AMPK Mediates Early Activation of the Unfolded Protein Response through a Positive Feedback Loop in Palmitate-Treated Muscle Cells

Background : Activation of the unfolded protein response (UPR) is closely related to the pathogenesis of many metabolic disorders. Accumulating evidence also shows that UPR and metabolic signaling pathways are interdependent. The AMP-activated protein kinase (AMPK) signal pathway controls the energy balance of eukaryotes. The aim of this study was therefore to investigate the possible interaction between AMPK signaling and UPR in muscle cells exposed to saturated fatty acids, as well as the potential mechanism. Methods : The saturated fatty acid palmitate was used to induce UPR in C2C12 myotubes. Compound C or knockdown of AMPK α with short hairpin RNA (shRNA) were used to inhibit the AMPK signaling pathway in palmitate-treated muscle cells. AMPK signaling in myotubes was activated using 5-amino-1-β -D-ribofuranosylimidazole-4-carboxamide (AICAR) or ex229. C2C12 myotubes were pre-treated with taurourdodeoxycholic acid (TUDCA) to inhibit UPR before adding palmitate. Real-time PCR and Western blotting were performed to evaluate the expression of UPR markers and activation of AMPK. Results : Palmitate treatment induced UPR in C2C12 myotubes while activating AMPK signaling. Inhibition of the AMPK pathway with compound C or AMPK shRNA reduced palmitate-induced activation of UPR, while inhibition of UPR with TUDCA reduced palmitate-induced AMPK activation. This indicates a positive feedback loop between UPR and AMPK. Furthermore, activation of the AMPK pathway with AICAR or ex229 caused a dose-dependent upregulation of UPR markers, including activating transcription factor 4 (ATF4), binding immunoglobulin protein (BIP), and growth arrest and DNA damage-inducible 34 (GADD34) protein. Conclusions : These results provide the first evidence that AMPK signaling is involved in the early activation of UPR caused by saturated fatty acids in skeletal muscle. Furthermore, they indicate that physiological or pharmacological activation of the AMPK pathway (e.g., by exercise or phenformin, respectively) can promote muscle health and function, thereby improving the quality of life in individuals with metabolic disorders due to a high-fat diet or obesity.


Introduction
Adult skeletal muscle shows considerable plasticity that allows it to respond rapidly under a variety of physiologic and pathologic conditions [1]. This is facilitated by the sarcoplasmic reticulum, a specialized form of the endoplasmic reticulum (ER) [2]. Environmental or cellintrinsic stimuli such as nutrient or oxygen deprivation, exposure to toxic substances, and oxidative stress can disrupt cellular homeostasis and induce ER stress, thereby activating the unfolded protein response (UPR) [3][4][5]. The canonical UPR in mammals is initiated by activation of three major ER transmembrane sensors: PKR-like endoplasmic reticulum kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE1) [6][7][8]. These trigger the expression of downstream transcription factors including ATF4, ATF6c, C/EBP homologous protein (CHOP), and spliced X-box binding protein 1 (XBP1s). The main outcome of UPR signaling, particularly in the early phase, is the restoration of ER homeostasis through inhibition of protein synthesis or upregulation of ER chaperone proteins [9,10]. However, prolonged UPR due to continuous stress can lead to the induction of apoptotic cell death [11,12]. Thus, the UPR is a cellular mechanism that controls cell fate in response to stress.
ER stress and the UPR can also be activated in skeletal muscles that are exposed to metabolic stress, as occurs in diabetic patients [13], or by the consumption of a highfat diet [14,15]. The high concentration of free fatty acids and especially of saturated fatty acids (SFAs) in plasma under these conditions is one of the main triggers for UPR in skeletal muscle [16][17][18]. The UPR is closely associated with SFA-induced inflammation, insulin resistance, and apoptosis in skeletal muscle [18][19][20], indicating there is crosstalk between the UPR and signaling pathways that regulate metabolism [21,22]. The AMP-activated protein kinase (AMPK) pathway is conserved across eukaryotes and integrates signals from multiple sources to control cellular energy balance [23]. Alterations to AMPK also contribute to the metabolic adaptations and progress of insulin resistance in muscles exposed to SFAs [24]. Given their critical influence on muscle health and metabolism, elucidating the interplay and synergisms (or antagomisms) between AMPK and UPR is prerequisite for the development of novel therapeutics or strategies to ameliorate muscle metabolic disorders. Although there is increasing evidence for interactions between AMPK signaling and the UPR [25][26][27][28][29][30][31][32], the mechanistic basis for crosstalk between these two pathways has yet to be elucidated in different models of ER stress.
Therefore, in the present study we investigated whether there is crosstalk between AMPK signaling and the UPR following induction by palmitate in skeletal muscle cells, as well as the possible underlying mechanism. We found that AMPK was activated in myotubes in response to palmitate treatment. Moreover, we showed that AMPK signaling was associated with early activation of the UPR via a positive feedback mechanism. Additionally, pharmacologic activation of AMPK led to the induction of UPR. These findings provide novel insights into the interactions between metabolic signals and homeostatic mechanisms in skeletal muscle cells that may be perturbed in metabolic disorders.

C2C12 myoblast cells (cat.
no. 1101MOU-PUMC000099) were purchased from National Infrastructure of Cell Line Resource (NICR) (Beijing, China). This cell line has been authenticated by NICR with flow cytometry. Venor ® Gem OneStep Kit (Minerva Biolabs ® , Berlin, Germany; cat. no.11-8025) was used to confirm that there is no mycoplasma contamination in C2C12 cells.

RNA Interference
Adenoviral constructs containing short hairpin RNA (shRNA) against AMPKα1 (PRKAA1, NM_001013367) or scrambled shRNA (negative control) were constructed by Genechem (Shanghai, China). C2C12 myotubes were transfected with the adenoviral constructs (multiplicity of infection of 100) for 12 h, and then treated with palmitate or left untreated as described above.

RNA Extraction and Real Time-PCR (RT-PCR)
Total RNA was extracted from C2C12 myotubes using TRIzol reagent (Invitrogen, Carlsbad, CA, USA; cat. no. 15596-026) and then 500 ng of total RNA was reverse transcribed into cDNA using the PrimeScript RT reagent kit (Takara Bio, Otsu, Japan; cat. no. RR037A) according to manufacturer's instructions. RT-PCR was performed using a StepOnePlus RT-PCR system (Invitrogen) with fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA; cat. no. 4385612). Each RT-PCR mixture contained 21 µL sterile water, 25 µL SYBR Green, 2 µL cDNA (500 ng/µL), and 1 µL each of forward and primers (10 pmol/µL). The reaction was performed by an initial denaturation step at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 10 s, annealing at the melting temperature of the specific primer set for 10 s, elongation at 72°C for 15 s, and concluded with a melting curve step. Target gene expression levels were normalized to those of the 18S rRNA gene. The sequences of the primers used are listed in Supplementary Table 1.

Immunoblotting
C2C12 myotubes were lysed in RIPA buffer containing 20 mM Tris·HCl (pH 7.4), 150 mM NaCl, 2.5 mM EDTA, 1% Triton X-100, 10% glycerol, 0.1% Sodium dodecyl sulfate (SDS), 1% deoxycholate and 1% protease inhibitor cocktail (Roche, Basel, Switzerland; cat. no. 04693132001). The supernatant was collected by centrifugation at 12,000 ×g for 10 min at 4°C, and the protein concentration was determined using Bradford protein assay reagents (Bio-Rad, Hercules, CA, USA; cat. no.500-0203). Equal amounts of extracted proteins (30 µg per sample) were denatured in SDS loading buffer and separated on SDS-polyacrylamide gels. The proteins were then transferred to a nitrocellulose membrane, which was blocked for 2 h in 5% non-fat milk or bovine serum albumin diluted in TBST. The membrane was then incubated overnight at 4°C with primary antibodies against CHOP (cat. no. 5554), ATF4 (cat. no. 11815), AMPKα (cat. no. 2532), and p-AMPKα (cat. no. 2531) (all from Cell Signaling Technology, Danvers, MA, USA), and β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA; cat. no. sc130656). The next day, the membrane was washed 3 times with TBST and incubated for 2 h at room temperature with secondary antibodies in 5% nonfat milk, followed by incubation with enhanced chemiluminescence reagent (Thermo Fisher Sci-entific; cat. no. 34580) in a dark room. Protein bands were quantified using Image-Pro Plus v6.0 software (Media Cybernetics, Rockville, MD, USA) and the densitometry of the protein signal was normalized to that of β-actin.

Statistical Analysis
Data are presented as mean ± SD. One-way or twoway analysis of variance followed by the Tukey's post hoc test was used to compare the means of multiple groups using Prism version 7.0 software (GraphPad, La Jolla, CA, USA). p ≤ 0.05 was considered statistically significant.

AMPK Signaling is Activated in the Early Stages of the UPR in Myotubes
We investigated the AMPK phosphorylation status and the expression of UPR markers in C2C12 myotubes treated with palmitate (a major component of dietary saturated fats) for different times. While the total AMPKα levels remained constant over time, AMPKα phosphorylation increased after 3 h of palmitate treatment and reached a peak after 12 h. The expression of AMPK phosphorylation after 12 h of palmitate treatment showed a significant difference compared to any other groups. However, after 24 h the p-AMPKα level was lower than in the control group (Fig. 1A). The gene expression level of UPR markers such as CHOP, ATF4, and XBP1s was also upregulated after 3 h of palmitate treatment and peaked at 12 h (Fig. 1B). These results indicate that activation of AMPK signaling induced by palmitate in myotubes is closely associated with the early activation of UPR.

AMPK Signaling is Involved in Palmitate-Induced UPR in Myotubes
To further investigate the interaction between the AMPK pathway and UPR, C2C12 myotubes were treated with palmitate for 12 h with or without compound C, a widely used specific inhibitor of AMPK. As expected, palmitate caused activation of AMPK, as well as increased expression of the ATF4 and CHOP proteins ( Fig. 2A). Compound C completely abolished the AMPK activation induced by palmitate ( Fig. 2A) and reduced the palmitateinduced upregulation of ATF4 and CHOP protein levels ( Fig. 2A). In agreement with the above findings, palmitate induced the upregulation of gene expression for multiple components of UPR, including ATF4, CHOP, GADD34, chaperone BIP, XBP1u and XBP1s (Fig. 2B). This upregulation was also attenuated by treatment with compound C (Fig. 2B).
To more specifically inhibit the AMPK signaling pathway, C2C12 myotubes were infected with adenovirus that expressed AMPKα shRNA. This confirmed the reduction of both total and phosphorylated AMPKα levels in C2C12 myotubes with or without palmitate treatment (Fig. 3). Moreover, the palmitate-induced upregulation of ATF4 and CHOP was also significantly blocked by knocking down AMPKα (Fig. 3). These data demonstrate that AMPK signaling contributes to early activation of the UPR induced by palmitate. Fig. 3. AMPKα knockdown attenuates palmitate-induced UPR in C2C12 myotubes. C2C12 cells were infected either with control shRNA or AMPKα shRNA for 12 h and then treated with or without 0.5 mM palmitate for a further 12 h. Total protein was then extracted and the protein levels for AMPKα, p-AMPKα, CHOP and ATF4 were evaluated by Western blot analysis (n = 6). All data are shown as the mean ± SD. *p < 0.05, ***p < 0.001 vs Con shRNA group; ##p < 0.01, ###p < 0.001 vs Con shRNA+Palmitate group (two-way analysis of variance).

Inhibition of the UPR with TUDCA Attenuates Palmitate-Induced AMPK Activation
We further investigated whether inhibition of UPR alters AMPK activation in C2C12 myotubes. Myotubes were pretreated with the UPR inhibitor TUDCA for 1 h before the addition of palmitate for 12 h. TUDCA significantly attenuated the palmitate-induced upregulation of ATF4 and CHOP (Fig. 4), thereby demonstrating pharmacologic inhibition of the UPR. Interestingly, TUDCA also abolished palmitate-induced AMPKα phosphorylation (Fig. 4), suggesting a positive feedback loop between the UPR and AMPK pathway in the early stage of palmitate treatment in muscle cells. Data are shown as the mean ± SD. ***p < 0.001 vs control group; ###p < 0.001 vs palmitate group (two-way analysis of variance).

Pharmacologic Activation of AMPK Signaling Induces UPR
Given the finding that AMPK activation contributes to palmitate-induced UPR, we speculated that pharmacologic activation of AMPK would be sufficient to induce the UPR in C2C12 myotubes. To test this hypothesis, C2C12 myotubes were treated with the AMPK agonist AICAR at concentrations ranging from 0.125-2 mM for 12 h to activate AMPK signaling. AMPK phosphorylation increased with increasing AICAR concentration, and was highest at 1 mM AICAR (Fig. 5A). Moreover, ATF4 and CHOP protein levels were upregulated by AICAR in a dose-dependent manner at concentrations <1 mM (Fig. 5A). A higher concentration of AICAR (2 mM) failed to induce AMPK activation and ATF4 and CHOP expression to a greater extent than 1 mM AICAR (Fig. 5A). AICAR was also found to induce gene expression of UPR markers such as BIP, ATF4, and GADD34 in a dose-and time-dependent manner (Supplementary Figs. 1,2). Ex229 is another allosteric activator of the AMPK pathway that is more potent than AICAR [34]. Indeed, ex229 treatment was observed to activate AMPKα phosphorylation at much lower concentrations than AICAR (Fig. 5B). As expected, the protein expression of ATF4 and CHOP was also upregulated by ex229 in a dose-dependent manner (Fig. 5B). Taken together, these results indicate that pharmacologic activation of AMPK is sufficient to induce the UPR in C2C12 myotubes.

Discussion
The results of this study provide novel evidence of the interaction between the AMPK pathway and UPR in muscle cells exposed to palmitate, a major component of dietary saturated fats [35]. Specifically, we first observed the unexpected activation of AMPK signaling within 12 h of palmitate treatment, which was accompanied by acute induction of the UPR. In support of these findings, a previous re-port showed that peroxisome proliferator-activated receptor gamma coactivator (PGC)-1α, one of the main downstream target genes of the AMPK pathway, was transiently upregulated after 4 and 8 h of palmitate treatment [36]. However, another study found that the AMPK pathway was inhibited in cells treated with palmitate for 16 h [18]. We speculate this discrepancy may be due to differences in the duration of treatment, especially since the changes induced by palmitate in vitro are stronger and occur more rapidly than those observed during clinical obesity or after induction by a high-fat diet. Indeed, we also found that p-AMPKα was downregulated in cells exposed to palmitate for 24 h.
The association between the UPR and AMPK has been investigated previously [27,[37][38][39][40]. Several studies on palmitate-induced ER stress have demonstrated an inhibitory effect of AMPK signaling on the UPR in different tissues and cells [27,37,40]. For example, pharmacologic activation of AMPK with AICAR was shown to suppress palmitate-induced ER stress in rat vascular endothelial cells [29]. In C2C12 myotubes, both GW501516 Fig. 6. Bidirectional crosstalk between AMPK signaling and unfolded protein response (UPR) in muscle cells exposed to saturated fatty acids (SFAs). During the early stages after palmitate treatment of muscle cells, AMPK signaling and UPR are activated and crosstalk through a positive feedback mechanism, thereby facilitating restoration of ER homeostasis. Physiological (e.g., exercise) or pharmacological (e.g., phenformin) interventions can accelerate the restoration of ER homeostasis and muscle health by promoting bidirectional crosstalk between AMPK signaling and UPR during the early stages of regular high-fat diet. ER, endoplasmic reticulum.
(a peroxisome proliferator-activated receptor [PPAR]δ receptor agonist) and oleate inhibited palmitate-induced ER stress through an AMPK-dependent mechanism [18,41]. Similarly, 5-lipoxygenase protected C2C12 myotubes from palmitate-induced ER stress via the activation of AMPK [42]. AMPK activation with AICAR (2 mM) effectively reduced palmitate-induced ER stress in muscle cells [18]. However, our results showed that inhibiting the AMPK pathway in myotubes attenuated the UPR induced by palmitate, thus demonstrating a stimulatory effect of the AMPK pathway on palmitate-induced UPR. In agreement with our findings, the antidiabetic drug phenformin was shown to activate ER stress in an AMPK-dependent manner, while AMPK deficiency completely abolished phenformininduced UPR [43]. Similarly, it was reported that AMPK activation induced mild UPR in C3H10T1/2 mouse mesenchymal stem cells [31]. Moreover, we demonstrated that inhibiting the UPR resulted in mitigation of palmitateinduced AMPK activation, indicating the existence of a positive feedback loop between AMPK and the UPR in the early stages of palmitate treatment in muscle cells. To our knowledge, this is the first report of a positive feedback regulatory mechanism between the AMPK pathway and UPR.
Interestingly, we also found that pharmacologic activation of AMPK with AICAR or ex229 was sufficient to induce upregulation of UPR components in myotubes. In line with this finding, PGC-1α was shown to induce the expression of several UPR-related genes in skeletal muscle [32]. Moreover, ER stress markers (e.g., ATF3 and CHOP) and chaperones (e.g., BIP and GRP94) were significantly unregulated in the gastrocnemius muscle of transgenic mice with muscle-specific overexpression of PGC-1α [32]. PGC-1α overexpression also induced the expression of genes related to protein folding and the UPR in primary myotubes [32]. The increased expression of BIP and GADD34 caused by exercise was abolished in musclespecific PGC-1α knockout mice, demonstrating that PGC-1α is important for the UPR in skeletal muscle [32]. Given the essential role of PGC-1α as an effector of the AMPK signaling pathway and of its upregulation shortly after palmitate treatment [36], we speculate that PGC-1α is involved in the early activation of palmitate-induced ER stress in skeletal muscle.

Conclusions
In summary, we have provided evidence of bidirectional crosstalk between AMPK signaling and early activation of the UPR in muscle cells exposed to SFAs (Fig. 6). We also showed that pharmacologic activation of AMPK was sufficient to induce mild UPR in skeletal muscle cells.
These findings demonstrate an essential role for the AMPK pathway in restoring ER homeostasis via activation of the UPR in response to metabolic stress. Furthermore, they may guide the development of new strategies for the treatment of diseases such as obesity and diabetes through improvements in skeletal muscle metabolism.

Availability of Data and Materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author Contributions
PZ and XC designed the research study. JG, LW, WT, ZL, XP, and WL performed the research. PZ, JG, WL, and SL analyzed the data. PZ, JG, LW, and XC wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

Ethics Approval and Consent to Participate
Not applicable.

Acknowledgment
Not applicable.

Funding
This work was supported by grants from the National Natural Science Foundation of China (81871522, 32171173) and the State Key Laboratory Grant of Space Medicine Fundamentals and Application (SMFA18B01, SMFA20A02).