使用命運追蹤方式研究間皮細胞，間皮下纖維母細胞，及纖維球細胞在腹膜纖維化過程的變化

Abstract

慢性腎臟病病患近年來在台灣逐年增加，而多數慢性腎臟病患者最後會進展至末期腎病，須藉由進行透析將體內產生的毒素及多餘的水分排出。目前的透析模式有腹膜透析和血液透析兩種。利用腹膜透析的患者約占台灣末期腎病病患中約10-15%。腹膜透析患者因長期使用高葡萄糖透析液，使得腹膜長期浸泡在高濃度葡萄糖溶液中，促使氧化因子GDPs (Glucose degradation products) 和AGEs (Advanced glycated end-products) 產生，使得腹膜通透性發生改變，若加上反覆的腹膜炎，都會促使腹膜纖維化的發生。雖然高葡萄糖透析液與反覆腹膜炎是最可能的致病因素，然而包囊性腹膜硬化症(encapsulating peritoneal sclerosis)的確實病因並不十分清楚。正常腹膜是由單一層細胞的間皮細胞(mesothelial cells)，基底膜(basement membrane)，和間皮下組織(submesothelial layer)等組成。間皮細胞是由中胚層發育而來具有上皮及間質細胞的特性。間皮下組織內含間皮下纖維母細胞(submesothelial fibroblasts)，微血管，淋巴球及結締組織。腹膜發生纖維化過程為喪失間皮細胞，間皮下組織增生纖維化，血管新生以致血管通透性增加。臨床上患者會出現腹膜超過濾率降低 (ultrafiltration failure)，腸阻塞，營養不良，嚴重者腹腔內器官整個被包裹沾黏在一起。造成腹膜纖維化的主要細胞是肌纖維母細胞，肌纖維母細胞的來源一直有爭議。目前最廣為接受的理論是間皮細胞受到葡萄糖透析液或感染等的刺激，使得間皮細胞轉分化為肌纖維母細胞，即所謂的上皮-間質細胞轉分化(epithelial-mesenchymal transition，EMT)。且間皮細胞進入間皮下組織轉變為肌纖維母細胞的過程為轉分化transdifferentiation。但有關這部分的動物實驗都是間接性證據，沒有直接動物實驗的結果。 我們利用基因改造小鼠與譜系追蹤技術發現小鼠的腹膜在損傷後會引發腹膜間皮細胞與腹膜間皮下纖維母細胞產生截然不同的命運。在餵食Col1a2-CreERTTg;Rosa26fstdTomato/+小鼠tamoxifen以誘導其腹膜間皮下纖維母細胞表現出紅色螢光蛋白(RFP)以作為譜系標誌之後，不論是利用次氯酸鈉、高糖腹膜透析液、或腺病毒表現乙型轉分化生長因子來傷害腹膜，我們都證實腹膜間皮下纖維母細胞是腹膜肌肉纖維母細胞的主要來源。而在WT1CreERT2/+;Rosa26fstdTomato小鼠利用相同的譜系追蹤技術研究則發現WT1-RFP+腹膜間皮細胞則主要負責修補受傷害的間皮層。但我們同時發現一小部份的WT1-RFP+細胞位於腹膜基底膜之下，也會表現第一型膠原蛋白，但並不表現cytokeratin；而且受傷增厚的間皮下層結疤組織中則有少數的肌肉纖維母細胞源自WT1-RFP+細胞。因此WT1促進子活化的CreERT2重組脢並無法專一性地活化間皮細胞的標記螢光蛋白或修改間皮細胞的表現基因。因此腹膜間皮下纖維母細胞雖是腹膜肌肉纖維母細胞的主要來源。我們也觀察到當腹膜受傷時是由存活下來的間皮細胞進行修補。 亦有其他研究指出來自骨髓的纖維細胞會在腹膜纖維化中轉變為肌纖維母細胞，為更瞭解肌纖維母細胞的來源，我利用parabiosis及骨髓移植技術將collagen I(α1)-GFP基因轉植小鼠的骨髓細胞移植到野生型小鼠，誘發腹膜纖維化後發現，受傷的腹膜上僅有少許綠色螢光(GFP)細胞且同時表現骨髓標記因子CD45，表示這些綠色螢光細胞是由骨髓來的纖維細胞，但藉由螢光染色發現這些來自骨髓的纖維細胞在腹膜纖維過程中並沒有表現肌纖維母細胞特定因子(α-smooth muscle actin ; αSMA)。此同時我也發現有許多巨噬細胞在受傷的腹膜上，我再利用基因譜系追蹤研究，並且將巨噬細胞去除以瞭解它在腹膜纖維化過程的角色，發現將巨噬細胞去除時，原本因纖維化而增厚的腹膜厚度不僅減少，連肌纖維母細胞的數量也減少了。此表示巨噬細胞對於間皮下纖維母細胞轉變為肌纖維母細胞的過程有特定的作用，這些仍是我需要再進一步研究瞭解的。 經由這些腹膜纖維化研究發現，間皮下纖維母細胞是肌纖維母細胞的主要來源，非來自骨髓的纖維細胞或間皮細胞。而巨噬細胞在間皮下纖維母細胞轉變為肌纖維母細胞過程中扮有一定角色，可能透過釋放細胞激素或因子作用在間皮下纖維母細胞。然而當腹膜受傷時，是透過存活的間皮細胞所修復的。

Peritoneal dialysis (PD) is one modality of renal replacement therapy for end-stage renal disease. Preservation of peritoneal membrane function is important for peritoneal dialysis. However, there are many challenges for peritoneal dialysis, such as peritonitis and peritoneal fibrosis. Peritoneal fibrosis means adhesion, calcification and fibrosis between peritoneal contents, leads to ultrafiltration failure and loss peritoneal function, then finally cocooning intra-abdominal organs. The severe type of peritoneal fibrosis is encapsulating peritoneal sclerosis (EPS). There is high mortality for EPS due to ileus, malnutrition, and intestinal obstruction. The risk factors to predisposing peritoneal fibrosis are peritonitis, high glucose PD solution, and glucose degradation products (GDPs), etc. Normal peritoneum is composed of single layer of mesothelial cells, basement membrane, and submesothelial layer. There are submesothelial fibroblasts, capillaries, lymphocytes, vessels and connective tissue in submesothelial layer. Peritoneal fibrosis is characterized by peritoneal thickening, and progressive fibrosis, which can be seen as with thickening of basal lamina and accumulation of α-smooth muscle actin (αSMA)+ myofibroblasts. There are three origins of myofibroblasts in peritoneal fibrosis, like mesothelial cells, submesothelial fibroblasts, and circulating fibrocytes. The major theory is epithelial-mesenchymal transition (EMT). Mesothelial cells are mesodermal origin and have both epithelial and mesenchymal characteristics. EMT means that mesothelial cells transform to myofibroblasts and/then migrate into submesothelial layer after injury. But there are most in vitro studies to support EMT and in vivo study is lacking. We used tamoxifen-inducible Cre/Lox techniques to genetically label and fate map mesothelial cells and submesothelial fibroblasts. We used sodium hypochlorite, TGF-β1, 4.25% PD solution plus methylglyoxal (MGO) to induce mice models of peritoneal fibrosis. After pulse labeling induced by tamoxifen, the cell number of Col1α2-tdTomato+ submesothelial fibroblasts increased markedly in fibrotic peritoneum. And more than 60% of Col1α2-tdTomato+ submesothelial fibroblasts are αSMA positive. However, WT1-tdTomato+ mesothelial cells are cytokeratin positive, Ki67 positive but αSMA negative in injured peritoneum. We also found surviving Ki67+ mesothelial cells repair injured peritoneum. For studying roles of circulating fibrocytes during peritoneal fibrosis, we used bone marrow transplantation and parabiosis to create chimeric mice. The bone marrow cells of Collagen I (α)1-GFP transgenic mice were transferred to wild type C57BL/6 mice. We induced peritoneal fibrosis models by sodium hypochlorite. In immunofluorescence stain, only few Col1a1-GFP+ fibrocytes expressing CD45, not αSMA, were seen in fibrotic peritoneum, however, many Col1a1-GFP-;CD11b+ cells were seen. Thereafter we also used tamoxifen-inducible Cre/Lox techniques to genetically label and fate map monocytes/macrophages. After pulse labeling induced by tamoxifen, the cell number of Csf1r-tdTomato+ macrophages increased markedly in fibrotic peritoneum. Ablation of macrophages in vivo in Csf1r-Cre/Esr1;tdTomato;iDTR mice were performed by diphtheria toxin during peritoneal fibrosis. The thickness of injured peritoneum and the number of myofibroblasts were decreased markedly under macrophage depletion. Thereafter, we conclude that submesothelial fibroblasts are major origin of αSMA+ myofibroblasts during peritoneal injury. Few circulating fibrocytes are recruited into injured peritoneum but do not transform into myofibroblasts directly. The significance of recruited fibrocytes in peritoneal fibrosis needs further study. Macrophages may cross talk with submesothelial fibroblasts that will transit to myofibroblasts during injury. And injured peritoneum is repaired by surviving mesothelial cells.