Editorials
Lipoprotein(a): Revisiting a Next Therapeutic Target
2016 Volume 80 Issue 2 Pages 329-331
Details
2016 Volume 80 Issue 2 Pages 329-331
Fifty years ago, Dr Kare Berg described for the first time two main features of lipoprotein(a) or Lp(a), namely inheritance of its increased levels and the association with premature atherosclerotic cardiovascular diseases (ASCVD).1 Since then it has been regarded as an intriguing analog of low-density lipoprotein (LDL), present in plasma in much lower concentrations than LDL but lacking in proof of pathogenicity.2 However, recently published studies have unequivocally established Lp(a) as a causal and independent risk factor for ASCVD,3,4 even under maximal intensity statin treatment.5 Furthermore, a possible causal role of Lp(a) in aortic valve stenosis6 or heart failure development7 has been reported as well.
Article p 512
Lp(a) consists of an LDL particle to which is attached a glycoprotein, apolipoprotein(a) (apo(a)), covalently linked to the apolipoprotein B (apoB) moiety of LDL by a disulfide bond (Figure 1). The presence of apo(a) increases the density of Lp(a) and greatly reduces its affinity for the LDL receptor (LDLR) (Figure 2).8 This could explain why raised Lp(a) levels in plasma are unaffected by statin treatment. Apo(a) has close structural similarity with plasminogen, which endows Lp(a) with antifibrinolytic properties via its competitive inhibition of tissue-type plasminogen activator. These attributes give Lp(a) the potential for promoting both atherosclerosis and thrombosis. Apo(a) consists of a number of pleated structures, Kringles, one of which, Kringle IV type 2, is repeated a variable number of times from 2 to >40; this results in considerable interindividual variation in the size and molecular weight of apo(a) (Figure 1). Based on this molecular background, Lp(a) levels are predominantly determined genetically, and are unaltered by other risk factors or by lifestyle.2
Lipoprotein(a) consists of an low-density lipoprotein (LDL)-like particle to which apolipoprotein(a) is covalently linked. (Reprinted with permission from Oxford University Press 3772170402908.2)
Particle size and density of various lipoproteins including lipoprotein(a). HDL, high-density lipoprotein; IDL, intermediate-density lipoprotein; LDL-C, low-density lipoprotein cholesterol; VLDL, very low-density lipoprotein. (Reprinted with permission from Elsevier Limited 1982084.8)
The only evidence that lowering Lp(a) reduces risk has come from studies using lipoprotein apheresis.9 However, this costly procedure requires weekly or bi-weekly extracorporeal circulation on a long-term basis. One of the few drugs that has been shown to reduce Lp(a) levels significantly, niacin can reduce Lp(a) levels, but has failed to show benefit in outcome studies such as the AIM-HIGH trial.10 Cholesteryl-ester transfer protein inhibitor also reduces Lp(a) levels, but this class of drug has been stuck in development.10,11 In contrast, monoclonal antibodies to proprotein convertase subtilisin/kexin-9 (PCSK9), which has recently become available in the USA and Europe, and expected in Japan shortly, has been reported to lower Lp(a) by 40%.12 This could promptly raise an important clinical question as to whether Lp(a) levels in patients with gain-of-function mutation of PCSK9 (FH-PCSK9) might differ from those with impairment of the LDLR itself (FH-LDLR), both of which show genetically determined increased levels of LDL (ie, familial hypercholesterolemia [FH]). In other words, contrary to the circumstance in PCSK9 inhibition, opposite alterations in Lp(a) level would be expected in patients with a PCSK9 gain-of-function variant.
In this issue of the Journal, Tada and colleagues13 investigate this issue with molecular analyses in a large patient cohort. They report that Lp(a) levels in FH-LDLR patients were comparable to those in FH-PCSK9 patients, both of whom show significantly greater Lp(a) levels as compared with non-FH hospitalized subjects, using a propensity score-matching technique. It has already been reported that Lp(a) levels in heterozygous FH are 3-fold greater than in apo(a) phenotype-matched non-FH subjects in Western countries,14 but considerable ethnic variation in Lp(a) levels has also been described.15 Therefore, the study done by Tada et al advances our understanding of Lp(a) levels in Japanese FH and non-FH subjects.
Despite this progress, some of their observations should be interpreted with great caution. Firstly, assay-dependent variability may exist in the influence of the number of Kringle IV domains of the apo(a) molecule,16 as the authors discuss themselves, and if so, different results would be obtained if other assay methods are used. Secondly and most importantly, the “control” subjects in the present study could not be considered as true controls. The authors retrospectively collected data from subjects for whom Lp(a) values were available. This surely indicates the existence of selection bias. Even after propensity score-matching, this limitation greatly affected the scientific validity of the comparison between FH and “control” groups. This issue should be also taken into account when judging the multiple regression analysis data obtained from the “control” group. From this point of view, comparison with previous population studies done among Japanese17,18 would be required. Thirdly, variability in the molecular background of FH-LDLR in the study appears to have greatly influenced the Lp(a) values. A recent study from Europe, which enrolled subjects only with LDL receptor gene mutations (FH-LDLR), suggested interactions among the more severe LDLR mutations, increased Lp(a) levels, and ASCVD development in FH.19 This study goes one step forward from the current study, which failed to demonstrate the association between the type of molecular defect in the LDLR and Lp(a) levels. Lastly, another important issue to be noted is that current study could not provide an answer to the paradox regarding the role of LDLR in Lp(a) metabolism. As mentioned earlier, statins inhibiting intracellular cholesterol biosynthesis and thereby upregulating the LDLR do not reduce Lp(a) levels. By contrast, PCSK9 inhibition, sharing the mode of action in LDL-lowering as statins, reportedly reduces Lp(a) levels.12 It remains unclear whether the lack of difference in Lp(a) levels between FH-LDLR and FH-PCSK9 is a chance observation or truly meaningful; in other words, Lp(a) metabolism could be altered in the circumstances of PCSK9 inhibition, but not in the case in its gain of function.
Further studies, such as antisense inhibition of apo(a) gene expression,20 could provide better understandings of Lp(a) metabolism, leading to efficient and safe therapeutic interventions to reduce circulating Lp(a) levels. Future clinical and basic studies, including currently ongoing randomized controlled trials using PCSK9 antibodies, could contribute to moving forward in our strategy for ASCVD prevention.
