Diabetogenic effect of pravastatin is associated with insulin resistance and myotoxicity in hypercholesterolemic mice

Low-density lipoprotein receptor knockout (LDLr^−/−) female mice with a C57BL/6J background, originally from the Jackson Laboratory (Bar Harbor, ME), were obtained from the breeding colony of State University of Campinas Multidisciplinary Center for Biological Research in Laboratory Animals (CEMIB/UNICAMP, Campinas, Brazil). Animal experiments were approved by the University’s Committee for Ethics in Animal Experimentation (CEUA/UNICAMP, protocol # 3819-1), and all experiments were performed in accordance with the national Brazilian guideline number 13 for “Control in Animal Experiments”, published on September 13th, 2013 (code 00012013092600005, available at 〈http://​portal.​in.​gov.​br/​verificacao-autenticidade〉). The mice were kept under standard laboratory conditions (at 20–22 °C and a 12 h/12 h light/dark cycle) in the local (conventional) animal facility in individually ventilated cages (3–5 mice/cage), with free access to filtered tap water and regular rodent AIN93-M diet (standard laboratory rodent chow diet, Nuvital CR1, Colombo, PR, Brazil). Female mice (4 weeks old) were randomly assigned to six groups (3 treated and 3 control groups) according to the time of treatment with pravastatin (dissolved in the drinking water, 400 mg/L) for three (n = 6), six (n = 6) and ten (n = 5) months. Controls received filtered tap water without pravastatin. The pravastatin sodium (Medley, Sanofi, SP, Brazil) dose of 40 mg/kg body weight per day was based on drink consumption rates (3 mL/day). One out of 5 mice that belonged to the 10-month pravastatin treatment group died in the 7th month of treatment. Based on the current literature, the 40 mg/kg dose is considered a moderate effective dose of pravastatin for mice [31, 32], among protocols employing doses from 10 to 300 mg/kg body weight [33]. We previously characterized defective insulin secretion and beta cell death in this model treated with 40 mg/kg of pravastatin for 2 months [8, 9]. We chose to study female LDLr^−/− mice for several reasons, including increased susceptibility to atherosclerosis, increased plasma cholesterol levels and increased response to statin treatment in females compared with males. In addition, recent reports on the diabetogenic effects of statins demonstrate an increased risk for women compared with men. In terminal experiments, mice were anaesthetized via intraperitoneal injection of ketamine and xylazine (50 and 10 mg/kg) and euthanized by exsanguination through the retro-orbital plexus. All animal experiments were performed between 8:00 and 11:00 p.m.

Blood glucose was measured using a glucose analyser (Accu-Chek Advantage; Roche Diagnostics, Basel, Switzerland). Plasma cholesterol and triglycerides were measured using standard commercial kits (Roche Diagnostics GMbH, Mannheim, Germany) according to the manufacturer’s instructions. Plasma insulin was measured using an Ultra-Sensitive Mouse Insulin ELISA Kit (Crystal Chemical, Illinois, USA). Fasting blood samples were obtained at 9:00 a.m. after a 12 h fasting period.

Conscious and fasted mice (food removed at 8:00 p.m. and blood obtained between 8 and 9:00 a.m.) received an oral dose of glucose solution (1.5 g/kg body weight) by gavage. Blood samples were collected directly from the tail tip cut by the glucose analyser before (t = 0 min) and 15, 30, 60, and 90 min after the glucose load. One week after the GTT, the same mice were submitted to the ITT. Blood samples were collected from fasted mice that had been refed for 3 h (t = 0 min) and 5, 10, 15, 30 and 60 min after an intraperitoneal insulin injection (0.75 U/kg body weight, regular human insulin, Eli Lilly Co.) for glucose analysis.

Pancreatic islets were isolated from overnight fasted and anaesthetized mice by collagenase type V (0.8 mg/mL; Sigma) digestion and were then selected under a microscope. Four replicates of 4 islets/well in each condition (basal and glucose stimulated) from each mouse were used for the insulin secretion assay. Islets were pre-incubated for 30 min at 37 °C in Krebs-bicarbonate buffer (KBB) composed of the following: 115 mmol/L NaCl, 5 mmol/L KCl, 2.56 mmol/L CaCl₂, 1 mmol/L MgCl₂, 10 mmol/L NaHCO₃, and 15 mmol/L HEPES, supplemented with 5.6 mmol/L glucose and 0.3% BSA and equilibrated with a mixture of 95% O₂/5% CO₂ at pH 7.4. The islets were further incubated for 1 h in KBB containing glucose (2.8 or 11.1 mmol/L). At the end of the incubation period, insulin content was measured by radioimmunoassay [8, 9].

Protein synthesis rates were measured in gastrocnemius muscle excised from anaesthetized and exsanguinated mice using phenylalanine incorporation into proteins, and protein degradation rates were measured by tyrosine released into incubation medium, as previously described [34]. Briefly, right leg gastrocnemius muscle was used for protein synthesis, and left leg gastrocnemius was used for the protein degradation assay. For synthesis, muscle was pre-incubated for 30 min at 37 °C in Krebs–Henseleit bicarbonate (KHB) buffer (in mM: 110 NaCl, 25 NaHCO₃, 3.4 KCl, 1 CaCl₂, 1 MgSO₄, 1 KH₂PO₄, 5.5 glucose and 0.01% albumin, pH 7.4) with continuous aeration (95% O₂ and 5% CO₂) and shaking. Next, KHB buffer was changed, supplemented with 5 µCi L[³H]-phenylalanine and incubated for 2 h. The muscles were then dried in filter paper, weighed, homogenized in trichloroacetic acid (TCA, 1:3 w/v), and centrifuged at 10,000 rpm for 15 min at 4 °C. The pellet was suspended in 1 M NaOH and incubated at 40 °C for 30 min. The supernatant was used to measure the total protein concentration and to quantity radioactivity in liquid scintillation using β-counter equipment (Beckman LS 6000 TA, Fullerton, CA, USA). Left gastrocnemius muscle was used for protein degradation analysis. Muscle was pre-incubated for 30 min at 37 °C in RPMI medium without phenol red under 95% O₂–5% CO₂ and shaking. Subsequently, muscles were incubated with KHB buffer containing cycloheximide, a protein synthesis inhibitor (130 µg/mL), for 2 h at 37 °C. Then, medium aliquots were treated with 30% TCA and centrifuged at 4000 rpm for 10 min to assess the tyrosine content using a fluorometric assay utilizing 20% 1-nitroso-2-naphthol reagent and nitric acid [34].

Chymotrypsin-like activity was determined in aliquots of homogenized muscle tissue by incubation with the fluorogenic substrate Suc LLVY-AMC (0.167 µg/mL in Tris–HCl, pH 7.4; excitation 360 nm, emission 460 nm). Cathepsin B activity was assayed using the fluorogenic substrate Z-Phe-Arg-NMec (0.02 mM with 0.1% Brij; excitation: 340–380 nm, emission: 460 nm) and cathepsin H, using Arg-NMec as substrate. Enzyme activities were expressed as units of fluorescence/µg protein/min [34].

A murine myogenic cell line, C2C12, was obtained from the ATCC (ATCC.^®. No. CRL-1772™; passage # 16–19). The cells were maintained in growth medium consisting of high glucose (4.5 g/L) Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum (FBS) and 1% penicillin–streptomycin in a humidified cell culture incubator for 2 days or until 100% confluence. Afterwards, cell differentiation into myotubes was induced using high glucose DMEM supplemented with 2% horse serum for 5 days. After that, the cells were incubated with 1, 10, or 50 µM pravastatin sodium salt (Sigma) for 12 h. Then, the expression of proteins of interest was determined by Western blot analysis, and H₂O₂ and superoxide production rates were determined as described below. Since pravastatin is one of the least toxic statins due to its hydrophilic nature [21, 29, 35, 36], it was expected that we would need high doses of pravastatin to induce and observe both therapeutic and toxic effects. Sun et al. [36] showed that 1 and 10 µM pravastatin does not affect glucose metabolism in C2C12 cells. Thus, we used a wide dose range (1–50 µM) in the in vitro C2C12 studies.

Cell homogenates were treated with Laemmli loading buffer containing dithiothreitol. After heating to 95 °C for 5 min, the proteins were separated by electrophoresis (30 µg protein/lane, 8 or 12% acrylamide/bisacrylamide gel) and were then transferred to nitrocellulose membranes. The nitrocellulose membranes were treated for 1.5 h with a blocking buffer (5% BSA, 10 mmol/L Tris, 150 mmol/L NaCl, and 0.02% Tween 20) and were incubated with the following primary antibodies: Caspase-3 (Millipore), Bax (Cell Signaling), Bcl-2 (Cell Signaling), phospho-ser⁴⁷³-AKT (Cell Signaling) and total AKT (Santa Cruz Biotechnology). GAPDH detected with rabbit polyclonal antibody and tubulin detected with mouse polyclonal antibody (Santa Cruz Biotechnology) were used as internal controls. Membranes were then incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000, Invitrogen). Detection was performed using enhanced chemiluminescence (SuperSignal West Pico, Pierce, Rockford, IL). Band intensities were quantified by optical densitometry (Scion Image, Frederick, MD).

Myotube C2C12 cells (10⁴) were plated and differentiated in 96-well plates and treated with pravastatin 1, 10, or 50 µM in DMEM for 12 h. The cells were then washed and incubated in the dark for 30 min at 37 °C with 15 µM dihydroethidium (DHE) in DMEM. When oxidized by a superoxide anion, DHE produces 2-hydroxyethidium, which intercalates DNA, staining the nucleus with red fluorescence. After staining, the cells were washed with phosphate-buffered saline (PBS) followed by high glucose PBS (4.5 g/L). Fluorescence was monitored in a temperature-controlled (37 °C) SpectraMax M3 Microplate Reader (Molecular Devices) using the excitation and emission wavelengths of 535 and 635 nm, respectively. Images were taken by FLoid^® Cell Imaging Station (ThermoFisher).

H₂O₂ release from cells was monitored by measuring the conversion of Amplex Red to the highly fluorescent resorufin in the presence of added horseradish peroxidase. A total of 10⁴ cells were incubated in 96-well plates, and after differentiation, the myotubes were treated with 1, 10 and 50 µM pravastatin for 12 h. Cells were incubated in a mixture containing 50 μM Amplex Red reagent (Invitrogen) and 0.1 U/mL horseradish peroxidase in Krebs–Ringer phosphate buffer (in mM: 145 NaCl, 5.7 sodium phosphate, 4.86 KCl, 0.54 CaCl₂, 1.22 MgSO₄, and 22 glucose, pH 7.35). This assay was conducted in the presence and absence of catalase (500 U/mL) for 1 h. Fluorescence was monitored over time in a temperature-controlled (37 °C) SpectraMax M3 Microplate Reader (Molecular Devices) using the excitation and emission wavelengths of 560 and 590 nm, respectively.

Data are presented as the mean ± the standard error (SE), and the number (n) of mice or cell passage is indicated in each figure. Comparisons between two groups were analysed by unpaired Student’s t-test. Multiple comparisons were tested using one-way analysis of variance (ANOVA) with LSD (Fisher’s least significance difference) post hoc test (GraphPad Prism, RRID: SCR_002798; URL: http://​www.​graphpad.​com). Animal and cell sample size (n) for each experiment was chosen based on similar studies in the literature. The level of significance was set at p ≤ 0.05.

Glucose intolerance is a feature of type 2 diabetes. Therefore, we determined glucose tolerance (GTT) after 3, 6 and 10 months of pravastatin treatment. The mean areas under the glycaemia curves after an oral load of glucose were significantly increased in the groups of mice treated for 6 (35%) and 10 (55%) months with pravastatin compared with their respective non-treated control groups (Fig. 1). Thus, long-term pravastatin treatment induced glucose intolerance in LDLr^−/− mice. Global body insulin sensitivity was evaluated by the insulin tolerance test (ITT) performed in fasted mice that had been refed for 3 h. After insulin administration, the blood glucose disappearance rate during the first 10 min (K_itt) markedly decreased (Fig. 2), and the area under the glycaemia curve significantly increased (~ 50%) in mice treated with pravastatin for 10 months (5496 ± 261 vs. 8150 ± 1086, n = 4–5, p < 0.05), characterizing an insulin resistance state induced by long-term treatment with pravastatin.

We have previously shown that ex vivo glucose-stimulated insulin secretion is impaired after 2 months of pravastatin treatment in LDLr^−/− mice [8]. Here, we observed that glucose (11.1 mM)-stimulated insulin secretion by isolated islets was decreased by 57%, 43% and 30% in mice treated with pravastatin for 3, 6 and 10 months, respectively, compared with non-treated mice (Fig. 3). Together, these results suggest that tissue insulin resistance needs to be established to induce glucose intolerance (6- and 10-month treatment), since pancreatic dysfunction without glucose intolerance was observed in the 3-month treatment.

To evaluate myotoxicity, we measured indicators of muscle protein turnover, such as protein synthesis and degradation rates. We observed that LDLr^−/− mice treated with pravastatin for 10 months presented disrupted muscle protein turnover due to increased gastrocnemius protein degradation as detected by greater tyrosine release (Fig. 4a) and no changes in protein synthesis rates (Fig. 4b). Increased protein degradation in the muscles of LDLr^−/− mice in the 10-month pravastatin treatment group was confirmed by markedly elevated chymotrypsin-like proteasomal activity (923 ± 89 vs 3470 ± 508 units/µg protein/h, n = 4, p = 0.008) and lysosomal cathepsin H (1.1 ± 0.2 vs 2.1 ± 0.2 units/µg protein/min, n = 4, p = 0.02). Accordingly, this long period of pravastatin treatment induced a decrease in the total gastrocnemius muscle protein content (Fig. 4c). The latter finding might explain the lower body weight found during this period (Table 1).

Next, we investigated possible mechanisms involved in pravastatin-induced muscle protein degradation using a cell culture model, C2C12 myotubes. The differentiated C2C12 mouse cell line (myotubes) is a well-established in vitro model of skeletal muscle [37]. Impaired insulin activation of AKT reflects muscle insulin resistance and induces atrophy and apoptosis [17]. To test this possibility, C2C12 myotubes were pretreated with increasing concentrations of pravastatin (1, 10 or 50 µM) for 12 h and then exposed to insulin (100 nM) for 10 min. The effect of insulin on the phosphorylation of the serine 473 residue of AKT was significantly decreased in a pravastatin dose-dependent manner. Total AKT levels were not affected (Fig. 5a, b). These results indicate that pravastatin decreases insulin sensitivity in myotubes, confirming the observation in whole organisms (ITT, Fig. 2e, f) and previous studies with other statins. These findings may explain the loss of muscle mass in LDLr^−/− mice treated for a long period of time (Fig. 4c).

To evaluate whether pravastatin induced muscle dysfunction, including cell death, we determined key markers of apoptosis in pravastatin-treated C2C12 myotubes. Cells treated with 50 µM pravastatin for 12 h showed an increased level of cleaved (activated) caspase-3 protein (Fig. 6a, b), the final apoptosis effector. In addition, the pro-apoptotic protein Bax was upregulated by pravastatin treatment (Fig. 6a, c), and the Bax/Bcl-2 ratio, used as an indicator of apoptotic potential, was increased in cells treated with 50 µM pravastatin (Fig. 6e).

Several studies have shown that statins might impair mitochondrial function in skeletal muscle cells [20, 22, 38], and mitochondrial dysfunction is associated with both apoptosis and oxidative stress. Superoxide anions may be generated at several cell sites, including mitochondrial complex I and III of the electron transport chain. Thus, we measured global cellular superoxide using dihydroethidium (DHE), a probe attacked by superoxide that generates the fluorescent product hydroxyethidium. We observed an increase in the hydroxyethidium signal when C2C12 cells were treated with pravastatin for 12 h in a dose-dependent manner (Fig. 7a, b). Superoxide may be converted into other reactive radicals (hydroxyl or peroxynitrite) or H₂O₂ by cytosolic and mitochondrial superoxide dismutases. Thus, we also measured H₂O₂ cell release using the specific Amplex red cell impermeable probe. No differences in the release of H₂O₂ from pravastatin-treated cells were observed (Fig. 7c).