Gosha-jinki-Gan (GJG) shows anti-aging effects through suppression of TNF-α production by Chikusetsusaponin V

Frailty develops due to multiple factors, such as sarcopenia, chronic pain, and dementia. Go-sha-jinki-Gan (GJG) is a traditional Japanese herbal medicine used for age-related symptoms. We have reported that GJG improved sarcopenia, chronic pain, and central nervous system function through suppression of TNF-α production. In the present study, GJG was found to reduce the production of TNF-α in the soleus muscle of senescence-accelerated mice at 12 weeks and 36 weeks. GJG did not change the differentiation of C2C12 cells with 2% horse serum. GJG signicantly decreased the expression of MAFbx induced by TNF-α in C2C12 cells on real-time PCR. TNF-α signicantly decreased the expression of PGC-1α and negated the enhancing effect of GJG for the expression of PGC-1α on digital PCR. Examining 20 chemical compounds derived from GJG, cinnamaldehyde from cinnamon bark and Chikusetsusaponin V (CsV) from Achyrantes Root dose-dependently decreased the production of TNF-(cid:0) in RAW264.7 cells stimulated by LPS. CsV inhibited the nuclear translocation of NF-κB p65 in RAW264.7 cells. CsV showed low permeability using Caco-2 cells. However, the plasma concentration of CsV was detected from 30 minutes to 6 h and peaked at 1 h in the CD1 (ICR) mice after a single dose of GJG. In 8-week-old SAMP8 mice fed 4% (w/w) GJG from one week to four weeks, the plasma CsV concentration ranged from 0.0500 to 10.0 ng/mL. The evidence that CsV plays an important role in various anti-aging effects of GJG via suppression of TNF-(cid:0) expression is presented. TNF-(cid:0)


Introduction
Frailty is becoming one of the most important problems for clinical practice and public health in the world 1 . In particular, the population of elderly people in Japan is increasing more rapidly than that of any other country 2 . It is an urgent task to establish strategies against frailty to prevent an increase in the bedridden and nursing care population. First, the physical aspects of the frailty phenotype require attention. Fried et al reported that frailty is composed of weak grip strength, slow gait speed, low physical activity, exhaustion, and unintentional weight loss 3 . After various discussions, the concept of frailty has changed. Currently, frailty is thought to be multidimensional, involving both physical and psychosocial factors. As to the physical aspects of frailty, sarcopenia has been gaining attention. Sarcopenia denotes the decrease in muscle strength and muscle mass associated with aging 4 . Sarcopenia is thought to be caused by a variety of factors, such as aging-related decreases in physical activity, in addition to changes in nutrient intake, hormones, and the in ammatory response 5 . At least 50% of individuals ≥80 years old reportedly have sarcopenia 6 . Additionally, chronic pain control due to lumbago and arthralgia is important for maintaining continuity of rehabilitation and exercise to prevent sarcopenia. With respect to the psychosocial aspects of frailty, depression and dementia induce withdrawal and loneliness, resulting in the development of sarcopenia 1 . Given the multidimensional aspects of frailty, new pharmacotherapies need to be developed.
In traditional Japanese herbal medicine (Kampo medicine), Go-sha-jinki-gan (GJG) has been used to alleviate various types of age-related conditions, e.g. lumbago, numbness, and nocturia 7,8 . In Japan, it is well known that GJG has a certain level of clinical effect for the above age-related symptoms. However, the pharmacological mechanism of GJG has yet to be elucidated in detail. Anticipating the anti-aging effect of GJG, the anti-sarcopenic effect of GJG was evaluated in senescence-accelerated mice (SAMP8), which are widely used in aging research 9 . We reported that GJG had a clear anti-sarcopenic effect on SAMP8 mice by improving the insulin-like growth factor (IGF-1) signaling pathway, increasing peroxisome proliferator-activated receptor-γ coactivator-1 alpha (PGC-1α) in the mitochondrial network, and decreasing the concentrations of tumor necrosis factor-alpha (TNF-) in muscle 10 . Focusing on the decreased production of TNF-, we examined the pain-relieving effect of GJG using chronic constriction injury (CCI) model mice. GJG signi cantly improved allodynia and hyperalgesia on the von Frey test, the cold-plate test, and the hot-plate test through suppression of TNF-expression derived from activated microglia in the spinal cord of CCI model mice 11 . Next, we investigated whether GJG affects the central nervous system. GJG signi cantly prevented the paralytic symptoms in experimental allergic encephalomyelitis (EAE) model mice by decreasing TNF-production derived from activated microglia in the brain 12 .
Taken together, GJG improved sarcopenia, chronic pain, and central nervous system function through the suppression of TNF-α production from macrophage cells. However, it is unclear which botanical herbs or chemicals derived from GJG are mainly involved in decreasing TNF-production. GJG consists of ten medicinal herbs in xed proportions. Examining various chemical compounds derived from the component herbs of GJG, we found that Chikusetsusaponin V (CsV), which is derived from Achyranthes Root, dose-dependently decreased the production of TNF-and the NF-κB signaling pathway. Moreover, we con rmed that CsV was detected in mice plasma after oral administration. In this study, the evidence that CsV plays an important role in the various effects of GJG through suppression of TNF-expression is presented.

Results
GJG decreased the expression of TNF-in the soleus muscle of SAMP8 mice at 12 and 36 weeks.
We previously reported the anti-sarcopenic effect of GJG in SAMP8 mice. GJG decreased the expression of TNF-in the soleus muscle of SAMP8 mice at 38 weeks 10 . We found that GJG suppressed the production of TNF-α from activated microglia cells 11,12 . We investigated the time-course of TNF-α and IL-6 expressions after GJG administration in SAMP8 and SAMR1 mice from 12 weeks to 48 weeks (Fig. 1a).
No signi cant difference was observed in body weight and blood sugar regardless of GJG administration, the same as before the examination (data not shown). The soleus muscles were collected in SAMP8 and SAMR1 mice at 12, 16, 24, 36 and 48 weeks. Western blotting analysis showed that the expression of TNF-in SAMP8 soleus muscles increased from 12 weeks to 36 weeks compared to that in SAMR1 mice and decreased with GJG administration at 12 weeks and 36 weeks (Fig. 1b,c). On the other hand, the expression of IL-6 in SAMP8 soleus muscle decreased slightly at 12 weeks, but was not changed from 16 weeks to 48 weeks, the same as in SAMR1 (Fig. 1d,e).

GJG decreased the expression of MAFbx in C2C12 cells after TNF-stimulation
We previously reported that GJG improved the IGF-1 signaling pathway and increased PGC-1α in the soleus muscle of SAMP mice 10 . We examined whether GJG has a direct effect on muscle cells using mouse skeletal muscle cell line C2C12 cells. The C2C12 cells differentiated to myotube cells with DMEM containing 2% horse serum 13 . At the same time, they were treated with 10 µg/mL, 100 µg/mL, or no GJG. We expected that GJG would directly affect muscle atrophy, but GJG did not change the differentiation of C2C12 cells to myotubes with 2% horse serum (Fig. 2a). It is well known that MAFbx and MuRF-1 suppress the degradation of skeletal muscle protein 14,15 . We investigated the expression levels of MAFbx and MuRF1 using real-time RT-PCR 3 h after 10-100 ng/ml of TNF-α or no TNF-α; 10 ng/mL of TNF-α signi cantly induced the expression of MAFbx compared to non-stimulated cells (1.12 ± 0.047 vs 2.90 ± 0.10 fold changes, P < 0.0001 Tukey-Kramer test). GJG signi cantly decreased the expression of MAFbx induced by TNF-α from 2.90 ± 0.10 to 1.63 ± 0.003 fold changes (P < 0.0001 Tukey-Kramer test). The same results were obtained with stimulation by 100 ng/mL TNF-α (Fig. 2b). On the other hand, TNF-α stimulation did not change the expression of MuRF-1 with or without GJG (Fig. 2c). Next, the expression of PGC-1α with various stimuli was examined in C2C12 cells using real-time RT-PCR. GJG increased the expression of PGC-1α by less than 1.5 times, and TNF-α decreased the expression of PGC-1α by more than 0.5 (data not shown). In general, it is di cult to distinguish the expression of the target gene less than 2 times or more than half using real-time RT-PCR. Therefore, a digital RT-PCR assay was used to evaluate the expression of PGC-1α. GJG signi cantly increased the expression of PGC-1α from 0.980 ± 0.08 to 1.183 ± 0.10 fold changes (P < 0.05 Dunnett's test, Fig. 2d). TNF-α signi cantly decreased the expression of PGC-1α and negated the enhancing effect of GJG on the expression of PGC-1α. These results suggest that TNF-α plays a central role in the pathogenesis of sarcopenia. These results are consistent with the reports related to sarcopenia and in ammation, especially TNF-α 16− 19 . We hypothesized that GJG shows an anti-sarcopenic effect through suppression of TNF production. Next, whether 20 chemicals derived from GJG and one control chemical derived from another botanical herb have an effect on TNF-α production was examined using RAW 264.7 cells, a mouse macrophage cell line.
GJG-derived chemicals decreased TNF-α expression by inhibition of nuclear translocation of NF-κB p65 in RAW 264.7 cells GJG is composed of ten medicinal herbs in xed proportions (Fig. 3a). It is well known that RAW 264.7 cells produce TNF-α with stimulation by lipopolysaccharide (LPS). We have previously reported that GJG prevented the production of TNF-α by RAW 264.7 cells stimulated with LPS in a dose-dependent manner 11 . On the other hand, GJG did not show a clear dose-dependent inhibitory effect of IL-6 in the same assay system (Fig. 1S). Thus, 20 chemical compounds derived from GJG and a negative control compound from another herbal medicine were selected based on analysis of the three-dimensional highperformance liquid chromatography pro le and LC/MS analysis of GJG 10 . It was found that cinnamaldehyde from cinnamon bark and CsV from Achyrantes Root decreased the production of TNF-α from RAW 264.7 cells stimulated with LPS without cytotoxicity (Fig. 3b). It was con rmed that cinnamaldehyde and CsV signi cantly inhibited the expression of TNF-α in a dose-dependent manner compared to control with DMSO (7064.1 ± 598.9 pg/mL vs 3792.4 ± 598.9 pg/mL, 3942.4 ± 598.9 pg/mL, P < 0.05 Tukey-Kramer test, Fig. 3c). Rokumi-Gan (RJG) consists of six medicinal herbs without cinnamon bark and Achyrantes Root (Fig. 3a), which means that RJG does not prevent the production of TNF-α. It was con rmed that RJG did not have an anti-sarcopenic effect in SAMP8 mice ( Figure S2). These results suggest that cinnamaldehyde and CsV play important roles via the suppression of TNF-α levels. It is di cult to evaluate cinnamaldehyde in human blood because cinnamon bark is a common daily food.
CsV is also detected from Panax japonicus, recognized as a very important medical herb 20,21 . Therefore, we focused on CsV as a drug e cacy marker of GJG.
Toll like receptor (TLR) 4 is well known as a receptor for LPS 22 . Therefore, whether GJG affected the expression of TLR4 signaling molecule was examined. GJG did not change the protein levels of TLR4 and MyD88 on Western blotting analysis (Fig. 4a). Next, whether GJG has an effect on the nuclear translocation of NF-κB p65 was examined, because NF-κB p65 is downstream of the TLR4 signaling pathway. In general, LPS stimulation activates the phosphorylation of IκBα Ser 32 23 . As a result, IκBα is degraded via ubiquitination, and NF-κB p65 translocates to the nucleus 23 . The same results were observed in the present study, and GJG inhibited the phosphorylation of IκBα Ser 32 and nuclear translocation of NF-κB p65 in a dose-dependent manner (Fig. 4b). In almost the same manner, CsV showed nuclear translocation of NF-κB p65 in a dose-dependent manner (Fig. 4c). These results indicated that CsV has an anti-in ammatory effect on RAW 264.7 cells through its inhibitory effect on nuclear translocation of NF-κB p65. These results suggest that CsV is a pharmacological marker for the antiin ammatory effect of GJG. Then, the characteristics of CsV as a drug compound were investigated.
CsV showed high solubility, low permeability, and high metabolic stability CsV has a saponin structure, as shown in Fig. 5a. Saponin has multiple hydrophilic sugars bound to a hydrophobic skeleton. To examine the solubility of CsV with PBS, the supernatant of 1% or 5% DMSO/PBS solution of CsV was examined using an LC/MS assay system, and the solubility was calculated based on the area under the curve (AUC). Figure 5b shows the single peak of the supernatant of 5% DMSO/PBS solution of CsV; 500 µM of CsV with 5% DMSO showed an AUC about 10 times greater than that of 50 µM of CsV with 5% DMSO, and 500 µM of CsV with 1% DMSO showed almost the same AUC as that of 500 µM of CsV with 5% DMSO (Fig. 5c). These results show that CsV has more than 95% solubility with PBS.
Next, the membrane permeability of CsV was examined, because herbal medicine is generally given orally. The caco-2 permeability assay system 24 was used, and the permeability coe cient (Papp) of CsV was calculated. The Papp of CsV was less than 0.1 cm/sec x 10 − 6 , which indicates that CsV has very low membrane permeability (Fig. 5d). Metabolic stability is said to be important to maintain blood concentrations of chemical compounds. To examine the metabolic stability of CsV, CsV was mixed with liver microsomes, coenzyme group (NADPH, G-6-P) and G6PDH, at 37 ˚C for 10 min and 60 min.
Unchanged CsV compounds were detected at above 80% at 10 and 60 min compared to the unchanged CsV before treatment as 100% (Fig. 5e). These results show that CsV has high metabolic stability. Next, whether Csv can be detected in the plasma of mice following oral administration of GJG was examined, because the permeability of CsV is very low.
CsV was detected in the plasma of CD1 (ICR) mice and SAMP8 mice following oral administration of GJG A single dose of GJG 1 g/kg was given to 8-week-old CD1 (ICR) mice fasted for 16 h, and changes in plasma CsV levels were examined. An LC-MS/MS assay system detecting a single peak of CsV in CD1 (ICR) mice plasma from 0.0200 to 10.0 ng/mL was established (Fig. 6a,b). After a single dose of GJG, the plasma concentration of CsV was detected from 30 min to 6 h and peaked at 1 h (Fig. 6c).
Finally, changes of plasma levels of CsV were examined in 8-week-old SAMP8 mice fed a normal diet or a normal diet supplemented with 4% (w/w) GJG from one week to four weeks. In SAMP8 mice, the plasma concentration of CsV ranged from 0.0500 to 10.0 ng/mL. Although there were variations depending on the individual mice, CsV was clearly detected in the plasma of SAMP8 mice fed a GJG-containing diet from one week to four weeks. On the other hand, CsV was not detected in mice fed a normal diet. These results clearly showed that CsV is absorbed into blood following oral administration of GJG and plays a central role in the anti-in ammatory effect of GJG.

Discussion
In this study, GJG was shown to inhibit TNF-α production in vivo and in vitro. Investigating the inhibitory mechanism of GJG using various chemical compounds, it was found that CsV derived from Achyranthes Root, a component of GJG, inhibited the nuclear translocation of NF-kB p65 after LPS stimulation in RAW264.7 cells. In addition, CsV was detected following oral administration of GJG in the plasma of ICR (ICR) mice or SAMP8 mice. Therefore, these results show that CsV plays a pivotal role in the pharmacological effects of GJG on muscle, spinal cord, and brain.
We previously reported that GJG improved the amounts of skeletal muscles bers and ber types in SAMP8 mice at 38 weeks 10 . We expected that GJG would have a direct effect on muscle atrophy, but GJG did not change the differentiation of mouse skeletal muscle cell line C2C12 cells to myotubes using 2% horse serum. These results suggest that GJG did not have a direct effect on C2C12 cells. Higher serum levels of TNF-α were reported in elderly people with sarcopenia compared to healthy controls 16 − 18 . It has been reported that TNF-α induces muscle atrophy in aged mice and TNF transgenic mice 19 . Thus, these reports suggest that TNF-α is required in the pathogenesis of sarcopenia. To examine the effect of GJG on C2C12 cells, TNF-α stimulation was added in the supernatant of the culture medium. It was found that GJG inhibited the expression of MAFbx induced by TNF-α at 3 h in C2C12 cells. MAFbx has been reported to be expressed by TNF-α via p38 in C2C12 cells from 2 h to 4 h 25 . These results were consistent with the present results. On the other hand, MuRF-1 was not induced by TNF-α in the present assay system using C2C12 cells. It has been reported that MuRF-1 was induced by TNF-α at 24 h after stimulation 26 . In the present study, the expression of MuRF-1 was evaluated 3 h after TNF-α stimulation. One of the reasons for the different results is that the expression of MuRF-1 was evaluated at a different time after TNF-α stimulation. We have previously reported that administration of GJG increased serum levels of IGF-1 in SAMP8 mice at 38 weeks. IGF-1 is required for skeletal muscle hypertrophy 14,15 . Plasma IGF-1 concentrations decrease with aging 27 . In an aging mice assay, the insulin/IGF-1 signaling pathway also regulates the expression of MuRF-1. With an in vitro assay using C2C12 cells, IGF-1 was not used in the supernatant of the culture medium, which might cause different experimental results.
We previously reported that GJG increased the protein level of PGC-1α in SAMP8 mice at 38 weeks. PGC-1α is known to be one of the main regulators of mitochondrial biogenesis and regulates muscle ber types 28, 29 . It has been reported that mitochondrial ssion causes muscle atrophy 30 . We examined whether GJG directly increases PGC-1α using a real-time PCR assay system, but the results were controversial because GLG changed the expression of PGC-1α slightly. Since mitochondrial DNA levels in peripheral blood mononuclear cells decrease gradually with aging, it is di cult to quantify mitochondrial genes using real-time PCR. Real-time PCR is less sensitive because it calculates the mitochondrial DNA copies compared to a housekeeping gene 31 . Digital PCR has been used in clinical microbiology, for example HIV, because it allows for absolute target measurement by directly counting single droplets 32,33 .
Digital PCR has been shown to be useful for measuring mitochondrial genes 31 . We showed that GJG increased the expression of PGC-1α slightly, but signi cantly, using a digital PCR assay system. TNF-α suppressed the expression of PGC-1α and negated the increasing effect of GJG. It has been reported that TNF-α decreased the mRNA of PGC-1α in C2C12 myotube cells 34,35 . The present results and former reports indicate that it is important to regulate the expression levels of TNF-α, which affects the mitochondrial function of muscle.
As a method of searching for active ingredients of herbal medicines, separation of a solution of a herbal medicine into fractions is performed, and the effects of each fraction are examined. The chemical compounds of GJG have been previously identi ed based on the three-dimensional high-performance liquid chromatography pro le and LC/MS analysis against GJG 10 . Although it is necessary to know the ingredients to some extent, the advantages of our methods are that they can examine the dose dependency of chemical compounds and con rm the reproducibility of experimental results. Two active compounds were detected from GJG, cinnamaldehyde and CsV. It has been reported that cinnamaldehyde inhibited the production of TNF-α stimulated by LPS in RAW 264.7 cells 36 . The present results do not contradict previous reports. Cinnamon bark is one of the components of Kakkon-to, a traditional herbal medicine, which has been used for the treatment of in uenza and the common cold 37 . CsV is noted as one of the active components of Panax japonicus, the most famous herbal medicine because of its anti-aging effect. Total saponin of Panax japonicus contains Cs II, IV, IVa, and V 21 . It has been reported that CsV and CsIVa have anti-in ammatory effects by inhibiting the NF-κB signaling pathway in RAW 264.7 cells stimulated by LPS 21 . The present results and previous reports show that cinnamaldehyde and CsV prevent TNF-α production by macrophage cells stimulated with LPS. In fact, RJG, composed of six traditional herbs without cinnamon bark and Achyrantes Root, did not have an anti-sarcopenic effect in SAMP8 mice. These results suggest that cinnamaldehyde and CsV might be biomarkers of the pharmacological effect of GJG. However, cinnamon bark is commonly used as a food ingredient and might be detected in the plasma without administration of GJG. On the other hand, Achyrantes Root is only used as a herbal medicine. Therefore, we focused our attention on CsV, one of the main active compounds of Achyrantes Root. The characteristics of CsV as a chemical compound remain unknown. We established a CsV detection system with LC-MS and showed that CsV has low permeability and high metabolic stability. As far as we know, this is the rst report of the characteristics of the chemical compound of CsV. Based on permeability experiments, low absorption of CsV following oral administration is expected. However, CsV was clearly detected in the plasma of CD1 (ICR) mice with single administration and SAMP8 mice with continuous administration. It has been reported that herbal medicines are quickly absorbed by transporters 38 . Therefore, unknown transporters are likely involved in the absorption of CsV. Further study is needed to clarify the precise mechanism of CsV absorption.
Thus, this study demonstrated that GJG inhibited the production of TNF-α in the soleus muscle of SAMP8 mice biphasically. GJG directly inhibited the expression of TNF-α from RAW 264.7 cells via the derived compounds, cinnamaldehyde and CsV. CsV was clearly detected in the plasma of SAMP8 mice following oral administration of GJG. These results suggest that CsV is a promising candidate for monitoring the pharmacological effects of GJG. (ChemFaces Biochemical Co., Ltd., Wuhan, China). Animals SAMP8 and SAMR1 mice (8 weeks old), purchased from SLC, Inc. (Shizuoka, Japan), were divided into two groups of 5 each one week after they were acclimated based on the diet that they were fed: a normal diet (powdered mouse food; Oriental Yeast Co. Ltd. (Tokyo, Japan; P8 + N group); and a normal diet supplemented with 4% (w/w) GJG (P8 + GJG group). The control group, which consisted of 8-week-old male SAMR1 mice also purchased from SLCm was also divided into two groups of 10 each based on the diet that they were fed: a normal diet (R + N group); and a normal diet supplemented with 4% (w/w) GJG (R + GJG group). The general conditions and body weight of all mice were recorded. Housing care and the experimental protocol were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with the approval of the Animal Care and Use Committee of Osaka University. We con rmed the study was carried out in compliance with the ARRIVE guidelines Cell Culture And Immuno uorescence Staining Mouse skeletal muscle cell line C2C12 cells were obtained from DS Pharma Biomedical (Osaka, Japan).

Materials And Methods
C2C12 cells were seeded in a 6-well plate at 5×10 4 cells/ml and cultured in DMEM containing 10% FBS, 100 U/mL penicillin, and 100 ug/mL streptomycin (Nacalai, Kyoto, Japan) for 48 h. After 100% con uence, the medium was replaced with DMEM containing 2% horse serum (Gibco BRL, Grand Island, NY, USA) and cultured for 96 h. Differentiation of C2C12 cells from myoblasts to myotubes was con rmed using immuno uorescence staining of sarcomeric α actinin 13 . C2C12 cells were xed with 4% paraformaldehyde solution in PBS at room temperature for 10 min and permeabilized with 1 mL of 0.25% Triton/PBS for 20 min. After blocking with 5% skim milk/PBS at room temperature for 30 min, anti-Sarcomeric Alpha Actinin antibody (ab9465) (Abcam, Cambridge, UK) with 5% skim milk/PBS was added to the dilution and left overnight at 4°C. The secondary antibody was diluted with PBS, the diluted solution was added, shielded from light with aluminum foil, and allowed to stand at room temperature for 1 h. Nuclear staining was performed with DAPI-containing encapsulant (ProLong ™ Gold Antifade Mountant with DAPI: Invitrogen) on a glass slide 13 .

Real-time Quantitative PCR And Droplet Digital PCR
The differentiated C2C12 myotubes underwent treatment with 100 µg/mL GJG, and then, 24 h later, the medium was replaced with a serum-free DMEM medium containing 100 µg/mL GJG and 10 or 100 ng/mL murine TNF-α (Peprotech, Rocky Hill, NJ). Then, 0, 3, and 6 h after the medium was changed, total RNA was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Subsequently, SuperScript VILO Master Mix (Invitrogen, Carlsbad, CA) was used to synthesize cDNA from 200 ng of RNA. The ViiA 7 real-time PCR system (Life Technologies, Carlsbad, CA, USA) was used for real-time quantitative PCR, with the following cycling conditions: 50 ºC for 2 min, 95 ºC for 20 sec, followed by 40 cycles of 95 ºC for 1 sec and 60 ºC for 20 sec at annealing temperature. TaqMan Fast Advanced master mix (Applied Biosystems, Foster City, CA) was used to amplify the target genes from 10 ng of cDNA. Gene-speci c primers and FAM-labeled probes (mouse MAFbx: Mm00499523_m1, mouse MuRF1: Mm01185221_m1, mouse PGC1-α: Mm01208835_m1, mouse B2M: Mm00437762_m1) were purchased from Applied Biosystems. Relative expression levels relative to beta2 macroglobulin (B2M) expression were then calculated.
A QX100 droplet digital PCR (ddPCR) system (Bio-Rad Laboratories, Hercules, CA) was used for ddPCR in a total volume of 20 µL, with 10 µL 2x ddPCR Supermix for the probes (Bio-Rad Laboratories), 1 µL primers and FAM-labeled probes (mouse B2M: Mm00437762_m1, mouse PGC1-α: Mm01208835_m1), and 5 µL sample cDNA (10 ng total RNA). After droplets were generated with a QX100 Droplet Generator (Bio-Rad Laboratories), a C1000 Touch Thermal Cycler (Bio-Rad Laboratories) was used to amplify each cDNA. The thermal cycling conditions were as follows: a pre-cycling hold at 95 ºC for 10 min, followed by 40 cycles of 94 ºC for 30 sec and 58 ºC for 1 min, with nal heating at 98 ºC for 10 min. A QX100 Droplet Reader (Bio-Rad Laboratories) was used to measure droplets, and QuantaSoft software version 1.6.6 (Bio-Rad Laboratories) was used to analyze the target copy number, with normalization of the target copy number by B2M.
TNF-α ELISA and cell viability assay A mouse macrophage cell line (RAW264.7), purchased from Dainippon Pharmaceutical (Osaka, Japan), was grown in DMEM with 10% FBS, 100 µg/mL streptomycin, and 100 U/mL penicillin at 37°C in a humidi ed 5% CO 2 atmosphere. LPS was purchased from Sigma-Aldrich (St. Louis, MO, USA). The RAW264.7 cells were seeded in triplicate at 5 × 10 4 cells/mL in 96-well plates and left for 48 h. At that point, the cells were cultured for 24 h in DMEM with or without 10 or 100 µg/mL GJG ( nal 1% FBS). This was followed by stimulation of the cells by 10 ng/mL LPS for 16 h. The TNF-α concentration in the supernatant of the medium was then evaluated with a mouse TNF-α Quantikine ELISA kit (R & D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions 11 . Cell viability, evaluated using Cell Counting kit-8 (Dojindo, Kumamoto, Japan) according to the manufacturer's protocol, was calculated by comparison to that of cells without GJG and LPS. Each of the above experiments was performed at least three times, and representative data are reported.
Western Blotting Analysis SAMP8 and SAMR1 mice were sacri ced at 12, 16, 24, 36, and 48 weeks (each group, n = 3), and soleus muscles were obtained, with the muscle samples cut into small pieces and homogenized in RIPA buffer (Nacalai) with phosphatase inhibitor cocktail (Nacalai) and protease inhibitor cocktail (Thermo Fisher Scienti c, Tokyo, Japan). After the lysates were centrifuged for 10 min at 14,000 g and 4°C, the supernatants were collected. The BCA Protein Assay Kit (Sigma) was then used to measure protein, followed by Western blot analysis according to previously described methods 10,11 .
After seeding of RAW264.7 cells in triplicate at 1 × 10 5 cells/mL in 6-well plates, they were grown in DMEM with 10% FBS, 100 µg/mL streptomycin, and 100 U/mL penicillin at 37°C in a humidi ed 5% CO 2 atmosphere. Then, 48 h later, the cells were cultured for 24 h in DMEM with or without 10 or 100 µg/mL GJG ( nal 1% FBS). Dexamethasone(Dex) (Sigma, St Louis, MO, USA) was added for 24 h at 10-100 µM. To detect MyD88 and TLR4, cell stimulation was performed with 10 ng/mL LPS for 24 h, followed by homogenization of whole cell lysate in RIPA buffer with phosphatase inhibitor cocktail and protease inhibitor cocktail. To detect NF-κB p65, p-IκBα, and Ser32 IκBα, cell stimulation with 10 ng/mL LPS was performed for 30 min, followed by separation of nuclear extracts using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scienti c). The method described above was then used for Western blot analysis. Speci c primary antibodies were then used for the following: TNF-α (Cell Signaling  Technology #3707 Eight-week-old, male SAMP8 mice were fed a normal diet (n = 3 from 1 week to 3 weeks, n = 6 at 4 weeks) or a normal diet supplemented with 4% (w/w) GJG for four weeks (n = 6 before and at each time point). Plasma samples were collected before and after administration at 1, 2, 3, and 4 weeks. Acetonitrile was added to the plasma samples, and the supernatants were collected after centrifugation. The supernatant was dried and dissolved in 100 µL of solution (10 mmol/L ammonium formate solution/acetonitrile (80:20)). The concentration of chikusetsusaponin V was measured by LC-MS / MS (MS / MS: QTRAP 5500; AB Sciex, Framingham, MA, HPLC: Agilent 1260; Agilent Technologies, Santa Clara, CA). The plasma concentration was calculated by adding a standard solution of chikusetsusaponin V (PhytoLab GmbH & Co. KG, Vestenbergsgreuth, Germany) to plasma to prepare a calibration curve. For analysis of CD1(ICR) mice analysis, the range of the calibration curve was from 0.0200 to 10.0 ng/mL and for SAMP8 mice, the range of the calibration curve was from 0.0500 to 10.0 ng / mL.

Statistical analysis
The real-time PCR results and the TNF-α ELISA results were analyzed using Tukey-Kramer's post hoc test, and the results of droplet digital PCR were examined by Dunnet's test using JAMP Pro version 14. A pvalue less than 0.05 was considered signi cant.
Declarations Figure 1 Suppressive effect of GJG on TNF-α production in the soleus muscle of SAMP8 mice. Experimental design of GJG administration from 12 weeks to 48 weeks in SAMP8 and SAMR1 mice (a), time course study of TNF-α expression in the soleus muscle of SAMP8 mice (b) and in the soleus muscle of SAMR1 mice (c), time course study of IL-6 expression in the soleus muscle of SAMP8 mice (d) and in the soleus muscle of SAMR1 mice (e). The left panel shows the western blotting analysis of TNF-α from 12 weeks to 48 weeks. GAPDH is shown as a loading control.
Immuno uorescence staining with anti-sarcomeric actinin antibody (a). The expression levels of MAFbx using a real-time PCR system (b). Data are presented as means ± SD and analyzed using the Tukey-Kramer test (* P<0.0001 indicates a signi cant difference, N = 3), the expression levels of MuRF-1 using a real-time PCR system (c), the expression levels of PGC-1a using digital PCR (d). Data are presented as means ± SD and analyzed using Dunnett's test (** P<0.001, * P<0.05 indicates a signi cant difference vs.
non-treated, n = 3).    ChikusetsusaponinV was detected in the plasma following oral administration of GJG. No administration of GJG (a) Single peak of chikusetsusaponinV is detected in the plasma of CD1 mice (ICR) (b) After 16 h of fasting, GJG extract powder 1 g/kg was given orally to 8-week-old male CD1 (ICR) mice. Plasma samples were collected before and after administration at 0.5, 1, 2, 4, 6, 8, 10, 12, and 24 h (n = 5 at each time point) (c) Eight-week-old male SAMP8 mice were fed a normal diet (n=3 from 1 week to 3 weeks, n=6