicate specific apoA‐I cleavage patterns by well identified macrophage‐derived metalloproteinases (MMPs) 15. These proteolytic activities, exerted at the arterial wall level on apoA‐I‐containing HDL may, on the one hand, potentially reduce cholesterol efflux capacity 21; on the other hand, proteolysis of apoA‐I may protect the arterial wall matrix from degradation, by offering an alternative suicide substrate. Among the cell types found in atheromas, activated macrophages secrete several MMPs (‐3, ‐7, ‐9, ‐12) 15; rat peritoneal macrophages (RPM) cultured 24 h in Medium 99 are a handy in vitro model of MMP production 15. We report in the following some unpublished results from our laboratories.

When HDL containing apoA‐I_M and HDL containing apoA‐I are incubated for 3 days with control or macrophage‐conditioned media, the only component to be degraded in HDL appears to be apoA‐I; a 22‐kDa proteolytic fragment accumulates with time (Figure 1A). By densitometry, the level of the intact apoA‐I band in an HDL sample incubated in a RPM culture supernatant (C3) is 70% compared to a sample incubated in nonconditioned medium. Conversely, apoA‐II concentrations are not modified upon incubation.

Figure 1 MMP proteolysis on HDL containing apoA‐I (from wt blood bank donors) and HDL containing apoA‐I and apoA‐I_M (from heterozygous carriers): Time course and size of the proteolytic peptides. A: Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS‐PAGE) and Coomassie staining of HDL samples; run under nonreducing (left) and reducing conditions (right): 0 = control; 3 = incubation for 3 days in nonconditioned medium; C1, C3 = incubation for 1, or 3 days, in medium conditioned for 24 h by 1×10⁶ rat peritoneal macrophages (RPM)/mL. Band identifications are marked sidewise. (Our unpublished results.) B: SDS‐PAGE as above and immunoblotting of HDL samples containing apoA‐I/apoA‐I_M; run under nonreducing (left) and reducing conditions (right). The same nitrocellulose membranes were probed with an anti‐apoA‐I_M/apoA‐I_M (left), then with an anti‐apoA‐I (middle) and with an anti‐apoA‐II antiserum (right). (Our unpublished results.) MMP = matrix metalloprotease; HDL = high density lipoprotein; apoA‐I = apolipoprotein A‐I; wt = wild‐type; CBB = Coomassie Brilliant Blue.

In contrast, all apolipoproteins in the HDL particle containing apoA‐I_M are extensively degraded: the homodimers of both apoA‐I_M (apoA‐I_M/apoA‐I_M) and A‐II (apoA‐II/apoA‐II) as well as the apoA‐I_M/apoA‐II heterodimer, and the C peptides (Figure 1A). A comparison of densitometric data for HDL incubated in RPM culture supernatant (C3) versus HDL incubated in nonconditioned medium 3, shows that under nonreducing conditions apoA‐I_M/apoA‐II concentration decreases by 85%, whereas under these conditions the two proteins migrate together.

By immunostaining apoA‐II/apoA‐II appears to be more susceptible to proteolysis than apoA‐I_M/apoA‐II (Figure 1B). A few novel bands, all of low intensity, are detected by the antisera (arrows). One band, 22 kDa in size, is likely to correspond to the proteolytic fragment already described for HDL containing apoA‐I 15. In the apoA‐I_M sample the band may derive from apoA‐I and monomeric apoA‐I_M. No novel spots are detected even after high‐sensitivity silver staining of the pattern of MMP‐incubated HDL containing apoA‐I_M (not shown). This finding implies that the degradation of apoA‐I_M‐containing HDL does not stop at intermediate, higher M_r products, but proceeds to lower relative molecular mass (M_r) peptides.

After incubation in RPM‐conditioned medium, apoA‐I‐containing HDL particles of smaller size (<8.8 nm) are extensively degraded whereas apoA‐I‐containing HDL particles of larger size (>8.8 nm) are only slightly modified (sample C3, Figure 2A, left). In contrast, proteolysis of apoA‐I_M‐containing HDL results in drastic lipoprotein degradation (>95% in sample C3, Figure 2A, right).

Figure 2 MMP proteolysis on HDL containing apoA‐I and HDL containing apoA‐I and apoA‐I_M: Size of the modified HDL/discrimination between lipid‐bound and lipid‐free apolipoproteins. A: Gradient gel electrophoresis (GGE) run to pore‐size equilibrium. Samples of HDL containing apoA‐I (left) and of HDL containing apoA‐I and apoA‐I_M (right): 0 = control; C3 = incubation for 3 days in conditioned medium, alongside with size markers (M, leftmost lane). Protein detection by Coomassie stain. (Our unpublished results.) B: GGE run for 3 h 30 min. After electroblotting, the protein pattern was stained with Red Ponceau (left), then immunodetection was performed with anti‐apoA‐I_M/apoA‐I_M, anti‐apoA‐I and anti‐apoA‐II antisera. Samples 0 and C3 as above, from HDL containing apoA‐I and apoA‐I_M. (Our unpublished results.). MMP = matrix metalloprotease; HDL = high density lipoprotein; apoA‐I = apolipoprotein A‐I.

Degradation of apoA‐I_M‐containing HDL thus appears to be more extensive when judged from the native gradient gel electrophoresis (GGE) (Figure 2A) than from the sodium dodecyl sulfate (SDS‐PAGE) patterns (Figure 1A). This discrepancy may be explained by the results obtained after short electrophoretic runs on polyacrylamide (PAA) gradients under native conditions (nonequilibrium GGE) (Figure 2B). After 3 h 30 min, both Red Ponceau stain (left) and immunostain for apoA‐I (middle‐right) on blotted samples detect, in proteolyzed HDL, a number of components with mobility either intermediate between intact HDL and lipid‐free apoA‐I_M/apoA‐I_M (I–III) or higher than lipid‐free apoA‐I_M/apoA‐I_M (IV–VII). Most fast immunoreactive bands (I–V) are also detected by anti‐apoA‐II (right), but none by anti‐apoA‐I_M/apoA‐I_M antiserum (middle). Bands in the proteolyzed sample thus correspond to mixtures of apoA‐I, apoA‐I_M, and apoA‐II fragments, as well as to degradation products of the apoA‐I_M/apoA‐II heterodimer. In apoA‐I_M/apoA‐I_M‐derived peptides the epitope characteristic of the dimeric form, to which the monoclonal antibody is directed, appears to be degraded to below detection limits.

Proteolysis of a native protein is kinetically regulated by the differential accessibility of the various target amino acids to the enzyme. The apparent unrestrained availability to proteolytic enzymes of a larger number of target sites in homo‐ and heterodimeric apoA‐I_M forms with respect to wild‐type apoA‐I thus suggests relevant structural differences. To explain this peculiar behavior, we built a model for the heterodimeric form of apoA‐I_M22. In silico and experimental structural and dynamic properties of such a synthetic HDL, containing apoA‐I_M, apoA‐II, and L‐α‐palmitoyloleoyl phosphatidyl choline (POPC), account for its high sensitivity to MMPs. Circular dichroism data (as associated with secondary structure arrangement of the protein) and root‐mean‐square fluctuation data (RMSF), as an index of protein backbone flexibility, consistently point to a relevant reduction of alpha‐helical content in comparison with wild‐type apoA‐I, as shown by Figure 3. This feature validates the finding of a higher susceptibility to proteolytic enzymes.

Figure 3 Model structure for(apoA‐I_M/apoA‐II)₂ in a sHDL containing 112 POPC molecules. Top: before molecular dynamics (MD); bottom: after 15 ns MD. The two heterodimers are color‐coded, red or cyan; labels are near the N‐termini; side view. Protein structure is rendered (with Discovery Studio 1.5, Accelrys, San Diego, CA, US) as α‐carbon ribbon, POPC with Corey‐Pauling‐Koltun (CPK) colors (similar to Figure 2 in 21). sHDL = synthetic high density lipoprotein; POPC = L‐α‐palmitoyloleoyl phosphatidyl choline.

Apolipoprotein pharmaceuticals

ApoA‐I, involved in most of the beneficial vascular effects of HDL, appears a prime candidate for the development of apolipoprotein‐like pharmaceuticals in the treatment of cardiovascular diseases. Recombinant DNA technology is highly suitable to produce large amounts of apolipoproteins as drugs, also providing a system to create site‐specific mutants 23–28, with improved antiatherogenic potential, compared to the wild‐type proteins.

A major issue in the development of apolipoprotein pharmaceuticals is the formulation of the recombinant proteins. Human apoA‐I circulates in plasma mainly as a component of mature HDL; only a minor fraction is present as lipid‐free/‐poor protein 29. ApoA‐I should be administered as a therapeutic component of an HDL‐like plasma particle containing synthetic or natural phosphatidyl cholines (PCs). These HDL‐like liposomes mimic most properties of native HDL and appear to be the most appropriate formulation for therapeutically active apoA‐I. Once injected in vivo, they undergo extensive remodeling in plasma, due to the interaction with cells, other lipoproteins, enzymes, and lipid transfer proteins 30.

The recombinant dimeric form of apoA‐I_M/apoA‐I_M has been produced in Escherichia coli 25 and, when formulated as synthetic HDL with phospholipids, it has been successfully used in the clinic 8.

Experimental data supporting the use of apoA‐I_M in the treatment of vascular disease

Animal data

The effect of the recombinant apoA‐I_M dimer on intimal thickening after balloon injury was tested in cholesterol‐fed rabbits 31. Synthetic HDL, made of apoA‐I_M/apoA‐I_M complexed with phosphatidyl choline (PC), was administered intravenously (40 mg of protein per dose), according to a schedule of evenly spaced injections beginning 5 days before and continuing 5 days after angioplasty. The synthetic HDL produced a substantial reduction in intimal thickening compared to placebo (−71%) and PC‐only treated rabbits (−57%). Similar results were obtained when the same schedule was applied to rabbits in which neointima formation was induced by a perivascular manipulation 32. The synthetic HDL particles containing apoA‐I_M are very effective in inhibiting intimal thickening only when administered before injury, and these effects are achieved without any significant variation in plasma and aortic cholesterol content. These observations postulate a role for apoA‐I_M in the early events of lesion formation. HDL have been shown to modulate endothelial cell activation both in vitro and in vivo33–35, and this effect was indeed convincingly demonstrated for reconstituted HDL containing apoA‐I_M34.

The hypothesis that apoA‐I_M‐containing synthetic HDL would inhibit aortic atherosclerosis was tested in apoE‐deficient mice 36. Animals received 18 injections over 5 weeks of either 40 mg/kg of dimeric apoA‐I_M/phospholipids (PC) or PC alone, and lesion progression was evaluated versus a contro group. Despite no changes in cholesterolemia, apoA‐I_M/PC treatment prevented progression of atherosclerotic lesions and reduced the lipid and macrophage content of plaques. More recent studies tested the hypothesis of inducing a significant regression/stabilization of atherosclerotic plaques by a single administration of synthetic HDL. A single, very high dose of dimeric apoA‐I_M complexed with dipalmitoyl phosphatidyl choline (DPPC) (400 mg/kg of protein), given to apoE‐deficient mice 37, promoted a potent cholesterol mobilization from tissues, as demonstrated by a massive total and free cholesterol increase in plasma already 1 hour after treatment. Plasma from infused mice was also more efficient than that of controls in promoting cholesterol efflux, based on ex vivo experiments using FU5AH cells 37.

A crucial study was instead performed with the use of a different model developed in rabbits 38. This consists of a perivascular manipulation by electrical injury at the common carotids, followed by a 1.5% cholesterol diet for 90 days 38. At the end of the dietary treatment, lipid‐rich, macrophage‐rich plaques develop at the site of injury. The advantage of this model resides in the possibility of detecting atherosclerotic lesions in vivo by the application of the intravascular ultrasound (IVUS) methodology 39. Rabbits were infused with saline, DPPC, or three different doses of dimeric apoA‐I_M/DPPC in the right carotid artery, where the atherosclerotic plaque develops 40. Infusions lasted for 90 minutes, and plaque areas were evaluated before, during, and at the end of the infusion. The dramatic result was a reduction in plaque size, up to 30%, in apoA‐I_M‐treated rabbits, within the time frame of the infusion 40. A striking increase in HDL cholesterol, proportional to the infused apoA‐I_M dose, was observed, the elevation being maintained until sacrifice (72 hours from the infusion). Histological analysis confirmed the IVUS data, also noting a significant effect of plaque regression, not only at the site of infusion, but also in vascular beds distant from the administration site. The rapid plaque reduction observed by IVUS monitoring during infusion suggests a direct cholesterol removal from lipid‐rich atherosclerotic lesions 41.

This study, together with the one performed on apoE‐deficient mice, provides direct evidence for the potential therapeutic value of synthetic HDL in inducing regression and/or stabilization of atheromatous plaques. The antiatheromatous effect appears to be associated to a dramatic increase of free cholesterol in HDL, as separated by a highly sensitive high‐pressure liquid chromatography (HPLC) technique. This accumulation (over 30‐fold above controls) appears to provide a mechanistic ground to the observed reduction in arterial cholesterol. The high efficiency in promoting cholesterol efflux from cells, together with the long half‐life of dimer versus the wild‐type counterpart (40 versus 20 hours in rabbits), clearly indicates a significantly improved efficacy of the former.

Clinical data on wild‐type Apo‐I and Apo‐I_M

In the earliest clinical study with wild‐type apoA‐I (using recombinant apoA‐I and soybean PC), four patients (two with preexisting coronary disease), all with very low HDL cholesterol levels, were injected once with the recombinant product 42. Administration proved easy, only 10 minutes being necessary to administer intravenously 1.6 g of pro‐apoA‐I (pro‐apoA‐I contains more amino acids at the N‐terminus), side effects were apparently minimal or nonexisting, and no antibody could be detected. HDL‐cholesterol levels increased rapidly after administration of the apoA‐I liposomes, staying elevated for at least 3 days after injection, thus suggesting an improved tissue cholesterol removal. In a subsequent study with the same preparation, a single infusion of 4 g of pro‐apoA‐I liposomes into patients with familial hypercholesterolemia increased total fecal steroid excretion 43, again indicating enhanced cholesterol mobilization. When, however, lipid‐free apoA‐I was infused or given as bolus injection to patients with low HDL cholesterol no increase in free or esterified HDL cholesterol could be detected 44, 45. When the same authors instead infused apoA‐I/PC or pro‐apoA‐I/PC discs into healthy humans, a remarkable rise in HDL cholesterol was observed 44, 45.

More recently, reconstituted HDL (rHDL) was used also as tool to evaluate the effects of HDL/apoA‐I on vascular function. In hypercholesterolemic individuals, characterized by an impaired vasodilator response to acetylcholine (ACh), infusion of rHDL (80 mg of protein) led to a remarkable improvement of endothelial‐dependent vasodilation 46. Similar findings were described in individuals with HDL deficiency consequent to mutations of the ABCA1 transporter 47.

Initial studies on synthetic HDL containing apoA‐I_M/apoA‐I_M dimers were carried out in healthy volunteers at doses ranging from 5 to 100 mg/kg in order to assess tolerability and safety. The treatment was well tolerated and did not give rise to any sign of immunogenic responses. Most interestingly, the content of free cholesterol in HDL rose up to 3‐fold, proportional to the administered dose, and this effect lasted for 1 week. The trial on coronary patients was carried out thereafter.

The very consistent findings of preclinical studies, as well as the clinical phase I data indicative of a potent cholesterol removal activity of apoA‐I_M/apoA‐I_M synthetic HDL (sHDL) in humans, represented the basis of a pilot clinical study designed to evaluate the ability of the recombinant product to reduce coronary plaque burden in patients with acute coronary syndromes 48. Patients aged 30–75 years requiring diagnostic coronary angiography for clinical indications within 14 days after an acute coronary syndrome were eligible for intravascular ultrasound (IVUS) study. Patients were randomized to three treatment groups: placebo (0.9% saline), or a low dose (15 mg/kg), or a high dose (45 mg/kg) of apoA‐I_M liposomes. Fifty‐seven of the 123 patients screened for inclusion in the study were randomly assigned; a total of 47 completed the protocol, 11 in the placebo group, 21 in the low‐dose group, and 15 in the high‐dose apoA‐I_M group. The primary efficacy end point was the change in percent atheroma volume in the target vessels for the combined treatment groups (i.e. the 15 mg/kg plus 45 mg/kg apoA‐I_M liposomes). Compared with pretreatment, the absolute reduction in atheroma volume in the apoA‐I_M‐treated groups was −14.1 mm³, a 4.2% decrease from baseline (P<0.001) 48. This phase II study clearly demonstrates that a short‐term treatment with apoA‐I_M liposomes causes a regression of coronary atherosclerosis, compared to long‐term clinical trials in statin‐treated patients, showing that statins can stop or slow atherosclerosis progression but do not lead to regression of the lesions. The apoA‐I_M trial had several limitations, in particular a small sample size, which limited interpretation of both safety and efficacy data. However, the relationship between coronary disease progression rates and clinical events has been previously established in trials using angiographic end points 49, and it is thus likely that the rapid and large reduction in disease burden observed in the apoA‐I_M trial will improve long‐term clinical outcome.

Very recently Tardif et al. tested in a similar protocol a natural, blood‐extracted, reconstituted HDL 50 given at higher doses and only 4 versus 5 times, compared to the study with the apoA‐I_M dimer. This study, carried out in a larger number of patients with some coronary atheroma burden, showed a lesser percentage change in atheroma volume (−3.4%) versus the apoA‐I_M study and particularly lower volumetric plaque changes (−5.3 mm³). Some liver function abnormalities were noted with the higher dose of the extractive product (80 mg/kg) that led to early discontinuation of this study group. This second study with extractive HDL infusions confirms that this route of administration is potentially very effective. It also gives a clear indication that the apoA‐I_M protein is likely to be a more powerful agent for coronary atheroma reduction.

Conclusions

Many experimental data suggest that a therapeutic strategy based on HDL administration could be useful against vascular diseases. Direct evaluation of regression of atherosclerotic plaques in the clinic, i.e. by direct monitoring of the plaque by technologies such as IVUS, has generally indicated a modest effect with LDL‐lowering treatments. Studies in statin‐treated individuals undergoing repeated IVUS examinations showed that plaque volume was only reduced to a modest extent, whereas a pharmacological treatment of hyperlipidemic patients, with a drug combination significantly raising HDL levels, markedly reduced coronary plaque volume compared to no significant effects (only a stabilization) of LDL cholesterol lowering induced by statins 51. On the basis of the experimental and clinical data reviewed here, a direct administration of HDL with recombinant apoA‐I_M/apoA‐I_M could promote not only plaque stabilization by lipid and macrophage removal, but also regression of established lesions. The apoA‐I_M/apoA‐I_M treatment may further possibly represent a valuable approach for restenosis prevention.

Acknowledgements

This investigation was supported in part by grants from Università degli Studi di Milano (FIRST) and from MIUR (FIRB 2003).