Regulation of lipoprotein metabolism by ANGPTL3, ANGPTL4, and ANGPTL8

Abstract

Triglyceride-rich lipoproteins deliver fatty acids to tissues for oxidation and for storage. Release of fatty acids from circulating lipoprotein triglycerides is carried out by lipoprotein lipase (LPL), thus LPL serves as a critical gatekeeper of fatty acid uptake into tissues. LPL activity is regulated by a number of extracellular proteins including three members of the angiopoietin-like family of proteins. In this review, we discuss our current understanding of how, where, and when ANGPTL3, ANGPTL4, and ANGPTL8 regulate lipoprotein lipase activity, with a particular emphasis on how these proteins interact with each other to coordinate triglyceride metabolism and fat partitioning.

LIPOPROTEIN LIPASE AND TRIGLYCERIDE PARTITIONING

Fatty acids are a vital energy source for many oxidative tissues, including heart and skeletal muscle. Most fatty acids are obtained from diet, but fatty acids can also be synthesized de novo, primarily in the liver and adipose tissue. Both dietary and liver-synthesized fatty acids are esterified into triglycerides (TG) and packaged into triglyceride-rich lipoproteins. These lipoproteins, chylomicrons for intestinally packaged dietary fat and VLDL for liver-derived fatty acids, circulate in the bloodstream allowing for the distribution of fatty acids to various peripheral tissues.

In most tissues, extraction of fatty acids from circulating triglyceride-rich lipoproteins is primarily carried out by lipoprotein lipase (LPL). Acting on the luminal surface of capillaries, LPL binds circulating triglyceride-rich lipoproteins and hydrolyses the triglycerides within, releasing fatty acids for uptake into tissues (1). LPL is expressed highly in tissues that consume or store fat, including heart, skeletal muscle, all adipose depots, and the mammary gland (2–5). LPL is secreted by the parenchymal cells of these tissues, i.e., the adipocytes, cardiomyocytes, and myocytes (6) and then transported across capillary endothelial cells by the endothelial cell protein glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1 (GPIHBP1) (7). GPIHBP1 also anchors LPL to the capillary wall during lipolysis of lipoproteins (8, 9). Without active vascular LPL, very little lipoprotein triglyceride is delivered to peripheral tissues. In mice, LPL-deficiency leads to neonatal death due to the inability of pups to properly utilize the triglyceride-rich milk from their mothers (10). Humans deficient in LPL manifest severe hypertriglyceridemia due to the lack of triglyceride clearing (11).

The fatty acids liberated by LPL-mediated lipolysis are largely taken up by the tissue in which LPL is located. Thus, LPL acts as a fatty acid gateway and the relative levels of LPL activity in different tissues are a major determinant of the relative distribution of triglyceride-derived fatty acids to those tissues. The tissue distribution of fatty acids, often referred to as fatty acid partitioning, shifts according to metabolic state and is important for maintaining metabolic homeostasis. Ectopic deposition of fat to tissues such as liver and skeletal muscle can drive metabolic dysfunction and lead to metabolic disease (reviewed in Ref. 3). Thus, proper regulation of LPL activity is critical for maintaining metabolic health. Although LPL activity is regulated on many levels, including transcriptional, translational, and posttranslational, one of the most prominent forms of regulation is inhibition by extracellular proteins. Here, we examine three such extracellular inhibitors, ANGPTL4, ANGPTL3, and ANGPTL8, which have emerged as key regulators of LPL, acting in specific tissues and at specific times to modulate LPL activity and adjust lipid partitioning.

THE ANGPTL FAMILY OF PROTEINS

The angiopoietin-like (ANGPTL) family of proteins are named such because of their structural similarity to angiopoietin proteins. Like angiopoietin proteins, ANGPTL proteins, with the exception of ANGPTL8, contain a coiled-coil domain and a large globular C-terminal fibrinogen-like domain (12, 13). Unlike angiopoietin proteins, ANGPTL proteins do not bind the receptor tyrosine kinase Tie2, but may still contribute to angiogenesis (12, 14). Each of the ANGPTL proteins contain a signal peptide and can be secreted from cells. The family currently has eight known members, ANGPTL1-8, with pleiotropic functions (13). Relevant to this review, several family members contribute to triglyceride metabolism. ANGPTL4, ANGPTL3, and ANGPTL8 have been subject to detailed study using both animal models (Table 1) and human genetics. These three proteins will be discussed in depth below. In addition to these three proteins, ANGPTL5 and ANGPTL6 may also play a role in lipid metabolism. Single nucleotide polymorphisms (SNPs) in ANGPTL5 are associated with plasma triglyceride levels in humans (40). In humans, ANGPTL5 is expressed in several tissues, including high expression in adipose tissue (40). However, ANGPTL5 is not found in mice, which has hampered its further study. Mice deficient in ANGPTL6 (also known as angiopoietin-related growth factor) become obese, insulin resistant, and accumulate lipid in skeletal muscle and liver (41). However, despite these marked changes, ANGPTL6 does not appear to modulate plasma TG levels (40, 41).

Table 1.

Animal models used in ANGPTL studies

	Generated Animal Models
ANGPTL3	KK/San mice with natural ANGPTL3 mutation (15) Adenovirus overexpression mouse (15) Whole body knockout mice (16–19) ANGPTL3-deficient mouse created with CRISPR base editing (20) ANGPTL3/ANGPTL8 double knockout mouse (21)
ANGPTL4	Adenovirus overexpression mice (22–24) Mouse with adenovirus overexpression of only the fibrinogen-like domain (25) Mouse with transgenic overexpression from the ApoE promoter (16) Mouse with transgenic overexpression from the AP2 promoter (26) Mouse with transgenic overexpression from the α-MHC promoter (27) Mouse with transgenic overexpression from the NPHS2 promoter (28) Whole body knockout mice (16, 29, 30) Mouse with ANGPTL4 E40K mutation (17) Adipose-specific (Adipo-Cre) knockout mice (31, 32) Brown adipose-specific (UCP1-ER-Cre) knockout mouse (33) ANGPTL4 hypomorphic mouse (EUCOMM/KOMP allele) (34)
ANGPTL8	Adenovirus overexpression mice (19, 35) Whole body knockout mice (36, 37) Whole body knockout rat (38) Mouse with human ANGPTL8 (164) Adipose-specific (Adipo-Cre) knockout mice (39) Liver-specific (Alb-Cre) knockout mice (39) ANGPTL3/ANGPTL8 double knockout mouse (21)

ANGPTL, angiopoietin-like.

In addition to their role in regulating plasma triglyceride levels, ANGPTL3, ANGPTL4, and ANGPTL8 have been associated with numerous physiological roles including stem-cell renewal, angiogenesis, and inflammation (42–45). However, this review will focus specifically on their roles in triglyceride metabolism.

ANGPTL4

Discovery

In 2000, two groups described the discovery of a secreted PPAR target gene expressed highly in adipose tissue (46, 47). Yoon et al. (47) named the protein PGAR (PPARγ angiopoietin related). Kersten et al. (46) named the protein fasting-induced adipose factor (FIAF) owing to their observation that the protein was highly induced by fasting. They also found that fasting induced liver expression of FIAF. Both groups speculated that this newly uncovered protein was involved in metabolism in some way. This conjecture was supported by a study in 2002 from Yoshida et al. (48), who renamed FIAF/PGAR to ANGPTL4 based on its homology to ANGPTL3. They showed that injection of ANGPTL4 could increase plasma triglycerides and that recombinant ANGPTL4 could inhibit LPL in vitro (48). The role of ANGPTL4 in regulating LPL was further solidified in a 2005 study by Köster et al. (16). They also found that ANGPTL4 could inhibit LPL in vitro. Moreover, they reported that transgenic overexpression of ANGPTL4 increased plasma triglycerides and reduced LPL activity measured in postheparin plasma. Conversely, ANGPTL4-deficient mice had lower than normal plasma triglycerides and higher postheparin-plasma LPL activity (16). In the years since, multiple laboratories have confirmed the ability of ANGPTL4 to inhibit LPL (49–55). With the ability of ANGPTL4 to inhibit LPL firmly established, attention has turned to the molecular mechanisms and the physiological roles of this inhibition.

Physiological Roles in Lipid Metabolism

ANGPTL4 clearly influences circulating triglyceride levels. It inhibits LPL (16, 48, 49, 55). ANGPTL4 loss-of-function mutations in humans are associated with lower plasma triglycerides (40, 56–58). ANGPTL4 deficiency in mice also results in lower plasma triglycerides (16, 17, 59) and ANGPTL4 overexpression increases triglycerides (16, 22, 26). However, despite the clear effect of ANGPTL4 on circulating triglycerides and despite the presence of ANGPTL4 in the circulation (26), most evidence suggests that the influence of ANGPTL4 on circulating TGs is a result of it acting locally in the tissue in which it is expressed to control local uptake of triglyceride-derived fatty acids. In fact, several studies have found that vascular LPL, bound to its receptor GPIHBP1, is protected from ANGPTL4 inhibition (51, 52, 60). As this is the context in which circulating ANGPTL4 would encounter LPL, these observations suggest that ANGPTL4 can only effectively inhibit LPL if it does so before it binds to GPIHBP1 on endothelial cells.

The most well characterized example of ANGPTL4 acting locally to control triglyceride partitioning is in adipose tissue during fasting. In 2002, after ANGPTL4 was discovered, but before its characterization as an LPL inhibitor, Bergö et al. (61) found that the decrease in adipose LPL activity that occurs with fasting was mediated by an independent protein that acted on LPL posttranslationally. It is now generally accepted that this protein is ANGPTL4. Angptl4 is highly expressed in adipose tissue (46, 47) and expression in adipose is quickly induced by fasting (16, 46, 62, 63). In adipocytes, ANGPTL4 appears able to inactivate LPL both intracellularly (64, 65) and extracellularly (29, 66). Thus, under fasting conditions, ANGPTL4 would inhibit LPL and reduce triglyceride delivery to adipose tissues. Studies with knockout mice confirm that in the absence of ANGPTL4, LPL activity and TG uptake in adipose increases under fasting conditions (31, 32, 63). In adipose-specific ANGPTL4 knockout mice, LPL activity is increased in adipose tissue but not in other tissues (31, 32), supporting the idea that adipose ANGPTL4 acts almost exclusively in paracrine/autocrine fashion to regulate LPL activity. Recent studies in human subjects confirm that fasting induces ANGPTL4 inhibition of LPL in human adipose as well (67).

Adipose-derived ANGPTL4 also plays a role in the triglyceride partitioning changes that occur with cold exposure. Cold exposure induces large increases in triglyceride uptake into brown adipose tissue (BAT) (68, 69). These increases require LPL activity, suggesting a shift in LPL regulation (68). Angptl4 gene expression decreases in brown adipose tissue in the cold (33, 70, 71). The resulting decrease in ANGPTL4 protein leads to increased LPL activity in BAT, increasing uptake of triglycerides for thermogenic oxidation (71). Cold exposure did not increase LPL activity in white adipose tissue, nor was ANGPTL4 expression decreased (71), indicating that these changes were specific to brown adipose tissue. Consistent with a role for ANGPTL4 in thermogenic triglyceride partitioning, either whole body or BAT-specific deletion of ANGPTL4 resulted in increased LPL activity and triglyceride uptake in BAT (33, 71). It is important to note, however, that modulation of ANGPTL4 expression is not essential for cold-induced thermogenesis or even cold-induced changes in triglyceride partitioning. Dijk et al. (71) found that greater uptake of TGs into BAT was induced by cold both in ANGPTL4-deficient mice and in mice overexpressing ANGPTL4. Dijk et al. (71) also observed that neither whole body knockout of ANGPTL4 nor ANGPTL4 overexpression had a significant effect on body temperature after prolonged cold exposure. Singh et al. (33) found that mice lacking ANGPTL4 specifically in BAT did have slightly higher body temperatures after acute cold exposure, but did not test the effects of prolonged exposure.

ANGPTL4 is also involved in the shifting of metabolic fuel in other tissues and under other conditions. Although Angptl4 is most highly expressed in adipose, it is also expressed in several other tissues including liver, heart, intestine, and skeletal muscle (40, 46, 47, 72). Catoire et al. (73) found that acute exercise increased ANGPTL4 expression in skeletal muscle, but only in nonexercising muscle. Using cell culture models they further found that free fatty acids, which increase in the serum with exercise, induce ANGPTL4 expression in myocytes. However, AMPK activation, as would occur in exercising muscle, counteracted fatty acid induction of ANGPTL4. These data suggested a model in which ANGPTL4 is induced in nonexercising skeletal muscle to prevent lipotoxicity and to divert fuel to exercising muscle where fatty acid oxidation is increased (73).

ANGPTL4 also plays an important role in limiting LPL activity in macrophages. LPL expression is high in macrophages and LPL activity can increase lipid uptake and foam cell formation when macrophages are exposed to triglyceride-rich lipoproteins (74). Strong evidence suggests that ANGPTL4 regulates this activity, blunting lipid uptake into macrophages. ANGPTL4-deficient mice fed a diet high in saturated fat accumulate Touton giant cells, cells that result from the fusion of lipid-laden macrophages, in their mesenteric lymph nodes (34, 59). These mice eventually die, likely as a consequence of the inflammation that results from unfettered lipid uptake into lymphatic macrophages (59). Peritoneal macrophages isolated from ANGPTL4-deficient mice exhibit significantly more lipid uptake and inflammatory cytokine release than those from normal mice when incubated with chyle (chylomicron-containing lymphatic fluid) (59). Treating these cells with either recombinant ANGPTL4 or with LPL inhibitors greatly reduces both lipid uptake and cytokine release, suggesting that ANGPTL4 protects from lipid-induced inflammatory responses (59). microRNA-134m, which suppresses ANGPTL4 expression in macrophages, increases foam cell formation in an LPL-dependent manner (75, 76). ANGPTL4 deficiency specifically in hematopoietic cells also increases both foam cell formation and atherosclerosis (77). Conversely, overexpression of ANGPTL4 in a mouse atherosclerotic model reduces foam cell formation and atherosclerosis (78).

ANGPTL4 may play a role in protecting against lipotoxicity in cardiac tissue as well. In the heart, ANGPTL4 is localized to the surface of cardiomyocytes (79). In mice, overexpression of ANGPTL4 reduces triglyceride uptake into the heart and protects against oxidative stress after either oral fat loading or high-fat diet feeding (27, 79).

ANGPTL4 is also expressed in the intestine (30, 80), where, according to one report, it may regulate fat absorption, possibly by inhibiting pancreatic lipase (81). In the intestine, there also appears to be an interaction between ANGPTL4 and the gut microbiome (30, 82–85). Conventionalizing germ-free mice (colonizing germ-free mice with microorganisms) suppresses intestinal ANGPTL4 expression (30, 83). In some studies, germ-free mice were protected from weight gain, and ANGPTL4 appeared to be necessary for this protection (30, 82). However, in other studies, germ-free mice were not protected from weight gain (83). The microbiota have also been implicated in the improved glucose tolerance observed in ANGPTL4-deficient mice fed a diet high in unsaturated fat (84). Despite these intriguing findings, the interactions between the microbiota and intestinal ANGPTL4 have not been clearly defined, nor have the mechanisms by which modulation of intestinal ANGPTL4 changes metabolic homeostasis been delineated. Future studies, likely employing intestinal-specific ANGPTL4 knockout mice, will be needed.

ANGPTL4 is robustly expressed in the liver in mice, and in humans liver is the tissue with greatest ANGPTL4 expression (86, 87). Yet the role of ANGPTL4 in the liver remains mysterious. Little LPL is expressed in the liver (3–5), and whole body ANGPTL4 deficiency does not alter triglyceride uptake into the liver (63), suggesting that hepatic ANGPTL4 does not target LPL-mediated lipolysis in the liver. Early studies showed that much of liver-derived ANGPTL4 is proteolytically cleaved (88). As proteolytic cleavage is known to increase the ability of ANGPTL4 to inhibit LPL (51, 89, 90), it is possible that cleaved, liver-derived ANGPTL4 is secreted into the circulation and regulates systemic triglyceride metabolism in an endocrine manner. Studies with liver-specific ANGPTL4 knockout mice are likely needed to fully understand the role of ANGPTL4 in this important tissue.

ANGPTL4 deficiency in humans is associated with higher circulating HDL-cholesterol (HDL-C) levels (56, 58, 91, 92), but the mechanism behind this change is not clear. In humans, HDL levels correlate with LPL activity levels (93, 94). Thus, increased HDL levels in ANGPTL4-deficiency individuals could simply be the indirect consequence of increased LPL activity. It is also possible that ANGPTL4 plays a more direct role in regulating HDL levels. ANGPTL4 can be found associated with HDL particles (26, 95), and a study by Yang et al. (95) reported that ANGPTL4 associated with HDL protects HDL phospholipids from hydrolysis by endothelial lipase (EL). However, further investigation is necessary to adequately understand the relationship between ANGPTL4 and HDL metabolism.

Mechanism of Action

The N-terminal coiled-coil domain of ANGPTL4 mediates the formation of oligomers, mostly dimers and tetramers (22, 89, 96). Oligomerization appears to be required for LPL inhibition (22, 89), though the structural basis for this requirement is not known. ANGPTL4 can be cleaved by furin-like proteases, separating the N-terminal coiled-coil domain from the C-terminal fibrinogen-like domain (96). When cleaved, the C-terminal domain dissociates to monomers, but the N-terminal region can remain oligomerized (89, 96). Several studies have shown that not only is the N-terminal region of ANGPTL4 necessary and sufficient for LPL inhibition, but also that, once cleaved from the fibrinogen-like domain, the N-terminal region inhibits LPL more efficiently than full-length ANGPTL4 (49, 51, 72, 89, 90). ANGPTL4 expressed in the liver is cleaved at a much higher rate than ANGPTL4 expressed in the adipose (88), but the physiological role of ANGPTL4 cleavage and how it is regulated are not clear.

The mechanism by which ANGPTL4 inhibits LPL has been the subject to much study and some debate. Several studies have indicated that ANGPTL4 catalytically inactivates LPL (49, 51, 52, 97, 98). For many years, LPL was thought to be active only as a homodimer, and initial studies suggested that ANGPTL4 inactivated LPL by converting active dimers to inactive monomers (49). Recent studies have shown that LPL is catalytically active as a monomer (99), making a dimer-to-monomer mechanism less likely. Indeed, a recent series of studies from the laboratory of Michael Ploug have shown that ANGPTL4 can act on monomeric LPL, and that ANGPTL4 increases the inherent instability of LPL resulting in the unfolding of LPL’s catalytic domain (52, 97, 98). This mechanism has not yet been universally accepted and studies from the laboratory of Saskia Neher have suggested that ANGPTL4 acts as a reversible, uncompetitive inhibitor of LPL (53, 100).

The ability of ANGPTL4 to inhibit LPL is also modulated by the environmental milieu. Vascular LPL is normally bound to its endothelial cell receptor GPIHBP1 (7–9). LPL bound to GPIHBP1 is largely protected from ANGPTL4 (51, 52, 60, 98). LPL is also protected from ANGPTL4 by the presence of fatty acids and lipoproteins (101, 102).

ANGPTL3

Discovery

In 1999, Conklin et al. (103) searched expressed sequence tag (EST) databases for signal sequences and amphipathic helices and found a 460-amino acid protein with the conserved secondary structure of the ANGPTL family (a signal peptide followed by an N-terminal coil-coil domain with a linker leading to a C-terminal fibrinogen-like domain). They found that this protein, which they named ANGPTL3, was expressed almost exclusively in the liver in both mice and humans (103). The role of ANGPTL3 in lipid metabolism was uncovered through a substrain of KK-obese mice. Although KK mice manifest obesity, insulin resistance, and hyperlipidemia (104), this substrain (KK/San) presented with strikingly lower levels of plasma triglycerides (15). Koishi et al. (15) identified a homozygous loss of function mutation in ANGPTL3 as the causal factor. They further showed that overexpression of wild-type ANGPTL3 could increase plasma triglycerides in these mice (15). Soon after, Shimizugawa et al. (105) made a direct connection between ANGPTL3 and LPL, finding that the KK/San mice had increased clearance of triglyceride-rich lipoproteins and showing that ANGPTL3 could inhibit LPL in vitro.

Physiological Roles in Lipid Metabolism

Since its initial discovery, the connection of ANGPTL3 to lipid metabolism has been extensively validated. ANGPTL3 null mice have decreased cholesterol and triglyceride levels and increased LPL activity in postheparin plasma (16, 18, 106). Overexpression or injection of ANGPTL3 into mice increases plasma triglycerides (15, 107). In humans, loss of functional ANGPTL3 protein also leads to decreased triglyceride levels (40, 108–110).

ANGPTL3 is expressed in and secreted from the liver and circulates in the bloodstream (15, 40, 72, 103). The physiological data would therefore suggest that circulating ANGPTL3 encounters and interacts with LPL in the vasculature, inhibiting its lipase activity and reducing clearance of TG-rich lipoprotein particles. In the absence of ANGPTL3, LPL would remain active and clearance of VLDL and chylomicrons would increase. In reality, the action of ANGPTL3 is more nuanced. Wang et al. (106) found that ANGPTL3 deficiency did increase TG-derived fatty acid uptake into heart and skeletal muscle, but that ANGPTL3 deficiency simultaneously decreased uptake into white adipose tissue. Moreover, this effect was observed primarily in the fed state (106). Curiously, although ANGPTL3 expression is regulated by LXR, insulin, leptin, and thyroid hormone (111–114), circulating levels of ANGPTL3 are relatively constant across feeding states (72, 106, 115).

In addition to decreased plasma triglycerides, mice and humans lacking ANGPTL3 also have lower plasma cholesterol levels (16, 18, 106, 108, 109, 116–118). The effect of ANGPTL3 on cholesterol levels is largely attributed to the ability of ANGPTL3 to inhibit endothelial lipase (EL). EL belongs to the same lipase family as LPL, but targets phospholipids on the shell of high-density lipoproteins (HDL), leading to decreased HDL levels (119–122). The relationship between ANGPTL3 and EL was uncovered by Shimamura et al. (123), who found that ANGPTL3 inhibited EL in a dose-dependent manner and that reexpression of ANGPTL3 in ANGPTL3-knockout mice increased plasma HDL-C levels. Subsequent studies confirmed the ability of ANGPTL3 to inhibit EL (124). More recent evidence indicates that EL regulates LDL-cholesterol (LDL-C) levels by hydrolyzing the phospholipids of VLDL remnant particles and reducing their conversion to LDL, suggesting that the lower LDL-C levels observed in ANGPTL3-deficient individuals may also arise from reduced EL inhibition (125, 126). Alternatively, lower LDL-C levels might be mediated by LPL. LPL is known to facilitate the uptake of lipoproteins in the liver (127–130). It is possible that in the absence of ANGPTL3, this nonenzymatic function of LPL might also be increased, thus lowering circulating levels of LDL-C.

Mechanism of Action

ANGPTL3 contains a protease cleavage site, and in vivo can be cleaved into two domains both intracellularly by PCSK3 and extracellularly by PACE4 (107, 124, 131, 132). The N-terminal domain contains the conserved SE1/LPL inhibitory domain (amino acids 32–55), a domain found in ANGPTL3 and ANGPTL4 known to be important for the binding and inhibition of LPL (17). Several studies have identified the N-terminal cleavage product to be necessary and sufficient for lipase inhibition (107, 123, 124). The necessity of proteolytic cleavage remains unclear. In vitro studies of the effect of cleavage on ANGPTL3-mediated inhibition have reported increased inhibition, decreased inhibition, and no effect on inhibition (107, 115, 124). As with ANGPTL4, ANGPTL3 oligomerizes, appearing to form trimers or hexamers, and oligomerization is mediated by the N-terminal domain (72, 100).

Aside from proteolytic cleavage, the ability of ANGPTL3 to inhibit LPL is modulated by a number of environmental factors. LPL’s endothelial cell receptor GPIHBP1 protects LPL from ANGPTL3 inhibition as does heparin (60, 115). The presence of triglyceride-rich lipoproteins also reduces ANGPTL3 inhibition (102).

The mechanism by which ANGPTL3 inhibits EL has not been investigated. The mechanism by which ANGPTL3 inhibits LPL has been studied, but has not yet been identified. Several possible, and somewhat contradictory, mechanisms have been proposed. One proposed mechanism, the dissociation of LPL dimers into inactive monomers (133), seems less likely given recent evidence that LPL is catalytically active as a monomer (99). LPL can be cleaved into an inactive state by furin-like proprotein convertases (134, 135). Liu et al. (132) found that ANGPTL3 increased the cleavage of LPL by PACE4, but not PCKS5. Interestingly, this cleavage was not prevented by the presence of GPIHBP1 or heparan sulfate proteoglycans (HSPGs) (132). Whether the LPL cleavage they observed was the direct mechanism of inhibition or LPL that has already been inhibited by ANGPTL3 is simply more susceptible to cleavage is not clear. Shan et al. (55) found that LPL inhibition by ANGPTL3 was distinct from ANGPTL4-mediated LPL inactivation. They found that ANGPTL3 reduced the catalytic activity of LPL, but did not increase the rate of self-inactivation. They also determined that ANGPTL3 inhibition of LPL was reversible and that heparin protected LPL from inhibition (55). In contrast, Mysling et al. (52) found that, like ANGPTL4, ANGPTL3 could inhibit LPL by unfolding or sterically changing the active domain of LPL, though its potency was far less than that of ANGPTL4. In fact, several studies have reported that ANGPTL3 is a less potent inhibitor than ANGPTL4 (54, 55, 60, 133, 136), and most of the studies showing that ANGPTL3-mediated LPL inhibition in vitro required supraphysiological concentrations of ANGPTL3 to achieve significant inhibition (60, 105, 107, 115). These observations were curious given that ANGPTL3-deficiency lowers plasma TG just as markedly as ANGPTL4 deficiency (16). As discussed below, more recent studies have shown that ANGPTL3 forms of complex with ANGPTL8 and that this complex is most likely the physiologically relevant LPL inhibitor (19, 54, 115, 137, 138). As none of the ANGPTL3 inhibition mechanistic studies discussed here were performed in the presence of ANGPTL8, their significance is unclear.

ANGPTL8

Discovery

ANGPTL8 is the most recently identified member of the ANGPTL family in mammals. In 2012, three groups independently identified ANGPTL8 and published studies indicating its importance in triglyceride metabolism (19, 35, 139). All three studies recognized the homology of ANGPTL8 with other ANGPTL family members especially ANGPTL3. All three studies also found that ANGPTL8 is expressed primarily in the adipose tissue and liver of mice and that expression was much higher in the fed state than in the fasted state. The Smas laboratory named the protein RIFL (refeeding induced fat and liver) based on its expression pattern (139). They reported that mice lacking ANGPTL8 had much lower than normal plasma triglyceride levels and that lipid storage into adipocytes was impaired in the absence of ANGPTL8. The Zhang laboratory first identified ANGPTL8 in an RNA-seq screen looking for nutritionally regulated genes (35). In addition to expression of ANGPTL8 in the adipose and liver of mice, they found that ANGPTL8 was expressed in human liver. Consistent with the decreased plasma TG levels observed in ANGPTL8-deficient mice, they found that overexpression of ANGPTL8 increased plasma TG levels. Moreover, they reported that recombinant ANGPTL8 was able to inhibit LPL activity and thus named the protein lipasin (35). Like the Zhang laboratory, the laboratory of Jonathan Cohen and Helen Hobbs also reported that ANGPTL8 overexpression increased plasma TG levels in mice (19), and also later showed that plasma TG levels were decreased in ANGPTL8-deficient mice (36). They found that similar to mice, ANGPTL8 was expressed in the liver and adipose of humans (19). Importantly, they first reported the interactions of ANGPTL8 with ANGPTL3, a finding that was crucial for the understanding of ANGPTL8 function, as will be discussed below. Based on its homology with other ANGPTL proteins, they named the protein ANGPTL8 and this is the name that has now been adopted in the field. A year after these foundational papers were published, another study renamed ANGPTL8 as betatrophin due to its purported ability to induce β-cell proliferation (140). This finding could not be reproduced (36, 141–143) and the initial manuscript has been retracted. Thus, the name betatrophin should not be used.

Physiological Roles in Lipid Metabolism

ANGPTL8 was initially identified as a regulator of triglyceride metabolism and this remains its clearest and most characterized physiologic function. However, it does not appear to act independently in this role. Instead, ANGPTL8 exerts its effects through its interactions with ANGPTL3 and ANGPTL4.

Interactions with ANGPTL3.

Although ANGPTL8 was initially reported to inhibit LPL on its own (35), later studies found that this not to be the case (115, 137, 138). Instead, ANGPTL8 acts on LPL and on triglyceride metabolism through its interaction with ANGPTL3. ANGPTL8 overexpression only increases plasma triglyceride levels if ANGPTL3 is present (19, 137). ANGPTL8 forms a complex with ANGPTL3 (19, 54, 115, 138) with an apparent stoichiometry of 3 ANGPTL3:1 ANGPTL8 (54). Complex formation is necessary for efficient secretion of ANGPTL8 (115) and appears to require intracellular co-folding, as mixing ANGPTL3 and ANGPTL8 outside the cell does not result in complex formation (115, 138). In vivo, ANGPTL3-ANGPTL8 complexes form in the liver as inferred by the observations that almost all circulating ANGPTL8 comes from the liver (39) and that most circulating ANGPTL8 is complexed with ANGPTL3 (54). Importantly, in vitro several studies have found that the ability of ANGPTL3-ANGPTL8 complexes to inhibit LPL is orders of magnitude better than that of either ANGPTL3 or ANGPTL8 alone (54, 115, 137, 138), suggesting that these complexes are the physiologically relevant inhibitory unit.

The idea that the ANGPTL3-ANGPTL8 complex is the functional unit of LPL inhibition for both ANGPTL3 and ANGPTL8 is supported by several in vivo observations. For example, the necessity of the ANGPTL3-ANGPTL8 complex explains why deletion of either ANGPTL3 or ANGPTL8 results in lower plasma TGs (15, 16, 18, 36, 39, 139), but overexpression of ANGPTL8 in the absence of ANGPTL3, or vice versa, has limited ability to increase plasma triglycerides (19, 115, 137), as absence of either protein would preclude the formation of the complex and limit LPL inhibition. The requirement of the ANGPTL3-ANGPTL8 complex and the strong induction of ANGPTL8 by feeding also explain why disruption of ANGPTL3 primarily effects triglyceride metabolism and lipid partitioning in the fed state despite the fact that ANGPTL3 expression itself is not feeding induced (72, 106).

Recent data suggest that the ANGPTL3-ANGPTL8 functional complex is also the target of the ApoA5. ApoA5 is an atypical apolipoprotein that plays a critical role in regulating plasma triglycerides in mice and humans (144–151). However, the mechanism of its triglyceride-lowering action has remained unknown. Chen et al. (152) recently reported that ApoA5 binds ANGPTL3-ANGPTL8 complexes, but not ANGPTL3, ANGPTL4, or ANGPTL8 alone, and greatly reduces the ability of this complex to inhibit LPL.

The exact mechanism by which the formation of ANGPTL3-ANGPTL8 complexes allows efficient inhibition of LPL is not clear. Although ANGPTL8 was initially reported to promote cleavage of ANGPTL3 (19), follow-up studies showed that this was not the case (36, 39). Part of the increase in LPL inhibition can be attributed to the increased ability of the complex to bind LPL compared with ANGPTL3 alone (54, 115). ANGPTL8 has a domain homologous to the lipase inhibitory domain of ANGPTL3 and ANGPTL4 (137). Haller et al. (137) found that this domain was incapable of inhibiting LPL when ANGPTL8 was alone, but was necessary for LPL inhibition by ANGPTL3-ANGPTL8 complexes. They hypothesized that complex formation unmasks the LPL inhibitory domain of ANGPTL8 allowing ANGPTL3-ANGPTL8 complexes to efficiently inhibit LPL. The precise role of the LPL inhibitory domain in inhibition is not known, so it may be that one of the primary functions of this domain is to facilitate binding to LPL. Thus, unmasking of this domain might also explain why the complex has much higher affinity for LPL than either ANGPTL3 or ANGPTL8 alone. Although it seems likely, based on homology, that ANGPTL3-ANGPTL8 complexes catalytically inactivate LPL in a manner similar to ANGPTL4, such a mechanism has not yet been shown directly.

Interactions with ANGPTL4

Both ANGPTL8 and ANGPTL3 are expressed in the liver, but ANGPTL8 is also expressed in adipose tissue where ANGPTL3 is not expressed (15, 19, 35, 40, 72, 103, 139), suggesting an ANGPTL3-independent role for ANGPTL8. Recently, it has become clear that ANGPTL8 also interacts directly with ANGPTL4, but the consequences of this interaction are quite different from those of ANGPTL3-ANGPTL8 complexes. Like with ANGPTL3, ANGPTL4 can be coimmunoprecipitated with ANGPTL8 in both in vitro and in vivo samples (39, 54, 138), but the stoichiometry of ANGPTL4-ANGPTL8 complexes appears to be 1:1 (54). Unlike the situation with ANGPTL3, where ANGPTL8 greatly enhances LPL inhibition, ANGPTL4-ANGPTL8 complexes are substantially poorer at inhibiting LPL than ANGPTL4 alone (39, 54, 138). These observations suggest that not only is ANGPTL8 unnecessary for ANGPTL4-mediated inhibition, it actively suppresses inhibition by ANGPTL4. The mechanism by which ANGPTL8 suppresses ANGPTL4 inhibition is not yet clear. Oldoni et al. (39) found that ANGPTL8 reduced secretion of ANGPTL4, suggesting that ANGPTL8 reduced ANGPTL4 inhibition simply by reducing levels of ANGPTL4. Chen et al. (54), however, did observe secretion of ANGPTL4-ANGPTL8 complexes and both Chen et al. (54) and Kovrov et al. (138) found that the specific inhibitory activity of ANGPTL4-ANGPTL8 complexes was greatly less than that of ANGPTL4 alone. Interestingly, Chen et al. (54) found that this was not because these complexes could not bind LPL, in fact they bound tightly, but that binding of the complex did not lead to substantial inhibition. Moreover, the binding of ANGPTL4-ANGPTL8 complexes appeared to prevent ANGPTL3-ANGPTL8 complexes and uncomplexed ANGPTL4 from inhibiting LPL (54). The extent to which ANGPTL4-ANGPTL8 complexes regulate LPL activity in vivo remains to be determined. Upon feeding, ANGPTL8 expression is induced, but ANGPTL4 expression is suppressed. The in vivo half-life of ANGPTL4 protein is not known. As ANGPTL4-ANGPTL8 complexes can be immunoprecipitated in vivo (54), there appears to be at least some overlap of ANGPTL4 and ANGPTL8 protein. One possibility is that induction of ANGPTL8 serves to neutralize remaining ANGPTL4 inhibition as the body transitions from fasting to feeding.

REVISITING THE A3-4-8 MODEL OF LPL REGULATION

Although ANGPTL3, ANGPTL4, and ANGPTL8 all regulate lipid metabolism, they clearly differ in the location and timing of their expression and their functional activity (summarized in Table 2). In 2016, Ren Zhang (153) outlined what he termed the A3-4-8 model of LPL regulation. In this model, ANGPTL4 is induced by fasting, inhibiting LPL activity locally in adipose tissue, and thus diverting triglycerides to oxidative tissues. Conversely, in the fed state, ANGPTL8 is induced and activates ANGPTL3. Acting as an endocrine factor, ANGPTL3 then inhibits LPL activity in oxidative tissues, thus funneling triglycerides to adipose tissue. Five years later, this model continues to be supported by the data. Moreover, many of the mechanistic and physiological details have come into sharper focus (Fig. 1).

Table 2.

Comparison of ANGPTL3, ANGPTL4, and ANGPTL8

	ANGPTL3	ANGPTL4	ANGPTL8
Tissue expression	Liver	Widely expressed, particularly high in liver and adipose	Liver Adipose
Known targets	EL LPL (w/ANGPTL8)	LPL PL	LPL (w/ANGPTL3)
Feeding regulation	Not strongly regulated by feeding state	Fasting induced	Feeding induced
Likely mode of action	Endocrine	Paracrine/autocrine	Endocrine (liver) Paracrine/autocrine (adipose)

LPL, lipoprotein lipase.

Figure 1.

The ANGPTL3-4-8 model revisited. In the fasted state (left), ANGPTL4 is expressed in adipose tissue and inhibits lipoprotein lipase (LPL) locally, reducing LPL activity and fatty acid uptake in adipose. ANGPTL3 is secreted by the liver, but has little ability to inhibit LPL on its own. In the fed state (right), both ANGPTL3 and ANGPTL8 are expressed in the liver, forming a complex that is secreted into the circulation and can inhibit LPL in oxidative tissues. In the adipose tissue, ANGPTL4 expression is repressed and ANGPTL8 expression is induced, leading to formation of ANGPTL4-ANGPTL8 complexes that have little ability to inhibit LPL and may also protect LPL from inhibition by circulating ANGPTL3-ANGPTL8 complexes.

Fasted State

As described in detail above, ANGPTL4 in adipose tissue is induced within the first few hours of fasting. ANGPTL4 inhibits LPL by promoting its degradation within adipocytes and/or by inactivating LPL extracellularly. This results in a decrease in LPL activity and reduced adipose uptake of triglyceride-derived fatty acids from the circulation. ANGPTL3 is also expressed during fasting, but without ANGPTL8 it has little ability to inhibit LPL. Expression of ANGPTL8 under fasting conditions is limited. Thus, few ANGPTL3-ANGPTL8 complexes are secreted into the circulation. In the absence of ANGPTL3-ANGPTL8 complexes, LPL is not inhibited in oxidative tissues such as heart and skeletal muscle. The combination of ANGPTL4 inhibition in adipose tissue and the lack of systemic LPL inhibition by ANGPTL3-ANGPTL8 complexes results in the preferential uptake of triglyceride-derived fatty acid fuel in heart and skeletal muscle where it can be used to fuel, allowing the organism to continue to function in the fasted state. It is important to note that although triglyceride delivery to adipose tissue is slowed during fasting, it does not stop.

Fed State

In the fed state, ANGPTL8 is induced both in the liver and in the adipose tissue. In liver, ANGPTL8 forms a complex with ANGPTL3 and this complex is secreted into the circulation as an endocrine factor. Circulating ANGPTL3-ANGPTL8 complexes inhibit LPL in oxidative tissues, such as heart and skeletal muscle, blunting triglyceride uptake into these tissues and thus allowing more uptake into adipose tissue. Circulating ANGPTL3-ANGPTL8 complexes could potentially inhibit LPL in adipose tissue as well, however, there is a concurrent reduction in ANGPTL4-mediated inhibition that results in a net increase in LPL activity in adipose in the fed state. This reduction in ANGPTL4 inhibition is mediated both by a decrease in ANGPTL4 expression and by an increase in adipose ANGPTL8 expression. The induction of ANGPTL8 expression results in the formation of ANGPTL4-ANGPTL8 complexes, which greatly reduce the ability of ANGPTL4 to inhibit LPL. Moreover, ANGPTL4-ANGPTL8 complexes, which retain the ability to bind LPL, appear to protect LPL from inhibition by the much more potent ANGPTL3-ANGPTL8 complexes. In combination, these events increase LPL activity in adipose tissue, allowing increased partitioning of triglyceride-derived fatty acids to adipose tissue for storage. The tissue-specific roles of ANGPTL8 outlined by this model are clearly apparent in adipose- and liver-specific ANGPTL8 knockout mice. Oldoni et al. (39) found that mice lacking liver ANGPTL8 had significantly lower plasma triglycerides than wild-type mice in the fed state, consistent with the absence of the inhibitory ANGPTL3-ANGPTL8 complexes that normally form in the liver. Conversely, in the absence of adipose ANGPTL8, fed-state plasma triglycerides were elevated compared with wild-type mice, consistent with the ability of adipose ANGPTL8 to diminish ANGPTL4 and/or ANGPTL3-ANGPTL8 inhibition in fat depots (39).

Although the model as a whole has come into greater focus, there are still details that remain obscure. For example, fasting induces ANGPTL4 expression in tissues other than adipose, including liver and some oxidative tissues (40, 46, 47, 72). Unlike the situation in adipose tissue, lack of ANGPTL4 does not appear to increase triglyceride partitioning to these other tissues under fasting conditions (63), and the physiological significance of ANGPTL4 induction in these tissues during fasting remains unclear. The C-terminal fibrinogen-like domain of ANGPTL4 has been shown to stimulate adipose tissue lipolysis (25). Therefore, it is possible that fasting induction of ANGPTL4 in tissues like liver serves to increase mobilization of lipid stores. The molecular mechanism by which ANGPTL proteins interact with one another and with lipoprotein lipase remain a work in progress. Understanding these mechanisms and the physiological timeline over which these interactions occur will provide additional insights into feeding-state induced shifts in LPL activity and fat partitioning.

Thermogenesis

LPL-mediated shifts in fat partitioning are also important for nonshivering thermogenesis. With cold exposure, large amounts of triglyceride are funneled to brown adipose tissue in an LPL-dependent manner (68, 69, 71). Although there are many gaps in our understanding of this process, a model for how ANGPTL4, ANGPTL8, and ANGPTL3 participate in facilitating thermogenesis is beginning to take shape. With prolonged cold exposure, expression of ANGPTL4 decreases in BAT and increases in WAT (33, 70, 71). In BAT, this results in increased LPL activity in BAT, increasing uptake of triglycerides for thermogenic oxidation (71). At the same time, ANGPTL8 expression is induced in the brown adipose tissue during cold exposure (70). This would increase ANGPTL4-ANGPTL8 complex formation, further reducing ANGPTL4 inhibition and increasing LPL activity in BAT. Chen et al. (154) showed that ANGPTL4-ANGPTL8 complexes may actually stimulate LPL activity at 22°C. However, the relevance of this observation to thermogenesis is not clear as it seems unlikely that temperatures in BAT ever drop this low. ANGPTL3 likely also plays some role in thermogenesis, as ANGPTL3-ANGPTL8 double knockout mice have higher body temperatures and appear to be more sensitive to sympathetic stimulation than wild-type mice (21). However, the mechanisms and pathways by which ANGPTL3 contributes to thermogenesis have not been elucidated.

It is important to note that ANGPTL proteins are likely not the primary driver of the alterations in lipid partitioning that occur during thermogenesis. The alterations in LPL activity and lipid partitioning observed in ANGPTL4 knockout mice and mice overexpressing ANGPTL4 only occurred after sustained cold exposure, and even then, the cold-induced increase of triglyceride uptake in BAT still occurred even when ANGPTL4 was absent or when it was overexpressed (71). Moreover, even though cold exposure increased ANGPTL4 expression in white adipose tissue, the partitioning of triglycerides to these depots did not appear to decrease (71).

Chronic High-Fat Diet and Obesity

Chronic overfeeding and obesity can lead to shifts in fat partitioning, ectopic lipid deposition, and metabolic disease (155–157). Surprisingly, despite the importance of altered lipid partitioning to the onset of metabolic disease, the contribution of ANGPTL proteins to these obesity-induced alterations have not been well characterized. The ANGPTL-regulated shifts in triglyceride partitioning for feeding and fasting and for thermogenesis have been delineated in nonobese, chow-fed mice. Whether the same shifts occur in the setting of obesity is not clear. As noted above, early experiments testing high-fat feeding in ANGPTL4-deficient mice were hampered by the lethal chylous ascites, intestinal injury, and lymphatic nodule inflammation induced by high-fat feeding in these mice (59). These lethal phenotypes can be avoided with a diet high in unsaturated fatty acids (158). Using such a diet, Janssen et al. (84) found that ANGPTL4-deficient mice gained more weight but were more glucose tolerant than wild-type mice. This finding is consistent with the observation that overexpression of ANGPTL4 in mice impaired glucose intolerance (26). These studies suggested that with high-fat feeding, ANGPTL4 would normally direct triglycerides away from adipose, increasing ectopic lipid deposition and glucose intolerance. More recent studies with tissue-specific ANGPTL4 knockout mice have provided additional insights. Two independent studies found that when high-fat diet was fed to mice without adipose ANGPTL4, these mice had increased adipose triglyceride deposition and improved glucose tolerance compared with wild-type mice (31, 32). Interestingly, both studies found that this effect was lost with chronic (5+ mo) of high-fat feeding (31, 32). Interestingly, even though increased triglyceride uptake in adipose disappeared with chronic high-fat feeding, adipose LPL activity remained higher in adipose-specific ANGPTL4 knockout mice, suggesting that after chronic high-fat feeding, LPL activity is no longer the limiting factor for triglyceride deposition in adipose (32). How the presence or absence of ANGPTL3 or ANGPTL8 effects fat partitioning after chronic high-fat feeding has not been examined but may provide key insights into the progression of ectopic lipid deposition and metabolic disease.

ANGPTL PROTEINS IN DISEASE AND DISEASE TREATMENT

Metabolic diseases such as obesity, metabolic syndrome, and type 2 diabetes mellitus often present with elevated plasma triglycerides and elevated triglyceride levels represent a cardiovascular risk factor (159–162). Recent human genetic evidence, including data from individuals with ANGPTL deficiency, has strengthened the idea that lowering plasma triglycerides could be beneficial for treating dyslipidemia and cardiovascular disease. Like mice, humans deficient in ANGPTL3, ANGPTL4, or ANGPTL8 have lower plasma triglycerides (56–58, 163–166). In the case of ANGPTL3, loss of ANGPTL3 results not only in hypotriglyceridemia, but in familial combined hypolipidemia (108, 109, 116–118). Importantly, multiple studies found that individuals deficient in one of these three ANGPTL proteins have a lower risk of cardiovascular disease (56, 57, 163–166). Interestingly, there is evidence that loss of ANGPTL function can also improve glucose homeostasis (110, 163, 167).

The protective effects of ANGPTL loss of function mutations has prompted studies into whether targeting these proteins could be used to treat dyslipidemia and metabolic disease. Preclinical studies using antibodies or antisense oligos showed that targeting ANGPTL3, ANGPTL4, or ANGPTL8 can lower plasma triglycerides (29, 56, 60, 165, 168–172). Unfortunately, targeting ANGPTL4 with an anti-ANGPTL4 antibody in mice fed a high-fat diet led to intestinal abnormalities, chylous ascites, lymph node inflammation, and premature death (29), consistent with observations for ANGPTL4-deficient mice fed a high-fat diet (29, 59). Lymph node inflammation was also observed in cynomolgus monkeys treated with a monoclonal antibody against ANGPTL4 (56).

Similar detrimental effects have not been observed in preclinical studies targeting ANGPTL3. In fact, an Italian town known for the longevity of its residents was found to have an unusually large number of individuals with the S17X loss-of-function ANGPTL3 mutation (108, 118, 173). Therefore, ANGPTL3 has become a key target for drug therapies that treat dyslipidemia. In particular, two targeting approaches, monoclonal antibodies and antisense oligonucleotides, have shown therapeutic promise.

In initial studies, Evinacumab, a human anti-ANGPTL3 monoclonal antibody, lowered triglyceride levels by 76%, LDL-cholesterol by 23.2%, and HDL-cholesterol by 18.4% in healthy individuals (165). Further studies confirmed the ability of Evinacumab to lower plasma triglyceride levels in healthy individuals (174), hypertriglyceridemic subjects (175), and patients with familial hypercholesterolemia (176–178). As the ANGPTL3-ANGPTL8 complex is the functional unit of LPL inhibition, Evinacumab likely acts to lower plasma triglycerides by binding the ANGPTL3 of ANGPTL3-ANGPTL8 complexes. Indeed, Jin et al. (179) showed that Evinacumab could robustly block inhibition by ANGPTL3-ANGPTL8 complexes. Similar to ANGPTL3-deficient individuals (discussed above in ANGPTL3, Physiological Roles in Lipid Metabolism), Evinacumab may lower plasma LDL-C levels either by preventing inhibition of EL by ANGPTL3 (125) or by preventing inhibition of LPL by ANGPTL3-ANGPTL8 complexes (54). In February 2021, Evinacumab, under the trade name Evkeeza, was approved by the Food and Drug Administration for the treatment of patients with homozygous familial hypercholesterolemia.

Vupanorsen, an antisense oligonucleotide designed to prevent the expression of ANGPTL3 in hepatocytes, represents an independent approach for therapeutically targeting ANGPTL3. In initial studies, treatment with Vupanorsen was associated with a 33%–63% decrease in triglycerides as well as reductions in Apo-C III, remnant and total cholesterol, HDL-C, and LDL-C (169). Vupanorsen also lowered plasma triglycerides in patients with diabetes, hepatic steatosis, and hypertriglyceridemia (180). Although these results are promising, it remains to be seen if targeting ANGPTL3 can improve cardiovascular outcomes.

The cardiovascular protection and the lack of deleterious effects associated with ANGPTL3 deficiency (40, 108, 163, 165, 166) have led to the proposal of a third, more radical method of targeting ANGPTL3 in humans, gene editing (20, 181, 182). Chadwick et al. (20) used CRISPR-mediated base editing to induce ANGPTL3 deficiency in 5-wk-old mice and found that this gene editing reduced ANGPTL3 expression, plasma triglycerides, and total cholesterol levels in both normal and hyperlipidemic mice. However, the bioethics of gene editing in humans remains controversial and both the ability to successfully gene edit ANGPTL3 in adult humans and the ability of ANGPTL3 gene editing to improve cardiovascular outcomes without side effects remain unproven.

The therapeutic benefit of targeting ANGPTL4 and ANGPTL8 remains to be determined. Given the detrimental effects observed in mice and monkeys, targeting ANGPTL4 in humans certainly warrants caution, but it should be noted that thus far these detrimental effects have not been reported in humans with ANGPTL4 deficiency. If targeting ANGPTL4 or ANGPTL8 does prove safe, it is possible that the efficacy in combating metabolic disease that comes from targeting these proteins could be quite different. Lowering ANGPTL3, ANGPTL4, or ANGPTL8 would each be expected to lower plasma triglycerides, but as these proteins act at different times and at different places, systemic effects could vary substantially. For example, targeting ANGPTL4 could potentially increase adiposity and increase glucose tolerance, as it does in mice (31–33, 84), an effect that might not be observed when targeting ANGPTL3 or ANGPTL8.

OPEN QUESTIONS

Despite the rapid progress in ANGPTL research, there is much about their function, their interactions with each other, and their physiological roles that remain unknown. Some of these open questions and how they might be addressed are discussed below.

Many questions regarding the structure and mechanisms of ANGPTL interactions remain unanswered. Although the mechanism of ANGPTL4-mediated LPL inhibition is coming into focus (52, 97, 98), the mechanism by which ANGPTL3-ANGPTL8 complexes inhibit LPL remains to be elucidated. One recent study suggested that ANGPTL3-ANGPTL8 complexes promote furin-mediated cleavage of LPL, but the levels of cleavage observed were not sufficient to explain the level of LPL inhibition (179). The structural basis for ANGPTL8’s interaction with ANGPTL3 and ANGPTL4 is not known, nor is it known why ANGPTL8 has opposite effects on ANGPTL3 and ANGPTL4. Such studies would likely be aided by solved protein structures. However, only the C-terminal fibrinogen domains of ANGPTL3 and ANGPTL4 have been solved (183), leaving the structure of the more relevant N-terminal LPL interacting domains unknown.

The ANGPTL3-4-8 model outlined above details how the interactions between ANGPTL3, ANGPTL4, and ANGPTL8 modulate fat partitioning according to feeding state. There are, however, many more metabolic states that alter fat delivery and could potentially be regulated by ANGPTL proteins. Two of these, cold-induced thermogenesis and high-fat feeding were discussed above. Other states include exercise, starvation, and chronic inflammation. One physiological state that deserves far more attention is age. Triglyceride metabolism changes with age, but the mechanisms behind these changes are largely unknown (reviewed in Ref. 184). The studies that form the basis of the ANGPTL3-4-8 model were largely performed in young mice. However, the risk of metabolic disease increases with age, especially in connection with chronic high-fat feeding. Studies in adipose-specific ANGPTL4-deficient mice, wherein some of the benefits of ANGPTL4 deficiency were lost with age and high-fat feeding (31, 32), highlight the importance of investigating the intersection of age and diet. It is likely that ANGPTL proteins both contribute to and are affected by age-induced metabolic changes.

It is worth noting that despite the profound effects that ANGPTL3, ANGPTL4, and ANGPTL8 have on circulating lipids, deficiency for these proteins has only minor effects on body weight, particularly in animals fed a normal chow diet. In some studies, ANGPTL8-deficient mice fed normal had lower body weights than wild-type littermates (36), but ANGPTL3- and ANGPTL4-deficient mice fed normal chow appear to have normal body weights (16). This is not necessarily unexpected. Even strong disruptions of plasma triglyceride clearance and uptake do not necessarily impact body weight, especially on diets that are not high in fat. For example, GPIHBP1-deficient mice exhibit severe hypertriglyceridemia and have very little LPL-mediated triglyceride uptake (8). Yet, these mice have normal body weights on normal chow diets (185). Why profound changes in circulating triglycerides do not translate to similarly profound changes in body composition is not clear? Likely, compensatory mechanisms help maintain gross body composition. However, additional metabolic challenges may make underlying partitioning defects more apparent. For example, when ANGPTL4 knockout mice are fed a high-fat diet, they gain significantly more weight and adiposity than wild-type littermates (59, 84). Increased study of ANGPTL proteins in the context of different metabolic challenges, including disease states will help unravel the physiological consequences of misregulated triglyceride delivery.

Although this review has largely focused on ANGPTL regulation of lipoprotein lipase, other lipases are also targeted by ANGPTL proteins. ANGPTL4 may inhibit pancreatic lipase (81) and there are conflicting reports of its ability to inhibit hepatic lipase (16, 50). ANGPTL3 is a known inhibitor of endothelial lipase (123, 124). The ability of ANGPTL8 to alter ANGPTL3- or ANGPTL4-mediated inhibition of these other lipases is not yet known. As endothelial lipase, pancreatic lipase, and hepatic lipase all contribute to plasma lipoprotein metabolism, a complete understanding of ANGPTL3, ANGPTL4, and ANGPTL8 in metabolic homeostasis will require incorporating other lipase targets into the ANGPTL3-4-8 model.

In summary, ANGPTL3, ANGPTL4, and ANGPTL8 have emerged as critical regulators of lipid homeostasis and promising therapeutic targets. Investigating their critical role in metabolic health and disease and their interactions with each other continues to produce important insights into the regulation of lipid partitioning and lipid homeostasis.

GRANTS

This work was supported by grants from the National Institutes of Health [R01HL130146 and R01HL134787 (to B.S.J.D.)].

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

B.S.J.D. prepared figures; K.L.S-D. and B.S.J.D. drafted manuscript; K.L.S-D. and B.S.J.D. edited and revised manuscript; K.L.S-D. and B.S.J.D. approved final version of manuscript.