急性骨髓性白血病的基因變異及臨床運用

Genetic alterations and their clinical implications in acute myeloid leukemia

侯信安(Hsin-An Hou)

指導教授 ： 田蕙芬

共同指導教授 ： 林亮音

國立臺灣大學/醫學院/臨床醫學研究所/博士(2013年)

https://doi.org/10.6342/NTU.2013.01303

摘要

急性骨髓性白血病是一種惡性的骨髓造血功能異常，在病理機轉、臨床表現及治療效果上都是一個相當分歧的疾病。分子研究上的進步可進一步擴展我們對於急性骨髓性白血病致癌機轉的瞭解。傳統上染色體的核型異常已被證實在診斷上及預後上為重要的生物標記。儘管如此，約有 60% 至 75% 的病患屬於中度風險染色體核型變化，這個族群的基因變異甚大且預後也不盡相同，如何進一步剖析這群病患是最重要的臨床課題。近年來由於分子生物技術的進步，次世代定序的蓬勃發展，揭露出更多基因變異在急性骨髓性白血病的角色。
本研究的主題在於釐清不同基因變異在急性骨髓性白血病的致癌角色、臨床表徵與治療預後，譜劃出急性骨髓性白血病的基因譜，繼而彙集這些基因變異與染色體核型變化的結果，建構整合預後評估風險系統，當作病患風險評估、預後判斷及治療選擇的標的。首先我們擬針對目前尚不明瞭的 3 個重要基因，包括 RUNX1 基因突變、WT1 基因突變及 DNMT3A 基因突變，研究其在急性骨髓性白血病的致癌角色及其特有的臨床生物表徵。同時分析 CEBPA 單雙基因突變對於急性骨髓性白血病病患預後的影響。次而我們將整合完成現有 17 個不同基因變異在急性骨髓性白血病的角色，搭配染色體核型變化的結果，針對急性骨髓性白血病病患，尤其是中度風險染色體核型變化族群做風險評估的標的及未來治療選擇的依循。最後根據不同基因在診斷時及復發時的變化，研究其穩定性及評估是否可能成為未來微量殘餘疾病分析的生物標記。
針對第一部份 (RUNX1 基因突變的角色)：我們的研究成果是第一篇針對大規模病患做全面 exon 3至 exon 8 的完整分析。RUNX1 基因突變在非 M3 亞型急性骨髓性白血病的發生率為 13.2%。RUNX1 基因突變好發於男性、年長者、擁有 FAB M0/M1 亞型或 +8 染色體核型變化的病患中。大多 runt homology domain (RHD) 的突變屬於 missense 基因突變 (64.5%)，相對地 transcription activation domain (TAD) 的突變則多為 frameshift 基因突變 (76.5%)。Missense基因突變預測會影響到去氧核醣核酸的結合能力或與 CBFβ 形成異合體 (heterodimerization) 的能力。而 frameshift 基因突變則會造成 TAD 部分或全部的遺失，失去 transactivation potential。在 62 位具有 RUNX1 基因突變的病患中， 31 位 (50%) 會伴隨其他的分子基因變異；其中最常見的是 Class I 基因變異 (83.9%)，包括 FLT3/ITD、 FLT3/TKD 與 NRAS基因突變等。在存活分析方面，我們確認了 RUNX1 基因突變是一個獨立的不良預後標的，病患會有較差的完全緩解率、無病況之存活時間與整體存活時間。
針對第二部份 (CEBPA 基因突變的角色)：我們的研究指出三分之二病患的 CEBPA 基因突變為雙基因突變，同時影響 TAD1 及 basic leucine zipper domain (bZIP) 的區域。我們的結果顯示擁有 CEBPA 雙基因突變的病患比那些 CEBPA 單基因突變或是沒有 CEBPA 基因突變之病患，他們的完全緩解率較高且整體存活時間及無病況之存活時間較長。進一步在多變項分析的結果顯示，具有 CEBPA 雙基因突變，但並非 CEBPA 單基因突變，是一個獨立預後好的存活因子。由於單基因突變比較容易存在於芽細胞有 CD56 表現的病患，同時常與一些不好的基因變異存在，諸如 FLT3/ITD、FLT3/TKD、MLL/PTD及 RUNX1 基因突變。或許可以部分解釋為何 CEBPA 單基因突變的病患預後較差的原因。
針對第三部份 (WT1 基因突變的角色)：這是第一篇收納最多染色體核型異常變化之急性骨髓性白血病病患研究，我們的研究顯示 WT1 基因突變在所有病患的發生率為 6.8%，在年輕且染色體核型正常變化的病患中其發生率為 8.3%；而突變率在染色體核型正常變化與染色體核型異常變化的兩族群中無差別。 WT1基因突變好發於年輕者或擁有 FAB M6 亞型的病患，而較少發生在 M0 亞型的病患中。除了發現 t(7;11)(p15;p15) 常見於 WT1 基因突變的病患外，大多數 (72%) 擁有 WT1 基因突變的病患會合併其他分子基因的變異，常見於 Class II 基因變異 (69.6%) 與 Class I 基因變異 (56.5%)，其中 FLT3/ITD 與 CEBPA 基因突變是最常與 WT1 基因突變同時出現的。在存活分析方面，不論在全部或是染色體核型正常變化的病患， WT1 基因突變皆是一個獨立的不良預後標的。病患會有較高的復發率及較短的無復發病況之存活時間與整體存活時間。藉著整合 4 個生物預後標記，包括年紀、WT1 基因突變、NPM1/FLT3-ITD 與 CEBPA 基因突變，我們可將病患做更有效的危險預後分組。
針對第四部份 (DNMT3A 基因突變的角色)：我們的研究指出 DNMT3A 基因突變會出現在 30 個不同的位置，最常見在 MTase domain。此處所有的 nonsense、frameshift 及 in-frame 基因突變預測會造成 truncated peptide，會破壞這個酵素的活性。其中最常見的 R882 突變會造成酵素功能的失調。DNMT3A 基因突變的發生率在所有的病患、非 M3 亞型的病患、帶有中度風險染色體核型變化的病患及染色體核型正常變化的病患，各為 14%、15.2%、19.5% 及 22.9%。 DNMT3A基因突變好發於年長者、擁有 FAB M4/M5 亞型或染色體核型正常變化的病患中。有趣地，在急性骨髓性白血病的病人身上，DNMT3A 基因突變很少單獨存在，高達 97.1% 會合併其他分子基因變異。除了 DNMT3A 基因突變會與 NPM1 基因突變或 FLT3/ITD 基因變異密切相關外，發現擁有 DNMT3A 基因突變的病患比那些沒有 DNMT3A 基因突變的病患，同時具有較高的 IDH2 或 PTPN11 基因突變；相反地， CEBPA 基因突變很少出現在擁有 DNMT3A 基因突變的病患身上。在存活分析方面，不論在全部或是染色體核型正常變化的病患， DNMT3A 基因突變皆是一個獨立的不良預後的標的，病患有較高的復發率及較短的無復發病況之存活時間與整體存活時間。藉著整合 8 個生物預後標的，包括年紀、白血球數目、 NPM1/FLT3-ITD、 CEBPA 基因突變、DNMT3A 基因突變、WT1 基因突變、RUNX1 基因突變與 IDH2 基因突變，我們可進一步將病患做更好的危險預後評估分組。
針對第五部份 (整合預後風險評估系統)：針對 318 位非 M3 亞型的急性骨髓性白血病且接受標準化學治療的病患，我們進行 17 個基因的檢測，發現 RUNX1 基因突變、WT1 基因突變及 DNMT3A 基因突變皆是獨立預後不良之存活因子。最後採用 8 個與預後有關的基因變異包括 FLT3/ITD、CEBPA、NPM1、RUNX1、WT1、IDH2、ASXL1 及 DNMT3A 基因突變，可將非常分歧的中度風險染色體核型變化之病患分成預後截然不同的三組。同時可將傳統上用染色體核型變化分類中占 72% 的預後中等族群，減少到利用整合分析而來之 24.5%。進一步發現：當病患屬於中度風險染色體核型變化卻同時擁有預後良好的分子基因變化 (在沒有 FLT3/ITD 基因變異下，具有 NPM1 基因突變、IDH2 基因突變或是 CEBPA 雙基因突變)者之治療成績與那些擁有預後良好之染色體核型變化者一樣好。相反地，當病患屬於中度風險染色體核型變化卻同時擁有預後不佳的分子基因變化 (具有 RUNX1 基因突變、WT1 基因突變、ASXL1 基因突變或 DNMT3A 基因突變) 者之治療成績與那些擁有預後不佳之染色體核型變化者一樣差。
針對第六部份 (基因變異的配對檢體系列檢驗分析)：我們利用系列性的病患檢體 (診斷、完全緩解及復發時) 做基因的檢測，來探討常見基因突變在急性骨髓性白血病發生及病程進展時所扮演的角色，尤其是 RUNX1、WT1、CEBPA 及 DNMT3A 等 4 個基因。我們發現在所有 117 位病患中，61 位病患出現基因改變 (包括基因遺失或基因獲得) 的現象。Class I 基因遠比 Class II 基因更容易出現基因改變的現象 (50.8% vs. 15.6%, P=0.001)。此外，復發時有 15 位 (12.8%) 病患獲得新的基因變異，其中多數是獲得新的 Class I 基因變異，其中 3 位獲得新的 WT1 基因突變。然而卻沒有任何 1 位獲得新的 Class II 基因變異。這部分所收納的病患皆是目前研究數目最多且最完整的系列檢驗。我們的研究發現相對於 FLT3/ITD 基因變異的不穩定性， DNMT3A 基因突變、NPM1 基因突變、CEBPA 基因突變及 IDH1/2 基因突變皆維持在癌化過程中的穩定地位。此外 RUNX1 基因突變也被證實在此疾病進展時扮演著重要角色，但在復發時的角色較不明顯。相反地，WT1 基因突變與 TET2 基因突變會在復發時消失或是出現。這些結果可協助判斷何種基因變異可做為微量殘餘疾病分析時的理想生物標記。
本研究的成果首先提供不同基因變異在急性骨髓性白血病病患的發生率、臨床表徵、與染色體核型變化及合併不同基因變異的關聯性，次而論及這些基因變異對於病患的預後影響。綜合這些資料我們可以譜出急性骨髓性白血病病患詳細的基因變異圖表，進一步可瞭解在診斷及復發時基因變異的狀況，建構整合預後評估風險系統，當做個人化醫療的基石及未來轉譯醫學的重要參考。

關鍵字

急性骨髓性白血病 ； 基因變異 ； 預後

並列摘要

Acute myeloid leukemia (AML) is a heterogeneous group of neoplastic disorders with great variability in the pathogenesis, clinical course and response to therapy. Advances in molecular research have greatly improved our understanding of the leukemogenesis in AML. The aim of the present doctoral thesis is to first investigate the detailed molecular genetic alterations in patients afflicted with AML. We plan to elucidate the role of RUNX1, WT1 and DNMT3A mutations, three molecular alterations not well studied before, and single or double CEBPA mutation (CEBPAsingle-mut or CEBPAdouble-mut) in the leukemogenesis of AML. Second, it is of utmost significance to understand the clinic-biologic features, associated cytogenetic and molecular abnormalities and clinical relevance of these molecular aberrations in AML patients. Further, we can depict the comprehensive picture of total 17 frequent molecular genetic alterations in AML patients from the project and other researches from our team. In addition to the importance of these genetic alterations in the development of AML, we also perform sequential studies to explore the role of these events in the disease progression. As we aware, the integrated prognostic system is more power than a single biomarker to risk-stratify the heterogeneous AML patients into different groups with distinct outcome. Therefore, the integration of cytogenetic and molecular changes we design will improve the prognostic stratification of AML patients, especially those with intermediate-risk cytogenetics, and may lead to better decision on therapeutic strategy.
In the subproject one (RUNX1 mutation in AML), we found that RUNX1 mutations could be detected in a substantial proportion (13.2%) of patients with de novo non-M3 AML. RUNX1 mutations were closely associated with male gender, older age, immature FAB subtypes (M0 and M1) and trisomy 8. Most runt homology domain (RHD) mutations were missense mutations (64.5%), while most transcription activation domain (TAD) mutations were of frameshift mutations (76.5%). Among the 62 patients with RUNX1 mutations, 31 (50%) showed additional molecular abnormalities at diagnosis. They were mutually exclusive with CEBPA and NPM1 mutations, but were closely associated with MLL/PTD. Furthermore, the RUNX1 mutation predicted lower complete remission (CR) rate and was an independent poor-risk factor for overall survival (OS) and disease-free survival (DFS).
In the subproject two (CEBPA mutation in AML), about two third of patients with CEBPA mutations had CEBPAdouble-mut both at TAD1 and bZIP domains. CEBPAdouble-mut patients had a higher CR rate and a significant longer DFS and OS than those with CEBPAwild or CEBPAsingle-mut. CEBPAsingle-mut was closely associated with CD56 expression but inversely correlated with HLA-DR, CD7 and CD15 expression. Compared to the patients with CEBPAdouble-mut, those with CEBPAsingle-mut had a higher incidence of concurrent FLT3/ITD, FLT3/TKD, MLL/PTD or RUNX1 mutation and had a poorer prognosis. This study provides evidences independently from previous ones, stressing the differences in biologic characteristics between CEBPAsingle-mut and CEBPAdouble-mut AML and their possible prognostic implication.
In the subproject three (WT1 mutation in AML), 29 different kinds of WT1 mutations were detected in 32 (6.8%) patients. We showed that WT1 mutations occurred with similar frequencies in patients with normal cytogenetics (CN-AML, 7.0%) and those with abnormal cytogenetics (6.3%). The mutation was closely associated with younger age, FAB M6 subtype and t(7;11)(p15;15), but inversely related to M0 subtype. Among the 32 patients with WT1 mutations, 23 (72%) showed additional molecular abnormalities at diagnosis; sixteen (69.6%) of them had at least one concurrent Class II mutation and 13 (56.5%), Class I mutation. Furthermore, the WT1 mutation predicted higher relapse rate and was an independent poor-risk factor for OS and relapse-free survival (RFS) among total cohort and CN-AML patients. Incorporation of the these gene mutations, including NPM1/FLT3-ITD, CEBPAdouble-mut, WT1 mutation and age at diagnosis that are closely associated with prognosis, into survival analyses can better stratify patients into different risk groups.
In the subproject four (DNMT3A mutation in AML), DNMT3A mutations at 30 different positions, most commonly in the MTase domain, were demonstrated. All the nonsense, frameshift, and in-frame mutations generated truncated peptide with complete or partial deletion of the MTase domain and were suggested to abolish the catalytic activity of this enzyme. The missense R882 mutations, the most common DNMT3A mutations, resulted in impaired enzyme activity. DNMT3A mutations could be detected in a substantial proportion of patients with de novo AML (14% of total patients; 15.2% of non-M3 AML; 19.5% of intermediate-risk cytogenetics and 22.9% of CN-AML). DNMT3A mutations were closely associated with older age, FAB M4/M5 subtypes and intermediate-risk cytogenetics and CN-AML. Among the 70 patients with DNMT3A mutations, 68 (97.1%) showed additional molecular abnormalities at diagnosis. They were mutually exclusive with CEBPA mutation, but were closely associated with FLT3/ITD, NPM1, PTPN11 and IDH2 mutations. Furthermore, the DNMT3A mutation was an independent poor-risk factor for OS and RFS among total cohort and CN-AML patients. Incorporation of DNMT3A mutation with eight other prognostic factors, including age, WBC counts, cytogenetics, NPM1/FLT3-ITD, CEBPA, RUNX1, WT1, and IDH2 mutations, into survival analyses can better stratify AML patients into different risk groups.
In the subproject five (integrated prognostic system in AML patients), early assessment of cytogenetics and mutational profiling of eight relevant genes, including CEBPA, NPM1, FLT3, RUNX1, WT1, IDH2, ASXL1 and DNMT3A, may provide a framework for risk stratification in 318 non-M3 AML patients. Using the proposed classification by integrating cytogenetic and mutational profiles, we reduced the proportion of patients in intermediate-risk group, as defined by cytogenetics alone, from 72% to 24.5%, defined by both cytogenetic and molecular changes. Three fourth of intermediate-risk patients defined by cytogenetics alone can be reliably reclassified into favorable- (35.2%) or unfavorable-risk (40.2%) group according to the molecular genotype. The patients with intermediate-risk cytogenetics but favorable molecular genotype (mutation of NPM1, IDH2 or CEBPAdouble-mut in the absence of FLT3/ITD) had similar treatment response to those with favorable-risk cytogenetics. Similarly, the prognosis of the patients with intermediate-risk cytogenetics but unfavorable molecular genotype (mutation of RUNX1, WT1, ASXL1, or DNMT3A) was as poor as those with unfavorable-risk cytogenetics. The integrated cytogenetic and molecular classification further refines the prognostic prediction models and may guide the therapeutic decision.
In the subproject six (comparative analyses of genetic alterations in paired samples), we found that mutational shifts, including mutational loses and gains, occurred in 61 (52.1%) of 117 patients at relapse. Class I mutations, which activate signal transduction, were lost more frequently (50.8%) than Class II mutations (15.6%), which cause transcriptional deregulation. Genetic evolution with acquisition of novel mutations at relapse were identified in 15 individuals (12.8%), all involving Class I, WT1 or ASXL1 mutations, but not Class II or NPM1 mutations. Sequential study showed that mutations of CEBPA, RUNX1, IDH1/2 or DNMT3A remained quitely stable during the clinical course. In contrast, the instability of WT1 and TET2 mutations was noticed during disease progression. Our findings distinctly help judge which genetic alteration can be used as a biomarker for minimal residual disease (MRD) monitoring.
In conclusions, the present doctoral thesis combined genetic association studies, molecular results and clinical data to demonstrate how the genetic alterations, especially RUNX1, WT1 and DNMT3A mutations are involved in the leukemogenesis of AML. Incorporation of cytogenetic changes and molecular alterations as an integrated prognostic system can better risk-stratify the heterogeneous AML patients, especially intermediate-risk cytogenetics into different prognostic groups with distinct outcome. Finally, the gene mutations which are stable during treatment courses can also be used as biomarkers to monitor MRD.

並列關鍵字

AML ； genetic alterations ； prognosis

參考文獻

Abbas, S., Lugthart, S., Kavelaars, F.G., Schelen, A., Koenders, J.E., Zeilemaker, A., van Putten, W.J., Rijneveld, A.W., Lowenberg, B. & Valk, P.J. (2010) Acquired mutations in the genes encoding IDH1 and IDH2 both are recurrent aberrations in acute myeloid leukemia: prevalence and prognostic value. Blood, 116, 2122-2126.

Alberta, J.A., Springett, G.M., Rayburn, H., Natoli, T.A., Loring, J., Kreidberg, J.A. & Housman, D. (2003) Role of the WT1 tumor suppressor in murine hematopoiesis. Blood, 101, 2570-2574.

Ariyaratana, S. & Loeb, D.M. (2007) The role of the Wilms tumour gene (WT1) in normal and malignant haematopoiesis. Expert Rev Mol Med, 9, 1-17.

Bachas, C., Schuurhuis, G.J., Hollink, I.H., Kwidama, Z.J., Goemans, B.F., Zwaan, C.M., van den Heuvel-Eibrink, M.M., de Bont, E.S., Reinhardt, D., Creutzig, U., de Haas, V., Assaraf, Y.G., Kaspers, G.J. & Cloos, J. (2010) High frequency type I/II mutational shifts between diagnosis and relapse are associated with outcome in pediatric AML: implications for personalized medicine. Blood.

Bacher, U., Haferlach, T., Alpermann, T., Kern, W., Schnittger, S. & Haferlach, C. (2013) Molecular mutations are prognostically relevant in AML with intermediate risk cytogenetics and aberrant karyotype. Leukemia, 27, 496-500.