Glucose transporter expression in human skeletal muscle fibers

Transcription

1 Am J Physiol Endocrinol Metab 279: E529 E538, Glucose transporter expression in human skeletal muscle fibers M. GASTER, 1,2 A. HANDBERG, 1 H. BECK-NIELSEN, 1 AND H. D. SCHRØDER 2 Departments of 1 Endocrinology and 2 Pathology, Odense University Hospital, DK-5000 Odense, Denmark Received 22 July 1999; accepted in final form 22 March 2000 Gaster, M., A. Handberg, H. Beck-Nielsen, and H. D. Schrøder. Glucose transporter expression in human skeletal muscle fibers. Am J Physiol Endocrinol Metab 279: E529 E538, The present study was initiated to investigate GLUT-1 through -5 expression in developing and mature human skeletal muscle. To bypass the problems inherent in techniques using tissue homogenates, we applied an immunocytochemical approach, employing the sensitive enhanced tyramide signal amplification (TSA) technique to detect the localization of glucose transporter expression in human skeletal muscle. We found expression of GLUT-1, GLUT-3, and GLUT-4 in developing human muscle fibers showing a distinct expression pattern. 1) GLUT-1 is expressed in human skeletal muscle cells during gestation, but its expression is markedly reduced around birth and is further reduced to undetectable levels within the first year of life; 2) GLUT-3 protein expression appears at 18 wk of gestation and disappears after birth; and 3) GLUT-4 protein is diffusely expressed in muscle cells throughout gestation, whereas after birth, the characteristic subcellular localization is as seen in adult muscle fibers. Our results show that GLUT-1, GLUT-3, and GLUT-4 seem to be of importance during muscle fiber growth and development. GLUT-5 protein was undetectable in fetal and adult skeletal muscle fibers. In adult muscle fibers, only GLUT-4 was expressed at significant levels. GLUT-1 immunoreactivity was below the detection limit in muscle fibers, indicating that this glucose transporter is of minor importance for muscle glucose supply. Thus we hypothesize that GLUT-4 also mediates basal glucose transport in muscle fibers, possibly through constant exposure to tonal contraction and basal insulin levels. age; GLUT-1; GLUT-2; GLUT-3; GLUT-4; GLUT-5; immunocytochemistry; muscle fibers; ontogenesis Address for reprint requests and other correspondence: M. Gaster, Dept. of Pathology, Odense University Hospital, DK-5000 Odense, Denmark. THE FACILITATIVE TRANSPORT of glucose across the plasma membranes of mammalian cells is catalyzed by a family of six glucose transport proteins, designated GLUT-1 through -7, where GLUT-6 is a pseudogene (38). In human adult skeletal muscles, the mrna and the protein of GLUT-1 and GLUT-3 through -5 have been described (reviewed in Refs. 2, 4, 38, 40). GLUT-4 mrna and protein are the most abundant of the glucose transporter isoforms in human skeletal muscle, where GLUT-4 is believed to be responsible for insulinstimulated glucose disposal (23). GLUT-4 immunoreactivity is located mainly in intracellular vesicles in the basal non-insulin-stimulated state and is translocated to the sarcolemmal membrane and T-tubules as a response to insulin or contraction (28, 32, 33). In human skeletal muscle, GLUT-1 mrna is expressed at a substantially lower level than GLUT-4 mrna. In adult rat muscles, GLUT-1 immunoreactivity has been demonstrated in perineurial sheaths and connective tissue, as well as in the sarcolemma in some (15 17, 34) but not all (12, 25) studies. In a crude membrane fraction of rat skeletal muscle, 60% of GLUT-1 has been described as originating from intramuscular perineurial sheaths, and the remainder is predominantly from muscle fibers (16). GLUT-1 has been suggested to be of importance for basal glucose uptake in muscle cells (4, 37). The expression of GLUT-1 in adult human muscle has been investigated by methods detecting the protein (35, 50) and the mrna (2, 11, 24, 35, 39). However, interpretation of the results obtained from homogenates of human muscle tissue has been hampered by the presence of erythrocytes and perineurial sheaths within the muscle specimens, because both these tissues express very high concentrations of GLUT-1. Thus it has not yet been clarified whether GLUT-1 is expressed in human skeletal muscle cells. GLUT-3 was originally cloned from a fetal skeletal muscle library (27). GLUT-3 protein was below detection limits in vivo (14, 48). Recently, Stuart et al. (49) concluded from a study of muscle from human autopsies that the lack of GLUT-3 protein in previous studies was caused by a rapid degradation of the GLUT-3 protein. However, the cellular source of GLUT-3 was not addressed. GLUT-5 mrna is expressed at very low levels in human skeletal muscles (26). Hundal et al. (20) and Shepherd et al. (47) demonstrated GLUT-5 protein in human skeletal muscle by Western blotting. Immunofluorescence staining has indicated that GLUT-5 protein is localized in the plasma membranes of human skeletal muscle fibers (20). Thus much is known about glucose transporter expression in human skeletal muscle; however, the ex- 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 /00 $5.00 Copyright 2000 the American Physiological Society E529

2 E530 pression of GLUT-1 through -5 protein during fetal development has not been investigated until now. Moreover, further studies in adult muscle regarding the localization of expression mainly of GLUT-1 and GLUT-3, and possibly of GLUT-5, would improve our understanding of the mechanisms underlying glucose uptake in skeletal muscle tissue. The present study was initiated to investigate GLUT-1 through -5 expression in developing and mature human skeletal muscle. To bypass the problems inherent in techniques using tissue homogenates, we applied an immunohistochemical approach, employing the sensitive enhanced tyramide signal amplification (TSA) technique, to detect the localization of GLUT expression in human skeletal muscle. METHODS Study design. Fetal skeletal muscle specimens from 12, 14, 15, 18, 21, 22, 23, 24, and 40 (two cases) wk of gestation and muscle from 1- and 4-yr-old children were obtained from autopsies. Adult muscle biopsies were obtained according to the method of Bergström (3) from healthy adults aged 29 (n 7) and 64 (n 8) yr, respectively. Muscle tissue was mounted in Tissue Tek OCT-compound (Sakura, Torrance, CA) and frozen. The protocol was approved by the local ethical committees of Funen and Vejle counties, and informed consent was received from all adult subjects before participation. Muscle tissue was also fixed in phosphate-buffered formaldehyde and embedded in paraffin. Before the sections were subjected to the immunostaining protocols, antigen retrieval was performed by 15 min of microwave heating (600 W) in 10 mmol/l Tris with 0.5 mmol/l EGTA at ph 9.0 (T-EG buffer). Antibodies. The following primary antibodies were used: rabbit polyclonal antibody anti-glut-1 (AB1341, Chemicon, Temecula, CA), mouse monoclonal anti-glut-1 antibody (F18) raised against human erythrocyte membranes (42), and rabbit polyclonal anti-glut-2 (AB1342), rabbit polyclonal anti human-glut-3 (AB1345), rabbit polyclonal anti- GLUT-4 (AB1346), and rabbit polyclonal anti-glut-5 (AB1048). Erythrocytes were identified by the erythrocytespecific glycophorin A antibody (Dako, Glostrup, Denmark) (30). Immunocytochemical techniques. Glucose transporter immunoreactivity was visualized with the labeled streptavidin biotin (LSAB) technique and/or with the enhanced tyramide signal amplification (TSA) technique. LSAB. Immunostaining of GLUT-1 through -5 and glycophorin A was performed with horseradish peroxidase (HRP)- LSAB (41). In brief, 5- m cryosections of muscle were washed in Tris HCl-buffered saline (TBS), ph 7.4, and pretreated with 2% BSA in TBS for 10 min. Incubation with primary antibodies (dilutions: AB1341, 1:1,600; F18, 1:200; AB1342, 1:8,000; AB1345, 1:8,000; AB1346, 1:1,600; AB1048, 1:500; and glycophorin A antibody, 1:50) at room temperature for 30 min was followed by a 30-min incubation with biotinylated secondary antibodies: goat anti-rabbit Ig (E0432, Dako) or goat anti-mouse Ig, (E0433, Dako), both diluted 1:200. Finally, sections were incubated with HRPconjugated streptavidin (P0397, Dako), diluted 1:300 for 30 min using 3-amino-9-ethylcarbazole (AEC) as chromogen, followed by counterstaining with Mayers hematoxylin. Sections were dehydrated and mounted with Aquatex (Merck, Darmstadt, Germany). Antibodies were diluted in 1% BSA (Sigma, St. Louis, MO), and TBS was used for rinsing of sections between incubations. TSA. GLUT-1 immunoreactivity was also detected by a modification of the TSA technique described by Adams (1). TSA-Indirect (NEN Life Science, Boston, MA) in combination with LSAB was used as a detection system. In brief, 5- m cryosections of skeletal muscle biopsies were fixed in 4% phosphate-buffered formaldehyde for 5 min. After they were rinsed in tap water, blocking of endogenous biotin was achieved with the Dako-Kit X0590 (Dako), followed by incubation with 0.5% blocking reagent (NEN Life Science) in TBS buffer for 25 min at room temperature. Sections were incubated overnight at 4 C with anti-glut-1 (AB1341, 1:10,000), anti-glut-1 (F18, 1:10,000), anti-glut-2 (1:32,000), anti-glut-3 (1:25,000), anti-glut-4 (1:5,000), or anti-glut-5 (1:8,000). Subsequently, muscle sections were incubated for 30 min with biotinylated secondary antibodies [biotinylated goat anti-rabbit Ig (E0432), and biotinylated goat anti-mouse Ig (E0433), both diluted 1:200], followed by 10-min blocking of endogenous peroxidase activity using the ChemMate peroxidase blocking solution (S2023, Dako). Next, sections were incubated with HRP-conjugated streptavidin (P0397, diluted 1:300) for 30 min, followed by a 5-min reaction with biotinylated tyramide and a secondary 30-min incubation with HRP-conjugated streptavidin. HRP activity was developed for 20 min using AEC as chromogen. Immunostaining was followed by a brief nuclear counterstaining in Mayers hematoxylin, and finally, coverslips were mounted with AquaTex (Merck). Thorough rinsing between incubation steps was done in TBS with 0.05% Tween-20 (Sigma). The antibodies and HRP-streptavidin were diluted in the 0.5% blocking reagent in TBS. Controls. In control sections, the primary antibodies were omitted, as well as being replaced, by species-isotypical primary antibodies (Coulter, Miami, FL). Specificity of the GLUT-1 TSA immunoreactions was ensured by the fact that no immunostaining was found after preincubation of the anti-glut-1 antibodies with highly purified human erythrocyte membrane. All antibodies were tested on multiblock sections containing a variety of tissues besides muscle. Sensitivity. To compare the sensitivity of the LSAB and the TSA detection systems, we used the membrane test system described by Scopsi and Larsson (46) with a few modifications. In brief, PVDF (polyvinylidine difluoride) membrane strips (0.22 m pore size, Millipore, Glostrup, DK) were activated with methanol and then washed with MilliQ water and transfer buffer. Two-microliter droplets of highly purified erythrocyte membrane solution in decreasing concentrations (100, 20, 4, 0.8, 0.14, etc., g protein/ml) were spotted on the strips and dried at room temperature. The strips were first incubated in TBS (MilliQ water) with 0.1% Tween-20 for 15 min as a blocking step, followed by overnight incubation with the GLUT-1 antibodies diluted 1:5,000 at 4 C. The next day, the strips were rinsed with TBS (MilliQ water) with 0.1% Tween-20, fixed in 0.02% glutaraldehyde and stained by either the LSAB or the TSA methods. The sensitivity of the staining systems was determined as the minimum detectable amount of antigen. Detection efficiency. Detection efficiencies of the LSAB and the TSA methods used for visualization of immunoreactivity were compared by incubation of consecutive cryosections of adult human skeletal muscle with decreasing concentrations of primary antibody, followed by staining by use of either LSAB or TSA. Detection efficiency was determined as the maximal dilution of primary antibody allowing visualization of GLUT-1. Dilutions of primary anti-glut-1 antibody were (1:200, 1: :25,600) for LSAB and (1:1,600, 1:3, :409,600) for TSA. The detection efficiency of TSA was tested on unfixed and formaldehyde-fixed cryosections. Spec-

3 E531 ificity of the TSA immunoreactions was ensured by the fact that no immunostaining was found after preincubation of the anti-glut-1 antibodies with highly purified human erythrocyte membrane. Erythrocyte membrane preparation. Erythrocytes were washed 3 times with cold TBS and placed in hypotonic KCl solution [75 mm KCl containing 200 mol/l phenylmethylsulfonyl fluoride (PMSF), 1 mmol/l leupeptin, and 1 mmol/l pepstatin]. After centrifugation at 350 g for 10 min, the supernatant was centrifuged at 5,000 g for 1 h at 4 C. The pellet, containing crude membranes, was resuspended in membrane buffer (containing 250 mmol/l sucrose, 1 mmol/l EDTA, 200 mol/l PMSF, 1 mmol/l leupeptin, 1 mmol/l pepstatin, 5 mmol/l Na 2 HPO 4, 5 mmol/l NaH 2 PO 4, ph 7.4) and centrifuged at 10,000 g for 1 h. After resuspension of the pellet in membrane buffer, protein content was determined, and the preparation was stored at 80 C. Western blotting. Samples of 6 g protein of purified erythrocyte membrane were heated for 5 min at 90 C in sample buffer without -mercaptoethanol and separated on 12% SDS-polyacrylamide gels (Ready-gels, Bio-Rad) and then transferred to PVDF membranes (Millipore). The membranes were first incubated in TBS with 0.1% Tween-20 containing 5% nonfat dried milk for 1 h and then shaken overnight with GLUT-1 antibodies diluted 1:5,000 at 4 C. The next day, the membranes were rinsed three times with TBS with 0.01% Tween-20, and immunoreactivity was visualized on the PVDF membranes by the LSAB method, as described above. Samples of a crude membrane fraction of muscle prepared as decribed in (7) were separated on 10% SDSpolyacrylamide gels (Bio-Rad) and blotted to PVDF membranes. After blocking with TBS containing 5% BSA and 0.1% Tween-20, the membranes were incubated overnight at 4 C with GLUT-1 antibody (F18, 1 g/ml), GLUT-2 antibody (1:37,000), GLUT-3 antibody (1:10,000), GLUT-4 antibody (1:1,000), or GLUT-5 antibody (1.25 g/ml). Immunoreactivity was visualized by species-specific secondary antibodies (1:1,000) coupled to a chemiluminescence detection system (Pierce, Rockford, IL), in accordance with the directions given by the manufacturer. Methodological considerations. GLUT-1 immunoreactivity was studied in human skeletal muscle from young and elderly adults and at different developmental stages of gestation. Two methods for visualization of immunoreactivity were compared to evaluate the gain in sensitivity obtained by employing the TSA technique compared with the traditional LSAB. We tested the two methods on PVDF strips loaded with antigen (erythrocyte membrane) dots in decreasing concentrations. The sensitivity of GLUT-1 detection by TSA compared with LSAB was increased 5- to 25-fold when using the polyclonal antibody, from 1.6 pg/dot to pg/dot (Fig. 1A). Furthermore, we determined the detection efficiency of GLUT-1 in muscle cryosections and found that the maximal dilution of the polyclonal anti-glut-1 antibody was 1:102,400 for TSA and 1:3,200 for LSAB (ratio TSA-to-LSAB, 1:32), and it was 1:51,200 for the TSA and 1:3,200 for the LSAB technique when monoclonal antibodies (ratio TSA-to- LSAB, 1:16) were used. When comparing unfixed and formaldehyde-fixed cryosections of muscle stained by TSA, we observed a slight decrease in signal intensity and a similar detection efficiency. Western blot analysis of purified erythrocyte membranes demonstrated GLUT-1 immunoreactivity with a band of 55 kda for both antibodies (Fig. 1B). Specificity of the TSA immunoreactions was ensured by the fact that no immunostaining was found after preincubation of the Fig. 1. Polyvinylidine difluoride (PVDF) membrane immunocytochemical models. A: PVDF immunocytochemical models containing 2- l spots of purified erythrocyte membrane solutions (final protein content per spot in pg, is shown at the left). Strips were immunocytochemically stained as described in METHODS, with the polyclonal GLUT-1 antibody. GLUT-1 immunoreactivity was detected by the tyramide signal amplification (TSA) technique on strip 1 and the streptavidin biotin (LSAB) technique on strip 3. Strips 2 and 4 serve as controls for strips 1 and 2, respectively. In control strips, the primary antibodies were omitted. B: representative Western blot of the purified erythrocyte membranes with the polyclonal GLUT-1 antibody. Localization of molecular mass markers is indicated. anti-glut-1 antibodies with purified human erythrocyte membranes. RESULTS GLUT-1 immunoreactivity in adult muscle. In cryosections of adult human skeletal muscle from both young and elderly subjects, GLUT-1 immunoreactivity related to muscle fibers was not detectable with LSAB; however, GLUT-1 immunoreactivity was found in erythrocytes and perineurial sheaths. Applying TSA on cryosections of adult human skeletal muscles, we found GLUT-1 immunoreactivity in the perineurial sheaths (Fig. 2A), in erythrocytes, and as a scattered weak staining of the sarcolemma in apposition to labeled capillaries (Fig. 2B). No labeling of the sarcoplasma could be detected. The distribution pattern of GLUT-1 immunoreactivity was similar in skeletal muscle from younger and elderly subjects (data not shown). In par-

4 E532 Fig. 2. Immunostaining of human skeletal muscle. Cryosections of muscle biopsies were immunostained as described in METHODS. GLUT-1 immunoreactivity was visualized by the TSA technique and glycophorin A immunoreactivity by the LSAB method. A: GLUT-1 immunoreactivity is located in perineurial sheaths and capillaries. B: sporadic GLUT-1 immunoreactivity is present partially at the sarcolemma in a few muscle fibers in apposition to vessels. No staining was detected in the sarcoplasma. C: GLUT-1 immunoreactivity in paraffin-embedded section. No immunoreactivity is present in muscle fibers in apposition to stained capillaries. D: glycophorin A immunoreactivity. Arrow points to capillary with staining extending to muscle fibers. No staining is present in the perineurial sheaths. N, peripheral nerve. affin-embedded muscle sections, staining of the sarcolemma in apposition to capillaries could not be revealed (Fig. 2C). No differences were found in the staining patterns obtained by the two primary antibodies. To explore the specificity of this sarcolemmal GLUT-1 immunoreactivity, we studied the immunoreactivity to glycophorin A, an erythrocyte-specific antigen. We found reactivity to erythrocytes within the capillaries and a scattered weak staining of the sarcolemma in apposition to some of the capillaries similar to the GLUT-1 immunoreactivity pattern (Fig. 2D) and no staining of the perineurial sheaths (Fig. 2D). Control stainings were blank. GLUT-1 immunoreactivity during development. To explore changes in GLUT-1 expression related to development, sections of human skeletal muscle obtained from weeks of gestation and from children aged 1 and 4 yr were immunostained by the TSA method. At (Fig. 3A) and at (Fig. 3B) wk of gestation, there was a general GLUT-1 immunoreactivity of muscle fibers in addition to the immunoreactivity associated with the vessels (erythrocytes). Around birth (40 wk) a focal, intense GLUT-1 immunoreactivity was found in muscle fibers associated with vessels (Fig. 3C), whereas at 1 and 4 yr (data not shown), the staining pattern was as presented in adults. GLUT-2 immunoreactivity in fetal and adult muscle. Immunohistochemical staining of striated muscle from fetal and adult subjects did not reveal any expression of GLUT-2 in muscle fibers, except in a few sections where a weak GLUT-2 immunoreactivity was found (results not shown). Sections of multiorgan tissue blocks revealed background staining in most tissues, more prominently with TSA than with LSAB (results not shown), and thus the GLUT-2 immunoreactivity detected in a few muscle sections was considered to be nonspecific. GLUT-3 immunoreactivity in adult muscle. Immunohistochemical staining of striated muscle from younger and older subjects did not reveal any expression of GLUT-3 in muscle fibers (Fig. 4A), whereas capillaries were positive. We found GLUT-3 immunoreactivity in spermatides (Fig. 4B). GLUT-3 immunoreactivity during development. At 18 (Fig. 4C), 24 (Fig. 4D), and 40 (Fig. 4E) wk of gestation, general GLUT-3 immunoreactivity of all muscle fibers was found in addition to the immunoreactivity associated with capillaries, whereas at 1 and 4

5 E533 GLUT-4 immunoreactivity in adult muscle. GLUT-4 immunoreactivity in sections from striated muscle was confined to the muscle fibers. A dense perinuclear reaction was common. Moreover, a more scattered and distinct granular reaction was seen in association with the cell surface (Fig. 5A). GLUT-4 immunoreactivity during development. GLUT-4 protein expression was present throughout the gestation period, with a homogeneous staining in the muscle fibers [at 18 (Fig. 5B), 24 (Fig. 5C), and 40 (Fig. 5D) wk of gestation]. After birth, GLUT-4 immunoreactivity obtained a characteristic subcellular location, as seen in adult muscle fibers (see above). GLUT-5 immunoreactivity in fetal and adult muscle. Immunohistochemical staining of striated muscle from fetal and adult subjects did not reveal any expression of GLUT-5 in muscle fibers (Fig. 6, A and C), whereas capillaries were positive. We found spermatides to be GLUT-5 immunoreactive (Fig. 6B). Western blotting of adult muscle. To ensure specificity of the glucose transporter antibodies, Western blottings of crude membranes from adult human muscle were subjected to immunolabeling with antibodies of GLUT-1 through -5. Anti-GLUT-1 antibody labeled a broad band of kda, anti-glut-4 a band of kda, and anti-glut-5 a band of 48 kda (Fig. 7). Anti-GLUT-2 labeled numerous bands corresponding to the major protein bands of the blots and was thus considered not to be immunoreactivity specific, and no immunoreactivity for anti-glut-3 could be detected (results not shown). DISCUSSION Fig. 3. Immunostaining of fetal skeletal muscle GLUT-1 during gestation. Cryosections of muscle biopsies obtained at gestation weeks 18, 24, and 40 were immunostained for GLUT-1 by the TSA technique as described in METHODS. Muscle fibers at gestation weeks 18 (A) and 24 (B) are immunoreacting with GLUT-1. At gestation week 40 (C) a focal, intense immunoreactivity is seen in muscle fibers associated with vessels. yr, the staining pattern was as presented in adults. From the series of immunostaining at different concentrations of antibody, it appears that the reactivity was weakest at 18 wk of gestation. The present study demonstrates that fetal skeletal muscle fibers express GLUT-1, GLUT-3, and GLUT-4 proteins, but not GLUT-5 protein. In contrast to fetal muscle fibers, adult muscle fibers seem to express only GLUT-4. The presence of the GLUT-1 protein in human muscle cells is a matter for debate. Because of an unavoidable contamination of muscle biopsies with erythrocytes and perineurial sheaths, which express GLUT-1 at very high concentrations, the interpretation of studies on GLUT-1 protein in muscle homogenates is hampered. Few studies of GLUT-1 mrna in human muscle have been published (2, 11, 24, 35, 39), some of which report very low levels of GLUT-1 expression (2, 11). Furthermore, interpretation of the results is complicated by contamination with GLUT-1 mrna from other intramuscular sources, particularly perineurial sheaths. To circumvent the problems of contamination with nonmuscle GLUT-1 and putative low expression level of GLUT-1 in muscle cells, we studied the distribution pattern of GLUT-1 in human skeletal muscle with two immunocytochemical techniques, the LSAB and the powerful TSA. The latter has been reported to be times more sensitive than the LSAB technique (10, 22, 51). In our hands, the gain in sensitivity of GLUT-1 immunoreactivity detection with the TSA vs. LSAB methods was 5- to 25-fold when estimated on

6 E534 Fig. 4. GLUT-3 immunostaining of human skeletal muscle. Cryosections of muscle biopsies and testis tissue were immunostained as described in METHODS. GLUT-3 immunoreactivity was visualized by the TSA technique. A: adult muscle without staining of muscle fibers. B: testis with staining of spermiogenic cells. C: fetal muscle at gestation week 18. D: fetal muscle at gestation week 22. E: stained muscle at term (week 40). a model of antigens loaded onto strips (46), and the gain in detection efficiency was 16- to 32-fold, which is in accord with studies using different primary antibodies (10, 22, 51). GLUT-1 expression in adult human muscle was investigated in the present study with different variations of the applied immunocytochemical techniques, as described in METHODS. In accord with previous studies in rat muscles, we were able to visualize GLUT-1 immunoreactivity in erythrocytes and in perineurial sheaths (16, 25, 34). Only the sensitive TSA technique applied on cryosections revealed a sporadic capillaryassociated sarcolemmal immunoreaction. However, a similar distribution of glycophorin A immunoreactivity could be demonstrated with the LSAB technique. In paraffin-embedded muscle, the scattered perivascular GLUT-1 immunoreactivity was absent. Thus we suggest that the capillary-associated sarcolemmal staining is a result of contamination by fragments of erythrocytes, and we propose that the GLUT-1 glucose transporter is expressed in adult human skeletal muscle only in insignificant amounts. In accordance with the findings in rodents, it has been proposed that the function of GLUT-1 in human skeletal muscles is to mediate a basal glucose supply independent of insulin (4, 37). Because we were unable to detect GLUT-1 in the sarcolemma, even when using a very sensitive labeling protocol, and because no previous studies have directly demonstrated GLUT-1 in human skeletal muscle cells, it seems unlikely that GLUT-1 plays a major role in mature human muscle glucose uptake under nonpathological conditions. It could be speculated that GLUT-4 mediates basal glucose transport in muscle fibers through the constant exposure of skeletal muscle fibers to insulin and to tonal contraction, both of which are believed to induce GLUT-4 translocation to the sarcolemma (9, 19, 29, 31). In contrast to the results in adult muscle, our immunodetection system clearly demonstrates GLUT-1 expression in fetal skeletal muscle cell membranes. GLUT-1 expression is decreased with increasing duration of gestation, markedly reduced but still present at birth, and further reduced within the first year of life. Studies concerning developmental regulation of glucose transporters in human skeletal muscles are scarce. Western blotting analyses of crude membranes of myoblasts (45) and fused cultured human muscle cells show GLUT-1 and GLUT-4 in both (18, 45), and they express higher ratios of GLUT-1 to GLUT-4 than adult human muscle (45). GLUT-4 and GLUT-1 are under developmental control as the ratio of GLUT-4 to GLUT-1 increases after cell fusion in cultures of rat (36) and human (45) skeletal muscles. Furthermore,

7 E535 Fig. 5. GLUT-4 immunostaining of human skeletal muscle. Cryosections of muscle biopsy tissue were immunostained as described in METHODS. GLUT-4 immunoreactivity was visualized by the TSA technique. A: adult striated muscle. GLUT-4 immunoreactivity is confined to muscle fibers with dense perinuclear and scattered granular surface-associated reactions. The different stainability exhibited by the fibers reflects the known difference in type 1 and type 2 fiber GLUT-4 expression. B: fetal skeletal muscle at week 18 with a homogeneous staining. C: fetal muscle at week 22 with a homogeneous staining. D: muscle at term (week 40) with a combined diffuse and granular staining. there is convincing evidence for a reciprocal regulation of GLUT-1 and GLUT-4 in developmental rat muscles in vivo (43, 44). Postic et al. (43) found that GLUT-1 protein expression in fetal hindlimb muscle became undetectable after birth and that GLUT-4 protein became detectable in 15-day-old rats. Doria-Medina et al. (8) demonstrated a low level of GLUT-1 immunoreactivity in the sarcolemma of rat skeletal muscles (8). Our observations of GLUT-1 protein expression in fetal muscle fibers and its disappearance during development are in accord with the above findings and suggest that GLUT-1 mediates a constant high glucose supply during growth. GLUT-4 is expressed exclusively in muscle and fat. Insulin and contraction induce translocation of GLUT-4 from intracellular storage vesicles to the plasma membrane and then to the transverse tubules (19, 33, 34). The developmental expression pattern of GLUT-4 in humans has not previously been described. We found GLUT-4 protein expression in muscle fibers throughout gestation and in adult muscle fibers. The characteristic adult expression pattern was obtained only after birth. We found no evidence for a reciprocal regulation of GLUT-1 and GLUT-4 in human skeletal muscle. GLUT-3 was originally cloned from a fetal skeletal muscle library (27), suggesting a role for GLUT-3 in muscle development. GLUT-3 protein has been reported in cultures of human satellite cells (5). Guillet- Deniau et al. (13) found that GLUT-3 mrna and protein rose during cell fusion and disappeared shortly thereafter in primary myoblast cultures from rat fetuses, indicating that GLUT-3 may be important during myoblast fusion. We found GLUT-3 protein expression in muscle cells from 18 wk of gestation until birth. In accord with previous studies, GLUT-3 protein was below the detection limit in adult human muscles (14, 48). In both fetal and adult muscle, GLUT-3 expression was present in capillaries. In an immunoblotting study of muscles from human autopsies, Stuart et al. (49) suggested that the lack of detection of GLUT-3 protein in previous studies was caused by degradation of GLUT-3 protein; the cellular sources of GLUT-3 were not addressed. Based on our findings, GLUT-3 expression in muscles from adult human autopsies seems to originate from capillaries. The GLUT-3 expression pattern in human fetal muscle fibers resembles that in differentiating rat myoblasts, suggesting that GLUT-3 is important for muscle development (13). The significance of our observations of GLUT-3 protein expression

8 E536 in developing fetal muscle fibers needs to be further established. GLUT-5 protein was not detectable in either fetal or adult skeletal muscle fibers. We propose that the GLUT-5 detected by Western blotting of human muscles (20, 47, and the present study) seems to originate in intramuscular capillaries. In contrast, Hundal et al. Fig. 7. Western blot of crude membranes. Representative Western blot of crude membranes isolated from muscle biopsy from younger subject using the GLUT-1 antibody F18 (A), GLUT-4 antibody (B), and GLUT-5 antibody (C). Localization of molecular mass marker is indicated. Fig. 6. GLUT-5 immunostaining of human skeletal muscle. Cryosections of muscle biopsies and testis tissue were immunostained as described in METHODS. GLUT-5 immunoreactivity was visualized by the TSA technique. A: adult skeletal muscle with no immunoreactivity in muscle fibers and with stained vessels. B: testis with immunoreactive spermiogenic epithelium. C: fetal muscle at week 24 with stained vessel but unstained fibers. (20) described GLUT-5 immunoreactivity in the plasma membrane of human muscle fibers. GLUT-5 transports fructose rather than glucose (21). The functional role of GLUT-5 in muscles, even if expressed, is not readily understood, because the physiological concentration of fructose in blood is very low ( mm) (6). Thus we propose that GLUT-5 is expressed in adult human skeletal muscle only in insignificant amounts. We tested GLUT-2 expression in human muscle with a commercially available anti-glut-2 antibody, but the interpretation of the immunohistochemical staining was hampered by a slight variable background reactivity. In line with this, the anti-glut-2 antibody labeled numerous bands corresponding to the major protein bands in Western blots of muscle, and thus the specificity of this antibody could be questioned. Previous reports did not demonstrate GLUT-2 mrna expression in human muscle (2). In summary, we have found that in striated muscle, the expression of glucose transporter isoforms is dependent on developmental stage. Although GLUT-4 appears to be continuously present from fetal life through mature adult muscle, albeit with changes in intracellular localization, GLUT-1 and GLUT-3 are found in fetal life from at least 18 wk of gestation but disappear around birth. This dependence of expression patterns on development reflects differences in metabolic conditions in developing and mature muscles. Thus, at present, GLUT-4 appears to be the major transporter in functional contractile muscle, and GLUT-1 and GLUT-3 may be associated with metabolic activity coupled to formation, differentiation, and maturation of muscle fibers. GLUT-1 immunoreactivity is expressed only in very limited amounts in adult muscle fibers, indicating that this glucose transporter is of minor importance for muscle glucose supply. We hypothesize that GLUT-4 also mediates the basal glucose transport in muscle fibers through constant exposure to tonal contraction and basal insulin levels. Irene Lynfort, Kirsten Dahl, and Ole Nielsen provided excellent technical assistance. P. Poulsen kindly provided adult muscle biopsies, and J. Vinten generously donated the F18 antibody. This study was supported by the Clinical Institute, Department of Health Science, Odense University.

9 E537 REFERENCES 1. Adams JC. Biotin amplification of biotin and horseradish peroxidase signals in histochemical stains. J Histochem Cytochem 40: , Bell GI, Kayano T, Buse JB, Burant CF, Takeda J, Lin D, Fukumoto H, and Seino S. Molecular biology of mammalian glucose transporters. Diabetes Care 13: , Bergstrom J. Percutaneous needle biopsy of skeletal muscle in physiological and clinical research. Scand J Clin Lab Invest 35: , Burant CF, Sivitz WI, Fukumoto H, Kayano T, Nagamatsu S, Seino S, Pessin JE, and Bell GI. Mammalian glucose transporters: structure and molecular regulation. Recent Prog Horm Res 47: , Ciaraldi TP, Carter L, Nikoulina S, Mudaliar S, McClain DA, and Henry RR. Glucosamine regulation of glucose metabolism in cultured human skeletal muscle cells: divergent effects on glucose transport/phosphorylation and glycogen synthase in nondiabetic and type 2 diabetic subjects. Endocrinology 140: , Darakhshan F, Hajduch E, Kristiansen S, Richter EA, and Hundal HS. Biochemical and functional characterization of the GLUT5 fructose transporter in rat skeletal muscle. Biochem J 336: , Dela F, Handberg A, Mikines KJ, Vinten J, and Galbo H. GLUT 4 and insulin receptor binding and kinase activity in trained human muscle. J Physiol (Lond) 469: , Doria-Medina CL, Lund DD, Pasley A, Sandra A, and Sivitz WI. Immunolocalization of GLUT-1 glucose transporter in rat skeletal muscle and in normal and hypoxic cardiac tissue. Am J Physiol Endocrinol Metab 265: E454 E464, Douen AG, Ramlal T, Rastogi S, Bilan PJ, Cartee GD, Vranic M, Holloszy JO, and Klip A. Exercise induces recruitment of the insulin-responsive glucose transporter. Evidence for distinct intracellular insulin- and exercise-recruitable transporter pools in skeletal muscle. J Biol Chem 265: , Erber WN, Willis JI, and Hoffman GJ. An enhanced immunocytochemical method for staining bone marrow trephine sections. J Clin Pathol 50: , Flier JS, Mueckler M, McCall AL, and Lodish HF. Distribution of glucose transporter messenger RNA transcripts in tissues of rat and man. J Clin Invest 79: , Froehner SC, Davies A, Baldwin SA, and Lienhard GE. The blood-nerve barrier is rich in glucose transporter. J Neurocytol 17: , Guillet-Deniau I, Leturque A, and Girard J. Expression and cellular localization of glucose transporters (GLUT1, GLUT3, GLUT4) during differentiation of myogenic cells isolated from rat foetuses. J Cell Sci 107: , Haber RS, Weinstein SP, O Boyle E, and Morgello S. Tissue distribution of the human GLUT3 glucose transporter. Endocrinology 132: , Handberg A, Kayser L, Hoyer PE, Micheelsen J, and Vinten J. Elevated GLUT 1 level in crude muscle membranes from diabetic Zucker rats despite a normal GLUT 1 level in perineurial sheaths. Diabetologia 37: , Handberg A, Kayser L, Hoyer PE, and Vinten J. A substantial part of GLUT-1 in crude membranes from muscle originates from perineurial sheaths. Am J Physiol Endocrinol Metab 262: E721 E727, Handberg A, Megeney LA, McCullagh KJ, Kayser L, Han XX, and Bonen A. Reciprocal GLUT-1 and GLUT-4 expression and glucose transport in denervated muscles. Am J Physiol Endocrinol Metab 271: E50 E57, Henry RR, Abrams L, Nikoulina S, and Ciaraldi TP. Insulin action and glucose metabolism in nondiabetic control and NIDDM subjects. Comparison using human skeletal muscle cell cultures. Diabetes 44: , Hirshman MF, Goodyear LJ, Wardzala LJ, Horton ED, and Horton ES. Identification of an intracellular pool of glucose transporters from basal and insulin-stimulated rat skeletal muscle. J Biol Chem 265: , Hundal HS, Ahmed A, Guma A, Mitsumoto Y, Marette A, Rennie MJ, and Klip A. Biochemical and immunocytochemical localization of the GLUT5 glucose transporter in human skeletal muscle. Biochem J 286: , Hundal HS, Darakhshan F, Kristiansen S, Blakemore SJ, and Richter EA. GLUT5 expression and fructose transport in human skeletal muscle. Adv Exp Med Biol 441: Hunyady B, Krempels K, Harta G, and Mezey E. Immunohistochemical signal amplification by catalyzed reporter deposition and its application in double immunostaining. J Histochem Cytochem 44: , James DE, Strube M, and Mueckler M. Molecular cloning and characterization of an insulin-regulatable glucose transporter. Nature 338: 83 87, Kahn BB, Rosen AS, Bak JF, Andersen PH, Damsbo P, Lund S, and Pedersen O. Expression of GLUT1 and GLUT4 glucose transporters in skeletal muscle of humans with insulindependent diabetes mellitus: regulatory effects of metabolic factors. J Clin Endocrinol Metab 74: , Kahn BB, Rossetti L, Lodish HF, and Charron MJ. Decreased in vivo glucose uptake but normal expression of GLUT1 and GLUT4 in skeletal muscle of diabetic rats. J Clin Invest 87: , Kayano T, Burant CF, Fukumoto H, Gould GW, Fan YS, Eddy RL, Byers MG, Shows TB, Seino S, and Bell GI. Human facilitative glucose transporters. Isolation, functional characterization, and gene localization of cdnas encoding an isoform (GLUT5) expressed in small intestine, kidney, muscle, and adipose tissue and an unusual glucose transporter pseudogene-like sequence (GLUT6). J Biol Chem 265: , Kayano T, Fukumoto H, Eddy RL, Fan YS, Byers MG, Shows TB, and Bell GI. Evidence for a family of human glucose transporter-like proteins. Sequence and gene localization of a protein expressed in fetal skeletal muscle and other tissues. J Biol Chem 263: , Kennedy JW, Hirshman MF, Gervino EV, Ocel JV, Forse RA, Hoenig SJ, Aronson D, Goodyear LJ, and Horton ES. Acute exercise induces GLUT4 translocation in skeletal muscle of normal human subjects and subjects with type 2 diabetes. Diabetes 48: , Klip A, Ramlal T, Bilan PJ, Cartee GD, Gulve EA, and Holloszy JO. Recruitment of GLUT-4 glucose transporters by insulin in diabetic rat skeletal muscle. Biochem Biophys Res Commun 172: , Loken MR, Shah VO, Dattilio KL, and Civin CI. Flow cytometric analysis of human bone marrow. I Normal erythroid development. Blood 69: , Lund S, Holman GD, Schmitz O, and Pedersen O. Contraction stimulates translocation of glucose transporter GLUT4 in skeletal muscle through a mechanism distinct from that of insulin. Proc Natl Acad Sci USA 92: , Lund S, Holman GD, Zierath JR, Rincon J, Nolte LA, Clark AE, Schmitz O, Pedersen O, and Wallberg HH. Effect of insulin on GLUT4 cell surface content and turnover rate in human skeletal muscle as measured by the exofacial bis-mannose photolabeling technique. Diabetes 46: , Marette A, Burdett E, Douen A, Vranic M, and Klip A. Insulin induces the translocation of GLUT4 from a unique intracellular organelle to transverse tubules in rat skeletal muscle. Diabetes 41: , Marette A, Richardson JM, Ramlal T, Balon TW, Vranic M, Pessin JE, and Klip A. Abundance, localization, and insulininduced translocation of glucose transporters in red and white muscle. Am J Physiol Cell Physiol 263: C443 C452, Miele C, Formisano P, Condorelli G, Caruso M, Oriente F, Andreozzi F, Tocchetti CG, Riccardi G, and Beguinot F. Abnormal glucose transport and GLUT1 cell-surface content in fibroblasts and skeletal muscle from NIDDM and obese subjects. Diabetologia 40: , Mitsumoto Y, Burdett E, Grant A, and Klip A. Differential expression of the GLUT1 and GLUT4 glucose transporters during differentiation of L6 muscle cells. Biochem Biophys Res Commun 175: , 1991.

10 E Mueckler M. Family of glucose-transporter genes. Implications for glucose homeostasis and diabetes. Diabetes 39: 6 11, Mueckler M. Facilitative glucose transporters. Eur J Biochem 219: , Pedersen O, Bak JF, Andersen PH, Lund S, Moller DE, Flier JS, and Kahn BB. Evidence against altered expression of GLUT1 or GLUT4 in skeletal muscle of patients with obesity or NIDDM. Diabetes 39: , Pessin JE and Bell GI. Mammalian facilitative glucose transporter family: structure and molecular regulation. Annu Rev Physiol 54: , Pfeiffer P, Grabau DA, Nielsen O, and Clausen PP. Immunohistochemical bulk staining of slides using a rack peroxidaselabelled streptavidin-biotin technique. Appl Immunohistochem 4: , Ploug T, Stallknecht BM, Pedersen O, Kahn BB, Ohkuwa T, Vinten J, and Galbo H. Effect of endurance training on glucose transport capacity and glucose transporter expression in rat skeletal muscle. Am J Physiol Endocrinol Metab 259: E778 E786, Postic C, Leturque A, Printz RL, Maulard P, Loizeau M, Granner DK, and Girard J. Development and regulation of glucose transporter and hexokinase expression in rat. Am J Physiol Endocrinol Metab 266: E548 E559, Santalucia T, Camps M, Castello A, Munoz P, Nuel A, Testar X, Palacin M, and Zorzano A. Developmental regulation of GLUT-1 (erythroid/hep G2) and GLUT-4 (muscle/fat) glucose transporter expression in rat heart, skeletal muscle, and brown adipose tissue. Endocrinology 130: , Sarabia V, Lam L, Burdett E, Leiter LA, and Klip A. Glucose transport in human skeletal muscle cells in culture. Stimulation by insulin and metformin. J Clin Invest 90: , Scopsi L and Larsson LI. Increased sensitivity in peroxidase immunocytochemistry. A comparative study of a number of peroxidase visualization methods employing a model system. Histochemistry 84: , Shepherd PR, Gibbs EM, Wesslau C, Gould GW, and Kahn BB. Human small intestine facilitative fructose/glucose transporter (GLUT5) is also present in insulin-responsive tissues and brain. Investigation of biochemical characteristics and translocation. Diabetes 41: , Shepherd PR, Gould GW, Colville CA, McCoid SC, Gibbs EM, and Kahn BB. Distribution of GLUT3 glucose transporter protein in human tissues. Biochem Biophys Res Commun 188: , Stuart CA, Wen G, and Jiang J. GLUT3 protein and mrna in autopsy muscle specimens. Metabolism 48: , Vogt B, Muhlbacher C, Carrascosa J, Obermaier KB, Seffer E, Mushack J, Pongratz D, and Haring HU. Subcellular distribution of GLUT 4 in the skeletal muscle of lean type 2 (non-insulin-dependent) diabetic patients in the basal state. Diabetologia 35: , von WR, Mengel M, Gignac S, Wilkens L, Werner M, and Georgii A. Tyramine amplification technique in routine immunohistochemistry. J Histochem Cytochem 45: , 1997.