| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Pediatrics, 2Department of Pathology, 3Department of Physiology and Functional Genomics, 4Department of Surgery, and 5Department of Physiological Sciences, University of Florida, Gainesville, Florida; and 6Department of Molecular and Integrative Physiology and 7Orthopaedic Research Labs, University of Michigan Medical School, Ann Arbor, Michigan
Submitted 22 April 2007 ; accepted in final form 15 June 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
-catenin signaling in skeletal muscles of adult diet-induced obese (DIO) rats. To this end, we generated a recombinant adeno-associated virus (rAAV) vector encoding murine Wnt10b cDNA. The long-term expression of rAAV1-Wnt10b was tested after intramuscular injection in the female DIO rat. Animals fed high-fat diet and treated with rAAV1-Wnt10b showed a sustained reduction in body weight compared with controls, and expression of Wnt10b was accompanied by a reduction in hyperinsulinemia and triglyceride plasma levels as well as improved glucose homeostasis. Nuclear magnetic resonance methods revealed that ectopic expression of Wnt10b resulted in a decrease in both global and muscular fat deposits in DIO rats. The long-range effect of locally expressed Wnt10b was also manifested through the increased bone mineral density. The detailed analysis of molecular markers revealed fibroblast growth factor-4 and vascular endothelial growth factor as possible mediators of the systemic effect of Wnt10b transgene expression. Our data demonstrate that altering Wnt/-catenin signaling in the skeletal muscle of an adult animal invokes moderate responses with favorable metabolic profile, bringing the notion of alternative therapeutic modality in the treatment of obesity, diabetes, and osteoporosis. Wnt10b; gene therapy; obesity; diabetes
Perhaps as many as 3% of all human genes are associated with energy homeostasis (37). A considerable fraction of these genes mediate adipose tissue development and maintenance (38). There is experimental and epidemiological evidence suggesting that mutations in some genes mediating adipogenesis are associated with human obesity and type 2 diabetes. Among such genes are members of wingless-type mouse mammary tumor virus integration site family (Wnts). Wnts are a family that regulate diverse developmental processes and adult tissue homeostasis. The interaction between secreted Wnt glycoproteins and their frizzled receptors inhibits phosphorylation of -catenin by GSK3 and prevents ubiquitination and proteosomal degradation of -catenin. The subsequent accumulation of free -catenin and its nuclear translocation followed by interaction with the lymphoid enhancer factor/T cell transcription factor (LEF/TCF) family induces transcription of various target genes.
Wnt10b, a recently described member in the Wnt family, is classified as a canonical Wnt, which activates the -catenin-TCF pathway. It plays an important role in the negative regulation of adipocyte differentiation in vitro and in vivo by shifting the balance of transcription factors in favor of osteoblastogenesis and/or myogenesis during development of precursor mesenchymal stem cells (2, 3, 16, 27, 40). FABP4-Wnt10b transgenic (TG) mice resist accumulation of adipose tissue on a high-fat (HF) diet, concomitant with improved insulin sensitivity compared with wild-type mice (27). In C57Bl/J mice, susceptibility to diet-induced obesity was correlated with a significant upregulation in adipose tissue of four genes (SFRP5, Dapper, Dickkopf, and Naked1) with demonstrated functions as inhibitors of Wnt signaling (21). In humans, mutations in Wnt10b are associated with obesity (6), and the variants of transcription factor 7-like 2 (TCF7L2, formerly TCF4) gene, another member of Wnt signaling pathway, confer risk of type 2 diabetes (10, 13, 42).
To elucidate the role of Wnt10b signaling in adult animals, and to explore the possibility of targeting adipogenesis as a long-term weight-reducing strategy, we used virus vector-mediated expression of Wnt10b in skeletal muscles of a diet-induced rat model of obesity. If Wnt10b, as a secreted factor, had a long-range systemic effect on adipogenesis, it would phenocopy FABP4-Wnt10b TG mice, with resistance of adult rats to weight gain on a HF diet and increased bone mass (2). If, on the other hand, Wnt10b acts locally, it would remodel skeletal muscle structure via activation of myogenic program (49). Surprisingly, in the face of a systemic body weight-reducing and insulin-sensitizing effect, we found little evidence of distal Wnt10b action and minimal effect on local muscle or bone structure.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Figs. 2A and 3A). BW was also measured at the time the rats were killed (210 days postinjection).
|
|
|
-actin promoter (Fig. 1A). This promoter has been selected to avoid transcriptional silencing of cytomegalovirus promoter (4), another strong promoter of choice in gene therapy applications. The overall vector design was validated in vitro in 3T3-L1 cells, whereupon the transduction with the vector at high multiplicity of infection (MOI) completely blocked the differentiation of fibroblasts into adipocytes (Fig. 1). rAAV vectors carrying transgenes were prepared using heterologous baculovirus system (48). Transgene cassettes derived via FastBac shuttle vectors were pseudotyped in AAV1 serotype capsids, using Baculovirus helpers generously provided by Dr. Kotin (National Institutes of Health). Vectors were purified by iodixanol gradient centrifugation and HighTrap Q HP column chromatography, as described earlier (56). Virus was then concentrated, and the buffer was exchanged to 100 mM sodium citrate and 10 mM Tris-Cl, pH 4.75, using centrifugal spin concentrators (Apollo, 150-kDa cutoff, 20-ml capacity). Physical rAAV particle titers were determined by dot blot assay. A single total dose of 1012 DNase I-resistant particles was administered per animal by injection into the medial area of the gracilis muscle of a hind leg, 250 µl per injection site, bilaterally. Isolation of mesenchymal stem cells from stromal vascular fraction of WAT. Mesenchymal stem cells (MSC) from WAT were isolated essentially as described by Negrel and Dani (31). The first pellet of stromal vascular cells was either flash-frozen in dry ice-EtOH bath or resuspended in RNAlater buffer for subsequent RNA isolation.
Bone histomorphometry. Left distal femur and the second lumbar vertebra were fixed in 10% phosphate-buffered formalin for 24 h and subsequently dehydrated in ethanol and embedded undecalcified in modified methyl methacrylate (1). Bone samples were then sectioned longitudinally with Leica/Jung 2,065 and 2,165 microtomes at 4 µm and stained according to the Von Kossa method with a tetrachrome counterstain (Polysciences, Warrington, PA). Histomorphometric measurements were performed in cancellous bone tissue of the distal femoral and lumbar vertebrae metaphyses in a sample area beginning 0.5 mm proximal to the growth plate-metaphyseal junction, which excluded the primary spongiosa. Bone measurements were performed with the Osteomeasure and Trabecular Analysis systems (Osteometrics, Atlanta, GA). The data collected include cancellous bone volume, cancellous bone surface, osteoblast and osteoclast numbers, osteoclast and osteoclast surfaces, osteoid volume, osteoid surface, osteoid thickness, trabecular thickness, trabecular number, trabecular separation, and other indexes of trabecular connectivity. The terminology used was that recommended by the Histomorphometry Nomenclature Committee of the American Society of Bone and Mineral Research (35).
Microcomputed tomography. Microcomputed tomography was performed on femoral bone essentially as described previously (2). A defined 3-mm3 cube of bone proximal to the growth plate in the distal femoral metaphysis and a 3-mm segment of cortical bone were evaluated at 18 micron voxel resolution using the GE Medical Systems Microview software.
RNA isolation. Total RNA from the white fat and liver was isolated by using TRIzol reagent (Invitrogen), following on-column of DNA digestion and concentration by using RNase-free ANase Set and RNeasy Mini Elute Cleanup Kit (Qiagen, Valencia, CA). Total RNA from the skeletal muscle was isolated by using RNeasy Fibrous Tissue mini kit (Qiagen). Tissues were homogenized by using Matrix D in FastPrep (FP120) homogenizer (Qbiogene, Carlsbad, CA). RNA integrity was verified by agarose gel (1.2%) electrophoresis with ethidium bromide staining.
Oxygen consumption. Oxygen consumption was assessed in up to six rats simultaneously with an Oxyscan analyzer (OXS-4; Omnitech Electronics, Columbus, OH). Flow rates were 2 l/min with a 30-s sampling time at 5-min intervals. The rats were placed into the chamber for 90 min with the oxygen consumption values for the last 30 min (when their oxygen consumption had reached stable levels) used in the calculations. Food was not available during the measurement. Results are expressed as mass-adjusted consumption (ml·min–1·kg0.67).
Relative quantitative RT-PCR analysis.
Unless specified, equal amounts of total RNA (in µg) isolated from a particular tissue and validated by gel electrophoresis were pooled for each treatment group (4 samples/group). Reaction Ready First Standard cDNA Synthesis kit (SuperArray Bioscience, Frederick, MD) was used for cDNA synthesis. cDNA was amplified by PCR using RT2 Real-Time SYBR Green PCR master mix (SuperArray Bioscience). Relative expression values between experimental and control animals were determined by the following rule: for each sample, we calculated difference between the CT (
CT) values for the gene of interest and housekeeping genes (GAPDH and -actin). For each pairwise set of samples, DIO/AAV1-Wnt10b and DIO/vehicle difference in CT values (CT) for gene of interest were normalized by the housekeeping gene. Finally, fold change of interrogated gene expression in each tissue was calculated as fold change = 
.
The rat Wnt Signaling Pathway RT2 Profiler PCR Array was utilized to interrogate the expression of 84 genes related to Wnt-mediated signal transduction. This set included cell surface receptors (members of the frizzled-1 signaling pathway), regulators of the Wnt-signaling pathway and competitive Wnt-binding antagonists, intracellular signaling molecules and target genes involved in growth regulation and proliferation, and genes involved in protein modification downstream of Wnt signaling, including genes involved in kinase and phosphatase activity and ubiquitination (SuperArray Bioscience) (Supplemental Table S1; Supplemental Material for this article is available at the AJP-Endocrinology and Metabolism web site). To interrogate specific genes mediating myogenesis and osteoblastogenesis pathways, as well as lipid and glucose metabolism, RT PCR primers were designed using Primer3 algorithm, available at the Whitehead Institute for Biomedical Research website.
Total protein purification and Western blot analysis.
Fast-frozen samples of skeletal muscle, liver, and white fat were homogenized using Matrix A and Matrix D in FastPrep in lysis buffer containing 50 mM Tris·HCl, pH 7.5, 120 mM NaCl, 1% Nonidet P-40, 10% glycerol, 10 mM Na4P2O7, 1 mM phenyl-methylsulfonyl fluoride, 1 mM EDTA, and 1 mM EGTA supplemented with protease inhibitor cocktail (set 3), and phosphatase inhibitor cocktail (set 2; Calbiochem, San Diego, CA). The homogenized lysates were incubated on ice for 1 h and clarified by centrifugation for 30 min at 14,000 rpm, 4°C. Normalized-for-protein concentration samples were separated using 12% polyacrylamide gel-SDS electrophoresis, transferred to a nitrocellulose membrane, and probed with the anti--catenin (1:1,500; BD Biosciences, San Jose, CA) or anti-GAPDH (1:1,500; Abcam, Cambridge, MA) followed by ECL anti-mouse IgG, horse radish peroxidase-linked, secondary antibodies (1:1,000, Amersham Biosciences, Little Chalfont, Buckinghamshire, UK). For fibroblast growth factor-4 (FGF4) expression analysis, incubation was performed in 5% BSA with anti-FGF-4 (1:2,000; Santa Cruz Biotechnology, Santa Cruz, CA), followed by HRP-conjugated chicken anti-goat IgG secondary antibodies (1:2,000; Immunology Consultants Laboratory, Newberg, OR).
The effect of Wnt10b-mediated signaling on adipogenesis in vitro. In vitro studies were performed on immortalized 3T3-L1 preadipocytes. Cells were incubated in DMEM-F-12 medium (GIBCO) with 10% FBS. Upon reaching confluence, cells were induced to differentiate by using the same medium supplemented with 0.5 mM 3-isobutyl-1-methylxanthine (Sigma), 1 µM dexamethasone (Sigma), and 1 µg/ml insulin (GIBCO) for 2 days, followed by 1 µg/ml insulin alone for 2 additional days. The effect of Wnt10b was tested upon infecting cells at 90% confluence 2 days prior to starting differentiation protocol.
Plasma hormone and lipid levels. Animals were fasted overnight, and a baseline tail blood sample was taken for hormone and lipid level measurement. Plasma leptin, insulin, and adiponectin levels were measured by using Rat Leptin ELISA kit, Rat/Mouse Insulin ELISA kit, and Rat Adiponectin ELISA kit (Linco Research Immunoassay, St. Charles, MO), respectively. Plasma triglyceride, nonesterified fatty acids (NEFA), total cholesterol levels, and corticosterone were determined by Serum Triglyceride Determination kit (Sigma), NEFA C kit, Total Cholesterol E kit (Wako, Richmond, VA), and Corticosterone DA (MP Biomedicals, Solon, OH), respectively.
Levels of VEGF and IGF-I were analyzed by Quantikine Rat VEGF Immunoassay and Quantikine Mouse IGF-I Immunoassay (R&D Systems, Minneapolis, MN), respectively.
Intraperitoneal glucose tolerance test. Glucose tolerance test (GTT) was performed at week 23 postvector injection. Testing was applied in rats that were fasted for 16 h prior to glucose loading (1.5 g/kg body wt). Unanesthetized rats were injected intraperitoneally, and blood samples were obtained from the tail vein before glucose challenge and at 15, 30, 60, 90, and 120 min after glucose challenge. Blood glucose concentration was measured with an Elite XL (Bayer, Elkhart, IN). Homeostasis model assessment (HOMA) index was calculated using the following formula: fasting serum insulin (µU/ml) x fasting serum glucose (mmol/l)/22.5.
High-resolution magnetic resonance imaging and spectroscopy. Body composition was determined on 4.7-T Bruker Avance spectrometer (Bruker BioSpin, Rheinstetten, Germany), using a custom built quadrature transmit/receive birdcage coil with 11.4 cm ID and 14 cm length. Respiratory-gated images were acquired from Wnt-treated (n = 5), DIO (n = 5), and DR (n = 5) rats using a 3D fast-spin echo imaging sequence [matrix = 256 x 256 x 128, no. of excitations (NEX) = 1, minimal repetition time (TR) = 300 ms, echo time (TE) = 7.5, echo train length = 16, field of view (FOV) = 15 x 10 x 5 cm2]. Following data acquisition, fatty tissue was segmented using 3D-Doctor algorithm (Able Software Corp) to determine overall body composition.
Muscle lipid levels of total creatine (TCr), intramyocellular (IMCL), and extramyocellular (EMCL) lipids were measured in both the tibilias anterior (TA) and soleus muscles of DIO, rAAV1-Wnt10-treated DIO, and control DR rats using localized magnetic resonance spectroscopy (MRS) at 11.1 T (Bruker BioSpin). All MRS measurements were made on a single rat hindlimb, using a custom-made quadrature transmit receive birdcage coil with 3.8 cm ID and single tuned to 470.5 MHz. High-resolution, T2-weighted multiplanar fast-spin echo images (TR = 3,000 ms, TE = 5 ms, echo train length = 8, matrix = 256 x 256, FOV = 28 x 28 mm2, slice thickness = 1.2 mm, NEX = 2) were acquired for the precise localization of a volume of interest (VOI) within the TA or soleus muscles. Nonwater-suppressed and water-suppressed localized MRS were acquired using point-resolved spectroscopy (VOI = 1.25 x 1.25 x 3.00 mm3, TR = 2,000 ms, TE = 16 ms, NEX = 512, NP = 4,096, sweep width = 5,952 Hz). Magnetic field homogeneity (<50 Hz) and water suppression were optimized over the VOI using localized shimming and chemical shift-selective water suppression. TCr, IMCL, and EMCL levels were determined using either Xwin-nmr (Bruker BioSpin) or jMRUI (http:/mrui.uab.es/mrui) using the AMARES algorithm. Metabolite chemical shifts were assigned on the basis of a chemical shift of 3.05 ppm for TCr and 1.3 and 1.5 ppm for IMCL and EMCL, respectively (32). In addition, TCr content was also determined on the basis of the ratio of TCr peak to the total proton density based on an unsuppressed point-resolved spectroscopy spectrum from the same volume. Extensive in vitro phantom work and in vivo imaging of VOI confirmed the accuracy of the VOI placement.
Statistical analysis.
Statistical analysis was conducted using unpaired Student's t-test with significance at P 
0.05. Area under the curve (AUC) in the GTT was calculated using SAS software, PROC EXPAND, trapezoid method (SAS 9.1.3 Help & Documentation, 2000–2004; SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
rAAV1-Wnt10b blocks adipogenesis in vitro. Vector was tested on immortalized 3T3-L1 fibroblasts, which differentiate with adipocytes recapitulating metabolic and endocrine functions of adipocytes in vivo. To investigate the effect of Wnt10b-mediated signaling on adipogenesis, cells cultivated at 90% confluence were infected with rAAV1-Wnt10b or rAAV1-GFP (negative control) at the MOI of 10,000. At 2 days postconfluence, cells were induced to differentiate by using appropriate hormonal cocktail. The differentiation status was monitored by Oil Red O staining of cells fixed in PBS containing 4% formaldehyde. We have shown that rAAV-mediated Wnt10b completely blocked differentiation of 3T3-L1 fibroblasts to adipocyte in vitro (Fig. 1, B and C) and that this effect is specific to Wnt10b transgene because rAAV1-GFP-infected cells retain the ability to differentiate (Fig. 1D).
rAAV1-Wnt10 treatment reduces BW and improves energy homeostasis.
To test metabolic effect of Wnt10b expression in vivo, vector was injected intramuscularly in 3-mo-old female rats. We have shown that a single intramuscular injection of 1012 physical particles of rAAV1-Wnt10b resulted in sustained (210 days), statistically significant (P = 0.029) reduction in BW accumulation (Fig. 2A). At the time the rats were killed (30 wk postinjection), the average BW of the treated group was 16% less than DIO control. Because of the initial BW difference between the DR and DIO cohorts, rats in Wnt10b group at the end of the experiment were heavier than in DR group despite similar accumulated BW (Fig. 2, A and B). These data are consistent with total fat volume obtained from 3D-MRI sequence (Fig. 2C) and weight of visceral fat pad (Supplemental Fig. S1). However, there was no significant difference in FI between DIO/AAV1-Wnt10b and DIO control group (Fig. 3A). DIO/AAV1-Wnt10b consumed, on average, 70.4 kcal/day, whereas DIO consumed 71.2 kcal/day.
Whole body oxygen consumption was assessed at day 140 postinjection by indirect calorimetry to estimate energy expenditure. Oxygen consumption was slightly elevated in the AAV1-Wnt10b-treated group compared with DIO control, not reaching significance or the levels in the DR control group (Fig. 3B). However, there was a significant difference in the amount of weight gained as a function of energy intake (feeding efficiency) between DIO and DIO/AAV1-Wnt10b groups (Fig. 3C). Feeding efficiency is the efficiency of the energy intake-body mass conversion, and the rate of this conversion was lower in DIO/AAV1-Wnt10b group.
Histopathology analysis. In rAAV1-Wnt10b-treated rats there was no inflammation in the muscles at the injection site, nor were there degenerative changes in myocytes. In contrast to DIO or DR control rats, there were no visible adipocytes in the muscle of Wnt10b-treated animals. Similarly, control rats (all DIO and 2 of 6 DR) had micro- and macrosteatosis of variable score in their livers, whereas none of the Wnt10b-treated rats had either. Average weight of the liver in DIO/rAAV1-Wnt10b rats was significantly lower compared with DR control group (Supplemental Fig. S1). Between three groups, there were no obvious differences in morphology of white (WAT) or brown (BAT) adipose tissue (data not shown), except for the total volume and weight (Fig. 2C and Supplemental Fig. S1).
rAAV1-Wnt10b treatment improves bones' quality. In the Wnt10b-treated rats, trabecular bone thickness displayed an upward trend (significant, P < 0.05 as determined by histomorphometry protocol, and not significant, P = 0.08 by microcomputed tomography; Table 1). This phenotype is consistent with that observed in the FABP4-Wnt10b transgenic mice. There were no apparent differences in cortical parameters, with the exception of bone mineral density (BMD) (P = 0.01; Table 2). These data suggest that Wnt10b treatment resulted in little to no change in geometric distribution and quantity of bone but that the quality of the bone had improved.
|
|
-catenin.
Wnt10b is a secreted protein and is a member of a family of growth factors acting locally in a paracrine or autocrine mode. In tissues, Wnt signaling is typically measured through Wnt-stimulated accumulation of -catenin, a surrogate marker, as well as downstream target gene expression. Western blotting analysis of -catenin revealed significantly higher levels of total -catenin in skeletal muscle at the injection site (Fig. 4, A and B) but not in the liver, fat, or MSC isolated from stromal vascular fraction in WAT (data not shown). The selective increase of -catenin in the muscle but not in other tissues was consistent with vector biodistribution analyzed indirectly by RT-PCR in RNA isolated from selected tissues. There was no evidence of vector-mediated expression in either WAT, liver, or MSC, whereas in the muscle there was a robust expression of the transgene that was interrogated with vector-specific primers (Fig. 4C). Because RT-PCR assay is a function of both vector distribution and the promoter activity in a particular cell or tissue type, we conclude that the systemic BW-reducing effect of rAAV-Wnt10 treatment is attributed mostly to the local ectopic expression of the transgene in the skeletal muscle of hind limbs.
|
Fig. 6B and Table 3) levels were also significantly lower in DIO/rAAV1-Wnt10b than in the DIO control group. Surprisingly, VEGF was reduced 2.7 times in the plasma of rAAV1-Wnt10b-treated rats, as determined by the ELISA assay (Table 3). However, the treatment did not significantly reduce plasma levels of glucocorticoids, NEFA, or cholesterol while drastically diminishing plasma lipids (Table 3).
|
|
|
It had previously been documented (19, 24, 43) that IMCL levels are associated with insulin resistance. In all three treatment groups and in both muscles tested, the EMCL/TCr levels were equal to or greater than IMCL/TCr levels (Fig. 5, A and C). In all cases, lipid levels were higher in the soleus compared with the TA (Fig. 5C). Rats expressing Wnt10b had the significantly lower IMCL/TCr (1.84 ± 1.23, P < 0.05) levels in the soleus compared with both DR (17.3 ± 7.5, P < 0.05) and DIO (4.56 ± 0.83, P < 0.05) groups. In the TA, a significantly lower IMCL/TCr level was found in the DIO-Wnt10b compared with DR (1.13 ± 0.13 vs. 3.6 ± 0.86, P < 0.05) and similar levels compared with DIO-GFP (1.13 ± 0.13 vs. 1.90 ± 0.49). Surprisingly, the DR group had the highest levels of IMCL in both soleus and TA. Because the measurements were conducted within the muscles distal to the injection site, the IMCL's decline in TA and soleus could be associated with systemic reduction in adiposity and not necessarily with ectopic Wnt10b expression in gracilis muscle.
rAAV1-Wnt10b improves glucose homeostasis. In fasted rats, basal glucose levels documented an impaired glucose tolerance in both DIO and DR groups (Table 3 and Fig. 6). Intermediately raised glucose levels in response to bolus injection were significantly higher in both DIO and DR control groups compared with the rAAV-Wnt10b-treated group (Fig. 6A). These differences in GT are consistent with the BW and insulin blood level (Fig. 6B) data for the respective groups of animals. The AUC vs. BW-scattered plot revealed overall metabolic improvement of Wnt10-treated rats (Fig. 6C).
HOMA has been widely employed in clinical research to assess insulin sensitivity. The HOMA value correlates well with clamp techniques and has been used frequently to assess changes in insulin sensitivity after treatment. Here we document improvement of insulin sensitivity in DIO/rAAV1-Wnt10b group as measured by significant (P = 0.029) decrease of HOMA index (Table 3).
Effect of Wnt10b treatment in obese rats.
Above, we have demonstrated the prophylactic effect of AAV1-Wnt10b treatment applied in young rats prior to development of obesity and glucose intolerance. To determine whether the same treatment could be successfully applied in already fat animals, we first induced an obese phenotype in rats by feeding a HF diet for 6 mo. When rats reached, on average, 470 g, we injected rAAV1-Wnt10b vectors intramuscularly as described above. At 
3.5 mo postinjection, the rate of BW accumulation in Wnt10b-treated rats had diminished, and, by the end of the experiment at 6 mo postinjection, the treated group lost, on average, 10% of BW compared with the control group (Fig. 7). Because of the small group size (n = 4 each) and the heterogeneity of the response, the difference in BW didn't quite reach statistical significance (P = 0.08). These data suggest that AAV-mediated expression of Wnt10b might induce a reduction in BW in preobese rats.
|
We also failed to document significant changes in myogenesis markers in muscle as well as osteoblastogenesis markers in both muscle and MSC. On the other hand, among genes mediating lipid and glucose metabolism in muscle we noticed the upregulation of sterol regulatory element-binding protein-1 (SREBP-1) and downregulation of phosphoenolpyruvate carboxykinase (PEPCK). Surprisingly, the increase in SREBP-1 expression did not lead to significant elevation of the expression of downstream target genes (among those tested were acetyl-CoA carboxylase, fatty acid synthase, stearoyl-CoA desaturase). In WAT, two genes were significantly downregulated: lipoprotein lipase (LPL) and Hsd11b1.
To validate the significance of the upregulated expression of FGF4 gene in muscle, WAT, and MSC, we analyzed FGF4 protein in plasma. Concurrently, rats treated with rAAV1-Wnt10b vector had significantly more FGF4 protein in the plasma, as assayed by Western blotting analysis (Fig. 8, A and B).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
-catenin signaling cascade, therefore, represents an attractive pathway for therapies targeting obesity, type 2 diabetes, and osteoporosis. Given the potential therapeutic implications, it is of considerable interest to develop experimental approaches allowing the modulation of Wnt10b in adult animal model. In the current report, rAAV vector had been utilized to upregulate Wnt10b in skeletal muscle of DIO rat.
We chose to target skeletal muscle in DIO rat to investigate whether sustained expression of the transgene might reduce the accumulation of ectopic fat in the form of IMCL. It was anticipated that Wnt10b transgene would act locally at the injection site to increase the amount of -catenin in myocytes and stem cells. To our surprise, Wnt10b-treated rats displayed systemic effect of diminished BW gain (Fig. 2A). The lower BW in the DIO/rAAV1-Wnt10b group was related to lower adiposity, as determined by MRI analysis in live rats prior to being killed (Fig. 2C) as well as measuring weight of WAT harvested at the time they were killed (Supplemental Fig. S1).
Another paradoxical observation was the fact that lower BW gain in Wnt10b group was not related to either food intake (Fig. 3A) or total oxygen consumption as an index of energy expenditure, with both parameters being not different among DIO the control and Wnt10b groups (Fig. 3B). The unchanged O2 was consistent with the weight of BAT that was similar in both the control and DIO/rAAV-Wnt10b group (Supplemental Fig. S1). Expression of Wnt10b from the adipocyte-specific promoter in FaBP4-Wnt10b TG blocks development of brown adipose tissue (27). There was no change in BAT in rAAV1-Wnt10b-injected adult rats, indicating either local mechanism of effector action or reflecting distinct cell lineages of WAT and BAT origin (44) that may have been differentially transduced with rAAV1 vectors.
Our analysis of the endocrine panel revealed that Wnt10b-treated rats had normalized levels of plasma triglycerides, cholesterol, glucose, and fasting insulin (Fig. 6B) while maintaining NEFA at the same or slightly elevated level (Table 3). Levels of circulating adipokines (leptin and adiponectin) in plasma are consistent with the diminished WAT mass, suggesting that the expression of these genes were not changed on a per-adipocyte basis. In the fa/fa rat model, increased expression of adiponectin in BAT seems to negate the reduction in WAT in some circumstances. Thus, the circulating adiponectin concentration can actually be increased in this obese rat vs. the lean controls (33). The decrease of adiponectin level in Wnt10b-treated mice documented in this report is consistent with the results published earlier by Wright et al. (53) and Longo et al. (27).
To elucidate the mechanism of improved glucose homeostasis, we measured muscle levels of TCr, IMCL, and EMCL in the tibialis anterior and soleus muscles using localized MRS. These muscles were probed because of the limitations imposed by hindlimb anatomy and the custom-made birdcage coil (3.8 cm ID) for MRS data acquisition. Nevertheless, even at the distal muscles relatively far from the injection site, there was a significant reduction in IMCL in both muscles relative to DR control group and in soleus compared with DIO control. A two- to threefold higher IMCL triglyceride content was found in the soleus muscle relative to that in tibialis anterior, a finding consistent with previous observations (36). Curiously, in the DR control group the IMCL content in muscle was considerably higher compared with the DIO group and apparently resulted from ectopic accumulation of lipids in muscle, causing impaired glucose tolerance (Fig. 6, A and C). The improved overall metabolic status was also consistent with much improved glucose tolerance, even compared with DR group (Fig. 6A). The scattered plot of AUC vs. BW showed that Wnt10b-treated rats had reduced adiposity and improved glucose tolerance. The insulin sensitivity, as assessed by the HOMA index, was also significantly improved in rats in the DIO/rAAV1-Wnt10b group.
Relatively large numbers of genes display an altered expression in response to Wnt signals in microarray analysis of tissue culture cells (39, 52). Modulating Wnt at early embryogenesis in TG or knockout animal models also resulted in dramatic effects, underscoring the functional relevance of these factors in a specific developmental context (3, 27, 39, 49). However, chronic experiments using vector-mediated transgene delivery in adult animals are quite different, whereupon only a few genes were shown to be modulated in response to ectopic expression of Wnt10b (Supplemental Table S1). For example, among Wnt-signaling pathway genes, only FGF4 had been significantly upregulated in WAT and MSC. Consistent with the enhanced expression, the amount of FGF4 protein in plasma had been shown to be increased significantly as well (Fig. 8). FGF4 had been previously identified as a direct downstream target for LEF1 and Wnt signaling (22). It is conceivable that Wnt10b-activated FGF4 may also function upstream of other signals, thereby linking Wnt and other distinct signaling pathways. An increase in BMD documented in rAAV1-Wnt10b-injected rats could be an example of such cascade initiated by Wnt10b through upregulated FGF4. Indeed, consistent with our data, Kuroda et al. (25) recently described an increase in BMD as well as cancellous bone mass in rats treated with exogenous FGF4. Curiously, other members of the diverse FGF family had recently been implicated in maintenance of energy balance. Specifically, FGF19 (9, 45) and FGF21 (17, 51) have demonstrated an ability to beneficially regulate metabolism.
Among tested genes mediating lipid and glucose metabolism, only a few (LPL, Hsd11b1, SREBP1c, and PEPCK) were significantly modulated. LPL is the major enzyme responsible for the hydrolysis of circulating triglycerides. The enzyme is found predominantly in adipose tissue, cardiac muscle, and skeletal muscle. Increased LPL activity increases the propensity for obesity and insulin resistance in a mouse model (7, 18). Generally, adipose tissue LPL correlates with serum insulin levels and the degree of insulin sensitivity. Overexpression of human LPL in mouse skeletal muscle resulted in elevated tissue triglyceride levels in two independent studies (18, 50). The muscle lipid content also correlates with muscle insulin resistance in rats and humans (46), and overexpression of LPL in muscle caused a reduction of muscle glucose disposal (8, 18). The reduced levels of the adipose tissue LPL expression and unchanged muscle LPL in Wnt10b-treated rats, therefore, correlate with the overall improved glucose homeostasis and reduction of adiposity in rats treated with Wnt10b.
Searching for a systemic effector positioned downstream of Wnt10b/-catenin signaling cascade, we assayed glucocorticoids in serum (Table 3) and also analyzed the expression of 11-hydroxysteroid dehydrogenase type 1 (11-HSD1) gene in WAT. Even though the circulating levels of the hormone had not changed, the tissue glucocorticoid exposure may have been altered independently of circulating levels by 11-HSD1, an enzyme that generated active glucocorticoid (cortisol in humans, corticosterone in rodents) within tissues, especially in adipose tissue. Adipose-specific overexpression of 11-HSD1 produces a phenotype analogous to the metabolic syndrome (with central obesity, insulin resistance, and dyslipidemia) (28). Knockout mice, homozygous for a deleted 11-HSD1 allele, are conversely protected from the metabolic consequences of obesity (20, 30). They have lower intracellular corticosterone levels, despite mildly elevated circulating levels (11, 55), and a favorable lipid profile (lower serum triglycerides and high HDL cholesterol) and resist hyperglycemia induced by stress and during high-fat feeding. Changes in glucocorticoid-dependent local gene expression are also seen in the organs responsive to glucocorticoids, resulting in decreased levels of gluconeogenic enzymes such as PEPCK. The overall lower adiposity in Wnt10b rats, therefore, is consistent with the reduction in the expression of 11-HSD1.
Another candidate for secondary messenger mediating systemic effect is VEGF, which was reduced in Wnt10b-treated animals 2.7-fold (Table 3). VEGF is necessary to initiate the formation of immature vessels by either vasculogenesis or angiogenic sprouting in the adult and during development, accounting for most of the angiogenic activity of adipose tissue. Mick et al. (29) had shown that VEGF synthesis in WAT is stimulated by insulin within physiological concentrations range. It is conceivable, therefore, that an almost twofold reduction in insulin levels in Wnt10b-treated animals caused the subsequent decline in VEGF, which, in turn, resulted in reduced levels of adipogenesis. Several diverse lines of evidence indicate that blood vessel development may influence adipocytes or adipogenesis. For example, treatment of mice with antiangiogenic factors decreased fat pad weights by 12 to 22% and decreased body weights in a dose-dependent and reversible manner (41), whereas the differentiation of vasculature precedes adipocyte differentiation (12, 14).
Pharmacological and genetic treatments that activate Wnt/-catenin signaling in mesenchymal precursors repress adipogenesis and stimulate osteoblastogenesis by repressing adipocyte transcription factors, stimulating osteoblast transcription factors, or both (23). In this report, in the face of pronounced physiological effect of reduced adiposity, improved glucose homeostasis, and increased BMD, we found little molecular evidence of a distal action of ectopically expressed Wnt10b. Altering Wnt/-catenin signaling early in ontogenesis causes profound developmental effects, whereas modulating it in adult organism, as shown in this report, invokes moderate response with favorable metabolic profile. Whether Wnt10b gene therapy could be applied in human trials depends on the safety considerations. Although to date no reports connecting human tumors to mutation or dysregulation of genes encoding Wnt ligands or receptors have been made, certain components within the Wnt pathway have been implicated. Pending subsequent comprehensive studies of the mechanisms and safety of the described gene therapy approach, our findings present an alternative therapeutic modality in the treatment of obesity, type 2 diabetes, and osteoporosis.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
, Diet-induced (DIO) group; 
, diet-resistant (DR) group; 
, DIO/rAAV1-Wnt10b group. Arrow shows the time of ip glucose tolerance test (GTT). B: actual BW at the time rats were killed (210 days postinjection). C: total volume of adipose tissue as determined by MRI analysis before the time when rats were killed. *P < 0.05 vs. DIO group.
