- 學位論文
脂肪細胞發炎之指標於代謝疾病與末期腎病之變化
The Role of Biomarkers Derived from Adipocyte Inflammation in Metabolic Health and End-stage Renal Disease
摘要
台灣肥胖的盛行率在過去20年間逐漸增加，伴隨著人口高齡化，肥胖成為孕育代謝疾病，如糖尿病、代謝症候群、非酒精性脂肪肝、慢性腎病，心血管疾病與多項癌症之溫床。當身體處於營養或能量過剩之狀態，脂肪組織所出現的變化包括脂肪細胞肥大，巨噬細胞浸潤增加且轉為以M1分類占上風。脂肪組織除了負責以三酸甘油脂之形式儲存身體過剩的營養外，也是一個具有內分泌功能的器官，透過分泌游離脂肪酸及多種脂肪激素，而可影響脂肪組織自身及其他負責能量攝取或代謝之器官，如下視丘、肝臟、骨骼肌等，藉此以維持身體能量的恆定。肥胖時，脂肪組織及脂肪細胞發生上述變化時，脂肪激素之分泌也隨之改變。 本研究利用小鼠3T3-L1前脂肪細胞培養，引導其分化為成熟脂肪細胞後，加入IL-1β引發脂肪細胞之發炎及胰島素阻抗作為模擬脂肪細胞發炎之模型。加入IL-1β之脂肪細胞作為實驗組及未做處理之細胞作為對照組，兩組細胞以微陣列實驗檢驗基因表現，並分析兩組間基因表現之差異。相較對照組，在實驗組中基因表現顯著上升並達六倍以上之44個基因，被挑選為脂肪發炎生物指標之候選者。這些候選基因之表現利用定量即時聚合酶鏈鎖反應加以驗證，而排除2個相對表現量在實驗中並未顯著上升之基因。 候選的42個基因利用STRING資料庫加以分析建立一蛋白質-蛋白質關係之網路。此網路中有267對蛋白質-蛋白質關聯，其中以防禦機轉、細胞激素、發炎反應、化學趨向之路徑聚集最為顯著，而兩條與脂質相關之路徑也呈現基因顯著集中之現象。LCN2及SLPI出現於脂質相關途徑，且皆與重要的脂肪激素IL-6間有中度的關聯性，而另一基因LRG1，與上述LCN2及SLPI皆在STRING資料庫預測有相關性。在K-means群聚分析中，SLPI及LRG1持續出現於同一亞群。因此，本研究挑選LCN2（其蛋白質產物為NGAL），SLPI及LRG1為有潛力之生物指標，可作為脂肪發炎之代表，並藉由兩不同特色的臨床群組，進一步探討此三指標之臨床意義。 第一個臨床案例群組共收案175位女性受試者，他們皆無心血管疾病或糖尿病。這群女性受試者於收案時接受健康史調查，血糖血脂等代謝指標檢驗，包括75克葡萄糖耐受試驗。此外他們亦接受雙能量X光吸光式測定儀測定身體組成。此小群之研究目的為檢驗脂肪細胞所衍生之指標與代謝異常及身體組成之相關性。在這個群組中，超過半數女性腰圍超過80公分之標準，17.1%受試者達到糖尿病之診斷標準，27.4%和代謝症候群。這個橫斷性的研究顯示，不同的代謝異常中身體組成之型態其實有個別特徵。在肥胖或中心肥胖者，全身脂肪量皆同步增加，而在代謝症候群或胰島素阻抗者，脂肪增加只發生於上臂及中心區域。在三個脂肪細胞指標中，NGAL與肥胖，中心肥胖及異常的葡萄糖代謝，如糖尿病及胰島素阻抗皆呈顯著相關。NGAL亦與全身的脂肪量以及脂肪中心分佈之型態呈現正相關。再者，本群組中有140位女性之身體質量指數並不超過27 kg/m2，根據2017年提出的代謝風險診斷標準，將這群未達肥胖標準的女性，區分為代謝健康及代謝不健康兩組。此診斷標準之核心在除了葡萄糖，血壓，血脂代謝異常之外，加入胰島素阻抗以及隱性發炎兩項，及排除代謝症候群的腰圍標準。在這群未達肥胖的女性中，代謝不健康者亦有特有之身體組成分佈。當與健康者相比，不健康者雖有較高的身體質量指數，但其下肢與骨盆區域之肥肉瘦肉組織量皆未隨之增加；更重要的是不健康者之脂肪中心分佈程度與肥胖者並無顯著差異，顯示未達肥胖但代謝病不健康者其脂肪分布特別集中於身體中央，而在周邊部位未見增加。而在代謝不健康的非肥胖女性，SLPI呈現顯著的增加，且此關聯不受年紀或中心肥胖之影響。 第二群臨床案例探討末期腎病接受慢性血液透析治療的169為患者，其中45.3%患有糖尿病。這群患者自2016年4月收案並接受追蹤至2019年10月。於收案時，患者接受過去病史詢問，生化血液檢驗及脂肪細胞指標。橫斷式分析顯示LRG1與患者罹有周邊動脈疾病及缺血性中風有顯著相關，且LRG1對周邊動脈疾病在調整共病因子後仍呈正相關。而三項脂肪細胞指標與患者之身體質量指數或代謝指標皆無關聯。38個月的追蹤期間中，34位患者死亡，14位歸因於心血管疾病，其餘20位歸因於非心血管疾病包括腸胃道出血，肝臟疾病及感染。在這群末期腎病患者中，收案時較高齡，較低的血紅素與白蛋白，較高的鹼性磷酸酶及肝指數，罹有糖尿病，冠狀動脈疾病，週邊動脈疾病，缺血性中風，冠狀動脈繞道手術，癌症及肝硬化皆與較高的死亡風險相關。而脂肪指標中，LRG1與總體死亡風險呈現正相關，SLPI及NGAL則與死亡風險呈現負相關。而將死亡區分為心血管事件及非心血管事件兩類後，心血管事件並未與任一脂肪細胞指標有相關，顯示前述橫斷性研究LRG1與大血管疾病之相關性並未影響患者之存活。而針對非心血管事件造成之死亡，三個指標皆與死亡風險呈現與前述總體死亡率相同之趨勢。而在校正影響死亡的共病因子之後，唯有SLPI與較低的死亡率仍保有顯著的相關性。 本研究利用模擬脂肪細胞發炎的細胞模型，以微陣列實驗分析脂肪細胞發炎所誘發之基因表現差異後，挑選出轉錄顯著上升超過6倍之基因後，將候選基因利用STRING資料庫建成蛋白質關聯之網路，挑選出有潛力作為脂肪細胞發炎指標之三個基因。臨床分析顯示不同的代謝異常其身體組成及生物指標之表現皆有個別特色，因此在面對肥胖等代謝疾病，體內所存在之全身性發炎及多重器官影響，本研究利用系統生物學工具來分析疾病表現以篩選診斷甚或診斷的標的，在未來疾病的診治必定會扮演更要的角色。而在本研究所發現生物指標與不同代謝表現型之相關，該指標是否參與代謝異常之病生理機轉及是否能監測疾病程度或預後的指標，有賴後續之追蹤。
並列摘要
Metabolic health is closely linked with risk of cardiovascular disease, diabetes mellitus (DM), metabolic syndrome (MS) and many types of cancers. The morbidity and mortality associated with obesity pose a significant burden on public health. Body fat excess is characterized by adipocyte inflammation and an altered adipokine profile. Adipose hypertrophy, increased infiltration of macrophages and switch to M1 phenotype, and secretion of proinflammatory cytokines contribute to development of insulin resistance and long-term cardiometabolic complications. Adipokines released by adipocytes and macrophages in the stromal vascular fraction of adipose tissue can affect adipose tissue per se as well as distant effect organs. Obesity is characterized by increase in IL-6, TNF-α, IL-1, resistin and a reduction of adiponectin. Obesity is most commonly defined by excess body mass index (BMI) and assessment of metabolic health will also include measurement of body composition, along with parameters of lipid and glucose metabolism. Investigation of adipokines in different metabolic traits may provide diagnostic and interventional clues and adoption of systems approach may provide better overview of adipose inflammation in dealing with a complex, multi-system alteration triggered by obesity. Mouse 3T3-L1 preadipocyte is a well-established model for studying adipogenesis. 3T3-L1 preadipocyte was induced to differentiated adipocytes, and was treated with IL-1β 40 mg/ml for 24 hours to induce insulin resistance and inflammation in vitro. The cultured adipocytes with and without IL-1β treatment were submitted for Affymetrix GeneChip Mouse Genome 430 2.0 microarray analysis. There were 42 genes upregulated by more than 6-fold and adiponectin expression was reduced by 50% in the IL-1β treated adipocytes. Validation with real-time qPCR was performed. STRING database was employed to created protein-protein interaction network and enriched pathways were reviewed to provide insights of the molecular functions of our candidate genes. Three candidate genes, leucine-rich alpha-2 glycoprotein 1 (LRG1), secretory leukocyte peptidase inhibitor (SLPI), and neutrophil gelatinase-associated lipocalin (NGAL) encoded by LCN2 were not a part of well-recognized inflammatory and chemotactic pathways. They were selected as potential novel biomarkers and further investigation was performed to explore their association with metabolic health and inflammation. One clinical cohort enrolled 175 women aged 37 to 67 years who did not have established cardiovascular disease. Body composition was assessed by dual energy X-ray absorptiometry (DXA) and biochemical assay and adipocyte biomarkers were measured. We found out that obesity and central obesity were characterized by a general increase in body fat, while in MS and insulin resistance, fat gain only occurred in arms and central regions. NGAL was positively correlated with the size of total and central fat depots, as well as obesity and central obesity phenotypes. NGAL was also associated with dysregulated glucose metabolism, and the presence of DM and insulin resistance. Among 140 women with BMI < 27 kg/m2, 50 of them had 2 or more components of metabolic risk criteria and were categorized as metabolically obese non-obese (MONO) phenotype. The components of metabolic risk criteria included MS criteria except central obesity and inclusion of insulin resistance and subclinical inflammation. Serum ferritin was used as a surrogate marker of inflammation. MONO women had a comparable degree of central adiposity to those with BM ≥ 27 kg/m2. One of the features unique to MONO was the lack of expansion of fat and lean mass in leg and gynoid region. SLPI was found to be an independent risk factor for metabolically unhealthy status in non-obese women after adjusting for age and android-gynoid fat mass ratio. End-stage renal disease (ESRD) is characterized by chronic inflammation. The second cohort enrolled 169 ESRD patients treated with hemodialysis, including 77 diabetics (45.3%). BMI was available in 134 patients at baseline and 16.4% of them were obese. Comorbid conditions including coronary artery disease (CAD), heart failure, ischemic stroke, peripheral artery disease (PAD), liver cirrhosis, and malignancy were recorded at baseline. The patients were followed from April, 2016 until death or study end in October, 2019 At baseline, higher LRG1 was associated with increased frequency of PAD and ischemic stroke. The correlation of LRG1 with PAD remained significant after multivariate adjustment. The biomarkers were not correlated with obesity. We further investigate if adipocyte-derived biomarkers were associated with patient survival. The mean follow-up duration was 1142.65 ± 305.52 days. Thirty-four patients (20.1%) expired, 14 patients died of cardiovascular events, including 4 out-of-hospital cardiac arrests (OHCA). Lower baseline SLPI and NGAL levels were associated with better overall survival and inverse relationship was observed for LRG1. Mortality was further divided into cardiovascular and non- cardiovascular causes. Cox-regression analysis showed higher LRG1, lower SLPI and lower NGAL were associated with poor non-cardiac survival, identical to the trend of all-cause mortality. After adjusting for age, albumin, hemoglobin, and key comorbidities, baseline lower SLPI and NGAL levels were predictive better all-cause mortality. By transcriptomic analysis of cultured adipocyte treated with IL-1β, we identified three markers associated with adipose inflammation. In the cohort of otherwise healthy women, circulating NGAL levels were proportional to total and central adiposity as well abnormal glucose metabolism. SLPI was associated with another aspect of metabolic health; it was linked with metabolic unhealthy state and subclinical inflammation rather than fatness or insulin resistance. In the second cohort of hemodialysis patients, there was no association between biomarkers and BMI or lipid profiles. LRG1 was correlated with inflammatory markers, CRP and IL-6, as well as increased prevalence of PAD, but this was not translated to cardiovascular mortality. Lower SLPI and NGAL were associated with better overall survival, which may be explained by reduced non-cardiac mortality risk. In this study, we demonstrated the effectiveness of transcriptome profiling by microarray in combination with network analysis. Three biomarkers, NGAL, SLPI, LRG1, reflecting different aspects of metabolic health were identified and they may be employed in further disease phenotyping or risk stratification in metabolic disorders.