Uncoupling protein-3 (UCP3) mRNA expression in reconstituted human muscle after myoblast transplantation in RAG2-/-/gamma c/C5(-) immunodeficient mice.

Uncoupling protein-3 (UCP3), which is expressed abundantly in skeletal muscle, is one of the carrier proteins dissipating the transmitochondrial electrochemical gradient as heat and has therefore been implicated in the regulation of energy metabolism. Myoblasts or differentiated muscle cells in vitro expressed little if any UCP3, compared with the levels detected in biopsies of skeletal muscle. In the present report, we sought to investigate UCP3 mRNA expression in human muscle generated by myoblast transplantation in the skeletal muscle of an immunodeficient mouse model. Time course experiments demonstrated that 7-8 weeks following transplantation fully differentiated human muscle fibers were formed. The presence of differentiated human muscle fibers was assessed by quantitative PCR measurement of the human alpha-actin mRNA together with immunohistochemical staining using specific antibodies for spectrin and the slow adult myosin heavy chain. Interestingly, we found that the expression of UCP3 mRNA was dependant on human muscle differentiation and that the UCP3 mRNA level was comparable with that found in human muscle biopsies. Moreover, the human UCP3 (hUCP3) promoter seems to be fully functional, since triiodothyronine treatment of the mice not only stimulated the mouse UCP3 (mUCP3) mRNA expression but also strongly stimulated the hUCP3 mRNA expression in human fibers formed after myoblast transplantation. To our knowledge, this is the first time that primary myoblasts could be induced to express the UCP3 gene at a level comparable of that found in human muscle fibers.

Uncoupling protein-3 (UCP3) 1 is a mitochondrial membrane protein that is predominantly expressed in human and rodent skeletal muscle and brown fat. The gene encoding UCP3 is located on human chromosome 11 and mouse chromosome 7 (1). The UCP3 gene encodes for a protein with 60% homology to the brown fat-specific mitochondrial uncoupling protein UCP1 (2,3). The sequence similarity between UCP1 and UCP3 suggested to us that UCP3, like UCP1, might be a mitochondrial uncoupling protein involved in adaptive thermogenesis and energy expenditure in muscle. Moreover, biochemical experiments based on the overexpression of UCP3 in yeast or reconstituted transport activity of UCP3 in liposomes have confirmed the uncoupling activity of the UCP3 protein (4,5). Furthermore, a correlation between the increase in UCP3 protein in skeletal muscle and the nonphosphorylating mitochondrial respiration rates was demonstrated in rats by modulating their metabolic status via triiodothyronine (T3) levels.
Despite accumulating evidence that UCP3 plays a role as a mitochondrial uncoupling protein, UCP3 does not appear to be involved in adaptive thermogenesis in rodents in response to cold exposure and diet. Moreover, the role of UCP3 in energy expenditure and uncoupling effect is subject to controversy. First, homozygous UCP3 Ϫ/Ϫ mice lacked the obese phenotype and displayed an unchanged metabolic rate in response to cold exposure, fasting, stress, or T3 (6,7). Second, transgenic mice overexpressing human UCP3 (hUCP3) in skeletal muscle are hyperphagic and nonobese with increased energy expenditure (8), a phenotype that could result from some sort of UCP3 expression and uncoupling artifact (9) as had already been described in yeast (10). In humans, even though their impact on these phenotypes seems to be modest, several genetic analyses have shown an association between UCP3 gene polymorphism in basal lipid oxidation reduction (11) or body mass index in relation to morbid obesity or physical activity (12,13).
A new proposed role for UCP3 arose from several studies in rodents and humans showing that the UCP3 gene was upregulated in physiological situations associated with high free fatty acid plasma levels such as starvation (14,15), postnatal development (16), or high fat diets (17). Fatty acids themselves when administered in vivo induce UCP3 gene expression in both rodents and humans (18,19). Altogether, these data favor the hypothesis that the primary function of UCP3 in skeletal muscle may be to either passively or actively regulate the utilization of lipids as a fuel substrate.
In muscle, UCP3 gene expression is regulated both by to the metabolic and the hormonal status. Multiple hormonal treatments in vivo, such as leptin, thyroid hormones, glucocorticoids, and especially peroxisome proliferator-activated receptors agonists, cause dramatic changes in UCP3 gene expression in muscle (14) (18). Although studies have been carried out on the UCP3 promoter (20), the hormonal signals and responsive elements regulating UCP3 gene transcription in skeletal muscle have not yet been identified. One of the problems in char-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
acterizing this promoter is due to the fact that virtually all cultured muscle cells express little if any UCP3. These data suggest that specific transcriptional programs generating full muscle phenotype are exclusively activated in vivo but not in cultured muscle cells. Alternatively, in vivo muscle fibers might receive external signals triggering UCP3 gene expression (16).
To study the mechanisms regulating hUCP3 gene transcription and identify cis-/trans-acting DNA elements, we searched for an experimental model that would express a hUCP3 mRNA at levels similar to that found in human skeletal muscle biopsies. In the present work, we have injected human myoblasts into the tibialis anterior (TA) of immunodeficient mice RAG2 Ϫ/ Ϫ/ ␥ c/C5 Ϫ (21). In this in vivo environment, injected cells proliferate and fuse to form muscle fibers that become innervated and contract. In these reconstituted muscles, we were able to detect normal levels of hUCP3 gene expression. Moreover, the hUCP3 promoter appears to be fully functional, since the effect of T3 on hUCP3 mRNA expression was similar to that found on mouse UCP3 (mUCP3) mRNA expression. To our knowledge, this is the first model that permits both the characterization of the hUCP3 promoter and the physiological function of the putative uncoupling protein UCP3.

EXPERIMENTAL PROCEDURES
Human Myoblasts Origin and Culture-A biopsy from the quadriceps muscles of a 5-day-old infant was obtained during autopsy in accordance with the French legislation on ethical rules. From this biopsy, satellite cell populations were isolated from biopsies as described previously (22,23). These cells were expanded in growth medium, which consists of Ham's F-10 (Invitrogen) supplemented with 50 g/ml of gentamycin and 20% fetal calf serum (Biomedia), and were called CHQ5B. The cells displayed a myogenic purity of 80%, as assessed by desmin staining (21,24). To induce myotube formation, confluent cultures were cultivated for 10 days in Dulbecco's modified Eagle's medium supplemented with 10 g/ml of insulin (Sigma) and 100 g/ml of transferrin (Invitrogen).
Animals-Fifty-five immunodeficient RAG2 Ϫ/Ϫ / ␥ c Ϫ mice, 2-3 months old, were used in this study as recipients for human myoblast implantation. All experiments were carried out in the specific pathogen-free animal facilities at the Pasteur Institute.
Cell Preparation and Myoblast Transplantation-The cells used for transplantation (CHQ5B) were expanded to 20 population doubling levels. Prior to injection the cells were trypsinized, centrifuged (360 ϫ g) and resuspended in growth medium. Sufficient aliquots of cells were prepared in siliconized Eppendorf tubes. Following additional centrifugation, the supernatant was aspirated leaving a pellet containing 5 ϫ 10 5 cells per injection. For implantation, the mice were anesthetized with hypnorm/hypnovel. The TA muscles of both hindlimbs were exposed, and the proximal and distal portions of the muscle received a localized frozen lesion for 10 s (25). This was repeated three times before injecting the cells using a 5-l Hamilton syringe. The cells were injected into three to five different sites along the length of the muscle. The skin was then closed with fine sutures. Mice were sacrificed at different weeks post-implantation, and the TA muscles were dissected, mounted in gum tragacanth (6% in water; Sigma), and frozen in isopentane precooled in liquid nitrogen or kept for mRNA extraction. Surgical procedures were performed under sterile conditions and in accordance with the legal regulations in France.
Immunofluorescence-Eight weeks after implantation, immunofluo- Images were digitalized using the Photometrics and MetaView image analysis system. Treatment by T3-Mice received a single T3 injection at 2.5, 10, and 100 g/100 g body mass (in 100 l), whereas the control received the vehicle only (5 mM NaOH). Injections were given intraperitoneally in the morning. The mice were sacrificed 24 h later, and the TA muscles were dissected and stored frozen for mRNA extraction.
RT-PCR Experiments-Total RNA was isolated from the TA of mice using the RNAXel kit (Eurobio, Paris, France). One microgram of total RNA was reverse-transcribed using the Moloney murine leukemia virus reverse transcriptase in 20 l of its own buffer (Invitrogen) and oligo(dT) at 37°C for 1 h. Quantitative real-time PCR was performed with a LightCycler using FastStart DNA Master SYBR GreenI (Roche Molecular Biochemicals). The cDNA product was amplified in a total volume of 20 l with 0.5 M of each primer, 4 mM MgCl 2 (final concentration). All the primers were designed to anneal and amplify only mRNA. Moreover, primers were chosen to distinguish between human and mice sequences. The sequence primers are presented in Table I. A negative control for PCR consisted of omitting cDNA in the reaction tube. hUCP3 and ␣-actin levels were normalized to human ␤2-microglobulin and mUCP3 was normalized to mouse ␤2-microglobulin. This was noted REL for relative expression level.

Comparative Study of UCP3 mRNA Expression in Cultivated
Human Myotubes and Skeletal Muscle Biopsies-UCP3 mRNA levels were evaluated in cultures of human myoblasts and myotubes as well as in human muscle biopsies using real-time RT-PCR. As reported in Table II, the levels of UCP3 expressed in both myoblasts and myotubes were very much lower than that expressed in muscle biopsies. Indeed, in human muscle biopsies, the level of UCP3 mRNA was 500-and 423-fold greater than that measured either in myoblasts or myotubes. There was virtually no difference between undifferentiated and differentiated myoblast. This would suggest that the in vitro culture conditions do not allow sufficient maturation to induce UCP3 expression to the level measured in mature skeletal muscle fibers.
Human Muscle Regeneration in Immunodeficient RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ Mice after Myoblast Transplantation-Human myoblasts were injected into the TA muscle of an immunodeficient RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ mouse. The presence of human muscle fibers in regenerated TA muscle was verified by immunohistochem-   istry using a species-specific (human, non-mouse) antibody against spectrin that specifically stains the basal lamina fibers (Fig. 1A). All mouse muscles studied contained donor-derived (human) tissue after 1 week of transplantation (data not shown). The formation of muscle fibers containing human nuclei was examined by immunohistochemistry 8 weeks after cell transplantation, using the differentiation marker adult slow MHC and mRNA quantification of the ␣-actin marker using real-time RT-PCR. The co-expression of MHC and spectrin in skeletal muscle sections could be clearly seen 8 week after transplantation (Fig. 1B). The human ␣-actin was amplified by real-time RT-PCR using human-specific primers that did not cross-react with the mouse ␣-actin. Fig. 2 shows that 8 weeks after transplantation, the ␣-actin mRNA level reached a level similar to that measured in the human muscle biopsies, whereas in myoblast and myotubes CHQ5B cultured in vitro, the ␣-actin mRNA level was, respectively, 1272-and 315-fold less than in human muscle biopsies. PCR amplification of the ␣-actin product was not due to genomic contamination, since PCR amplification without reverse transcriptase did not produce any fluorescent signal. The specificity of the PCR products was assessed by Southern blotting with an internal oligonucleotide probe.
Together these results demonstrate that transplantation of human myoblasts into RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ mouse muscles produces fully differentiated human muscle fibers that express UCP3.
Recovery of hUCP3 mRNA Level in Implanted Human Myoblasts-The hUCP3 mRNA level was measured immediately after myoblast transplantation and at 1-, 2-, 4-, 6-, 7-, 8-, 9-, and 10-week intervals (Fig. 3). These results show that the hUCP3   4. Effect of T3 on UCP3 mRNA level. T3 was injected intraperitoneally into RAG2 Ϫ/Ϫ / ␥ S Ϫ/Ϫ mice, and 24 h later transplanted tibialis muscles were removed. Mice received a single T3 injection at 2.5, 10, and 100 g/100g body mass (in 100 l), whereas the other received vehicle (5 mM NaOH). UCP3 mRNA expression was measured by real-time RT-PCR. m␤2m, mouse ␤2-microglobulin; h␤2m, human ␤2-microglobulin. shows an example of a mature spectrin-positive human fiber that also expresses the slow MHC. The arrows indicate an immature human fiber that is negative for slow MHC. Bar ϭ 60 m. mRNA level increased between 2 and 7 weeks after transplantation and then reached a plateau at 7-8 weeks, at which time hUCP3 mRNA levels in human implants and human muscle biopsies were comparable (Fig. 3).
T3 Has a Similar Effect on hUCP3 and mUCP3 mRNA Expression-Eight weeks after transplantation, mice were injected intraperitoneally with T3, and 24 h later the TA muscles were removed. The in vivo effect of T3 on both human and mouse UCP3 mRNA expression is presented in Fig. 4. T3 induced a dose-dependent increase in hUCP3 mRNA levels with 2-, 6.7-, and 8.5-fold induction at 0.1, 2.5, and 10 mg/kg T3, respectively (Fig. 4A). Similarly, T3 induced a dose-dependent increase in mUCP3 mRNA levels with 3.6-, 4.6-, and 6.9-fold induction at 0.1, 2.5, and 10 mg/kg T3, respectively (Fig. 4B). DISCUSSION In this study we described for the first time an efficient model study UCP3 gene expression in human myoblasts. In this model, following transplantation of human CHQ5B myoblasts into the TA muscle of immunodeficient mice RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ , human muscle fibers were formed. When human myoblasts were cultured using standard in vitro culture conditions, UCP3 gene expression was hardly detectable compared with its expression in vivo. Results from quantitative RT-PCR experiments showed that in human muscle biopsies the level of UCP3 mRNA was 500-and 423-fold greater than that measured in either myoblasts or myotubes, respectively. Such a low level of expression of UCP3 mRNA is a classical feature of muscle cells in culture and has also been observed in rat L6 and murine C2C12 myotubes (26,27). In vitro differentiated skeletal muscle cells form large multinucleated myotubes, but their gene expression profiles are markedly different to that of mature muscle fibers in vivo. For example, GLUT-4 is poorly expressed in myotubes (28), whereas strong expression can be measured in muscle tissue.
The most important finding in the present study is that the hUCP3 mRNA level was fully restored by transplantation of human CHQ5B myoblasts in the TA of an immunodeficient mouse. This model permits transplantation tolerance for periods of time sufficient to allow maturation of donor-derived muscle cells. In this study, we first confirmed that transplantation of human myoblast into regenerating TA muscle of RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ mice produced fully differentiated human muscle fibers. The expression of human spectrin and differentiation markers, including the human ␣-actin and the slow adult MHC 8 weeks post-transplantation, confirmed the presence of regenerated mature human muscle fibers muscles in the TA of RAG2 Ϫ/Ϫ / ␥ c/C5 Ϫ mice. Second, we showed that transplantation of human myoblasts restores the expression of the human UCP3 gene from a very low level in myoblast to a level similar to that found in human muscle biopsies. In the reconstituted human skeletal muscle UCP3 gene expression followed a kinetic pattern that reached a plateau at 7-8 weeks posttransplantation. Taken together these results indicate that an increase in UCP3 gene expression at 7-8 weeks following human myoblast transplantation was associated with in vivo human muscle cell differentiation and maturation. The results also suggest that an additional specific in vivo transcriptional program is required to trigger UCP3 gene expression. Alternatively, in vivo biological signals absent in vitro might participate in UCP3 gene expression. Numerous studies have shown that T3 treatment strongly stimulates UCP3 gene expression in rodent and human skeletal muscles. As expected, we found that T3 treatment increased UCP3 mRNA in a dose-dependent manner in human implants, thus confirming the stimulatory effect of T3 on UCP3 gene expression. In addition, UCP3 gene regulation by T3 in humans and mice likely shares common features, as T3 treatment increases both human and mouse UCP3 transcripts to a similar level. Several mechanisms could account for the effect of T3 on UCP3 gene expression. The hUCP3 promoter contains a sequence that varies by a single base from the canonical thyroid response element (TRE) (20), but it is not known whether this putative TRE is functional. Interestingly, the mUCP3 promoter might also contain a similar TRE motif. Therefore the transplant of human myoblasts stably expressing the hUCP3 promoter region should help to functionally characterize the putative TRE-responsive element.
Despite the increasing number of reports on UCP3, the molecular and/or cellular mechanism controlling UCP3 gene expression in human skeletal muscle remains poorly understood. This might be due to the fact that all studies aiming at characterizing UCP3 promoter were carried out in vitro where the UCP3 gene is poorly expressed, thus making it difficult to identify the cis-regulatory elements controlling UCP3 gene transcription. The means to restore UCP3 gene expression in differentiated human myoblast in vitro has not yet been found, despite considerable research on this topic. Therefore, myoblast transplantation into immunodeficient mouse skeletal muscles appears to be the model of choice to study hUCP3 gene regulation and its promoter. Screening assays based on promoter knowledge should help to select compounds that stimulate hUCP3 gene expression with a view to potential application in obesity therapy.