To the Editor,

Despite recent advances in the treatment of acute myeloid leukemia (AML), a subset of patients remains incurable, suggesting an unmet effective therapeutic need [1]. Dysfunction of p53, a renowned tumor suppressor, is often associated with poor outcomes in AML, thus activation of p53 roles is intensely studied [2, 3]. Many studies endeavor to pharmacological inhibition of MDM2, a p53 negative regulator, while p53 activation with MDM2 inhibitor alone only results in limited clinical activities in AML [4, 5]. Thus, how to potentiate the anti-AML efficacy of MDM2 inhibitor is worthy of further investigation.

Triptolide, a traditional Chinese medicine, exerts potent anti-tumor activities in many malignant disorders [6]. Others and our prior studies have reported that low-dose (LD) of Triptolide could sensitize AML to multiple conventional chemotherapies and targeted agents [7,8,9]. In this study, we sought to examine the antileukemia effects of LD Triptolide in conjunction with MDM2 inhibitor Nutlin-3a on AML. Our results showed that LD Triptolide (~ 20 nM) synergized with Nutlin-3a to inhibit cell proliferation, promote caspase-cascades apoptosis, and increase nuclear instability in wild-type p53 (p53 wt) cells (Fig. 1A, B; Additional file 1: Fig. S1A–D). The synergy was evidenced by the combination index (CI) < 1, ranging from 0.18 to 0.89. Mitochondrial membrane potential (MMP) usually collapses before intrinsic apoptosis takes place. We found that LD Triptolide cooperated with Nutlin-3a to induce an obviously decreased MMP levels, indicating that mitochondrial-mediated apoptosis partially contributes to the synergistic effect (Additional file 1: Fig. S2A, B). In agreement with these in vitro data, LD Triptolide cooperated with Nutlin-3a to induce apoptosis in primary p53 wt AML cells, and to slow tumor growth and ameliorate leukemia burden in mice without unacceptable adverse effects (Fig. 1C–F; Additional file 1: Fig. S2C, D; Additional file 2: Table S1). Mechanically, Nutlin-3a transcriptionally upregulated p53 downstream targets (p21 and PUMA) rather than p53 (Fig. 1G). LD Triptolide downregulated MDM2 mRNA levels and decreased the transcriptional abundances of two anti-apoptotic components, Mcl-1 and XIAP (Fig. 1H).

Fig. 1
figure 1

The synergistic interaction of LD Triptolide and MDM2 inhibitor in p53 wt AML cells. AML cell lines harboring wild type p53 were treated with control, Nutlin-3a (N3), LD Triptolide (TPL) or their combination for 48 h (OCI-AML3, MV4-11) or 24 h (MOLM13). A Cell viability was determined with the CCK-8 kit. B Annexin-V/PI dual staining was used to analyze cell apoptosis. A schematic plan for the in vivo study. C Analysis of apoptotic cells in p53 wt primary AML patients (n = 19). D A schematic plan for the in vivo study. E Tumor volumes and weight of each group were separately measured. F Images of subcutaneous tumors were captured at the end of the experiment. G, H qRT-PCR analysis of the mRNA levels of p53, PUMA, P21, MDM2, XIAP and MCL-1 in OCI-AML3 cells exposed to Nutlin-3a or LD Triptolide treatment for 2, 8 and 16 h. Data were presented as mean ± S.D. ns indicates not significant, *p < 0.05, **p < 0.01, ***p < 0.001, colored by corresponding single drug treatment in statistical tests

Full size image

To ascertain the role of p53 status in this combined regimen, we silenced p53 using shRNA in p53 wt cells and found that Nutlin-3a failed to induce apoptosis in the p53 knockdown cells. Knocking down p53 only partially impeded the proapoptotic activity of LD Triptolide monotherapy or in combination with Nutlin-3a (Fig. 2A), implying that p53-independent molecular bases might play roles in Triptolide cytotoxicity. We indeed confirmed this implication, as Triptolide exhibited a robust cytotoxicity against all five p53 deficient primary AML cells and three p53 deficient cell lines (Fig. 2B, C; Additional file 2: Table S1). The synergistic effect of Triptolide and Nutlin-3a was observed in Kasumi-1 carrying p53 p.R248Q mutation and 2 out of 5 p53 deficient primary AML cells, while the synergy was not evident in p53 null cells(THP-1 and HL-60) (Fig. 2B, C; Additional file 2: Table S1).

Fig. 2
figure 2

The p53 independent role of Triptolide in p53 deficent AML cells. A Apoptotic Analysis of OCI-AML3 cells with vehicle (red lines) or p53 shRNA (blue lines) treated with control, Nutlin-3a, Triptolide, or their combination for 24 h. B Analysis of apoptotic cells in p53 deficient primary AML patients (n = 5) treated as indicated doses for 48 h. C Cells viability was detected in AML p53 mut/null cell lines treated as indicated doses for 48 h (mut: Kasumi-1; null: THP-1, HL-60). D The GO and hallmark pathway analysis was performed to reveal the enriched signaling for the 482 DEGs. E GSEA analysis of the annotations related to ER stress pathway. F Heatmap displayed hierarchical clustering of ER related genes among control,Triptolide alone or Triptolide in combination with Nutlin-3a. G, H Immune blot assessed the protein levels of MYC and ATF4 in OCI-AML3 (p53 wt) and THP-1 (p53 deficiency) cells exposed to LD Triptolide and Nutlin-3a alone or in combination. ns indicates not significant, *p < 0.05, **p < 0.01, ***p < 0.001, colored by corresponding single drug treatment in statistical tests

Full size image

To better understand the p53 independent function involved in Triptolide treatment, we performed RNA-Seq in OCI-AML3 cells treated with Triptolide with or without Nutlin-3a. We identified 482 shared differential expression genes (DEGs, fold change > 2, FDR < 0.05) (Additional file 1: Fig. S3A, B; Additional file 3) and observed that Triptolide significantly attenuated the activity of several pathways related to protein synthesis processes (UPR, mTORC1, MYC targets) (Fig. 2D, E). The ER stress associated genes, especially MYC and ATF4 targeted subsets (e.g., MYC, ATF4, TRIB3, CHAC1, DITT4), were remarkably blocked by Triptolide and the combined treatment (Fig. 2F). The proto-oncogene MYC enhances protein synthesis that is essential to induce an adaptive ER stress response(unfolded protein response, UPR) in tumor progression [10]. As a vital piece of UPR, PERK-eIF2a-ATF4 signaling is selectively activated to clear misfolded proteins with MYC hyperactivation [10, 11]. Our subsequent functional assays demonstrated that MYC and ATF4 was extremely suppressed by Triptolide and its combination with Nutlin-3a, while PERK and DDIT4 exhibited no significant changes in both OCI-AML3 and THP-1 cells. Interestingly, the synergistic effect was observed in MYC deletion at early stage(16 hs) and in ATF4 deletion at later stage(24–48 h) of treatment (Fig. 2G, H). We suppose that ATF4 probably served as a key UPR effector in response to Triptolide treatment.

In summary, our study offered a novel potent therapeutic regimen by combining LD Triptolide with MDM2 inhibitor against p53 wt AML. More importantly, this drug combination was also active in some AML models with p53 deficiencies. The synergism was closely related to activation of p53 relevant proapoptotic cascades in p53 wt AML. Beyond the p53 reliant functions, Triptolide also exerted its antileukemia efficacy through interfering with the MYC-ATF4 axis, a p53 independent pathway. These data provided a potential rationale for clinical administration for the treatment of patients with AML [12], while prospective studies are warranted to further validate this hypothesis.