CTRP9 transgenic mice are protected from diet-induced obesity and metabolic dysfunction

Antibodies and chemicals.

Mouse monoclonal anti-FLAG M2 antibody was obtained from Sigma, and rat monoclonal anti-HA (clone 3F10) antibody was obtained from Roche. Rabbit antibodies recognizing phospho-AMPKα (Thr-172), AMPKα, and COX IV were obtained from Cell Signaling Technology. Mouse anti-GAPDH monoclonal antibody was obtained from Novus Biologicals. HRP-conjugated rabbit polyclonal anti-tubulin antibody was obtained from Abcam. Rabbit polyclonal antibody used to detect endogenous CTRP9 was generated previously (47).

Animals.

C57BL/6J male mice (The Jackson Laboratory) were used to evaluate diet-induced changes in CTRP9 mRNA and circulating levels. Sera from wild-type male mice subjected to ad libitum fed, fasted, and refed were collected, and CTRP9 levels were quantified by Western blot analysis. For ad libitum-fed group, serum samples were obtained at 2–3 h into the light cycle. For the fasted group, food was removed for 16 h (beginning at 10 h into the light cycle), and sera were collected at ∼2–3 h into the light cycle. For the refed group, mice were fasted for 16 h and refed with chow pellets for 2 h before sera were collected. For all other experiments, mice were bred and weaned at The Johns Hopkins University School of Medicine animal facilities. Four-week-old CTRP9 transgenic mice and WT control littermates were housed in polycarbonate cages on a 12:12-h light-dark photocycle with ad libitum access to water throughout the study period. Male mice were used throughout the study. Mice were fed ad libitum a high-fat diet (HFD) (60% kcal derived from fat, Research Diets; D12492) or an isocaloric-matched low-fat diet (LFD) (10% kcal derived from fat, Research Diets; D12450B). HFD was provided for a period of 14 wk. Blood samples were collected for serum analysis. Tissues were collected, snap frozen in liquid nitrogen, and kept at −80°C. All animal protocols were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University School of Medicine.

Generation of CTRP9 transgenic mice.

The C-terminal HA epitope-tagged CTRP9 was cloned into the XhoI site of pCAGGS vector (25). Expression of CTRP9 transgene was driven by the ubiquitous CAG promoter, which consists of a CMV enhancer element with a chicken β-actin promoter. The plasmid construct was digested with SalI and NotI restriction enzymes, and resulting DNA fragments (∼3.5 and 2.5 kb) were separated on a 1% agarose gel. The ∼3.5 kb linear DNA fragment containing the CAG promoter and enhancer, CTRP9-HA transgene, and the rabbit β-globin polyA adenylation signal were excised from the agarose gel, purified, and verified by DNA sequencing. Pronuclear injections were performed; several transgenic founder lines were obtained. Multiple founder lines were obtained, and all showed a lean phenotype. However, only one line was maintained and expanded for phenotypic analysis. The transgenic line was maintained on a (C57BL/6 × FVB) mixed genetic background. Transgene-negative littermates were used as WT control mice throughout the study.

Semiquantitative PCR analysis.

Total RNAs from mouse tissues were isolated with TRIzol (Invitrogen). Two micrograms of total RNA were reverse-transcribed using Superscript II (Invitrogen). Thirty-cycle PCR was carried out using Hot Start Taq Blue polymerase (Denville); the cycling conditions were as follows: 15 s denaturation at 95°C, 15 s primer annealing at 60°C, and 45 s primer extension at 72°C. Primers used included the following: CTRP9-HA forward 5′-CCCAGATGCACCCATTAAATTCG-3′ and reverse 5′-TCAAGCGTAGTCTGGG ACGTCGTATGG-3′; and β-actin forward 5′-CGTGACATTAAGGAGAAGCTGTGC-3′ and reverse 5′-CTCAGGAGGAGCAATGATCTTGAT-3′.

Mouse serum and blood chemistry analysis.

Mouse serum samples were harvested by tail bleeding after overnight fast (∼16 h) or at an indicated time point. Serum samples were separated using microvette CB 300 (Sarstedt). Glucose concentration was determined at time of collection with a glucometer (BD Bioscience). Serum samples were prepared according to manufacturer's instructions for individual assay or diluted 1:20 in SDS loading buffer (50 mM Tris·HCl, pH 7.4, 2% SDS wt/vol, 6% glycerol wt/vol, 1% 2-mercaptoethanol vol/vol, and 0.01% bromophenol blue wt/vol) and subjected to Western blot analysis. Serum/tissue triglyceride (ThermoFisher Scientific), nonesterified free fatty acid (NEFA; Wako), insulin, leptin, and adiponectin (Millipore) were determined according to kit manufacturer's instructions.

Intraperitoneal glucose and insulin tolerance tests.

Separate cohorts of 8–10 13-wk-old transgenic (Tg) and control littermates were injected intraperitoneally with glucose (1 g/kg). Animals were fasted for 6 h prior to the glucose tolerance test (1). For the insulin tolerance test, food was removed 2 h prior to insulin injection. Serum samples were collected at the indicated time points shown in the results.

Protein purification.

Recombinant full-length mouse CTRP9, containing a C-terminal FLAG-tagged epitope, was produced in HEK 293T mammalian cells, as previously described (47).

Body composition analysis.

Body compositions of Tg and control littermates were determined using a whole body NMR instrument (EchoMRI) at the metabolic phenotyping core facility at Johns Hopkins University School of Medicine. EchoMRI analysis provided values for fat mass, lean mass, and water content.

Quantitative real-time PCR analysis.

RNAs were isolated from tissues using TRIzol (Invitrogen) and reverse transcribed using Superscript II RNase H-reverse Transcriptase (Invitrogen), according to the manufacturer's protocol. PCR primers used included Lcad forward 5′-TCTTTTCCTCGGAGCATGACA-3′ and reverse 5′-GACCTCTCTACTCACTTC TCCAG-3′; Mcad forward 5′-AGGGTTTAGTTT TGAGTTGACGG-3′ and reverse 5′-CCCCGCTT TTGTCATATTCCG-3′; CoxII forward 5′-GCCG ACTAAATCAAGCAACA-3′ and reverse 5′-CAA TGGGCATAAAGCTATGG-3′; CytoB forward 5′-CATTTATTATCGCGGCCCTA-3′ and reverse 5′-T GTTGGGTTGTTTGATCCTG-3′; and 18 S rRNA forward 5′-GCAATTATTCCCCATGAACG-3′and reverse 5′-GGCCTCACTAAACCATCCAA-3′. Analyses were performed on an Applied Biosystems Prism 7500 Sequence Detection System. Samples were analyzed in 25-μl reactions, according to the standard protocol provided in the SyBR Green PCR Master Mix (Applied Biosystems). All expression levels were normalized to the corresponding 18 S rRNA levels.

Indirect calorimetry.

CTRP9 Tg mice and control littermates (n = 7 or 8 per group) were used for simultaneous assessments of changes in daily body weight, energy intake (corrected for spillage), and whole body metabolic profile in an open-flow indirect calorimeter (Oxymax, Columbus Instruments), as described previously (30). LFD- and HFD-fed mice were evaluated in separate studies. Data were collected for 3 days to confirm acclimation to the calorimetry chambers (stable body weights and food intakes), and data from the fourth day in the Oxymax were analyzed. Rates of oxygen consumption (V̇o₂, ml·kg⁻¹·h⁻¹) and carbon dioxide production (V̇co₂) were measured for each chamber every 16 min throughout the studies. Respiratory exchange ratio (RER = V̇co₂/V̇o₂) was calculated by Oxymax software (v. 4.02) to estimate relative oxidation of carbohydrate (RER = 1.0) vs. fat (RER approaching 0.7), not accounting for protein oxidation. Energy expenditure was calculated as EE = V̇o₂ × [3.815+ (1.232 × RER)] (20), and normalized to lean body mass (kcal·kg⁻¹·h⁻¹) as recommended by Butler and Kozak (4). Average metabolic values were calculated within subjects, then averaged across subjects for statistical analysis by Student's t-test, with P ≤ 0.05, indicating significant group differences.

Physical activity.

Mice (n = 8 per group) were tested in open-field chambers with infrared beam arrays to detect movement [photobeam activity system (PAS)-open field; San Diego Instruments, San Diego, CA]. Locomotor activity was monitored for 24 h, in 12 bins of 2-h duration, and reported as total ambulatory activity (ambulatory-like beam break patterns in the center plus periphery).

Cell culture.

GripTite HEK 293 cells (Invitrogen) were cultured in DMEM containing 10% FBS, 2 mM l-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin. Rat H4IIE hepatocytes and rat L6 myocytes were cultured in DMEM containing 10% FBS and antibiotics. L6 myotubes were differentiated as previously described (32). Each experiment represents the combined data from three independently performed experiments. For the in vitro assay, free fatty acid (palmitate)/BSA conjugates were prepared as previously described (19). Briefly, a 20-mM solution of free fatty acids in 0.01 M NaOH were incubated at 70°C for 30 min, and fatty acid soaps were then complexed with 5% BSA in PBS at an 8:1 ratio of fatty acid to BSA. The conjugates were administered overnight to cultured H4IIE hepatocytes at concentrations indicated.

Fatty acid oxidation.

To measure fatty acid oxidation, we adapted the protocol as described by Buzzai et al. (5), in which oxidation of [9,10-³H]-palmitic acid results in formation of [³H]H₂O. In brief, rat L6 myotubes were incubated for 2 h in serum-free DMEM containing 0.2% BSA with recombinant CTRP9 (5 μg/ml) or vehicle buffer. Next, 0.2 μCi/ml [9,10-³H]palmitic acid (Moravek Biochemical) were added to the media and incubated for 60 min. The tritiated palmitate was oxidized to CO₂ and [³H]H₂O. After incubation, the medium was transferred to a tube containing equal volume of chilled (4°C) 10% trichloroacetic acid. Samples were mixed and incubated for 10 min at 4°C, and then centrifuged for 30 min at 4°C. After centrifugation, 400 μl of the supernatant were collected and combined with 55 μl of 6 N NaOH, and then transferred to a Micro Bio-spin chromatography column (Bio-Rad; cat. no. 732-6204), containing 0.5 g Dowex ion exchange resin (Sigma 217425). The [³H]H₂O would be selectively retained by the resin, while the hydrophobic, nonoxidized [9,10-³H]palmitic acid in the supernatant would pass through the column. The bound [³H]H₂O in the column was eluted with 1 ml dH₂O, and the elution was transferred to a liquid scintillation vial. The amount of ³H radioactivity was determined with a Beckman Coulter counter (model LS6000SC). The amount of [³H]H₂O collected indicates the extent of fatty acid oxidation.

Western blot analysis.

Tissue lysates were prepared in T-PER buffer (Pierce Biotechnology, Rockford, IL) with phosphatase and protease inhibitors added (Calbiochem). Protein concentrations were determined using the Coomassie Plus protein assay reagent (Thermo Scientific). Ten micrograms of protein from tissue lysates or 1 μl of serum were loaded and separated on a 10% bis-Tris NuPAGE gel (Invitrogen). Western blotting and quantification were carried out as previously described (43).

Statistical analysis.

All results are expressed as means ± SE. Statistical analysis was performed with Prism 5 software (GraphPad). Blood chemistry data were analyzed with two-tailed Student's t-tests between CTRP9 Tg and control littermates. Repeated-measures ANOVA were performed on body weights, as well as serum glucose and insulin measurements in various tolerance tests. Values were considered to be significant at P < 0.05.

Metabolic state affects circulating levels of CTRP9.

Expression and circulating levels of many adipokines change in response to alterations in energy state. Therefore, we first determined whether changes in nutritional states acutely affect circulating levels of CTRP9. Mice fed ad libitum and mice that were fasted/refed had three-fold higher transcript levels of CTRP9 in adipose tissue and four-fold higher circulating levels compared with overnight fasted animals (Fig. 1, A and B). In diet-induced obese male mice fed an HFD for 12 wk, we observed an ∼50% reduction in circulating levels of CTRP9 (Fig. 1D). However, reduced serum levels of CTRP9 did not result from a reduction in mRNA expression in the epididymal (visceral) adipose tissue (Fig. 1C). This suggests that diet-induced alteration in serum CTRP9 levels may be mediated by a posttranscriptional mechanism. Together, these data indicate that short-term changes in nutritional state, as well as chronic metabolic stress induced by a high fat-diet, alter circulating levels of CTRP9.

Generation of CTRP9 transgenic mice.

To address the in vivo metabolic function of CTRP9, we generated a Tg mouse model overexpressing HA epitope-tagged CTRP9. In mice, endogenous CTRP9 mRNA is expressed predominantly by adipose tissue, with lower expression levels in other tissues. Within adipose tissue, both adipocytes and cells of the stromal vascular fraction express CTRP9 mRNA (47). Therefore, expression of the CTRP9 transgene was driven by a ubiquitous promoter (Fig. 2A). In the Tg mouse line, CTRP9 expression was detected in the adipose tissue, skeletal muscle, heart, brain, and kidney (Fig. 2B). Despite comparable expression of CTRP9 transgene mRNA in various mouse tissues, a substantially higher level of CTRP9-HA protein was detected in the skeletal muscle and heart compared with other tissues (Fig. 2D). As expected, overexpression also resulted in a significant increase in circulating levels of CTRP9, ∼5-fold greater than baseline serum levels in wild-type mice (Fig. 2C). This gain-of-function mouse model enabled us to explore the long-term metabolic consequences of elevated circulating levels of CTRP9.

CTRP9 Tg mice are lean and resistant to diet-induced weight gain.

Although comparable in body weight at weaning (4 wk of age), Tg mice fed an LFD consistently gained less body weight over time (Fig. 3A). By 18 wk of age, Tg mice fed an LFD were ∼22% lighter than control littermates. Remarkably, when fed an HFD, the body weight phenotype of the animals became overtly pronounced. While wild-type control littermates became progressively obese over time on an HFD as expected, Tg mice were resistant to weight gain (Fig. 3, B and C). In fact, by 18 wk of age, Tg mice (26.9 ± 2.4 g) were just half the body weight of control littermates (50.4 ± 3.7 g). Strikingly, Tg mice on an HFD gained less body weight and accumulated similar amounts of fat mass (percent of body weight) compared with Tg mice on an LFD (Fig. 3, A and B). These data have been independently confirmed in three separate cohorts of Tg mice and their WT littermate controls. Thus, the beneficial and protective metabolic function of CTRP9 was revealed when mice were challenged with an HFD to induce metabolic stress.

Reduced adiposity and adipocyte size in CTRP9 Tg mice.

Differences in body weight between WT and Tg mice fed an HFD could result from reduced accumulation of fat mass. Indeed, body composition analysis using quantitative NMR revealed that Tg mice had significantly less fat mass compared with WT controls (Fig. 3D), accounting for the lower body weight of Tg mice. In contrast, we observed an increase in percent lean mass (∼9%) in Tg mice fed an HFD relative to littermate controls (Fig. 3E). Reduction in adiposity appeared more striking in subcutaneous (inguinal fat pad) compared with visceral fat depots (gonadal fat pad) in Tg mice relative to littermate controls (Fig. 4, A and C). Further, the size of adipocytes was significantly smaller in Tg mice (Fig. 4, B and D); again, the difference was more pronounced in subcutaneous fat pad. We quantified the number of adipocytes in a given random ×20 magnification field. In the gonadal fat pad, there were 79 ± 14 (WT) vs. 183 ± 27 (Tg) adipocytes per ×20 field (n = 9 for WT and n = 6 for Tg mice). In the subcutaneous fat pad, there were 101 ± 9 (WT) vs. 346 ± 39 (Tg) adipocytes per ×20 field (n = 9 for WT and n = 6 for Tg mice). On the basis of these data, we estimated that the adipocytes found in the gonadal and subcutaneous fat depots of Tg mice are, on average, 2.3 and 3.4 times smaller than the WT counterparts, respectively. Smaller adipocytes are associated with improved metabolic profile and insulin sensitivity of white adipose tissue (6, 16, 34, 38). We also looked for potential “browning” of the adipose compartment; we observed no differences in the mRNA levels of key marker genes (UCP-1, Elovl3, Otop1, Cox7a1) previously shown to be associated with brown adipocyte-like cells within white adipose tissue (3).

CTRP9 Tg mice show enhanced energy expenditure.

Several mechanisms could account for the lean phenotype of CTRP9 Tg mice, such as differences in food intake, voluntary physical activity levels, and/or energy expenditure. Indirect calorimetry analyses were carried out to determine the consequences of CTRP9 overexpression on whole-body energy balance. On an LFD, WT and Tg mice consumed comparable amounts of food pellets and showed no differences in physical activity levels, V̇o₂, RER, and energy expenditure (Fig. 5). However, Tg mice fed an HFD consumed fewer calories (Fig. 6A) and were physically less active during the dark photocycle compared with littermate controls (Fig. 6B). Further, Tg mice fed an HFD had increased oxygen consumption (Fig. 6C) and increased carbon dioxide production (Fig. 6D) relative to littermate controls, indicating an enhanced metabolic rate. As expected, mice fed an HFD had lower RER compared with mice fed an LFD, due to a greater oxidation of lipid over carbohydrate substrates. Within the HFD group, CTRP9 Tg mice had a lower RER compared with littermate controls (Fig. 6E), indicating greater fat oxidation. Because of increased metabolic rate, Tg mice fed an HFD had a modest overall increase in whole body energy expenditure compared with WT mice (Fig. 6F). These data suggest that a combination of reduced food intake and increased basal metabolism over a 14-wk period on HFD could account for the remarkable differences in body weight and percent fat and lean mass between WT and CTRP9 Tg mice (Fig. 3).

CTRP9 enhances AMPK activation and promotes skeletal muscle fat oxidation.

To uncover the mechanism by which CTRP9 promotes energy expenditure, we examined possible signaling pathways and enzymes that regulate fat oxidation in skeletal muscle. Consistent with enhanced fat oxidation, we observed a three-fold increase in the expression of fatty acid oxidation enzyme genes (Lcad and Mcad) in the skeletal muscle of Tg mice relative to control littermates (Fig. 7A). Further, there was over a twofold increase in the expression of mitochondrion-specific genes (CoxII and CytoB) in the skeletal muscle of Tg mice (Fig. 7B). As with the mRNA expression data, mitochondrion-specific protein (e.g., COX IV) levels were also significantly increased (approximately twofold) (Fig. 7C), indicating greater mitochondrial content in the skeletal muscle of Tg mice on an HFD. Increased fat oxidation consequently resulted in lower triglyceride content in the skeletal muscle of Tg mice relative to control littermates fed an HFD (Fig. 7D).

AMPK phosphorylation (at Thr-172) and activation are known to increase mitochondrial biogenesis and muscle fat oxidation (50, 53). In the skeletal muscle of CTRP9 Tg mice, AMPKα (the catalytic subunit) was hyperphosphorylated at Thr-172 relative to WT controls (Fig. 8A), indicating enhanced AMPK activation. In accordance with a direct effect of CTRP9 in promoting muscle fat oxidation in vivo, purified recombinant CTRP9 likewise induced AMPKα (Thr-172) phosphorylation (Fig. 8B) and enhanced fatty acid oxidation in differentiated rat L6 myotubes (Fig. 8C). Treatment of myotubes with compound C, an AMPK inhibitor (52), abolished the ability of CTRP9 to enhance fatty acid oxidation (Fig. 8C), confirming that CTRP9 activates AMPK signaling to control fat oxidation. Together, these data indicate that CTRP9 promotes energy expenditure in vivo by enhancing skeletal muscle fat oxidation via AMPK activation and by increasing mitochondrial content.

It has been increasingly appreciated that brown adipose tissue also plays an important role in whole body energy expenditure by promoting fat oxidation (36). Although the white adipose tissue mass was greatly reduced in Tg mice (Fig. 3D), we did not observe any changes in interscapular brown adipose tissue mass (normalized to body weight) nor any differences in the expression levels of uncoupling protein-1 (UCP-1) mRNA and protein in brown fat isolated from WT and Tg mice fed an HFD (data not shown), indicating minimal contribution by brown fat to overall differences in energy expenditure observed between WT and Tg mice.

CTRP9 decreases hepatic lipid accumulation.

While HFD leads to hepatic triglyceride accumulation and, consequently, fatty liver (hepatic steatosis) in WT mice, overexpressing CTRP9 substantially reduced hepatic triglyceride levels by ∼30% (Fig. 9, A and B). Circulating ketone levels, a product of hepatic fatty acid oxidation, were elevated in Tg mice relative to WT controls (Table 1), indicating enhanced fat oxidation in the liver. This may account, in part, for lower hepatic triglyceride content in Tg mice. Consistent with the in vivo data, recombinant CTRP9 treatment reduced basal, as well as palmitate-induced lipid accumulation in cultured H4IIE hepatocytes (Fig. 9C). Together, these results suggest that skeletal muscle and liver are target tissues of CTRP9 in vivo and are likely responsible for oxidizing a substantial amount of the available lipid substrates.

Improved metabolic profiles of mice overexpressing CTRP9.

Enhanced fat oxidation coupled with reduced hepatic and skeletal muscle triglyceride levels are predicted to improve the metabolic profiles of CTRP9 Tg mice. Indeed, on an HFD, overnight-fasted Tg mice had markedly reduced insulin and glucose levels compared with littermate controls (Table 1), indicating enhanced insulin sensitivity in Tg animals. As expected from a reduction in fat mass, serum leptin levels in Tg mice were correspondingly decreased relative to WT controls. Serum glucagon levels were also reduced in Tg mice, likely a consequence of decreased insulin levels. On an HFD, Tg mice have higher circulating levels of triglycerides and no significant difference in NEFA. Circulating levels of adiponectin were comparable between WT and Tg mice. However, for the LFD-fed mice, only fasting insulin levels were different between WT and Tg mice (Table 1).

CTRP9 prevents HFD-induced insulin resistance.

Assessment of whole-body insulin sensitivity using the homeostatic model assessment insulin resistance index (23) revealed a substantial reduction in insulin resistance in CTRP9 Tg mice relative to WT controls fed an HFD (data not shown). To confirm that Tg mice were, indeed, more insulin-sensitive, an intraperitoneal glucose tolerance test (GTT) was performed. Mice overexpressing CTRP9 were much better at handling a glucose load, with a significantly higher rate of glucose disposal in the peripheral tissues (Fig. 10, E and F). The rate of glucose excursion in GTT is dependent on the magnitude of insulin secretion from the pancreatic β-cells in response to glucose challenge, as well as peripheral tissue insulin sensitivity. Therefore, insulin levels were measured. The magnitude of insulin secretion in Tg mice in response to the GTT was significantly lower (Fig. 10, G and H), despite a much greater rate of glucose disposal in these animals. These results indicate that CTRP9 Tg mice on an HFD are, indeed, more insulin-sensitive compared with littermate controls. Interestingly, although both Tg and WT mice fed an LFD showed comparable glucose disposal rate in GTT (Fig. 10, A and B), Tg mice did so with reduced insulin secretion during GTT (Fig. 10, C and D), indicative of enhanced insulin sensitivity in the peripheral tissues of these animals.