- 學位論文
開發偵測voriconazole所引發肝毒性之分析平台
Development of analytical platforms for the detection of voriconazole-induced liver injury
摘要
Voriconazole是治療真菌感染之重要藥物，而肝毒性是其主要副作用之一，嚴重肝毒性可能造成治療中止。目前臨床上使用肝功能指數(liver function index或liver function tests)作為評估依據，但肝功能指數可能會受到生理因素或其他組織的傷害與病變所影響，無法清楚歸因是否voriconazole所造成。為了專一地偵測voriconazole引發的肝毒性，急需發展準確有效的voriconazole肝毒性生物指標。 本論文採用標的代謝體學 (targeted metabolomics) 研究方法，開發了liquid chromatography - triple-quadrupole mass spectrometer (LC-QqQ MS) 與gas chromatography - mass spectrometer (GC-MS) 之分析方法，並以之尋找反映voriconazole引發肝毒性之代謝物生物指標 (metabolite biomarkers)。我們挑選文獻報導與藥物引發肝毒性有關之代謝物作為標的代謝物 (target metabolites)，以之為基準在LC-QqQ MS調整了樣品回溶溶劑與體積、管柱選擇、移動相組成、離子源選擇、MRM離子選擇與層析梯度等參數；在GC-MS平台則探討了凍乾步驟之影響、SIM (selective ion monitoring) mode與scan mode之比較、衍生物穩定性與dwell time對訊號強度之影響。我們以調整後LC-QqQ MS與GC-MS分析條件，進行voriconazole肝毒性研究。 我們使用小鼠作為藥物引發肝毒性之模式生物 (model organism)，分別給予三種藥物和兩種溶劑：acetaminophen、valproate、voriconazole、去離子水或poly ethylene glycol，於24小時後收集血漿，比較急性肝毒性小鼠與控制組血漿內選定的代謝物濃度之差異。 實驗結果顯示本研究所挑選之標的代謝物確實對acetaminophen、valproate與voriconazole三種藥物皆具有分辨實驗組與控制組之能力。在給予acetaminophen 24小時後，小鼠血漿中之threonine、aspartate、glutamate、serine、tryptophan、ornithine、cholesterol、ascorbate等八種代謝物之血中濃度顯著高於控制組；而trimethylamine-N-oxide (TMAO)、urea、urocanate、3-methylhistidine、1-methylhistidine則顯著較低；另一方面valproate造成實驗組之acetoacetate、inosine和ursodeoxycholate三種標的代謝物的血中濃度顯著低於控制組。此兩種已知引發肝毒性之模式藥物(model drugs)之結果證實本研究之研究策略確實可用於藥物引發肝毒性之研究。 在voriconazole的部分，給藥後24小時之小鼠其血中citrate、3-hydroxybutyrate、chenodeoxycholate與cytidine之濃度顯著高於poly ethylene glycol組，lysine、alanine、asparagine、glycine、5-hydroxylysine、methionine、serine、threonine與1-methylhistidine之血中濃度則顯著低於控制組。由此可以推測voriconazole可能影響了醣類之使用，使得脂質與蛋白質被作為替代能量。 在本研究所使用的兩種平台中，LC-QqQ MS平台之樣品前處理較簡單方便，分析時間亦較短。在本研究觀察到顯著變化之代謝物之中，有三個代謝物是在GC-MS平台分析，其餘23個代謝物皆於LC-QqQ MS平台分析。綜合而言，LC-QqQ MS平台較適用於本研究所挑選之標的代謝物。 本研究成功開發LC-QqQ MS之藥物肝毒性偵測平台應用於藥物引發肝毒性小鼠之血漿樣品，並以小鼠模式找到具有偵測voriconazole引發肝毒性生物指標潛力之代謝物，進一步的臨床研究將有助於確認這些代謝物對偵測voriconazole引發肝毒性之適用性。
並列摘要
Voriconazole is an antifugal drug commonly used for the treatment of invasive fungal infections. Hepatotoxicity is one of its major side effects, which may limit its use. Liver function tests are clinically used to evaluate hepatotoxicity. Nevertheless, liver function tests may be influenced by other physiological or pathological factors and can not distinguish the causes of liver injuries. To detect voriconazole-induced liver injury specifically, it is necessary to investigate new hepatotoxicity biomarkers. In this study, we used a mass based targeted metabolomics approach to identify metabolites associated with voriconazole induced liver injury. Sixty-five metabolites, which were reported to be associated with hepatotoxicity were selected as target metabolites. Both liquid chromatography - triple-quadrupole mass spectrometer (LC-QqQ MS) and gas chromatography - mass spectrometer (GC-MS) methods were developed to investigate hepatotoxicity biomarkers. The adjustment parameters of the LC-QqQ MS method included reconstitution solvent and volume, columns, mobile phase composition, ionization sources, elution gradient, and MS transition ions. In the GC-MS method, we investigated the influence of lyophilization, stability of the metabolite derivatives, and MS parameters. We used the adjusted LC-QqQ MS and GC-MS methods to investigate voriconazole-induced hepatotoxicity markers. Liver toxicity was induced in mice through the administration of two model drugs, acetaminophen and valproate, and our investigation drug, voriconazole. The results showed that the metabolite profile of mice plasma is able to discriminate treatment groups from the control groups. In the acetaminophen treatment group, plasma concentrations of threonine, aspartate, glutamate, serine, tryptophan, ornithine, cholesterol and ascorbate were significantly higher than the control group, and plasma concentrations of trimethylamine-N-oxide (TMAO), urea, urocanate, 3-methylhistidine and 1-methylhistidine were significantly lower than the control group. In the valproate treatment group, plasma concentrations of acetoacetate, inosine and ursodeoxycholate were significantly lower than the control group. The results of the two model drugs reveal that our developed MS based analytical platforms are feasible for the study of drug induced liver injury. In the voriconazole treatment group, plasma concentrations of citrate, 3-hydroxybutyrate, chenodeoxycholate and cytidine were significantly higher than the control group, and plasma concentrations of lysine, alanine, asparagine, glycine, 5-hydroxylysine, methionine, serine, threonine and 1-methylhistidine were significantly lower than the control group. This result suggests that voriconazole may affect the use of carbohydrate in the generation of energy, resulting in the compensatory consumption of amino acids and lipids. Comparing the two platforms used in this study, the sample preparation procedure of LC-QqQ MS method is simpler, and the analytical time of LC-QqQ MS method is shorter. Among the statistically significant metabolites, 3 metabolites are analyzed by the GC-MS platform, and the other 23 metabolites are analyzed by the LC-QqQ MS platform. In general, the developed LC-QqQ MS platform is more suitable for investigating hepatotoxicity markers. In conclusion, we successfully developed an LC-QqQ MS platform to detect drug-induced hepatotoxicity, and used mice models to elucidate the metabolic change that occurs after voriconazole-induced hepatotoxicity. We used the significant metabolites to propose possible mechanisms of voriconazole-induced hepatotoxicity. Further clinical study is required to improve our understaning on voriconazole induced hepatoxicity.