A pharmacokinetic and safety study of oral arsenic trioxide in patients with acute promyelocytic leukemia

Acute promyelocytic leukemia (APL) has evolved from a highly fatal disease to the most curable form of acute myeloid leukemia using a chemotherapy-free regimen of arsenic trioxide (ATO) IV and oral all trans retinoic acid (ATRA) with cure rates >90%.^1-5 However, there remains a significant unmet need for these patients due to the treatment burden associated with receiving daily ATO infusions and the risk of complications associated with indwelling central catheters. In addition, patient access to ATO may be limited by the lack of an oral formulation. Patients receive up to a total of 140 infusions of ATO, delivered for nearly a year of treatment, with up to 60 daily doses during induction and then 20 doses per month in alternating months during 4 cycles of consolidation therapy.⁶ An oral ATO formulation that can achieve comparable pharmacokinetic (PK) exposure to ATO IV that is safe and tolerable offers the opportunity to improve quality of life, increase access to treatment, and reduce treatment burden without compromising the high efficacy established by ATO IV.

A 100% hematologic complete remission (CR) rate and a 2-year event-free survival rate of 97% were reported in the pivotal study supporting the approval of ATO IV in combination with ATRA in adults with newly diagnosed low-risk APL characterized by the presence of the t(15;17) translocation or promyelocytic leukemia–retinoic acid receptor alpha (PML/RARα) gene expression.⁵ In addition, ATO IV is indicated for induction of remission and consolidation in patients with relapsed and refractory APL.^6,7 Two primary mechanisms through which ATO may exert potent inhibitory activity in APL include degradation of the oncogenic PML RARα fusion protein, thereby restoring RARα-dependent transcription and tumor cell differentiation and promoting restoration of PML nuclear bodies, which leads to p53 activation and loss of the self-renewing capacity of leukemia-initiating cells.^8,9 

SY-2101 (previously ORH-2014) includes the same active pharmaceutical ingredient as ATO IV (As₂O₃) but is formulated for oral administration as a novel powder-filled capsule with an improved dissolution profile compared with unprocessed ATO. A previous phase 1 study evaluated the PK and safety of SY-2101 in 12 participants with advanced hematologic disorders.¹⁰ Participants received daily dosing of SY-2101 with a median duration of treatment of 7.5 weeks. SY-2101 was well tolerated with no dose-limiting toxicities. Administration of SY-2101 at 15 mg daily had comparable PK with the approved ATO IV dose of 0.15 mg/kg based on a comparison to historical ATO IV PK data; therefore, it was selected for use in this study. A prospective, randomized, intrapatient crossover PK study of SY-2101 vs ATO IV in patients with APL was recommended for future study.

Availability of an oral ATO formulation has been long awaited. No oral formulations of arsenic have been approved in the United States, although clinical studies of other oral arsenic formulations in other countries, including some with long-term follow-up, have reported similar efficacy and safety to ATO IV.^11-14 The oral arsenic Realgar-Indigo naturalis formula, containing tetraarsenic tetrasulfide (As₄S₄), has been commercially available in China since 2009.¹⁵ In Hong Kong, a liquid solution of ATO has been reported to have bioavailability comparable with ATO IV and has been used in several phase 2 studies in both relapsed and newly diagnosed APL.^11,12,16-20 

Here, we report the results from a phase 1 study in adult patients with APL designed to evaluate the PK, safety, and tolerability of SY-2101 compared with ATO IV. This study was intended to generate comprehensive PK data for ATO and its metabolites for both SY-2101 and ATO IV. Given the limited PK data available for ATO IV, with data on only 6 patients included in the prescribing label,⁶ these data are paramount to select a dose of SY-2101 comparable to the approved dose of ATO IV for further development in a treatment study in patients with APL. Selecting a dose that provides comparable exposure to ATO IV is critical to ensure that SY-2101 provides similar high efficacy outcomes to those established by ATO IV.

This phase 1, open-label, randomized, intrapatient crossover study enrolled participants across 3 sites in the United States between September 2021 and September 2023. Adult patients with APL in a morphological CR after receiving ATO IV with ATRA induction therapy and who were planning to receive consolidation therapy or had completed therapy in the past 6 months with ATO IV and ATRA were eligible. Patients must have been able to tolerate full-dose ATO IV (0.15 mg/kg) and could not have experienced grade ≥3 nonhematologic toxicities associated with ATO administration.

This study included 2 distinct modules designed to characterize the PK, safety, and tolerability of SY-2101 and ATO IV. Participants were able to participate in 1 or both modules, as each module was optional. The single-dose PK module compared the PK of SY-2101 with ATO IV and the impact of food on SY-2101. Participants received a single dose of SY-2101 administered in the fasted state, a single dose of SY-2101 administered in the fed state, and a single-dose of ATO IV. Standard-of-care (SOC) consolidation treatment consists of 4 consolidation cycles, 8 weeks each, including initial 4 weeks of ATO per each cycle. Participation in this module took place during weeks 6 to 8 of any SOC consolidation cycle (≥7 days after the participant’s last dose of ATO in that consolidation cycle) or within 6 months of completion of consolidation therapy. Participants were treated with a single dose of ATO IV at 0.15 mg/kg for 2 hours on day 1 and then randomized to a single dose of SY-2101 at 15 mg in fed or fasted states on day 8 or 15 (Figure 1). During the food effect portion of the module, participants were randomized to receive SY-2101 within 30 minutes after consumption of a high-fat and high-calorie meal (fed state) or after an overnight fast for at least 8 hours (fasted state). A sample size of ∼10 participants was planned based on precedent set by other PK feasibility studies of similar nature and was not based on power calculations.

The multiple-dose oral module evaluated steady-state PK and safety of longitudinal dosing of SY-2101 for a month of treatment. Participants who were in a molecular CR before their fourth cycle of consolidation therapy were eligible and received a 15 mg SY-2101 dose in place of ATO IV (0.15 mg/kg) 5 days per week for 4 weeks and ATRA 45 mg/m² per day divided into 2 equal doses for the first 2 weeks of the fourth consolidation cycle. Participants remained on the study with safety assessments to 30 days after the last dose of ATO. This module provided supplemental safety and steady-state PK assessments of SY-2101, and enrollment was not based on a statistically determined sample size.

The primary objective of the study was to characterize the PK of SY-2101. As(III) was selected for the primary PK evaluation as it is the active form of arsenic. Total arsenic and its metabolites arsenic acid [As(V)], dimethylarsinic acid (DMA), and methylarsonic acid (MMA) were also analyzed to compare the overall metabolism between SY-2101 and ATO IV. The PK of oral SY-2101 was compared with ATO IV, and the impact of food on exposure to SY-2101 was evaluated. For comparisons of the oral to IV formulations of ATO, a dose of 15 mg SY-2101 was compared with the weight-based dose of 0.15 mg/kg of ATO IV. Secondary objectives were to characterize the safety and tolerability of SY-2101 using Common Terminology Criteria for Adverse Events version 5.0.

In the single-dose PK module, ATO IV was administered for 2 hours with serial blood samples obtained before administration, at 0.5 and 1 hour after start of infusion, immediately after the end of the infusion, and at 3, 4, 6, 8, 24, 48, 72, 96, and 168 hours after the start of the infusion. After oral administration of SY-2101, serial blood samples were obtained for PK analysis before administration and at 0.5, 1, 2, 3, 4, 6, 8, 24, 48, 72, 96, and 168 hours after administration. In the multiple-dose oral module, serial blood samples were obtained at baseline and at 2, 4, and 6 hours after SY-2101 administration on days 5 and 26 of the consolidation cycle.

After ATO IV, PK parameters of maximum concentration (C_max), time of maximum concentration (T_max), last concentration (C_last), time of last concentration (T_last), area under the plasma concentration (AUC) from 0 hour to last time point (AUC_0-last) and from 0 hour to infinity (AUC_0-inf), clearance (CL), volume distribution (Vd), and elimination half-life (t_1/2) for total arsenic and its metabolites were estimated using noncompartmental methods. After SY-2101, the same PK parameters were estimated except for apparent CL (CL/F) and Vd (Vd/F). PK samples were analyzed for total arsenic, As(III), As(V), DMA, and MMA through liquid chromatography with inductively coupled plasma mass spectrometry assays.

The safety population includes all participants who received any amount of ATO. Demographics, baseline characteristics, disposition, and analyses of safety were summarized using the safety population.

For the primary As(III) PK end point, statistical comparisons used the natural log (ln)–transformed data for C_max, AUC_0-last, and AUC_0-inf. To assess the bioavailability of SY-2101 in the fasted state to ATO IV, a linear mixed model repeated measures analysis was performed to estimate the geometric mean ratios (GMRs) of C_max, AUC_0-last, and AUC_0-inf. To assess the impact of food on PK of SY-2101 in the fed state to SY 2101 in the fasted state, the same linear mixed model analysis was used to estimate the GMRs of C_max, AUC_0-last, and AUC_0-inf. The analysis of variance was performed using SAS PROC MIXED on the ln-transformed C_max, AUC_0-last, and AUC_0-inf for As(III). GMRs were calculated using the exponentiation of the difference between treatment least squares means from the analyses on the ln-transformed C_max, AUC_0-last, and AUC_0-inf. These GMRs were expressed relative to the reference treatment. The 90% confidence intervals (CIs) for the GMRs were derived. Bioequivalence was supported if the 90% CI for the ratio was between 80% and 125%.²¹ 

To evaluate whether the PK disposition of ATO IV, SY-2101 fed, and SY-2101 fasted was different, statistical comparisons were performed using the Kruskal-Wallis test in SAS to determine whether the As(III) and total arsenic CL, Vd, t_1/2, C_max, T_max, C_last, T_last, AUC_0-last, and AUC_0-inf differed between the treatment groups. Dwass, Stell, Critchlow-Fligner multiple-comparison analysis, which is based on pairwise 2-sample Wilcoxon comparisons, was performed using SAS to determine where statistical difference exists.²² In addition, the sign rank test was performed using SAS to determine whether the As(III) and total arsenic ratios of C_max, AUC_0-last, and AUC_0-inf for SY-2101 fasted state to ATO IV and SY-2101 fed state to SY-2101 fasted state were different than 1. Statistical results were adjusted for multiple comparisons using the Dwass, Stell, Critchlow-Fligner multiple-comparison test.

The protocol was conducted under an Investigational New Drug application. The protocol was approved by the applicable institutional review boards, and all participants provided informed consent before any study procedures. The study was registered at ClinicalTrials.gov (identifier: NCT04996030) and conducted in accordance with the Declaration of Helsinki.

A total of 15 participants were enrolled across the 2 study modules. Thirteen participants were enrolled in the single-dose PK module comparing ATO IV with SY-2101 in fed and fasted states, of which 12 were treated with SY-2101 and 10 completed all 3 single-dose treatments. Three participants discontinued early due to withdrawal of consent (1 withdrew after day 1 ATO IV and the other 2 participants discontinued after receiving day 1 ATO IV and day 8 SY-2101). Five participants were enrolled and completed treatment in the multiple-dose oral module, of which 4 participants completed the planned PK studies.

For the 12 participants randomized and treated with SY-2101 in the single-dose PK module, the median age was 42 years (range, 21-70) and the median weight was 85 kg (range, 52-143). Most participants were female (9/12 [75%]), and the median time since diagnosis of APL was 2.5 months (range, 2.0-14.6; Table 1).

For the comparison of oral SY-2101 fasted vs ATO IV, the plasma concentration vs time profiles of As(III) were similar in 168 hours (Figure 2). The PK parameters for SY-2101 fasted state and ATO IV demonstrated similar disposition based on CL, Vd, and t_1/2 (Table 2). Notably, the As(III) exposures for SY-2101 in the fasted state and ATO IV were nearly identical, with GMRs for the ratios of SY-2101 in fasted state to ATO IV for AUC_0-last and AUC_0-inf of 1.00 (90% CI, 0.86-1.17) and 1.00 (95% CI, 0.86-1.15), respectively, and thus were within the bioequivalence range of 80% to 125%.²¹ The individual and mean As(III) C_max and AUC_0-inf were similar (Figure 3), with ratios of C_max, AUC_0-last, and AUC_0-inf for SY-2101 in the fasted state to ATO IV that were not statistically different (Table 3).

These ratios of As(III) PK parameters after SY-2101 in the fasted state to ATO IV also represent the estimate of oral bioavailability as measured by the fraction of drug absorbed after oral administration. For As(III) AUC_0-last and AUC_0-inf, the mean ± standard deviation fraction of drug absorbed after oral administration for SY-2101 in the fasted state was 1.06 ± 0.36 and 1.04 ± 0.31, respectively. The variations in the data preclude the ability to make an exact determination of oral bioavailability; however, the SY-2101 As(III) fraction of drug absorbed after oral administration of ∼1 indicates that the bioavailability is high with ∼100% of the drug being absorbed orally.

There was a difference in As(III) T_max and C_max between SY-2101 in the fasted state and ATO IV (P = .04 and P = .07, respectively; Table 2). The As(III) C_max of SY-2101 in the fasted state was lower than that of ATO IV, with a GMR of 0.76 (90% CI, 0.59-0.97; Table 3). However, this difference was minor and not clinically relevant given the exposure difference was <25%. The PK values of total arsenic, As(V), DMA, and MMA PK were similar for both treatments (supplemental Tables 1-5).

When evaluating the impact of food on SY-2101 PK, the plasma concentration vs time profiles of As(III) after administration of SY-2101 in the fed and fasted states were similar in 168 hours (Figure 2). SY-2101 administered in the fed and fasted states provided similar exposures based on GMRs of SY-2101 fed to SY-2101 fasted for As(III) C_max, AUC_0-last, and AUC_0-inf of 0.85 (90% CI, 1.08-1.12), 1.08 (90% CI, 0.93-1.27), and 1.12 (90% CI, 0.97-1.30), respectively. However, the 90% CIs for AUC_0-last and AUC_0-inf were not within the bioequivalence limits of 80% to 125% (Table 3). The individual and mean C_max and AUC_0-inf after administration of SY-2101 in the fed and fasted states were similar (Figure 4). There was no effect of treatment sequence on the comparison of SY-2101 in the fed state to the fasted state. The PK parameters for total arsenic, As(V), DMA, and MMA PK were similar to that of As(III) after administration of SY-2101 in the fed and fasted states (supplemental Tables 1-5).

PK analysis after multiple-dose administration of SY-2101 in 4 PK-evaluable participants demonstrated that the disposition of SY-2101 was similar on days 5 and 26. The PK values of total arsenic, As(III), As(V), DMA, and MMA were similar after repeated dosing (supplemental Tables 6-8).

In the single-dose PK module, the planned dose of ATO IV at 0.15 mg/kg and SY-2101 at 15 mg was administered to all participants. There were 3 participants (23%) who experienced at least 1 adverse event after ATO IV administration and 4 (33%) who experienced at least 1 adverse event after SY-2101 administration. No adverse events were reported for >1 participant in this module. In addition, 1 participant (8%) had a treatment-related adverse event of vomiting and 1 participant (8%) had serious adverse events of bacteremia and sepsis after SY-2101 administration that were not considered related to the study treatment. All adverse events were of low grade except for grade 3 sepsis and hypertension (1 participant each after SY-2101 administration); neither of the events was assessed as related to the study treatment.

In the multiple-dose oral module, all 5 participants enrolled received the planned 20 doses of SY-2101 (dosed 5 days per week for 4 weeks). There were 4 participants (80%) who received a dose of 15 mg SY-2101 for all doses and 1 participant (20%) who was initially treated with 15 mg SY-2101 and had a dose reduction to 10 mg on day 15 due to grade 1 QT prolongation with QT interval corrected for heart rate using Fridericia’s formula (QTcF) of 472 milliseconds. This participant had a baseline QTcF of 463 milliseconds at screening (3 days posttreatment with ATO IV) which had decreased to 432 milliseconds before treatment with SY-2101 (31 days posttreatment with ATO IV) and was receiving concomitant treatment with fluoxetine. There were 4 participants (80%) who experienced at least 1 adverse event, all adverse events were of low grade, and no serious adverse events were reported. Adverse events reported in >1 participant in this module included headache (3 participants) and nausea (2 participants). Across all study modules, there were no deaths reported and no adverse events leading to treatment discontinuation.

SY-2101 is a novel oral formulation of ATO which has the potential to improve quality of life for patients with APL by reducing the treatment burden associated with the need for daily ATO infusions in nearly a year of treatment. Treatment with ATO IV is highly efficacious, making it critical to demonstrate that similar PK exposures can be achieved with SY-2101. In this study, SY-2101 demonstrated comparable systemic exposure to ATO IV based on As(III) AUC_0-last and AUC_0-inf, with GMRs of 1.00 and 90% CI within the bioequivalence range of 80% to 125%. These comparable exposures were achieved using a flat dose of 15 mg of SY-2101 compared with a weight-based dose of 0.15 mg/kg of ATO IV. As expected, due to the differences in oral compared with IV administration, the As(III) T_max was later (median 3.05 vs 2.02 hours) and the C_max was lower with a ratio of 0.76 with SY-2101 in the fasted state compared with ATO IV. However, differences in C_max were minor and not clinically significant given that the difference was <25%. The mean F_PO of SY-2101 in the fasted state was ∼1, indicating nearly complete oral absorption.

Comparisons of SY-2101 administered in the fed and fasted states also demonstrated similar exposure. Although the 90% CIs for the GMR for AUC_0-last and AUC_0-inf were not within the bioequivalence limits, the effect of food on the oral absorption of SY-2101 was considered minor. C_max was generally lower with oral administration of ATO as compared with IV, and therefore a safety concern is not expected. Similarly, although the CIs were wide for the AUC, the overlap in exposure between oral and IV administration may suggest similar efficacy if compared head-to-head using these same doses in patients with APL. On the basis of these results, SY-2101 can be evaluated for treatment of patients with APL when administered with or without food, potentially allowing for concurrent administration with ATRA, which is taken orally with food. The PK values of total arsenic and its related metabolites, As(V), DMA, and MMA, were evaluated to provide a complete understanding of the absorption and distribution of SY-2101. Although As(III) is the primary active component of SY-2101, evaluation of all metabolites confirmed that the metabolite profile of SY-2101 aligns closely with that of ATO IV. The similar PK exposures and disposition of SY-2101 and ATO IV further support the likelihood of clinical equivalence between the 2 formulations of ATO.

The safety profile of SY-2101 demonstrated in this study is encouraging and consistent with what has been previously reported.¹⁰ The 15 mg dose of SY-2101 was well tolerated without the need to evaluate lower dose levels due to observed toxicity. Adverse drug reactions associated with ATO IV have been well documented and include serious and potentially fatal risks, including differentiation syndrome, hepatotoxicity, neurotoxicity, and QTc prolongation.^3,6,7 Notably, no new safety signals were identified in this study and no serious cases of hepatotoxicity or QTc prolongation were reported for SY-2101. The lower C_max with SY-2101 compared with ATO IV may mitigate some of the toxicity associated with ATO IV use.^10,12 The assessment was limited to participants in consolidation who had previously tolerated ATO IV contributing to the observed favorable safety profile of SY-2101 in this study. Further evaluation during induction therapy, when patients are most critically ill and at risk for complications, including differentiation syndrome, will provide a more complete assessment of the safety across the full range of the APL patient treatment plan from induction to consolidation. The optimal dose of SY-2101 is based on matching exposure to ATO IV, which is based on weight, with 0.15 mg/kg IV well established as the SOC in APL treatment. The extent to which this dose is optimal to balance clinical efficacy and toxicities, particularly at extremes of weight, remains a future area of research.²³ 

The PK and safety data generated from this study support further development of SY-2101 in APL. Given the nearly identical PK exposures by AUC of SY-2101 and ATO IV, a future study evaluating the use of SY-2101 to treat patients with APL may allow the opportunity to demonstrate clinical benefit using bioequivalence and more expeditiously advance the development of this long-awaited oral ATO formulation. This approach has recently been used in the development of other therapies for hematologic malignancies.^24-26 Although this study was not designed or powered to formally demonstrate bioequivalence, the comparable exposures observed along with clinical outcomes reported with other formulations of ATO demonstrate that oral SY-2101 can be expected to provide equivalent efficacy to ATO IV and would significantly reduce the treatment burden for patients, allowing for an all-oral treatment regimen for APL.

The authors acknowledge and thank all participants, families, investigators, coinvestigators, and study staff who contributed to this clinical study.

This study was supported by research funding from Syros Pharmaceuticals, Inc.

Contribution: D.A.R., E.W., K.B., and M.K. designed the research; F.R., G.G., S.R., H.K., G.G.-M., and M.Y. performed the research; P.R. and B.A.Z. performed the statistical analysis; T.H. and W.C.Z. performed the pharmacokinetic analysis; K.B., J.G., M.K., T.H., and W.C.Z. wrote the manuscript; and all authors had full access to the study data, analyzed and interpreted the data, reviewed the manuscript, and approved the final version of the publication.

Conflict-of-interest disclosure: D.A.R., J.G., K.B., M.K., P.R., and T.H. are shareholders and employees at Syros Pharmaceuticals, Inc. E.W. is a former employee of Syros Pharmaceuticals, Inc. B.A.Z. and W.C.Z. are consultants at Syros Pharmaceuticals, Inc. F.R. received research funding and honoraria (membership of advisory meetings) from Syros Pharmaceuticals, Inc. G.G. received research funding from AbbVie Inc, Kinomica Inc, Menarini Richerche Inc, and Arcellx. H.K. receives honoraria from, serves on the advisory board of, and acts as a consultant for AbbVie, Amgen, Ascentage, Ipsein Biopharmaceuticals, KAHR Medical, Novartis, Pfizer, Shenzhen Target Rx, Stemline, and Takeda; and receives research grants from AbbVie, Amgen, Bristol Myers Squibb, Daiichi Sankyo, Immunogen, and Novartis. The remaining authors declare no competing financial interests.

Correspondence: Farhad Ravandi, Department of Leukemia, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Unit 428, Houston, TX 77030; email: fravandi@mdanderson.org.