ADORE: an open platform study of ruxolitinib in combination with other novel therapies in patients with myelofibrosis Open Access

Myelofibrosis (MF) is a rare blood cancer that leads to the development of extramedullary hematopoiesis and fibrosis.¹ The current standard of care for symptomatic MF involves the use of the Janus kinase (JAK)1/JAK2 inhibitor ruxolitinib.^1-3 Ruxolitinib provides symptomatic relief, reduces spleen volume, and improves the quality of life and overall survival of patients with MF.^4-7 Despite initial positive responses, higher risk patients may lose their response to ruxolitinib after 2 to 3 years of treatment.^4,7-9 In clinical practice, ruxolitinib discontinuation rates are variable, ranging from ∼40% to 70% during the initial year of therapy.^9-11 Approximately 35% of discontinuations are attributed to a lack or loss of response.⁸ These poor outcomes are multifactorial. Patients with higher MF disease burden, as indicated by higher Dynamic International Prognostic Scoring System scores, are more likely to experience suboptimal response.¹² There is an urgent need to explore alternative treatment options to improve clinical outcomes. Combination therapies have the potential to overcome the limitations of monotherapy, offering a promising avenue to attain an additive or synergistic effect that can result in a disease-modifying effect and improved clinical outcomes.^9,13 Indeed, previous studies on approaches involving ruxolitinib combination therapy have revealed promising findings.^14-19 

ADORE (NCT04097821) was a phase 1b/2 open-label, multicenter platform study that investigated the safety, pharmacokinetics (PK), and preliminary efficacy of combination treatment of ruxolitinib with 5 novel compounds in patients with MF. Using ruxolitinib treatment as the backbone, this study evaluated the combined effect of compounds that can alter disease drivers and may deliver clinical benefits. Considering the number of potential targets and emerging data from compounds in development, an open platform design was used. Part 1 of the study focused on safety and tolerability, and parts 2 and 3 were planned to assess the preliminary efficacy of ruxolitinib combinations. The open study design was chosen to assess multiple combinations in the same trial and allowed for accelerated selection of combination partners, flexibility in prioritization, and expedited phase 2 study design. Five compounds were initially selected for analysis: siremadlin, rineterkib, crizanlizumab, sabatolimab, and NIS793. Siremadlin is a small molecule inhibitor that targets human double minute-2 and restores p53-mediated apoptosis, which may promote clonal regression and spleen volume reduction (SVR).^20,21 Rineterkib is a small-molecule inhibitor of extracellular signal-regulated kinase 1 and 2,²² a pathway required for the maintenance of JAK-mutated clones.²³ Crizanlizumab is an antibody that targets P-selectin on activated endothelial cells and platelets, and sabatolimab is an antibody to T-cell immunoglobulin and mucin domain-3 (TIM-3), an immune checkpoint molecule present on myeloproliferative neoplasm progenitor cells.²⁴ NIS793 is an antibody with high-affinity binding to transforming growth factor-β1 (TGFβ1) and TGFβ2, which are components in pathways required for propagation of myeloid cancers,^25-27 and low-affinity binding to TGFβ3.

Owing to the rapidly evolving landscape in drug development for MF, the sponsor halted recruitment in June 2022. The protocol was amended to define a core treatment phase and introduced an extension phase for part 1. Here, we report the results of the primary analysis, which include data up until the cutoff date when the last patient completed the end of treatment (EOT) visit for the core treatment phase (part 1), for the siremadlin cohort, to which most patients were allocated. Available data from other cohorts are reported in the supplemental Data.

ADORE was an open-label, multicenter, 3-part phase 1b/2 open platform study (Figure 1). After a screening period of up to 28 days, eligible patients began treatment for a planned duration of 24 weeks (8 cycles for treatment arms containing NIS793, 6 cycles for all other arms). In the context of the enrollment halt, the study design was limited to part 1 dose escalation and safety run-in, and an extension phase. Patients who received the study treatment in part 1 for a minimum of 24 weeks and benefitted from the treatment were eligible to enter the extension phase after the protocol amendment was implemented.

The part 1 core treatment phase comprised the following 5 combination arms: ruxolitinib plus siremadlin, ruxolitinib plus crizanlizumab, ruxolitinib plus sabatolimab, ruxolitinib plus rineterkib, and ruxolitinib plus NIS793. Figure 1 describes the dosing schedules. Within part 1, dose escalation was performed for siremadlin and rineterkib to determine the recommended phase 2 dose (RP2D) for combination treatment with ruxolitinib. For crizanlizumab, sabatolimab, and NIS793, the proposed RP2D was known before study initiation, and a safety run-in was conducted among 6 patients to assess unexpected toxicities and confirm pharmacodynamic and PK properties in combination with ruxolitinib.

Only 1 patient entered part 2 (in the ruxolitinib monotherapy control arm) at the time of enrollment halt. This patient had already discontinued before the primary analysis cutoff date.

The inclusion criteria for this study included patients aged ≥18 years with primary MF, post–polycythemia vera (PV) MF, or post–essential thrombocythemia MF with splenomegaly, defined as a palpable spleen measuring ≥5 cm from the left costal margin or a spleen volume of ≥450 cm³ as determined by magnetic resonance imaging/computed tomography scan. A hemoglobin level of <11 g/dL (or <10 g/dL as per an earlier protocol) and a platelet count of ≥75 × 10⁹/L were required. The exclusion criteria included splenic irradiation within 6 months or platelet transfusion within 28 days before study initiation. Patients were required to have received ruxolitinib for a minimum of 12 weeks (or 24 weeks as per an earlier protocol) at a stable dose for at least 4 weeks (or 8 weeks as per an earlier protocol) before the first combination dose. Once clinical benefit was demonstrated in the core treatment phase, as determined by the investigator’s assessment, patients became eligible for the extension treatment phase. This study was designed and conducted in accordance with the guidelines for Good Clinical Practice of the International Council for Harmonisation, the principles of the Declaration of Helsinki, and local laws and regulations. The study protocol and all amendments were reviewed and approved by the independent ethics committee and/or institutional review board for each study center.

The primary end point of part 1 was the assessment of the incidence and severity of dose-limiting toxicities (DLTs) within the first 2 treatment cycles. Secondary safety end points included the incidence and severity of adverse events (AEs), and secondary efficacy end points were change in SVR and change in total symptom score (TSS). Serious AEs, on-treatment deaths, AEs leading to study discontinuation, most frequently reported AEs, and PK were also assessed. Change in spleen volume (by magnetic resonance imaging/computed tomography) from baseline was defined as the proportion of patients who achieved (1) ≥35% SVR and (2) ≥25% SVR. Spleen length reduction was assessed by palpation, requiring ≥50% reduction for baseline splenomegaly of >10 cm or becoming not palpable for baseline splenomegaly of ≥5 to ≤10 cm below the left costal margin. Symptom response was defined as the proportion of patients who achieved ≥50% reduction from baseline in the Myelofibrosis Symptom Assessment Form (MFSAF) version 4.0. PK parameters were assessed for all patients who received at least 1 dose of any of the 6 study drugs. PK parameters were derived from the individual concentration vs time profile using a noncompartmental method as implemented in Phoenix WinNonlin (Pharsight, Mountain View, CA).

JAK2V617F mutational burden and TP53 status were centrally assessed at baseline and follow-up. JAK2V617F mutational burden was measured using a droplet digital polymerase chain reaction assay with 0.5% variant allele frequency (VAF) limit of detection and TP53 status through next-generation sequencing with the Archer VARIANTPlex Core Myeloid panel at ∼1000× sequencing depth and validated to a limit of detection of 2% VAF. Levels of the pharmacodynamic marker growth differentiation factor 15 (GDF-15) were evaluated for patients on ruxolitinib plus siremadlin.

Primary analyses were performed on all 44 enrolled patients at the end of the treatment visit for the core treatment phase (part 1) and had completed at least 24 weeks of treatment or discontinued earlier. Primary analyses were performed using DLT data within the first 2 treatment cycles. Identification of the RP2D of the combination treatment was based on the estimation of the probability of DLT within the first 2 cycles and all available safety, clinical, PK, and PD data up to the cutoff date (March 16, 2023). Dose escalation was guided by a Bayesian logistic regression model analysis of DLT data from cycle 1 and cycle 2 with the escalation overdose control principle.^28,29 

The impact of ruxolitinib and siremadlin, either alone or in combination, was also evaluated in vitro on apoptosis and colony formation of normal donor and MF CD34⁺ cells. See the supplemental Data for further methods and results.

A total of 44 patients were enrolled in part 1 of the study, including 23 in the ruxolitinib plus siremadlin cohorts in part 1, all of whom received treatment. Ruxolitinib plus siremadlin data are reported here, and available data for other combinations are captured in the supplemental Data.

In the ruxolitinib plus siremadlin cohorts, 3 patients completed the treatment in the core phase and progressed into the extension phase and 20 patients discontinued the treatment in the core phase. The most common reasons for treatment discontinuation were AEs (n = 10, 43.5%), physician’s decision (n = 7, 30.4%), and progressive disease (n = 2, 8.7%; Table 1).

The median (range) age was 68 (37-86) years. Most patients were male (n = 15, 65.2%) and all of White ethnicity. Most patients had an initial diagnosis of primary MF (n = 15, 65.2%), whereas the remaining patients had secondary MF after PV (n = 7, 30.4%) or essential thrombocythemia (n = 1, 4.3%). Median spleen volume was 1940 mL, and palpable spleen length was 11 cm. Median MFSAF TSS was 12.0, and median duration of previous ruxolitinib treatment was 67.9 weeks, with the highest median duration in the ruxolitinib plus siremadlin 30 mg arm (165.1 weeks; Table 2).

Among patients with available mutational data at baseline, 14 patients treated with ruxolitinib plus siremadlin had mutant JAK2 (JAK2V617F; 20 mg, n = 3; 30 mg, n = 6; 40 mg, n = 5) and 5 patients had mutant calreticulin (CALR; 20 mg, n = 3; 30 mg, n = 2). Ten patients (siremadlin 20 mg, n = 2; 30 mg, n = 6; 40 mg, n = 2) possessed a high-risk genetic profile, defined as ≥1 mutation in ASXL1, SRSF2, EZH2, IDH1, or IDH2. One additional patient in the siremadlin 20 mg cohort had a mutation in U2AF1 (Q157R).^30-33 No patient treated with ruxolitinib plus siremadlin had loss-of-function mutations in TP53 at baseline (supplemental Figure 1).

Patient disposition, baseline characteristics, and mutations in other cohorts can be found in supplemental Tables 1 and 2 and supplemental Figure 1, respectively.

In the siremadlin cohort, the median (range) duration of exposure in part 1 was 32.0 weeks (3.9-156.4) and 52.2% of patients were treated for >24 to 48 weeks (supplemental Table 3). No DLTs were reported in the ruxolitinib plus siremadlin 20 mg cohort. In the ruxolitinib plus siremadlin 30 mg cohort, 1 patient experienced grade 4 thrombocytopenia and grade 4 neutropenia. In the ruxolitinib plus siremadlin 40 mg cohort, 2 patients experienced grade 3 thrombocytopenia and 1 had grade 4 thrombocytopenia. The proportion of patients with DLTs was 40.0% in the ruxolitinib plus siremadlin 40 mg cohort, which exceeded the maximum tolerated DLT rate of 33%. Therefore, siremadlin 30 mg was considered the RP2D.

All patients (n = 23) experienced an AE, and 91.3% (n = 21) had a grade ≥3 AE (Table 3). The most common grade ≥3 AEs were hematological; 60.9% (n = 14), 47.8% (n = 11), and 47.8% (n = 11) of patients experienced grade ≥3 anemia, thrombocytopenia, and neutropenia, respectively. Most patients (9/12) who developed grade ≥3 neutropenia did not have neutropenia at baseline. However, 8 of the 12 patients who experienced grade ≥3 thrombocytopenia had grade 1 at baseline. Serious AEs were reported in 34.8% (n = 8) of patients, with 17.4% (n = 4) considered treatment related. Furthermore, 11 patients (47.8%) had an AE leading to discontinuation. The exposure (plasma drug concentration) increased as siremadlin doses increased; however, no association was observed between increased exposure and AE occurrence. In the ruxolitinib plus siremadlin 40 mg cohort, there were 3 deaths attributed to underlying MF, multiple organ dysfunction, and hemorrhage. The hemorrhage was preceded by grade 4 thrombocytopenia and was accompanied by a fatal event of infection; both fatal AEs were considered related to ruxolitinib and siremadlin. An overview of AEs in other cohorts is in supplemental Table 4.

A ≥35% SVR at week 24 was achieved in 7 patients (30.4%), with the highest proportion in the ruxolitinib plus siremadlin 30 mg cohort (n = 6, 60.0%; Figure 2). Most patients had discontinued the study before week 48, but 1 patient at each siremadlin dose level attained a ≥35% SVR at week 48 (supplemental Table 5). One patient in the ruxolitinib plus siremadlin 20 mg cohort had a ≥25% reduction from baseline at weeks 24 and 48 (Figure 2; supplemental Table 5). The overall median percentage change from baseline in spleen volume at week 24 was −22.8%, with the greatest change observed with ruxolitinib plus siremadlin 30 mg (−38.7%). Improvements were sustained for patients receiving treatment up to week 48 (overall median change from baseline, −34.6%). Changes in palpable spleen length at week 24 are found in supplemental Figure 2.

Five patients (21.7%) achieved a ≥50% reduction in MFSAF TSS from baseline at week 24 (Figure 3). Although most patients (n = 17, 73.9%) had discontinued treatment before week 48, 1 patient in each ruxolitinib plus siremadlin cohort had a ≥50% reduction from baseline at week 48.

Change in spleen volume and change in MFSAF TSS from baseline at week 24 in other cohorts are found in supplemental Figures 3 and 4, respectively.

For patients with JAK2V617F at baseline, a median VAF decrease of −12.6% was observed at week 24 in ruxolitinib plus siremadlin patients. The largest decrease occurred with ruxolitinib plus siremadlin 20 mg (−14.0%). One patient in the ruxolitinib plus siremadlin 20 mg cohort tested negative for JAK2 mutation at baseline but was positive at week 24 (Figure 4).

At baseline, mutations in CALR were identified in 5 patients treated with ruxolitinib plus siremadlin (20 mg cohort, 3 patients; 30 mg cohort, 2 patients). Three patients had missing baseline samples but had CALR mutations detected at week 24. A decrease in CALR VAF from baseline on treatment with siremadlin was assessed in 2 patients from the ruxolitinib plus siremadlin 20 mg cohort (−35.8% and −16.0%) and 1 patient from the ruxolitinib plus siremadlin 30 mg cohort (−17.9%).

An increase from baseline in the serum protein level of GDF-15, a downstream target of TP53, was observed for all patients treated with ruxolitinib plus siremadlin. At week 1, mean change from baseline in the GDF-15 concentration was 13.3 ng/mL for patients on ruxolitinib plus siremadlin. The greatest increase occurred with siremadlin 30 mg (15.8 ng/mL), followed by the siremadlin 40 mg (13.2 ng/mL) and siremadlin 20 mg (10.9 ng/mL) cohorts (supplemental Figure 5).

Moreover, 6 patients had emergence of TP53 mutations at week 24 or at the EOT visit: 1 patient in the ruxolitinib plus siremadlin 20 mg cohort and 3 patients in the ruxolitinib plus siremadlin 30 mg cohort at week 24, and 1 patient each in the ruxolitinib plus siremadlin 30 mg and 40 mg cohorts at the end-of-treatment visit (supplemental Table 6). Five of these patients had >1 TP53 mutation. The VAF of the treatment-emergent TP53 mutations at these time points was low (2%-8%).

The acquisition of other new mutations was observed on treatment in 10 patients in the siremadlin cohorts: ASXL1; ABL1 and U2AF1 (Q157R); ASXL1 and GATA2; DNMT3A; KRAS and CBL; MPL; NRAS and PHF6; PTPN11; RUNX1 and STAG2; RUNX1 and ZRSR2 (supplemental Figure 1).

The area under the curve of ruxolitinib remained consistent at all doses between days 1 and 5 for all combinations, suggesting no effect of the combination on the PK of ruxolitinib (supplemental Table 7). For siremadlin, an increase in exposure was observed with increasing doses at 20, 30, and 40 mg, with the mean (SD) corresponding area under curve from time zero to the last measurable concentration sampling time values (on day 5) of 2230 ± 1110, 3340 ± 2120, and 4390 ± 2270 ng∗h/mL, respectively, with moderate variability, and a mean accumulation ratio of 1.03 to 1.47 across doses (supplemental Table 8). PK for other cohorts is found in supplemental Table 9.

The ADORE study assessed the potential benefit of combination therapy for patients with MF who had a suboptimal response to ruxolitinib alone. Using an innovative open platform design, the addition and/or prioritization of treatment arms as efficacy and safety data emerge can accelerate the development of combination treatments. This design may help overcome the downsides to the standard approach of separate phase 1/2 dose-finding and expansion studies before a phase 3 trial, which include needing more patients to be recruited across separate studies, and requiring a ruxolitinib monotherapy control. Moreover, emerging compounds may not be available at the start of the study. Although the sponsor permanently halted patient enrollment, ADORE successfully executed an open platform study design, validating this approach for future studies in this field and beyond.

A total of 44 patients were treated with ruxolitinib combination treatments in part 1 with 23 treated with ruxolitinib plus siremadlin. In this group, the most frequent AEs were gastrointestinal toxicity (generally of low grade) and hematological toxicity (thrombocytopenia, anemia, and neutropenia). Interestingly, the incidence of GI toxicity reported with siremadlin was lower than that observed for other MDM2 inhibitors; however, a direct comparative study would be required to confirm.^34-37 Importantly, the increase in exposure with increased doses of siremadlin did not correlate with any adverse safety observations.

The greatest improvements in spleen size and MFSAF TSS were in the ruxolitinib plus siremadlin 30 mg arm. SVR ≥35% was observed in a third of ruxolitinib plus siremadlin–treated patients, with the highest proportion noted in those treated with siremadlin 30 mg. Combination with siremadlin also resulted in symptomatic benefit with ∼22% reporting a ≥50% reduction in TSS at week 24. These benefits were achieved through addition of siremadlin to patients already experiencing a suboptimal response on existing ruxolitinib therapy with anemia, low symptom burden, and prolonged previous treatment with ruxolitinib at baseline.

Long-term ruxolitinib monotherapy reduces JAK2V617F allele burden in patients with MF and PV,^7,38-40 and in this study, further reductions in JAK2V617F allele burden at week 24 were observed when ruxolitinib was used in combination with siremadlin. An increase in GDF-15 was noted, and although observations need to be confirmed in further studies, these data demonstrate the potential modulation of downstream P53 targets and overall disease-modifying activity of ruxolitinib plus siremadlin.

We identified the emergence of low-VAF TP53 mutations in 6 patients on siremadlin treatment. These mutations were located in the DNA-binding region, leading to inactivation of P53 activity. Owing to the low sample size and low allele frequency, the potential impact of these mutations could not be assessed with the current data set. As TP53 mutation in the neoplastic clone is associated with poorer prognosis and increased risk of disease progression, the identification of treatment-emergent TP53 mutations in this study is important. We were unable to determine whether TP53 mutations arose in subclones of myeloproliferative neoplasm or represented independent clonal hematopoiesis. Our findings are consistent with those of a phase 1 trial of idasanutlin in patients with PV, where expansion of TP53-mutant subclones was observed during therapy but remained stable or reduced after treatment discontinuation.⁴¹ These data suggest that therapies targeting the MDM2-p53 interaction have the potential to induce the expansion of TP53-mutant subclones. This highlights the need for further studies to assess the type of treatment-emergent TP53 mutations using single-cell approaches⁴² and evaluate their prevalence and potential association with disease progression, secondary neoplasia, and therapeutic response. The detection of other new mutations in siremadlin cohorts is of unknown clinical significance. The biomarker analysis and assays used in this study were exploratory and not able to clearly delineate whether new mutations were already present at low VAF at baseline or truly arose on treatment.

The major limitation of this study lies in the permanent halt of enrollment and subsequent early termination. Consequently, not all preplanned study objectives for part 1 could be completed, and none of the objectives for parts 2 and 3 were pursued. The enrollment halt also led to a small study population across treatment arms, limiting the sample size for analysis. Nonetheless, ADORE represents an innovative approach in the development of combination treatments, and available data illustrate the potential benefits of combining agents with ruxolitinib in patients with suboptimal response to ruxolitinib alone. Among the combinations investigated, siremadlin 30 mg and ruxolitinib demonstrated the most promising safety and efficacy profile.

The authors thank the patients and their families who made this trial possible. The authors acknowledge Soracha Ward of Novartis Ireland and Timothy Harries of Novartis Pharmaceuticals UK for providing medical writing support during the preparation of this manuscript, funded by Novartis Pharma AG, Basel, Switzerland, in accordance with the Good Publication Practice 2022 guidelines (https://www.ismpp.org/gpp-2022).

This study was supported by research funding from Novartis.

Contribution: All authors provided substantial contribution to the study design and/or collection, analysis, interpretation of data, were involved in the drafting and/or critical reviewing of the manuscript, provided final approval of the version to be published, and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Conflict-of-interest disclosure: D.M.R. has received honoraria from Merck and Novartis, and has had an advisory role for Keros, Merck, Menarini, Novartis, and Takeda. F.H.H. has participated in advisory roles for Novartis, AOP Orphan, Bristol Myers Squibb (BMS), GlaxoSmithKline (GSK), Merck, AbbVie, and Kartos, and received research funding from Novartis, BMS, and CTI BioPharma Corporation. A.C.P. has received consultation fees from Novartis and GSK, and honoraria from Novartis, CTI BioPharma Corporation, BMS, GSK, and AbbVie. A.R. reports receiving research support from AbbVie, AOP, AstraZeneca, Blueprint Medicines Corporation, BMS, Cogent Therapeutics LLC, GSK, Incyte, and Novartis; consultancy for AOP, Blueprint Medicines Corporation, GSK, and Novartis; honoraria for AOP, Blueprint Medicines Corporation, GSK, and Novartis; and travel reimbursement from AOP, Blueprint Medicines Corporation, GSK, and Novartis. H.B. received honoraria from Astellas, BMS, GSK, MSD, Novartis, and Servier. T.L. reports consultancy and advisory board participation for Novartis and AbbVie; and research funding from Novartis. O.V. reports receiving research funding and honoraria from Novartis. A.M.V. received fees from Novartis, Incyte, AbbVie, AOP, and GSK, and is on advisory boards for Novartis, Incyte, GSK, and AOP. V.G. reports institutional grants from AbbVie and Novartis; consulting fees from Novartis, BMS Celgene, GSK, AbbVie, Pfizer, and Daiichi Sankyo; honoraria from Novartis, BMS Celgene, and AbbVie; and participation on a data safety or advisory board for BMS Celgene, GSK, AbbVie, Pfizer, and Daiichi Sankyo. M. Wondergem reports participating in steering committee for Novartis; a data safety monitoring board for Keros therapeutics; and educational talks for Novartis and AOP. J.-J.K. reports consulting fees from GSK, Novartis, and AbbVie; honoraria from AOP Health and PharmaEssentia; support for attending meetings and/or travel from Novartis; and participation on a data safety monitoring board or advisory board for Incyte and BMS. G.C. is an employee of Novartis Ireland Limited, Dublin, Ireland. A.Z., A.P., and M. Wroclawska are employees of Novartis Pharma AG, Basel, Switzerland. J.K. is an employee and stockholder of Novartis Healthcare Pvt Ltd, Hyderabad, India. C.N.H. has received institutional grants from Constellation, GSK, and Novartis; received consultation fees from Novartis, MSD, Karyopharm, AOP, GSK, BMS, Sobi, Galecto, and CTI; received honoraria fees from Novartis, MSD, Karyopharm, Sobi, GSK, and BMS; received support for attending meetings and/or travel from Novartis; participated on a data safety monitoring board or advisory board for BMS and Galecto; leadership role with Blood Cancer UK (trustee; unpaid), European Hematology Association (Deputy Editor-in-Chief; remunerated), and MPN (myeloproliferative neoplasms) Voice (Medical Director; unpaid); and stock or stock options with Chakana Medical Limited. The remaining authors declare no competing financial interests.

Correspondence: Claire N. Harrison, Haematology, Guy's and St Thomas’ NHS Foundation Trust, Great Maze Pond, London, United Kingdom SE1 9RT; email: claire.harrison@gstt.nhs.uk.