Chronic kidney disease induces autophagy leading to dysfunction of mitochondria in skeletal muscle

Abstract

Chronic kidney disease (CKD) causes loss of lean body mass by multiple mechanisms. This study examines whether autophagy-mediated proteolysis contributes to CKD-induced muscle wasting. We tested autophagy in the muscle of CKD mice with plantaris muscle overloading to mimic resistance exercise or with acupuncture plus low-frequency electrical stimulation (Acu/LFES) treatment. In CKD muscle, Bnip3, Beclin-1, and LC3II mRNAs and proteins were increased compared with those in control muscle, indicating autophagosome-lysosome formation induction. Acu/LFES suppressed the CKD-induced upregulation of autophagy. However, overloading increased autophagy-related proteins in normal and CKD muscle. Serum from uremic mice induces autophagy formation but did not increase the myosin degradation or actin break down in cultured muscle satellite cells. We examined mitochondrial biogenesis, copy number, and ATP production in cultured myotubes, and found all three aspects to be decreased by uremic serum. Inhibition of autophagy partially reversed this decline in cultured myotubes. In CKD mice, the mitochondrial copy number, biogenesis marker peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), mitochondrial transcription factor A (TFAM), and mitochondrial fusion marker Mitofusin-2 (Mfn2) are decreased. Both muscle overloading and Acu/LFES increased mitochondrial copy number, and reversed the CKD-induced decreases in PGC-1α, TFAM, and Mfn2. We conclude that the autophagy is activated in the muscle of CKD mice. However, myofibrillar protein is not directly broken down through autophagy. Instead, CKD-induced upregulation of autophagy leads to dysfunction of mitochondria and decrease of ATP production.

chronic kidney disease (CKD) impairs the control of muscle protein metabolism resulting in muscle atrophy. Our previous study showed that a major pathway for the degradation of protein in CKD-induced muscle wasting is the ubiquitin-proteasome system (UPS) (39, 45). Activation of protein degradation by the UPS can be triggered by defects in insulin and/or insulin-like growth factor (IGF-1) signaling. When insulin/IGF-stimulated intracellular signaling is interrupted, the forkhead (FoxO) transcription factors are activated to translocate into the nucleus where they promote expression of E3 ubiquitin ligases TRIM63/MuRF1 and FBXO32/atrogin-1. These E3 ligases specifically recognize and facilitate muscle protein degradation through the UPS (30, 44). Although FoxO1, FoxO3, and FoxO4 are each present in skeletal muscle, we found that FoxO1 is the major mediator of CKD-induced muscle wasting (30, 44).

FoxO transcription factors also can activate macroautophagy (hereafter referred to as autophagy), a cellular response that is reportedly associated with muscle atrophy in several catabolic conditions, including starvation, denervation, disuse, sepsis, and cancer (23–25). Autophagy is essential for intracellular protein and organelle quality control and it is a normal process in muscle. There is evidence that FoxO stimulates the promoters of several autophagy-related genes (21, 46). Zhao et al. (46) showed that FoxO directly binds to the promoters of LC3b, Ganarapl1, and autophagy-related gene 12 (Atg 12) and stimulates their mRNA expression. Mammucari et al. (19) reported the presence of several FoxO response elements in the Bnip3 promoter region, and that overexpression of FoxO increases Bnip3 protein. Bnip3 can induce autophagy by interacting with Bcl-2 to release Beclin1 (also known as Atg 6) allowing formation of a Beclin1/VPS34 complex (10). The autophagy pathway begins with the formation of the phagophore around cytoplasmic degradation targets in the cytoplasm. The Beclin1/VPS34 regulatory complex upregulates phosphoinositide-3-kinase class III (PI3KC3) signals to recruit microtubule-associated protein 1A/1B-light chain 3 (LC3) to allow phagophore membrane elongation and autophagosome formation (16, 28). One of the most characteristic proteins that serve as markers for the autophagy process is LC3, which has two isoforms, LC3-I and LC3-II. Autophagosomes convert LC3-I to LC3-II, therefore, the ratio of LC3-II to LC3-I is indicative of initial autophagy activation. After formation, the autophagosome travels to and fuses with a lysosome where the contents of the autophagosome are degraded by acidic lysosomal hydrolases (22). Autophagosome substrates frequently include the protein p62/SQSTM1, so levels of this protein serve as an indicator of autophagosome clearance (31). Autophagy can selectively eliminate bulky elements in the cytoplasm, including abnormal protein aggregates and damaged organelles including mitochondria (27, 34). The precise mechanism of autophagy in the muscle of CKD mice is not yet elucidated.

Understanding the pathways that cause loss of muscle proteins in CKD is necessary for designing therapeutic approaches to limit muscle protein loss. Attempts to suppress CKD-induced muscle wasting have included efforts to prevent activation of the UPS, or to increase protein synthesis by exercise, or both. Removing the gastrocnemius and soleus muscles to create muscle overloading (OL) of the plantaris muscle is a resistance exercise model for studying muscle wasting. An additional method for chronically stimulating muscles is to administer acupuncture coupled with low-frequency electric stimulation (Acu/LFES). In mice with CKD, muscle overloading blunted CKD muscle loss by decreasing muscle proteolysis and stimulating muscle protein synthesis (37). In CKD mice treated with Acu/LFES, we showed that IGF-1 production was increased, which stimulated muscle regeneration (14). Both models have shown attenuation of the loss of muscle mass in CKD, indicating that exercise and Acu/LFES are viable treatments for muscle loss (14, 37). Neither study examined the effect on CKD-induced activation of autophagy.

We hypothesized that autophagy is induced in the muscle of CKD mice. In this current study, we evaluated autophagy signaling in muscles of CKD mice by comparing whether muscle overloading or Acu/LFES would regulate autophagosome formation. Our results identify how CKD-induced changes in autophagy signaling affect myofibrillar protein (actin, myosin heavy chains, or myosin light chains) abundance and mitochondrial function.

MATERIALS AND METHODS

Animals and CKD model.

Male C57/BL6 mice (2–4 mo old) were used to create CKD by subtotal nephrectomy, followed by surgery to produce muscle OL or Acu/LFES treatments (37). All experiments were approved by the institutional animal care and use committee of Emory University. CKD was induced by removing the right kidney and two poles of the left kidney in anesthetized mice (xylazine, 12 mg/kg; ketamine, 60 mg/kg). Hemostasis was achieved by cautery and pressure. Initially, mice were pair-fed 14% protein, 4% fat, 73% carbohydrate (Harlan Teklad) with a weight-matched mouse that underwent sham operation; the drinking water contained 0.45% NaCl. After 7 days, all mice were pair-fed a high-protein diet (40% protein) for an additional 2 wk. Blood urea nitrogen (BUN) was measured with a BUN Kinetic Procedure Kit (Thermo Electron, Louisville, CO). Soleus and plantaris muscle dry weights were measured after individual muscles were dehydrated at 60°C for 48 h. Gastrocnemius muscle was freeze-clamped and plunged into liquid nitrogen, then stored in −80°C freezers for later RNA and protein analysis.

Plantaris muscle overloading.

To mimic resistance exercise, we studied muscle overloading. Three days after subtotal nephrectomy, plantaris muscles of mice were overloaded by removing gastrocnemius and soleus muscles from both hindlimbs (37). After 2 wk, left plantaris muscles were harvested for measuring dry weight, and right plantaris were frozen in liquid nitrogen and stored at −80°C for RNA and protein measurements.

LFES treatment.

The mice were maintained in a specially designed restraint so they remained in a recumbent position during LFES treatment (14). Acupuncture points were chosen from the World Health Organization Standard Acupuncture Nomenclature (18). The positive point (Yang Ling Quan, GB34) is under the front head of the fibula ~6 mm deep (in mice that weigh 20 g). It is close to the superficial fibular nerve and deep fibular nerve. The negative point (Zu San Li, ST36) is outside of the knee joint and under the head of the fibula ~7 mm deep; it is close to the fibular nerve. Acupuncture started 3 days after subtotal nephrectomy. Needles were connected to an SDZ-II Electronic acupuncture instrument using consistent pulse, electric frequency of 20 Hz and electric current of 1 mA for 15 min every day. Disposable sterile needles with a diameter of 0.25 mm (Shen Li Medical & Health Material, Wujiang, China) were used to deliver the acupuncture. After 2 wk of LFES treatment, mice were killed, and soleus and plantaris muscle were dissected for dry weight measurement; gastrocnemius muscles were frozen in liquid nitrogen and stored at −80°C for RNA and protein measurements.

Quantitative real-time polymerase chain reaction (qPCR).

Total RNA was extracted using Tri-Reagent (Molecular Research, Cincinnati, OH). RNA concentrations were measured by spectrophotometry (Nano Drop Lite; Thermo Scientific, Wilmington, DE). To remove contaminating DNA, samples were treated with recombinant RNase-free DNase I (Thermo Fisher Scientific, West Palm Beach, FL). RNA was reverse transcribed to cDNA using the M-MLV Reverse Transcriptase (Invitrogen, Carlsbad, CA). Real-time qPCR was performed on CFX Connect Real-Time PCR Detection System (Bio-Rad, Hercules, CA) using SsoAdvanced Universal SYBR GRN SMX (Bio-Rad). The following cycle parameters were used: 94°C for 2 min and 40 cycles at 94°C for 15 s, 55°C for 30 s, and 72°C for 30 s with a final extension at 72°C for 10 min. The Cq (quantification cycle) was defined as the number of cycles required for the fluorescence signal to exceed the detection threshold. Individual expressions of mRNAs were standardized to 18S gene and the mRNA expression was calculated as the difference between the threshold values of the two genes (ΔΔcq). Melting curve analysis was always performed during real-time qPCR to analyze and verify the specificity of the reaction. Primers were designed to cross intron-exon boundaries. They were used to generate amplicon in their linear ranges listed in Table 1. For mitochondrial copy number, the 176-bp-long mitochondrial fragment within the mtND1 gene and 188 bp of the lipoprotein lipase (LPL) nuclear gene were amplified and inserted into pCNA3 vectors for measurement of copy number and generation of a standard curve. DNA was isolated from cultured myotubes or muscle using the proteinase K method (1). Real-time qPCR was performed at 94°C for 2 min and 40 cycles at 94°C for 15 s, 49°C for 30 s, and 72°C for 30 s with final extension at 72°C for 10 min. The primer sequences are listed in Table 1. Plasmid standards of known copy number were used to generate a log-linear standard curve for determining the copy number of ND1 and LPL. A quantitation of mitochondrial genome is expressed as the ratio of ND1 to LPL copy number.

Table 1.

Primer sequences

Gene	Amplicon	Gene Code
LC3II
Forward	171	NM_026160	CACTGCTCTGTCTTGTGTAGGTTG
Reverse		NM_026160	TCGTTGTGCCTTTATTAGTGCATC
Bnip3
Forward	151	NM_009760	TTCCACTAGCACCTTCTGATGA
Reverse			GAACACCGCATTTACAGAACAA
Beclin-1
Forward	272	M11154	TGAATGAGGATGACAGTGAGCA
Reverse			CACCTGGTTCTCCACACTCTTG
Atg12
Forward	242	NM_026217	TCCGTGCCATCACATACACA
Reverse			TAAGACTGCTGTGGGGCTGA
18S
Forward	106	X00686	CCA GAG CGA AAG CAT TTG CCA AGA
Reverse			TCG GCA TCG TTT ATG GTC GGA ACT
ND1
Forward	176	AY394057.1	GGCCCATTCGCGTTATTC
Reverse			GATGCTCGGATCCATAGG
LPL
Forward	188	NM_008509.2	GATGCCCTACAAAGTGTTCC
Reverse			CCACCTCCGTGTAAATCAAG
PGC-1α
Forward	112	NM_008904.2	ACTGAGCTACCCTTGGGATG
Reverse			TAAGGATTTCGGTGGTGACA
TFAM
Forward	171	NM_009360.4	TGAAGCTTGTAAATGAGGCTTGGA
Reverse			CGGATCGTTTCACACTTCGAC

Primary muscle satellite and C2C12 cell culture.

For primary culture of satellite cells, skeletal muscles from both hindlimbs were removed and gently parted using tissue forceps until the mixture was homogeneous (7). Single-cell suspensions from mouse skeletal muscles were generated with a Skeletal Muscle Dissociation Kit (Macs Miltenyi Biotec, Gaithersburg, MD). Satellite cells were separated via magnetic labeling assay with a satellite cell isolation kit (Miltenyi Biotec), then cultured on a dish with an ECL matrix gel (Upstate Biotechnology, Temecula, CA) coated in Ham’s F-10 Nutrient Mixture medium (Invitrogen) with 20% FBS, 100 u/ml penicillin, and 100 μg/ml streptomycin. Maintenance media for satellite cells contained 5 ng/ml human β-fibroblast growth factor (Atlanta Biologicals, Atlanta, GA). The growth factors were removed to let cells form to myotubes. Culture uniformity was evaluated by immunohistology and satellite cells were identified using anti-myoD antibodies. Cells collected were 99% MyoD positive. Satellite cells were not used beyond passage 5.

C2C12 myoblasts are cultured in high-glucose DMEM plus 10% FBS and antibiotics. To differentiate the cells into myotubes, the growth medium was switched to high-glucose DMEM plus 2% horse serum and antibiotics for 2–3 days. C2C12 cells were not used beyond passage 10.

Western blotting and antibodies.

To identify proteins in skeletal muscle, hindlimb muscles were homogenized in Gentle Lysis Buffer (10 mM Tris·HCl, 10 mM NaCl, 2 mM EDTA, 0.5% NP-40, and 1% glycerol; and then fresh added 1 mM Na₃VO₄; 10 μg/ml PMSF; 5 μg/ml aprotinin; and 1 μg/ml leupeptin) with phosphatase inhibitor cocktail 1 and 2 (Sigma). Protein concentration was measured using an RC-PC protein assay kit (Bio-Rad) and equal amounts of protein were loaded on acrylamide/bis SDS-PAGE gels (41). Proteins were transferred to a polyvinylidene difluoride membrane and blotted with the following specific primary antibodies: actin (catalog number A2033), LC3 (L7543), Beclin-1(B6186), and Bnip3 (B7931) from Sigma-Aldrich (St. Louis, MO); P62 (610832) from BD PharMingen (San Jose, CA); myoD (D7F2), MyHC2a (2F7), MyHC2x (6H1), and MyLC2 (MF5) from Developmental Studies Hybridoma Bank (Iowa City, IA); mtTFA (A-17) from Santa Cruz Biotechnology (Paso Robles, CA); Mfn2 (ab56889) and TRIM63/MuRF1 (ab77577) from Abcam (Cambridge, MA); Atrogin-1 (AP 2041) from ECM Biotechnology (Versailles, KY); GAPDH (14C10), mTOR (2972), pmTOR Ser²⁴⁴⁸ (2971), FoxO1 (75D8), p-FoxO1 Thr²⁴ (9464) and Atg12/5 (D88/H11) from Cell Signaling (Danvers, MA). We used secondary antibodies Alexa Fluor 680 goat anti-rabbit IgG or goat anti-mouse IgG from Invitrogen. Protein bands were scanned and quantified using the Li-Cor Odyssey infrared scanning system (Li-COR Biosciences, Lincoln, NE).

ATP production assay.

The cultured myotubes were solubilized in the lysis buffer provided in the ATP assay kit (K354-100; BioVision, Milpitas, CA). Samples were centrifuged to remove insoluble and the protein content of supernatant fractions was quantified with the DC Protein Assay kit (Bio-Rad). ATP levels in the cell lysates were determined by following the manufacturer’s instructions. The fluorescence was measured using a RF-5301PC Spectro-fluorophotometer (Shimadzu, Columbia, MD) with a 538-nm excitation wavelength and a 587-nm emission wavelength and quantification determined by comparison with a standard curve.

Autophagy detection assay.

Autophagy activities were detected by lysosomal/autophagic vacuoles fluorescent marker monodansylcadaverine (MDC) using the CYTO-ID Autophagy Detection kit (ENZ-51031) from ENZO Life Sciences (Farmingdale, NY) according to the manufacturer’s protocol. Images were visualized with an Olympus 1X51 inverted fluorescence microscope and captured with a DP73-1-51-17MP color camera with the CellSens Dimension 1.9 software (Olympus, Melville, NY).

Statistical analysis.

Values are presented as means ± SE. To identify significant differences between two groups, comparisons were made using the t-test. When multiple treatments were compared, a one-way ANOVA was performed with a post hoc analysis by the Student-Newman-Keuls test. Differences with values of P < 0.05 were considered significant.

RESULTS

CKD induces autophagy in mouse skeletal muscles.

CKD mice were used to prove our hypothesis, which is that autophagy is induced in the muscle of CKD mice. CKD was induced by subtotal nephrectomy; BUN was approximately threefold higher. The body weight and soleus or plantaris muscle dry weight were significantly decreased in the CKD mice compared with pair-fed sham-operated mice. Muscle OL and Acu/LFES reversed CKD-induced muscle atrophy (Table 2). In CKD mice, mRNA expression of Bnip3, Beclin-1, LC3II, and Atg12/5 was increased in the gastrocnemius muscles (Fig. 1), which indicates induction of autophagy. The protein levels of Bnip3, Beclin-1, and PI3KC3, which are the autophagosome formation inducers, also increased in CKD muscle. Because the formation of the autophagosome membrane requires interactions of several key autophagy proteins including LC3 and Atg12/5, we examined their abundance in CKD muscles and found that both were increased over sham-operated animals (Fig. 1B). The increase in the ratio of LC3II to LC3I proteins in CKD muscle indicates upregulation of autophagosome formation. P62 protein was reduced in CKD muscle, suggesting increased autophagic clearance. Because mTOR is a major negative regulator of autophagy, we examined levels of total and phospho-mTOR in CKD muscle. We found that mTOR Ser²⁴⁴⁸ phosphorylation is significantly decreased in the muscle of CKD mice (Fig. 1B), which could contribute to the increased autophagy.

Table 2.

Muscle dry weight and body weight

	Sham	CKD	OL	OL+CKD	Acu/LFES	CKD-Acu/LFES
Body weight, g	22.6 ± 1.2	19.6 ± 1.1*	22.8 ± 1.8	20.6 ± 1.4*#	22.1 ± 1.7	20.7 ± 1.7*#
Soleus, mg	2.2 ± 0.07	1.6 ± 0.05*	N/A	N/A	2.1 ± 0.09	1.9 ± 0.2*#
Plantaris, mg	2.4 ± 0.04	1.9 ± 0.08*	3.3 ± 0.2*	2.7 ± 0.3#	2.5 ± 0.09	2.1 ± 1.0#
BUN, mg %	29.3 ± 1.5	95.6 ± 3.8*	30.5 ± 1.9	97.8 ± 3.6*	28.6 ± 0.9	91.5 ± 3.6*

Values are means ± SE. CKD, chronic kidney disease; OL, overloading; Acu/LFES, acupuncture plus low-frequency electrical stimulation; BUN, blood urea nitrogen.

^*P < 0.05 is significant vs. sham, n = 9/group.

^#P < 0.05 vs. CKD;, n = 9/group.

Fig. 1.

The autophagosome-proteolysis pathway was activated in the muscle of CKD mice. A: total RNA isolated from gastrocnemius muscles of mice in the sham and CKD groups were assayed for autophagosome-related mRNA expression by real-time quantitative PCR. The bar graph shows mRNA from the muscles of CKD mice expressed as a fold change from mice in the sham group (represented by a line at onefold). Results are normalized to 18S RNA (values are means ± SE, n = 9, *P < 0.05 vs. sham). B: autophagosome related proteins Bnip-3, Beclin-1, PI3KC3, LC3, Atg12/5, P62, mTOR, and phospho-mTOR were measured by Western blotting in gastrocnemius muscle lysates from mice in the sham and CKD groups. The bar graph (bottom) compares the densities of protein bands in each group expressed as a fold change from levels in the sham group, which is represented by a line at onefold. All band densities were normalized to the density of GAPDH (values are means ± SE, n = 9, *P < 0.05 vs. sham).

Muscle overloading activates autophagy but prevents CKD-induced muscle loss.

In our previous study, we found that muscle OL prevented CKD-induced muscle wasting (37). To investigate whether OL also affects CKD-induced autophagy signaling, mice were randomly divided into four groups: healthy sham-operated (sham); sham-operated with plantaris muscle OL (sham/OL); CKD; and CKD plus plantaris muscles OL (CKD/OL). Mice in the sham, CKD/OL, and sham/OL groups were pair-fed with the CKD mice. The CKD mice exhibited a 21% decrease in dry weight of the plantaris muscle (P < 0.01) vs. the sham-operated mice. Mice in the CKD/OL group reversed this decrease resulting in significantly heavier muscles vs. mice with CKD (Table 2). To investigate whether OL alters CKD-induced autophagy in muscle, we measured autophagy markers. We found that muscle levels of Bnip3, Beclin-1, and the LC3II/LC3I protein ratio in sham-operated mice subjected to OL were significantly increased vs. the values from the sham-operated mice. Autophagy was increased in the CKD/OL mice, but this increase was not statistically higher than the autophagy marker level in mice with CKD alone. Muscle OL induced a 3.5-fold increase in p62 protein in the sham-operated mice, and a 1.9-fold increase in the CKD mice (Fig. 2A). Collectively, these results demonstrate that mice subjected to OL (a model of exercise) exhibit increased markers of autophagy initiation in sham-operated mice, but are not increased above CKD levels.

Fig. 2.

The impact of muscle overloading (OL) or acupuncture coupled with low-frequency electric stimulation (Acu/LFES) on autophagosomes in the muscle of CKD mice. A: muscle proteins lysates were prepared from plantaris muscles from mice in sham, muscle overloading (OL), CKD, or CKD/overload (OL) groups. Autophagosome-related proteins Bnip3, Beclin-1, LC3, and P62 were measured by Western blotting. The bar graph (bottom) compares the protein band densities in each group expressed as a fold change from levels in sham mice (represented by a line at onefold). All band densities were normalized to the density of GAPDH: OL, gray; CKD, white; and CKD/OL, black (values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. CKD). B: muscle protein lysates were prepared from gastrocnemius muscles from mice in the sham, Acu/LFES, CKD, and CKD plus Acu/LFES groups. Autophagosome-related proteins Bnip3, Beclin-1, LC3, and P62 were measured by Western blotting. The bar graph (top) compares the protein band densities in each group expressed as a fold change from levels in sham mice (represented by a line at onefold). All band densities were normalized to the density of GAPDH: Acu/LFES, gray; CKD, white; and CKD plus Acu/LFES, black (values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. CKD).

Acu/LFES prevents muscle loss and attenuates CKD-induced upregulation of autophagosome proteins.

Because we previously demonstrated that employing Acu/LFES can prevent CKD-induced loss of muscle protein (14), we examined whether Acu/LFES would have effect on the activation of autophagy process in CKD-induced muscle wasting. As with the OL model of exercise, we studied four groups of mice: sham, Acu/LFES, CKD, and CKD plus Acu/LFES. In the sham group, Acu/LFES resulted in decreased protein levels of Bnip3 monomer and Beclin-1 compared with results from mice in the sham group that did not undergo Acu/LFES, but the LC3II-to-LC3I ratio was unchanged. In the CKD group, Acu/LFES resulted in decreased proteins Bnip3 and Beclin-1, and the LC3II-to-LC3I ratio, but p62 protein was increased compared with the CKD group without Acu/LFES treatment (Fig. 2B). Thus, Acu/LFES appears to attenuate CKD-induced autophagy formation.

Uremic serum, but not acidification alone, promotes autophagy in cultured myotubes.

CKD is characterized by uremia. This condition includes metabolic acidosis as well as uremic toxicity (Fig. 1). Because metabolic acidosis stimulates the degradation of muscle proteins by activation of UPS, we examined whether acidification would activate autophagy in cultured myotubes (2). In these experiments, primary cultured satellite cells were cultured in normal growth medium for 48 h to form myotubes. The myotubes were then cultured in 1) 2% normal mice serum (control), 2) 2% serum pooled from uremic mice (pH 7.4), 3) acidified media (pH 7.1 in 2% normal mice serum), or 4) in 2% serum from uremic mice following acidification to pH 7.1. After an additional 48 h, the cells were lysed and marker protein levels were analyzed by Western blotting. Both monomeric and dimeric Bnip3 were substantially increased in uremic serum-treated cells. Uremic serum also increased Beclin-1 levels (eightfold). The LC3II-to-LC3I ratio and Atg12/5 were increased 1.7-fold or 2.5-fold, respectively, in myotubes treated with uremic serum, but not increased in the cells treated with acidified serum. However, acidification alone increased dimeric Bnip3 and Beclin-1 (Fig. 3A). To confirm these results, we detected autophagosome formation using monodansylcadaverine (MDC), a fluorescent marker for lysosomal/autophagic vacuoles. C2C12 myoblasts were treated with mouse uremic or acidification serum for 24 h, then cells were stained with MDC (CYTO-ID Autophagy Detection kit; ENZO, Farmingdale, NY). A limited number of lysosomal/autophagic vacuoles were visible in the 2% normal mouse serum-treated cells, but their presence was greatly increased in the uremic serum-treated cells. Acidification does not show a significantly increased presence of lysosomal/autophagic vacuoles (Fig. 3B). These results suggest that uremia toxins, but not acidosis, activate autophagy in vitro.

Fig. 3.

Uremic serum but not acidosis promotes activation of the autophagy pathway in cultured myotubes. A: representative Western blot of cultured myotubes cultured with or without uremic serum or acidification. Primary cultured satellite cells were grown in Ham’s F-10 Nutrient Mixture medium with 20% FBS, but without β-fibroblast growth factor to form myotubes. Myotubes were treated with 2% normal mouse serum, 2% uremic mouse serum (pH 7.4, blood urea nitrogen 95 mg/dl), acidic medium (pH 7.1) with 2% normal serum, or acidic medium (pH 7.1) with 2% uremic serum for 48 h. Total cellular protein was isolated and assayed for Bnip3, Beclin-1, LC3, and Atg12/5 by Western blotting. The bar graphs (bottom) compare the protein band densities in each treatment group expressed as a fold change from levels in control myotubes (represented by a line at onefold). All band densities were normalized to the density of GAPDH: uremic serum, gray; acidification, white; and uremic serum plus acidification, black (values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. uremic serum). B: autophagy formation was detected in cultured muscle cells by staining with monodansylcadaverine (MDC), a fluorescent marker for lysosomal/autophagic vacuoles. C2C12 myoblasts were treated with 2% normal mouse serum, 2% uremic mouse serum, acidic medium, or acidic medium with 2% uremic serum for 24 h. Representative images are shown from each group (green indicates MDC, blue indicates nuclei). The intensity of MDC was analyzed by CellSens Dimension 1.9 software. The bar graph compares the intensities from each group normalized to 2% mouse serum treatment (represented by a line at onefold). Values are means ± SE, n = 6, *P < 0.05 vs. control, #P < 0.05 vs. acidification; scale bar = 50 μm).

Myofibrillar proteins are substrates of the UPS rather than autophagy.

CKD-induced loss of muscle is closely linked to the degradation of myofibrillar proteins (5, 8, 41). Because our in vivo and in vitro results indicate that CKD can increase the formation of autophagosomes and the degradation of their “cargos” by lysosomes, we assessed whether evidence for autophagy would be coupled to an increase in the degradation of myofibrillar proteins. Myotubes derived from satellite cells were examined to evaluate changes in myosin heavy chain protein and the cleavage of actin. Our previous studies showed that an initial step in the degradation of myofibrillar proteins is activation of caspase-3 because it cleaves the myofibrillar proteins to provide substrates for the UPS (8). To determine whether caspase-3 is also involved in increasing autophagy we treated satellite myotubes with 25 nM staurosporine (activation of caspase-3) and compared the results with those found in cells treated with either 0.5 μM epoxomicin (to inhibit proteolysis in the UPS) (13, 38) or 5 mM 3-methyladenine (to inhibit autophagy). In addition, 40 μM tamoxifen was used to activate autophagy.

First, we examined changes in the level of the14-kDa actin breakdown fragment, which results from the cleavage of actin to peptides (Fig. 4A). As expected, activation of caspase-3 in the myotubes with staurosporine in combination with the UPS inhibitor epoxomicin resulted in a raised level of the 14-kDa actin fragment. Increased actin fragment is not observed with staurosporine alone due to efficient degradation by the proteasome. Actin cleavage was not increased in cells that were treated with caspase-3 plus the inhibitor of autophagy. Activation of autophagy with tamoxifen did not result in increased actin fragment. The results confirm that the initial degradation of the myofibrillar protein, actin, proceeds predominately by activation of caspase-3 and the UPS rather than autophagy.

Fig. 4.

Myofibrillar protein is a substrate of UPS, but not autophagy in myotubes from primary cultured satellite cells. A: myotubes formed from primary cultured satellite cells were treated with either 25 nM staurosporine (to initiate myofibril break down), 40 µM tamoxifen (to activate autophagy), 0.5 µM epoxomicin (proteasome inhibitor), 5 mM 3-methyladenine (autophagy inhibitor), or nothing (untreated control). The protein was assayed for actin cleavage by Western blotting using an anti-actin antibody against the COOH-terminal actin peptide. The bar graph (bottom) compares the 14-kDa peptide breakdown product band densities to the intact 42-kDa actin protein in each treatment group. Numbers are expressed as a fold change from levels in control myotubes (represented by a line at onefold). Values are means ± SE, n = 9, *P < 0.05 vs. control cells. B: myotubes formed from primary cultured satellite cells were treated with different doses of tamoxifen to activate autophagy for 24 h. The protein was assayed for LC3, myosin heavy chain 2a (MyHC2a), myosin light chain 2 (myLC2), and myosin heavy chain 2x (MyHC2x) by Western blotting. The bar graphs compare the protein band densities in each treatment group expressed as a fold change from levels in controls (represented by a line at onefold). All band densities were normalized to the density of GAPDH (values are means ± SE, n = 6, *P < 0.05 vs. control). C: cultured myotubes were treated with serum starvation for the indicated times. In separate groups, myotubes were treated with 0.5 µM proteasome inhibitor epoxomicin (myHC2a+Epo) or 5 mM autophagy inhibitor 3-methyladenine (myHC2a+Methy). LC3, TRIM63/MuRF1, and MyHC2a were assessed with Western blotting. The bar graph compares the protein band densities in each treatment group expressed as a fold change from levels in controls (represented by a line at onefold). All band densities were normalized to the density of GAPDH (values are means ± SE, n = 6, *P < 0.05 vs. control).

Second, we treated myotubes with different concentrations of tamoxifen to chemically induce autophagy and measured myofibrillar breakdown by observing changes in the levels of myosin heavy chain 2a (MyHC2a), myosin heavy chain 2x (MyHC2x) and myosin light chain 2 (MyLC2). There were no increases in the breakdown of these myosin proteins despite activation of autophagy, as evidenced by increased LC3 (Fig. 4B). Third, we cultured myotubes in serum-free media to produce an endogenous activation of autophagy as signaled by an increase in expression of the LC3 protein (Fig. 4C). This treatment also activates the UPS as evidenced by the increase in TRIM63/MuRF1 levels. After 8 h of serum starvation there were significant decreases in MyHC2a protein levels. There was no significant inhibition of this decrease in MyHC2a when autophagy was blocked by 5 mM 3-methyladenine; however, this decrease was blocked by adding epoxomicin to inhibit protein degradation in the UPS. The MyHC2a results provide additional evidence that the decrease in the myofibrillar proteins results from UPS-mediated proteolysis.

Activation of autophagy by uremic serum damages mitochondrial function in cultured myotubes.

To further identify the role of autophagy in muscle during CKD, we treated cultured myotubes with uremic serum and measured mitochondria copy number in isolated DNA from uremic serum-treated cultured myotubes with or without autophagy inhibitors. NADH-ubiquinone oxidoreductase chain 1 was used for quantification of mitochondrial DNA (mtDNA) copy numbers (12). The nuclear gene of lipoprotein lipase (LPL) was used to normalize results. The mitochondria copy numbers are decreased 18% by uremic serum vs. normal mouse serum. Inhibition of autophagy using 80 μM chloroquine (Sigma) or 5 mM 3-methyladenine (InvivoGen, San Diego, CA) partially blocked this decrease (Fig. 5A). We measured both mitochondrial transcription factor A (TFAM) and peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) to identify mitochondrial biogenesis. TFAM is a key activator of mitochondrial transcription as well as a participant in mitochondrial genome replication (9). PGC-1α is a master regulator of mitochondrial biogenesis (35). Treatment of the myotubes with uremic serum resulted in a 39% decrease in TFAM mRNA and a 31% decrease in PGC-1α, and inhibition of autophagy reversed this decline (Fig. 5, B and C). Mitochondrial fusion protein Mitofusin-2 (Mfn2) is a mitochondrial membrane protein that participates in mitochondrial fusion and contributes to the maintenance and operation of the mitochondrial network (47). Mfn2 protein was also decreased by treatment of the cells with serum from uremic mice (Fig. 5D). An additional indicator of mitochondrial function is ATP production. ATP production decreased 50% in myotubes treated with uremic serum compared with those treated with normal mouse serum. When supplemented with an inhibitor of autophagy, 3-methyladenine, uremic serum treatment resulted in a 21% decrease in ATP production compared with normal control serum (Fig. 5E). ATP production was restored by both autophagy inhibitors chloroquine (1.38-fold) and methyladenine (1.55-fold) compared with uremic serum alone. Together, these data suggest that uremic serum-induced mitochondrial dysfunction may involve activation of autophagy.

Fig. 5.

Activation of autophagy by uremic serum damages mitochondria function in cultured myotubes. Myotubes were treated with pooled serum from uremic mice (Ure) for 48 h or with two autophagy inhibitors, 5 mM 3-methyladenine (Ure/Meth, 48 h) and 80 µM chloroquine (Ure/Ch, added 4 h before harvest of the myotubes). A: DNA was isolated and assayed for mitochondria copy numbers by quantitative PCR (qPCR). The copy number of mtND1 gene were used to identify mitochondrial DNA, and lipoprotein lipase (LPL) was used as a nuclei DNA control. The bar graph compares the copy numbers in each treatment group expressed as a fold change from levels in control myotubes (represented by a line at onefold). Results are normalized to LPL copy numbers (values are means ± SE, n = 12, *P < 0.05 vs. control, #P < 0.05 vs. uremic serum). B and C: mitochondrial-related mRNAs mitochondrial transcription factor A (TFAM) (B), and peroxisome proliferator-activated receptor gamma coactivator-1α (PGC1α) (C) were measured by real-time qPCR in the RNA from cultured myotubes of various treatment groups. The bar graph shows mRNA from the different treatment groups expressed as a fold change from the control myotubes (represented by a line at onefold). Results are normalized to 18S RNA (values are means ± SE, n = 12, *P < 0.05 vs. control, #P < 0.05 vs. uremic serum). D: mitochondrial related proteins TFAM and Mitofusin-2 (Mfn2) were measured by Western blotting in the different treatment group lysates from cultured myotubes. The bar graph compares the densities of protein bands in each group expressed as a fold change from levels in control myotubes, which is represented by a line at onefold. All band densities were normalized to the density of GAPDH. Values are means ± SE, n = 12, *P < 0.05 vs. control, #P < 0.05 vs. uremic serum. E: ATP production was measured by fluorometric assay in the myotubes of various treatment groups. The bar graph compares the ATP levels in nmol/mg protein in each group (values are means ± SE, n = 9, *P < 0.05 vs. control, #P < 0.05 vs. uremic serum).

Muscle overloading and Acu/LFES can improve mitochondrial biogenesis in CKD-induced muscle wasting.

To identify whether similar mitochondrial changes occur in vivo we measured the mitochondria copy number in muscle of CKD mice. The mitochondria copy number was decreased 9% in the muscle of CKD mice vs. mice in the sham-operated groups. Acu/LFES increased mitochondria content by 18% and OL by 15% compared with the copy number in muscle from untreated CKD mice (Fig. 6A). Messenger RNA levels of TFAM decreased 26% and PGC-1α decreased 50% in the muscle of CKD mice compared with normal mice, and Acu/LFES reversed this decline (Fig. 6B). Skeletal muscle OL not only reversed the CKD-induced decrease in PGC-1α, but it increased it above control levels. Furthermore, protein levels of TFAM and Mfn2 were reduced in CKD muscle and these lower levels were reversed by either Acu/LFES or OL (Fig. 6D).

Fig. 6.

Muscle overloading and Acu/LFES can improve mitochondrial biogenesis in CKD-induced muscle atrophy. A: DNA was isolated from hindlimb muscles of mice in sham, CKD, CKD plus Acu/LFES (CKD/Acu), and CKD plus muscle overloading (CKD/OL) groups and assayed for mitochondria copy numbers by qPCR. The copy number of mtND1 gene was used to identify mitochondria DNA, and LPL was used as a nuclear DNA control. The bar graph compares the copy numbers in each treatment group expressed as a fold change from levels in sham (represented by a line at onefold). Results are normalized to LPL. (Values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. CKD). B and C: mitochondrial-related mRNAs, TFAM (B) and PGC1α (C) were measured by real-time qPCR in the RNAs isolated from hindlimb muscles of mice in the sham and CKD, CKD plus Acu/LFES (CKD/Acu), and CKD plus muscle overloading (CKD/OL) groups. The bar graph shows mRNA from the different treatment groups expressed as a fold change from sham (represented by a line at onefold). Results are normalized to 18S RNA (values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. CKD). D: protein was isolated from hindlimb muscles of mice in sham and CKD, CKD plus Acu/LFES (CKD/Acu), and CKD plus muscle overloading (CKD/OL) groups, and mitochondrial-related proteins TFAM and Mfn2 were measured by Western blotting. The bar graph compares the densities of protein bands in each group expressed as a fold change from levels in the sham group, which are represented by a line at onefold. All band densities were normalized to the density of GAPDH. (Values are means ± SE, n = 9, *P < 0.05 vs. sham, #P < 0.05 vs. CKD.)

DISCUSSION

Muscle atrophy is commonly present in patients with CKD (2, 26, 40). We and others have shown that FoxO transcription factors play a very critical role in CKD-induced muscle wasting (17, 37). FoxO1 can activate E3 ubiquitin ligases TRIM63/MuRF1 and FBXO32/atrogin-1, leading to upregulation of the UPS pathway to catalyze the myofibrillar proteins (30, 39). Activated FoxO3 also stimulates lysosomal proteolysis in muscle by activating the autophagy process by transcriptional regulation (19, 46). In this study, we found that CKD muscle loss is associated with activation of autophagy. In mice with CKD, the levels of autophagy-related proteins including Bnip3, Beclin-1, PI3KC3, and Atg12/5 are increased. Likewise, the ratio of the autophagy proteins LC3II-to-LC3I, which is associated with autophagosome formation, is increased in muscle of mice with CKD. Protein p62 is responsible for tagging different “cargos” (i.e., substrates of the autophagosome). We showed that p62 is decreased in muscles of mice with CKD, confirming increased autophagosome-mediated degradation. Overall, our results demonstrate that the autophagy process is activated in the muscles of mice with CKD undergoing muscle catabolism.

The first major finding of this study is that activation of autophagy can be independent from loss of muscle mass in CKD. We have three lines of evidence. First, although there is activation of autophagy and muscle loss in CKD mice, when these mice were exercised (skeletal muscle OL), they maintained their increased autophagy but their CKD-related losses of muscle mass were reduced. Second, in cultured myotubes, acidification does not increase autophagy markers. However, in animal studies, we found that acidosis does contribute to muscle wasting (2). Third, we demonstrated that in cultured myotubes, myofibrillar protein degradation in response to serum starvation proceeds predominately via the UPS rather than through activation of autophagy (Fig. 4C).

Previous evidence suggests that autophagy could contribute to the loss of muscle proteins. For example, literature citations claim that in certain catabolic conditions autophagy degrades nonmyofibrillar cytosolic proteins and organelles (e.g., mitochondria) (5, 29). Several muscle catabolic conditions including cancer, fasting, sepsis, disuse, and denervation (23–25) have been found to activate autophagy. However, those reports did not demonstrate that myofibrillar proteins are the substrates of autophagosomes. In fact, evidence from autophagy-deficient (Atg7 knockout) mice showed that myosin composition was not affected by deficiency of Atg7 in muscle (20). In exploring whether autophagy results in degradation of myofibrillar proteins we used primary cultured myotubes derived from normal mouse muscles. Interestingly, in serum-starved myotubes, myosin was degraded along with autophagosome formation. Inhibition of autophagy did not block the breakdown of myosin. In contrast, in serum-starved myotubes, when we inhibited the UPS, the loss of myosin heavy chain proteins was prevented (Fig. 4). Thus, it appears the substrates for autophagy differ from those of the UPS. There is emerging evidence that the myofibrillar proteins (myosin, actin, tropomyosin, troponins) are substrates for the E3 ligases of the UPS (4, 6). Caspase-3 initially cleaves myofibrillar proteins leaving a 14-kDa peptide that is an index of muscle protein degradation in CKD (8, 41). In the current study, the 14-kDa actin fragments accumulated when the UPS was inhibited but not when autophagy was inhibited, indicating that actin breakdown is mediated by the UPS. There is no current evidence that myofibrillar proteins are substrates for autophagy-mediated degradation.

A second major finding of this study is that uremic toxicity, not acidification alone, activates autophagy in CKD. In our previous study, adding NaHCO₃ to the diet of rats with CKD reversed acidosis and prevented muscle proteolysis (2). In the present experiments, we found that acidified medium decreased the ratio of LC3II/I (Fig. 3A), which theoretically, would suggest that CKD-associated acidosis could inhibit autophagy. Conversely, we also found that acidosis increases Bnip3 and Beclin-1, indicating increased autophagy. We also showed that acidosis does not alter LC3 puncta, which suggests there is no effect on autophagy. We do not have a good explanation for these apparently disparate findings. However, it must be remembered that CKD causes a combination of biochemical changes, acidosis being only one. Acidosis studies are performed in cells in a purer form than an animal would experience in an effort to determine whether this aspect of uremia is responsible for the autophagy response. What we can say from the cell experiments is that acidosis is very unlikely to be the sole mediator of the autophagy changes that are observed in the uremic animal. In addition, uremia in combination with acidosis showed an increase in LC3 puncta, and an increase in Bnip3, Beclin-1, and Atg12/5 proteins. This suggests that although acidosis might depress some aspects of the autophagy pathway the overall effect of the disease state is to show increased autophagy. Nevertheless, completely characterizing the mechanism of acidosis-mediated changes in autophagy would be interesting in a future study.

The third major finding of this study is that deterioration of mitochondrial function and decreased ATP production in the muscle of CKD mice is linked to levels of autophagy. We found that Bnip3 is increased in CKD muscle and cultured muscle cells treated with uremic serum. Bnip3 is a mitochondrial proapoptotic protein that is localized at the mitochondrial outer membrane. It regulates the opening of a permeability transition pore in the mitochondrial double membrane and is involved in mitochondrial quality control (43). This process allows the translocation of lysosomal proteins from the cytoplasm into the mitochondrial matrix (36, 42). Bnip3 has two molecular forms, monomer and dimer. Only dimeric Bnip3 induces apoptosis, at least in part, by inhibiting the suppression of apoptosis by Bcl-2 (3). The COOH-terminal domain of Bnip3 influences dimer formation and directs its expression to mitochondria (3). In our recurrent study, the mitochondrial dysfunction could be related with CKD-induced upregulation of Bnip3.

It is well known that mitochondrial clearance is an important function of autophagy. A proper balance of the autophagic activity is essential for maintaining healthy mitochondria and skeletal muscle. Too much autophagy could cause excessive removal of cellular components such as mitochondria that are needed for normal activities. Insufficient autophagy could lead to accumulation of damaged or dysfunctional cell components causing structural and functional impairment, which could result in muscle weakness (10). In our study, upregulation of autophagy was accompanied by disruption of normal mitochondrial biogenesis, which could lead to decreased ATP production under uremic conditions (Fig. 5E). Our results strongly suggest that accelerated autophagy in CKD would contribute to a negative outcome. We also observed that mitochondrial biogenesis would not have been regulated by autophagy alone since inhibition of autophagy only partially reversed the CKD-induced decline in PGC-1α and ATP production. Some of the mitochondrial damage could be caused by CKD toxins. Unlike the autophagy responses in CKD, activation of autophagy by muscle OL might remove damaged mitochondria produced by exercise and plays a protective role against muscle wasting (Fig. 6D). This occurs in both the healthy muscle and in muscles from CKD animals.

Although we see a correlation of increased autophagy with decreased mitochondria biogenesis in CKD muscle, a limitation of this study is that this is not direct evidence proving that the elevation in autophagy is the cause of decreased mitochondrial biogenesis in vivo. It is possible that they have a similar cause and shared effects. This field needs further study.

Studies have shown that autophagy is acutely activated by exercise (11). In this present study, we found that activation of autophagy by exercise has a beneficial effect on the mitochondrial network; however, upregulation of autophagy by CKD impairs mitochondrial function. Acu/LFES is another method of preventing loss of muscle mass in catabolic diseases such as CKD, diabetes, and denervation injury (14, 32, 33). The mechanism involves an increase in IGF-1 production and an increase in p-Akt in muscle. Increasing p-Akt inhibits FoxO1 transcription factor, which suppresses protein degradation in CKD mice (14). Notably, variations of Acu/LFES have been given to patients with CKD and there is evidence that it prevents atrophy of skeletal muscles (15). In the current study, we demonstrated that the increase in autophagy markers and the loss of proteins in muscles of mice with CKD can, at least in part, be attenuated by treatment of CKD mice with Acu/LFES.

In conclusion, we find that autophagy is activated in the muscles of mice with CKD undergoing muscle catabolism. Uremic toxicity, not acidification, induces autophagosome formation in cultured muscle cells. However, the increase in autophagy is not directly responsible for the break down in myofibrillar protein. Notably, CKD-induced upregulation of autophagy leads to deterioration of mitochondria function and decrease of ATP production.

GRANTS

Support for this study was provided by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant R01 AR-060268 to X. Wang; and by National Natural Science Foundation of China Grant 30871179, Zhejiang Provincial Natural Science Foundation of China Grant LY15H270017, and Zhejiang Provincial Program for the Cultivation of High-level Innovative Health Talents to Z. Su.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.Z. and X.H.W. conceived and designed research; Z.S. and F.H. performed experiments; Z.S., J.D., M.B.H., and X.H.W. analyzed data; Z.S., H.A.F., and X.H.W. interpreted results of experiments; Z.S., F.H., and X.H.W. prepared figures; H.W. drafted manuscript; J.D.K., L.Z., and X.H.W. edited and revised manuscript; X.H.W. approved final version of manuscript.