Targeting apoC-III and ANGPTL3 in the treatment of hypertriglyceridemia

ABSTRACT

Introduction

The prevalence of hypertriglyceridemia (HTG) is increasing. Elevated triglyceride (TG) levels are associated with an increased cardiovascular disease (CVD) risk. Moreover, severe HTG results in an elevated risk of pancreatitis, especially in severe HTG with an up to 350-fold increased risk. Both problems emphasize the clinical need for effective TG lowering.

Areas covered

The purpose of this review is to discuss the currently available therapies and to elaborate the most promising novel therapeutics for TG lowering.

Expert opinion

Conventional lipid lowering strategies do not efficiently lower plasma TG levels, leaving a residual CVD and pancreatitis risk. Both apolipoprotein C-III (apoC-III) and angiopoietin-like 3 (ANGPTL3) are important regulators in TG-rich lipoprotein (TRL) metabolism. Several novel agents targeting these linchpins have ended phase II/III trials. Volanesorsen targeting apoC-III has shown reductions in plasma TG levels up to 90%. Multiple ANGPLT3 inhibitors (evinacumab, IONIS-ANGPTL3-L_Rx, ARO-ANG3) effectuate TG reductions up to 70% with concomitant potent reduction in all other apoB containing lipoprotein fractions. We expect these therapeutics to become players in the treatment for (especially) severe HTG in the near future.

KEYWORDS:

• ANGPTL3
• apoC-III
• CVD risk
• evinacumab
• hypertriglyceridemia
• pancreatitis
• triglycerides

1. Introduction

Hypertriglyceridemia (HTG) is a highly frequent condition in developed countries [1]. The burden of HTG is rapidly increasing due to the growing obese population and associated atherogenic lipid profile with high plasma triglyceride (TG) levels [2,3]. HTG is associated with an increased risk of cardiovascular disease (CVD) and all-cause mortality in large population studies [4,5]. Mendelian randomization studies show evidence for a causal relationship with regard to increased risk for CVD [6,7]. However, this causal relationship could be confounded by possible pleiotropic effects of gene polymorphisms that are associated with triglyceride (TG) levels and can also influence VLDL particle number and HDL cholesterol [8].

Severe HTG in Familial Chylomicronemia syndrome (FCS), with plasma TG levels above 886 mg/dl (10 mmol/l), is often caused by either homozygous or compound heterozygous loss of function mutations in genes involved in TG lipolysis, leading to impaired Lipoprotein Lipase (LPL)-mediated lipolysis. Frequently these mutations occur in the LPL gene, or in other genes modulating LPL function such as apolipoprotein AV (APOAV), glycosyl phosphatidylinositol HDL binding protein 1 (GPIHBP1), apolipoprotein C-II (APOC2) and Lipase Maturation Factor 1 (LMF1) [9,10]. A major characteristic of this monogenic form of HTG is the young age at which the HTG phenotype manifests. As recently reported by Dron et al. [11]; when severe HTG occurs at later age, the underlying mechanism usually involves a polygenic origin in combination with lifestyle factors; termed the Multigenic Chylomicronemia syndrome. In two independent cohorts of patients with severe HTG, 32% had a high polygenic risk score based on common variants in the above mentioned genes whereas in 53% no genetic cause could be identified, elaborating to the complex nature of HTG involving both genetics and environmental factors [11,12].

Mild-to-moderate HTG, with plasma TG levels varying from 150 mg to 886 mg/dl, is a very frequent condition in the Western world, with a prevalence of around 30%. The epidemic of obesity and the strong increase in prevalence of type 2 diabetes in the last decades have a large share in this matter [13]. It has been shown that the atherogenic lipoprotein phenotype with mildly elevated plasma TG levels, is a hallmark in type 2 diabetes mellitus (DM) patients [14–16]. Additionally, alcohol abuse is another major risk factor for developing mild to severe HTG.

A major clinical issue with the HTG phenotype is the high risk of developing pancreatitis in patients with plasma TG levels above 886 mg/dl [17]. In these patients with severe HTG, the prevalence of acute pancreatitis may exceed 20%, which culminates into a > 350-fold increased risk in LPL-deficient patients compared to the general population [18]. Interestingly, recent data revealed that mild-to-moderate HTG is also associated with pancreatitis in a linear fashion, already starting from 500 mg/dL onwards [19]. To both prevent pancreatitis and lower residual CVD risk, there is an urgent clinical need for effective lowering of TG plasma levels. However, currently available therapeutic agents are not sufficient to substantially decrease plasma TG.

2. Current pharmacological strategies

In HTG, the treatment consists of two goals: reduce residual CVD risk and prevent acute pancreatitis; the latter is particularly important in severe HTG. During the last decades, multiple pharmacological strategies have been explored to effectively lower TG levels. Statins are the first choice of treatment to lower CVD risk in all hyperlipidemias, resulting in a reduction of plasma TG levels by 10–20% [20]. Fibrates currently are the most potent TG-lowering agents, achieving a TG lowering up to 50%, depending on baseline TG levels [20]. Large randomized clinical trials (RCT) have shown at best a modest effect of fibrates on CVD risk, evident in post hoc analyses of patients with elevated TG levels or when used without concomitant statin use [21–26]. Classic fibrates can also be associated with adverse effects regarding liver and renal function [27]. Currently, pemafibrate, a selective fibrate targeting peroxisome proliferator-activated receptor alpha (PPARα) with a more favorable side-effect profile, has entered a phase III outcomes study [28]. Its safety and efficacy was already extensively evaluated [29–32], whereas the PROMINENT trial will now provide results on CVD outcomes in primary prevention patients with diabetes [33].

Next to fibrate and statin regimens, n-3 fatty acids have become an interesting alternative for their TG-lowering capabilities in the last decades. In 2018, a large Cochrane meta-analysis (n = 112.059) showed no effect of omega 3 (n-3) supplement strategies on CVD risk or death [34]. However, the recent results of the REDUCE-IT trial evaluating n-3 eicosapentaenoic acid (EPA) in high dose (4 g/day) showed an impressive 25% reduction in major adverse cardiovascular events (MACE) in both primary and secondary prevention patients [35]. Even more impressive, the use of EPA resulted in a hazard ratio of 0.70 (95% CI 0.55–0.90; p = 0.004) for all-cause mortality in the US subgroup of the REDUCE-IT [36]. Importantly, the modest absolute TG lowering achieved in this study of 14 mg/dl is unlikely to fully explain the major reduction in MACE [37]. In support, the CVD risk benefit was independent of baseline TG levels. Pleiotropic effects of high-dose EPA, comprising anti-inflammatory and anti-thrombotic effects, as well as beneficial changes in endothelial function and membrane-stabilizing effects, have been postulated to contribute to the beneficial outcome [37]. In this context, the recent premature discontinuation of the STRENGTH study (using approximately 3 g of DHA/EPA offering comparable TG-lowering effects as 4 g EPA [38];) due to futility, further supports the notion that EPA cannot be readily classified as a TG-lowering intervention.

Important in the therapeutic approach are also nonpharmacologic interventions: dietary intervention, weight loss, exercise, and reduction of alcohol consumption. In mild-to-moderate HTG patients, dietary interventions should focus on weight reduction and reduction of high glycemic foods. However, in severe HTG (>886 mg/dl), dietary restrictions of fat to less than 25 to 40 g are essential to prevent fasting chylomicronemia.

3. Novel therapeutics

Here, we will focus on the two most promising targets for TG-lowering therapies: apolipoprotein C-III (apoC-III) and angiopoietin-like 3 (ANGPTL3). An overview of current and future therapeutic agents can be found in Table 1.

Table 1. Current and future TG-lowering agents in clinical trials.

Agent	Target	Phase	TG lowering potency	Important side effects	Reference
Fibrates	PPARα	Approved for HTG	~50%	Gastrointestinal symptoms, increased liver function tests, creatinine rise, myositis
Omega-3 fatty acids	Pleiotropic + liver	Approved adjunctive for CVD risk next to statins	~20%	Gastrointestinal symptoms
Statins	HMG-CoA reductase	Approved	~10–20%	Gastrointestinal symptoms, muscle symptoms
Volanesorsen	ApoC-III (ASO)	EMA: conditional authorization FDA: not approved yet	~70–90%	ISRs, thrombocytopenia	[62,63]
AKCEA-APOCIII-LRx	ApoC-III (GALNac3 ASO)	Phase IIa	~60–80%	Possibly mild ISRs	[66]
Evinacumab	ANGPTL3 (monoclonal antibody)	Phase III	~50%	Not known yet	NCT03452228, NCT03409744
IONIS-ANGPTL3-LRx	ANGPTL3 (ASO)	Phase II	~60%	Not known yet	NCT03371355
ARO-ANG3	ANGPTL3 (siRNA)	Phase IIa	~50–60%	Possibly mild ISRs	NCT03747224

TG, triglycerides; PPARα, peroxisome proliferator-activated receptor alpha; HTG, hypertriglyceridemia; ApoC-III, apolipoprotein C-III; ASO, antisense oligonucleotide; EMA, European Medicines Agency; FDA, U.S. Food and Drug Administration; ISR, injection site reaction; GALNac, N-acetyl galactosamine-conjugated; ANGPTL3, angiopoietin-like 3; siRNA, silencing RNA.

3.1. Targeting apolipoprotein C-III in hypertriglyceridemia

ApoC-III is a 79 amino acid peptide that is synthesized in the liver and intestine and is a major component of most circulating lipoproteins, although the majority of apoC-III is on high-density lipoprotein cholesterol (HDL-C) or TG-rich lipoproteins, particularly chylomicrons and very low-density lipoproteins (VLDL). It has to be emphasized that apoC-III is not evenly distributed among lipoproteins. For example, 40%-60% of VLDL, 10%-20% of LDL particles and up to 15% of HDL particles carry apoC-III [39]., which has implications for the understanding of lipoprotein metabolism. In HTG, due to an overload of these TG-rich lipoproteins (TRL), plasma apoC-III levels are elevated, resulting in an increased CVD risk [40,41].

Insights in the role of extracellular apoC-III in pathophysiological processes gradually expanded after its discovery 50 years ago [42]. More than a decade later, it was described that apoC-III (at supraphysiological concentrations ex vivo) inhibits lipolysis of TRLs by lipoprotein lipase (LPL), resulting in persistent atherogenic TRLs [43]. Underlining this concept, transgenic mice expressing 100 copies of the APOC3 gene are severely hypertriglyceridemic [43]. Independent of the LPL-pathway, kinetic studies in humans have shown that the hepatic catabolism of TRLs, involving three different receptors, is impaired in the presence of elevated levels of apoC-III [44]. Further evidence for participation of apoC-III in the modulation of hepatic uptake of TLR via both low-density lipoprotein (LDL) receptors and hepatic sulfate proteoglycans has been shown in targeted mouse models [45]. Another important, but less acknowledged, function of intracellular apoC-III is its effect on the hepatic secretion of TG-rich lipoprotein particles. Evidence comes from in vitro experiments showing that human apoC-III promotes the assembly and secretion of VLDL-like particles in cultured hepatic cells in a lipid-rich condition of excess oleate [46]. Human apoC-III in an apoc3-deficient mice resulted in an increased production of VLDL particles when fed a high-fat diet [47]. A recent kinetic study in type 2 diabetic patients with the atherogenic lipoprotein profile indeed establishes a function for apoC-III in the production and hepatic secretion of VLDL [48].

In 2008, Pollin et al. found that APOC3 mutation carriers in the Amish population had very low plasma TG levels and accompanying low levels of LDL-C and HDL-C [49]. Carriers of this mutation were less likely to have a detectable coronary artery calcium score (CAC) (OR = 0.35) and had a significantly lower 10-year Framingham CHD risk (RR = 0.68) than noncarriers. Later, these findings were replicated in two studies investigating the effect of loss-of-function variants of the APOC3 gene on CVD risk using Mendelian randomization, showing a 40% reduction in CAD [50,51]. This strong correlation does not per se prove causality, due to the earlier mentioned limitations of a Mendelian randomization study design [52]. Interestingly, more recent insights into lipoprotein composition revealed that apoC-III in LDL particles is strongly related to increased risk for CVD in type 2 diabetic patients [53,54]. To complicate our view on apoC-III, it has been shown that the amount of HDL particles containing apoC-III is also positively correlated with an increased CVD risk [55,56]. All these observations provide a rationale for apoC-III as a therapeutic target.

Where the conventional lipid lowering strategies mentioned earlier only modestly decrease apoC-III levels; after the introduction of volanesorsen, an antisense oligonucleotide (ASO) agent, there is a new potent agent inhibiting apoC-III synthesis. ASOs were first applied in vitro in 1978, and since then, multiple ASOs have been clinically implemented [57]. ASOs consist of single-stranded deoxyribonucleotides of approximately 20 nucleotides, which block translation through binding to the targeted mRNA [57]. These drugs are administered subcutaneously and have rapid systemic distribution, and for the newer ASOs, long elimination half-lives. Volanesorsen’s phase I study, completed in 2013 [58], showed promising reductions in apoC-III (up to 78.0%) and TG levels (up to 43.8%) in healthy subjects. Remarkably, 28% of subjects had transient C-reactive protein (CRP) elevations, dose dependent, however, not associated with symptoms. The first open-label phase II trial was performed in three patients with FCS due to LPL deficiency, which resulted in potent lowering of apoC-III levels (71% to 90%) and TG (56% to 86%) [59], providing solid evidence for an LPL-independent TG-lowering effect of apoC-III. In a larger, placebo-controlled phase II trial in 57 patients with HTG, comparable dose-dependent apoC-III and TG-lowering effects were seen [60]. A more recent trial, also placebo-controlled, in 15 type 2 diabetes patients also showed comparable apoC-III and TG reductions [61]. Interestingly, they also observed a 57% improvement in whole-body insulin sensitivity. In all studies, dose-dependent injection site reactions (ISRs) are the most frequent observed side effects [58–60], significantly more frequent compared to placebo. In the phase II study investigating volanesorsen in HTG patients (n = 57), more fatigue, musculoskeletal pain, nausea, and myalgia were observed in the treatment group compared to placebo. One patient developed a serum sickness-like reaction, however without autoantibodies or abnormalities in skin biopsy.

Two recent phase III trials, the APPROACH and COMPASS, with respectively 66 patients with FCS and 113 patients with HTG showed a reduction in plasma TG levels between 70% and 80% [62,63]. In a combined analysis of these studies, volanesorsen showed a significant reduction in pancreatitis, with 1 occurring pancreatitis in the volanesorsen group versus 9 pancreatitides in the placebo group (p = 0.0185) [64,65]. It should be noted that, in the APPROACH trial, declines in platelet counts led to discontinuation of volanesorsen in five patients, which resulted in complete recovery of platelet counts. In sum, volanesorsen is a promising potent TG-lowering agent with reductions up to 90%. However, the unfavorable side-effect profile will likely preclude routine use of this agent in larger groups of patients.

The introduction of the N-acetyl galactosamine-conjugated (GalNAc₃) ASO (AKCEA-APOCIII-LR_x), with a 20–30 fold higher potency and hence lower required dosages, is likely to resolve these adverse effects [66]. This kind of modified ASO establishes this higher potency through selectively targeting the asialoglycoprotein receptor in hepatocytes, minimizing systemic exposure and associated side effects.

Lastly, the dual apoC-II mimetic and apoC-III inhibiting peptide D6PV show potential to effectively lower TG. It mimics apoC-II by activating LPL, while also inhibiting the TG-raising effect of apoC-III [67]. Although it has not been tested in humans yet, D6PV rapidly reduces apoC-III and plasma TG in mice (up to 85% reduction in plasma TG) and has a good bioavailability in non-human primates [68]. Trials in humans will have to point out the clinical applicability of D6PV.

3.2. Inhibiting angiopoietin-like protein 3 to reduce triglyceride levels

After the discovery of ANGPTL3 [69], genetic studies showed that loss-of-function mutations of ANGPTL3 resulted in a hypobetalipoproteinemia associated with decreased plasma TG, HDL-C and LDL-C levels [70]. Subjects with this hypobetalipoproteinemia phenotype showed no coronary atherosclerosis [71]. These findings further accelerated development of ANGPTL3 inhibitors.

In fact, three ANGPTLs are linked to lipoprotein metabolism, i.e. ANGPTL3, 4 and 8. ANGPTL3 is a secreted protein, and exclusively expressed in the liver [69], while ANGPTL4 is synthesized in a large variety of tissues with a focus on adipose tissue. ANGPTL8 is synthesized in both adipose tissue and the liver. ANGPTL3 and ANGPTL4 share a molecular domain structure including a fibrinogen/angiopoietin-like domain which is lacking in ANGPTL8 [72]. ANGPTL3 and ANGPTL4 both inhibit TG hydrolysis through inhibiting LPL function; data mostly comprised from isolated in vitro studies. Recent evidence suggested that ANGPTL4, comparable to ANGPTL3, leads to unfolding and inactivation of the LPL hydrolase domain [73], rather than serving to dissociate catalytically active LPL homodimers into intrinsically unstable LPL monomers as presumed previously [74]. ANGPTL3 and ANGPTL8 form an intracellular complex which increases the function of ANGPTL3. ANGPTL8 itself does not exhibit a clear function [75]. It has been postulated that ANGPTL8 can also form a complex with ANGPTL4 but this is only observed in in vitro studies [76]. This implies that the effect of ANGPTL4 on LPL action may be more persistent [77]. Based on the phenotype of the ANGPTL3 loss-of-function carriers who have hypobetalipoproteinemia one would anticipate that intracellular ANGPTL3 may interfere with VLDL secretion. Indeed, inactivation of ANGPTL3, using specific inhibitory antibodies, in mice lead to a reduction in VLDL-TG but VLDL-apoB100 production was not altered [78]. From these and other studies it is evident that more research is required to elucidate the intracellular function of ANGPTL3 in the liver and its consequence on lipoprotein metabolism.

Inactivating variants in ANGPTL4, comparable with ANGPTL3, result in a lower CVD risk [79,80]. An ANGPTL4 fully human monoclonal antibody (REGN1001) was the first ANGPTL inhibiting treatment tested in both mice and non-human primates and resulted in an almost 50% reduction in TG levels [79]. However, mesenterial lymphadenopathy was observed in both primates and mice, most likely due to increased fatty acid uptake by macrophages in mesenteric lymph nodes [81]. This finding blocked further clinical research in humans, whereafter the focus shifted to the inhibition of ANGPTL3.

The first ANGPTL3 blocking agent used in a clinical trial was the monoclonal antibody evinacumab. In the phase I study (n = 83), promising reductions in both TG and LDL-C were seen [82]. Importantly, there were no serious adverse events, and no discontinuations due to adverse effects. Two participants had transient increase in alanine aminotransferase, which did not lead to discontinuation. In a phase II proof-of concept trial involving nine patients with familial hypercholesterolemia (FH), a median reduction in TG of 47% was achieved next to the LDL-C reduction between 25% and 90% [83]. This was in addition to guideline-based treatment with statins, ezetimibe, and PCSK9 inhibitors. As in the phase I study, there were no discontinuations due to adverse effects. Following these results, the FDA has designated evinacumab a breakthrough status, paving the way for fast and efficient drug approval. Ongoing phase II and phase III trials investigating long-term safety in severe HTG (FCS) and FH will determine whether evinacumab will be a new treatment option in these diseases in the near future (NCT03452228, NCT03409744).

Another therapeutic to reach clinical trials is IONIS-ANGPTL3-L_Rx, an ASO targeting ANGPTL3, which was evaluated in mice and humans [84]. The phase I study (n = 44) in humans showed dose-dependent reductions in TG (up to 63%) and LDL-C (up to 32.9%). There were no differences in adverse events between participants receiving placebo and treatment. A subsequent phase II trial involving 144 patients with NAFLD, type 2 diabetes, HTG, or fatty liver disease is ongoing and will provide further information about the lipid lowering effects and safety of IONIS-ANGPTL3-L_Rx (NCT03371355).

Recently, the first preliminary outcomes of a phase I/II trial (n = 94) with a silencing RNA therapeutic agent (ARO-ANG3) targeting ANGPTL3 were presented (NCT03747224). In the first 40 patients, receiving 4 different single doses compared with placebo, a dose dependent reduction in TG up to 66% was observed, which persisted throughout the study (16 weeks). There was also a dose dependent reduction in LDL-C, however not significant due to the small sample size. There were no serious adverse events or drop-outs in the participants receiving the single dose. Two participants (one active, one placebo) experienced mild transient alanine aminotransferase elevations, one patient had an ISR.

4. Conclusion

HTG is a highly prevalent disease with increasing burden due to the worldwide pandemic of obesity and diabetes. Conventional lipid lowering strategies are not able to efficiently decrease TRLs and lower associated plasma TG levels, leaving a residual CVD risk, and for severe HTG, a high risk of developing pancreatitis. Volanesorsen targeting apoC-III has now been approved in several countries as a therapeutic option for severe HTG in patients with molecularly confirmed FCS, yet, frequent side effects comprising injection site reactions and thrombocytopenia may limit implementation. Several therapeutic modalities targeting ANGPTL3 are catching up and may even be more promising due to lowering of all apoB containing lipoprotein fractions and a favorable side-effect profile (Table 1).

5. Expert opinion

HTG remains a highly frequent condition amongst Western adults. The focus of lipid lowering strategies has been on CVD risk reduction through apolipoprotein B (apoB) and LDL-C lowering strategies, which has resulted in highly potent LDL-C lowering drugs (statins, PCSK9-inhibitors). However, the aforementioned increasing prevalence of obesity and diabetes with associated HTG leaves a residual CVD risk in these high-risk individuals, despite the use of potent LDL-C lowering strategies. On top of that, we emphasized the very high risk of pancreatitis in severe HTG, uncovering an unmet need in current clinical practice.

Whereas the HTG burden keeps increasing, few potent TG-lowering strategies are available for clinicians. The use of fibrates, offering up to 50% TG reduction, failed to show benefit in the majority of RCTs, with CVD benefit confined to patients with (marked) HTG [21–26].

The need for agents reducing the burden of TG-rich lipoproteins in patients at increased CVD risk has resulted in the advent of two promising therapeutic strategies. First, the apoC-III ASO volanesorsen has reached phase III trials, with remarkable reductions in TG plasma levels (70–90%) coinciding with a marked reduction in pancreatitis episodes in patients with FCS [85]. Although the side effects (injection site reactions, flu-like symptoms, thrombocytopenia) hamper its use, the advent of the new-generation GalNAc₃ apoC-III ASO is likely to resolve these issues. Trials addressing the CVD benefit of apoC-III mediated reduction of TRL in high-risk CV patients are now awaited. Second, the ANGPTL3 lowering agents also hold a great promise, since they offer TG reductions up to 70% with concomitant reductions in LDL-C lowering up to 50%, with a favorable side-effect profile. Theoretically, the potent reduction of all atherogenic, apoB containing lipoproteins can be expected to translate into a superior CVD reduction. Also here, an RCT outcome study addressing the capacity of ANGPTL3 lowering to further reduce the residual CVD risk in patients is eagerly awaited.

An important consideration relates to the costs of these new high potential drugs. Whereas high costs may be acceptable for patients with rare diseases at high risk of pancreatitis such as FCS, it will preclude broader implementation in high-risk, mild-to-moderate HTG patients. Hopefully, the advent of newer, less costly platforms such as siRNA will help to reduce costs of new agents. Finally, the future introduction of more agents effectively reducing CVD risk in high-risk patients will draw attention to the importance of personalized medicine algorithms. ‘Brainless’ stacking of novel therapeutic agents on top of existing regimens is hampered by two drawbacks: 1. Reduced absolute benefit when added on top of effective regimens and 2. Reduced adherence due to stacked poly-pharmacy. The future is bright, but not without hurdles.

Article highlights

• Hypertriglyceridemia has a high burden in the modern Western world due to the pandemic of obesity and diabetes

• Hypertriglyceridemia is associated with an increased CVD and pancreatitis risk

• Current pharmacological strategies are not sufficient for TG lowering, especially in moderate to severe HTG patients

• Volanesorsen, an antisense oligonucleotide (ASO) targeting apoC-III, potently lowers TG levels up to 90%

• Evinacumab (monoclonal antibody) and IONIS-ANGPTL3-L_Rx (ASO) effectively lower TG levels up to 70% with concomitant lowering of all other atherogenic apoB containing lipoprotein fractions

• These and more novel therapeutics will be the cornerstone of future HTG treatment, especially in severe HTG patients.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.