Muscle-derived angiopoietin-like protein 4 is induced by fatty acids via peroxisome proliferator-activated receptor (PPAR)-delta and is of metabolic relevance in humans

Abstract

Objective: Long-chain fatty acids (LCFAs) contribute to metabolic homeostasis in part via gene regulation. This study's objective was to identify novel LCFA target genes in human skeletal muscle cells (myotubes).

Research design and methods: In vitro methods included culture and treatment of human myotubes and C2C12 cells, gene array analysis, real-time RT-PCR, Western blotting, ELISA, chromatin immunoprecipitation, and RNA interference. Human subjects (two cohorts) were characterized by oral glucose tolerance test, hyperinsulinemic-euglycemic clamp, magnetic resonance imaging and spectroscopy, and standard blood analyses (glucose, insulin, C-peptide, and plasma lipids).

Results: We show here that ANGPTL4 (encoding angiopoietin-like protein 4) represents a prominent LCFA-responsive gene in human myotubes. LCFA activated peroxisome proliferator-activated receptor (PPAR)-delta, but not PPAR-alpha or -gamma, and pharmacological activation of PPAR-delta markedly induced ANGPTL4 production and secretion. In C2C12 myocytes, knockdown of PPARD, but not of PPARG, blocked LCFA-mediated ANGPTL4 induction, and LCFA treatment resulted in PPAR-delta recruitment to the ANGPTL4 gene. In addition, pharmacological PPAR-delta activation induced LIPE (encoding hormone-sensitive lipase), and this response crucially depended on ANGPTL4, as revealed by ANGPTL4 knockdown. In a human cohort of 108 thoroughly phenotyped subjects, plasma ANGPTL4 positively correlated with fasting nonesterified fatty acids (P = 0.0036) and adipose tissue lipolysis (P = 0.0012). Moreover, in 38 myotube donors, plasma ANGPTL4 levels and adipose tissue lipolysis in vivo were reflected by basal myotube ANGPTL4 expression in vitro (P = 0.02, both).

Conclusions: ANGPTL4 is produced by human myotubes in response to LCFA via PPAR-delta, and muscle-derived ANGPTL4 seems to be of systemic relevance in humans.

FIG. 1.

ANGPTL4 production by human myotubes and its regulation by LCFAs. Cells were treated for 20 h with 1.25% BSA (control for myristate, palmitate, palmitoleate, oleate, and linoleate), 2.5% BSA (control for stearate and palmitate + linoleate), or 0.5 mmol/l of each LCFA. A and C: Induction of the PPAR target genes ANGPTL4 (A) and PDK4 (C) by LCFA. RNA was quantified by real-time RT-PCR (relative arbitrary units [RAU]). B: Intracellular ANGPTL4 protein levels. ANGPTL4 protein was measured by ELISA and normalized to cellular protein contents. Statistics: P = 0.0008 (A), P = 0.0133 (B), and P < 0.0001 (C); ANOVA; n ≥ 4; *significantly different from 1.25% BSA (post hoc P < 0.05); **significantly different from 2.5% BSA (post hoc P < 0.05).

FIG. 2.

Induction of ANGPTL4 production in human myotubes by pharmacological PPAR-δ activation. A: Induction of ANGPTL4 expression by isoform-specific PPAR agonists. Cells were treated for 20 h with 0.1% DMSO (carrier control), 10 μmol/l Wy-14,643, 60 μmol/l fenofibrate, 10 μmol/l troglitazone, 1 μmol/l rosiglitazone, or 1 μmol/l GW501516. RNA was quantified by real-time RT-PCR. Statistics: P < 0.0001; ANOVA; n ≥ 4; *significantly different from DMSO (post hoc P < 0.05). B and C: Time-dependent induction of ANGPTL4 (B) and PDK4 (C) by GW501516. Cells were treated for 48 h with 0.1% DMSO or 1 μmol/l GW501516. RNA was quantified by real-time RT-PCR. Statistics: *significant differences between treatment groups over time: P < 0.0001 (B) and P = 0.0021 (C); time versus treatment; MANOVA; n = 3. D: Time-dependent stimulation of ANGPTL4 secretion by GW501516. Cells were treated for 48 h with 0.1% DMSO or 1 μmol/l GW501516. ANGPTL4 secreted into the culture supernatant was quantified by ELISA. Statistics: *significant differences between treatment groups over time: P < 0.0001; time versus treatment; MANOVA; n = 3. E: Cleavage of secreted ANGPTL4 during GW501516 treatment. Cells were treated for 48 h with 1 μmol/l GW501516. Supernatants conditioned by human myotube cultures derived from two donors were subjected to immunoblot analysis.

FIG. 3.

Role of PPAR-δ in LCFA-induced ANGPTL4 expression in C2C12 myocytes. A: Regulation of C2C12 ANGPTL4 expression by LCFAs. Cells were treated for 20 h with 1.25% BSA, 2.5% BSA, or 0.5 mmol/l of each LCFA. RNA was quantified by real-time RT-PCR (relative arbitrary units [RAU]). Statistics: P = 0.0012; ANOVA; n = 3; *significantly different from 1.25% BSA (post hoc P < 0.05); **significantly different from 2.5% BSA (post hoc P < 0.05). B: Time-dependent induction of ANGPTL4 by GW501516. Cells were treated for 20 h with 0.1% DMSO or 1 μmol/l GW501516. RNA was quantified by real-time RT-PCR. Statistics: *significant differences between treatment groups over time: P < 0.0001; time versus treatment; MANOVA; n = 3. C: Knockdown of C2C12 PPARD expression by RNAi. Cells were left untreated (basal) or were treated for 8 h with siRNA directed against bacterial luciferase (for control) or PPARD, respectively. Cells were lysed after siRNA washout and incubation with fresh medium for 16 h. RNA was quantified by real-time RT-PCR. Statistics: P < 0.0001; ANOVA; n = 3; *significantly different from basal (post hoc P < 0.05). D: Oleate-induced ANGPTL4 expression of C2C12 cells after PPARD knockdown. Cells were left untreated or were treated for 8 h with siRNA directed against bacterial luciferase (for control) or PPARD, respectively. After siRNA washout and 16-h incubation with fresh medium, cells were treated for 20 h with 1.25% BSA or 0.5 mmol/l oleate. RNA was quantified by real-time RT-PCR. Statistics: P = 0.0217; ANOVA; n = 3; *significantly different from BSA (post hoc P < 0.05); **significantly different from oleate + control siRNA (post hoc P < 0.05). E: Oleate-induced recruitment of PPAR-δ to the ANGPTL4 gene. Cells were treated for 6 h with 1.25% BSA or 0.5 mmol/l oleate, respectively. After cross-linking with formaldehyde, cells were subjected to anti–PPAR-δ ChIP. The co-immunoprecipitated DNA was analyzed for the presence of ANPTL4 DNA using PCR amplification of a 310-bp fragment harboring the PPRE in intron 3.

FIG. 4.

ANGPTL4-dependent LIPE expression in C2C12 myocytes. A and B: Time-dependent induction of C2C12 LIPE (A) and PNPLA2 (B) by GW501516. Cells were treated for 48 h with 0.1% DMSO or 1 μmol/l GW501516. RNA was quantified by real-time RT-PCR (relative arbitrary units [RAU]). Statistics: *significant differences between treatment groups over time: P = 0.0007 (A) and P = 0.0002 (B); time versus treatment; MANOVA; n = 3. C: Knockdown of C2C12 ANGPTL4 expression by RNAi. Cells were treated for 8 h with 1 μmol/l GW501516 alone (control) or with GW501516 in combination with siRNA directed against bacterial luciferase or ANGPTL4, respectively. Cells were lysed after siRNA washout and incubation with fresh GW501516-containing medium for 16 h. RNA was quantified by real-time RT-PCR. Statistics: P < 0.0001; ANOVA; n = 3; *significantly different from control (post hoc P < 0.05). D: C2C12 LIPE expression after ANGPTL4 knockdown. Cells were treated as described above (see C). RNA was quantified by real-time RT-PCR. Statistics: P = 0.0063; ANOVA; n = 3; *significantly different from DMSO (post hoc P < 0.05); **significantly different from GW501516 + control siRNA (post hoc P < 0.05).

FIG. 5.

Association of ANGPTL4 with lipid metabolism in humans. A and B: Association of human plasma ANGPTL4 with fasting NEFA (A) and WAT lipolysis (B). Plasma ANGPTL4 and NEFA were determined in 108 subjects. A: Fasting plasma NEFAs were adjusted for sex, age, and BMI. B: As an estimate of WAT lipolysis, the area under the curve (AUC) of NEFA during OGTT was used and adjusted for sex, age, BMI, and the AUC of insulin during OGTT. Adjustments were achieved by multivariate linear regression modeling. C, D, and E: Association of basal human myotube ANGPTL4 expression with basal PPARD expression (C), WAT lipolysis of the donors (D), and plasma ANGPTL4 levels of the donors (E). RNA was quantified by real-time RT-PCR (relative arbitrary units [RAU]). Glycerol and ANGPTL4 levels were measured in plasma. Data derived from 38 human donors are plotted. C and E: Unadjusted data are shown. In D, as an estimate of WAT lipolysis, the AUC of glycerol during OGTT was used and adjusted for sex, age, BMI, and the AUC of insulin during OGTT. Adjustments were achieved by multivariate linear regression modeling.

FIG. 6.

Hypothetical model of mANGPTL4's role in lipid metabolism. In states of increased muscle PPAR-δ activity and/or PPARD expression, such as fasting and exercise, SKM produces and secretes ANGPTL4. Simultaneously, muscular fatty acid oxidation is increased by PPAR-δ–dependent induction of β-oxidative enzymes. Via the circulation, mANGPTL4 enhances WAT lipolysis and thus prevents too strong decrements of plasma NEFA levels and ensures ongoing fuel supply of the stressed (fasting or working) muscle. Together with ANGPTL4's inhibitory effect on LPL, this mechanism is expected to provoke loss of WAT mass. Furthermore, as derived from our in vitro data, ANGPTL4 could stimulate SKM lipolysis via LIPE induction in an autocrine/paracrine manner. This effect would constitute, in addition to PPAR-δ's inductive effect on β-oxidation, a synergistic mode of PPAR-δ action on muscle lipid catabolism (IMCL, intramyocellular lipids).