Repressing HIF-1α-induced HDAC9 contributes to the synergistic effect of venetoclax and MENIN inhibitor in KMT2Ar AML

Abstract

KMT2A-rearranged acute myeloid leukemia (KMT2Ar-AML) is an aggressive subtype of AML with poor response and prognosis. KMT2Ar-AML has been demonstrated to be sensitive to BCL2 inhibitor venetoclax (VEN), but these patients are unable to benefit from current VEN-based regimen (VEN plus azacitidine or low dose-cytarabine), so a novel and KMT2A rearrangement-specific targeting partner is required, and MENIN inhibitor (MEN1i) is a promising one. Herein, we investigated the effect and mechanism of VEN plus MEN1i in KMT2Ar-AML. Our results showed that VEN and MEN1i exhibited a striking synergistic effect in KMT2Ar-AML cell lines (in vitro), primary KMT2Ar-AML cells (ex vivo), and MOLM13 xenotransplantation model (in vivo). Furthermore, we found that VEN plus MEN1i significantly enhanced apoptotic induction in KMT2Ar-AML cell lines. VEN or MEN1i monotherapy disrupted balance of BCL-2/BCL-XL or down-regulated HOXA9/MEIS1, respectively, but these mechanisms were not further strengthened by their combination. RNA-Sequencing identified that HDAC9 was specifically repressed by VEN plus MEN1i rather than monotherapy. We demonstrated that HDAC9 was indispensable for KMT2Ar-AML proliferation and its repression contributed to proliferation inhibition of VEN plus MEN1i. Moreover, we found that hypoxia induced HDAC9 expression in KMT2Ar-AML, and VEN plus MEN1i inhibited hypoxia pathway, especially HIF-1A, and its target HDAC9. As our results indicated, VEN plus MEN1i-mediated HDAC9 down-regulation was partially dependent on HIF-1A repression in KMT2Ar-AML. Hypoxia induction sensitized KMT2Ar-AML to VEN plus MI-503-mediated proliferation inhibition and apoptosis induction. Therefore, repressing HIF-1A-induced HDAC9 contributed to the synergistic effect of VEN and MEN1i in KMT2Ar-AML.

To the editor,

KMT2Ar-AML exhibits poor response and prognosis, and new therapy is urgently required. KMT2Ar-AML is sensitive to VEN [1], but these patients fail to benefit from current VEN plus AZA/LDAC regimen [2, 3], so a novel partner typically targeting KMT2A rearrangements is needed. MENIN-KMT2A complex is required for KMT2Ar-AML initiation and maintenance. Consistently, KMT2Ar-AML is sensitive to disruption of MENIN-KMT2A [4], so MEN1i is a promising candidate. Herein, we investigated VEN plus MEN1i in KMT2Ar-AML (supplemental Materials and Methods).

Initially, we tested VEN or MEN1i MI-503 monotherapy in AML cell lines, and found that KMT2Ar-AML (MV4-11, MOLM13 and THP-1) was sensitive to both of agents (Fig. 1A-B; Fig. S1). To exclude MI-503-specific effects, we tested additional MEN1is (MI-463 and VTP-50469) and demonstrated their sensitivity in KMT2Ar-AML (Fig. S2). NPM1-mutated AML is also dependent on MENIN-KMT2A complex [5], but OCI-AML3 (NPM1-mutated AML) was less sensitive to MEN1is than KMT2Ar-AML (Fig. 1B; Fig. S3), so KMT2Ar-AML is the best indication of MEN1i.

Fig. 1

VEN and MEN1i synergistically inhibited KMT2Ar-AML proliferation. A IC50 of VEN in different AML cell lines (72 h); B IC50 of MI-503 in different AML cell lines (7 days); C Growth inhibition and synergistic index (combination index [CI] was calculated via CalcuSyn software) in THP-1 (expressing KMT2A-MLLT3), MV4-11 (expressing KMT2A-AFF1), and MOLM13 (expressing KMT2A-MLLT3) cells after treatments with different concentrations of VEN, MI-503, and their combination (72 h); D Apoptotic induction (Annexin V-positive cells) was determined after treatments with VEN, MI-503, and their combination (72 h) [two-tailed Student's t-tests; **P < 0.01, ***P < 0.001]; E Growth inhibition and synergistic index of VEN plus MI-503 in primary bone marrow MNCs from KMT2Ar-AML patients (72 h); F Detecting the progression of tumor burden via fluorescence imaging after treating the CDX mouse model constructed by MOLM13-Luciferase with DMSO, VEN, MI-503 and their combination. G The relative luminescence unit (RLU) value was calculated for indicating leukemic burdens of treated mice (28 days) [two-tailed Student's t-tests; *P < 0.05, **P < 0.01, ***P < 0.001]. H Kaplan–Meier survival curves in the MOLM13 xenotransplantation model [two-tailed Student's t-tests; *P < 0.05, **P < 0.01, ***P < 0.001]

Then, we investigated combination therapy of VEN and MEN1i in KMT2Ar-AML. Notably, an effective synergy in proliferation inhibition was observed in KMT2Ar-AML but not non-KMT2Ar-AML cell lines (Fig. 1C; Fig. S4-S5). Furthermore, we found that VEN plus MI-503 consistently enhanced apoptotic induction, but not cell cycle arrest or cell differentiation in three KMT2Ar-AML cells (Fig. 1D; Fig. S6). Besides, the synergistic effect of VEN plus MI-503 was selectively exhibited in primary KMT2Ar-AML but not non-KMT2Ar-AML or NBM cells (Fig. 1E; Fig. S7; Tab. S1). In vivo study, our results revealed that compared to VEN or MI-503 monotherapy, their combination exhibited better effects of reducing the leukemic burden and prolonging the survival duration in MOLM13 xenotransplantation model (Fig. 1F-H; Fig. S8). Therefore, VEN and MEN1i showed a striking synergistic effect in KMT2Ar-AML.

As reported, VEN and MI-503 mainly inhibited leukemia via disrupting balance of BCL-2/BCL-XL and down-regulating HOXA9/MEIS1, respectively. When VEN and MI-503 were combined, imbalance of BCL-2/BCL-XL or down-regulation of HOXA9/MEIS1 were not further enhanced by MI-503 or VEN, respectively (Fig. S9). To uncover their synergistic mechanism, RNA-sequencing was displayed in KMT2Ar-AML (Fig. S10). To select possible candidates contributing to their synergy, we adapted indicated strategies, and 34 genes, specifically regulated by combination therapy, were identified. After screening above genes based on their function and correlation with KMT2Ar-AML, and we focused on HDAC9 (Fig. 2A; Tab. S2, S3, S4, S5, S6, S7, S8, S9 and S10). High HDAC9 expression is associated with KMT2Ar-leukemia [6]. Therefore, HDAC9 was a possible target for VEN plus MI-503 in KMT2Ar-AML. First, we demonstrated that HDAC9, but not other HDACs, was consistently down-regulated in KMT2Ar-AML specifically at 72 h treatment (Fig. 2B; Fig. S11-12), and its protein was down-regulated accordingly, but this phenomenon was not observed in OCI-AML3 (Fig. 2C; Fig. S13). Second, HDAC9 inhibition with TMP-269 or shRNA impaired KMT2Ar-AML proliferation, though their HDAC9 levels were not all higher than non-KMT2Ar-AML cell lines (Fig. 2D-E; Fig. S14-15). Third, combination-mediated proliferation inhibition was not further enhanced by HDAC9 knockdown in KMT2Ar-AML (Fig. 2F). Therefore, HDAC9 was required for KMT2Ar-AML maintenance, and VEN plus MI-503 exerted HDAC9 repression to enhance their proliferative inhibition in KMT2Ar-AML.

Fig. 2

VEN plus MEN1i impaired KMT2Ar-AML growth via repressing HIF-1A-induced HDAC9. A After treating KMT2Ar-AML cell lines THP-1, MV4-11, MOLM13 with DMSO, VEN, MI-503 and their combination for 72 h, RNA sequencing was performed and differentially expressed genes (DEGs) were analyzed; Overall strategy for screening key DEGs related to the synergistic mechanism of VEN plus MI-503 in KMT2Ar-AML was shown; Specific DEGs in VEN plus MI-503 group vs. DMSO group shared by three KMT2Ar-AML cell lines were listed; B The mRNA expression of HDAC9 in THP-1, MV4-11, and MOLM13 after treatments with DMSO, VEN, MI-503, and their combination [two-tailed Student's t-tests; **P < 0.01, ***P < 0.001 ****P < 0.0001]; C The protein expression of HDAC9 and HIF1A in THP-1, MV4-11, and MOLM13 after treatments with DMSO, VEN, MI-503 and their combination; D IC50 of TMP-269 in THP-1, MV4-11, MOLM-13, U937, HL-60, Kasumi-1, and OCI-AML3 cells (72 h); E Growth inhibition was mediated by HDAC9 knockdown in THP-1, MV4-11, and MOLM13 cells; F Different therapeutic effects of VEN plus MI-503 in normal control or HDAC9 knockdown THP-1, MV4-11, and MOLM13 cells were exhibited; Relative absorbance value (optical density [OD] ratio) to normal control for each group was presented and the difference value of OD ratio between two groups was marked as numbers between the histogram bars; G Hypoxia pathway gene set was enriched in the VEN plus MI-503 group compared to DMSO group, the horizontal axis represents the p-value, the vertical axis represents the standard enrichment fraction (NES), the size of the circle represents the -log (p adjusted) value, and different colors represent different gene sets; H The protein expression of HDAC9 and HIF1A in THP-1, MV4-11, and MOLM13 treated with DMOG. I Under HIF-1A inhibitor BAY 87-2243 treatment, HIF-1A mRNA (left) and protein (middle) levels were evaluated. Furthermore, HDAC9 mRNA was detected in DMSO or VEN (400 nM) plus MI-503 (700 nM)-treated THP1 cells with or without BAY 87-2243 3-day pre-treatment (72 h) [two-tailed Student's t-tests; *P < 0.05, **P < 0.01, ***P < 0.001]. J Under normoxia and hypoxia, THP-1 cells were treated by DMSO or VEN plus MI-503, and cell viability and cell apoptosis were evaluated [two-tailed Student's t-tests; **P < 0.01, ***P < 0.001 ****P < 0.0001]

To explore how HDAC9 was repressed, we conducted GSEA and showed that hypoxia pathway was inhibited by VEN plus MI-503 (Fig. 2G; Fig. S16; Tab. S11). HDAC9 was reported to be induced by hypoxia, especially HIF-1A [7]. In our study, hypoxia inducer DMOG stabilized HIF-1A and up-regulated HDAC9 in KMT2Ar-AML (Fig. 2H). VEN plus MI-503 repressed HIF-1A and its target expression, while HDAC9 repression was attenuated in THP-1 cells with HIF-1A inhibition (Fig. 2C, I; Fig. S17). Moreover, VEN plus MI-503 exerted a more remarkable proliferation inhibition and apoptosis induction in THP-1 under hypoxia than normoixa (Fig. 2J). Thus, VEN plus MI-503 down-regulated HDAC9 via suppressing HIF-1A, which contributed to their synergistic effect.

MEN1i monotherapy exerts its best inhibition taking more than 7 days, and easily induced MEN1 mutations to mediate resistance [8]. VEN works within 48-72 h, so it can overcome the long onset duration of MEN1i and reduce chances of resistance in combination. As reported, both of VEN and MEN1i are oral available and target LSCs [9, 10], so VEN plus MEN1i is possibly a bi-oral and bi-LSCs-targeted regimen, which will make evolutions in therapeutic concept for KMT2Ar-AML. How VEN and MI-503 cooperate still lacks in recent studies [11, 12], and we demonstrated that beyond each effect of monotherapy, repression of HIF-1A-induced HDAC9 mainly contributed to enhanced proliferative inhibition in combination.

Collectively, targeted therapy with VEN plus MEN1i is promising in KMT2Ar-AML.