去甲基化藥物對於急性骨髓性白血病治療的有效性與調控機轉

Abstract

急性骨髓性白血病(acute myeloid leukemia，以下簡稱血癌)定義為周邊血液或骨髓血液髓性芽細胞(myeloblast)數目超過20%。血癌主要治療為化學治療，若無法接受化療或是化療失敗的血癌病人預後很差。近年來研究致力於非化療藥物在血癌的治療，兩種去甲基化藥物 (hypomethylating agents): azacitidine和decitabine最早核准於治療骨髓再生不良症候群 (myelodysplastic syndrome, MDS)，隨後亦證實其在血癌的療效。去甲基化藥物結構類似核苷(nucleoside)，可以併入DNA，抑制DNA甲基轉移酶(DNA methyltransferases) 的作用而達到去甲基化之效果。血癌病人帶有相當多不同的基因突變，其中亦包含表觀遺傳(epigenetics) 相關的基因突變，例如: DNMT3A, TET2, ASXL1, IDH1, IDH2, EZH2。表觀遺傳機轉包含了DNA甲基化 (DNA methylation)、組蛋白調控 (histone modification)、核小體修飾 (nucleosome modeling)。越來越多的研究顯示表觀遺傳可能是血癌重要的致病機轉之一。 與一般化療藥物不同的是，臨床上評估去甲基化藥物療效至少要在治療兩個療程，也就是兩個月後。如果能夠在治療之初就能預測療效，對於臨床照顧病人將會有莫大的好處。直到最近，一篇大型研究發現帶有TP53突變的血癌對於decitabine治療反應很好(反應率高達100%)。可惜的是，其他幾個研究團隊並沒有辦法驗證帶有TP53突變對治療的影響。去甲基化藥物對於帶有TP53突變血癌有效的分子機轉也仍然不明瞭。TP53是一個重要的抑癌基因(tumor suppressor gene)，研究已經發現表觀遺傳與p53蛋白的調控關係。p53蛋白與其下游基因的甲基化程度會影響基因表達，進而影響化療或標靶藥物之有效性。P53蛋白功能異常(p53 dysfunction)是很多癌症，包括血癌重要的致病機轉之一。P53蛋白功能異常可以出現在mutant p53也可以出現在wild-type p53，因此無論p53突變與否都有可能因為這個抑癌基因失去功能導致癌症。使用藥物恢復P53蛋白功能異常是目前重要的研究主題。有鑑於去甲基化藥物對血癌的有效性機轉仍不明瞭，本研究試圖研究去甲基化藥物對p53路徑的調控。我們的研究假設是，臨床上去甲基化藥物對血癌的療效，可能是因為去甲基化藥物活化p53 pathway來促進細胞死亡。 在研究的第一部份，我們希望驗證TP53突變與去甲基化藥物對血癌的療效是否相關。我們回溯性收集113位使用過azacitidine的血癌病人，挑選出50位初診斷即接受Azacitidine治療病人的檢體，以次序代定序(next generation sequencing)的方式找出病人突變的基因，探討TP53以及其他的基因突變，是否可以預測azacitidine臨床治療的反應。我們的研究發現，20%血癌病人帶有TP53突變，對於Aza治療反應在TP53突變與非突變病人分別為50%和31.4%，並未達到顯著差異。 第二部分，我們將使用血癌細胞株(cell lines)來測試其對兩種去甲基化藥物的敏感性以及有效的機轉。我們將分別選用p53突變細胞株Kasumi-1和p53未突變的細胞株THP-1, OCI-AML3，來研究去甲基化藥物是否能恢復p53蛋白功能異常，以及是否能活化p53下游路徑的基因。不同的細胞株對於去甲基化藥物敏感性不同，我們可由生長曲線和細胞聚落生成實驗(colony formation assay)中觀察到，p53突變細胞株Kasumi-1對於去甲基化藥物最敏感。然而，去甲基化藥物無法改善Kasumi-1細胞株p53蛋白功能異常，且無論是p53蛋白與p53下游蛋白表現皆不受到影響。因此去甲基化藥物對Kasumi-1細胞的有效性並非來自於影響p53調控的凋亡和細胞週期路徑。 我們的研究結果可提供臨床醫師使用AZA治療血癌的參考。收集更多病人數目後，亦可發展重要的生物標記(biomarker)來預測臨床使用AZA治療的有效性。實驗室已建立去甲基化藥物處理細胞株的模型，奠定未來深入研究的基礎。將來，我們可利用分子生物技術改變p53的表達狀態，進一步了解去甲基化藥物和p53分子路徑的調控機轉。

Acute myeloid leukemia (AML) is defined as myeloblast more than 20% in either peripheral blood or bone marrow. The prognosis for acute myeloid leukemia (AML) patients who are unfit or refractory to chemotherapy is dismal. Hypomethylating agents (HMA), azacitidine (AZA) and decitabine (DAC), had been approved by United States Food and Drug Administration for treatment of high-risk myelodysplastic syndrome (MDS) and AML patients. HMA is incorporated into DNA and acts as a direct and irreversible inhibitor of DNA methyltransferases (DNMTs), thereby reversing the transcriptional silencing of tumor-suppressor genes. Genome-wide studies in AML identified several mutations in genes related to epigenetic modifications, such as DNMT3A, IDH1, IDH2, and EZH2, as well as altered methylation across the entire epigenome. These data highlight the importance of epigenetic mechanisms in the pathogenesis of AML. Different from chemotherapy agents, the onset of clinical response to HMA is insidious, which should be assessible at least 2 months later. It is clinically important to identidy suitable clinical and molecular determinants at initial diagnosis that may predict HMA response to tailor individualized treatment strategy. Recently, TP53 mutations were identified as crucial molecular determinants of response to DAC in AML and MDS patients in one study, with 100% of TP53-mutated patients respond to DAC. However, other studies failed to find the correlation between TP53 mutation status and clinical responses. Besides, the detailed molecular mechanisms underlying the sensitivity of AML to HMA remain to be elucidated. TP53 is a well known tumor suppressor gens, and studies have demonstrated the interplay between epigenetic changes and the p53 protein expression. Aberrant methylation of p53 target genes is inversely correlated with gene expression and has been linked to tumor response to HMA. It was well documented that p53 dysfunction, despite its mutation status, was rather frequent in various AML entities. Reversal of p53 dysfunction is an area of active research. We hypothesize that HMA exert their anti-leukemic effects via p53 activation and regulation of p53 target genes. We postulated that HMA treatment may increase the expression of p53 target genes. In the first part of our study, we aimed to investigate whether TP53 mutation was a true molecular determinant of HMA response in AML patients. We performed a retrospective analysis of 113 AML patients who received AZA treatment at National Taiwan University Hospital. Sixty-seven patients received AZA as first line theray, and the other 46 patients received AZA as salvage treatment. Next generation sequencing of 54 target genes that are commonly mutated in AML was performed in 50 patients receving AZA as first line therapy. We found that TP53 mutation occurred in 20% of the patients. The overall response rate was not significantly different between TP53-mutated (50%) and TP53 wild-type (31.4%) patients. In the second part, we investigated the effects of DAC and AZA on p53 pathways in vitro using human AML cell lines— OCI-AML3 and THP-1 cells (wild-type TP53), and Kasumi-1 cell line (mutant TP53). The cells were treated with nanomolar doses of AZA or DAC for 3 consecutive days, and then harvested for subsequent analysis. The sensitivity of HMA on individual cell lines was determined by the growth curves and methylcellulose colony formation assays. We investigated whether p53 and its target genes, such as those related to pro-apoptosis (Bax, Bmf, Bak1, Bad), cell cycle (Gadd45a), p53 regulator (mdm2) were activated following DAC or AZA treatment. Further, the results will be validated in primary leukemic blasts from patients. Different cell lines exhibit different sensitivity to HMA. Kasumi-1, a p53 mutant cell line, is the most sensitive one to HMA based on growth curve and colony formation assay. However, reversal of mutant p53 dysfunction is not evident in Kasumi-1 cells. There was no decline in p53 protein, and no changes in gene expression of p53 target genes related to cell cycles and apoptosis. Therefore, the inhibitory effects of nanomolar HMA on Kasumi-1 cells appeared to be independent of p53-mediated apoptosis or cell cycle pathways. Our data is useful as a guidance for clinicians to use AZA. We believe that our study will eventually provide useful clinical parameters and molecular determinants as important biomarkers to predict HMA response. The HMA treated cell line model is useful for future mechanistic studies. It is feasible to modify the p53 status and study the associated pathway changes in the future.