從轉錄體分析中發現具潛力的藥物能加強disulfiram對急性骨髓性白血病的抗癌效果並探討其機制

Abstract

急性骨髓性白血病 (acute myeloid leukaemia, AML) 的主要治療挑戰是復發或難治性，有高達57％的患者在完全緩解後出現此情況。而Disulfiram (DSF) 雖然於臨床前試驗中，在癌症治療方面取得令人鼓舞的結果，但其後發現，它在臨床上的應用會受到其不穩定性而影響，導致在血液中無法形成足夠的濃度的diethyldithiocarbamate-copper complex (CuET)。而DSF之所以可以抗癌，就是依賴這一含銅(copper, Cu)代謝物。因此，在本篇論文中會嘗試找出合適的藥物與DSF/Cu進行合併，以增強其在低濃度下的細胞毒殺作用，並從而增加DSF的臨床價值。 首先，我們分別評估了低濃度和高濃度DSF/Cu對兩種AML細胞株 (OCI AML 3和HEL) 的影響。低濃度組的作用是模擬臨床上可達到的CuET水平，而高濃度組則用來比較。結果顯示低濃度組的細胞毒殺作用輕微或不確定，而高濃度則表現出強大的細胞毒殺作用。此外，我們發現DSF/Cu的細胞死亡機制比較複雜，會誘導不同類型的細胞死亡。因此，我們認為於這方面進行更探入的探討可能不會對找出合適的藥物有所幫助。所以我們轉移焦點，探討DSF/Cu的藥物一開始的作用機制，即產生氧化壓力。結果如預期，我們發現能夠對細胞產生足夠的氧化壓力是DSF/Cu可以導致細胞死亡至關重要的一個機制。 得知這個機制後和經實驗室一些轉錄組的數據分析，我們選出PIM447作為第一個合併的候選藥物。然而，結果顯示與DSF/Cu合併後，其細胞毒殺效果不佳，僅表現出細胞生長的抑制效應，而非導致細胞死亡。 為更清楚瞭解DSF/Cu的分子機制，我們對OCI AML 3細胞進行了RNA定序分析 (RNA sequencing, RNA-seq)。基因集豐富分析(GSEA)結果與目前的研究結果一致，上調的基因路徑包括氧化壓力的反應、金屬或銅有關的反應，以及蛋白質錯誤摺疊的反應。而下調的路徑包括DNA複製的路徑和粒線體功能的路徑。 為了找到合適的藥物合併，我們利用RNA-seq數據構建了蛋白質交互作用(protein-protein interaction, PPI)網路。並從中得出49個關鍵基因相信是與DSF/Cu所誘導的細胞死亡有關。通過檢索不同藥物數據庫，結果得出了三種候選藥物，rapamycin (sirolimus), zinc protoporphyrin (ZnPP)，以及APR-246 (Eprenetapopt)，適合與低濃度的DSF/Cu 合併。 在這三種候選藥物中，APR-246表現出較好的效果，在對細胞死亡和誘導粒線體ROS的產生的測試中都得出有協同效應。進一步的分析顯示，這兩個結果是可以通過影響硫醇/二硫化物(thiol/disulfide)的平衡和誘導未折疊蛋白反應（UPR）來連結，產生因果關係。

A major challenge in treating acute myeloid leukaemia (AML) is relapsed or refractory, which affects up to 57% of patients after complete remission (CR). Despite promising preclinical results in cancer therapy, the clinical application of disulfiram (DSF) was limited by its poor stability and the inability to form the diethyldithiocarbamate-copper complex (CuET), an ultimate anticancer copper-containing metabolite of DSF, at sufficient concentrations. Hence, combining DSF/Cu with an appropriate drug to enhance its cytotoxicity at low concentration may be a way to enhance the clinical value of DSF. In this study, we examined the effects of both low and high concentrations of DSF/Cu on two AML cell lines (OCI AML 3 and HEL). The low concentration group (L group) simulated the clinically achievable levels of CuET, while the high concentration group (H group) serving as a direct comparison. The results demonstrated minimal or ambiguous cytotoxicity in the L group, while potent cytotoxic effects were observed in the H group. Furthermore, we revealed a complex cell death mechanism of DSF/Cu that is capable of inducing different types of cell death. Therefore, it prompted us to shift our focus to investigating the drug mechanism of DSF/Cu from the beginning, that is oxidative stress effects. As expected, we found that induction of an overwhelming oxidative stress is a critical mechanism to contribute to the cell death of DSF/Cu. Based on this perspective and transcriptomic analysis, we firstly evaluated PIM447 as a candidate for combination. However, the results indicated suboptimal cytotoxic effects, with only a cytostatic effect rather than cell death. To better understand the mechanism of DSF/Cu, RNA-sequencing (RNA-seq) was performed in OCI AML 3 cells. Gene set enrichment analysis (GSEA) results were consistent with current studies, the upregulated pathways including response to oxidative stress, response to metal or copper ions and protein misfolding response. Additionally, our GSEA results indicated that the DNA replication pathways and mitochondrial function pathways were downregulated. To identify an appropriate druggable target for combination, a protein-protein interaction (PPI) network was constructed using the RNA-seq data. A total of 49 essential genes were identified from PPI network that presumably involved in DSF/Cu-induced cell death. By searching the drug database, three drug candidates were suggested for combination with low concentration of DSF/Cu, which are rapamycin (sirolimus) zinc protoporphyrin (ZnPP) and APR-246 (Eprenetapopt). Among these candidates, APR-246 in combination with a low concentration of DSF/Cu exhibited the most pronounced effectiveness, demonstrating a synergistic effect on the induction of mitochondrial ROS and cell death. Further analysis revealed that these two effects were interconnected through the perturbation of thiol/disulfide homeostasis and the induction of unfolded protein response (UPR).