Introduction

Autosomal dominant hypercholesterolemia (ADH) is a heterogeneous genetic disorder characterized by a selective increase of LDL cholesterol (LDL-C) levels in plasma, giving rise to tendon and skin xanthomas, arcus cornea, and vascular deposits, leading to progressive and premature atherosclerosis, coronary heart disease (CHD), and death. The first two genes implicated in the disease are the gene that encodes the LDL receptor (LDLR at 19p13.3; OMIM 606945, 143890) [1] and the apolipoprotein B (apoB) gene (APOB at 2p23–p24; OMIM 107730, 144010), encoding the ligand of the LDL receptor [2]. The existence of a greater level of genetic heterogeneity in ADH and the involvement of a third locus named HCHOLA3 (formerly FH3; OMIM 603776) were reported by our team. In 2003, we discovered [3] that PCSK9 was the third gene implicated in ADH. This pioneering work revealed a new major player in cholesterol homeostasis and was the first step of the adventure involving proprotein convertase subtilisin/kexin type 9 (PCSK9) as a promising therapeutic target in lowering LDL-C levels and reducing the risk of cardiovascular diseases. In this review we will follow the PCSK9 adventure from the involvement of its mutations and variants in cholesterol disease and CHD to the several clinical trials that have been launched.

Discovery of the Involvement of PCSK9 in Cholesterol Metabolism

Through the French Research Network for ADH (Réseau National de Recherche sur les Hypercholestérolémies Familiales), families with hypercholesterolemia were recruited from several regions of France [3]. After the exclusion of the LDLR and APOB genes, a positional cloning strategy was used to identify the genetic region linked to the disease. Using this classic genetic approach, HCHOLA3 was mapped to 1p34.1–p32 in a French multiplex family (HC2) [4]. A year later Hunt et al. [5] confirmed this localization in an ADH family originating from Utah. Segregation analysis, genetic mapping, and sequencing studies performed helped in excluding several genes, and in refining the boundaries of the region through the identification, by Abifadel et al., of a new French multiplex family, HC92, linked to the same HCHOLA3 locus. Extensive sequencing studies of several candidate genes expressed in the liver allowed us the detection, on the 13 September 2002 in the HC2 and HC92 families, of a common mutation, p.S127R, in the PCSK9 gene and another mutation, p.F216L, in a third French family with ADH [3]. The PCSK9 gene, the ninth member of the proprotein convertase subfamily, had been characterized in 2003 by Seidah et al. [6], who identified it from a patented database in a BLAST search to find proteins related to a recently identified proprotein convertase called SKI-1 (site-1 protease). PCSK9 was formerly designated as neural apoptosis regulated convertase 1 (NARC1) as it was discovered in 2001 by Millenium Pharmaceuticals through the cloning of complementary DNA upregulated after apoptosis induced by serum deprivation in primary cerebellar neurons. It was also designated as LP251, which was identified by Eli Lilly and Co. in 2002 via the cloning of secretory proteins [6]. The mammalian serine proprotein convertase family is responsible for the proteolytic maturation of secretory proteins, including neuropeptides, prohormones, cytokines, growth factors, receptors, serum, and cell surface proteins [6, 7].

PCSK9 Protein: Structure and Function

PCSK9 complementary DNA (NM_174936.2) spans 3,617 bp over 12 exons that encode the 692 amino acid protein PCSK9 (NP_777596.2). It is particularly expressed in the liver, gut, kidney, and nervous system [6, 8]. The detailed structure and processing of PCSK9 are given in Fig. 1 [6, 9–19]. The 60-kDa mature form and the furin-cleaved form of PCSK9 are present in the circulation [12, 17].

Fig. 1
figure 1

Structure, processing of proprotein convertase subtilisin/kexin type 9 (PCSK9), and impact of PCSK9 main variants and mutations. The PCSK9 structure is characterized by a signal sequence (amino acids 1–30), a prodomain (amino acids 31–152), and a catalytic domain, followed by a 243 amino acid cysteine-rich and histidine-rich C-terminal region. PCSK9 is synthesized as an inactive proenzyme and contains a triad of residues (Asp-186, His-226, and Ser-386) that are required for catalytic activity. The approximately 74-kDa precursor form of PCSK9 undergoes intramolecular autocatalytic cleavage in the endoplasmic reticulum (ER), which produces an approximately 60-kDa catalytic fragment. Autocatalytic cleavage of the zymogen in the ER is essential for transport from this compartment and for secretion. The PCSK9 crystal structure shows that the cleaved prodomain of approximately 14 kDa remains associated with the catalytic domain, blocking the PCSK9 active site, which could explain why no other proteolytic activity has been reported for PCSK9. The 60-kDa mature and secreted form is cleaved at the motif RFHR↓218 into an approximately 53-kDa inactivated or less efficient fragment by other proprotein convertases, particularly furin and/or proprotein convertase C5/6A (PC5/6A). PCSK9 degrades LDL receptor (LDLR) independently of its catalytic activity by involving mainly extracellular and possibly intracellular pathways. PCSK9 might work in a post-ER compartment, where it might target LDLR for degradation in lysosomes. The binding site for the LDLR EGF-A domain resides on the surface of PCSK9 that is formed primarily by residues 367–381. Key interactions with EGF-A are made by Arg-194 and Asp-238 of PCSK9. Several gain-of-function mutations are reported: The p.S127R variant interferes with autocatalytic cleavage, which is crucial for secretion from the cell. The p.D374Y variant binds LDLR 25 times more tightly than does wild-type PCSK9 at neutral pH, remains in a high-affinity complex at acidic pH, and is approximately tenfold more active in reducing LDLR levels than the wild-type protein. The p.R218S, p.F216L, and p.D374Y mutations result in total (p.R218S) or partial loss of the furin/PC5/6A processing of PCSK9, which increases the stability of PCSK9. Loss-of-function mutations are also represented: no protein was detected with the p.Y142X mutation, probably owing to nonsense-mediated messenger RNA decay. Some mutants associated with hypocholesterolemia either remain in the ER (p.C679X and the p.G106R mutations) or do not sort to endosomes (p.L253F and p.Q554E), resulting in loss of function (Benjannet et al. [9, 12], Lagace et al. [17], Cunningham et al. [10], McNutt et al. [11], Piper et al. [15], Nassoury et al. [19], Zhang et al. [14], Kwon et al. [16], Poirier et al. [13])

Full size image

PCSK9 Mutation in Hypercholesterolemia

The p.S127R mutation in a highly conserved region between species in exon 2 was found in the first two French families studied: HC2 and HC92. The second mutation, p.F216L, in a conserved region in exon 4, was identified in a French family in which the proband died from myocardial infarction at the age of 49 years with a total cholesterol level of 441 mg/dl and an LDL-C level of 356 mg/dl [3, 20, 21]. These two mutations allowed us to identify for the first time the involvement of PCSK9 in ADH and cholesterol metabolism [3]. The third mutation, p.D374Y, was reported in 2004 in the hypercholesterolemic Utah kindred [22] previously linked to the 1p32 region [5]. The same mutation was found in three Norwegian families [23] and in three English families, with 12 affected patients having severe hypercholesterolemia and a family history of premature CHD [24].

Other mutations adjacent to these mutations were also reported: p.D374H in Portuguese patients with severe hypercholesterolemia [25]; p.R218S, which we identified in a French family whose proband at the age of 45 years had an LDL-C level of 293 mg/dl and presented with tendinous xanthoma and arcus corneae [26]; p.R215H in two families from Norway [27]; and p.D129G in a family originating from New Zealand [28]. A novel missense mutation of the PCSK9 gene, p.R306S, was found in a Chinese population [29]. More recently, we identified two gain-of-function mutations of PCSK9 in French families: (1) p.L108R, in a black family originating from Mauritius whose proband at the age of 41 years had an LDL-C level of 302 mg/dl and tendon xanthomas; (2) p.D35Y in a family’s proband who had an LDL-C level of 234 mg/dl at the age of 55 years [30]. The PCSK9 mutations inducing ADH are very rare, but well documented (familial segregation analysis, in vitro mutagenesis, etc.). The clinical findings that have been reported in PCSK9 heterozygote carriers are those related to hypercholesterolemia: tendon xanthomas, CHD, premature myocardial infarction, and stroke. Most enzymopathies are recessively inherited, and thus the dominance of the ADH trait associated with PCSK9 was in favor of a gain-of-function mechanism [3]. This was confirmed by cellular and animal models showing that these gain-of-function mutations decreased the number of LDL receptors at the cell surface, leading to hypercholesterolemia [17, 31, 32].

In vitro studies showed that the two gain-of-function mutations p.S127R and p.D374Y resulted in a 23 % decreased level of cell surface LDL receptors and a 38 % decreased level of internalization of LDL compared with wild-type PCSK9 [33]. It was shown more recently that the p.L108R mutant exhibited a marked approximately twofold enhanced degrading activity towards LDL receptor, resulting in a clear and significant gain-of-function in this assay [30]. The mechanisms of action of the gain-of-function mutations are depicted in Fig. 1.

PCSK9 and Hypocholesterolemia

Two years after our first report of the involvement of PCSK9 in cholesterol metabolism and disease, two nonsense mutations in PCSK9, p.Y142X and p.C679X, were identified in subjects with low plasma levels of LDL-C (below 58 mg/dl) from the Dallas Heart Study, a multiethnic population of Dallas County, Texas, USA [34]. Subjects with nonsense mutations had significantly lower plasma levels of total cholesterol and LDL-C, but not all of them were hypocholesterolemic [34]. In the USA, one in every 50 African Americans has a nonsense mutation in PCSK9. In the Atherosclerosis Risk in Communities (ARIC) study, comprising 3,363 black and 9,523 white participants aged 45–64 years from four American communities [35], the nonsense mutations occurred in 2.6 % of the black subjects examined and were associated with a 28 % reduction in mean LDL-C level and an 88 % reduction in the risk of CHD. These mutations were found at this same high frequency in a Nigerian population [36], in 3.7 % of African women from Zimbabwe and associated with a 27 % reduction in LDL-C levels [37], but were very rare in Americans of European origin (less than 0.1 %) [36]. However, another variant, p.R46L, was found in 3.2 % of the white subjects examined in the ARIC study and was associated with a 15 % reduction in LDL-C levels and a 47 % reduction in the risk of CHD [34, 35, 38]. The p.Q152H mutation of PCSK9 was identified in a French Canadian, with mean decreases in circulating PCSK9 and LDL-C concentrations of 79 % and 48 %, respectively, compared with unrelated noncarriers [39]. The p.G106R mutation segregated with low LDL-C levels in a Norwegian family [18]. The impacts of these variants on CHD have been studied and are reported in Fig. 1.

A woman originating from Zimbabwe, homozygous for p.C679X, was reported [37] with a very low LDL-C level (15 mg/dl). Furthermore, Zhao et al. [40] reported a compound heterozygote for the p.Y142X mutation and an in-frame 3-bp deletion (c.290_292delGCC) that deletes an arginine at codon 97. She had no immunodetectable circulating PCSK9. This 32-year-old African American woman with an LDL-C level of only 14 mg/dl was apparently healthy, fertile, and normotensive, with grossly normal hepatic, neuronal, and renal function test results [40]. A 49-year-old Caucasian man with a heterozygous double PCSK9 mutation, undetectable circulating PCSK9, and profound familial hypobetalipoproteinemia (FHBL) (LDL-C level 16 mg/dl) was also reported. A monoallelic PCSK9 double-mutant R104C/V114A cosegregated with FHBL, with no mutation found at other FHBL-causing loci [41]. Two nonsense mutants, p.A68fsL82X and p.W428X, have been identified in Sicilian and Japanese hypocholesterolemic patients [42, 43], respectively. One proband heterozygous for a novel single nucleotide deletion in exon 1 (c.202delG), which causes a frameshift in messenger RNA (mRNA), leading to a premature stop codon (A68fsL82X), was a 34-year-old white overweight male (body mass index 30 kg/m2) who had been referred to the clinic for fatty liver. This loss-of-function mutation was also identified in two healthy blood donors who had no clinical or laboratory signs of liver disease; the results of other routine laboratory tests were normal [42]. In the Dallas Heart Study, no significant difference in the median content of hepatic triglycerides or in the prevalence of hepatic steatosis between the subjects with and without an LDL-lowering mutation in PCSK9 was observed in either ethnic group [36]. Hypocholesterolemia due to a deficiency of PCSK9 appears to be benign, in contrast to other Mendelian forms of severe hypocholesterolemia such as abetalipoproteinemia (OMIM 200100) and homozygous hypobetalipoproteinemia (OMIM 107730), which are both associated with malnutrition, hepatic steatosis, steatorrhea, and manifestation of fat-soluble vitamin deficiency [40].

PCSK9 in CHD and Large Population Studies

PCSK9 variants have variable frequencies in different populations, and their impact on cholesterol levels and CHD was analyzed in African [37], American [35], and European [18, 44] populations and in different studies (ARIC [35], PROSPER [45], LCAS [46], TEXGEN [46], PLIC [47]) by evaluating either the protection of the loss-of-functions variants or the severity of coronary atherosclerosis associated with gain-of-functions polymorphisms (mainly p.E670G). These studies, their objectives, their results, and their conclusions are summarized in Table 1 [35, 45–59]. They showed that genotype is a better predictor of lifelong exposure to LDL-C than LDL-C measured in adult life. But the impact on LDL may not be the only effect of PCSK9 on atherogenesis [60]. It is noteworthy that several genome-wide association studies identified an association of the PCSK9 locus and of some PCSK9 variants with the variability of LDL-C levels or early-onset myocardial infarction [61].

Table 1 Major studies of the impact of PCSK9 variants in different populations and diseases, specifically coronary heart diseases (CHD)
Full size table

Genotype–Phenotype Correlation

PCSK9 polymorphisms account for cholesterol variability not only in normolipemic subjects but also among familial hypercholesterolemia (FH) patients sharing the same mutation of LDLR [62]. We showed that PCSK9 might constitute a modifier gene in FH: in Lebanese FH patients sharing the LDLR p.C681X mutation, p.Leu21dup, in exon 1 of PCSK9, known to be associated with lower LDL-C levels in general populations, is also associated with a reduction of LDL-C levels in FH [62]. Furthermore, additive effects of mutations of LDLR and gain-of-function mutations of PCSK9 on the phenotype of FH have been reported in several studies [26, 63] and might be associated with a severe phenotype. It is noteworthy that the p.R496Q variant in PCSK9 was identified [33] in a subject homozygous for apolipoprotein E2 who presented with type III hyperlipoproteinemia.

We identified PCSK9 p.L21tri (p.L15_L16ins2L) mutation in two French-Canadian families with familial combined hypercholesterolemia (FCHL) and in one French-Canadian woman and her father with hypercholesterolemia [64]. Our report of the involvement of the L11 variant of PCSK9 in FCHL was the first report of the involvement of PCSK9 in this disease. This was confirmed by Brouwers et al. [65], who showed that PCSK9 levels were higher in FCHL patients than in normolipidemic relatives and spouses. They also reported that PCSK9 levels were related to markers of cholesterol synthesis in FCHL [66].

PCSK9 and ApoB

In vivo kinetics of apoB100-containing lipoproteins studied in two subjects carrying the p.S127R mutation in PCSK9 showed that PCSK9 mutation increased the production rate of apoB100 by threefold compared with controls or LDLR-mutated subjects, which is related to direct overproduction of VLDL (threefold), intermediate-density lipoprotein (threefold), and LDL (fivefold) [67]. Expression of the PCSK9 p.D374Y variant increases secretion of apoB100-containing lipoproteins from the cells by twofold to fourfold probably by reducing the degradation of nascent protein [24]. This also suggests that the variants of PCSK9 found in FH influence the secretion of apoB-containing lipoproteins. The same team produced transgenic mice expressing the p.D374Y variant of the human PCSK9 gene at physiological levels and showed that the phenotype closely matched that found in heterozygous p.D374Y patients and that reduced LDL receptor activity is not the sole cause of their hypercholesterolemia. The p.D374Y mice secreted more triglyceride-rich lipoproteins into the circulation than did wild-type mice [68]. Recently Sun et al. [69] studied the impact of PCSK9 overexpression (approximately 400-fold above the baseline) on apoB synthesis and secretion in mouse models. They demonstrated that endogenous PCSK9 interacted with apoB in hepatocytes. The physical interaction of PCSK9 with apoB acts to shunt apoB away from autophagosomes and degradation. In turn, most of the apoB would be destined for assembly and secretion as VLDL from hepatocytes. This observation is consistent with increased apoB production on overexpression of PCSK9. They thus proposed a new role for PCSK9 that involves shuttling between apoB and LDL receptor.

PCSK9 Expression

PCSK9 expression seems regulated by nutritional and hormonal status. PCSK9 is upregulated and increased by overexpression of sterol responsive element binding protein 2 (SREBP-2), cholesterol depletion [70], inflammation, administration of insulin, and statin therapy [71]. PCSK9 is downregulated by the suppression of SREBP-2, cholesterol feeding, and berberine but also by glucagon [72], ethinylestradiol [72], chenodeoxycholic acid, and farnesoid X receptor agonist [73]. It is now established that several antihyperlipidemic drugs such as statins, fibrates, and ezitimibe induce an increase of PCSK9 levels. This might attenuate their cholesterol-lowering effect by reducing LDL receptor abundance at the cell surface. In 2004 Dubuc et al. [71] showed for the first time that the expression of PCSK9 mRNA was strongly induced by statins in a dose-dependent manner and that human, mouse, and rat PCSK9 promoters contain two typical conserved motifs for cholesterol regulation: a sterol regulatory element and an Sp1 site. Cellular and animal studies by several teams showed that statins increase SREBP-2 levels and lead to an increase of LDL receptor levels but also of the levels of PCSK9, which decreases the abundance of LDL receptor on the cell surface, limiting the hypocholesterolemic action of statins. Several studies in humans showed that different statin (atorvastatin, simvastatin, rosuvastatin, etc.) treatments caused an increase in serum PCSK9 levels. The increase of PCSK9 levels caused by atorvastatin was 47 % for 80 mg versus 14 % for 10 mg. These data suggest that the explanation for why increasing doses of statins fail to achieve proportional LDL-C lowering is that statins increase PCSK9 levels in a dose-dependent fashion, and that the increased PCSK9 levels largely negate further statin-induced increases in hepatic LDL receptor levels [74]. Thus, it was suggested that a combination of a statin with a PCSK9 inhibitor could overcome this effect and enhance reduction of cholesterol levels. An initial proof-of-concept was provided by statin administration to Pcsk9 −/− mice that produced an exaggerated increase in LDL receptors levels in liver and enhanced LDL clearance from plasma [75]. This has been confirmed in nonhuman primate models and humans.

Furthermore, when added to statin therapy, ezetimibe leads to a further increase of PCSK9 levels (77 % vs 45 % with statins alone) [76, 77]. Several studies have investigated the impact of fibrates on the circulating levels of PCSK9, but the results are conflicting [78–80]. This might be due to the use of different analytical techniques to measure circulating PCSK9 levels. However, there is more evidence currently that fibrates increases serum PCSK9 levels and that these increases are highly correlated with fenofibrate-induced changes in LDL-C levels [81].

PCSK9 Levels in Blood

PCSK9 is present in human plasma, but the factors that contribute to differences in plasma concentrations are not very well known. Several teams have developed an enzyme-linked immunosorbent assay (ELISA) to measure PCSK9 in plasma. Plasma levels of PCSK9 vary at least 100-fold [82]. Serum PCSK9 levels measured by ELISA seem to be directly correlated with serum LDL-C and total cholesterol levels [83]. In hypercholesterolemic subjects, PCSK9 levels were higher than in control subjects, and increased in proportion to the dose of statin, combined or not combined with ezetimibe [71]. Plasma PCSK9 levels are positively associated with LDL-C levels in FH patients, and might contribute to the phenotypic severity in this disorder [84]. Serum PCSK9 levels display a diurnal rhythm that closely parallels that of cholesterol synthesis [85]. PCSK9 concentrations were lower with a polyunsaturated fatty acid diet [86], a Mediterranean diet [87], administration of estrogens [88], and administration of growth hormone [88]. The PCSK9 level was found to be associated with the γ-glutamyl transferase level in diabetic patients [89] and with carotid intima–media thickness in hypertensive patients [90]. The plasma level of lipoprotein-associated phospholipase A2 is inversely correlated with PCSK9 levels [91]. The plasma level of PCSK9 was increased at the baseline in proteinuric subjects, predicted lipoprotein responses to proteinuria reduction, but remained unchanged after proteinuria reduction [92]. At physiological levels observed in human obesity, it was shown that resistin increases cellular expression of PCSK9, which enhances intracellular LDL receptor lysosomal degradation [93]. Nevertheless, no positive association of plasma PCSK9 with resistin was found in lean and moderately obese individuals [94].

Therapeutic Strategies to Reduce PCSK9 Levels or Inhibit PCSK9

Several strategies to inhibit PCSK9 or lower PCSK9 levels have been investigated. Specific inhibition of PCSK9 via a classic pharmaceutical approach such as orally active molecules targeting PCSK9 seems difficult. Strategies known to target proteins not accessible to small molecules have been tested. Gene silencing by RNA interference and specific antibodies or competing peptides targeting PCSK9 have been developed. The details of these molecules or antibodies, and the results obtained in cellular models or animal models (mice or monkeys) and the related patents were reviewed in a previous article [95]. Clinical studies have been launched by several pharmaceutical companies. The details of these studies, their results, and the adverse reactions are given in Table 2 [96, 97•, 98–108, 109••, 110, 111, 112••, 113–115]. The first strategies based on gene silencing that targets PCSK9 intracellular and extracellular functions consisted in a subcutaneous administration of antisense oligonucleotide (ASO) targeting PCSK9 or small interfering RNA (siRNA). ASO studies have been conducted mainly with a second-generation ASO produced by Isis Pharmaceuticals, or with a 13-mer locked nucleic acid (LNA) ASO or a 14-mer LNA-ASO specific for a human PCSK9 sequence from Santaris Pharma. They showed in cellular, mouse, and monkey models a significant reduction of hepatic Pcsk9 mRNA expression and of total cholesterol and LDL-C levels. These ASOs were well tolerated in animals. The most frequent adverse event with this approach was injection-site erythema that seems to resolve spontaneously. To determine whether injection of these compounds results in toxic effects in humans, a clinical trial has been launched by Bristol-Myers Squibb using BMS-844421 (BMS-PCSK9Rx), which is an ASO developed by Isis. Nevertheless, the clinical study has been discontinued and no data are available. The clinical trial launched by Santaris Pharma to assess the safety, tolerability, pharmacokinetics, and pharmacodynamics of SPC5001 (a 14-mer LNA-ASO specific for a human PCSK9) has also been discontinued and no data are available either (Table 2). PCSK9 gene silencing in mice and monkeys has also been achieved using siRNA. Active, cross-species siRNAs capable of targeting murine, rat, nonhuman primate, and human PCSK9 have been developed by Frank-Kamenetsky and coworkers [97•, 116]. Delivery of the PCSK9 siRNA to the liver was facilitated by a lipidoid nanoparticle, minimizing toxicity. A phase I clinical trial was conducted by Alnylam Pharmaceuticals to determine the safety, tolerability, pharmacokinetics, and pharmacodynamics of a single dose of ALN-PCS02. The results are given in Table 2.

Table 2 Major clinical trials targeting PCSK9: details of the published studies using RNA interference or anti-PCSK9 monoclonal antibodies to lower PCSK9 levels or inhibit PCSK9, the results, and adverse reactions
Full size table

Other molecules that are currently being studied are adnectins (BMS-962476) [96], which are in phase I trials (Table 2), and small molecule inhibitors (SX-PCSK9, detailed at http://www.serometrix.com/pipeline.html).

Several antibodies or competing peptides targeting PCSK9 have been developed and studied in cellular and animal models (mice and monkeys). Clinical studies are being performed by pharmaceutical companies: LGT209 by Novartis is in a phase II study, LY3015014 by Eli Lilly is in a phase II study, RG7652 (MPSK3169A) [98] by Genentech (Roche) is in a phase II study, and RN316 (bococizumab) by Pfizer has undergone phase I studies and is now in phase II [99] and phase III studies. Available published results of these studies are reported in Table 2.

Many phase I and phase II studies have been published recently in several interesting articles for two antibodies targeting and inhibiting PCSK9 interaction with LDL receptor: AMG 145 (evolocumab) developed by Amgen (Thousand Oaks, CA, USA), and SAR236553/REGN727 (alirocumab) developed by Regeneron Pharmaceuticals (Tarrytown, NY, USA) and Sanofi-Aventis (Paris, France). These antibodies and related patents were given in our previously published review on PCSK9 patents [95], but the details of the clinical trials, the doses given every 2 or 4 weeks subcutaneously, the results, and the adverse events are given in Table 2 [100–108, 109••, 110, 111, 112••, 113–115].

Phase III studies have been initiated by Amgen and Sanofi and Regeneron. The results of two of these phase III studies with evolocumab have been published and are detailed in Table 2, and several other phase III studies have been launched but have not been published yet. For alirocumab, an important program (ODYSSEY) concerning a large number of patients in short-term or long-term trials and targeting several populations has also been initiated. The design of these studies is summarized in Table 2 as well. Long-term studies that will involve 20,000 patients for both evolocumab and alirocumab will provide results regarding the long-term efficacy, safety, and tolerability of these anti-PCSK9 antibodies that are eagerly awaited.

Other PCSK9 Interactions and Studies in Other Diseases

PCSK9 interactions and the possibility of the involvement of PCSK9 in several diseases such as liver diseases, obesity, Alzheimer disease, cognitive performance [58] and cancer [59] have been studied (Table 1). Jonas et al. [117] showed that overexpression of PCSK9 in cells decreased cellular levels of BACE1, a membrane protease responsible for the production of toxic β-amyloid peptides that accumulate in neuritic plaques of Alzheimer disease brains. However, Liu et al. [118] found that PCSK9 does not have a role in regulating LDL receptor family members or BACE1 protein levels in the adult mouse brain and that the development of PCSK9 therapies for CHD is probably not to be hampered by potential CNS adverse effects. Devay et al. [119] discovered recently that PCSK9 interacts via its C-terminal domain directly and in a pH-dependent manner with amyloid precursor protein as well as amyloid-precursor-protein-like protein 2. It is notable that no genetic association was found between PCSK9 polymorphisms and Alzheimer disease and plasma cholesterol level in Japanese patients studied by Shibata et al. [120]. PCSK9 reduces the protein levels of LDL receptor in mouse brain during development and after ischemic stroke [121]. In vivo, endogenous PCSK9 regulates VLDL receptor protein and triglyceride accumulation in visceral adipose tissue. In a clinical perspective, because Pcsk9 −/− mice do not develop liver steatosis and are not prone to obesity, the administration of a PCSK9 inhibitor developed for hypercholesterolemia treatment should not result in adverse effects [122]. A potential role of PCSK9 in the pancreas is also controversial. PCSK9 deficiency reduces liver metastasis by its ability to lower cholesterol levels and by possibly enhancing TNFα-mediated apoptosis [123]. Furthermore studies in Xenopus oocytes and in epithelia showed that PCSK9 noncatalytically reduced the abundance of the epithelial Na+ channel, a major contributor to blood pressure control [124]. PCSK9 interacts with annexin A2 [125]. Possible other unknown functions of PCSK9 and unidentified binding partners could exist; thus, it is important for the safety of new cholesterol-lowering therapy to target specifically PCSK9 action on the LDL receptor. An antiviral effect of circulating liver PCSK9 on hepatitis C virus in cells has recently been shown, and PCSK9 downregulates in vitro the level of expression of mouse liver CD81, a major hepatitis C virus receptor [126]. Conditional knockout mice lacking PCSK9 in hepatocytes have impaired liver regeneration after a partial hepatectomy, suggesting that on hepatic damage, patients lacking PCSK9 could be at risk [127].Thus, liver problems, hepatitis, or muscle problems are taken into consideration before inclusion or exclusion and are closely monitored during clinical trials. In clinical trials, anti-PCSK9 antibodies seem well tolerated, with no clinically significant safety findings in phase I and phase II/III studies, the most commonly reported adverse events being nasopharyngitis, injection-site pain, headache, skin burning sensation, upper respiratory tract infection, influenza, and back pain [100–108, 109••, 110, 111, 112••, 113–115]. Longer-term studies will provide the highly awaited long-term efficacy, safety, and tolerability of these anti-PCSK9 antibodies.

Conclusions

Reduction of PCSK9 levels or inhibition of PCSK9 is especially interesting in patients with hypercholesterolemia or an atherogenic lipid profile who fail to reach their individual cholesterol goal from classic lipid-lowering treatment, patients at high risk of developing side effects from statins, poor responders to statin therapy alone, and patients with severe hypercholesterolemia, particularly some carriers of a mutation of the LDLR, APOB, or PCSK9 gene. The tremendous commitment from all the centers of the French Research Network for Hypercholesterolemia that helped us in recruiting French patients and the enormous amount of genetic and molecular work we performed were very important in our pioneering step linking PCSK9 to LDL-C metabolism and paving the way for the work of several other teams. Finally, the PCSK9 story is a wonderful example of how collaboration between teams (Boileau’s and Seidah’s teams) conducting research in completely different fields can be initiated and prove to be highly successful. It is also a fine example of the power of genetic research strategies in revealing new therapeutic targets.

The results of the phase III studies using the anti-PCSK9 antibodies with or without statins or other hypocholesterolemic drugs are highly awaited, with the hope that this new class of blockbuster candidates will keep its promises in helping lowering cholesterol levels and fighting against cardiovascular disease.