High-Density Lipoproteins as an Emerging Therapeutic Target for Atherosclerosis

Plasma levels of high-density lipoprotein cholesterol (HDL-C) and its major protein apolipoprotein A-I (apoA-I) are consistently inversely associated with coronary heart disease (CHD) risk in observational studies.¹ Furthermore, studies in animals over the last 2 decades have established that intravenous infusion of HDL or apoA-I or genetic overexpression of apoA-I can substantially reduce the progression and even induce regression of preexisting atherosclerosis.² Based on these data, HDL and apoA-I have become a major target for the development of new therapies for atherosclerosis.^{3 - 4}

Finding small molecules that upregulate the gene expression of apoA-I has been an elusive goal of pharmaceutical research for more than 2 decades. Despite intensive investigation, few compounds that upregulate apoA-I have been found and none have entered large-scale clinical trials, leading to increasing interest in other modalities to deliver apoA-I therapeutically. Given that administration of recombinant proteins is an established therapeutic approach for many acute and chronic disorders,⁵ administering recombinant apoA-I as an infusion is a logical approach. Indeed, chronic bolus infusion of apoA-I in rabbits has been shown to inhibit atherosclerosis⁶ and single-dose studies of apoA-I infusion in humans have been performed.⁷ However, because the concept of using normal wild-type apoA-I for therapeutic purposes is in the public domain, there is little economic incentive to develop this approach for commercialization and clinical use.

Accordingly, the story of apoA-I Milano has achieved almost mythical proportions within the field of lipoprotein metabolism. In 1980, investigators from Milan reported a family from the town of Limone sul Garda in northern Italy in which 3 members had very low HDL cholesterol levels (<15 mg/dL [0.39 mmol/L]) but had no clinical signs of atherosclerotic vascular disease.⁸ These individuals were found to have a mutant form of apoA-I, dubbed apoA-I Milano, in which a cysteine is substituted for an arginine at position 173. This cysteine confers very different properties to this protein compared with normal apoA-I, including the ability to form disulfide-bonded dimers with other apoA-I Milano molecules and with other HDL proteins such as apoA-II. The catabolism of apoA-I Milano is very rapid compared with normal apoA-I,⁹ accounting for the low levels of HDL-C and apoA-I in carriers of this mutant apoA-I.

Although initial reports suggested that carriers of the apoA-I Milano gene may be protected from CHD and prone to longevity,¹⁰ subsequent studies with appropriate controls have not supported this contention. Indeed, apoA-I Milano carriers have a degree of carotid atherosclerosis similar to matched related controls¹¹ but are clearly not at the increased risk of premature CHD that might be predicted on the basis of their very low HDL-C levels. This, however, is not unique to apoA-I Milano carriers; certain other genetic conditions characterized by very low HDL-C levels and increased catabolism of apoA-I are also not associated with increased cardiovascular risk.^{12 - 13} Nevertheless, there has been substantial interest in the potentially unique biological properties of the apoA-I Milano protein and the possibility that it could be used as a therapeutic approach for atherosclerosis. Animal studies using intravenous infusion of recombinant apoA-I Milano have demonstrated reduced atherosclerosis in mice and rabbits and reduced restenosis after angioplasty in pigs.¹⁴ These data led to the development of recombinant apoA-I Milano as a potential therapeutic approach for atherosclerosis in humans.

The study by Nissen and colleagues¹⁵ in this issue of THE JOURNAL represents the first clinical trial of the administration of apoA-I Milano, and indeed any form of apoA-I, in humans with regard to effects on atherosclerosis. The investigators recruited patients who were hospitalized with acute coronary syndromes and who required coronary angiography clinically. After a baseline angiogram that included intravascular ultrasound (IVUS) of a segment within a target vessel that did not undergo revascularization, study participants received weekly infusions of recombinant apoA-I Milano complexed with phospholipids for 5 weeks. At week 6, a repeat coronary IVUS study was performed to measure changes in the amount of atherosclerosis within the target segment. The results of this study are surprising to even the most optimistic supporters of the concept of targeting HDL as a therapy for atherosclerosis. Patients who received the 5 infusions of apoA-I Milano experienced statistically significant regression in coronary atheroma volume in the target segment compared with baseline measurements. A smaller placebo group that received only saline infusions had no significant change in atheroma volume.

There are some major limitations to the interpretation of this study. The study was not designed or powered for direct comparison between the active treatment group and the placebo group, and there was no statistically significant difference between the 2 groups in the change in atheroma volume. Furthermore, the administration of a saline placebo, rather than phospholipids alone, leaves open the possibility that the regression seen in the active treatment group was due to the phospholipid rather than the apoA-I Milano component. In addition, 2 different doses of the apoA-I Milano/phospholipid infusion were used but no evidence of a dose-response was seen. Finally, the study was rather small (45 patients receiving apoA-I Milano/phospholipids and 12 receiving placebo) and the results certainly must be replicated in a larger study. Nevertheless, the results suggest that the regression observed in the active treatment group was likely to be a direct result of the apoA-I Milano/phospholipid infusions. This study provides the best example to date that directly targeting HDL can have an impact on atherosclerosis in humans.

The mechanisms by which apoA-I Milano reduces atherosclerosis in animals and humans are not known. The most popular hypothesis to explain the relationship between apoA-I and atherosclerosis is that of reverse cholesterol transport, whereby apoA-I removes excess cholesterol from macrophages within the atherosclerotic plaque and by way of HDL returns it to the liver for excretion in the bile.¹⁶ Indeed, overexpression of normal apoA-I in mice has been shown to promote the rate of reverse cholesterol transport from macrophages to liver to feces in vivo.¹⁷ Measuring rates of reverse cholesterol transport in humans is difficult, although a single intravenous infusion of normal apoA-I was shown to increase fecal sterol excretion as a potential marker of increased reverse cholesterol transport.¹⁸ Studies measuring reverse cholesterol transport rates should be performed in animals and humans after administration of apoA-I Milano and in studies of new therapies that target HDL metabolism.

In addition, HDL and apoA-I have been proposed to have other properties that could contribute to their antiatherogenic effect, including antioxidant, antiinflammatory, nitric oxide–promoting, prostacyclin-stabilizing, and platelet-inhibiting properties.^{1 - 4} This issue is important because if these properties contribute substantially to protection from atherosclerosis by HDL, in theory a treatment that increases HDL-C and apoA-I levels could be beneficial even if it does not increase the rate of reverse cholesterol transport. Determining the cellular and molecular mechanisms by which 5 weekly infusions of apoA-I Milano induced quantitative regression of coronary atherosclerosis in humans will be of great scientific and clinical importance; such insights could very well lead to new therapies for inducing regression of atherosclerosis.

Whether apoA-I Milano has biological properties that make it even more effective than normal wild-type apoA-I in preventing or regressing atherosclerosis is hotly debated. Despite numerous animal studies involving administration or genetic expression of either normal apoA-I or apoA-I Milano, no published article has directly compared these 2 forms of apoA-I in animals in a head-to-head fashion with regard to their effects on atherosclerosis. Therefore, from a scientific standpoint it remains an unanswered question as to whether apoA-I Milano has unique properties that result in greater antiatherogenic potential than normal apoA-I. Resolving this issue would provide insight into the structure-function properties of apoA-I that affect atherosclerosis. Importantly, it is possible that if the same study by Nissen et al had been performed with normal apoA-I instead of apoA-I Milano, a similar (or even better) result might have been obtained.

One of the critical issues that the report by Nissen et al raises is whether change in coronary atheroma burden as measured by IVUS is a reliable surrogate for clinical benefit. Nonobstructive coronary atherosclerosis, as assessed indirectly through angiography, is a predictor of future coronary events and changes in luminal diameter after medical intervention also are related to future events.¹⁹ Although it is tempting to assume that changes in coronary atheroma volume by IVUS would be an effective intermediate end point for predicting ultimate clinical benefit of a new therapeutic intervention for atherosclerosis, this has not been established. As the interest in the use of IVUS for assessing the therapeutic efficacy of new antiatherogenic therapies increases, it will be critical to establish whether short-term changes in coronary atheroma detected by IVUS actually do predict long-term clinical benefit.

Historically, atherosclerosis has been considered a chronic, slowly progressive disease for which the treatments are largely preventive in nature and exert their effects over long periods. Over the last decade, investigators have found that coronary atherosclerosis may be much more dynamic in nature; many acute coronary events are caused by the rupture of an inflammatory unstable plaque.²⁰ Indeed, aggressive cholesterol-lowering therapy with statins appears to have clinical benefit within 4 to 6 months of initiation of therapy,²¹ suggesting rather rapid effects on the stability of preexisting coronary lesions.

The rapid regression of atherosclerosis in animals induced by HDL infusion²² or apoA-I expression²³ led to the concept of HDL-based therapies for patients with acute coronary syndromes. The study by Nissen et al demonstrates that this concept in fact may be reasonable. If this concept is confirmed in future studies, some day patients with acute coronary syndromes may receive "acute induction therapy" with HDL-based therapies for rapid regression and stabilization of lesions, followed by long-term therapy to prevent the regrowth of these lesions. In this model, long-term HDL-based therapies will still be needed as a vital component of the preventive phase. Importantly, despite the potential of newer quantitative measures of atherosclerosis burden, such as IVUS and magnetic resonance imaging, to help assess efficacy of new therapies, ultimately all HDL-based therapies must be proven to reduce cardiovascular events.

The apoA-I Milano story is remarkable, beginning with the discovery of rare individuals in a small Italian town and leading to the development of a potential new therapy for atherosclerosis. In fact, rare genetic conditions have provided major insight into the physiology and regulation of HDL metabolism and have spurred development of other new therapies targeted toward HDL, such as inhibition of the cholesteryl ester transfer protein²⁴ or up-regulation of the ABCA1 transporter.²⁵ If the pace of these discoveries continues, the next 2 decades may be to HDL what the last 2 decades were to LDL: an era in which the development of new therapies may permit the unequivocal demonstration of the clinical benefit of targeting HDL to reduce the burden of atherosclerotic cardiovascular disease.