肌抑素之調控及其與肌肉萎縮之研究

Abstract

Myostatin (肌抑素，又稱作growth and differentiation factor 8，GDF8)隸屬於transforming growth factor beta family，具有負向調控小鼠之胚胎及成體骨骼肌細胞之能力。剔除肌抑素基因的小鼠（myostatin knock-out mice），其肌肉大小約為正常鼠的三倍，脂肪量則會減少。若在成鼠體內持續分泌肌抑素，則觀察到肌肉及脂肪組織的同時減少，類似人類惡病質症候群（cachexia syndrome）。肌抑素基因位於人類染色體2q32.2，其突變會造成肌肉量的增加；而包括牛、綿羊、狗、斑馬魚等物種亦發現類似突變及多肉(double-muscle)表現型間之相關性。肌抑素蛋白在物種間具保守性，在具活性的C端序列的部分，嚙齒目、雞、火雞、豬及人皆完全相同。其蛋白產物為375個胺基酸之前原胜肽，經移除21個胺基酸的信號序列，成為全長肌抑素；全長肌抑素雙體，經過蛋白酶切割成為N端原胜肽和C端110個胺基酸的活性分子。C端活性分子與原胜肽、卵泡抑制素(follistatin)或卵泡抑制素相關基因產物以非共價鍵相結合成為不活化態。肌抑素作用方式同時包括自泌(autocrine)及內泌(endocrine)方式，與位於細胞膜表面的activin第二型受器(ActRII)結合，並藉由磷酸化進而活化胞內Smad2及Smad3等轉錄因子。 由於肌抑素具有抑制肌肉生長的作用，其於肌肉相關疾病與代謝症候群之治療前景值得探究。肌肉相關疾病包括：神經肌肉疾病(肌肉失養症、肌肉病變等)、老化造成之肌質缺乏症(sarcopenia)、因代謝性疾病或惡性腫瘤造成之惡病質(cachexia)、廢用性或去神經性肌肉萎縮(disuse/denervation muscle atrophy)、橫紋肌肉瘤(rhabdomyosarcoma)等等。它們雖然有不同的病因，但最終表現均為肌肉萎縮。在2002年兩個實驗室分別利用不同方法來抑制mdx小鼠（杜馨氏及貝克氏肌肉失養症之模式動物）的肌抑素基因，以減輕病鼠症狀並增進其功能。Minetti等人於2006年以腹腔注射trichostatin A (TSA)有效減緩肌肉失養症病鼠之症狀，並發現肌抑素表現受調控。將側索硬化症模式病鼠之肌抑素基因予以抑制，雖然無法延緩疾病發生或延長存活，但能改善橫膈膜的組織病理學特徵，並顯著但短暫的增加肌力及肌肉量。老化過程中肌抑素表現是增加或減少，目前仍未有定論。惡病質指的是嚴重創傷、惡性腫瘤、感染或其他慢性內科疾病造成的脂肪和肌肉組織同時流失，有些癌症病患肌肉質量減少可高達75%，病理變化主要是肌纖維直徑變小，進而造成肌力衰退、疲勞、廢用、乃至死亡。惡病質是不佳的預後因子，除死亡外，還會增加復健需求及醫療支出。目前有探討肌抑素表現的惡病質模式包括：慢性腎衰竭、心臟衰竭、後天免疫不全症候群、慢性阻塞性肺病、愛迪生氏症、庫欣氏症候群、肝硬化、癌症、熱量限制及燙傷等。整體認為在肌肉萎縮過程中肌抑素mRNA表現會增加，但慢性期表現則尚未有定論。罹患全身疾病的住院病患往往在原發疾病治癒後，卻因次發的肌肉萎縮而造成重大失能及社會經濟的直接或間接損失。目前的治療多侷限於復健運動、電療或保守的姑息療法，肌抑素的調控為此類病患提供了一線曙光。坐骨神經截斷為常用之去神經肌肉萎縮模式，大鼠腓腸肌內肌抑素表現量在術後第一天升高、第14天則反而降低，此時肌肉萎縮為術前的50%；腓腸肌內肌抑素蛋白量則大致與mRNA變化相同。顯示肌抑素在去神經肌肉萎縮過程中有複雜的調控機轉，且與發生的時間有密切關係。後肢懸吊則為廢用性肌肉萎縮的模式之一，大鼠經過10天的後肢懸吊，肌肉內肌抑素mRNA及蛋白皆增加，肌肉萎縮明顯；回復行走4天，上述指標皆回復正常。若在懸吊期間每天行走30分鐘，雖然肌抑素mRNA仍會增加，但可有效避免肌肉萎縮，顯示單純肌抑素的增加未必會造成健康成鼠肌肉萎縮。慢性髖關節退化的病患肌抑素mRNA表現增加，同時IIA及IIX纖維面積減少。橫紋肌肉瘤是一群未分化的增生肌肉細胞，肌抑素雖能抑制肌肉生長及分化，但對此疾病的作用仍無定論。肌抑素不單影響肌肉生長，亦影響身體脂肪量。肌抑素剔除小鼠身體的脂肪堆積減少，有可能是身為脂肪燃燒器的肌肉量增加所致。全身性攝取肌抑素造成白脂肪組織量降低，可能係透過和BMP7競爭第二型受器而抑制其作用，進而抑制脂肪生成。 測量血漿中肌抑素濃度可作為肌抑素基因表現之指標。目前有零星報告利用放射免疫分析法(RIA)定量血漿中肌抑素濃度，發現後天免疫不全症候群造成的肌肉萎縮病患，其肌抑素濃度增加且全身肌肉量下降；在臥床的靜止病人其血中肌抑素濃度亦會增加。然放射線物質的使用會造成環境及職業安全的顧慮，廢棄物的處理較為困難且成本高昂，使得RIA的應用受到限制。另一方面，由於酵素免疫分析無需進行肌肉切片以獲取檢體，又不具放射線廢棄物處理問題，其經濟性及可及性將使肌肉相關疾病之臨床試驗更方便進行。過去已有不少研究利用酵素免疫法分析血中肌抑素濃度，然由於抗原抗體種類及特異度在各實驗間皆不同，其結果歧異甚大，不易比較；亟需選用高特異性抗原抗體對，自行開發酵素免疫分析系統以測量血漿中之肌抑素濃度。 組蛋白去乙醯基酶抑制劑(Histone deacetylase inhibitor, HDACI)可促成組蛋白乙醯基化，打開纏繞的染色絲，進而活化基因。在肌肉組織的作用上，可以增加肌母細胞融合，進而增加肌纖維截面積。本實驗室發現HDACI—包括TSA及valproic acid --可有效活化已分化肌肉細胞之肌抑素表現，增加mRNA表現量達40倍。這是目前已知對肌抑素調控最強的藥物，然作用機轉仍不清楚。 本研究目的在(1) 分析TSA活化肌抑素的機轉及所牽涉的訊息傳遞鏈，進而找出訊息傳遞鍊中合適的分子以調控肌抑素，未來可用於了解肌肉生物學及治療肌肉相關疾病。(2) 嘗試開發以ELISA測量血漿中之肌抑素濃度，並進行臨床試驗，測量血液透析病患及接受酵素補充治療的龐貝氏症病患之血中肌抑素濃度，確立肌抑素為肌肉生長/分化之生物指標。 本研究提出假說：(1) TSA係透過MAPK及Akt/mTOR路徑活化肌抑素，若抑制上述兩路徑，則肌抑素之表現亦會降低；若直接活化上述兩路徑，則肌抑素亦會被直接活化。(2) 血清肌抑素濃度和肌力/肌肉量相關，使用高通量透析管病患較使用低通量透析管者有較低的肌抑素濃度，因此擁有較大肌力/肌肉量。龐貝氏症患者接受酵素補充療法前，由於肌肉萎縮，其血中肌抑素及IGF-1濃度較同年齡對照組為低；接受酵素補充療法後，血中肌抑素及IGF-1濃度將明顯增加，可作為追蹤療效的指標分子。 肌抑素活化機轉研究以小鼠C2C12肌母細胞株為模式，含10%胎牛血清之DMEM (Dulbecco’s modified Eagle’s medium)培養基為生長培養基，加入2%馬血清之DMEM為分化培養基。分化4天後，先以p38 MAPK 抑制劑SB203580、ERK抑制劑PD98059、JNK抑制劑SP600125、PI3K抑制劑PIK-75、Akt抑制劑LY294002及AKT INH VIII、或mTOR抑制劑rapamycin作一小時的前處理，再加入50nM TSA處理24小時。收集mRNA並以Q-PCR定量。RNA合成酶抑制劑(Actinomycin D)以前處理方式加入分化之細胞以驗證轉錄階層控制。p38 MAPK及JNK之活化劑(anisomycin, 200ng/mL)則用以代替TSA誘發肌抑素之生成。同時收集細胞萃取液，以西方墨點法驗證MAPK及PI3K/Akt/mTOR路徑中成員被磷酸化之比例增加。為求更專一的基因抑制，利用穩定表現p38 MAPK、Akt1、Akt2專一shRNA的C2C12肌母細胞先予以分化三天，再加入TSA處理，分析肌抑素基因的表現量是否被抑制。結果顯示，經TSA處理24小時之C2C12肌母細胞其肌抑素mRNA表現增加達40倍，給予actinomycin D作為前處理可減少肌抑素之誘發93%，證明此過程係透過轉錄階層控制。若給予p38 MAPK、JNK、PI3K、Akt及mTOR抑制劑作前處理，則分別可降低肌抑素誘發達72%、43%、66%、82%及90%；而ERK抑制劑則無此效應。若利用RNA干擾技術抑制p38 MAPK、Akt1或Akt2表現，則可在C2C12肌母細胞中分別降低肌抑素誘發達77%、56%及82%。利用西方墨點法發現經TSA處理之C2C12肌母細胞其磷酸化p38 MAPK、JNK、Akt及S6K蛋白較未處理組增加。若直接以anisomycin活化p38 MAPK及JNK，可誘發肌抑素表現達4倍。至於p38 MAPK及JNK的上游分子，包括MKK3/4/6及ASK1，其磷酸化蛋白亦在TSA處理後增加。我們認為： 經TSA處理24小時之C2C12肌母細胞其肌抑素mRNA表現增加主要透過ASK1-MKK3/6-p38 MAPK、ASK1-MKK4-JNK及PI3K/Akt/mTOR三條路徑達成，p38 MAPK、JNK及Akt的活化是肌抑素誘發過程中必要但不充分之要素。 在血液透析研究部分，收錄41位健康對照組，及使用高通量與低通量透析管血液透析超過半年之慢性腎衰竭病患各29及31人。健康對照組接受抽血及以生物阻抗分析儀分析身體組成；腎衰竭病患則在透析前、後分別接受抽血，並於透析後測量身體組成。此外每位受試者都接受慣用手最大握力測量。血清中肌抑素濃度使用本實驗室自行開發之ELISA進行測量。IGF-1濃度則依據ELISA製造商之建議流程量測。肌抑素濃度與其他參數間關係以皮爾森相關分析檢定。線性迴歸分析則用以找出肌抑素及握力之獨立決定因子。二項邏輯斯迴歸用以決定低握力之勝算比。結果顯示，與健康對照組相比，接受透析病患的身體質量指數較低、血中IGF-1濃度較低、握力亦較低。與使用低通量透析管的病患相比，使用高通量透析管的病患其握力較佳(25.5 vs. 19.2 kg)、洗前血中肌抑素濃度較低(31.0 vs. 18.5 μg/ml)。但在透析後兩者濃度截然不同：高通量透析管透析後降低肌抑素濃度36%，然而低通量透析管反而增加25%。在線性迴歸分析中，肌抑素濃度與年齡以及使用高通量透析管呈負相關；握力則和年齡、女性、肌肉量、肌抑素濃度、及洗腎呈負相關，而與使用高通量透析管呈正相關。若以二項邏輯斯迴歸分析，在校正年齡、性別、肌肉量、透析與否和使用透析管種類後，血中肌抑素濃度高於平均者，其握力有7.6倍機會將低於平均值。我們認為：透析管型式可調控血中肌抑素濃度，血中肌抑素濃度高則肌肉功能較差。肌抑素在其他臨床狀況所扮演角色值得進一步探究。 在龐貝氏症研究部分，收錄16位龐貝氏症之病童，分別在給予酵素補充治療(ERT)前與治療一年後進行抽血及肌肉切片，平均治療追蹤期間為11.7月(全距: 6-23月)。病患組成包括有治療前後完整樣本的6位嬰兒期發病及4位較晚期發病病患，以及僅有治療後樣本的6位嬰兒期發病病患。對照組取自無肌肉疾病、年齡及性別相符之新生兒篩檢個案。分別以ELISA試劑測量血清肌抑素、IGF-1及卵泡抑制素濃度，病患粗動作及精細動作之發展則以皮巴迪動作發展量表評估。肌肉切片取自股四頭肌，由一位具經驗之病理科醫師於高倍光學顯微鏡下觀察，並計算具有細胞肝醣液泡之肌細胞比率。病童及對照組兩組間之人口學、生化學、病理學及臨床變項平均值以Wilcoxon rank-sum檢定，ERT前後之血清指標均值則以Wilcoxon signed-rank檢定。結果發現與對照組相比，龐貝氏症患者在接受酵素補充治療前血中肌抑素及IGF-1濃度較低。但在接受治療後，肌抑素、IGF-1及卵泡抑制素濃度分別上升129%、74%及62%，與對照組無異。同時在治療後，肌纖維內有液泡百分比降低60%、皮巴迪動作發展量表商數降低(95.2降至79.5)。我們認為：血中肌抑素及IGF-1濃度可反映肌肉再生，它們能作為神經肌肉疾病療效之生物指標。 本研究之主要貢獻包括：(1) 提出了嶄新的肌抑素調控訊息路徑。不同於過去研究著重於肌抑素下游訊息(像是ActRIIB、Smad、MEF、atrogin等)，此研究發現MAPK中的P38 MAPK及JNK路徑，以及PI3K/Akt/mTOR路徑與肌抑素的活化有關。(2) 提出肌抑素可作為臨床疾病(接受血液透析病患及龐貝氏症)肌肉生長/分化之生物指標。 未來藉由抑制肌抑素，可活化衛星細胞而增加肌肉再生，進而改善肌肉相關疾病。利用個別路徑的抑制劑(包括SB203580、LY294002、PD98059、PIK-75、rapamycin、AKT INH VIII、shRNA質體等)可以部分抑制肌抑素的活化，而HDACI、anisomycin、IGF-1等則可活化肌抑素。完整個體中除了已分化肌細胞外，尚有不同階段的肌母細胞及衛星細胞，故有待動物或人體in vivo研究證實這些藥物的效用。另一方面，肌肉質量在許多疾病模式中都可預測罹病率或死亡率，因此未來除了在其他肌肉相關疾病中，驗證肌抑素是否可作為肌肉量及功能之指標之外，還需設計大樣本的縱貫性研究來探究肌抑素與罹病率或死亡率間之關聯，以確認血中肌抑素之臨床應用價值。

Introduction Myostatin (growth and differentiation factor 8, GDF8), a member of transforming growth factor beta family, is a potent negative regulator of muscle growth and differentiation. Myostatin knock-out mice have 3 times larger muscle mass and reduced adiposity than wild type. On the other hand, increased secretion of myostatin protein systemically causes cachexia-like phenotype. Human myostatin gene is located on 2q32.2, and its mutation causes increased muscularity. Myostatin mutation also causes double-muscle phenotype in cattle, sheep, dog, and zebrafish. Myostatin peptide sequence is conserved among vertebrates. The amino acid sequence of the active C-terminal peptide is identical in rodents, chicken, turkey, pig, and human. The translated product, prepropeptide, is composed of 375 amino acids. It becomes full-length myostatin after removing the 21-amino-acid signal peptide. The full-length myostatin is cleaved by peptidase into N-terminal propeptide and C-terminal 110-amino-acid active peptide. The C-terminal active peptide binds non-covalently with propeptide, follistatin, and product of follistatin-related gene to become latent. The myostatin acts through autocrine and endocrine fashion. It binds with activin type II receptor on the membrane, and phosphorylates intracellular transcription factors-- Smad2 and Smad3. Since myostatin can inhibit muscle growth, it may have therapeutic role in the muscle related diseases and metabolic syndrome. The muscle related diseases include neuromuscular diseases (eg. muscular dystrophy, myopathy), sarcopenia, cachexia caused by metabolic diseases or malignancy, muscle atrophy caused by denervation or disuse/immobilization, and rhabdomyosarcoma. These diseases have different etiology, but the common final presentation is muscle atrophy. In 2002, two groups of researchers applied different methods to inhibit myostatin in mdx mice, a rodent model of Duchenne and Becker muscular dystrophy, and found improved symptoms and function in these animals. Minetti injected trichostatin A (TSA) intraperitoneally into mdx mice, and also found improved symptoms and decreased myostatin expression. Inhibition of myostatin in mouse model of amyotrophic lateral sclerosis improved pathological characteristics in the diaphragm and increased muscle power and mass temporally and significantly, but did not lengthen survival. There are still debates on the expression of myostatin during ageing. Cachexia is the clinical condition of losing muscle and fat tissue, caused by severe trauma, malignancy, infection, or other medical diseases. The decrease of muscle mass can reach 75% in patients with malignancy, which will cause weakness, fatigue, disuse, and death. The typical pathological manifestation is smaller muscle fiber diameter. Cachexia is also a poor prognostic factor, which increases medical expenses and demand for rehabilitation. The secondary muscle atrophy usually incurred major disability and direct or indirect socioeconomic loss, even if the primary disease is cured. The mainstream of treatment for cachexia is rehabilitation, electric stimulation, or palliative therapy. Myostatin modulation might shed light for the treatment of cachexia. The relationship between cachexia and myostatin has been explored in chronic renal failure, heart failure, AIDS, COPD, Addison’s disease, Cushing’s syndrome, liver cirrhosis, cancer, calorie restriction, and burn. Generally speaking, the myostatin mRNA expression increased during the process of muscle wasting; however, its expression in chronic stage was still unclear. On the first day after sciatic neurectomy, the myostatin expression in gastrocnemius of rat increased 31%. It decreased 34% 14 days after operation, when the muscle size decreased 50%. The change of myostatin protein level in gastrocnemius was similar to that of mRNA. The results implied that the mechanism governing denervation-induced muscle atrophy was complex. The effect of myostatin might be variable in the different time frame. After hind limb suspension in rat for 10 days, the myostatin mRNA and protein increased 110% and 37%, and the muscle mass decreased to 84%. Those parameters normalized after reload for 4 days. Reload by walking for 30 minutes every day could prevent muscle atrophy caused by high-limb suspension, but could not prevent myostatin elevation. This suggested that increase of myostatin would not cause muscle atrophy in healthy adult rat. Myostatin mRNA expression increased and the area of type IIA and IIX fiber decreased in the patients having hip joint osteoarthritis. Rhabdomyosarcoma is the overgrowth of the undifferentiated muscle cells. The effect of myostatin on rhabdomyosarcoma is still in debate. Myostatin not only inhibited muscle growth, but also affected adipose tissue. Myostatin knock-out mice had decreased fat deposit. The cause might be the increased fat burner—muscle. On the other hand, systemic administration of myostatin caused decrease of white adipose tissue. The possible reason is that myostatin competes with ActRIIB for BMP7. Decreased BMP7 expression then inhibits lipogenesis. Serum myostatin level related to the intramuscular expression of this gene. Some reports using radioimmunoassay (RIA) as a measurement for serum myostatin showed that the level of serum myostatin was inversely related with the total skeletal muscle mass in patients having acquired immunodeficiency syndrome. The serum myostatin level also increased in those bed-ridden patients. Due to environmental and occupational issues on radioactive substance, the application of RIA is limited in many ways. On the other hand, enzyme-linked immunosorbent assay (ELISA) is widely used because of its high accessibility and economy, and no radioactive waste. Some studies employed ELISA to quantify myostatin, but the discrepancy between them is obvious. Using antibody-antigen pair with high specificity is crucial to ELISA. Histone deacetylase inhibitor (HDACI) facilitates the acetylation of histone, unwinds packed chromatin, and activates related genes. It also increases myoblast fusion, and myofiber cross-section area. Our lab found the HDACI, including trichostatin A (TSA) and valproic acid, could activate the expression of myostatin over 40 times in differentiated mouse myoblast than the untreated cells. However, the mechanism is still unclear. Our specific aims are to elucidate the mechanism and involving signal pathways governing myostatin expression in order to found suitable modulators to treat muscle related diseases, and to establish ELISA system for serum myostatin to investigate whether it could be a new biomarker for systemic muscle atrophy and regeneration in chronic renal failure and Pompe disease. There are two hypotheses. First, TSA activates myostatin through MAPK and Akt/mTOR pathway, and we can decrease myostatin expression by inhibiting those pathways. Besides, serum myostatin level is related to muscle strength/mass. Patients using high-flux dialyzer will have lower myostatin level and thus higher muscle strength/mass than those using low-flux dialyzer. Before enzyme replacement therapy (ERT), Pompe disease patients will have lower myostatin and IGF-1 level due to muscle atrophy. After ERT, the level of those serum markers will increase. Material and Method We used differentiated mouse C2C12 myoblast as a model to measure the expression of myostatin in different treatment condition with quantitative real-time PCR. The Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal calf serum was the maintaining media, and the DMEM containing 2% horse serum was the differentiation media. After 4-day differentiation, the C2C12 myotubes were pretreated with p38MAPK inhibitor (SB203580), ERK inhibitor (PD98059), JNK inhibitor (SP600125), PI3K inhibitor (PIK-75), Akt inhibitor (LY294002 and AKT INH VIII), or mTOR inhibitor (rapamycin) for 1 hour, and subsequently treated with 50nM TSA for 24 hours. The RNA polymerase inhibitor (Actinomycin D) was added as pretreatment to verify transcriptional level control. Besides, p38 MAPK and JNK activator (anisomycin) will be added to induce myostatin expression without TSA. The phosphorylated form of the components in the two signal pathways were quantified with western blot. In order to inhibit respective pathway more specifically, we employed C2C12 stably transfected with short hairpin RNA plasmids aimed at p38MAPK, Akt1, and Akt2 to see if the induction of myostatin would be inhibited. In the study of chronic renal failure, we recruited 41 healthy controls and 60 patients receiving maintenance hemodialysis (MHD) for over half year. Twenty-nine of them used high-flux dialyzer, and 31 used low-flux dialyzer. The healthy controls received blood sampling, and body composition determination with multi-channel bioimpedance analyzer. The hemodialysis patients accepted blood sampling before and after single episode of hemodialysis session, and body composition determination. Dominant hand grip strength was measured in all subjects. The serum myostatin level was determined by home-made ELISA kit. IGF-1 was measured with commercial ELISA kit. Pearson correlation analysis was used to find the relationship between myostatin and other parameters. The linear regression analysis was applied to find the major determinants of myostatin and grip strength. Binary logistic regression was employed to find the odds ratio of low grip strength. In the study of Pompe disease, we recruited 16 patients with Pompe disease and 16 sex-, age-matched neonate screening subjects as control. We performed quadriceps muscle biopsy and blood sampling to measure the myostatin, follistatin, creatine kinase, and IGF-1 level to monitor the change of serum markers before and one year after ERT. The median duration of follow-up was 11.7 months (range: 6-23 months). There were 10 patients having complete muscle and blood samples. Six were infant-onset Pompe disease, and 4 were late-onset Pompe disease. The gross and fine motor development was evaluated with Peabody Developmental Motor Scale (PDMS). The muscle sample was biopsied, processed and stained, and was evaluated by an experienced pathologist. The percentage of intramuscular vacuoles was calculated. The test of mean between two variables was performed by Wilcoxon rank-sum test nonparametrically. The pre- and post-ERT comparison of serum markers was performed with Wilcoxon signed-rank test. Results TSA increased myostatin mRNA expression up to 40-fold after treatment for 24 hours. Pretreatment with actinomycin D reduced the TSA-induced myostatin mRNA by 93%, suggesting TSA induced myostatin expression mainly at the transcriptional level. Pretreatment with p38 MAPK and JNK inhibitors, but not ERK inhibitor, blocked TSA-induced myostatin expression respectively by 72% and 43%. Knockdown of p38MAPK by RNAi inhibited the TSA-induced myostatin expression by 77% in C2C12 myoblasts. The protein levels of phosphorylated p38 MAPK, JNK, but not ERK, increased with TSA treatment in differentiated C2C12 cells. Direct activation of p38 MAPK and JNK by anisomycin in the absence of TSA increased myostatin mRNA by 4-fold. The phosphorylated form of the kinase MKK3/4/6 and ASK1, upstream cascades of p38 MAPK and JNK, also increased with TSA treatment. We concluded that the induction of myostatin by TSA treatment in differentiated C2C12 cells is in part through ASK1-MKK3/6-p38 MAPK and ASK1-MKK4-JNK signaling pathways. Activation of p38 MAPK and JNK axis is necessary, but not sufficient for TSA-induced myostatin expression. We also explored the role of the PI3K-Akt-mTOR axis in myostatin induction. We confirmed that phosphorylated Akt was induced after TSA treatment. Pretreatment with respective PI3K, Akt, and mTOR inhibitors partially blocked the TSA-induced myostatin expression 66%, 82%, and 90%, respectively. The shRNA aimed to knock-down Akt1 and Akt2 also inhibited TSA-induced myostatin expression 56% and 82%, suggesting that TSA-induced transcription activation of myostatin is in part through Akt pathway. The MHD patients had lower body mass index, IGF-1 level, and grip strength than the normal controls. The patients using the high-flux dialyzer had better grip strength than those using the low-flux (25.5 vs. 19.2 kg). The pre-dialysis myostatin level was higher in low-flux dialyzer than high-flux (31.0 vs. 18.5 μg/ml). Interestingly, the high-flux dialyzer reduced the serum myostatin by 36%, whereas low-flux dialyzer increased it by 25%. The myostatin was inversely related to age and the use of high-flux dialyzer. Furthermore, the grip strength was negatively related to age, female gender, muscle mass, myostatin levels and hemodialysis, but positively to the use of high-flux dialyzer in linear regression. The risk of low grip strength was 7.6 times higher in those with higher serum myostatin with the adjustment of age, gender, muscle mass, hemodialysis and mode of dialysis in a logistic regression. The mode of dialyzer modulates the blood levels of myostatin. Higher myostatin is associated with lower muscle function. The use of myostatin assay in various clinical settings merits further investigation. Pompe disease patients had lower serum myostatin and IGF-1 levels than controls before ERT. However, myostatin, IGF-1, and follistatin levels increased 129%, 74%, and 62% after ERT respectively, and have fell into normal ranges. In contrast, these were not changed in the controls during follow-up. At the same time, the percentage of muscle fibers having intracytoplasmic vacuoles decreased 60%, and the quotient of PDMS declined significantly from 95.2 to 79.5. The increase in myostatin and IGF-1 levels may reflect muscle regeneration. They are potential therapeutic biomarkers for Pompe disease. Discussion The main contribution of this study included revealing novel signaling pathways modulating myostatin, and establishing myostatin as a biomarker for muscle growth/differentiation in different clinical scenarios. The p38MAPK and JNK in MAPK pathway, and PI3K/Akt/mTOR axis are involved in the activation of myostatin. Dissection of these pathways may help design new therapeutic modality for the treatment of muscle related diseases. We may treat muscle related diseases through myostatin inhibition, satellite cell activation, and muscle hypertrophy. The small molecules including SB203580, LY294002, PD98059, PIK-75, rapamycin, AKT inhibitor VIII, and shRNA plasmids will be employed to inhibit myostatin. On the other hand, HDACI (eg. TSA and valproic acid), anisomycin, and IGF-1 can activate myostatin. There are differentiated myotubes, undifferentiated myoblasts, and quiescent satellite cells at different stages in the whole body. We should perform experiment in animal or human model to establish the role of these molecules in the complex body. Besides, since muscle mass can predict morbidity and mortality in many diseases, we should perform large scale longitudinal studies to see the correlation between myostatin level and morbidity and mortality. In the future, we will collect the blood samples and anthropometric parameters from elite sportsmen performing weight training, frail patients performing strength training, patients performing acute bout exercise and chronic training, and stroke patients performing rehabilitation. The correlation between serum myostatin and body composition will be sought. We hope to establish the serum myostatin as a new biomarker of systemic muscle atrophy and regeneration.