Icaritin requires Phosphatidylinositol 3 kinase (PI3K)/Akt signaling to counteract skeletal muscle atrophy following mechanical unloading.

Counteracting muscle atrophy induced by mechanical unloading/inactivity is of great clinical need and challenge. A therapeutic agent that could counteract muscle atrophy following mechanical unloading in safety is desired. This study showed that natural product Icaritin (ICT) could increase the phosphorylation level of Phosphatidylinositol 3 kinase (PI3K) at p110 catalytic subunit and promote PI3K/Akt signaling markers in C2C12 cells. This study further showed that the high dose ICT treatment could significantly attenuate the decreases in the phosphorylation level of PI3K at p110 catalytic subunit and its downstream markers related to protein synthesis, and inhibit the increases in protein degradation markers at mRNA and protein levels in rat soleus muscle following 28-day hindlimb unloading. In addition, the decreases in soleus muscle mass, muscle fiber cross-sectional area, twitch force, specific force, contraction time and half relaxation time could be significantly attenuated by the high dose ICT treatment. The low dose ICT treatment could moderately attenuate the above changes induced by unloading. Wortmannin, a specific inhibitor of PI3K at p110 catalytic subunit, could abolish the above effects of ICT in vitro and in vivo, indicating that PI3K/Akt signaling could be required by ICT to counteract skeletal muscle atrophy following mechanical unloading.


Results
Effect of ICT on PI3K/Akt signaling in C2C12 cells. The C2C12 myoblasts were cultured and treated with 5 μ M, 10 μ M and 20 μ M ICT for 24 hours. The phosphorylated and total protein levels of PI3K at p110 catalytic subunit (PI3K-p110) and its downstream proteins, including Akt, the mammalian target of rapamycin kinase (mTOR), ribosomal protein S6 kinase (p70S6K) and eukaryotic translation initiation factor 4E binding protein 1 (4EBP1) 32 , were detected by Western blotting. The distribution of forkhead box O (FOXO) family transcription factors (FOXO1 and FOXO3a) in the nuclear fraction and total cell were also determined by western blotting. The mRNA levels of muscle atrophy F-box protein (Atrogin-1) and muscle RING finger protein 1 (MuRF-1) were examined by real-time PCR. Comparing to the Control Group, the phosphorylation level of PI3K-p110 in the C2C12 cells incubated with 5 μ M, 10 μ M and 20 μ M ICT was 1.2-fold, 2.3-fold and 4.5-fold higher, respectively. Similar patterns were also observed in the phosphorylation levels of the downstream proteins of PI3K (Akt, mTOR, p70S6K and 4EBP1) (Fig. 1B). On the other hand, the expression levels of FOXO1 and FOXO3a in nuclei fraction in the ICT treatment groups (5-20 μ M) were 10-35% lower than those in the Control Group in a dose-dependent manner (Fig. 1C). Similar data were also found in the mRNA levels of Atrogin-1 and MuRF-1 (Fig. 1D). In addition, the C2C12 cells were cultured and treated with 20 μ M ICT for 1, 3, 5 and 7 days. Cell viability, morphology and myogenic markers (MyoD and myogenin) levels were determined by cell counting kit-8 (CCK-8) assay, cell morphology examination and Western blotting. The differentiation rate of C2C12 cells with ICT treatment was significantly higher than vehicle control (Supplementary Figure 1).

Involvement of PI3K-p110 catalytic subunit for effect of ICT in C2C12 cells.
To further determine whether PI3K-p110 catalytic subunit is required for the ICT-induced activation of PI3K/Akt signaling in C2C12 cells, a specific inhibitor of PI3K at p110 catalytic subunit wortmannin (Wort, 5nM) was pre-incubated with C2C12 cells for 1 hour followed by the treatment with ICT (20 μ M) (ICT+ Wort Group) or routine culture medium (Wort Group) for further 24 hours. Cells only treated with ICT (20 μ M) or routine culture medium served as the ICT Group or the Control Group, respectively. The phosphorylation levels of PI3K-p110 and its downstream proteins related to protein synthesis in the Wort Group were significantly lower than those in the Control Group. The phosphorylation levels of the proteins in the ICT+ Wort Group were significantly lower than those in the ICT Group, but a little higher than those in the Wort Group (P > 0.05). No significant differences were found between the ICT+ Wort Group and the Control Group ( Fig. 2A). Significantly higher expression levels of FOXO1 and FOXO3a in nuclei fraction were present in the Wort Group compared to the Control Group. The expression levels of FOXO1 and FOXO3a in nuclei fraction were significantly lowered in the ICT Group compared to the Control Group, which was obviously blocked in the ICT+ Wort Group (P < 0.05). There were no significant differences between the ICT+ Wort and the Control Group (Fig. 2B). Similar data were also found in the mRNA levels of Atrogin-1 and MuRF-1 (Fig. 2C).

Effect of ICT treatment on muscle weight and muscle fiber cross-sectional area (CSA).
To investigate the effect of ICT treatment on unloading-induced skeletal muscle atrophy, two oral dosages of ICT (80 or 120mg/kg/day) were administered daily to 3-month old male Sprague-Dawley (SD) rats with or without daily injection of wortmannin (15 μ g/kg/day) during 28-day hindlimb suspension (HS). The animals were assigned to eight groups including Baseline, Age-mateched Control, HS, Low-ICT, High-ICT, Wort, Low-ICT+ Wort, and High-ICT+ Wort groups as described in Materials & Methods. The muscle-to-body weight (mw/bw) ratio of soleus in the Age-matched Control Group was not changed when compared to the Baseline Group. After 28-day HS, the ratio showed a significant decrease from the baseline in the HS Group. The ratio in the Low-ICT Group was a little higher than that in the HS Group (P > 0.05), whereas the ratio was significantly higher in the High-ICT Group than that in the HS Group. However, the ratio in the High-ICT Group was still significantly lower than that in the Baseline Group. In addition, the ratio in the Wort Group was not significantly different from that in the HS Group. Interestingly, the ratio was significantly lower in the High-ICT+ Wort Group than that in the High-ICT Group (Fig. 3A). Similar pattern was also found in the results of muscle fiber CSA (Fig. 3B).
There was no significant difference observed in the data of blood biochemical evaluation and histological analysis of the major organs among the HS, Low-ICT and High-ICT Groups (Supplementary Figure 2).

Effect of ICT treatment on muscle fiber type distribution.
To identify the muscle fiber type distribution, the muscle myosin heavy chain (MHC) type I (slow-twitch type fibers), IIa (fast-twitch type fibers) and IIb (faster-twitch type fibers) were detected by immunohistochemistry staining. As shown in Fig. 4A, the majority of the muscle fiber type in soleus muscle was MHC I fiber. MHC IIb fiber was not detected in soleus muscle while hybrid MHC I/IIa fibers were observed in all groups. There was no significant difference in the proportion of the MHC I fiber between the Baseline Group and the Age-matched Control Group. After 28-day HS, the proportion of the MHC I fiber showed a significant decrease in the HS Group compared to the Baseline Group, while the proportion of the MHC IIa fiber increased instead when compared to the Baseline Group. The proportion of the MHC I fiber in the High-ICT Group was significantly higher than that in the HS Group, which was blocked by the wortmannin treatment (P < 0.05). The proportion of MHC I fiber in the Low-ICT Group was moderately higher than that in the HS Group (P > 0.05). There were no significant differences observed among three wortmannin-treated groups and the HS Group (Fig. 4B).
Effect of ICT treatment on muscle function. The influence of ICT treatment on muscle functional properties was evaluated with in vitro muscle functional testing. The twitch force and specific force in the HS Group were significantly lower than those in the Baseline Group. Both parameters were moderately higher in the Low-ICT Group (P > 0.05), but significantly higher in the High-ICT Group compared to the HS Group. The twitch force in the High-ICT was still lower than that in the Baseline Group, whereas the specific force in the High-ICT Group was not significantly different from that in the Baseline Group. The contraction time and half relaxation time were shorter in the HS Group than those in the Baseline Group, while both parameters were prolonged in the High-ICT Group, but still shorter compared to the Baseline Group. Wortmannin treatment led to a moderately lower twitch force and specific force (P > 0.05) and significantly shorter contraction time and half relaxation time compared to the High-ICT Group. All the parameters in the three wortmannin treatment groups were not significantly different from those in the HS Group (Fig. 5).
Effect of ICT treatment on protein synthesis markers. The influence of ICT treatment on the phosphorylation state (activation) of PI3K-110 and its downstream proteins (Akt, mTOR, p70S6K, 4EBP1) that contribute to the protein synthesis in soleus muscle was determined by western blotting. Compared to the Baseline Group, the phosphorylation levels of above markers were significantly lower in the HS Group. The phosphorylation levels of the proteins in the Low-ICT Group were a little higher than those in the HS Group (P > 0.05). The High-ICT Group demonstrated significantly higher phosphorylation levels than those in the HS Group, which were significantly attenuated by the wortmannin treatment. The phosphorylation levels of those proteins in the three wortmannin-treated groups were not statistically different from those in the HS Group (Fig. 6).
Effect of ICT treatment on protein degradation markers. FOXO1 and FOXO3a would be dephosphorylated and enter into the nucleus when Akt was inactivated, then activate atrophy-induced genes, such as muscle-specific E3 ubiquitin ligases Atrogin-1 and MuRF-1 and trigger protein degradation 19 . The cellular distribution of FOXO1/FOXO3a proteins and the mRNA levels of Atrogin-1 and MuRF-1 were detected by western blotting and real-time PCR, respectively. The expression levels of FOXO1 and FOXO3a in nuclei fraction were significantly higher in the HS Group compared to the Baseline Group. The expression levels in the Low-ICT Group were slightly lower than those in the HS Group (P > 0.05), while the levels in the High-ICT Group were significantly lower than those in the HS Group. Wortmannin treatment led to the significantly higher levels of FOXO1 and FOXO3a in nuclei fraction compared to the High-ICT Group. No differences were found between the Wort Groups and the HS Group. Similar data were also found in the mRNA levels of Atrogin-1 and MuRF-1 (Fig. 7).

Discussion
To our knowledge, the current study is the first one to demonstrate that ICT could require PI3K/Akt signaling to attenuate the skeletal muscle atrophy induced by mechanical unloading in a dose-dependent manner in adult rats. After 28-day HS, declines of 42.7% in muscle mass and 54.7% in muscle fiber CSA were observed, which were consistent with the previous findings 11 . An attenuation effect of ICT treatment on muscle atrophy during mechanical unloading was demonstrated. The high dose ICT treated animals showed 25.4% higher in muscle mass and 48.1% higher in muscle fiber CSA compared to the HS animals, the low dose ICT treatment showed moderate attenuation effect. It indicated that ICT could attenuate the decreases of muscle mass and muscle CSA induced by mechanical unloading in a dose-dependent manner.
As an antigravity muscle, soleus muscle consists of mostly type I (slow-twitch) fibers 33 . During unloading-induced muscle atrophy, a slow-to-fast fiber type (type I to II) transition was noticed in our study, which was consistent with the previous reports 5, 34 . The low dose ICT treatment moderately attenuated the transition compared to the HS Group, whereas the high dose ICT significantly attenuated the transition in the current study. Consistently, the muscle functional analysis showed that the twitch contraction time, half relaxation time, twitch peak force and tetanic peak force decreased during unloading-induced muscle atrophy, which were all attenuated by the ICT treatment. These evidences indicated that ICT treatment could attenuate this slow-to-fast fiber type transition and the decrease of muscle function in a dose-dependent manner.
Inconsistently, the specific force in the High-ICT Group was close to that of the Baseline Group, indicating ICT treatment could prevent the decrease of the specific force during unloading-induced muscle atrophy. The specific force reflects the functional quality of the muscle, maximum specific force in single skeletal muscle fibers is dependent on the number of cross bridges per half sarcomere, the average force per cross bridge, and the fraction of cross bridges in the force generating state 35 . Regarding to the incomplete prevention effect of ICT on the structural features including the muscle mass, muscle fiber CSA and muscle fiber type distribution during  mechanical unloading, the inconsistence between the structural and functional consequences might be due to the involvement of more than one mechanism, for example, the regulation of ICT on the excitation-contraction (E-C) coupling properties during unloading-induced muscle atrophy.
Furthermore, the safety evaluation showed no dominant side effects on blood biochemical parameters and major organs during the ICT treatment. ICT could be a potential candidate for counteracting muscle atrophy induced by mechanical unloading.
PI3K/Akt signaling pathway is critical for regulating the balance between protein synthesis and degradation. Activation of PI3K leads to activation of the serine/threonine kinase Akt, which in turn phosphorylates and activates its downstream proteins (mTOR, p70S6K, 4EBP1 and eIF4E) and result in increased protein synthesis 36,37 . Our in vitro study showed that ICT treatment could promote the expression of PI3K/Akt signaling markers. Consistent with previous reports 38, 39 , the phosphorylation levels of PI3K at p110 catalytic subunit, Akt and the downstream markers related to protein synthesis decreased after 28-day HS, which could be attenuated by ICT treatment in a dose-dependent manner in our study. On the other hand, PI3K/Akt pathway also could inhibit the process of protein degradation 19 . Deactivation of Akt would lead to relocation of FOXO family transcription factors into the nucleus and activates atrophy-induced genes, such as muscle-specific E3 ubiquitin ligases Atrogin-1/MAFbx and MuRF-1, which contributes to the protein degradation during the progression of muscle atrophy 19,40,41 . In our study, ICT treatment could attenuate the relocation of FOXO1 and FOXO3a to the nucleus and then depress the elevation of mRNA expression of Atrogin-1 and MuRF-1. Taken together, ICT could counteract the reduction in the activity of PI3K/Akt signaling pathway during unloading.
Wortmannin is a specific PI3K inhibitor, which binds to the p110 catalytic subunit, noncompetitively and irreversibly inhibiting the enzyme 42,43 . The expression level of phosphorylated p110 catalytic subunit of PI3K in the ICT+ Wort Group in vitro (High-ICT+ Wort Group in vivo) was significantly lower than that in the ICT Group in vitro (High-ICT Group in vivo). Consistently, the counteraction effect of ICT treatment on unloading-induced muscle atrophy could also be abolished by wortmannin, as evidenced by the lower muscle mass, reduced muscle fiber CSA, more slow-to-fast transition and decreased muscle function in the High-ICT+ Wort Group when compared to the High-ICT Group. These evidences indicated that PI3K at p110 catalytic subunit in the PI3K/Akt signaling was required by ICT to attenuate unloading-induced muscle atrophy.
Theoretically, the muscle atrophy should be more serious once the wortmannin treatment and HS were combined. Interestingly, there was no significant difference in all the examined parameters between the Wort Group and the HS Group. PI3K/Akt signaling pathway is crucial for regulating both protein synthesis and degradation during muscle atrophy. Although there are other signaling pathways contributing to muscle atrophy as well [44][45][46] , PI3K/Akt signaling pathway might play the dominant role during the muscle atrophy induced by mechanical unloading 47 . There's no significant difference between the activation levels of PI3K in HS group and HS plus wortmannin group, indicating wortmannin treatment plus HS does not have a superimposition effect. It implied that PI3K/Akt signaling pathway inhibition might be mainly required by HS to do harm to muscle.
In summary, our findings demonstrated that ICT treatment had an attenuation effect on mechanical unloading-induced muscle atrophy in a dose-dependent manner and PI3K/Akt signaling was required by ICT to exert its therapeutic effect.

Materials & Methods
Cell Culture. C2C12  Animal model and group designation. Eighty 3-month old male SD rats were recruited and divided into the following eight groups (n = 10 per group): 1) Baseline Group: sacrificed before HS; 2) Age-matched Control Group: weight bearing throughout experiment; 3) HS Group: HS plus vehicles served as treatment control; 4) Low-ICT Group: HS plus ICT at low dose (80mg/kg/day); 5) High-ICT Group: HS plus ICT at high dose (120mg/ kg/day); 6) Low-ICT+ Wort Group: HS plus Low-ICT plus Wortmannin (15 μ g/kg/day); 7) High-ICT+ Wort Group: HS plus High-ICT plus Wortmannin; 8) Wort Group: HS plus Wortmannin. ICT was orally administered daily throughout HS. Wortmannin was intraperitoneally injected daily throughout HS 48 . All the procedures used in animal studies were carried out in accordance with the Animal Experimentation Regulations at The Chinese University of Hong Kong and The Hong Kong Code of Practice for Care and Use of Animals for Experimental Purposes. All experimental protocols were reviewed and approved by the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong (Ref No. 12/019/GRF-5). At the end of hindlimb suspension, all the animals were sacrificed with overdose of sodium pentobarbital (35mg/kg) (Sigma-Aldrich, Saint Louis, MO, USA). One side of soleus muscle was subjected to the in vitro functional testing and subsequent histology and immunohistochemistry analysis. The other side of soleus muscle was quickly frozen in liquid nitrogen for protein and total RNA extraction followed by the western blotting and real-time PCR analysis, respectively. For safety evaluation, the blood sample was obtained from the HS, Low-ICT and High-ICT groups in order to determine the biochemical parameters. The heart, liver, kidneys, lungs and brain from the above three groups were dissected out, washed with saline and conserved in 10% formalin for histological analysis.
Hindlimb suspension procedure. The animals were subjected to hindlimb suspension for 28 days following established procedure 11,49 . Briefly, a strip of adhesive tape (15 cm × 0.5 cm) was applied to the animal's tail, which was suspended by passing the tape through a fish-line swivel that was attached to a metal bar on the top of the cage. This allowed the forelimbs to have contact with the grid floor and allowed the animals to move around the cage for free access to food and water. The suspension height was adjusted to prevent the hindlimbs from touching any supporting surface while maintaining a suspension angle of approximately 30°. The animal's overall appearance, drinking and eating habits and tail were monitored four times per day. The distal tip of the tail was examined to ensure that the procedure did not occlude blood flow to the tail by maintaining the tail pink.
In vitro muscle functional testing. The in vitro muscle functional testing was performed following our established protocol 50,51 . Dissected muscle with intact tendons was mounted in a specific chamber bathed with 95%O 2 /5%CO 2 -bubbled Krebs' solution (In Vitro Muscle Test System 1205, Aurora Scientific Inc., Aurora, ON, Canada). One end of the tendon was attached to a hook connected to the lever arm of a position feedback motor, and the other end was attached to a force transducer. The muscle was adjusted to the optimum length (L 0 ) at which the maximal twitch force was produced. The muscle was stimulated with 0.5 ms pulses of supramaximal intensity through two platinum plates that were parallel to the muscle. Peak twitch force at 1Hz and peak tetanic force at 100Hz for 400 ms were recorded, and twitch contraction time and half relaxation time were calculated with the ASI Dynamic Muscle Control Software (DMC v5.1 beta, Aurora Scientific Inc.). Specific force was expressed as peak tetanic force normalized to the muscle physical cross-sectional area 52 .
Histology and immunohistochemistry. The dissected soleus muscles were snap frozen in liquid nitrogen-cooled isopentane and then embedded in OCT medium. Serial cross-sections (6 μ M thickness) were cut from the mid-belly of the muscles on a cryostat at − 20 °C for histological and immunohistochemical staining. Hematoxylin and eosin staining was performed on sections to examine the general morphology and to determine the mean cross-sectional area (CSA) of the muscle fiber.
Heart, liver, kidneys, lungs and brain were conserved in 10% formalin for histopathological studies. Tissues were processed by conventional techniques. The paraffine embedded sections of 4-5 μ m thickness were prepared with the rotary microtome, stained with hematoxylin and eosin for microscopic examination.
Western blot analysis. It was to test the activation of PI3K/Akt signaling markers related to protein synthesis and myogenic markers, protein extracted from cell or muscle homogenates was quantified using Bradford Assay (Bio-Rad). Protein samples were separated by SDS-PAGE and transferred into nitrocellulose membranes (Perkin Elmer). After blocking with 5% nonfat dry milk in tris buffered saline (TBS) containing 0.05% Tween 20, the membranes were then probed with primary antibodies against total and phospho-PI3K p110 (Tyr 485) (Santa Cruz), total and phospho-Akt (Ser473), total and phospho-mTOR (Ser2481), total and phospho-p70S6K (Thr389), total and phospho-4EBP1 (Thr37/46) (Cell Signaling Technology, Beverly, MA, USA), MyoD and myogenin (Santa Cruz). Membranes were incubated with specific horseradish peroxidase-conjugated secondary antibodies (Santa Cruz). Immunodetection was performed using enhanced chemiluminescence consistent with the manufacturer's protocol (BioRad). Membranes were also probed with an antibody for GAPDH (Santa Cruz) to serve as a loading control.
It was to test the cellular distribution of FOXO1 and FOXO3a related to protein degradation, frozen muscle pieces were fractionated into nuclear fraction using a Subcellular Proteo-Extract proteome kit (EBD Biosciences). The proteins extracted from nuclear fraction and total cell were then subjected to the western blot analysis mentioned above.
Real-time PCR analysis. Total RNA of soleus muscles was extracted using SV total RNA isolation system (Promega), and cDNA was synthesized from 0.5 μ g total RNA of each soleus sample using SuperScript III First-Strand Synthesis Kit (Invitrogen) according to manufacturer's instructions. TaqMan ® gene expression assay with primers and MGB probes that are specific for rat Atrogin-1 and MuRF-1 (Assay ID: AF441120 for Atrogin-1 and AY059627 for MuRF-1; Applied Biosystems) were used for real-time PCR analysis. The amplification was performed in a real-time PCR system (StepOnePlus, Applied Biosystems). GAPDH was used as an endogenous control for normalization. Each sample was assessed in triplicate. Relative changes in Atrogin-1 and MuRF-1 gene expression were determined using the 2 −ΔΔCt (normalized expression ratio) method of analysis.
Blood biochemical parameters evaluation. Blood samples without anticoagulant were centrifuged (1000 rpm, 30 min) and the obtained serum was kept at − 4 °C for four days, at most, before the biochemical parameters determination. Serum creatinine, urea, uric acid, total protein, alanine aminotransferase (ALT), aspartate aminotransferase (AST) and alkaline phosphatase were determined. Creatinine and total proteins were measured by colorimetric assay, while the other parameters were measured by enzymatic colorimetric assay. The kits used in these tests were obtained from Abcam (UK) and the analysis was carried out using a Multiparametric Autoanalyzer (Lisabio B652, USA). Data analysis. The results were presented as mean ± standard error of the means (SEM). The data were analyzed by one way analysis of variance (ANOVA). LSD post-hoc test was made for multiple comparisons among all eight groups to determine the differences at a significance level of P < 0.05.