Inhibiting angiopoietin-like protein 3: clear skies or clouds on the horizon?

This editorial refers to ‘Genetic inhibition of angiopoietin-like protein-3, lipids, and cardiometabolic risk’, by É. Gobeil et al., https://doi.org/10.1093/eurheartj/ehad845.

‘All that I can say is now it’s getting so much clearer’ Taylor Swift (from Today was a Fairytale)

‘So many things I would have done, but clouds got in my way’ Joni Mitchell (from Both Sides Now)

The more we study the angiopoietin-like protein (ANGPTL) family proteins, the cloudier our understanding becomes. Three family members—ANGPTL3, ANGPTL4 and ANGPTL8 - affect lipid metabolism mainly by modulating energy flux in fed and fasted states. The most familiar is ANGPTL3,¹ which was discovered >20 years ago as the causal gene for hypotriglyceridemia in KK/San mice. Human genetic studies showed that rare loss-of-function variants in ANGPTL3 were causal for familial combined hypolipidemia, a rare semi-dominant phenotype, also called familial hypobetalipoproteinemia type 2 (Mendelian Inheritance in Man identifier 605019), characterized by reduced plasma triglycerides and low-density lipoprotein (LDL) cholesterol.² Plasma high density lipoprotein (HDL) cholesterol is also very depressed in homozygotes.² Population studies show that carriers of disabling ANGPTL3 variants are protected against atherosclerotic disease (ASCVD) events without apparent ill effects.³

These observations motivated development of agents to recapitulate the seemingly favorable natural deficiency of ANGPTL3.¹ These efforts progressed rapidly despite an incomplete understanding of the normal roles of ANGPTL3 and of the mechanisms through which pharmacologic knock-down might impart potential clinical benefits. Targeting ANGPTL3 was assumed to offer a panacea for all hyperlipidemias. Because humans with ANGPTL3 loss-of-function variants had apparently consequence-free panhypolipidemia, it was reasoned that patients with either hypercholesterolemia, hypertriglyceridemia or combined hyperlipidemia could benefit from ANGPTL3 inhibition. At least three agents targeting ANGPTL3 have reached advanced stages of development: evinacumab (Regeneron), vupanorsen (Akcea-Ionis and Pfizer) and zodasiran (Arrowhead) (see Graphical Abstract).

Graphical Abstract.

Angiopoietin-like protein 3 (ANGPTL3) is encoded by the ANGPTL3 gene which lies within an intron of the much larger DOCK7 gene on the short arm of chromosome (Chr) 1. Expression in highest in the liver, where the 7 exons of the gene are transcribed into messenger ribonucleic acid (mRNA) transcripts, yielding a mature protein of 460 amino acids that circulates in plasma. Several pharmaceutical agents inhibit ANGPTL3 production or block its activity at various stages, including genome editing with VERVE-201, RNA interference by the allele specific oligonucleotide (ASO) vupanorsen and small interfering RNA (siRNA) zodasiran, and binding to protein epitopes in plasma by the monoclonal antibody evinacumab. These drugs have been variably given to a non-human primate (NHP) model of familial hypercholesterolemia (FH), and to patients with combined hyperlipidemia (HLP), homozygous familial hypercholesterolemia (HoFH), familial chylomicronemia syndrome (FCS) and multifactorial chylomicronemia syndrome (MCS). Impact of these treatments on plasma concentrations of ANGPTL3, low-density lipoprotein cholesterol (LDL-C) and triglycerides are shown. ND means not determined in the clinical trial

Evinacumab is a fully human monoclonal antibody that binds extracellular ANGPTL3 in plasma and diverts it to the reticuloendothelial system for elimination. This drug, given intravenously every four weeks, is highly effective in patients with homozygous familial hypercholesterolemia (HoFH) who lack LDL receptor activity. The extreme LDL cholesterol elevations in HoFH are refractory to statins and proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors, and respond only to apheresis and lomitapide.⁴ Evinacumab reduces LDL cholesterol by up to 55% in HoFH patients⁵ and also in patients with severe heterozygous FH (HeFH).⁶ Triglyceride reductions of >50% were also seen,^5,6 although these were small in absolute terms. Evinacumab is approved in Europe and North America for treatment of HoFH.

However, evinacumab has inconsistent effects in patients with hypertriglyceridemia. In familial chylomicronemia syndrome (FCS) due to biallelic disabling variants in genes encoding lipoprotein lipase (LPL) or its cofactors, severe hypertriglyceridemia was essentially unchanged with evinacumab, while LDL cholesterol increased by 15%.⁷ In multifactorial chylomicronemia syndrome (MCS) evinacumab reduced triglycerides by up to 80%, but absolute levels often remained markedly elevated,⁷ while LDL cholesterol increased by >50%.⁸ These observations prove that plasma ANGPTL3 interacts with LPL and indicate that targeting ANGPTL3 is likely futile in patients with absolutely lipolytic deficiency. In contrast, evinacumab retains efficacy in patients with a partial lack of LPL activity, i.e. those with MCS. However, by alleviating the backlog of triglyceride-rich lipoproteins (TGRL) evinacumab yields more substrate for hepatic synthesis and secretion of apo B-containing lipoproteins, ultimately raising LDL cholesterol, which may prove to be counterproductive for cardioprotection in this context.

Vupanorsen is an antisense oligonucleotide (ASO) that targets ANGPTL3 messenger ribonucleic acid (mRNA) in the nucleus and cytoplasm. In patients with metabolic syndrome and moderate combined hyperlipidemia, vupanorsen given at doses that suppressed ANGPTL3 by ∼55% reduced triglycerides and LDL cholesterol by ∼40% and ∼7%, respectively.⁹ However, in such patients, vupanorsen at doses that suppressed ANGPTL3 by >95% reduced triglycerides and LDL cholesterol by ∼55% and ∼9%, respectively, but also markedly increased hepatic fat by ∼76% and raised serum transaminase levels.¹⁰ Lower doses of vupanorsen that suppressed ANGPTL3 by only 60%–70% showed similar efficacy on lipids but with more modest liver effects.¹⁰ Vupanorsen development was halted due to the unfavorable benefit-to-harm ratio.

Zodasiran (formerly AROANG3) is a small interfering RNA (siRNA) molecule that targets ANGPTL3 mRNA in the cytoplasm. In healthy volunteers zodasiran reduced plasma ANGPTL3, triglycerides and LDL cholesterol by ∼80%, ∼55% and ∼9%, respectively, with no increase in intrahepatic fat.¹¹ Among patients with mild-to-moderate combined hyperlipidemia (i.e. triglycerides 1.7 to 5.6 mmol/L and LDL cholesterol ≥1.8 mmol/L) zodasiran given subcutaneously at doses up to 200 mg reduced ANGPTL3, triglycerides and LDL cholesterol up to 71%, 59% and 32%, respectively, with no increased hepatic fat and even decreased hepatic fat among those with hepatosteatosis at baseline.¹²

Can these diverse clinical trial findings be reconciled? First, the clearest benefit of ANGPTL3 inhibition is seen in HoFH patients with severely elevated LDL cholesterol and normal triglycerides. In this situation, evinacumab reduces LDL cholesterol by a receptor-independent mechanism within plasma. In addition, LPL or endothelial lipase are probably not important for this effect, since their conventionally understood activities are remote from LDL particles. Also, because evinacumab effectively reduces LDL cholesterol in patients without HoFH but with refractory hypercholesterolemia, this agent, particularly subcutaneously could help a wider population with isolated LDL cholesterol elevation. Preliminary data suggest that zodasiran similarly reduces LDL cholesterol in HoFH patients.

Second, among FCS patients with severe hypertriglyceridemia and impaired clearance of TGRL, inhibiting circulating ANGPTL3 has essentially no effect on triglycerides. But since individuals with MCS experience triglyceride lowering with evinacumab, this indicates that the extracellular function of ANGPTL3 requires at least some lipolytic capacity. It has been suggested that failure to respond to ANGPTL3 inhibition may actually differentiate FCS from other forms of severe hypertriglyceridemia such as MCS, but this remains to be shown. Furthermore, we are lacking data in these patients of the effect of targeting ANGPTL3 mRNA.

Third, in patients with moderate combined hyperlipidemia who tend to be obese, insulin resistant and prone to over-secrete hepatic TGRL, reducing ANGTPL3 by 60%–70% via RNA interference by either vupanorsen or zodasiran is associated with moderate and mild reductions in triglycerides and LDL cholesterol, respectively, although hepatic fat is increased with vupanorsen. In contrast, intense >95% reduction of ANGPTL3 levels by vupanorsen delivered no incremental lipid benefit but was associated with marked hepatic steatosis. As proposed by Osstveen and colleagues, essentially complete inhibition of ANGPTL3 in the hepatocyte could attenuate the unloading of apoB-containing lipoproteins intracellularly, leading to intrahepatic lipid accumulation⁸ as also suggested by silencing experiments in liver cell lines¹³ and in subjects with ANGPTL3 loss-of function mutations.² This intracellular mechanism was also encountered by Kathiresan and colleagues¹⁴ who used CRISPR-Cas9 silencing to demonstrate an intrahepatic role for ANGPTL3 in secretion and assembly of apo B-containing lipoproteins, independent of its extracellular effects on plasma lipases.

A final piece of evidence that further clouds the rationale for targeting ANGPTL3 is now reported in the European Heart Journal by Gobeil, Bourgault and colleagues.¹⁵ These clever investigators imputed the effect of ANGPTL3 inhibition at the genomic DNA, RNA and protein levels by evaluating, respectively, truncating DNA variants in ANGPTL3 in populations (as a surrogate for genomic editing), single nucleotide polymorphisms (SNPs) associated with depressed ANGPTL3 RNA levels in liver tissue from obese patients (as a surrogate for RNA knock-down), and SNPs associated with plasma ANGPTL3 levels (as a surrogate for antibody inhibition of ANGPTL3 protein). They found that: 1) common SNPs associated with ANGPTL3 mRNA or plasma protein levels showed strongest association with triglycerides, less with LDL cholesterol, marginal with apo B and none with ASCVD; and 2) rare ANGPTL3 truncating variants showed a similar gradient of association with largest effects on triglycerides, less with LDL cholesterol, marginal with apo B and none with ASCVD. Associations with most clinical traits were negative, except for an unexpected positive association with pancreatitis episodes that appeared not to be mediated via hypertriglyceridemia. The authors conclude that ANGPTL3 inhibition to reduce TG will have only a modest effect on ASCVD outcomes unless reductions in LDL cholesterol and apo B-containing lipoproteins attained in real world clinical applications are greater than those suggested by the indirect genetic models.

To date, studies with ANGPTL3-lowering agents indicate a clear benefit on high LDL cholesterol in HoFH patients with absent LDL receptors and normal triglyceride metabolism. However, when hypertriglyceridemia enters into the phenotype, antibody binding of ANGPTL3 protein in plasma has no effect if lipolytic function is completely compromised and inconsistent benefits when lipolytic function is partially compromised. Furthermore, increasingly intense intracellular ANGPTL3 inhibition by ASO RNA silencing is associated with diminishing returns for plasma lipid lowering at the expense of marked hepatic fat accumulation with the highest doses of vupanorsen in patients with combined hyperlipidemia who oversecrete TGRL. Finally, the contrasting apparent absence of liver fat accumulation with zodasiran might be due either to its different mechanism of RNA silencing (siRNA vs. ASO) or because ANGPTL3 reductions >70% were not evaluated.

Have the recent new data clarified or clouded our understanding of ANGPTL3? This target is obviously biologically complex without straightforward explanations for the diverse clinical trial findings. The benefit of targeting ANGPTL3 is clear in HoFH patients without hypertriglyceridemia, but becomes cloudy when the clinical picture includes hypertriglyceridemia. The impact of ANGPTL3 inhibition depends on the site of inhibition (i.e. liver vs. plasma), the mode by which inhibition is mediated (i.e. siRNA vs. ASO vs. monoclonal antibody vs. future genome editing) and also the genetic and metabolic context of the patient (i.e. HoFH vs. FCS vs. complex polygenic MCS vs. obese patient with combined hyperlipidemia). Future work could help scatter the clouds, allowing us to see ANGPTL3 from both sides and to maximize the benefit of inhibiting this interesting and challenging target.

Declarations

Disclosure of Interest

R.A.H. reports consulting fees from Acasti, Aegerion, Akcea/Ionis, Amgen, Arrowhead, HLS Therapeutics, Novartis, Pfizer, Regeneron, Sanofi and UltraGenyx.

Funding

R.A.H. is supported by the Jacob J. Wolfe Distinguished Medical Research Chair, the Edith Schulich Vinet Research Chair, and the Martha G. Blackburn Chair in Cardiovascular Research. R.A.H. holds operating grants from the Canadian Institutes of Health Research (Foundation award), the Heart and Stroke Foundation of Canada (G-21-0031455) and the Academic Medical Organization of Southwestern Ontario (INN21–011).