Development of a highly potent Menin-MLL1 inhibitor

Through high throughput screening (HTS) and derivatization from the lead scaffolds, we developed DS-1594a·succinate (the succinic acid salt of DS-1594a; (1R, 2S, 4R)-4-(amino)-2-cyclopentan-1-ol; Fig. 1A), which contained unique chemical structures such as (1R, 2S, 4R)-cyclopentan-1-ol and 5,6-dimethoxypyridazine moieties compared with those of other Menin-MLL1 inhibitors [33, 34]. These structures enabled DS-1594a·succinate to display the most favorable drug-like properties among our derivatives and potent inhibitory activity against the Menin-MLL1 interaction. For crystallographic study of the interaction of DS-1594a·succinate with Menin, human Menin with deletion of residues 460–519 was expressed in Escherichia coli and purified as described previously with some modifications [35]. Thereafter, the crystal structure of Menin bound to DS-1594a·succinate was solved to reveal the molecular basis of DS-1594a·succinate-mediated inhibition (Fig. 1B). The structure contains 4 molecules in the asymmetric unit. However, we focused on the chain A interaction mode, as no significant interaction mode changes were observed among chains A-D. The structure demonstrated a number of hydrophobic interactions between DS-1594a·succinate and surrounding amino acids that likely contributed to DS-1594a·succinate’s strong inhibitory activity. It also demonstrated that DS-1594a·succinate forms a hydrogen bond with Tyr276 and engages in π-π or CH-π interactions with Phe238, Met278, Tyr319, Met322, Tyr323, and Trp341 (Additional file 1: Figure S1A), which likely contributes to its target specificity. Notably, the 5,6-dimethoxypyridazine moiety seems to contribute to both inhibitory activity and target specificity through the formation of typical CH-π interactions with the Trp341 side chain (Fig. 1C) and through extremely good shape complementarity with the surrounding pocket (Additional file 1: Figure S1B).

Fig. 1

Structure and in vitro activity of the Menin-MLL1 inhibitor DS-1594a·succinate. A Chemical structure of DS-1594a·succinate (succinic acid salt of DS-1594a). B Overall structure of the DS-1594a·succinate/human Menin complex showing 4 chains of Menin included in 1 asymmetric unit (left) and Fo-Fc electron density map of DS-1594a·succinate bound to chain A of Menin contoured at 3 sigma (right). C Inhibition of the Menin-MLL1 interaction by DS-1594a·succinate in a cell-free assay (AlphaLISA assay). The inhibition (%) of the Menin-MLL1 interaction by DS-1594a·succinate is shown as the mean ± SD at each concentration of DS-1594a·succinate (1000, 170, 28, 4.6, 0.77, 0.13, 0.021, and 0.0036 nM; n = 4). D Inhibition of the Menin-MLL1 interaction by DS-1594a·succinate in a cellular assay. Co-immunoprecipitation (Co-IP) was performed with an anti-HA antibody or control IgG in human embryonic kidney 293 T (HEK293T) cells transfected with HA-MLL1N1-1395 upon treatment with DMSO (0) or DS-1594a·succinate (10, 30, 100, 300, 1000 nM). The input represents lysate samples before Co-IP treatment. E In vitro growth-inhibitory effects of DS-1594a·HCl (the HCl salt of DS-1594a) against leukemia cell lines (left). Titration curves from the cell viability assay (CellTiter-Glo) were created after 7 days of treatment of human leukemic cell lines with DS-1594a·HCl. The horizontal axis is the common logarithm of the concentration of DS-1594a·HCl, and the mean (± SD) survival rates at each concentration are shown (n = 3). The GI50 concentrations of DS-1594a·HCl for each cell line are shown (right). B-ALL, B-cell acute lymphoblastic leukemia; CML, chronic myeloid leukemia; APL, acute promyelocytic leukemia; Co-IP, coimmunoprecipitation

DS-1594a·succinate demonstrated inhibitory activity against the Menin-MLL1 interaction in the subnanomolar range (with a half-maximal inhibitory concentration of 1.4 nM) in a cell-free AlphaLISA assay using a full-length recombinant Menin protein and MLL1 peptide (Fig. 1D and Additional file 2). DS-1594a·succinate was characterized in detail in human embryonic kidney cells (HEK293T) expressing an HA-human MLL1​N​ truncated protein (HA-MLL1N1-1395) to assess its ability to block the Menin-MLL1 interaction, as indicated by coimmunoprecipitation of Menin and HA-MLL1N1-1395. Our results demonstrated that DS-1594a·succinate could reach the target protein and inhibit endogenous Menin protein and the exogeneous HA-MLL1N1-1395 interaction at more than 30 nM in a dose-dependent manner (Fig. 1E).

Thereafter, we tested DS-1594a·HCl (HCl salt of DS-1594a) activity in leukemic cell lines with and without MLL1-r and NPM1c. DS-1594a·HCl markedly inhibited the cellular growth of leukemic cells harboring various MLL fusion proteins (MV-4–11, MOLM-13, KOPN-8) and NPM1c (OCI-AML3), with 50% growth inhibition (GI50) values between 2.5 and 28.5 nM. In contrast, DS-1594a·HCl did not inhibit leukemic cell growth (exhibiting > 350-fold selectivity) without MLL1-r or NPM1c (OCI-AML5, KG-1, HL-60 and K-562; Fig. 1F).

The Menin-MLL1 inhibitor induced differentiation and loss of clonogenic potential of human MLL-AF9–evoked murine AML-like cells

We evaluated the effects of DS-1594a·HCl and DS-1594a·succinate on differentiation and population of leukemia-initiating cell (LIC) fractions in MLL1-r AML-like cells, using RT‒qPCR, MGG staining, FCM, serial colony formation and growth assays. MLLfusion target gene expression in human MLL-AF9–evoked murine AML-like (MA9) cells was assessed after 4 days of treatment with DS-1594a·HCl. RT‒qPCR confirmed that DS-1594a·HCl exhibited concentration-dependent inhibition of Meis1, Hoxa9, Mef2c, and Pbx3 expression (Fig. 2A).

Fig. 2

The Menin-MLL1 inhibitor induced differentiation and loss of LICs of human MLL-AF9–evoked murine AML-like cells A RT‒qPCR was performed in MLL-AF9–evoked murine AML-like cells after treatment with DS-1594a·HCl (1.5, 4.6, 14, and 41 nM) for 4 days. The expression levels of Meis1, Hoxa9, Mef2c, and Pbx3 were normalized to that of Actb and compared to those in the DMSO-treated control (mean ± SD; n = 3). B FCM analysis with the indicated antibodies (cKit, CD11b, Ly-6G, and Annexin V) for MLL-AF9–evoked murine AML-like cells after 7 days of treatment with DMSO, DS-1594a·succinate (10, 20, and 40 nM), or Ara-C (50 and 100 nM). The bars represent the mean ± SD; n = 3. C Serial replating in the MethoCult M3434 assay. The colonies were counted 4–5 days after seeding with DMSO, DS-1594a·HCl (10, 20, and 40 nM), or Ara-C (50 and 100 nM). The first colony (MC1) was counted, collected, and replated (MC2–MC5) in fresh MethoCult M3434 with DMSO or test compounds (n = 2, represented by each circle). D The cellular growth and colony formation activity (MethoCult M3434 assay) of MLL-AF9–evoked murine AML-like cells were monitored after washout of DS-1594a·HCl or Ara-C pretreatment for 7 days. Ara-C, cytarabine

This downregulation of MLL-fusion target gene expression occurred with an increase in the number of differentiated cells on microscopic examination of MGG–stained murine MA9 cells. The number of cells differentiated into granulocytic lineages with lobulated cell nuclei and loss of cytoplasmic color increased in a DS-1594a·HCl concentration-dependent manner at final concentrations of 10, 20, and 40 nM for 7 days, in contrast to the findings among cells treated with dimethyl sulfoxide (DMSO) or standard chemotherapy drug cytarabine (Ara-C; Additional file 1: Figure S2A). The proportion and extent of differentiation induction were assessed by counting of the differentiated and undifferentiated cells. Compared with the percentage after treatment with DMSO (control; 87%), the percentage of blasts, the most undifferentiated cell type, decreased in a concentration-dependent manner to 51%, 20%, and 5.7% after treatment with 10, 20, and 40 nM DS-1594a·HCl, respectively. Furthermore, the percentage of neutrophils, the most differentiated cell type, increased in a concentration-dependent manner to 1.4%, 6%, and 33% after treatment with 10, 20, and 40 nM DS-1594a·HCl, respectively, while the percentage remained 0% after treatment with DMSO (control, 0%; Additional file 1: Figure S2B).

FCM analysis of murine MA9 cells treated with DMSO, DS-1594a·succinate, or Ara-C for 7 days was performed with antibodies against the LIC-associated marker c-KIT [12, 36], the granulocytic and myeloid differentiation markers CD11b and Ly-6G, and the apoptosis marker Annexin V (Fig. 2B, Additional file 1: Figure S2C, S2D and S2E). Our results showed that the CD11b + Ly-6G + and Annexin V + fractions increased by 50% and 19%, respectively, with 40 nM DS-1594a·succinate, while the c-KIT + fraction decreased by 11%. No significant increases or decreases in the CD11b + Ly-6G + , c-KIT + , and Annexin V + fractions were observed with 100 nM Ara-C, the standard-of-care drug for AML treatment, compared with the DMSO control.

The colony-forming and self-replicating ability of murine MA9 cells was tested using a MethoCult assay. Unlike for DMSO-treated cells (100%), no colonies were detected in the fifth serial replating of 40 nM DS-1594a·HCl-treated cells, while ⁓50% of colonies were detected in 100 nM Ara-C-treated cells (Fig. 2C). To verify the durable efficacy after DS-1594a·HCl treatment termination, MethoCult assays were performed after compound washout in murine MA9 cells treated with DMSO, DS-1594a·HCl, or Ara-C for 7 days. On day 7 after washout (day 14), the proliferative and colony-forming ability of the 40 nM DS-1594a·HCl–treated cells was eliminated, whereas that of the 200 nM Ara-C–treated cells was only slightly suppressed and not suppressed compared with that of the DMSO-treated cells on days 10 and 14, respectively (Fig. 2D). Overall, our results indicate that the Menin-MLL1 inhibitor might deplete cKit + fractions reflecting LICs via strong differentiation-inducing activity with suppression of Menin-MLL1 regulatory gene expression even at very low concentrations.

The Menin-MLL1 inhibitor induced differentiation and loss of CD34 + /CD38 − cells in patient-derived MLL1-r + AML cells (AML#8531, AML676, and NCCHD010) in vitro

We evaluated the effect of DS-1594a·succinate on induction of differentiation and loss of CD34 + /CD38 − cells, which are a fraction of potential LICs, in patient-derived MLL1-r + AML cells (AML#8531, AML676, and NCCHD010) in vitro. Compared with DMSO, DS-1594a·succinate induced strong CD34 downregulation in a dose-dependent manner, leading to decreased CD34 + /CD38 − cell frequencies (Fig. 3A and Additional file 1: Figure S3A). Consistent with the loss of the CD34 + /CD38 − fraction, an increased proportion of cells expressing the monocytic differentiation marker CD11b was observed after DS-1594a·succinate in AML#8531, AML676 and NCCHD010 cells (Fig. 3B and Additional file 1: Figure S3B). The DS-1594a·succinate-mediated downregulation of CD34 and upregulation of CD11b were simultaneously confirmed by RT‒qPCR in AML#8531 and AML676 cells (Additional file 1: Figure S3C).

Fig. 3

The Menin-MLL1 inhibitor induced differentiation and loss of CD34 + /CD38 − cells in patient-derived MLL1-r + AML cells (AML#8531, AML676, and NCCHD010) in vitro. Patient-derived primary AML cells (MLL1-r AML#8531, NCCHD010, AML676) were treated for 7 days with DMSO, DS-1594a·succinate, or Ara-C at the indicated concentrations. A FCM analysis with anti-human CD34/CD38 antibodies. The bars represent the mean ± SD; n = 3. B FCM analysis with anti-human CD11b. The bars represent the mean ± SD; n = 3. C RT‒qPCR to detect the expression levels of MEIS1 and HOXA9 (mean ± SD; n = 3) normalized to that of ACTB

In parallel, DS-1594a·succinate greatly reduced expression of MLL-fusion target genes, MEIS1 and HOXA9, dose dependently in patient-derived MLL1-r + AML cells (Fig. 3C), as in NPM1c AML patient-derived cells (AML#7789, AML#7915, AML#7919; Additional file 1: Figure S4). These results suggest that enhancement of differentiation and potential loss of LIC fraction via reductions in MEIS1 and HOXA9 expression are major mechanisms of action for DS-1594a antitumor activity.

The Menin-MLL1 inhibitor induced gene expression changes through Menin-MLL1 chromatin-associated complexes

To investigate the effects of Menin-MLL inhibitors on global gene expression in MLL1-rearranged leukemias, we performed RNA-seq with RNA from a human MLL1-r AML cell line (MOLM-13) and patient-derived primary AML with MLL1-r (NCCHD010) cells treated with DS-1594a·succinate for 3 and 7 days, respectively. Significant gene expression changes were found for 560 vs 186 genes and 737 vs 271 genes were ≥ twofold upregulated vs downregulated in MOLM-13 (100 nM DS-1594a·succinate) and NCCHD010 (500 nM DS-1594a·succinate) cells (p ≤ 0.05). GSEA was also used to determine changes in MLL-fusion target–related signature expression in DS-1594a·succinate–treated MOLM-13 and NCCHD010 cells. We used the gene list from prior publications; of 139 listed MLL-AF9 target gene members [37], 105 genes that could be mapped in humans were used, and 69 and 85 target genes that were upregulated and downregulated, respectively, in hematopoietic precursor cells conditionally expressing HOXA9 and MEIS1 were used [38, 39]. GSEA demonstrated strong downregulation of MLL-AF9 downstream targets and HOXA9- and MEIS1-upregulated genes with DS-1594a·succinate in MOLM-13 and NCCHD010 cells. GSEA also showed upregulation of HOXA9- and MEIS1-downregulated genes in MOLM-13 and NCCHD010 cells (Fig. 4A). DS-1594a·succinate concentration-dependently reduced the expression of various MLL-fusion target genes, including MEIS1, MEF2C, PBX3, JMJD1C, HOXA7, and HOXA9, in RNA-seq studies (Additional file 1: Figure S5A), and these gene expression changes were validated by RT‒qPCR (Fig. 4B).

Fig. 4

The Menin-MLL1 inhibitor induced gene expression changes through Menin-MLL1 chromatin-associated complexes. A RNA-seq was performed in MOLM-13 and NCCHD010 cells treated with DS-1594a·succinate for 3 and 7 days, respectively. GSEA of MOLM-13 cells treated with 100 nM DS-1594a·succinate for 3 days and NCCHD010 cells treated with 500 nM DS-1594a·succinate for 7 days. B RT‒qPCR was performed in MOLM-13 and NCCHD010 cells treated with DS-1594a·succinate for 3 and 7 days, respectively, at the indicated concentrations. The graphs represent the expression levels of the indicated genes compared to those in the DMSO-treated control (mean ± SD, n = 3). C Changes in Menin/MLL1 occupancies in MOLM-13 and NCCHD010 cells treated with 100 nM or 5 μM DS-1594a·HCl or DS-1594a·succinate, respectively, for 3 days. Normalized coverage tracks of Menin, MLL1, H3K79me2, and H3K4me3 ChIP-seq signals (reads per million) at selected target genes in MOLM-13 or NCCHD010 cells are shown. The peaks around the TSS are shown. D Analysis of the leukemic stem cell score. LSC17 scores and pLSC6 scores in AML676 and NCCHD010 cells were calculated for each sample treated with DMSO or DS-1594a·succinate at the indicated concentrations (n = 3). The numbers on the plots are P values of the Jonckheere-Terpstra trend test with Bonferroni correction. For the GSEA, the t test results were used as the metrics for scoring and ranking genes. The FDR q value for each gene set was calculated with the permutation of the gene sets. LSC17, 17-gene adult LSC; pLSC6, 6-gene pediatric LSC; TSS, transcription start site

As DS-1594a·HCl and DS-1594a·succinate treatment disrupted Menin-associated protein complexes and MLL fusion-driven gene expression, we assessed the chromatin occupancy of Menin and MLL1 proteins and histone modifications important in MLL1-r leukemia, including H3K79me2 and H3K4me3. ChIP-seq was performed on the genomic binding sites and the dynamics of Menin and MLL1 interactions with chromatin after DS-1594a·HCl or DS-1594a·succinate treatment (3 days) in MOLM-13 (100 nM DS-1594a·HCl) and NCCHD010 (500 nM or 5 μM DS-1594a·succinate) cells. Menin binding analysis confirmed that DS-1594a·HCl and DS-1594a·succinate treatment led to local loss of Menin binding to chromatin. The sites and magnitude of Menin loss were similar between MOLM-13 and NCCHD010 cells. We observed reduced levels of MLL1, H3K79me2, and H3K4me3 at genes including MEF2C, MEIS1, PBX3, and JMJD1C, but not at the HOXA genes RUNX1 and MYB (Fig. 4C; Additional file 1: Figure S5B). Together, the data suggest that only a subset of MLL target genes are sensitive to inhibition of Menin, possibly by either genetic deletion or pharmacologic perturbation.

Finally, we evaluated stemness/LSC scores to assess whether DS-1594a·succinate impairs LSC-enriched gene expression in patient-derived primary MLL1-r AML cells (AML676, NCCHD010). Adult and pediatric stemness scores, assessed by 17-gene adult LSC (LSC17) and 6-gene pediatric LSC (pLSC6) scores, respectively, were significantly decreased in DS-1594a·succinate–treated cells, indicating the biological impact on potential fractions of LSCs or LICs (Fig. 4D).

The Menin-MLL1 inhibitor showed in vivo antitumor efficacy in a MOLM-13 xenograft model

DS-1594a·HCl’s in vivo antitumor efficacy was investigated using an intravenously inoculated MLL1-r AML leukemia xenograft model. MOLM-13 cells were transplanted into NSG mice to induce systemic leukemia, and disease progression was monitored. FCM analysis of bone marrow cells from the hindlimb bones showed that the human CD45 (hCD45) positivity rate in myeloid cells was ~ 76% in the vehicle group vs. ~ 16% in the 50 mg/kg DS-1594a·HCl group. Furthermore, the percentages of hCD45 + cells in the 100 and 200 mg/kg DS-1594a·HCl groups decreased to ≤ 1% (Fig. 5A).

Fig. 5

The Menin-MLL1 inhibitor showed in vivo antitumor efficacy in a MOLM-13 xenograft model. A FCM analysis to assess tumor burden (hCD45 +) in BM 1 day after 17 days of treatment (days 3–19). BM cells from NI NSG mice or NSG mice intravenously inoculated with MOLM-13 cells 1 day after 17-day treatment (day 20) with vehicle (CTL, QD × 17) or DS-1594a·HCl (50, 100, 200 mg/kg, QD × 17). The horizontal bars represent the mean tumor burden (n = 3 mice per group). *p = 0.0005, **p < 0.0001 (Dunnett’s test). B RT‒qPCR was performed for BM cells from MOLM-13 xenografts 6 h after dosing with vehicle (CTL, QD × 7) or DS-1594a·HCl (50, 100, 200 mg/kg, QD × 7). The expression levels of MEIS1, HOXA9, PBX3 and MEF2C were normalized to that of ACTB (mean ± SD, n = 3). C)Kaplan–Meier survival curves of vehicle- or DS-1594a·HCl–treated MOLM-13 xenograft mice (days 3–21, QD × 19; n = 6 mice per group). D Representative images from H&E staining of BM from NI mice (day 123), CTL mice (day 8), and 100 mg/kg DS-1594a·HCl-treated mice (day 123). The significance of the survival curves was analyzed by the Kaplan‒Meier log-rank test, with survival time as the indicator of the life-prolonging effect of the test compound.QD, once daily

RT‒qPCR analysis of RNA from hindlimb bone marrow cells revealed that the DS-1594a·HCl–treated group showed reductions in the expression of Menin-MLL1-regulated genes such as MEIS1, HOXA9, PBX3, and MEF2C by up to 96%, 83%, 96%, and 99%, respectively (Fig. 5B). These results indicate that DS-1594a·HCl suppresses the expression of Menin-MLL1–regulated genes in MOLM-13 cells in murine bone marrow, which is correlated with repression of MOLM-13 cell growth.

The persistence of DS-1594a·HCl–mediated AML cell growth inhibition in MOLM-13 xenograft mice was evaluated using survival tests. In the vehicle group, all mice died within 21 days of transplantation. In contrast, the DS-1594a·HCl 50, 100, and 200 mg/kg-treated groups showed a significantly pronounced survival benefit after treatment termination (increase in life span [ILS]: 82.5%, > 515%, > 515%, p = 0.0018; Fig. 5C). DS-1594a·HCl-treated mice (100 and 200 mg/kg) that were alive 123 days after transplantation remained in good condition without weight loss at all times after treatment termination. In these mice, the hCD45 positivity rate in the bone marrow for the 100 mg/kg DS-1594a·HCl dosage group was ≤ 1% and similar to that in the non-inoculated (NI) group (data not shown), suggesting that MOLM-13 cells may have been completely eradicated from the bone marrow.

In the control (vehicle-treated) MOLM-13 xenograft group, the femoral bone marrow cavities of all mice (n = 3) contained predominantly neoplastic cells, with minimal residual normal bone marrow. Tumor cell proliferation was also detected between the cortical bone and periosteum on the distal end. However, no neoplastic cells were found in the femoral marrow cavities of any mice that survived (n = 5) at day 123 in the DS-1594a·HCl 100 mg/kg group, as in the femurs in the NI group, indicating similar normal marrow status (Fig. 5D).

The Menin-MLL1 inhibitor showed in vivo antitumor efficacy in AML-ALL PDX models

The results obtained with MOLM-13 xenografts prompted further investigation into the efficacy of the Menin-MLL1 inhibitor on patient-derived xenograft (PDX) models. In vivo antitumor efficacy of DS-1594a·HCl in an NPM1c AMLPDX model (AM7577) was evaluated by treating with vehicle or DS-1594a·HCl (25, 50, 100, or 200 mg/kg) for 24 and 35 days, respectively. DS-1594a·HCl dosing groups showed complete elimination of hCD45 + AML blast cells in peripheral blood, which was sustained for 40 days after the dosing termination on day 35 (Fig. 6A). Similarly, DS-1594a·HCl induced significant reductions in the proportions of hCD45 + cells in the peripheral blood, spleen, and bone marrow in the DS-1594a·HCl group vs. the vehicle group (p < 0.05; Additional file 1: Figure S6) at the time of mouse termination. Survival was significantly prolonged in the DS-1594a·HCl 25 mg/kg group vs. the vehicle group (p < 0.01, ILS > 296%; Fig. 6B). No body weight loss was observed in the vehicle group or in the DS-1594a·HCl groups for any doses (data not shown).

Fig. 6

The Menin-MLL1 inhibitor showed in vivo antitumor efficacy in AML-PDX (A-B) and ALL-PDX (C-D) models. A FCM analysis to demonstrate the kinetics of tumor burden (percentage of hCD45 + cells) in PB from AM7577 mice 31 days after inoculation (vehicle [days 0–23, QD × 24] or DS-1594a·HCl [25, 50, 100, and 200 mg/kg; days 0–34, QD × 35], n = 6 mice per group). B Kaplan‒Meier survival curves of AM7577 mice (vehicle [days 0–23, QD × 24] or DS-1594a·HCl [25, 50, 100, and 200 mg/kg; days 0–34, QD × 35], n = 6 mice per group). C FCM analysis to assess tumor burden (percentage of hCD45 + cells) 1 day after 28 days of treatment (day 38) in the BM of NCCHD007 mice (vehicle or 12.5, 25, and 50 mg/kg DS-1594a·succinate, n = 3 mice per group). D Kaplan‒Meier survival curves of NCCHD007 mice (vehicle or 12.5, 25, and 50 mg/kg DS-1594a·succinate; days 10–37, BID × 28, n = 6 mice per group). The significance of the survival curves was analyzed by the Kaplan‒Meier log-rank test, with survival time as the indicator of the life-prolonging effect of the test compound. BID, twice daily; mCD, mouse CD; NI, non-inoculated; PB, peripheral blood; QD, once daily

DS-1594a·succinate’s in vivo antitumor efficacy in an MLL1-r ALL-PDX model (NCCHD007) was evaluated after the treatment for 28 days. On day 38, the hCD45 positivity rate in hindlimb bone marrow cells in NCCHD007 mice was ~ 87% in the vehicle-treated group vs. 4%, 0.19%, and 0.21% in the DS-1594a·succinate 12.5, 25, and 50 mg/kg groups, respectively; the latter two rates were close to the background level in the NI group (~ 0.11%; Fig. 6C). The vehicle-treated group showed ~ 18% hCD45 positivity in peripheral blood cells, while the DS-1594a·succinate groups (12.5, 25, and 50 mg/kg) showed ~ 0.00%–0.02% hCD45 positivity, which was comparable to the ~ 0.02% background level in the NI group (data not shown). Furthermore, RT‒qPCR analysis of bone marrow cells from day 38 showed that compared to the vehicle, DS-1594a·succinate markedly reduced the mRNA levels of the downstream MLL1-fusion target genes MEIS1 and MEF2C (Additional file 1: Figure S7).

Moreover, vehicle-treated NCCHD007 mice were dead within 47 days after transplantation. In contrast, mice treated with 12.5, 25, and 50 mg/kg of DS-1594a·succinate survived till day 47, showing a significant survival benefit of DS-1594a (ILS: 64%, 185%, > 254%, respectively; p = 0.0018). Although mice treated with 12.5 or 25 mg/kg DS-1594a·succinate died eventually (67% and 185% ILS, respectively), all mice treated with DS-1594a·succinate 50 mg/kg remained alive 160 days after inoculation (Fig. 6D) and did not exhibit weight loss. Notably, the hCD45 positivity rate in the bone marrow at termination (day 161) for the 50 mg/kg twice-daily DS-1594a·succinate dosage group (n = 6) was ≤ 0.4%, which was similar to that in the NI group (data not shown). This finding suggested that MLL1-r ALL cells may have been completely eradicated from the bone marrow. Consistently, the mean spleen weight per gram of body weight was 5.71 mg in the vehicle group vs. 0.80–1.00 mg in the DS-1594a·succinate group and 1.30 mg in the NI group (data not shown).

In the pharmacokinetics (PK) evaluation, the time to reach maximum plasma concentration (Tmax) values of DS-1594a in plasma remained constant from 1.0 h to 3.0 h at the dose range from 6.25 mg/kg to 100 mg/kg. The maximum plasma concentration (Cmax) and area under the plasma concentration–time curve up to 24 h (AUC24h) values of DS-1594a increased with the dose ranging from 6.25 mg/kg to 100 mg/kg (Additional file 1: Figure S8). To investigate the hematological toxicity of DS-1594a·succinate, the white blood cell count (WBC), neutrophil count (Neutro.), red blood cell count (RBC), hemoglobin concentration (HGB), and platelet count (PLT) were measured in Crl:CD1 (ICR) mice orally administered DS-1594a·succinate (30, 100, and 300 mg/kg) for 28 days. RBC, HGB, and PLT were slightly lower in the DS-1594a·succinate-treated groups, but these effects were completely reversed after cessation of dosing even in the 300 mg/kg dose group (Additional file 1: Figure S9). No toxicologically significant findings were noted with regard to clinical signs, body weight, or food consumption or during ophthalmology, urinalysis, or necropsy evaluations at any dose level during the dosing period (data not shown). In addition, the toxic effects of DS-1594a·HCl and Ara-C on the granulocyte/macrophage colony forming unit (CFU-GM) from human cord blood-derived CD34-positive cells were evaluated by MethoCult assay. The number of CFU-GM colonies exposed to 1,000, 3,000 or 10,000 nM of DS-1594a·HCl were comparable to that of DMSO control, however, no colony was detected in the 40 or 200 nM of Ara-C treatment (Additional file 1: Figure S10). The IC50 for the inhibitory effect of DS-1594a·succinate on hERG channel current was 0.3 uM (data not shown), more than 100-fold higher than the GI50 on MV-4–11 (Fig. 1E).