Perilipin 3 Differentially Regulates Skeletal Muscle Lipid Oxidation in Active, Sedentary, and Type 2 Diabetic Males

Abstract

Context:

The role of perilipin 3 (PLIN3) on lipid oxidation is not fully understood.

Objective:

We aimed to 1) determine whether skeletal muscle PLIN3 protein content is associated with lipid oxidation in humans, 2) understand the role of PLIN3 in lipid oxidation by knocking down PLIN3 protein content in primary human myotubes, and 3) compare PLIN3 content and its role in lipid oxidation in human primary skeletal muscle cultures established from sedentary, healthy lean (leans), type 2 diabetic (T2D), and physically active donors.

Design, Participants, and Intervention:

This was a clinical investigation of 29 healthy, normoglycemic males and a cross-sectional study using primary human myotubes from five leans, four T2D, and four active donors. Energy expenditure, whole-body lipid oxidation, PLIN3 protein content in skeletal muscle tissue, and ex vivo muscle palmitate oxidation were measured. Myotubes underwent lipolytic stimulation (palmitate, forskolin, inomycin [PFI] cocktail), treatment with brefeldin A (BFA), and knockdown of PLIN3 using siRNA.

Setting:

Experiments were performed in a Biomedical Research Institute.

Main Outcome Measures:

Protein content, 24-hour respiratory quotient (RQ), and ex vivo/in vitro lipid oxidations.

Results:

PLIN3 protein content was associated with 24-h RQ (r = −0.44; P = .02) and skeletal muscle–specific ex vivo palmitate oxidation (r = 0.61; P = .02). PLIN3 knockdown showed drastic reductions in lipid oxidation in myotubes from leans. Lipolytic stimulation increased PLIN3 protein in cells from leans over T2Ds with little expression in active participants. Furthermore, treatment with BFA, known to inhibit coatomers that associate with PLIN3, reduced lipid oxidation in cells from lean and T2D, but not in active participants.

Conclusions:

Differential expression of PLIN3 and BFA sensitivity may explain differential lipid oxidation efficiency in skeletal muscle among these cohorts.

Impaired lipid oxidation along with elevated levels of intramyocellular lipid (IMCL) are prominent characteristics of skeletal muscle pathophysiology associated with type 2 diabetes (T2D) (1). It has been demonstrated that reduced insulin sensitivity in skeletal muscle is strongly associated with increased IMCL content (2). However, Goodpaster et al (3) showed in 2001 that athletes who are highly insulin sensitive also have high levels of IMCL, known as the athlete's paradox. Efforts to understand lipid storage and utilization in skeletal muscle are important to provide insight into the athlete's paradox.

Storage and packaging of lipid droplets in skeletal muscle has recently become an important area of research. Much research has been extensively focused on perilipin 5 (PLIN5), expressed in skeletal muscle (4), and has been shown to be associated with lipid metabolism (4–11). Unfortunately, another lipid droplet protein highly expressed in skeletal muscle (4), perilipin 3 (PLIN3), has been neglected. Studies conducted in mouse skeletal muscle have shown PLIN3 to colocalize to lipid droplets during catecholamine stimulation and also after muscle contraction (12). Other studies have shown that adipose triglyceride lipase (ATGL) colocalizes to PLIN3-coated lipid droplets during lipolytic stimulation (13, 14). Though ATGL does not directly interact with PLIN3 (11), other investigations have shown that coatomer GTPases, such as ADP-ribosylation factor 1 (ARF1), ARF-related peptide 1 (ARFRP1), and Golgi-Brefeldin A–resistant factor 1 (GBF1), are involved in pathways that deliver ATGL to PLIN3-coated lipid droplets, thus facilitating lipid oxidation (15, 16). We have recently shown a possible role of PLIN3 on skeletal muscle lipid oxidation using both in vitro and ex vivo models of exercise (17). However, apart from this observational study, the role of PLIN3 for skeletal muscle lipid oxidation has not been heavily investigated. To the best of our knowledge, no studies have examined skeletal muscle lipid oxidation following knockdown of PLIN3 in skeletal muscle cells, nor has the differential expression of PLIN3 and PLIN5 been investigated in myotubes from actives, T2Ds, or metabolically healthy sedentary lean subjects following lipolytic stimulation.

We aimed, therefore, to perform an investigation in three phases to examine the effects of PLIN3 on skeletal muscle lipid oxidation. First, we performed a clinical investigation in 29 sedentary, healthy, nonobese, normoglycemic males to examine the relations between the skeletal muscle protein expression of PLIN3 to both whole-body fat oxidation as measured using a 24-hour metabolic chamber and ex vivo skeletal muscle specific lipid oxidation. Second, we cultured primary myotubes from five of these participants to determine the role of PLIN3 on lipid oxidation by knocking down PLIN3. Finally, we compared PLIN3 and PLIN5 protein content following in vitro pharmacological lipolytic stimulation (palmitate, forskolin, and ionomycin; PFI cocktail), as well as lipid oxidation following brefeldin A (BFA) treatment, a pharmacologic compound show to inhibit ARF1 and coatomers that associate with PLIN3 (15) in primary skeletal muscle myotubes cultures from active, sedentary lean, and T2D donors. We hypothesized that PLIN3 would be associated with both whole-body and muscle-specific lipid oxidation, and that knockdown of PLIN3 in primary human myotubes would result in lower lipid oxidation. We also hypothesized there would in fact be a differential expression of PLIN3 and PLIN5 in primary skeletal muscle cultures from these three cohorts, and that would be differential lipid oxidation response to BFA treatment from these cohorts.

Methods and Materials

Clinical studies and skeletal muscle biopsy procedure

Twenty-nine sedentary, normoglycemic male participants were recruited into the EAT trial (ClinicalTrials.gov No. NCT01672632) and underwent body composition measures (dual-energy x-ray absorptiometry, QDR 4500A, Hologics; and magnetic resonance imaging [MRI], 3.0T magnet, Excite HD System, General Electric), fasting blood serum measures and a 24-hour respiratory chamber stay to measure 24-hour energy metabolism. All anthropometric and metabolic characteristics are provided in Table 1. Vastus lateralis muscle biopsies were obtained to measure protein content (n = 26), ex vivo palmitate oxidation (n = 15), and to establish primary myotube cultures in five lean, healthy male donors (Lean). As a secondary study, skeletal muscle biopsies were obtained from the vastus lateralis from four lean, physically active male donors (Active), and four obese, type 2 diabetic male donors (T2D) to establish primary skeletal muscle cultures for cross-sectional comparison with sedentary, lean donors. Anthropometric and metabolic characteristics of the Active, T2D, and the subset of five Lean participants, in which primary muscle cultures were examined, are provided in Table 2. Active subjects were recruited based on their level of habitual physical activity level, if they were between 20 and 40 years of age, body mass index (BMI) between 20 and 30 kg/m², were nondiabetic, were taking no medications, and were otherwise healthy. Physical activity level was calculated from a 7-day physical activity questionnaire recall and a triaxial accelerometer worn for at least 4 days. Physical activity index (total daily energy expenditure/resting metabolic rate) was calculated using both methods and daily activity level was scrutinized from accelerometer data ensuring active subjects have an activity index of greater than 1.6. Another inclusion criterion for active participants was a VO_2max greater than 40 mL/kg/min. Active participants were enrolled onto the ACTIV trial (ClinicalTrials.gov No. NCT00401791). Participants with T2D were enrolled onto the TAKE TIME trial (ClinicalTrials.gov No. NCT00401791) if they had known T2D mellitus, were weight stable, were otherwise healthy, and were permitted to be taking metformin, insulin and/or sulfonylureas, but not thiazolididediones. All participants gave written informed consent, and all trials were reviewed and approved by the Institutional Review Board of Pennington Biomedical Research Center. Clinical anthropometric and metabolic characteristics of study participants were performed as follows: 24-hour whole-body metabolism was measured using in a metabolic chamber as previously described (18); basal and maximal in vivo assessment of skeletal muscle mitochondrial ATP production were performed under magnetic resonance spectroscopy (3T Signa Excite MRI; General Electric) as previously described (19, 20); fasting serum measures were assessed in a certified clinical chemistry laboratory. We reported cross-sectional measures of clinical data for 24-hour energy expenditure, and mitochondrial ATP production [which we have shown to associate with VO_2max (20)] to show that there were indeed differences between insulin sensitivity and physical fitness from the active participants compared with participants with T2D. Skeletal muscle samples were collected after an overnight fast using the Bergstrom technique with suction (Propper Manufacturing Co.) from the vastus lateralis following administration of local anesthetic of lidocaine/bupivicaine. Clinical measures were performed in the same order for all participants with MRI/magnetic resonance spectroscopy (MRS) measures performed first followed by a stent in the metabolic chmber for 24 hours. The participants then had skeletal muscle biopsies performed the morning they emerged from the metabolic chamber.

Table 1.

Anthropometric and Clinical Characteristics of Clinical Study

Anthropometric Characteristics	Mean ± sd
N	29
Age, y	26.8 ± 5.4
Weight, kg	81.9 ± 10.3
BMI, kg/m2	25.5 ± 2.3
Fat, %	19.4 ± 4.9
FM, kg	16.0 ± 4.8
FFM, kg	65.9 ± 7.3
SAT, kg	4.1 ± 1.5
VAT, kg	0.58 ± 0.49
Type-1 Fibers (vastus lateralis), %	35.0 ± 13.5
Type-2a Fibers (vastus lateralis), %	53.0 ± 14.3
Type-2× Fibers (vastus lateralis), %	12.0 ± 10.9
Metabolic characteristics
24-h RQ	0.89 ± 0.02
Serum measures
Glucose, mg/dL	91.0 ± 6.7
Insulin, μU/mL	5.4 ± 4.0
FFA, nmol/L	0.26 ± 0.08
Triglycerides, mg/dL	87 ± 42
Total cholesterol, mg/dL	171 ± 25
HDL-C, mg/dL	55 ± 12
LDL-C, mg/dL	99 ± 23
Cholesterol/HDL	3.27 ± 0.96
HDL/LDL	0.59 ± 0.20

Abbreviations: FM, fat mass; FFM, fat free mass; FFA, free fatty acids.

Table 2.

Anthropometric and Clinical Characteristics from Primary Myotubes

Characteristics	Active, Mean ± sd	Lean, Sedentary, Mean ± sd	T2D, Mean ± sd	P Value
N	4	5	4	—
Weight, kg	79.98 ± 8.88a	76.50 ± 8.19a	110.53 ± 18.31b,c	.005
FM, kg	10.58 ± 1.96a	14.36 ± 6.29a	31.41 ± 2.89b,c	<.001
FFM, kg	69.75 ± 7.50	62.14 ± 4.30	64.73 ± 5.37	.18
Fat, %	13.15 ± 1.95a	18.36 ± 6.39a	35.27 ± 5.37b,c	<.001
BMI, kg/m2	25.12 ± 2.55a	24.18 ± 0.55a	39.65 ± 6.98b,c	<.001
Serum glucose, mg/dL	88.75 ± 2.75a	87.80 ± 5.50a	117.25b,c	.02
Maximal ATP production, mm/min	1.17 ± 0.07a,b	0.66 ± 0.11c	0.59 ± 0.24c	<.001
Basal ATP production, μM/min	7.27 ± 2.18a	5.35 ± 1.42	3.37 ± 1.81c	.04
24-h energy expenditure, kcal	3854 ± 339.79a,b	2181 ± 132.23c	2675 ± 201.23c	<.001

Abbreviations: FM, fat mass; FFM, fat free mass.

P values are provided from one-way ANOVA. Tukey post-hoc test was used to evaluate differences between groups:

^aP < .05 compared with T2D.

^bP < .05 compared with Lean.

^cP < .05 compared with Active.

Assessment of in vivo mitochondrial capacity

Mitochondrial capacity was assessed both as the maximal rate of ATP production (ATPmax) and basal rate of ATP production (ATPase). These were both measured by ³¹P MRS. Detailed descriptions of these methodologies have been previously described (19, 20). ATPmax: maximal mitochondrial capacity was assessed in the vastus lateralis by measuring the time constant (tau) required for phosphocreatine (PCr) recovery after 45 seconds of isometric contractions of the quadriceps and the PCr level in resting oxygenated muscle (PCr_rest): ATPmax = PCr_rest/tau. Participants were asked to lie supine on the patient table of the 3.0 Tesla Magnetic Resonance Imager/Spectrometer (GE Excite) for approximately 45 minutes. After a scout scan to locate the vastus lateralis, ³¹P spectra were acquired every 60 seconds for 2 minutes to achieve baseline PCr, ATP, and phosphate ion concentrations; during acquiring ³¹P spectra, a blood pressure (BP) cuff was inflated around the upper thigh to 30 mm Hg above systolic BP. The cuff was inflated for approximately 16 minutes. The participants were then asked to perform isometric contractions of the quadriceps muscles for 45 seconds at a frequency of 1 contraction per second to deplete PCr stores. Replenishment of PCr stores were assessed following the 45-second isometric exercise. ATPase: ATPase rate was measured by the rate of PCr breakdown in anoxic muscle, which represents the basal ATP demand of the cell. Anoxia in the vastus lateralis was induced by inflating a BP cuff around the upper thigh to 50–60 mm Hg above systolic BP, thus haulting blood flow and O₂ delievery, and therefore requiring the muscle to deplete stores of Mb-O₂ and Hb-O₂ to meet aerobic respiration needs. Cuff pressure was controlled with a Hokanson Rapid Inflator/Deflator system. Anoxia/ischemia was induced in the participant's leg for 15 minutes with levels of PCr, ATP, and phosphate ion assessed by ³¹P MRS in the vastus lateralis for 60 seconds every 2 minutes. In addition, glycolytic ATP production was assessed by measuring changes in PCr and cellular pH by ¹H MRS.

Measurement of abdominal subcutaneous and visceral adipose tissue volumes

Abdominal subcutaneoussc adipose tissue (SAT) and visceral adipose tissue (VAT) volumes were assessed with a 3.0 T scanner (Excite HD System; General Electric). Between 240 and 340 images were obtained from the highest point of the liver to the pubic symphysis and analyzed by a single trained technician using Analyze software (AnalyzeDirect). The mean coefficient of variation for three readings of the same scan was 9.9% for VAT and 1.8% for SAT. Estimates of MRI volumes were converted to mass using an assumed density of 0.92 kg/L.

Establishment of primary human skeletal muscle cultures and in vitro treatment with PFI, BFA, and knockdown of PLIN3

Establishment of human primary muscle culture has been modified from protocols as previously described (17, 22), and distinct pooled cell lines for Lean, Active, and T2D groups were established for experiments using protocols previously described (23). Myotubes were treated with 30 μM palmitate, 4 μM forskolin, and 0.5 μM ionomycin (PFI cocktail) using techniques adapted as previously described (17, 24). BFA treatments were carried out in myotubes using techniques adapted from Soni et al (15). Briefly, myotubes were maintained in differentiation media for 7 days and then treated for 30 minutes with BFA at a concentration 1 μg/mL (purchased from Sigma). Knockdown of PLIN3 was achieved through siRNA silencing using Silencer Select Predesigned siRNAs (Cat No. s19952) according to the manufacturer's instructions (Life Technologies). Myotubes were differentiated for 7 days and treated with lipofectamine in serum-free media followed by siRNA incubation with either siPLIN3 or siScramble (siSCR; Life Technologies), or exposed merely to serum-free media for controls. Cultures were maintained under serum starvation for 24 hours per the manufacturer's specifications.

Skeletal muscle tissue measures of ex vivo and primary myotube measures of in vitro lipid oxidation

Lipid oxidation was performed in skeletal muscle tissues as previously described (25). Complete lipid oxidation in primary myotubes was measured as liberation of ¹⁴CO₂ using 100 μM [1-¹⁴C] oleate as previously described with slight modifications (26, 27). Cells were treated with 500 μL of radioactive media, which consisted of low-glucose DMEM supplemented with 12.5mM HEPES, 1mM L-carnitine, and 100μM [1-¹⁴C] oleate (1 μCi/mL), and maintained in an incubator (37°C, 5% CO₂) for 2 hours. At the end of the reaction period, the plates were placed on ice to stop the reaction, and 400 μL of the reaction medium was transferred to plastic tubes, which were sealed with rubber caps, with the addition of 100 μL of 70% perchloric acid to liberate ¹⁴CO₂, which was trapped in 200 μL of 1 N NaOH. NaOH-trapped ¹⁴CO₂ was detected via scintillation counting. Data were normalized to protein content of each individual well.

Gene and protein expression measures

Gene expression was performed as previously described (17) with Real-Time qPCR on a 7900HT Fast Real-Time PCR system (Life Technologies) using TaqMan Gene Expression Assays-on-Demand (Life Technologies). Catalogue numbers for each Assay-on-Demand product is provided as follows: ARF1 (Hs00796826_s1), Sec23a (Hs00197232_m1), CGI-58 (Hs01104373_m1), perilipin 2 (PLIN2) (Hs00765634_m1), and Ribosomal Protein, Large Protein 0 (Hs99999902_m1). Expression levels were determined against a standard curve and adjusted to the expression of Ribosomal Protein, Large Protein 0. Total protein for all experiments (both in skeletal muscle tissue and in human primary myotubes) was collected using RIPA buffer (Sigma) supplemented with 2% protease inhibitor cocktail (Sigma), 2% phosphatase inhibitor cocktail 2 (Sigma), and 2% phosphatase inhibitor cocktail 3 (Sigma). Protein content was assessed from total protein extracts using Western immunoblotting adjusted to GAPDH (Cat No. AB9484; AbCam). Imaging of Western blots was facilitated on the Odyssey infrared imaging system (LiCor). Antibodies against PLIN3 (Cat No. NB110–40764) and PLIN5 (Cat No. NB110–60511) were obtained from Novus Biologicals; the antibodies for ATGL (Cat No. AB109251) and ARFRP1 (Cat No. AB08199) were obtained from AbCam.

Statistical analysis

Data was analyzed using the PRISM GraphPad Software, version 6.0 (GraphPad Software). Pearson r correlations were used to assess associations between PLIN3 protein content and metabolic measures of lipid oxidation. One-way ANOVA with Tukey post-hoc tests were performed to compare differences between clinical and myotube measures across Active, Sedentary Lean, and T2D individuals (Table 2 and Figure 1D). Two-way, paired Student t tests were used to assess differences between control and treatment measures in myotubes (Figure 1, A–C). Myotube experiments were performed in triplicate. All graphical data are presented as the mean ± SEM, and a P < .05 was considered statistically significant.

Figure 1.

A, PLIN3 protein content from the vastus lateralis is inversely related to 24-h RQ indicating that increases in PLIN3 protein content is associated with increased whole-body fat oxidation (Person r = −0.44; P = .02; n = 26). B, PLIN3 protein content from the vastus lateralis is positively associated with ex vivo palmitate oxidation from muscle tissue from the vastus lateralis indicating that increases in PLIN3 protein content is associated with increased skeletal muscle–specific fat oxidation (Pearson r = 0.61; P = .02; n = 15). C, Representative images of Westerns blots probed for PLIN3 from primary human myotubes from lean, sedentary individuals under control conditions, following SCR treatment, and following siRNA knockdown of PLIN3. D, Complete lipid oxidation levels in myotubes established from lean participants following knockdown of PLIN3 compared with control and treatment with an SCR. A distinct decrease in lipid oxidation is noted in cells from leans following PLIN3 knockdown. Data represents mean ± SEM of experiments performed in triplicate. *, P < .05 vs control and vs SCR.

Results

PLIN3 protein is associated with whole-body and ex vivo lipid oxidation and PLIN3 knockdown results in marked decrease in complete lipid oxidation

The 29 male participants (age, 26.8 ± 5.4 y; weight, 81.9 ± 10.3 kg; BMI, 25.5 ± 2.3 kg/m²) were considered on average insulin sensitive (fasting glucose level, 91.0 ± 6.7 mg/dL, also confirmed by hyperinsulinemic euglycemic clamp [data not shown]) and nonobese with body fat of 19.4 ± 4.9%. Fasting serum lipids showed a total triglyceride level of 87 ± 42 mg/dL, free fatty acids of 0.26 ± 0.08 nmol/L, total cholesterol of 171 ± 25 mg/dL, high-density lipoprotein (HDL) cholesterol of 55 ± 12 mg/dL, low-density lipoprotein (LDL) cholesterol of 99 ± 23 mg/dL, and the ratio of total cholesterol to HDL of 3.27 ± 0.96 as well as HDL-to-LDL ratio of 0.59 ± 0.20. These results have been provided in Table 1. Skeletal muscle protein content of PLIN3 was inversely related to 24-hour respiratory quotient (RQ) measured in a metabolic chamber (Figure 2A) and was positively associated with ex vivo palmitate oxidation in skeletal muscle tissue (Figure 2B). There were no other significant correlations between PLIN3 protein content and other clinical variables: type-I fiber (r = −0.32; P = .11), body weight (r = −0.12; P = .57), fat-free mass (r = −0.09; P = .64), fat mass (r = −0.06; P = .77), percent body fat (r = −.02; P = .91), VAT (r = 0.12; P = .59), and SAT (r = 0.13; P = .53). Knockdown of PLIN3 in myotubes from lean, sedentary donors revealed an ∼85% reduction in complete oleate oxidation when compared with control and siSCR (Figure 2D). Figure 2C provides a representative immunoblot showing knockdown of PLIN3 in myotubes from lean donors.

Figure 2.

Primary myotubes taken from active (A); sedentary, lean (L); and type 2 diabetic (T or T2D) participants display differential protein content following PFI treatment (time course from immediately following 3 d of PFI treatment to 24 h post-PFI treatment). A, Image of Western immunoblotting of control and the time course following PFI treatment. B, Quantification of PLIN3 protein content reveals that PLIN3 is most prominently expressed in Ls followed by T2Ds with only minimal expression in actives. C, In contrast, PLIN5 is most prominently expressed in actives with PLIN5 only evident in T2Ds very minimally 24 h post-PFI treatment. D, ARFRP1 protein is most prominent in Ls over the other two cohorts throughout the duration of time course following PFI treatment. E, ATGL protein content is more prominent in actives immediately following, 30 min following, and 1 h following cessation of PFI treatment over the other two cohorts, with T2Ds having virtually no change in ATGL from control conditions, and Ls increasing expression at 1 and 24 h post-PFI treatment.

Differential expression and response of PLIN3 and intracellular transport proteins after PFI treatment in myotubes from active, lean, and T2D

In terms of PLIN3 protein content, cells cultured from sedentary lean donors revealed an increase of PLIN3 following PFI treatment and up to 24 hours after PFI treatment, similar to our previous experiment (17), whereas T2D myotubes continuously increase levels of PLIN3 protein content up to 24 hours (Figure 3B), both expressing higher levels than cells cultured from active donors. Instead, PLIN5 protein content was up-regulated in Actives from immediately following PFI treatment up to 24 hours after treatment (Figure 3C). Lean and T2D myotubes displayed virtually no PLIN5 protein content with PFI treatment (Figure 3C). ARFRP1 had the most prominent protein content following PFI treatment in Leans above both T2Ds and Actives (Figure 3D). ATGL levels were increased in active myotubes following PFI treatment up to the 1-hour time point, with increases in lean myotubes only at 1 hour and 24 hours post-PFI treatment, and virtually no changes in T2D myotubes (Figure 3E). At 24 hours post-PFI treatment, PLIN3 and ARFRP1 had higher protein content in lean and T2D myotubes (Figure 3, B and D); PLIN5, however, was greatly expressed only in active myotubes 24 hours post-PFI treatment (Figure 3C).

Figure 3.

Differential mRNA expression following PFI treatment (time course) of myotubes collected from actives, leans, and T2Ds. A, ARF1, a coatomer protein associated with ER-to-Golgi transport, is highly expressed in cells from lean following PFI with actives showing the highest decrease. B, Sec23a, part of the COPII complex, is differentially expressed in the three cohorts with actives having the highest increase in expression following PFI treatment. C, CGI-58, the coactivator of the lipase ATGL, is most highly expressed in actives following PFI treatment. D, PLIN2, a member of the perilipin family, displayed a distinct expression patter among the three cohorts with leans increasing their expression following PFI with actives having a nearly 80% decrease in mRNA expression, and T2Ds having between a 20–60% decrease following the time course post-PFI treatment. All experiments were performed in triplicate and represented as mean ± SEM for a percent change.

The mRNA expression of ARF1, a coatomer protein that interacts with GBF1 to facilitate intracellular transport (28), was elevated above control conditions following PFI treatments in lean myotubes only, whereas both T2D and active cells decreased their ARF1 expression, notably actives reducing their expression more than T2Ds (Figure 4A). Sec23a, a subunit of the Coatomer protein, type 2 (COPII) complex responsible for retrograde transport from Golgi-to-endoplasmic-reticulum (ER) (29), had increased expression in both lean and active myotubes following PFI treatment, whereas it remained only minimally elevated in T2D cells (Figure 4B). CGI-58, the coactivator of ATGL (30), was grossly elevated in active myotubes following PFI treatment over and above both T2Ds and lean myotubes (Figure 4C). Finally, PLIN2, a perilipin protein formerly known as ADRP, had drastic reduction in mRNA expression in active myotubes followed by T2D after PFI treatment, whereas cells from lean increased PLIN2 expression (Figure 4D).

Figure 4.

Complete fatty acid oxidation (CO₂ release) in myotubes established from lean (A), T2Ds (B), and actives (C) under control conditions and following 30 min treatment with BFA, a drug known to inhibit ARF1 and ER-to-Golgi transport. D, Percent changes in fatty acid oxidation compared with control conditions between leans, T2Ds, and actives. All data represents mean ± SEM of experiments performed in triplicate. *, P < .05 vs control.

Reduced in vitro lipid oxidation following BFA treatment in lean and T2D myotubes but not in actives

Treatment of myotubes with BFA, a drug that inhibits the activity of ARF1 (31), displayed reduced in vitro oleate oxidation for lean (Figure 1A) and T2D donors (Figure 1B). However, oleate oxidation was unaffected following BFA treatment in active myotubes (Figure 1C). Percent change of in vitro oleate oxidation compared with control conditions revealed a significant difference between cells from lean and active (P = .02) and trended toward a difference between T2Ds and Actives (P = .08), with no differences in oxidation between lean and T2D myotubes (Figure 1D).

Discussion

The relationship between skeletal muscle insulin resistance and intramyocellular lipid (IMCL) content gained considerable attention as both endurance-trained athletes and patients with T2D possess elevated IMCL yet represent the bilateral extremes across the insulin sensitivity spectrum. Our study has shown for the first time a differential protein expression of PLIN3 and PLIN5 in primary skeletal myotubes from active, lean and T2D following in vitro lipolytic treatment. Furthermore, our study has shown the novel potential role PLIN3 might play in facilitating skeletal muscle lipid oxidation as evident by the reduced lipid oxidation levels in myotubes from lean donors following PLIN3 knockdown and its associations with both-whole body in vivo lipid oxidation and skeletal muscle–specific ex vivo lipid oxidation.

Prior investigations into the PAT family of proteins in human skeletal muscle have focused extensively on the function of PLIN5. Repeated investigations have shown a direct protein–protein interaction between PLIN5 and ATGL (10, 11), and that PLIN5 overexpression in mice has enhanced lipid oxidation (9). Our investigation shows that PLIN5 is highly up-regulated in myotubes derived from active donors following stimulation with PFI that has previously been shown to increase lipid oxidation and been referred to as an exercise mimetic (Figure 3C) (24). Our findings support the hypothesis that skeletal muscle from endurance trained individuals utilizes PLIN5 for packaging of lipid droplets for faster lipid oxidation due perhaps in part to the direct interaction between PLIN5 and ATGL, which has been shown previously (10, 11). As we show in Figure 3E, ATGL is more responsive acutely to PFI treatment in active myotubes, and likewise, CGI-58, the coactivator of ATGL, greatly increases mRNA expression in active myotubes over and above sedentary, lean myotubes and T2D myotubes following PFI treatment (Figure 4C). In addition, it has been shown that PLIN5 packages lipid droplets composed exclusively of triacylglycerides (32). We have previously shown that primary myotubes from active donors have significantly elevated levels of triacylglycerides over both lean and T2Ds (33), thus providing an additional explanation for elevated levels of PLIN5 in actives following lipolytic stimulus.

However, the abundance of data concerning PLIN5 and its association with lipid oxidation has perhaps overshadowed other targets that may be relevant to muscle lipid storage for oxidation. PLIN3 has been shown, in our previous study, to be positively associated with both whole body in vivo and ex vivo skeletal muscle tissue fat oxidation following an endurance exercise bout in 20 healthy males (17). We also showed an increase in PLIN3 protein following epinephrine stimulation of primary human myotubes (17). Prior investigations have shown a colocalization of ATGL to PLIN3 coated lipid droplets in HeLa cells (13, 14), and a colocalization of PLIN3 to lipid droplets in isolated mouse skeletal muscle (12). All of these have provided circumstantial evidence to the involvement of PLIN3 with the facilitation of lipid oxidation in skeletal muscle. Our present study independently repeated our findings of PLIN3 protein content being associated with both whole body in vivo and skeletal muscle specific ex vivo lipid oxidation; this time, although, using a 24-hour metabolic chamber (Figure 2A) and using skeletal muscle taken under resting, basal conditions (Figure 2B). Importantly, our study with human myotubes also demonstrated a robust reduction in lipid oxidation following knockdown of PLIN3 (Figure 2, C and D). Previously, we have reported that in a naturally occurring knockdown of PLIN3 protein in primary adipose cultures taken from women with polycystic ovary syndrome, fat oxidation is reduced, but is increased following aerobic exercise, which is concomitantly associated with increases in PLIN3 protein content (34). Here, we show that induction of knockdown of PLIN3 protein content likewise results in reductions in fat oxidation. This evidence demonstrates a significant possibility that PLIN3 serves as a target responsible for facilitating lipid oxidation in sedentary, lean and in insulin-resistant individuals.

We further investigated whether the effects of exercise mimetic stimulation on PLIN3 and PLIN5 protein content in Actives, Leans and T2D donors. Our findings show a remarkable difference in PLIN3 and PLIN5 response upon exercise mimetic stimulation among these three cohorts. Actives almost exclusively expressed PLIN5 following exercise stimulation (Figure 3C) whereas Leans and T2Ds favor the expression of PLIN3 (Figure 3B). In addition, we investigated the expression of ER-to-Golgi transport coatomer GTPases. Prior investigations in HeLa cells have suggested the dependence of coatomer proteins in facilitating ATGL delivery to PLIN3-coated lipid droplets (15, 16). We previously reported an up-regulation of several coatomer factors following a long bof endurance exercise in human skeletal muscle tissue (17). We also previously reported the same phenomenon to occur in adipose tissue from women with polycystic ovary syndrome, a cohort that seemingly favors PLIN3 expression following aerobic exercise (34). Here, we have demonstrated a differential response of coatomer targets among cells from Actives, Leans, and T2Ds after in vitro lipolytic stimulus at the protein level (ARFRP1; Figure 3D) and mRNA level (ARF1 and Sec23a; Figure 4). In addition, we demonstrate the differences in related targets involved in lipid packaging and lipolysis among these populations (CGI-58 and PLIN2; Figure 4). Though these differences are novel, they alone were not fully insightful. Thus, we further investigated lipid oxidation levels in primary myotubes cultured from these three cohorts after treatment with BFA, an inhibitor of ARF1 activity (31). BFA has previously been shown to block ATGL delivery to PLIN3-coated lipid droplets (15), but has similarly been shown not to inhibit intracellular transport of PLIN3 (35). We demonstrated a reduction in lipid oxidation in myotubes from lean and in T2D donors following BFA treatment; therefore, suggesting that lipid oxidation is dependent in part on coatomer GTPases in these two groups (Figure 1, A and B). In contrast, myotubes from active donors did not show a reduction in lipid oxidation, suggesting a differing divergent pathway not requiring coatomer GTPases for facilitating lipid oxidation (Figure 1C). The novelty of this aspect of our study was the demonstration of differential up-regulation of PLIN3 and PLIN5 following exercise-mimetic stimulation, the differential expression of coatomer GTPases, and the differential lipid oxidation levels following BFA treatment. These data demonstrate the possibility of previously unexplored lipid oxidation pathways dependent on PLIN3 and coatomer GTPases that is more readily used in skeletal muscle of Sedentary, Lean, and possibly T2Ds compared with Actives. Furthermore, these data highlight that PLIN3 would be pharmacological target in individuals with T2D due to the fact that they seem to favor lipolytic-induced expression of PLIN3 and do not seem to express much PLIN5.

One of the major strengths of our study is the fact that we independently reproduced positive correlational associations between both whole-body in vivo lipid oxidation using results from a 24-hour metabolic chamber and skeletal muscle ex vivo lipid oxidation with PLIN3 protein content. This shows that in two separate, independent studies from nonactives that skeletal muscle expression of PLIN3 is associated with lipid oxidation (17). To further strengthen those results, we successfully knocked down PLIN3 protein expression in primary human myotubes collected from individuals who participated in the reported clinical investigation and showed that lipid oxidation in vitro is reduced with reduced PLIN3 protein content. A second major strength of our study is the fact that our subsequent data presented here is exclusively performed in human primary myotubes as opposed to cultured cell lines. We and others have shown that human myotubes reflect the phenotype of the donor (33, 36), thus allowing a powerful tool for investigating the skeletal muscle pathophysiology associated with T2D by using primary myotubes obtained from donors with T2D. Although there is no substitute for a rigorous clinical investigation from direct skeletal muscle tissue in these cohorts, we hold that our experiments in primary human myotubes offer conclusive insight into PLIN3 as facilitating muscle lipid oxidation.

The T2D participants who we were able to recruit for this study were taking medications that assist in regulating hyperglycemia (eg, metformin, sulfonylurea, and insulin). Therefore, the effects of medication on any differences in the results from T2D participants cannot be completely ruled out. We acknowledge that inclusion of an obese, nondiabetic cohort as well as T2D subjects without medication could have provided further insight to our results. However, based on the fact that our aims were to explore the effects of lipolytic stimulation on PLIN3 content and the inhibition of coatomer function on levels of lipolysis in myotubes from the three cohorts to see whether there were in fact differences, our data presented here are of great importance.

To conclude, the skeletal muscle expression of PLIN3 is associated with both in vivo and ex vivo lipid oxidation, and has been shown as a target for the facilitation of lipid oxidation in primary human myotubes taken from healthy, lean, sedentary male donors. PLIN3 is more highly up-regulated in myotubes taken from lean and T2D donors following in vitro exercise-mimetic stimulation. The ER-to-Golgi coatomer target ARF1 similarly seems to be involved with facilitating lipid oxidation in primary myotubes from lean and T2D donors. In contrast, primary myotubes from active donors do not seem to express much PLIN3, but rather express PLIN5, following in vitro lipolytic stimulation. In addition, active donors do not seem to be dependent on ARF1 or related coatomer proteins to stimulate lipid oxidation whereas sedentary leans and patients with T2D did. These data suggest a potential for two separate pathways (PLIN3 and PLIN5) in regulation of lipid oxidation in skeletal muscle that may depend upon training status of the individual, and thus may in part offer insight into the existence of the athlete's paradox.