肝臟移植病人及其捐贈者基因多型性對tacrolimus血中濃度之影響

Abstract

研究背景與目的： Tacrolimus（TAC）的藥動學在肝臟移植病人有非常大的個體內以及個體間的差異，研究顯示許多因素會影響TAC的藥動學。然而在過去的研究中並沒有同時考量臨床因素以及受贈者、捐贈者的基因型因素，更沒有任何臺灣人相關的研究，本研究是目前為止第一個探討臺灣肝臟移植病人在移植後3個不同時間所有可能會影響TAC劑量校正谷濃度（dose normalized trough concentrations，dnC0；與dosing weight and dose normalized trough concentrations，dnC0/DW）的臨床及基因因素。 方法： 本研究由2008年1月1日至2015年7月31日進行肝臟移植的307位臺灣病人中，選取活體移植，年齡介於20到65歲，在移植後持續使用TAC至少6個月作為免疫抑制劑，於研究期間持續追蹤。排除條件包含再移植或多重器官移植、以及人類免疫缺乏病毒反應呈陽性的病人。選取之病人及其捐贈者，簽署受試者知情同意書（informed consent）自願提供檢體做基因檢測者則納入本研究，收集到96個病人及其捐贈者的檢體，最後有70個病人及其捐贈者的基因定序結果可納入分析。本研究以dnC0及dnC0/DW代表TAC藥動學的變化，在迴歸分析時，dnC0及dnC0/DW經由對數轉換來確保資料符合常態分布。在三個評估點：移植手術出院前、移植後3個月以及移植後6個月，病人如有使用已知會和TAC產生交互作用的藥品則會在該時間點被排除。統計方法使用單變項分析篩選臨床因素及基因，當其p值＜0.2則放入多元迴歸分析，並用逐步選取法找出顯著影響（p＜0.05）TAC log-transformed dnC0（LN dnC0）及log-transformed dnC0/DW（LN dnC0/DW）的因素。Kolmogorov-Smirnov檢定用於確認常態分布；連續性資料使用獨立樣本t檢定或曼恩-惠尼U檢定；同一病人在不同時間點的TAC濃度資料與檢驗數值使用單因子重複量測變異數分析；類別資料如性別及基因多型性等則使用Chi-squared test以及Fisher’s exact test。 結果： 本研究共納入70個移植病人及其捐贈者，在移植後的三個評估點一致顯示影響TAC的LN dnC0或LN dnC0/DW最重要因素是捐贈者的CYP3A5*3，解釋力由移植手術出院前的18 %隨時間增加至移植後六個月的25 %。受贈者的CYP3A5*3及CYP3A4*18B在移植手術出院前以及移植後6個月會影響TAC的LN dnC0（R2 = 4-5 % for CYP3A5*3，R2 = 6-7 % for CYP3A4*18B）。我們也發現不論肝臟移植受贈者或捐贈者CYP3A5*3的變異與CYP3A4*1G有高度關聯性：當CYP3A5是表現型時（帶有CYP3A5*1），有約80 %會同時有至少一個CYP3A4*1G變異，兩種基因型都會使TAC的LN dnC0下降；反之，當CYP3A5是非表現型時（CYP3A5*3*3），則CYP3A4*1G變異較少（約15 %）。 移植後不同時間點影響TAC的LN dnC0或LN dnC0/DW的重要臨床因素均不相同，且整體解釋能力僅約10 %。移植手術出院之重要臨床因素有mycophenolate mofetil或mycophenolate sodium（MMF/MPS）以及proton pump inhibitor（PPI）的使用；移植後3個月有TAC的劑型（Prograf ® 或Advagraf ®）；移植後6個月有ALP、AST以及steroid的每日劑量。MMF或MPS的使用有較低的TAC LN dnC0可能源自於TAC的劑型使用分布不均，在有使用MMF或MPS的病人中有約75 %使用Advagraf ®；沒有使用MMF或MPS的病人則有約35 %使用Advagraf ®，經次族群分析後發現TAC的劑型可能才是影響TAC LN dnC0的因素。 隨著納入分析的因素增加，整體的解釋能力也有增加的趨勢；在只納入受贈者臨床因素時，整體解釋能力約10 %；納入受贈者基因因素後，整體解釋能力最多可以增加到27 %，而原先的臨床因素的解釋比例持平；再進一步納入捐贈者基因因素後整體解釋能力最多可達53 %，但受贈者基因的解釋能力會因捐贈者基因的納入而下降。 結論： 肝臟移植病人移植後6個月內，在三個時間點影響TAC的LN dnC0或LN dnC0/DW最重要且持續影響的因素為捐贈者的CYP3A5*3（R2 = 18-25 %）。重要的臨床因素整體解釋能力只有10 %左右，且會隨著時間而不同，包含TAC的劑型、MMF/MPS的使用、PPI的使用、steroid的每日劑量、ALP以及AST。MMF或MPS的使用對TAC LN dnC0影響可能來自於TAC的劑型使用分布不均所致。CYP3A5*3的變異與CYP3A4*1G之變異有高度關聯性，此關聯性對於TAC LN dnC0的影響是同方向的。建議將捐贈者以及受贈者的CYP3A5基因型納入肝臟移植術前評估，在已完成肝臟移植的病人，當移植後病人使用TAC時發現很難達到目標濃度時，可考慮檢測捐贈者CYP3A5基因。

Background： Large interindividual and intraindividual variations of tacrolimus (TAC) pharmacokinetics (PK) existed in liver transplant recipients. Many factors were reported to influence TAC PK. However, there was no study on the influence of both the clinical and different genetic factors of liver transplant recipients and the genetic factors of donors, not to mention study in Taiwanese. Up to our knowledge, this is the very first study investigated clinical and genetic factors of recipients and the genetic factors of donors that significantly influenced the TAC dnC0 (dose normalized trough concentrations) and the dnC0/DW (dosing weight and dose normalized trough concentrations) in Taiwanese liver transplant recipients at three different time points. Methods： From January 2008 to July 2015, 307 Taiwanese patients received liver transplantation in our medical center. Patients were included when they met the following criteria: received living related transplantation at the age of 20 to 65 years, on TAC-based regimen for at least 6 months, coupled with their donors, who agreed be genotyped for specific polymorphisms. All patients enrolled were followed up during study period and their blood samples were taken for genetic study after signing informed consent. Exclusion criteria were retransplantation, multiorgan transplantation, and immunodeficiency virus (HIV) positive patients. Blood or oral swab samples were collected from 96 patients and their matched donors to determine genotypes. Seventy recipients and their matched donors’ genotypes could be determined. TAC dnC0 and dnC0/DW were used as the PK marker, and were log-transformed to ensure a normal distribution. This study was performed at three time point: the last steady state C0 before discharge after transplantation, 3 months and 6 months post-transplant. Patients received drugs that interacted with TAC were exclude at the specific time point. Clinical factors and genetic factors with a p value of < 0.20 in a univariate regression analysis were included in the stepwise multivariate regression analysis to explore important determinents of TAC log-transformed dnC0 (LN dnC0) or log-transformed dnC0/DW (LN dnC0/DW). Kolmogorov-Smirnov test was used to test normal distribution. Independent samples t-test or Mann-Whitney U test were used for continuous data. One-way Repeated Measurement ANOVA (analysis of variance) was used to compare the dnC0 and dnC0/DW of TAC and lab data at 3 time points. Chi-squared test and Fisher’s exact test were used in categorical data. Result： A total of 70 liver transplant recipients, coupled with their donors, were enrolled in the study. Donors’ CYP3A5*3 was the most significant factors that consistently influenced TAC LN dnC0 or LN dnC0/DW at all three time points. The coefficient of determination (R2) increased from 18 % on discharge to 25 % at 6 months after transplantation. Recipients’ CYP3A5*3 and CYP3A4*18B was significantly associated with TAC LN dnC0 on discharge and six months after transplant (R2 = 4-5 % for CYP3A5*3, R2 = 6-7 % for CYP3A4*18B). There was linkage between CYP3A5*3 and CYP3A4*1G in both recipients and donors. About 80 % CYP3A5 expressers (CYP3A5*1 carrier) had CYP3A4*1G mutation. Both genotypes resulted in decreased TAC LN dnC0. On the other hand, CYP3A4*1G mutation was observed in only 15 % CYP3A5 nonexpressers (CYP3A5*3*3). Significant clinical factors that influenced TAC LN dnC0 or LN dnC0/DW, with a total R2 of 10 %, varied among three time points. These factors included mycophenolate mofetil or mycophenolate sodium (MMF/MPS) use and proton pump inhibitor (PPI) use on discharge after transplant; dosage form of TAC (Prograf ® or Advagraf ®) 3 months after transplant; and ALP, AST and steroid daily dose 6 months after transplant. MMF/MPS users had lower LN dnC0. About 75 % MMF/MPS users used Advagraf ®, and so did only 35 % MMF/MPS non-users. Subgroup analysis indicated that dosage form of TAC rather than MMF/MPS use could be the factor that influenced TAC LN dnC0. When using multiple regression to explore important determinations of LN dnC0 and LN dnC0/DW of TAC, R2 increased as number of factors increased. R2 was around 10 % when only recipients’ clinical factors were included. After including recipients’ genetic factors, the overall R2 increased up to 27 %. While R2 of the orginal clinical factors remained as 10 %. Overall R2 increased to 53 % when we included clinical and genetic factors of recipients and the genetic factors of donors. However, R2 of the genetic factors of recipients decreased. Conclusion： CYP3A5*3 of liver transplant donors was the most important factor that consistently influenced TAC LN dnC0 and LN dnC0/DW for all three time points within 6 months after transplant (R2 = 18-25 %). Clinical factors of recipients that influenced TAC LN dnC0 or LN dnC0/DW only accounted for a total 10 %, and varied at three time points. The finding that MMF/MPS use was associated with a decrease in TAC LN dnC0 could be due to variation in the use of dosage form of TAC. There was linkage between CYP3A5*3 and CYP3A4*1G in both recipients and donors. We suggested that recipient and donor CYP3A5*3 genotype be routine checked in pre-transplant assessment. For patients with difficulty in achieving target TAC C0, a genetic test for donor CYP3A5*3 could be considered.