第一部分：建立氣相層析質譜儀方法定量短鏈脂肪酸之濃度並研究糞便檢體樣品製備方法產生之誤差 第二部分：以液相層析質譜儀結合柱後注入內標法分析尿液中的生物胺並應用於研究乳癌化療藥物之療效反應

Abstract

代謝體學為近年新興之研究領域，主要目標為廣泛地分析生物系統中的代謝物，已展現出具有應用於個人化醫療的潛力。標的代謝體學探討特定標的代謝物與臨床疾病之間的相關性，而針對這些標的代謝物開發分析方法可以獲得更準確的結果。本論文運用質譜技術於標的代謝體學，分別測定兩種不同類型的小分子代謝物：短鏈脂肪酸與生物胺。 在本論文的第一部分，我們開發並確效一利用氣相層析飛行式質譜儀定量人體糞便中之六種短鏈脂肪酸濃度，包含：乙酸、丙酸、丁酸、異丁酸、異戊酸與戊酸。為達到去除基質干擾物的目的，我們使用丁醇從經酸化處理的糞便中萃取短鏈脂肪酸。在分析條件的最適化下，利用VF-WAXms毛細管層析管柱進行分離，能於十六分鐘內完成此六個短鏈脂肪酸之定量分析。此方法之準確度介於87.9至110.9% (n = 9)，同日 (n = 3) 與異日間 (n = 9) 精密度之相對標準偏差皆小於10%。我們進一步應用此方法研究糞便檢體製備方式對於分析短鏈脂肪酸可能造成之誤差，包括凍乾與取樣位置的影響。研究發現配對樣品中 (n = 6) ，經凍乾處理的糞便檢體與未經處理的檢體之間的短鏈脂肪酸濃度比值介於0.6至1.4。同時，我們也觀察到從同一糞便樣品 (n = 5) 的六個不同位置取樣所測定的短鏈脂肪酸濃度具有顯著的差異，六個短鏈脂肪酸濃度的相對標準偏差介於6.0％至66.3％。我們的結果指出，若僅使用60毫克的糞便樣品進行分析，將可能因為不同的取樣位置造成顯著的測量偏差。 第二部分我們利用液相層析質譜儀並結合柱後注入內標校正法建立分析人體尿液中14個生物胺的分析方法，包括兩種不同化學結構的多胺與兒茶酚胺。我們以丹磺醯氯作為衍生化試劑並使用Gemini C18作為層析管柱，此方法能在十二分鐘內完成14個標的生物胺的分離。我們利用經衍生的腐胺同位素 (putrescine-d6) 做為單一的柱後注入內標，能夠有效地校正不同尿液檢體中的基質效應而引起的測量誤差。我們進一步將此方法應用於研究尿液生物胺濃度與乳癌病人化療反應的關係。60個接受化學治療之乳癌病人分別在第一次化療治療前、第二次化療前、第三次化療前三個不同時間點採集尿液檢體，並利用此分析方法分析其尿液中生物胺含量。實驗結果顯示，14個被選為標的的生物胺中，N1,N12-diacetylspermine (N1N12) 與N1,N8-diacetylspermine (N1N8) 在化學治療前之尿液含量，在化療有效組 (n = 43) 中顯著高於 (P < 0.05) 化療無效組 (n = 17)。除此之外，N-acetylputrescine (NAP) 與N1N12的含量在化療有效組中之化學治療前、後有顯著改變。我們的結果顯示尿液中的N1N12含量具有潛力作為評估乳癌化療反應的生物指標，但未來仍需要更多的研究來驗證。 本研究成功建立分別用於測定短鏈脂肪酸與14個生物胺的分析平台，未來可更廣泛將這兩個分析方法應用於不同之臨床研究以探討這些代謝物在疾病機轉的角色與協助臨床診斷。

Metabolomics, an emerging field which aims at comprehensive analysis of metabolites in the living system, has shown its potential in personalized medicine. Targeted metabolomics evaluates the relevance of predefined analytes to a specific clinical problem, and the analytical method could be designed for those target analytes to obtain more accurate results. In this study, a mass spectrometry-based targeted metabolomics approach was employed to determine two different classes of small-molecule metabolites (short-chain fatty acids (SCFAs) and biogenic amines) In the first part of this thesis, we developed and validated a quantitative method to analyze six SCFAs, including acetic, propionic, butyric, isobutyric, isovaleric, and valeric acids, in human fecal samples by gas chromatography-time of flight mass spectrometry. To remove interfering species, we used butanol to extract SCFAs from acidified fecal suspensions. Six SCFAs could be quantified within 16 mins under optimal conditions by using a capillary VF-WAXms column for separation. The accuracy of the method was ranged from 87.9% to 110.9% (n = 9), and both the intra-day (n = 3) and inter-day precision (n = 9) were below 10% relative standard deviation (RSD). We further applied the method to study sample preparation errors including lyophilization and sampling position. Our findings revealed that the concentration ratios between the paired samples (n = 6), freeze-dried feces and wet feces (without lyophilization) were ranged between 0.6 and 1.4. Also, we observed different concentrations of SCFAs among samples collected from different positions of the same feces (n = 5). The RSD of six sampling from different positions for six SCFA was range from 6.0% to 66.3%. Our results indicated significant bias may be introduced by different sampling position if 60 mg of fecal sample was used for the analysis. In the second part, we analyzed 14 selected biogenic amines, including two distinct chemical structures—polyamines and catecholamines—in human urine using post-column infused internal standard (PCI-IS) strategy in liquid-chromatography mass spectrometry. The 14 target amines were derivatized with dansyl chloride, and the separation was achieved by a Gemini C18 column within 12 mins. The dansylated putrescine-d6 was utilized as a universal PCI-IS for the 14 target amines, and it effectively calibrated the errors caused by matrix effect in urine samples. The developed method was applied to investigate the relationship between the urinary biogenic amines and the chemotherapeutic response in breast cancer patients. Urine samples were collected from 60 patients at three time points: before chemotherapy was initiated (pre-C1), on the day before the second cycle (pre-C2), and the third cycle (pre-C3) of chemotherapy and they were analyzed by the optimal LC-MS method to investigate the level of biogenic amines. Among the 14 selected biogenic amines, N1,N12-diacetylspermine (N1N12) and N1,N8-diacetylspermine (N1N8) in the responder group (n = 43) were significantly higher (P < 0.05) than the nonresponder group (n = 17) prior to the chemotherapy. N-acetylputrescine (NAP) and N1N12 showed significantly change in response to chemotherapy in the responder group. Our results revealed urinary N1N12 showed high potential to serve as an indicator for assessment of the response to chemotherapy in breast cancer, though further studies are required to verify this finding. This study successfully established two analytical platforms for the SCFAs and the biogenic amines which could be used for various clinical studies to explore their potential roles for interpretation of disease mechanism and clinical diagnosis.