The HIV Protease Inhibitor Indinavir Decreases Insulin- and Contraction-Stimulated Glucose Transport in Skeletal Muscle

Indinavir was a gift from Merck Pharmaceuticals (Nutley, NJ). Pork insulin was purchased from Eli Lilly (Indianapolis, IN). 3-O-[³H]methyl-d-glucose (3MG) was obtained from American Radiolabeled Chemicals (St. Louis, MO), and [U-¹⁴C]mannitol and [γ-³²P]ATP were from Du Pont-NEN (Boston, MA). The anti-phosphotyrosine monoclonal antibody was obtained from Sigma Chemical (St. Louis, MO). Phosphatidylinositol was from Avanti Polar Lipids (Alabaster, AL). The anti-phospho-protein kinase B (Ser473) antibody was obtained from New England BioLabs (Beverly, MA). Horseradish peroxidase (HRP)-conjugated donkey anti-rabbit IgG was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Enhanced chemiluminescence (ECL) reagents were obtained from Amersham (Arlington Heights, IL). The bicinchoninic acid (BCA) protein assay kit was purchased from Pierce Chemical (Rockford, IL). 2-N-4-(1-azi-2,2,2-trifluoroethyl)-benzoyl-1,3-bis-[³H] (d-mannose-4-yloxy)-2-propylamine was purchased from Toronto Research Chemicals (North York, Ontario, Canada). The GLUT4 antiserum (G4 829) used for immunoprecipitation was kindly provided by Dr. Mike Mueckler (Washington University, St. Louis, MO). All other chemicals were obtained from Sigma Chemical.

Male Wistar rats (100–130 g) were obtained from Charles Rivers Laboratories (Wilmington, MA). Rats were maintained on a 12-h light/12-h dark schedule and given rat diet and water ad libitum. After overnight fasting, animals were anesthetized with an intraperitoneal injection of sodium pentobarbital (5 mg/100 g body wt), and the epitrochlearis and soleus muscles were removed. The epitrochlearis, consisting primarily of type IIb fibers, is a small thin muscle of the forearm that is suitable for measurement of glucose transport activity in vitro (16,17). The soleus muscle, which consists mainly of type I fibers, was split before 3MG transport measurements (18). The Washington University Animal Studies Committee approved all experimental procedures.

Epitrochlearis or split soleus muscles were incubated in a shaking incubator at 35°C for 4 h in 2 ml Krebs-Henseleit bicarbonate (KHB) buffer containing 0.1% radioimmunoassay grade bovine serum albumin (BSA), 8 mmol/l glucose, and 32 mmol/l mannitol, in the absence or presence of 12 nmol/l insulin, with or without indinavir at concentrations ranging from 1 to 40 μmol/l. All vials were incubated with shaking and a continuous gas phase of 95% O₂:5% CO₂.

Immediately after 4-h incubation in the absence or presence of 20 μmol/l indinavir, epitrochlearis muscles were stimulated to contract in vitro as described previously (19). A total of 10 tetanic contractions were elicited by stimulating at 100 Hz for 10 s at a rate of one contraction per minute for 10 min in the absence or presence of 20 μmol/l indinavir.

Glucose transport was measured using the nonmetabolizable glucose analog 3MG. Muscles were transferred to 2 ml KHB buffer with 0.1% BSA, 40 mmol/l mannitol, and the same concentrations of indinavir and insulin as in the previous step and incubated with shaking for 10 min at 30°C to wash glucose from the extracellular space before measurement of glucose transport. 3MG transport was measured during 10 min in 1 ml KHB buffer containing 0.1% BSA, 8 mmol/l 3MG (2.0 μCi/ml), 32 mmol/l [U-¹⁴C]mannitol (0.3 μCi/ml), and the same concentrations of insulin and indinavir present in the previous incubations. Immediately after incubation, muscles were blotted, trimmed of excess tissue, and clamp-frozen with aluminum tongs cooled in liquid N₂. Extracellular space and intracellular 3MG (μmol · ml⁻¹ · 10 min⁻¹) were determined as described previously (17,20).

Muscles were preincubated for 4 h in 2 ml of KHB containing 0.1% BSA, 8 mmol/l glucose, and 32 mmol/l mannitol, in the absence or presence of 12 nmol/l insulin, with or without 20 μmol/l indinavir. To wash glucose from the extracellular space before incubation with 2-N-4-(1-azi-2,2,2-trifluoroethyl)benzoyl-1,3-bis-[2-³H] (d-mannose-4-yloxy)-2-propylamine (ATB-[2-³H]BMPA), muscles were transferred to 2 ml KHB buffer with 0.1% BSA, 40 mmol/l mannitol, and the same concentrations of indinavir and insulin as in the previous step, and incubated with shaking for 10 min at 30°C. The ATB-[2-³H]BMPA exofacial photolabeling technique was used to measure GLUT4 at the cell surface as described previously (21), except that the labeled GLUT-4 was immunoprecipitated with a rabbit polyclonal antibody.

Epitrochlearis muscles were incubated for 4 h in the absence or presence of indinavir (20 μmol/l) with or without insulin (12 nmol/l) and then clamp-frozen. Muscles were stored at −80°C until processing. Muscle extracts were prepared using a modification of the method of Saad et al. (22). Muscle samples were homogenized in ice-cold buffer containing 50 mmol/l HEPES (pH 7.4), 150 mmol/l NaCl, 10% glycerol, 1% Triton X-100, 1.5 mmol/l MgCl₂, 1.0 mmol/l EDTA, 10 mmol/l Na₄P₂O₇, 100 mmol/l NaF, 2.0 mmol/l Na₃VO₄, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 0.5 μg/ml pepstatin, and 2 mmol/l phenylmethylsulfonyl fluoride. Homogenates were incubated with end-over-end rotation at 4°C for 1 h, then centrifuged at 200,000g for 50 min at 4°C. Protein concentration of each sample supernatant was determined using a BCA protein assay kit. Immunoprecipitation of phosphotyrosine and measurement of phosphatidylinositol (PI) 3-kinase activity were performed as described by Backer et al. (23).

For determination of phosphorylated PKB, aliquots of muscle extract were treated with 2× Laemmli buffer (24) containing 100 mmol/l dithiothreitol and heated in a boiling water bath for 5 min. A total of 75 μg of protein from each sample was subjected to SDS-PAGE (10% resolving gel), followed by transfer to nitrocellulose membranes at 90 V for 90 min using the method of Towbin et al. (25). The membrane was blocked overnight at 4°C in Tris-buffered solution (20 mmol/l Tris, pH 7.6, 13.7 mmol/l NaCl, 0.1% Tween 20) with 5% non-fat dried milk and used for detection of phosphorylated PKB using a polyclonal anti-phospho-PKB-antibody specific for phosphorylation at the Ser473 site, followed by HRP-conjugated donkey anti-rabbit-IgG. Band intensity was analyzed using densitometry.

Data are presented as mean ± SE of the mean. Group comparisons of 3MG transport were analyzed using analysis of variance (ANOVA). A Student-Newman-Keuls post hoc test was used to identify differences when indicated by ANOVA. Group comparisons of PI 3-kinase activity and cell surface GLUT4 were analyzed using a Kruskal-Wallis one-way ANOVA. Differences in insulin-stimulated phosphorylated PKB in the absence or presence of 20 μmol/l indinavir were analyzed using a Student’s t test. P < 0.05 was the accepted level for statistical significance.

As shown in Fig. 1, incubation of epitrochlearis muscles for 4 h with a maximally effective concentration of insulin (12 nmol/l) resulted in a sixfold increase in glucose transport above basal. When indinavir was included in the incubation medium, insulin-stimulated glucose transport was inhibited in a dose-dependent manner. The increase in glucose transport induced by insulin was reduced by 40% in the presence of 5 μmol/l indinavir, 58% with 20 μmol/l indinavir, and 72% in the presence of 40 μmol/l indinavir (Fig. 1). Indinavir (40 μmol/l) had no effect on basal glucose transport (0.12 ± 0.05 and 0.10 ± 0.04 μmol · ml⁻¹ · 10 min⁻¹ for basal and basal + indinavir, respectively). The inhibitory effect of indinavir on the stimulation of glucose transport by insulin is not muscle fiber type–specific, as a similar ∼65% reduction was observed in soleus muscles incubated in the presence of 20 μmol/l indinavir (Fig. 2).

The exofacial label ATB-[2-³H]BMPA was used to quantify GLUT4 transporters at the cell surface after insulin stimulation in the absence or presence of 20 μmol/l indinavir. Insulin induced an approximate sevenfold increase in GLUT4 labeling (Fig. 3). When 20 μmol/l indinavir was included in the incubation medium, the increase in cell surface GLUT4 labeling in response to insulin was reduced by ∼70% (Fig. 3).

3MG transport was increased approximately sevenfold in epitrochlearis muscle in response to stimulation of contractile activity (Fig. 4). The contraction-stimulated increase in glucose transport was reduced by ∼70% when muscles were incubated with 20 μmol/l indinavir for 4 h before stimulation (Fig. 4).

Activation of PI 3-kinase is an essential step in the signal-transduction pathway by which insulin stimulates glucose transport (26,27,28 29). Therefore, we determined whether indinavir inhibits PI 3-kinase activity. As shown in Fig. 5, incubation of muscles with 12 nmol/l insulin for 4 h resulted in an approximate fourfold increase in PI 3-kinase activity, and 20 μmol/l indinavir did not inhibit the stimulation of PI 3-kinase activity by insulin (Fig. 5).

The next step in the insulin-signaling pathway is mediated by PKB (30,31). PKB is activated by dual phosphorylation of threonine 308 and serine 473. Phosphorylation of the threonine 308 residue on PKB, by 3-phosphoinositide-dependent protein kinase (PDK)-1, with subsequent phosphorylation of serine 473, results in activation of the kinase (32,33). As shown in Fig. 6, insulin-stimulated phosphorylation of PKB, as measured using an antibody specific for the phosphorylated serine 473 residue of PKB, was not decreased in muscles incubated with 20 μmol/l indinavir (Fig. 6).

Our results show that the HIV protease inhibitor indinavir induces severe insulin resistance of muscle glucose transport within 4 h in rat skeletal muscle incubated in vitro. Skeletal muscle is responsible for ∼90% of insulin-stimulated glucose disposal (14,15). The present finding is, therefore, compatible with the hypothesis that the first event in the process by which the HIV protease inhibitor indinavir causes impaired glucose tolerance and type 2 diabetes is the development of insulin resistance of skeletal muscle. In this context, it also seems possible that the increase in central abdominal fat and the serum lipid and lipoprotein abnormalities that develop in patients with HIV treated with HIV protease inhibitors are mediated, at least in part, by hyperinsulinemia that develops in response to insulin resistance in muscles.

One potential mechanism by which indinavir could be decreasing insulin-stimulated glucose transport is inhibition of one or more steps in the signaling pathway by which insulin stimulates glucose transport. The first step in the insulin-signaling cascade is the binding of insulin to the insulin-receptor, which results in activation of the tyrosine kinase of the intracellular β-subunits of the receptor (34,35,36). The activated insulin receptor phosphorylates the insulin receptor substrates, IRS-1 and IRS-2, on multiple tyrosines. The phosphorylated insulin-receptor substrates bind the Src homology-2 (SH2) domain of PI 3-kinase, resulting in activation of the kinase. Involvement of PI 3-kinase activation in the stimulation of muscle glucose transport by insulin is evidenced by the finding that inhibition of PI 3-kinase with wortmannin blocks stimulation of glucose transport by insulin (27,28 29). In the present study, we found that activation of PI 3-kinase by insulin is not inhibited by a concentration of indinavir that markedly reduces insulin-stimulated glucose transport. This finding also provides evidence that the earlier, tyrosine phosphorylation insulin-signaling steps are not inhibited in the presence of indinavir.

The next step in the insulin-signaling pathway involves threonine/serine phosphorylation of PKB, resulting in its activation (37). PI-3,4-biphosphate and PI-3,4,5-triphosphate, produced by the action of PI 3-kinase, result in phosphorylation of threonine 308 of PKB by PDK-1, with subsequent phosphorylation of serine 473 and activation of the kinase (32,33,37). Our finding that insulin-stimulated phosphorylation of PKB is not inhibited by indinavir suggested that indinavir acts at a step beyond the insulin-signaling pathway. This possibility was supported by the finding that contraction-stimulated glucose transport was also inhibited in the presence of indinavir. Therefore, it seemed likely that indinavir is acting on a step common to both pathways, i.e., GLUT4 translocation to the cell surface.

Our finding that cell-surface labeling of GLUT4 in maximally insulin-stimulated muscle was decreased ∼70% in muscles treated with indinavir is compatible with the hypothesis that indinavir inhibits GLUT4 translocation. However, while this paper was in revision, Murata et al. (38) reported that indinavir inhibits glucose transport in adipocytes without affecting GLUT4 translocation to the plasma membrane. They used both a cell-fractionation procedure and the plasma membrane sheet method to show that GLUT4 translocation was not inhibited. On the basis of these results and their finding that indinavir reduced the glucose transport activity of GLUT4 expressed in oocytes, Murata et al. (38) concluded that indinavir acts by inhibiting GLUT4 intrinsic activity. In this context, our finding that cell-surface GLUT4 labeling by ATB-[2-³H]BMPA is markedly reduced by indinavir in insulin-stimulated muscle could be explained by a direct effect of indinavir on GLUT4 rather than on GLUT4 translocation. Such an effect could be mediated either by 1) binding of indinavir to the glucose- and ATB-[2-³H]BMPA-binding site on GLUT4, thus interfering with the ability of GLUT4 to interact with glucose and ATB-[2-³H]BMPA, or 2) by a conformational change in GLUT4 induced by indinavir that interferes with its ability to transport glucose and bind ATB-[2-³H]BMPA.

In conclusion, the results of this initial study show that exposure of skeletal muscle to the HIV protease inhibitor indinavir causes a large decrease in insulin responsiveness of glucose transport within 4 h. This finding raises the possibility that the first event in the development of the insulin-resistance syndrome that occurs in patients with HIV treated with the HIV protease inhibitor indinavir is insulin resistance of skeletal muscle glucose transport.

This research was supported by National Institutes of Health Grant DK-18986. L.A.N. was supported by National Institute on Aging Institutional National Research Service Award AG-00078. K.K. was supported by a mentor-based fellowship from the American Diabetes Association. K.E.Y. was supported by NIDDK 49393 and NIDDK 54163.

We thank William G. Powderly, MD, and Pablo Tebas, MD, for helpful conversations, Jung Hoon Kim and Cicely Miederhoff for excellent technical assistance, and Victoria Reckamp for expert assistance with preparation of the manuscript.