Silylation of Deoxynucleotide Analog Yields an Orally Available Drug with Antileukemia Effects

Myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) are characterized by impaired differentiation of hematopoietic stem cells (HSCs) or hematopoietic progenitor cells (1, 2). Somatic mutations and aberrant DNA hypermethylation that silence cancer-regulating genes (e.g., tumor-suppressor genes) in these cells (3) are involved in pathogenesis and progression of MDS and AML (4–8). Epigenetic alterations lead to reduced hypermethylation, which can in turn cause reexpression of silenced tumor-regulating genes. These alterations represent novel therapeutic targets for both of these cancers (9).

Although the clinically available DNA methyltransferase (DNMT) inhibitors azacitidine (AZA) and decitabine (DAC) have improved the prognosis of MDS and AML (10, 11), they are easily degraded by cytidine deaminase (CDA), limiting their bioavailability after oral administration; consequently, they must be administered intravenously or subcutaneously. The efficacy and safety of oral AZA or DAC, as well as the combination of these drugs with CDA inhibitors, has been evaluated in clinical trials (12–14), resulting in the approval of CC-486 (oral AZA) and ASTX727 (DAC plus cedazuridine, a CDA inhibitor; refs. 13, 14).

Resistance to CDA is crucial for achieving elevated plasma concentrations of these drugs, which would in turn increase efficacy (15). Silylation of glycoproteins is known to influence their stabilities and therapeutic effects (16, 17); however, the effects of silylation of deoxynucleotide analogs and enzymatic cleavage of silylation have not been fully elucidated. In previous work, we developed 5-O-trialkylsilylated DACs that were resistant to CDA and exerted anticancer effects against solid tumors and adult T-cell leukemia/lymphoma (18, 19). In this study, we assessed the oral bioavailability and efficacy of 5-O-trialkylsilylated DAC for treatment of MDS and AML both in vitro and in vivo.

5-O-trialkylsilylated DAC (OR2100, 2003, 2007, 2008, 2009, 2010, 2102, 2103, 2104, and 2201) was obtained from Ohara Pharmaceutical. AZA and DAC were purchased from Sigma-Aldrich. All reagents were dissolved in DMSO and stored at −20°C.

All procedures were performed at NISSEI BILIS. The study was compliant with Ohara Pharmaceutical policies on bioethics in animal work, and all procedures were compliant with the relevant national and international guidelines for the care and use of experimental animals. Procedures for pharmacokinetic studies are described in Supplementary Materials and Methods.

The MDS-L cell line was derived from the nonleukemic phase of a patient with MDS with refractory anemia-ringed sideroblasts (20). SKM1 cells (derived from the blast cells of a patient with MDS) were purchased from the Japanese Collection of Research Bioresources Cell Bank. HL60, THP-1, KG1a, and Kasumi-1 cell lines (derived from patients with AML) were purchased from the ATCC. AZA-resistant HL60 (HL60R) cells were established. Detailed cell culture methods are described in Supplementary Materials and Methods.

Primary samples were obtained from patients with MDS or AML. Experiments were approved by the Institutional Review Board of Saga University (Saga, Japan), and all procedures involving human participants were undertaken in accordance with the principles of the Declaration of Helsinki.

Western blot analyses were performed to assess the protein levels of DNMT1, UCK1, UCK2, and DCK. Procedures are described in Supplementary Materials and Methods.

Pyrosequencing assays of long interspersed nucleotide element-1 (LINE-1) methylation were performed to assess global methylation (19). Procedures are described in Supplementary Materials and Methods. The methylation status of γ-globin promoter status was assessed in samples (peripheral blood) sequentially collected on day 0, 1, 5, 8, 12, 15, 19, 22, 26, and 29 after initiation of 5-day duodenal administration of 10.9 μmol/kg OR21, as described previously (21).

Cell proliferation was evaluated by CCK-8 assay (Dojindo) and half-maximal inhibitory concentration values (IC₅₀) were determined by nonlinear regression using the CalcuSyn software (Biosoft). Procedures for cell proliferation analyses are described in Supplementary Materials and Methods.

Antibodies used for flow cytometry are listed in Supplementary Materials and Methods. Stained samples were assessed using a FACSVerse cytometer (BD Biosciences), and data were analyzed using the FlowJo software (Tree Star).

Genome-wide DNA methylation analysis was performed using Infinium Human Methylation EPIC Bead Chips (Infinium EPIC; Illumina) as described previously (18). β-values ranged from 0 (unmethylated) to 1 (fully methylated). DNA methylation data were deposited in Gene Expression Omnibus (GEO) under accession number GSE148314.

Procedures for gene expression microarray analyses are described in Supplementary Materials and Methods. Microarray data were deposited at GEO under accession number GSE141677.

All animal experiments were approved by the Institutional Review Board of Saga University (Saga, Japan) and performed according to the University's institutional guidelines. BALB/c Rag-2/JAK3 double-deficient (BRJ) mice (22) were housed in a specific pathogen-free barrier facility; BRJ mice were used for xenograft transplantation experiments. BRJ mice received twice-weekly administration of vehicle (1% DMSO), DAC, or OR21 after xenotransplantation. Procedures for animal experiments are described in Supplementary Materials and Methods.

All statistical analyses were performed using the EZR software package (Saitama Medical Center, Jichi Medical University, Saitama, Japan; ref. 23). P < 0.05 was considered statistically significant.

In this study, we developed 5-O-trialkylsilylated DACs (Supplementary Fig. S1A). Oral absorbability is assessed according to the log P value, defined as the ratio of a chemical's concentrations in the octanol versus aqueous phase of a two-phase octanol/water system (24). The log P values of 5-O-trialkylsilylated DACs were +1.64 to +5.20; the appropriate log P value for oral bioavailability is +0 to +3. Determination of LINE-1 methylation levels by pyrosequencing, which measures methylation (the ratio of T to C) at each CpG position within a sequence after bisulfite treatment, yields a representative picture of global genomic DNA methylation (25, 26). Treatment with 100 nmol/L OR2100 (OR21), 2003, 2008, 2009, 2102, or 2201 significantly decreased LINE-1 methylation levels in MDS-L (Supplementary Fig. S1B). On the basis of these drug screening experiments, we defined OR21 as a promising 5-O-trialkylsilylated DAC, as in previous reports (18, 19).

OR21, a DAC prodrug, is an oligonucleotide consisting of DAC nucleoside 5′-O-trisilylate, which is protected from degradation by CDA (Fig. 1A). The log P value of OR21 is +2.14, indicating likely oral availability. The stability of the base moiety of OR21 is almost the same (half-life, 20 hours) as that of DAC in an aqueous environment (pH 6.0–7.5), and the in vitro stabilities of AZA and DAC are also similar (27).

Initially, we investigated the pharmacokinetics of OR21 in six cynomolgus monkeys. Sequential plasma samples were taken from monkeys given the prodrug, and plasma drug concentrations were measured over time. OR21 is easily degraded by gastric acid because of its chemical structure, but direct duodenal administration of OR21 to cynomolgus monkeys resulted in absorption of the drug and subsequent release of DAC. After administration of 6.6 μmol/kg OR21, the area under the plasma drug concentration–time curve (AUC) was 0.30 (±0.18) μmol/L/hour, and after administration of 6.6 μmol/kg DAC, the AUC was 0.01 (±0.01) μmol/L/hour (Fig. 2A and B). A higher dose of DAC did not result in enteric bioavailability (Supplementary Fig. S2). Thus, OR21 exhibited higher enteric absorbability than DAC. We next evaluated the pharmacodynamics of OR21. After intravenous administration of 10.9 μmol/kg DAC, the AUC was 0.49 (±0.17) μmol/L/hour, and after duodenal administration of 10.9 μmol/kg OR21, the AUC was 0.25 (±0.14) μmol/L/hour (n = 4 monkeys; Fig. 2C). After 5-day intravenous administration of 10.9 μmol/kg DAC and duodenal administration of 10.9 μmol/kg OR21, γ-globin promoter methylation (average three CpG sites at −53, +6, +17) levels (Cont, 88.3 ± 1.35%) was significantly reduced on day 8 (DAC, 66.5 ± 2.52%, P = 0.02; OR21, 79.5 ± 4.50%, P = 0.02; Fig. 2D). In this setting, the demethylating effect of DAC (intravenous) was stronger than that of OR21 (duodenal administration), possibly due to differences in AUC: 9.0 mg/kg (26.3 μmol/kg) OR21 (duodenal administration) achieved an AUC of 0.45 ± 0.23 μmol/L/hour, comparable with the AUC of 1.67 mg/kg DAC (intravenous; Supplementary Fig. S2). These results indicated that OR21 has both enteral absorbability and hypomethylating effects. We next evaluated the in vitro and vivo efficacy of OR21 for AML and MDS. (At that time, the appropriately sized enteric-coated OR21 for administration in vivo was not available, as the enteric-coated OR21 developed for monkeys was too large to administer to mice.)

Hypomethylating agents (AZA and DAC) deplete DNMT1. Consistent with this, DNMT1 levels were markedly lower in two MDS (MDS-L, SKM1) and two AML (HL60, THP-1) cell lines treated with OR21 (100 or 500 nmol/L) than in untreated cells (Fig. 1B; Supplementary Fig. S3A–S3C). LINE-1 methylation levels were higher in both MDS-L and SKM1 cells (88.9% and 82.7%, respectively). Treatment with 100 nmol/L OR21 significantly decreased LINE-1 methylation levels in MDS-L [74.2% (P < 0.01)] and SKM1 [66.9% (P = 0.01)] cells (Fig. 1C); the effect was equivalent to that of 100 nmol/L DAC (75.2% reduction in MDS-L cells and 66.9% in SKM1 cells). In contrast, 100 nmol/L AZA did not decrease the LINE-1 methylation level in MDS-L; consistent with a previous report, 100 nmol/L of AZA had weak demethylating effects relative to DAC or OR21 (19). LINE-1 regions were also hypermethylated in AML cell lines (KG1a, Kasumi-1, HL60, and THP-1); in this case, 500 nmol/L OR21 tended to induce global hypomethylation (Fig. 1D). These results indicate that OR21 induces global genomic DNA hypomethylation in MDS and AML cell lines by depleting DNMT1, and that the effect is equivalent to that of DAC.

OR21 inhibited growth of both MDS and AML cells in a dose-dependent manner (Supplementary Fig. S4A); the effect was equivalent to that of DAC. The dose–response curves and IC₅₀ values for all three agents after 72 hours of treatment are shown in Supplementary Table S1. OR21 induced apoptosis of MDS and AML cells, again in a dose-dependent manner (Supplementary Fig. S4B). In addition, OR21 induced apoptosis of primary MDS and AML cells (Supplementary Fig. S4C). These results indicate that OR21 exerts antitumor effects against MDS and AML cells.

Vehicle-treated HL60 and MDS-L cells shared a high proportion of hypermethylated and hypomethylated sites, and OR21-treated cells had reduced hypermethylation β-values, indicating that OR21 induced DNA demethylation (Fig. 3A and B). In the three cell lines (HL60, SKM1, and MDS-L), a heatmap comparing log₂ fold changes in gene expression following OR21 treatment revealed several enriched genes, which played roles in immune system processes (e.g., PRAME and CT45A), tumor suppression (e.g., MT1E, F, TGFβ, and ALOX5AP), and cell differentiation (e.g., EVI2A, B, and TYROBO; Fig. 3C; Supplementary Fig. S5A–S5C). These genes were demethylated to a greater extent than others (Fig. 3D–F; Supplementary Fig. S6A and S6B), indicating that their expression levels were regulated by promoter DNA methylation. The demethylating effect of 500 nmol/L OR21 was similar to that of 500 nmol/L AZA (Supplementary Fig. S7A–S7C). OR21 is a prodrug of DAC; thus, the demethylating effect and gene expression profiles of OR21-treated cells were very similar to those of DAC-treated cells, as described previously (18). AZA was used as a control. MDS and AML are characterized by impaired cell differentiation of HSCs and hematopoietic progenitor cells (1); therefore, induction of cell differentiation is a potential therapeutic strategy for both forms of cancer.

CD11b is defined as a myeloid cell differentiation marker, whereas immature cells expressed HLA-DR. Hence, we used these surface markers to assess the potential of OR21 to induce cell differentiation in MDS or AML cells. OR21 upregulated expression of CD11b and downregulated expression of HLA-DR in MDS-L cells (Supplementary Fig. S8A). SKM1 and HL60 cells also exhibited upregulation of CD11b after OR21 treatment (Fig. 4A; Supplementary Fig. S8B). These cells also underwent morphologic changes, including a reduced nuclear–cytoplasmic ratio, larger cell size, and higher numbers of multilobed nuclei (Fig. 4B; Supplementary Fig. S8C), indicating that OR21 induced cell differentiation, particularly granulocytic differentiation. Hence, we used real-time quantitative reverse transcription PCR to identify the transcription factors responsible for cell differentiation. OR21 upregulated expression of the CEBPE and GATA-1 mRNAs in MDS-L and SKM1 cells (Fig. 4C; Supplementary Fig. S8D). These results indicate that OR21 induces cell differentiation, possibly by upregulating expression of the late cell differentiation drivers CEBPE and GATA-1.

Next, we used a mouse xenograft model to evaluate the anti-MDS efficacy of OR21. For this purpose, we intravenously injected SKM1 cells into BALB/c Rag-2/JAK3 double-deficient (BRJ) mice. Starting on day 7 posttransplantation, the mice were treated twice weekly with vehicle OR21 (Fig. 5A). On day 28, the number of human CD45-positive cells in peripheral blood (PB) was lower in OR21-treated mice than in vehicle-treated mice [vehicle, 6.40%; OR21, 0.31% (P = 0.0597); Fig. 5B]. In addition, OR21 prolonged survival [median overall survival; vehicle, 29 days; OR21, not reached at day 53 (P = 0.0246); Fig. 5C]. All OR21-treated mice survived with engraftment of SKM1 cells at day 53 (Supplementary Fig. S9).

Next, we used a mouse xenograft model to evaluate the anti-AML efficacy of OR21. In these experiments, BRJ mice were injected intravenously with HL60 cells and treated twice weekly with vehicle, DAC (1.0 mg/kg), or OR21(2.7 mg/kg) starting on day 5 posttransplantation (Fig. 5D). On day 37, the number of human CD45-positive cells in PB was significantly lower in OR21-treated mice than in vehicle-treated mice [vehicle, 2.96%; DAC, 1.67% (P = 0.48); OR21, 0.75% (P = 0.049)], and the OR21-treated mice exhibited prolonged survival [median overall survival; vehicle, 44 days; DAC, 46.5 days (P = 0.164); OR21, 49 days (P = 0.005); Fig. 5E and F]. In addition, OR21 induced significant hypomethylation of LINE-1 regions in bone marrow cells [vehicle, 83.7%; DAC, 56.3% (P < 0.01%); OR21, 63.0% (P < 0.01%)]. OR21-treated mice did not develop anemia, whereas DAC-treated mice were anemic at the time of sacrifice [hemoglobin levels: vehicle, 17.5 (±1.20) g/dL; DAC, 15.6 (±1.22) g/dL; OR21, 17.1(±0.31) g/dL; Fig. 5G and H]. OR21 (5.4 mg/kg) prolonged survival, whereas a high dose of DAC (2.0 mg/kg) did not [median overall survival; vehicle, 48 days; DAC, 32 days (P = 0.040); OR21, 57 days (P < 0.001); Fig. 5I]. Notably, DAC-treated mice died earlier than those in the other groups, possibly due to severe toxicity relative to the vehicle. These results suggest that the high dose of OR21 was safer and more efficacious than a high dose of DAC. Significant cell differentiation was observed in mice exposed to late treatment (from day +28) with OR21 (vehicle, 43.3%; DAC, 75.7%, P = 0.229; OR21, 65.5%; P = 0.038; Fig. 5J). These results confirmed the antitumor effects of OR21 in the MDS and AML mouse models.

Patients with MDS or AML who fail to respond to AZA treatment have limited therapeutic options, and their prognosis is extremely poor. Hence, we investigated the efficacy of OR21 against AZA-resistant AML. AZA-resistant HL60 (HL60R) cells survived in the presence of 10 μmol/L AZA. However, OR21 depleted DNMT1 protein, inhibited growth, and induced apoptosis and differentiation in these cells to a similar extent as DAC and more effectively than AZA in vitro (Supplementary Figs. S10A–S10C and S11). OR21 also significantly prolonged survival in the HL60R xenograft model (Supplementary Fig. S10D, P = 0.0188); in contrast, DAC-treated mice exhibited severe toxicity (Supplementary Fig. S12A and S12B).

To investigate the mechanism underlying AZA resistance, we compared the protein levels of nucleoside-metabolizing enzymes between HL60 and HL60R. HL60R had lower levels of uridine-cytidine kinase 2 (UCK2; Supplementary Fig. S13A). However, UCK2 mRNA levels and promoter methylation in proximal regions did not differ significantly between the two cell lines (Supplementary Fig. S13B and S13C). These observations suggested that depletion of UCK2 was associated with AZA resistance, but that the underlying mechanisms might not involve transcriptional regulation via altered methylation. The majority of downregulated genes in HL60R were involved in immune system processes or mitochondrial pathways (Supplementary Fig. S13D; Supplementary Table S3). HL60 methylome analysis revealed that HL60R cells were much more demethylated than HL60 (Supplementary Fig. S13E); meanwhile, the downregulated genes were less demethylated than the upregulated genes (Supplementary Fig. S13F and S13G). Pathway analysis of genes upregulated >2.0-fold revealed that genes associated with cell adhesion were upregulated in HL60R cells (Supplementary Fig. S14A and S14B). Among them, the chemokine receptor CXCR4 (Supplementary Fig. S14C and S14D) was upregulated at the mRNA and protein levels in HL60R. HL60R xenotransplant mice developed extramedullary disease (Supplementary Fig. S14E), suggesting that increases in cell adhesion activity are associated with elevated malignant potential. The changes in HL60R may be the consequence of transcriptional regulation via altered methylation.

We monitored the long-term administration of OR21 in BALB/c mice (not immunodeficient) treated with a high (2.5 mg/kg of DAC or 7.5 mg/kg of OR21) or low (1.25 mg/kg of DAC or 3.7 mg/kg of OR21) dose of the indicated drug. Although the plasma AUCs for high-dose DAC and OR21 were approximately equal, high-dose DAC resulted in greater hematopoietic toxicity, resulting in severe anemia and early death (Fig. 6A and B). In contrast, mice treated with a high dose of OR21 developed anemia by day 29, but recovered by day 57 while still on the drug (Fig. 6B). In addition, these mice did not die early: all were still alive after 80 days, whereas all mice receiving high-dose DAC died within 80 days. Neither DAC- nor OR21-treated mice exhibited signs of lethal toxicity in response to low-dose DAC or OR21; however, DAC-treated mice were more likely than OR21-treated mice to develop anemia (Fig. 6C). Together, these results indicate that OR21 may be less toxic than DAC, resulting in improved tolerability even at high doses.

Finally, we developed enteric-coated OR21 with the goal of making the drug resistant to gastric acid. We investigated the pharmacokinetic of enteric-coated OR21 in five ICR mice. Sequential plasma samples were taken from mice given the prodrug, and plasma drug concentrations were measured over time. Oral administration of enteric-coated OR21 to mice resulted in absorption of the drug and subsequent release of DAC. After administration of 5-mg enteric-coated OR21, the AUC of DAC (released from OR21) was 1.06 (±0.28) μmol/L/hour (Supplementary Fig. S15). Thus, enteric-coated OR21 exhibited high oral absorbability in a mouse model.

The findings of this study reveal that OR21 is a promising hypomethylating agent that can be administered orally and exerts antitumor effects against MDS and AML both in vitro and in vivo. In addition, OR21 is less toxic than DAC. DAC and OR21 had similar effects in vitro in terms of hypomethylation, cell growth inhibition, and apoptosis induction; these effects were distinct from those of AZA. These differences are consistent with previous results showing that AZA is incorporated predominantly into RNA, whereas DAC is incorporated predominantly into DNA (27). Hence, the in vitro activity of OR21, which was triggered by release of DAC from OR21 after removal of the 5′-O-trisilylate substituent, might be due to incorporation of DNA rather than RNA.

Resistance to CDA is crucial for achieving elevated plasma concentrations of DNMT inhibitors. Guadecitabine, which is resistant to CDA and exerts stronger hypomethylating activity than AZA (28), has a favorable survival outcome in patients with MDS and AML who are not responsive to AZA (15). Nevertheless, guadecitabine is not orally absorbable (29). In contrast, the recently approved agents CC-486 (13) and ASTX727 (14) can be orally absorbed. Silylation of DAC confers resistance to CDA (18); accordingly, OR21 was enterically bioavailable in a cynomolgus monkey model, and enteric-coated OR21 was orally bioavailable in a mouse mice model. In addition, silylation of DAC and enzymatic cleavage of silylation did not result in new toxicities in vivo. It is worth noting that ASTEX727 consists of two compounds, DAC and cedazuridine, a CDA inhibitor, whereas OR21 is a single compound (DAC prodrug). Consequently, OR21 has enormous advantages in terms of production, that is, it is easier and cheaper to manufacture. We have now prepared gelatin capsules filled with granules of enteric-coated OR21, and we are currently performing pharmacokinetic studies of this formulation in cynomolgus monkeys to determine the dose for a first-in-human clinical trial. Because formulation is a very important issue related to future development, we prefer not to present detailed pharmacokinetic data until the first in-human clinical trial is completed. The preliminary results suggest that OR21 could be used as an oral hypomethylating agent.

OR21-induced global DNA hypomethylation in MDS and AML cells could upregulate expression of silenced genes associated with tumor suppression, cell differentiation, and immune system processes. Among the genes that were upregulated after OR21 treatment, TGFβ is an important tumor-suppressor gene in hematologic malignancies that affects immune system process and cell differentiation (30); accordingly, TGFβ may play an important role in the antitumor effects of OR21. Although it is unclear how immune system process pathways are involved in the antitumor effects of OR21, these pathways were downregulated in AZA-resistant HL60 cells (31), suggesting that immune pathways play important roles in the antitumor effects of hypomethylating agents used to treat AML. Moreover, upregulation of genes associated with immune system processes and tumor suppression could alter immunogenicity or tumor recognition, thereby sensitizing various tumors (32, 33). Immune checkpoint inhibitors used to treat melanoma (34), lung cancer (35), or other solid tumors (36, 37) also exhibit antitumor effects in patients with MDS or AML (38, 39). To optimize promising combination therapeutic strategies involving OR21, subsequent administration of immune checkpoint inhibitors may improve efficacy against MDS or AML.

AZA-resistant HL60R cells exhibited downregulation of the UCK2 protein without alteration of UCK2 promoter methylation or mRNA levels (40, 41). UCK2 is a pivotal uridine-metabolizing enzyme (42). Furthermore, alteration of promoter methylation resulted in upregulation of genes related to cell adhesion in AZA-resistant HL60R cells, and OR21 overcame AZA resistance. Patients with MDS or AML who fail to respond to AZA treatment have limited therapeutic options, and their prognosis is extremely poor (43, 44). We showed that OR21 was effective agaist AZA-resistant AML in vitro and in vivo, whereas DAC exerted a high degree of toxicity in vivo. These results suggest that OR21 may yield clinically favorable outcomes in patients with MDS or AML who do not respond to AZA.

Patients with MDS or AML exhibit altered HSC differentiation (1, 2), and patients who have AML with low CEBPE gene expression level had significantly worse survival outcomes (Supplementary Fig. S16); therefore, induction of appropriate cell differentiation is a potential therapeutic strategy for MDS and AML (45, 46). In three different cell lines, OR21 upregulated genes associated with cell differentiation, resulting in induction of myeloid cell differentiation. Induction of global DNA hypomethylation triggered upregulation of the silenced key differentiation drivers CEBPE and GATA-1, which are inactivated in MDS and AML (47, 48), thereby triggering cell differentiation (49). OR21 induced antitumor effects in vivo by promoting both hypomethylation and cell differentiation. The effects were similar to those of DAC, although OR21 did not induce severe toxicities in mice. Long-term administration experiments confirmed that OR21 was less toxic than DAC at both high and low doses. Although it is not clear why the toxicity of OR21 is lower than that of DAC, one possible explanation is that OR21 achieves a lower peak plasma concentration: the peak plasma concentration of DAC in mice was 0.43 μmol/L after intraperitoneal administration of 1.59 μmol/L/kg OR21, versus 0.64 μmol/L after administration of 0.88 μmol/L/kg DAC (calculated as an AUC equivalent to that of released DAC, which had an AUC of 0.38 μmol/L/hour; Supplementary Table S4; ref. 18).

Although the safety of cleaved sialylated DAC has not been fully elucidated, we confirmed that OR21 is less toxic, possibly due to differences in peak plasma concentration and AUC profiles relative to DAC administration. Meanwhile, the efficacy of OR21 is mainly due to the release of DAC from OR21 after removal of the 5′-O-trisilylate substituent. Therefore, OR21 represents an orally bioavailable drug with antitumor effects against MDS and AML similar to those of DAC, and is therefore a potentially suitable treatment for patients with MDS and AML (10, 50).

In conclusion, silylation of deoxynucleotide analog can provide oral bioavailability without new toxicities. OR21 exhibits high oral availability and exerts strong antitumor effects against MDS and AML both in vitro and in vivo; moreover, it is less toxic than DAC. We anticipate that the efficacy and safety of OR21 for treatment of MDS and AML will be verified in planned early-phase clinical trials.

H. Ureshino reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; in addition, H. Ureshino has a patent for OR2100 issued. Y. Kurahashi reports personal fees from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; in addition, Y. Kurahashi has a patent with Ohara Pharmaceutical Co., Ltd. issued. T. Watanabe reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; in addition, T. Watanabe has a patent for Ohara Pharmaceutical Co., Ltd. issued. K. Kamachi reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; grants from Ohara Pharmaceutical Co., Ltd. outside the submitted work; in addition, K. Kamachi has a patent for OR-2100 issued. Y. Yamamoto reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; grants from Ohara Pharmaceutical Co., Ltd. outside the submitted work; in addition, Y. Yamamoto has a patent for OR-2100 issued. Y. Fukuda-Kurahashi reports personal fees from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; in addition, Y. Fukuda-Kurahashi has a patent for Ohara Pharmaceutical Co., Ltd. issued. N. Yoshida-Sakai reports a patent for OR2100 issued. N. Hattori reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; in addition, N. Hattori has a patent for OR-21 structure issued. T. Ushijima reports grants from AMED and Ohara Pharmaceutical Co., Ltd. during the conduct of the study; grants from AMED, grants from JSPS, and nonfinancial support from Sysmex outside the submitted work; in addition, T. Ushijima has a patent for OR21 structure issued. S. Kimura reports grants from Ohara Pharmaceutical Co., Ltd. during the conduct of the study; personal fees from Bristol Myers Squibb, Otsuka Pharmaceutical, Pfizer, Novartis; grants from Chugai Pharmaceutical, Kyowa Kirin, Nippon Shinyaku, Taiho Pharmaceutical, Ono Pharmaceutical, Takeda Pharmaceutical, CSL Behring, Otsuka, and Daiichi Sankyo outside the submitted work; in addition, S. Kimura has a patent. No disclosures were reported by the other authors.

H. Ureshino: Resources, data curation, formal analysis, validation, investigation, visualization, writing–original draft, writing–review and editing. Y. Kurahashi: Data curation, formal analysis, visualization. T. Watanabe: Data curation, supervision, visualization, writing–review and editing. S. Yamashita: Data curation, formal analysis. K. Kamachi: Data curation, writing–review and editing. Y. Yamamoto: Data curation. Y. Fukuda-Kurahashi: Data curation. N. Yoshida-Sakai: Data curation. N. Hattori: Data curation, supervision. Y. Hayashi: Supervision. A. Kawaguchi: Formal analysis. K. Tohyama: Resources, supervision. S. Okada: Resources, supervision. H. Harada: Resources, supervision. T. Ushijima: Supervision. S. Kimura: Conceptualization, supervision, funding acquisition, writing–review and editing.

This work was supported by Ohara Pharmaceutical Co., Ltd.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.