PTEN Expression Contributes to the Regulation of Muscle Protein Degradation in Diabetes

Experiments were approved by the Baylor College of Medicine Animal Care Committee. db/db mice aged 10 weeks (The Jackson Laboratory), models of insulin resistance, were compared with littermate, wild-type mice. Four months after 65 mg streptozotocin (STZ)/kg was injected intraperitoneally on 2 consecutive days, long-term diabetes was assessed in comparison with control age-matched C57B6 mice. The acute type 1 diabetic mice were injected intraperitoneally with 150 mg · kg⁻¹ · day⁻¹ STZ on 2 days and 5 days later were compared with control mice injected with the diluent and pair fed against the STZ-injected mice (2,7). Diabetes was confirmed by glucosuria (Urine Strip; Fisher, Fairlawn, NJ). We also studied PTEN^+/− mice and pair-fed littermate wild-type mice before and after STZ injection (2,21).

C2C12 myoblasts (ATCC; Manassas, VA) were grown in Dulbecco's modified Eagle's medium (DMEM) plus 10% fetal bovine serum (FBS). At 70% confluence, the media was changed to DMEM plus 2% horse serum and resulting myotubes treated with 2% horse serum and BSA or horse serum plus 0.4 mmol/l palmitate in BSA. We also examined C2C12 myotubes treated with 0.4 mmol/l palmitate with or without expression of a dominant-negative p38 produced by mutagenesis of two phosphorylation sites of p38α (180T/A and 182Y/F). Other C2C12 cells were treated with 10% FBS with or without the PI3K inhibitor, 10 μmol/l LY294002 (Calbiochem, San Diego, CA). After 0, 3, 6, and 12 h, cells were harvested for analyses of proteins or mRNAs.

Primary myoblasts were isolated from hindlimb muscles of PTEN^+/− and littermate wild-type mice as described by Rando and Blau (22). Growth media consisted of 1:1 DMEM:F10 media, 10% FBS, 5 ng/ml bFGF, 200 units/ml penicillin G, and 200 μg/ml streptomycin. The rate of protein degradation in these cells was calculated from the l-[U-¹⁴C]phenylalanine released from prelabeled cells as previously described (23).

Arterial blood from anesthetized (20 mg/kg ketamine; 4 mg/kg xylazine) mice was used to measure glucose. Tibialis anterior muscles were used to evaluate total protein content, and soleus and extensor digitorum longus muscles were used to determine protein synthesis and degradation. The lateral gastrocnemius muscle was immediately frozen in liquid nitrogen and stored at −80°C; the medial gastrocnemius was fixed at resting length and imbedded in HistoPrep Frozen Embedding Media (Fisher, Pittsburgh, PA), and 5-μm sections were stained for laminin or caspase-3. To determine the cross-sectional area, we used Image Pro software (Silver Spring, MD) and assessed 600 muscle cells in each section. Caspase-3 activity in muscle was detected by immunofluorescence using an anti–cleaved caspase-3 antibody (Cell Signaling, Beverly, MA) followed by Alex 488-secondary antibody (7). Myofibers were outlined by laminin staining, and nuclei were visualized by DAPI. The negative control was normal rabbit lgG instead of the primary antibody; autofluorescence was suppressed using autofluorescence eliminator (Chemicon, Temecula, CA). The percentage of caspase-3–positive fibers in a cross-section of the entire middle gastrocnemius muscle was calculated for each of six mice in each group. To measure the total protein content, the entire tibialis anterior muscle was grinded in 0.5 N NaOH; after complete dissolution, protein concentration was determined by the DC Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA). Protein synthesis in isolated, incubated muscles was measured as the incorporation of l-[U-¹⁴C]phenylalanine, while protein degradation was evaluated from the release of tyrosine (24).

The lateral gastrocnemius muscle (∼50 mg) was homogenized in 0.5 ml lysis buffer (50 mmol/l Tris [pH 7.4]; 1% NP-40; 0.25% Na-deoxycholate; 150 mmol/l NaCl; 1 mmol/l EDTA; 1 mmol/l phenylmethylsulfonyl fluoride; 1 μg/ml each of aprotinin, leupeptin, and pepstatin; 1 mmol/l Na₃VO₄; and 1 mmol/l NaF). After centrifugation (16,000g) at 4°C for 15 min, supernatant proteins were analyzed by Western blotting using the following antibodies: p-Akt (Thr437), p-FOXO, p-p38, and PTEN (Cell Signaling, Beverly, MA), as well as Akt1, FOXO1, p38, and glyceraldehyde-3-phosphate dehydrogenase (Santa Cruz Biotechnology, Santa Cruz, CA). The 14-kDa actin fragment as another index of muscle proteolysis was measured as described (6,7). Densities were determined with an infrared laser scanner (Odyssey LiCor, Lincoln, NB).

Gastrocnemius muscle RNA was isolated (TRI reagent; Sigma-Aldrich, St. Louis, MO), and 20 μg RNA was separated on a 1% agarose gel, transferred to a nylon transfer membrane (Amersham Pharmacia Biotech, Piscataway, NJ), and ultraviolet cross-linked. Hybridization was performed (Quick Hybrid; Stratagene, La Jolla, CA) for 2 h at 64°C before washing the blot twice with 2× and once with 0.5× SSC (saline sodium citrate buffer) (1× SSC: 0.15 mol/l NaCl plus 0.015 mol/l sodium citrate), containing 0.1% sodium dodecyl sulfate (7). α-³²P dCTP–labeled atrogin-1, MurF1, or PTEN cDNA probes were synthesized using a random primer labeling kit (Amersham) and purified with G-25 columns (Roche, Indianapolis, IN). Real-time PCR was performed as described using 2 μg total RNA and the following primers: mouse PTEN, forward 5′-AATTCCCAGTCAGAGGCGCTATGT-3′ and reverse 5′-GATTGCAAGTTCCGCCACTGAA CA-3′, and mouse GAPDH (glyceraldehyde-3-phosphate dehydrogenase), forward 5′-ACCCCCAATGTATCCGTTGT-3′ and reverse 5′-TACTCCTTGGAGGCCATGTA-3′ (3).

The lateral gastrocnemius muscle (∼50 mg) was homogenized for 5 min in 1 ml buffer (137 mmol/l NaCl, 20 mmol/l Tris-HCl [pH 7.4], 1 mmol/l CaCl₂, 1 mmol/l MgCl_2, 0.1 mmol/l Na₃VO₄, 2 μg/ml leupeptin, 2 μg/ml aprotinin, 1 mmol/l phenylmethylsulfonyl fluoride, and 1% Nonidet-P40) and kept on ice for 15 min before centrifuging at 13,000g for 10 min. Lysate (2 mg) was immunoprecipitated with 2 μg anti–IRS-1 antibody overnight at 4°C (3). Immunoprecipitates were washed three times with buffer and three times with TNE buffer (20 mmol/l Tris-HCl [pH 7.4], 100 mmol/l NaCl, and 0.5 mmol/l EGTA). We measured PI3K activity as described, and ³²P-labeled inositol lipids were visualized and analyzed by autoradiography using NIH Image J (7).

PTEN activity was measured by immunoprecipitating 600 μg muscle lysates with 20 μl anti-PTEN antibody overnight, followed by incubation with protein A/G-agarose for 1 h at 4°C. Precipitates were washed five times with lysate buffer containing 10 μm dithiothreitol. The enzyme activity was measured using an enzyme-linked immunosorbent assay kit (Echelon, Salt Lake City, UT).

PIP₃ in ∼100 mg fresh gastrocnemius muscle was measured by quickly homogenizing it in 1 ml 1:1 methanol:chloroform plus 50 μl 12 N HCl. After centrifugation (16,000g for 10 min), the organic phase was dried under vacuum, the residue was suspended in 40 μl chloroform, and the phospholipids were separated on activated thin-layer chromatography (TLC) plates (Sigma-Aldrich) using 7:10:1.5:2.5 chloroform:methanol:ammonium hydroxide:water over 3 h (25). Minutes after intravenously injecting insulin (150 μg/kg), muscle PIP₃ was measured as a positive control. After drying, TLC plates were treated with Odyssey blocking solution for 30 min and probed overnight with an anti-PIP₃ monoclonal antibody (Echelon). IR dye-800–labeled rabbit anti-mouse lgG antibody (Li-Cor) was incubated with TLC plates for 1 h at room temperature and scanned with the Odyssey system.

Results are means ± SE. Differences between two groups were analyzed by the t test; multiple comparisons were analyzed by ANOVA with a post hoc analysis by the Student-Newman-Keuls test for multiple comparisons. P < 0.05 was considered statistically significant.

In atrophying muscles from mice studied 5 days after STZ-induced insulin deficiency as a model of type 1 diabetes, PTEN protein levels were ∼50% lower than those in muscle of control mice (Fig. 1A). There was no change in PTEN mRNA, suggesting that acute insulin deficiency involves posttranslational modification of PTEN (Fig. 1B). Similar results occurred in muscle of mice fasted for 24 h (Fig. 1 C and D), indicating that results in acute diabetes were not simply a toxic response to STZ.

With prolonged abnormalities in insulin signaling (e.g., models of type 2 diabetes or long-term STZ injection), muscle PTEN expression differed markedly from responses to acute diabetes. There was a 2.5-fold increase in PTEN in muscle of insulin-resistant db/db mice and a 60% increase in muscle of mice injected 4 months earlier with STZ. In addition, PTEN mRNA was increased threefold in muscle of db/db mice and ∼50% in mice with long-term diabetes (Fig. 1 B and D). These changes suggest that PTEN is transcriptionally regulated in chronic insulin deficiency/resistance.

Wang et al. (26) reported that p38 can influence PTEN expression; we therefore evaluated changes in the p38 pathway. As shown in Fig. 2A, the level of phosphorylated p38 in muscle was not significantly changed in response to acute insulin deficiency or fasting but was increased in muscle of db/db mice and mice with long-term STZ-induced diabetes. To explore mechanisms for these differences, we incubated C2C12 myotubes with palmitic acid and found that it increased the level of activated p38 and levels of PTEN protein and mRNA. We also incubated C2C12 muscle cells with palmitic acid in the presence of dominant-negative p38; p38 inhibition blocked the increase in PTEN mRNA and protein (Fig. 2B and C).

Since acute diabetes or fasting is associated with decreased PI3K activity (7), we treated C2C12 cells with a specific PI3K inhibitor, LY294002 (10 μmol/l), and measured PTEN. There was a rapid decrease in PTEN over 6 h and a return toward control levels by 12 h (Fig. 2D). Note that there was no change in p-p38 or the mRNA of PTEN over 12 h, consistent with findings in vivo (Fig. 2E). These results suggest that downregulated PTEN in acute diabetes involves posttranslational mechanisms, while PTEN upregulation in chronic diabetes involves transcriptional mechanisms.

To determine the physiologic relevance of changes in PTEN expression on muscle proteolysis, we established primary cultures of muscle cells from wild-type mice and mice with partial PTEN deficiency. The cells were differentiated into myotubes by incubating in 2% horse serum DMEM. We stimulated proteolysis by serum withdrawal and measured the rate of protein degradation over 24 h. Protein degradation rates in wild-type and PTEN^+/− myotubes incubated in 2% horse serum were not different. However, when proteolysis was stimulated by serum starvation, the increase in protein degradation occurring in wild-type cells was significantly (P < 0.05) reduced in PTEN^+/− myotubes (Fig. 3A).

To examine the mechanism by which a decrease in PTEN could protect against excessive muscle protein losses, we measured p-Akt because we and others have shown that low p-Akt activity in muscle is associated with an increase in atrogin-1/MAFbx expression and protein degradation (7,27,28). p-Akt was not different in PTEN^+/− or wild-type myotubes cultured in 2% horse serum. But with serum starvation, p-Akt in PTEN^+/− myotubes was higher than in wild-type muscle cells (Fig. 3B). We also assessed the expression of atrogin-1/MAFbx. Serum starvation significantly increased atrogin-1 mRNA expression in both types of cultured myotubes, but the increase was significantly less in cells from PTEN^+/− mice than in those from wild-type mice (Fig. 3C).

Decreased IRS-1–associated PI3K activity increases muscle protein degradation in catabolic conditions including insulin resistance and acute insulin deficiency (2,3,21,27,29). To determine whether reduced PTEN expression can counteract muscle protein catabolism in vivo, we minimized changes in PI3K by injecting a high dose of STZ into wild-type and PTEN^+/− mice. After 5 days, the blood glucose in STZ-injected mice was approximately sixfold higher than in wild-type mice (33.5 ± 2.8 vs. 5.2 ± 0.7 mmol/l, respectively), and the islet damage in both types of mice was similar (data not shown). Without STZ injection, the cross-sectional area of muscle fibers and the total muscle protein from pair-fed wild-type and PTEN^+/− mice did not differ. With acute insulin deficiency, the muscle cross-sectional area and protein content in wild-type mice were significantly less than those in pair-fed PTEN^+/− mice (Fig. 4 A and B).

We examined whether changes in protein synthesis or degradation in red-fiber soleus or white-fiber extensor digitorum longus could account for the muscle loss. The rates of protein synthesis in each type of muscle from diabetic wild-type and PTEN^+/− mice were not different (Fig. 4C). In contrast, the rate of protein degradation was significantly suppressed in both types of muscle obtained from diabetic PTEN^+/− mice compared with control values from wild-type mice (Fig. 4D).

We then examined whether partial deletion of PTEN counteracts two components of muscle proteolysis: actomyosin cleavage mediated by caspase-3 and activation of the UPS. In acute diabetes, activated caspase-3 in individual myofibers was significantly less (P < 0.05) in muscle of PTEN^+/− mice versus values in pair-fed, wild-type mice (Fig. 5A). The resulting characteristic 14-kDa actin fragments were also decreased in muscle of diabetic PTEN^+/− mice (Fig. 5B) (P < 0.01). We evaluated UPS activation by Northern analyses of the E3 ubiquitin–conjugating enzymes atrogin-1/MAFbx and MurF1 mRNAs. Likewise, atrogin-1/MAFbx and MurF1 mRNAs were significantly (P < 0.01) lower in muscles of acutely diabetic PTEN^+/− mice (Fig. 6A). Finally, we examined the mechanism that stimulates the expression of atrogin-1/MAFbx by evaluating changes in Akt activity and the activation of FOXO1. We found higher levels of p-Akt (Fig. 6B) and phosphorylated FOXO1 in muscle of acutely diabetic PTEN^+/− mice (Fig. 6C). Thus, a decrease in PTEN will suppress two important steps in proteolysis, activation of caspase-3 and the UPS, to provide a mechanism by which decreased PTEN counteracts muscle protein breakdown induced by acute diabetes.

Acute insulin deficiency decreases IRS-1–associated PI3K activity, but this cannot account for the antiproteolytic influence of PTEN^+/− mice because PI3K activities were not statistically different in muscles of the two groups of acutely diabetic mice (Fig. 7A). The PTEN protein and the PTEN activity were both significantly lower in muscle of diabetic PTEN^+/− mice compared with diabetic pair-fed wild-type mice (Fig. 7,B and C). Consequently, the level of PIP₃ was significantly higher in muscle of diabetic PTEN^+/− mice compared with pair-fed diabetic wild-type mice (Fig. 7D). This increase in PIP₃ was consistent with an increase in both p-Akt and phosphorylated FOXO1 in muscles of acutely diabetic PTEN^+/− mice compared with identically treated wild-type mice (Fig. 6 B and C). These results indicate that regulation of muscle proteolysis in acute insulin deficiency results from changes in PTEN as well as PI3K.

Recent reports suggest that in many catabolic conditions, decreased signaling through the PI3K/Akt pathways activates muscle protein degradation (3,6,7,29). Contrariwise, when PI3K/Akt signaling is activated, two key proteolytic events are suppressed: actomyosin cleavage and expression of key E3 ubiquitin ligases atrogin-1/MAFbx and MurF1 (7,27,28). In addition, an endogenous mechanism changes the activity of PTEN to regulate changes in muscle protein turnover; a decrease in PTEN yields an increase in PIP₃ and suppression of muscle proteolysis. In muscle of mice with STZ-induced acute diabetes, we found that PTEN is, in fact, downregulated, and in cultured muscle cells when PTEN expression is suppressed, protein degradation is decreased. Interestingly, the mechanism that suppresses PTEN in cultured myocytes is linked to acute inhibition of PI3K activity.

In contrast to acute diabetes, in chronic diabetes, PTEN mRNA and protein in muscle are increased. This is relevant because an increase in PTEN in chronic diabetes provides an additional positive feed-forward stimulus of muscle protein degradation.

What factors change PTEN expression? PTEN in muscle can be downregulated via a posttranslational modification (30). When we suppressed PI3K by creating acute diabetes or treated C2C12 myotubes with LY294002, PTEN expression, but not PTEN mRNA, decreased (Figs. 1,A and B and 2 C and E). Suggested posttranslational mechanisms that could explain a decrease in PTEN include the study by Torres et al. (31), who found that caspase-3 can cleave the COOH-terminus of PTEN. This is relevant because an acute decrease in insulin activates caspase-3 in muscle (7). Another mechanism affecting the degradation of PTEN is also related to insulin. Birle et al. (32) reported that PTEN is stabilized by PIP₃-induced protein kinase C phosphorylation, suggesting that insulin deficiency could reduce PTEN stability, leading to its degradation.

The mechanism for the increase in PTEN expression with chronic diabetes also appears to be complex, but in this case, the mechanism involves an increase in PTEN mRNA. Wang et al. (26) found that incubating endothelial cells with free fatty acids increases the activity and expression of PTEN. They related the PTEN increase to p38 mitogen-activated protein kinase stimulation. Our findings in cultured C2C12 muscle cells are consistent with these results. We confirmed that palmitic acid increases p38 activation and the expression of PTEN (Fig. 2A, B, and D). In fact, the chronic insulin resistance models that we studied are characterized by higher free fatty acid levels compared with those in animals with acute insulin deficiency (33,34). Finally, we found that inhibition of p38 suppresses the palmitic acid–induced increase in both PTEN mRNA and protein (Fig. 2 A and D). Taken together, the results suggest that PTEN transcription and expression are related to p38 activation.

How do variations in PTEN change muscle protein degradation? The balance between PIP₃ formation and its conversion to phosphatidylinositol 4,5-bisphosphate in muscle changes when PTEN is reduced; PIP₃ levels rise, and protein degradation is suppressed (Figs. 6 and 7). These responses can explain how partial deletion of PTEN counteracts insulin deficiency–induced muscle atrophy even though IRS-1–associated PI3K activity remains low (Fig. 7A). The mechanism is due to suppression of actomyosin/myofibril cleavage (Fig. 5B) plus an increase in phosphorylated FOXO1, which decreases the expression of atrogin-1/MAFbx and reduces proteolysis in the UPS (Fig. 6). Our finding that the 14-kDa actin fragment was reduced in muscle of acutely diabetic PTEN^+/− mice is important because it signifies that an initial step in muscle proteolysis occurred to augment the degree of muscle proteolysis. The same result occurs in several catabolic conditions, including those affecting patients (3,6–8,11,29). Earlier, we found that there was increased caspase-3 activity and an increase in the 14-kDa actin fragment in muscle of rodents with acute diabetes (7). When we examined the influence of actomyosin cleavage in vitro, there was a large increase in degradation of actomyosin in response to small increases in actomyosin cleavage (6). Likewise, the decrease in atrogin-1/MAFbx in muscle of diabetic PTEN^+/− mice is important because Bodine et al. (10) reported that forced expression of atrogin-1/MAFbx in myotubes increases proteolysis and that knockout of the gene suppresses the protein breakdown induced by denervation.

Importantly, we found no difference in the size of muscles of mice with partial PTEN deletion compared with that of littermate wild-type controls unless the mice were challenged by insulin deficiency. These results reflect the equivalent levels of PIP₃ in muscles of nondiabetic PTEN^+/− and wild-type mice (Fig. 7D). Consistent with these results, Wijesekara et al. (17) reported that p-Akt is unchanged in mouse muscle even when there is muscle-specific knockout of PTEN. However, we found that acute insulin deficiency reduces PTEN levels in muscle of both wild-type and PTEN^+/− mice, resulting in higher levels of PIP₃ and p-Akt in mice with partial PTEN deletion (Figs. 6B and 7D). Taken together, these results demonstrate that a decrease in PTEN will counteract the muscle proteolysis usually triggered when there is downregulation of PIP₃/p-Akt levels (Figs. 4–7).

In contrast to the different mechanisms that change PTEN expression in acute and chronic diabetes, the mechanisms causing muscle proteolysis are similar. A low p-Akt leads to activation of caspase-3 and increased expression of atrogin-1/MAFbx; the latter response proceeds from a decrease in p-Akt–dependent phosphorylation of FOXOs (7,9,27,28).

How might the responses we identified play a physiologic role in overall metabolism? In normal subjects, short-term fasting could suppress PI3K/Akt signaling and stimulate proteolytic pathways in muscle, producing gluconeogenic amino acids. However, fasting also decreases PTEN to counteract the increase in muscle protein degradation. In this way, protein balance is minimally displaced. In contrast, prolonged insulin deficiency/resistance enhances gluconeogenesis from substrates derived in part from protein degradation in muscle. Consequently, an increase in PTEN acts in concert with a decrease in IRS-1–associated PI3K activity to lower PIP₃. This response would provide a positive forward feed to activate muscle protein breakdown.

In summary, we find that there is variability in the expression and activity of PTEN in different forms of diabetes. This variability exerts physiologically important responses in muscle protein metabolism. Potentially, PTEN could be a therapeutic target for reducing accelerated muscle atrophy.

This study was supported by grants from the National Institutes of Health (P50 DK64233, R01 DK37175, and DK62796), grant R01 HL70762, and American Diabetes Association Junior Faculty Award 1-04-JF-48. I.H.L. was supported by the Research Institute of Medical Science, Catholic University of Daegu, Daegu, South Korea.