Critical role of SREBP-1c large-VLDL pathway in environment-induced hypertriglyceridemia of apoA-V deficiency

Abstract

Objective:

APOA5 variants are strongly associated with hypertriglyceridemia (HTG) as well as increased risks of cardiovascular disease and acute pancreatitis. HTG in apoA-V dysfunction often aggravates by environmental factors such as high-carbohydrate diets or aging. To date, the molecular mechanisms by which these environmental factors induce HTG are poorly defined, leaving the high-risk HTG condition undertreated. Previously, we reported that LXR-SREBP-1c pathway regulates large VLDL production induced by LXR agonist. However, the pathophysiological relevance of the finding remains unknown.

Approach and Results:

Here, we reconstitute the environment-induced HTG phenotype of human APOA5 deficiency in Apoa5^−/− mice and delineate the role of SREBP-1c in vivo by generating Apoa5^−/−;Srebp-1c^−/− mice. The Apoa5^−/− mice, which showed moderate HTG on a chow diet, developed severe HTG upon high-carbohydrate feeding or aging as seen in human apoA-V deficient patients. These responses were nearly completely abolished in the Apoa5^−/−;Srebp-1c^−/− mice. Further mechanistic studies revealed that in response to these environmental factors, SREBP-1c was activated to increase TG synthesis and to permit the incorporation of TG into abnormally large VLDL particles, which require apoA-V for efficient clearance.

Conclusions:

Severe HTG develops only when genetic factors (apoA-V deficiency) and environmental effects (SREBP-1c activation) coexist. We demonstrate that the regulated production of large-sized VLDL particles via SREBP-1c determines plasma TG levels in apoA-V deficiency. Our findings explain the long-standing enigma of the late-onset HTG phenotype of apoA-V deficiency and suggest a new approach to treat HTG by targeting genes that mediate environmental effects.

Graphical Abstract

Introduction

Hypertriglyceridemia (HTG) arises from the accumulation of TG-rich lipoproteins (TRLs) in the circulation ¹. Triglycerides (TGs) of exogenous and endogenous origins are secreted from intestine and liver in the form of chylomicrons (CMs) and very low density lipoproteins (VLDLs), respectively. These lipoprotein-TGs are hydrolyzed by lipoprotein lipase (LPL) in the circulation to liberate free fatty acids as energy sources for peripheral tissues. Severe HTG increases the risk of acute pancreatitis, while moderate HTG increases the risk of cardiovascular diseases (CVD). Recent genome-wide association studies uncovered an association between plasma TG levels and CVD risk with the HTG pathway genes, such as apolipoprotein C-III (APOC3), angiopoietin-like 3 (ANGPTL3), angiopoietin-like 4 (ANGPTL4), asialoglycoprotein receptor 1 (ASGR1), LPL, and apolipoprotein A-V (APOA5) ^2–4. In spite of these advances, effective therapeutic modalities to rescue HTG are limited, and a better understanding of plasma TG metabolism is warranted ⁵.

Severe HTG, defined as plasma TG levels greater than 1,000 mg/dl, can be classified into monogenic and polygenic types ⁵. Monogenic HTG (MIM 238600) is a rare (1 in 1,000,000) disorder with early-onset in infancy or childhood, and caused by monogenic defects in LPL pathway genes such as LPL, apolipoprotein C-II (APOC2), glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1 (GPIHBP1), LPL chaperone lipase maturation factor 1 (LMF1), and APOA5. Polygenic HTG (MIM 144650) is a more common (1 in 600), late-onset disorder, which typically results from the accumulation of multiple genetic variants in LPL pathway genes. In either type, the severe HTG develops by the combination of genetic plus environmental/secondary factors (e.g., high-carbohydrate, lipid-rich diets, aging, obesity, diabetes mellitus, and excess alcohol intake) ⁶. Previous work has demonstrated that the accumulation of VLDL in response to these factors contributes to the development of severe HTG ⁷. In vivo kinetic studies have further implicated the contribution of the increased production of large-sized VLDL particles^8,9. However, a molecular understanding of how environmental factors induce VLDL accumulation remains elusive in the sense that few studies have successfully identified a single gene that dominantly mediates the environmental effects of severe HTG.

APOA5 encodes apolipoprotein A-V (apoA-V), which is secreted from the liver and circulates on CMs, VLDLs and high-density lipoproteins (HDLs) ^10,11. ApoA-V reduces plasma TG levels primarily by stimulating the LPL-mediated plasma TG clearance ^12–15. ApoA-V binds both TRL and HSPG/GPIHBP1 that tethers LPL on endothelial surfaces, thereby facilitating the interaction between TRLs and LPL ^12–14. ApoA-V has alternative functions to enhance the whole cell uptake of TRLs via its ligand activities for lipoprotein receptors ^13,16,17, or to inhibit VLDL synthesis via its intracellular role to promote lipid droplet formation ^18,19. The dominant role of apoA-V in regulating plasma TG levels in humans has been documented from the severe HTG phenotype in APOA5 deficient patients (MIM 606368) ^20,21. As in severe HTG of other genetic causes ⁶, severe HTG in APOA5 deficiency often develops in response to environmental/secondary factors such as high-carbohydrate diets, lipid-rich diets, diabetes, and aging ^20–23. Molecular understanding of the environment-induced HTG in apoA-V deficiency could lead to the development of new therapeutic options to treat severe HTG by intervening the gene-environment interactions.

Mice deficient in apoA-V (Apoa5^−/−) has been described previously¹¹. Apoa5^−/− mice manifest only moderate HTG (~400 mg/dl) on a chow diet. In humans, APOA5 deficient patients manifest variable levels of plasma TG, and severe HTG (~6,000 mg/dl) often develops in response to environmental/secondary factors as discussed above^20–23. Although this discrepancy has been commonly interpreted as a species difference, we rather reasoned that Apoa5^−/− mice might serve as a useful model to identify the environmental/secondary factors that induce severe HTG and to explore its underlying mechanisms in vivo.

Here we thoroughly examined the environmental factors that induce severe HTG in human APOA5 deficient subjects in Apoa5^−/− mice. We found that severe HTG was induced in Apoa5^−/− mice either by (i) administration of T0901317, a synthetic pharmacological activator of liver X receptor (LXR); (ii) feeding with high-carbohydrate diets; or (iii) aging. These severe HTG responses were all prevented in Apoa5^−/− mice that lacked SREBP-1c (Apoa5^−/−;Srebp-1c^−/−). Further study demonstrated that the activation of SREBP-1c is required to increase TG synthesis and to permit the incorporation of TG into abnormally large VLDL particles, which require apoA-V for efficient clearance. The apoA-V-SREBP-1c interplay expands our understanding of the gene-environment interaction in the pathogenesis of type V HLP, suggesting the potential therapeutic approach to treat severe HTG by targeting the SREBP-1c large-VLDL pathway that mediate the environmental effects.

Methods

The authors declare that all supporting data are available within the article (and its Online Data Supplement).

Animals

Apoa5^−/− mice ¹¹ were obtained from Mutant Mice Regional Resource Centers (Stock #011467-UCD, FVB.129X1(B6)-Apoa5^tm1Hgc/Mmucd). Srebp-1c^−/− mice ²⁴ were obtained from the Jackson Laboratory. The Apoa5^−/−;Srebp-1c^−/− double knockout mice were generated by breeding Apoa5^−/− mice with Srebp-1c^−/− mice. C57BL/6J and Ldlr^−/− mice (B6.129S7-Ldlr^tm1Her/J) were obtained from the Jackson Laboratory. Vldlr^−/− mice ²⁵ and MX1 cre⁺;Lrp1^fl/fl mice ²⁶ have been described previously. Mice were used for the experiments with age and sex-matched controls on the same genetic background. Animals were housed in colony cages with a 12-h light/12-h dark cycle and fed a standard chow diet (Teklad Mouse/Rat Diet 7002 from Harlan Teklad Premier Laboratory Diets, Madison, WI; CLEA Rodent Diet CE-2, CLEA Japan). Male mice of 2-4 month old were used for experiments unless otherwise stated. Only male mice were used to avoid sex-related confounding effects. The T0901317 (J-Star Research, South Plainfield, NJ) treatment studies were carried out as described previously ²⁷. For dietary studies, a rodent diet high in fructose (69 kcal% fructose, 10 kcal% fat, 20 kcal% protein; #D08040107, Research Diets, Inc.) was used. For olive-oil gavage studies, 16-hour fasted mice were given oral gavage of olive oil (300 μl/body), after the pretreatment with intraperitoneal injection of normal saline or Triton WR-1339 (500 mg/kg-body weight) ²⁸. For all the animal studies, blood was drawn from the retro-orbital sinus with heparinized capillary tubes, and plasma was separated immediately for analysis. Animal experiments described in this manuscript have been approved and conducted under the oversight of the UT Southwestern Institutional Animal Care and Use Committee, and the Animal Care Committees of the University of Tokyo.

Biochemical Analyses

Plasma levels of cholesterol and triglycerides (TG) were measured enzymatically as described previously using kits ²⁹. Plasma levels of insulin were assayed with the mouse insulin enzyme-linked immuno-sorbent assay (ELISA) kits (Ultra Sensitive Rat Insulin ELISA Kit from Crystal Chem Inc.; Morinaga, Tokyo, Japan) as described ²⁹.

Fast Performance Liquid Chromatography

For fast performance liquid chromatography (FPLC) in Figure 1CD and Figure 5E, pooled plasma (1.7 ml) of each group of mice in Figure 1AB and Figure 5D, respectively, was subjected to ultracentrifugation at a density (d) of 1.215 g/ml. Plasma of all mice in the same group was pooled. The resulting lipoprotein fractions (d < 1.215) were adjusted to a final volume of 3 ml with a buffer containing 0.15 M sodium chloride, 0.01% (wt/vol) sodium EDTA, and 0.02% (wt/vol) sodium azide at pH 7.2, after which 2 ml was subjected to FPLC using a Superose 6B column. Fractions (2 ml) were collected, and 0.1 ml from each fraction was used to determine the content of total cholesterol and triglycerides ^27,30.

Figure 1. Plasma lipids and lipoproteins in wild-type and Apoa5^−/− mice fed a chow diet with or without LXR agonist (T0901317).

(A and B) Plasma lipid levels in wild-type (WT) and Apoa5^−/− mice (2-5 months old, male) fed an ad libitum chow diet or a diet supplemented with 0.015% T0901317 (T). Aliquots of blood were obtained by retro-orbital bleeding at the indicated time after the start of the diet and used for the measurement of plasma triglycerides (A) and cholesterol (B) levels. Each value represents the mean ± SEM of data from 5-7 mice. (C and D) FPLC profiles of plasma lipoproteins from WT (C) and Apoa5^−/− mice (D) fed a chow diet with (filled circles) or without (open circles) 0.015% T0901317 for 6 days. Plasma from 5-7 mice of each group was pooled and subjected to ultracentrifugation at d = 1.215 g/ml. The lipoprotein fractions (d < 1.215 g/ml) were subjected to gel filtration by FPLC, and the content of triglycerides in each fraction was measured as described in Methods. FPLC, fast performance liquid chromatography. (E-G) Electron microscopy of negatively stained plasma VLDL (d < 1.006 g/ml) from WT and Apoa5^−/− mice fed a chow diet with or without 0.015% T0901317 (T) for 6 days. Plasma samples from 4 mice of each group were pooled, and VLDL was isolated for viewing by electron microscopy (E) as described in Methods. Magnification × 100,000. Size distribution of VLDL particles from wild-type (F) and Apoa5^−/− (G) fed a chow diet with (filled bars) or without (open bars) 0.015% T0901317 for 6 days. The diameters of more than 300 VLDL particles from each group were measured, and the percentages of VLDL particles of different size were shown. Statistical significance was determined by two-way ANOVA with post hoc test. *P < 0.05, between genotypes. See also Figure I in the online-only Data Supplement.

Figure 5. Effects of olive oil gavage and aging on plasma lipids and lipoproteins in wild-type (WT), Apoa5^−/−, and Apoa5^−/−;Srebp-1c^−/− mice.

(A) Wild-type and Apoa5^−/− littermates (10-14 weeks old, male, n = 4-7) (upper) or Apoa5^−/− and Apoa5^−/−;Srebp-1c^−/− littermates (13-14 weeks old, male, n = 4) (lower) were pretreated with saline (left) or Triton WR-1339 (500 mg/kg-body weight) (right), followed by gavage with olive oil (300 μl/body). Aliquots of blood were obtained by retro-orbital bleeding at the indicated time points and used to determine the plasma levels of triglycerides. Each value represents the mean ± SEM of data from 4 mice. (B and C), Apoa5^−/− (8~17 weeks old (n = 25, male), 18~27 weeks old (n = 30, male), 28~ weeks old (n = 23, male)) were fed an ad libitum chow diet, and aliquots of blood were obtained by retro-orbital bleeding. Plasma levels of triglycerides (B) and cholesterol (C) were determined. Data are presented as the mean ± SEM. (D and E), Wild-type (WT) (n = 10, black), Apoa5^−/− (n = 7, red), and Apoa5^−/−;Srebp-1c^−/− littermates (n = 7, blue) (20-30 weeks old, male) were fed an ad libitum chow diet, and aliquots of blood were obtained by retro-orbital bleeding. Plasma levels of triglycerides were determined (D). Data are presented as the mean ± SEM. Pooled plasma from each group was subjected to FPLC analysis, and the content of triglycerides in each fraction was measured (E) as described in Methods. FPLC, fast performance liquid chromatography. Statistical significance was determined by two-way ANOVA with post hoc test (for A), Kruskal-Wallis test with post hoc test (B), and one-way ANOVA with post hoc test (for C and D). *P < 0.05 versus wild-type mice (for A) or versus all other groups (for B-D). See also Figure I in the online-only Data Supplement.

High Performance Liquid Chromatography

The size distribution of plasma lipoproteins was analyzed by gel-permeation high-performance liquid chromatography (GP-HPLC) system as described previously ³¹ (LipoSEARCH®, Skylight Biotech, Inc., Akita, Japan). Briefly, plasma lipoproteins were separated with tandemly connected gel permeation columns, and TG levels in the column effluent were measured and expressed in mV. Lipoproteins were classified on the basis of their particle sizes by component peak analyses with the Gaussian curve fitting technique. The ranges of eluting positions for major lipoproteins are as follows: CM or large VLDL (>80 nm), 15~19.1 ± 1.0 min; VLDL (30~80 nm), 19.1~22.7 ± 1.0 min; LDL (16~30 nm), 22.7~25.1 ± 1.0 min; HDL (8~16 nm), 25.1~28.0 ± 1.0 min. The lipoprotein fraction of VLDL size (30~80 nm) was further classified into 3 subclasses on the basis of their particle sizes as shown in the inset of the figure.

Apolipoprotein analyses of plasma VLDL

Plasma VLDL (d < 1.006 g/ml) was isolated from the pooled plasma of each group of mice by ultracentrifugation. The protein concentrations in VLDL were measured as described previously ²⁷. VLDLs (6 μg protein) from each group were delipidated with ethanol/ether (1:1), precipitated, and dissolved in SDS loading buffer. Equal amounts of VLDL protein (7.5 ng per lane) were subjected to 4-15% SDS gradient gels and immunoblot analyses using primary antibodies at a dilution of 1:1000 (anti-apoB and apoE antibody, #K23300R, Biodesign International, Saco, Maine, USA).

Electron Microscopy of Plasma VLDL

Plasma VLDL (d < 1.006 g/ml) was isolated from the pooled plasma of each group of mice by ultracentrifugation. VLDL fractions were negatively stained and visualized by electron microscopy. The diameters of VLDL particles were measured by ImageJ 1.39u software (NIH) as described previously ²⁷.

Quantitative Real-Time PCR Analysis

Total RNA was prepared from mouse livers using a TRIzol reagent (Invitrogen Corp., Tokyo, Japan) and subjected to quantitative real-time PCR as previously described ^24,29. The relative amounts of mRNAs were calculated using the comparative CT method. Mouse cyclophilin mRNA was used as the invariant control. Most of the primers for real-time PCR analysis were described previously ^24,27. The primers used for the measurement of individual mRNAs are available on request.

Recombinant Adenoviruses

Recombinant adenoviruses that express β-galactosidase (lacZ) or APOA5 under the control of the cytomegalovirus promoter were constructed using the pAd/cytomegalovirus/V5-DEST Gateway system (Invitrogen Corp.) as described previously ³². The recombinant adenoviruses were expanded in HEK293 cells, purified by cesium chloride ultracentrifugation, and stored in 10% (v/v) glycerol in phosphate-buffered saline at −80 °C. A total of 0.5 × 10¹¹ particles of lacZ virus (Ad-lacZ) or APOA5 virus (Ad-APOA5) were injected into the jugular vein of mice anesthetized with 60 mg/kg of sodium pentobarbital.

Hepatic VLDL-TG Secretion

Secretion rates of VLDL-TG in vivo were estimated by the intravenous administration of Triton WR-1339 as described previously ²⁸. Briefly, mice were fasted overnight (16 hours) prior to intravenous injection of Triton WR-1339 (500 mg/kg-body weight). Plasma was obtained by retro-orbital bleeding at the indicated time points after injection. Plasma TG levels were measured as described above.

The Effect of Heparin on CM Clearance

Apoa5^−/− mice (11-16 weeks old, male) were fasted for 16 hours, followed by gavage with olive oil (10 μl/g-body weight). Four hours after the gavage, mice were randomly assigned to a group injected with normal saline (100 μl/body) or a group injected with heparin (50 IU/100 μl/body). Aliquots of blood were obtained by retro-orbital bleeding at the indicated time points (0, 3, 15, 30, 120 min) and used to determine the plasma levels of triglycerides.

Measurement of Plasma Lipase Activities

Activities of Lipoprotein lipase (LPL) and hepatic lipase (HL) in post-heparin plasma were determined as described ^33,34. Post-heparin plasma was collected after injection of heparin (50 IU/100 μl/body) via the jugular vein at the indicated time points ³⁵. To eliminate the possible contamination of plasma VLDL-TG into the lipase assay reaction, plasma was ultracentrifuged at 48,000 rpm (204,782 g) for 16 hours to remove the VLDL fraction (top). An aliquot of the bottom fraction (equivalent to 5 μl of plasma) was then diluted in a final 100 μl normal saline solution and used as an enzymatic source in the lipase activity assay ³⁴. LPL activity was calculated by subtracting the HL activity (assayed in the presence of 1M NaCl) from the total lipase activity (assayed in the absence of 1M NaCl) as described ³⁴.

Statistical Analyses

All values are given as mean ± SEM. Statistical analyses were performed using PRISM 8 for Mac OS X (GraphPad Software, Inc., La Jolla, CA). Data were tested for normality and equal variance to confirm the appropriateness of parametric tests. Data that followed a normal distribution were analyzed by Student t test, Welch t test, one-way ANOVA with Tukey post hoc test, or two-way ANOVA with Tukey or Sidak post hoc tests, where appropriate. Data that failed normal distribution were analyzed by Mann-Whitney U test or by Kruskal-Wallis test with Dunn post hoc test. A P value of less than 0.05 was considered significant.

Results

LXR-agonist induce severe HTG accompanying the accumulation of large VLDL particles in Apoa5^−/− Mice

We first characterized the lipid and lipoprotein profiles of Apoa5^−/− mice under chow-fed conditions as well as in the pharmacological conditions where the production of large VLDL particles are stimulated by LXR agonist (T0901317). Production of large-sized VLDLs has been suggested as a prominent feature of environment-induced HTG^8,9, and we and others have previously demonstrated that the activation of SREBP1a³⁰, LXR³⁶, or LXR-SREBP-1c pathway²⁷, produces large-sized VLDL particles of chylomicron (CM) size (>80 nm), which massively accumulates in the plasma of Ldlr^−/− mice. The large-sized VLDL particles produced by LXR agonist has been shown to be of hepatic origin by primary hepatocyte study²⁷. Immunoblot analyses of apoB-48 and apoB-100 in VLDL proteins (Figure IA in the online-only Data Supplement) were compatible with the notion that LXR-induced large VLDL particles are mainly of hepatic origin, as the apoB-48 to apoB-100 ratio of VLDL particles were not increased, but decreased by T0901317 treatment.

On chow-fed condition, these mice exhibited a moderate increase in plasma TG (wild-type;chow, 201 ± 14 mg/dl; Apoa5^−/−;chow, 479 ± 62 mg/dl, P < 0.01) (Figure 1A, chow) and a slight non-significant increase in plasma cholesterol (wild-type;chow, 129 ± 11 mg/dl vs. Apoa5^−/−;chow, 154 ± 13 mg/dl) (Figure 1B, chow), which is consistent with the previous report ¹¹. Elevated levels of plasma TGs are confined to VLDL fraction as revealed by fast performance liquid chromatography (FPLC) (Figure 1, C and D, chow). Notably, the median size of plasma VLDL particles (d < 1.006) in Apoa5^−/− (39.2 nm) is slightly larger than in wild-type mice (36.5 nm) as determined by negative stain electron microscopy (Figure 1, E-G, chow), which is consistent with previous reports ^16,37.

Treatment of mice with T0901317 resulted in a robust increase in the mRNA expression of SREBP-1c and its target genes in both wild-type and Apoa5^−/− mice, which is consistent with previous reports ²⁷(data not shown). When Apoa5^−/− mice were treated with T0901317, they exhibited a massive increase in plasma TGs, reaching over 4,000 mg/dl (Figure 1A), accompanying a modest increase in plasma cholesterol levels (Figure 1B). In clear contrast, in the presence of apoA-V, wild-type mice treated with T0901317 exhibited an only modest increase in plasma TG (Figure 1A) and a mild increase in plasma cholesterol (Figure 1B). FPLC analysis revealed that the elevated TGs in Apoa5^−/− were contained in VLDL particles (Figure 1D). T0901317 increased the size of VLDL particles in wild-type mice, where the median size of VLDL particles increased from 36.5 to 49.6 nm (Figure 1, E and F)^27,36, which is more pronounced in Apoa5^−/− (39.2 to 58.2 nm) (Figure 1, E and G).

The percentage of the large-sized VLDL particles of CM size (>80 nm) was increased in wild-type (1.4% to 12.8%) (Figure 1F), which is again more pronounced in Apoa5^−/− mice (2.8% to 17.5%) (Figure 1G).

These results suggest that large-sized VLDL particles are removed from the plasma in the presence of apoA-V, but not efficiently in the absence of apoA-V.

The Large-Sized VLDL Particles Are Cleared by ApoA-V

To determine if apoA-V clears the large-sized VLDL particles from plasma, we overexpressed apoA-V in several models of large VLDL accumulation by using recombinant adenovirus carrying human APOA5 (Ad-APOA5). We first confirmed the TG-lowering effect of Ad-APOA5 in Apoa5^−/− mice. Severe HTG was induced by 6 days of T0901317 treatment in Apoa5^−/− mice (694 ± 122 mg/dl (chow) vs. 5540 ± 823 mg/dl (T0901317)), which was reversed 3 days after the injection of Ad-APOA5 but not after the control virus injection (Ad-lacZ, 4849 ± 977 mg/dl vs. Ad-APOA5, 485 ± 122 mg/dl, P < 0.05 by Welch t test). We next tested another model of large VLDL accumulation (Ldlr^−/−;T0901317) ²⁷. As shown in Figure 2, the overexpression of human APOA5 by recombinant adenovirus resulted in a rapid reduction of plasma TG levels (Figure 2A) accompanying a delayed reduction in plasma TC levels (Figure 2B) in Ldlr^−/−;T0901317 mice. The slower reduction of the plasma levels of cholesterol than triglycerides suggests the role of apoA-V in TG hydrolysis. Measurement of the VLDL size revealed a reduction in the percentage of large VLDL particles (>80 nm) by APOA5 overexpression (Ldlr^−/−;chow;Ad-lacZ, 0.2%; Ldlr^−/−;T0901317;Ad-lacZ, 12.6%; Ldlr^−/−;T0901317;Ad-APOA5, 2.4%) (Figure 2, C and D). Together with additional supportive evidence in Ldlr^−/−;Vldlr^−/− mice (Figure 2E) and in MX1-Cre⁺;Lrp1^fl/fl;Ldlr^−/− mice (Figure 2F), where LXR-agonist induced HTG was rescued by Ad-APOA5 with closely similar time courses as in Ldlr^−/− mice (Figure 2, A and B), these results demonstrate that apoA-V enhances the clearance of large VLDL particles produced by LXR agonist.

Figure 2. Effects of APOA5 overexpression on plasma lipids and VLDL size in Ldlr^−/−, Vldlr^−/−;Ldlr^−/−, and MX1-Cre⁺;Lrp1^fl/fl;Ldlr^−/− mice treated with T0901317.

(A and B) Male Ldlr^−/− mice (12-20 weeks old) were fed with a chow diet (black, open circles), or a chow diet containing 0.015% T0901317 (T) (filled circles). Six days after the start of the T0901317 treatment, mice were randomly allocated to two groups and injected via the jugular vein with 0.5 × 10¹¹ particles per mouse of recombinant adenoviruses encoding lacZ (Ad-lacZ, black, filled circles) or human APOA5 (Ad-APOA5, red, filled circles). Blood was obtained from retro-orbital plexus at the indicated time and used for measuring the plasma levels of triglycerides (A) and cholesterol (B). Each value represents the mean ± SEM of data from 4 mice. (C) Electron microscopy of negatively stained plasma VLDL (d < 1.006 g/ml). Plasma samples from 4 mice of each group were pooled, and VLDL was isolated for viewing by electron microscopy as described in Methods. Magnification x100,000. (D) The percentages of large VLDL particles (>80 nm). The diameters of more than 300 VLDL particles from each group were measured, and the percentages of large VLDL particles (>80 nm) were shown. (E and F) Male Vldlr^−/−;Ldlr^−/− mice (8-16 weeks old) (E), or male MX1-Cre⁺;Lrp1^fl/fl;Ldlr^−/− mice (12-20 weeks old) (F) were fed with a chow diet (black, open circles), or a chow diet containing 0.015% T0901317 (T) (filled circles). Six days after the start of the T0901317 treatment, mice were randomly allocated to two groups and injected via the jugular vein with 0.5 × 10¹¹ particles per mouse of recombinant adenoviruses encoding lacZ (Ad-lacZ, black, filled circles) or human APOA5 (Ad-APOA5, red, filled circles). Blood was obtained from retro-orbital plexus at the indicated time and used for measuring the plasma levels of triglycerides (upper) and cholesterol (lower). Each value represents the mean ± SEM of data from 4 mice. Statistical significance was determined by Mann-Whitney U test. * P < 0.05, versus control virus.

Rescue of LXR Agonist-Induced HTG by SREBP-1c Deficiency

To define the molecular mechanisms underlying the LXR-agonist induced accumulation of large VLDL particles in Apoa5^−/− mice, we asked if the deletion of SREBP-1c could rescue this response. We crossbred Srebp-1c^−/− mice with Apoa5^−/− mice to generate doubly mutant mice (Apoa5^−/−;Srebp-1c^−/−). We confirmed that the hepatic mRNA expression of SREBP-1c was abolished and those of SREBP-1c target genes were largely blunted in Apoa5^−/−;Srebp-1c^−/− mice (Table I in the online-only Data Supplement, chow). On a chow diet, the deletion of SREBP-1c almost completely normalized the plasma TG levels of Apoa5^−/− mice to the levels of wild-type mice (wild-type;chow, 146 ± 17 mg/dl; Apoa5^−/−;chow, 638 ± 173 mg/dl; Apoa5^−/−;Srebp-1c^−/−;chow, 144 ± 16 mg/dl) (Figure 3A). Strikingly, the LXR-agonist induced HTG in Apoa5^−/− was nearly completely rescued in Apoa5^−/−;Srebp-1c^−/− mice (wild-type;T0901317, 158 ± 33 mg/dl; Apoa5^−/−;T0901317, 2780 ± 836 mg/dl; Apoa5^−/−;Srebp-1c^−/−;T0901317, 200 ± 51 mg/dl) (Figure 3A). The size of plasma VLDL was increased by LXR agonist in the presence of SREBP-1c (wild-type and Apoa5^−/−), but not in the absence of SREBP-1c (Apoa5^−/−;Srebp-1c^−/−) (Figure 3, B and C). The percentage of large VLDL particles (>80 nm) of chylomicron size was increased by LXR agonist in the presence of SREBP-1c (wild-type, 0.7% vs. 3.7% (chow vs. T0901317); Apoa5^−/−, 2.8% vs. 19.2% (chow vs. T0901317)), but not in the absence of SREBP-1c (Apoa5^−/−;Srebp-1c^−/−, 2.0% vs. 0.4% (chow vs. T0901317)) (Figure 3, B and C). These results demonstrate the essential role of the SREBP-1c large-VLDL pathway in severe HTG induced by LXR agonist in apoA-V deficiency.

Figure 3. Effects of treatment with LXR agonist (T0901317) on plasma lipids and lipoproteins in wild-type (WT), Apoa5^−/−, and Apoa5^−/−;Srebp-1c^−/− mice.

(A) Male mice with the indicated genotype were fed a chow diet with (filled symbols) or without (open symbols) 0.015% T0901317 (T). Aliquots of blood were obtained by retro-orbital bleeding at the indicated time (0, 2, or 6 days) after the start of the diet and used to determine the plasma levels of triglycerides. Each value represents the mean ± SEM of data from 4 mice. (B) Electron microscopy of negatively stained plasma VLDL (d < 1.006 g/ml). Plasma samples from 4 mice of each group were pooled, and VLDL was isolated for viewing by electron microscopy as described in Methods. Magnification x100,000. (C) Size distribution of VLDL particles from wild-type (WT) (black, top), Apoa5^−/− (red, middle), and Apoa5^−/−;Srebp-1c^−/− (blue, bottom) fed a chow diet with (filled bars) or without (open bars) 0.015% T0901317 (T) for 6 days. The diameters of more than 300 VLDL particles from each group were measured, and the percentages of VLDL particles of different size were shown. Statistical significance was determined by two-way ANOVA with post hoc test. *P < 0.05 versus all other groups.

Carbohydrate-Induced HTG and Its Rescue by SREBP-1c Deficiency

The important unanswered question here is if the SREBP-1c large-VLDL pathway plays a dominant role in the development of severe HTG in pathophysiological conditions. To this end, we thoroughly tested for environmental factors that induce severe HTG in human APOA5 deficient patients ^22,23 (Figures 4 and 5). Figure 4 shows an experiment comparing the responses of wild-type, Apoa5^−/−, and Apoa5^−/−;Srebp-1c^−/− mice to high-fructose feeding over a 14 days period. When wild-type mice were fed a high-fructose diet (69 kcal% fructose), they showed a marginal increase in plasma TG as well as plasma cholesterol (Figure 4, A and B). In marked contrast, we discovered that Apoa5^−/− mice treated with a high-fructose diet exhibited a massive increase in plasma TGs, reaching ~3,000 mg/dl at day 14, accompanying a mild increase in plasma cholesterol (Figure 4A, P < 0.001, two-way ANOVA; Figure 4B, P < 0.001, two-way ANOVA).

Figure 4. Effects of high-fructose feeding on plasma lipids and lipoproteins in wild-type, Apoa5^−/−, and Apoa5^−/−;Srebp-1c^−/− mice.

(A-C) Wild-type (WT, black) and Apoa5^−/− (red) littermates (9-13 months old, male) (A and B), or Apoa5^−/− (red) and Apoa5^−/−;Srebp-1c^−/− (blue) littermates (9-17 months old, male) (C) were fed a chow diet (open circles) or a high-fructose diet containing 69 kcal% fructose (filled circles). Aliquots of blood were obtained by retro-orbital bleeding at the indicated time and used to determine the plasma levels of triglycerides and cholesterol. Each value represents the mean ± SEM of data from 4-5 mice. (D) Pooled plasma from each group of mice was subjected to gel-permeation high-performance liquid chromatography (GP-HPLC) analysis, and the content of triglycerides in each fraction was measured as described in Methods. Bold dashed lines and bold lines represent chow and high-fructose feeding groups, respectively, and thin lines represent ± SEM of data from 4-5 mice per each group. The inset shows the content of triglycerides in each fraction of lipoproteins: Apoa5^−/− (SKO, red) and Apoa5^−/−;Srebp-1c^−/− (DKO, blue) fed a chow diet (open circles) or a high-fructose diet containing 69% fructose (filled circles). Component peak analyses with the Gaussian curve fitting technique were used to estimate the size of lipoproteins as described in Methods. Fraction #1 contains lipoproteins larger than 80 nm in diameter; #2, #3, #4 contains lipoproteins with diameters in the range of 30~80 nm. The fraction of lipoproteins contains larger particles in the numerical order (#1 (>80 nm) > #2 > #3 > #4). Data are presented as the mean ± SEM. Statistical significance was determined by two-way ANOVA with post hoc test (for A-C) and one-way ANOVA with post hoc test (for the inset of D). *P < 0.05 versus wild-type mice (for A and B) or versus all other groups (for C and D). See also Figure SI and Table I in the online-only Data Supplement.

Next, we tested the role of SREBP-1c in this fructose-induced HTG of apoA-V deficiency. Previous work demonstrated that high-fructose feeding induces hepatic lipogenesis partially via SREBP-1c ^38–41. Consistently, fructose-feeding increased the mRNA levels of all fatty acid and triglyceride biosynthetic genes in Apoa5^−/− mice, which were partially suppressed in Apoa5^−/−;Srebp-1c^−/− mice (Table I in the online-only Data Supplement). Consequently, the high-fructose feeding increased the TG content in livers of Apoa5^−/− mice (7.2 ± 0.8 vs. 43.3 ± 6.4 mg/g-liver, chow vs. high-fructose, P < 0.001), which was again partially suppressed in Apoa5^−/−;Srebp-1c^−/− mice (4.1 ± 0.3 vs. 26.9 ± 1.9 mg/g-liver, chow vs. high-fructose, P < 0.001) (Table 1).

Table 1.

Effects of high-fructose feeding in wild-type, Apoa5^−/−, and Apoa5^−/−;Srebp-1c^−/− mice.

	Apoa5−/−		Apoa5−/−;Srebp-1c−/−
	Chow n=5	High fructose n=5	Chow n=4	High fructose n=4
Body weight (g)	29.9 ± 0.3	29.9 ± 1.2	28.4 ± 1.3	30.3 ± 1.6
Food intake (g/day)	4.6 ± 0.1	4.5 ± 0.1	4.8 ± 0.1	4.5 ± 0.2
Liver weight/body weight (%)	5.3 ± 0.1	6.5 ± 0.2 a	5.2 ± 0.1	5.9 ± 0.1 a
Liver triglyceride content (mg/g)	7.2 ± 0.8	43.3 ± 6.4 a	4.1 ± 0.3	26.9 ± 1.9 a,b
Liver cholesterol content (mg/g)	12.2 ± 0.5	14.8 ± 1.2	12.6 ± 0.6	17.5 ± 3.5
Plasma glucose (mg/dl)	162.2 ± 4.8	179.0 ± 20.2	169.5 ± 8.7	180.5 ± 21.6

Mice were fed an ad libitum chow diet or a high-fructose diet for 14 days. Each value represents the mean ± SEM of 4-5 values. One-way ANOVA with Tukey post hoc test was used to evaluate statistical significance

^abetween dietary treatments within the same genotype (P < 0.05) and

^bbetween genotypes within the same dietary treatment (P < 0.05).

The deletion of SREBP-1c produced the more pronounced effect on the plasma TG levels: SREBP-1c deficiency nearly completely rescued the fructose-induced HTG in Apoa5^−/− mice (Figure 4C, P < 0.0001, two-way ANOVA). The complete rescue of the HTG phenotype by SREBP-1c deficiency can be explained if SREBP-1c is essential for the production of large VLDL particles not only by LXR agonist treatment (Figure 3) ²⁷ but also by high-fructose feeding. Indeed, lipoprotein analyses by HPLC methods demonstrated that the high-fructose feeding induced the accumulation of large VLDL particles (>80 nm) in Apoa5^−/− mice, which is almost completely rescued in Apoa5^−/−;Srebp-1c^−/− mice (Figure 4D). Furthermore, the component peak analyses with the Gaussian curve fitting technique revealed that VLDLs of larger size were preferentially accumulated by high-fructose feeding in Apoa5^−/− mice (Figure 4D, Inset), which was nearly completely abrogated in Apoa5^−/−;Srebp-1c^−/− mice.

Importantly, the rate of VLDL secretion was not different between Apoa5^−/− and Apoa5^−/−;Srebp-1c^−/− mice on a chow diet (Figure IB in the online-only Data Supplement), suggesting that the basal rate of VLDL secretion is not controlled by SREBP-1c at least on a chow-fed condition.

In the experiments not shown, we tested the effect of high-glucose and high-sucrose feeding and found no increase in plasma TG levels in Apoa5^−/−;Srebp-1c^−/− mice by any of these high-carbohydrate diets (data not shown).

Collectively, these results reveal the essential role of the SREBP-1c large-VLDL pathway in severe HTG induced by various carbohydrates.

SREBP-1c Deficiency Does Not Affect CM Accumulation in Apoa5^−/− Mice

Another diet that induces severe HTG in human APOA5 deficiency is a lipid-rich diet. As shown in Figure 5A (top, left), olive oil gavage in wild-type mice resulted in a transient increase in plasma TG levels that returned to baseline after 8 hours, reflecting the production of CM particles followed by rapid clearance. In marked contrast, olive-oil gavage in Apoa5^−/− mice resulted in a massive and persistent increase in the plasma TG levels, reaching over 5,000 mg/dl after 8 hours (Figure 5A, top, left panel, P < 0.0001 by 2-way ANOVA). When these mice were pretreated with Triton WR-1339, an inhibitor of LPL-mediated TG lipolysis, olive oil gavage induced CM accumulation both in wild-type and Apoa5^−/− mice at similar rates to similar levels, demonstrating the role of apoA-V in LPL-mediated CM-TG clearance (Figure 5A, top, right panel). Previous work suggested that apoA-V stimulates LPL pathway by enhancing the interaction between TRL and LPL via its dual binding affinity for TRL and GPIHBP1/HSPGs that tether LPL on endothelial cell surfaces ^12–14. If this is the case, one would expect that heparin injection would rescue TRL accumulation in apoA-V deficiency by releasing LPL from endothelial cell surfaces into circulation and restoring the interaction between LPL and TRL. As expected, the administration of heparin to olive-oil fed Apoa5^−/− mice rapidly cleared pre-accumulated CM-TG (Figure IC in the online-only Data Supplement, P < 0.0001, two-way ANOVA). Heparin injection also cleared pre-accumulated VLDL-TG in Apoa5^−/− mice (Figure ID in the online-only Data Supplement). Importantly, heparin-releasable LPL activity was not decreased in Apoa5^−/− compared to WT mice (Figure IE in the online-only Data Supplement). These results support the notion that apoA-V deficiency decreases the substrate (TRL) -enzyme (LPL) interaction but not the amount of enzyme (the heparin-releasable activity of LPL protein). We next tested the role of SREBP-1c in CM metabolism of Apoa5^−/− mice. In marked contrast to the rescue of VLDL accumulation in Apoa5^−/− by SREBP-1c deficiency (Figures 3 and 4), the rates of CM accumulation were not affected by SREBP-1c deletion either in the presence or absence of Triton WR-1339 (Figure 5A, lower panel), indicating that SREBP-1c deficiency affects neither clearance nor production of CMs after a fat load. These results suggest that SREBP-1c controls VLDL levels at least partially by regulating the size of VLDL particles, without directly affecting CM production or the CM/VLDL removal system.

Age-Related HTG in Apoa5^−/− Mice and Its Rescue by the SREBP-1c Deletion

Lastly, we tested whether aging aggravates HTG in Apoa5^−/− mice as is often observed in human APOA5 deficiency ^22,23, and, if so, whether the deletion of Srebp-1c could rescue the response. As shown, the age-related severe HTG phenotype in human APOA5 deficiency (~ 6,000 mg/dl) was successfully reconstituted in Apoa5^−/− mice (Figure 5B). In Apoa5^−/− mice, aging severely increased the plasma TG levels (8~17 weeks old, 959 ± 118 mg/dl; 18~27 weeks old, 2019 ± 193 mg/dl; 28~ weeks old, 3362 ± 388 mg/dl, Spearman r = 0.6715, P < 0.0001) (Figure 5B) and moderately increased the plasma TC levels (8~17 weeks old, 231 ± 13 mg/dl; 18~27 weeks old, 328 ± 16 mg/dl; 28~ weeks old, 410 ± 29 mg/dl, Spearman r = 0.6059, P < 0.0001) (Figure 5C). The plasma insulin levels also increased along with age (Figure IF in the online-only Data Supplement) (8~17 weeks old, 0.89 ± 0.12 ng/ml; 18~27 weeks old, 2.63 ± 0.43 ng/ml; 28~ weeks old, 3.65 ± 0.82 ng/ml, Spearman r = 0.6806, P < 0.0001). Notably, the plasma levels of TG correlate well with those of insulin in Apoa5^−/− mice (Figure IH in the online-only Data Supplement; Spearman r = 0.7121, P < 0.0001). Inasmuch as age-associated hyperinsulinemia (Figure IF in the online-only Data Supplement) activates SREBP-1c ^42,43, the age-related HTG in apoA-V deficiency would be mediated by SREBP-1c. Indeed, the deletion of SREBP-1c nearly completely abolished the age-related accumulation of TG in VLDL fraction in apoA-V deficiency (WT, 162 ± 11 mg/dl, Apoa5^−/−, 2225 ± 303 mg/dl, Apoa5^−/−;Srebp-1c^−/−, 286 ± 29 mg/dl, P < 0.0001) (Figure 5, D and E). We isolated plasma VLDL from each group of aged mice in Figure 5DE, and measured the size of VLDL particles. As in younger mice (Figure 1 and 3), plasma VLDL particles of Apoa5^−/− mice (median size, 39.9 nm; percentage of VLDL > 80 nm, 4.0%) was larger than that of WT mice (36.5 nm and 1.6%, respectively), and this phenotype was abolished in Apoa5^−/−;Srebp-1c^−/− mice (30.3 nm and 0.4%, respectively), suggesting that the SREBP-1c large-VLDL pathway may at least in part play a role in the rescue of age-related HTG by SREBP-1c deficiency.

Together, these works clearly demonstrate the pivotal role of the SREBP-1c large-VLDL pathway in environment-induced severe HTG of apoA-V deficiency (Figure 6).

Figure 6. Schematic representation of the pivotal role of the SREBP-1c large-VLDL pathway in environment-induced HTG.

(A-C) ApoA-V is required for the clearance of large TRL particles. By biding both TRLs and GPIHBP1/HSPG that tethers LPL, apoA-V facilitates the interaction between TRLs and LPL and stimulates LPL-mediated TRL-TG clearance. Under a normal non-stressed situation, TRL particles are scarce, resulting in only moderate increases in plasma TGs in apoA-V deficiency (A). Under metabolic stresses that increase the production of TRLs (B, C), these particles accumulate in the absence of apoA-V. SREBP-1c is essential for the accumulation of large VLDLs (B), but not CMs (C), via its essential role in the production of large VLDL particles. LPL, lipoprotein lipase; GPIHBP1, glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1; HSPG, heparan sulfate proteoglycan.

Discussion

In this report, we identify the pivotal role of the SREBP-1c large-VLDL pathway in the environment-induced HTG using a disease mouse model. We found that SREBP-1c is required for the accumulation of VLDLs, but not CMs, via its specific role in large VLDL formation. Our main findings are summarized in Figure 6. Without environmental stress, VLDLs and CMs are scarce (Figure 6A). Under conditions where SREBP-1c is activated (LXR activation, high-carbohydrate diets, hyperinsulinemia, etc.), large VLDL particles are increasingly produced via SREBP-1c. As large TRL particles (large VLDL, CM) require apoA-V for efficient clearance (Figures 1–5 and Figure I in the online-only Data Supplement), these large VLDLs massively accumulate in Apoa5^−/− (Figure 6B). Without SREBP-1c, large VLDL particles are not produced, resulting in clear rescue of severe HTG (Figure 6B). On the other hand, SREBP-1c does not affect CM metabolism (Figure 6C).

A number of genes have been identified as causative genes of severe HTG (LPL, APOC2, GPIHBP1, LMF1, and APOA5) ⁵. However, little is known about the genes that mediate environmental effects. The major drawback to investigate the gene-environment interactions has been the lack of suitable animal models. Although several different genetic mouse models of severe HTG have been described previously (e.g., Lpl^−/−44, Lmf1 mutant mice ⁴⁵, Gpihbp1^−/−46, and Apoc2 hypomorphic mice ⁴⁷), these mice manifest severe HTG without environmental stress ^46,47 or even neonatal death due to severe HTG at birth ^44,45,47. As a suitable experimental model to study gene-environment interactions of severe HTG, we focused on apoA-V deficient mice, which are unique among other mouse models in that they are neither lethal nor severely hypertriglyceridemic at a younger age ¹¹. By using Apoa5^−/− mice, we successfully reconstituted the severe type V HTG phenotype of human APOA5 deficiency in response to environmental stimuli, which enables us to delineate the underlying molecular mechanisms.

Severe HTG is still an orphan disease with limited therapeutic options ⁵. Newly developing approaches aim to target LPL pathway genes by inhibition (e.g., antisense or monoclonal antibody against APOC3 or ANGPTL4) or overexpression (e.g., adeno-associated virus (AAV) for LPL). However, these modalities are costly and still under development before wide and safe clinical use. Targeting apoA-V is promising: APOA5 variants are strongly associated with HTG as well as CAD risks ^48,49 and multiple studies in mice have demonstrated the potent TG-lowering and athero-protective effects of apoA-V, albeit by overexpression methods (e.g., recombinant proteins, adenoviruses, AAV or transgenic animals) ^18,50 (Figure 2). Our results offer a novel approach to target apoA-V pathway by inhibiting gene(s) that mediate the environmental effect (Figure 6). Current therapy for environment-induced HTG is only aimed at controlling environmental factors, which is often difficult to attain (strict fat/carbohydrate restriction, adequate control of diabetes mellitus) or impossible to manipulate (aging). Inhibition of SREBP-1c large-VLDL pathway provides a translatable approach to alleviate these environmental effects. Theoretically, reduction of VLDL by SREBP-1c inhibition may, in turn, decrease the level of CM, inasmuch as VLDL and CM compete for a common, saturable, plasma TG removal system of LPL pathway ⁶.

Our studies may have relevance not only to severe HTG but also to more common types of HTG such as postprandial or atherogenic dyslipidemia. Previous studies in humans have shown the association of APOA5 variants with postprandial HTG after a high-fat meal ^51,52. The indispensable role of apoA-V in postprandial TG clearance is clearly demonstrated in our data (Figures 4, 5, and S1). Variants of APOA5 has also been shown to be associated with atherogenic dyslipidemia, which is characterized by a triad of HTG, low HDL-C, and a predominance of small, dense LDL ^53,54. Our data suggest that the accumulation of large VLDLs in the face of apoA-V dysfunction (Figures 1–4) may increase small, dense LDLs, inasmuch as large VLDLs are the source of small, dense LDLs ⁹. The well-known association between metabolic syndrome and small, dense LDLs ⁵⁵ may be explained by the activation of SREBP-1c in metabolic syndrome ^42,43. Inhibition of SREBP-1c has therapeutic potential for these common, atherogenic type of dyslipidemia.

The current studies advance our understanding of the molecular pathophysiology of environment-induced severe HTG. In vivo kinetics studies in humans have indicated that the overproduction of large VLDL is a prominent feature of environment-induced severe HTG^8,9. Molecular studies have identified several genes that control large VLDL production including LXR, SREBP-1a, and SREBP-1c^27,30,36. However, whether these molecular pathways control HTG in physiological conditions has not been tested. Our data for the first time demonstrate that a single gene (SREBP-1c) plays an essential role in environment-induced severe HTG using a disease model of apoA-V deficiency.

Inhibitors of SREBP is currently under clinical development⁵. CAT-2003 and CAT-2054 are conjugates of eicosapentaenoic acid and niacin and inhibit the maturation of both SREBP-1 and SREBP-2⁵⁶. These compounds are now tested for its efficacy and safety for treating hypercholesterolemia as well as HTG. More specific inhibitors of SREBP-1c might be more effective in treating severe HTG, which deserves further study.

Our study may answer another long-standing enigma of apoA-V: how can an apolipoprotein with much lower plasma concentration than VLDL plays a dominant role in VLDL metabolism ⁵⁷. The plasma concentration of apoA-V (~100-400 ng/ml) ^58,59 is estimated as only 4% of that of total VLDL particles in plasma ⁵⁷. This enigma can be solved, providing that apoA-V is required primarily for the clearance of large VLDL particles of chylomicron size (> 80 nm) (Figures 1–4). As large VLDL particles (> 80 nm) constitute only a proportion of total VLDL particles (estimated as < 3% of total VLDLs in our data) (Figures 1–3), the concentrations of apoA-V (~4% of total VLDLs) and large VLDLs (~3% of total VLDLs) are within the same range. As a molar basis, Merkel et al. estimated that about 1 apoA-V molecule is present in 24 VLDL particles. Given that large VLDLs constitute ~3% of total VLDLs (Figures 1–3), 1 apoA-V molecule is present in ~1 large VLDL particle.

As to the mode of action of apoA-V, our results are compatible with the notion that apoA-V enhances LPL action by facilitating the interaction between TGRL and LPL. However, our results do not preclude the possibility that apoA-V enhances whole particle uptake of TGRL^13,16,17, or inhibits VLDL secretion via its intracellular role to promote lipid droplet formation ^18,19.

Our study may have several limitations. First, the results presented were performed in male mice, and further validation in female mice is required to assess the possible effect of sex-related confounding factors that affect lipid metabolism. Second some studies used relatively small numbers of mice (n=4-5) per each group, although we repeated experiments to confirm the results. Third, the underlying molecular mechanisms on the clear rescue of age-related HTG by SREBP-1c deficiency may involve not only large VLDL formation but also other factors such as VLDL secretion rate, which deserves further study. Fourth, our data are compatible with the role of SREBP-1c in large VLDL formation but do not preclude the possibility that SREBP-1c may play a role in CM formation as well. Further study in tissue-specific mouse models is required to rule out this possibility. Fifth, although our data are consistent with the model whereby apoA-V clears large VLDL of chylomicron size, the direct proof by in vitro biochemical approaches requires further study.

In conclusion, our studies uncover a pivotal role of the SREBP-1c large-VLDL pathway in environment-induced severe HTG of type V HLP. The molecular dissection of gene (apoA-V)-environment (SREBP-1c-mediated large VLDL production) interactions not only expands our understanding of plasma TG homeostasis and HTG pathogenesis but also offers a brand-new approach to treat severe HTG by inhibiting genes that mediate environmental effects.

Highlights:

• Environment-induced HTG in human APOA5 deficiency was successfully reconstituted in Apoa5^−/− mice.

• The deletion of SREBP-1c in Apoa5^−/− mice nearly completely rescued severe HTG induced by carbohydrates or aging.

• In response to environmental stimuli, SREBP-1c is activated to produce large VLDL particles, which are cleared from plasma in the presence of apoA-V but massively accumulate in apoA-V deficiency.

• Targeting a single gene (SREBP-1c) that mediates environmental effects may be an effective approach to treat severe HTG.

Nonstandard Abbreviations and Acronyms

SREBP

sterol regulatory element-binding protein

VLDL

very low density lipoprotein

HDL

high-density lipoprotein

apoA-V

apolipoprotein A-V

HTG

hypertriglyceridemia

TG

triglyceride

TRL

triglyceride-rich lipoprotein

CM

chylomicron

LPL

lipoprotein lipase

CVD

cardiovascular disease

GPIHBP1

glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1

LMF1

LPL chaperone lipase maturation factor 1

LXR

liver X receptor