First-in-Human Dose-Escalation Study of the First-in-Class PDE3A–SLFN12 Complex Inducer BAY 2666605 in Patients with Advanced Solid Tumors Coexpressing SLFN12 and PDE3A

Abstract

Purpose:

The study aims to evaluate the safety, tolerability, and pharmacokinetics of BAY 2666605, a velcrin that induces complex formation between the phosphodiesterase PDE3A and the protein Schlafen 12 (SLFN12), leading to a cytotoxic response in cancer cells.

Patients and Methods:

This was a first-in-human phase I study of BAY 2666605 (NCT04809805), an oral, potent first-in-class PDE3A–SLFN12 complex inducer, with reduced PDE3A inhibition. Adults with advanced solid tumors that coexpress SLFN12 and PDE3A received BAY 2666605 at escalating doses starting at 5 mg once daily in 28-day cycles. Forty-seven patients were prescreened for SLFN12 and PDE3A overexpression, and five biomarker-positive patients received ≥1 BAY 2666605 dose.

Results:

The most common adverse event was grade 3 to 4 thrombocytopenia in three of the five patients treated. The long half-life (>360 hours) and associated accumulation of BAY 2666605 led to the selection of an alternative schedule consisting of a loading dose with a once-daily maintenance dose. The maximum tolerated dose was not established as the highest doses of both schedules were intolerable. No objective responses were observed. Due to the high expression of PDE3A in platelets compared with tumor tissues, the ex vivo dose-dependent inhibitory effect of BAY 2666605 on megakaryocytes, and the pharmacokinetic profile of the compound, alternative schedules were not predicted to ameliorate the mechanism-based thrombocytopenia.

Conclusions:

Despite the decreased PDE3A enzymatic inhibition profile of BAY 2666605, the occurrence of thrombocytopenia in treated patients, an on-target effect of the compound, precluded the achievement of a therapeutic window, consequently leading to trial termination.

Translational Relevance.

Several solid tumors, such as melanoma, sarcoma, glioblastoma, and ovarian cancer, coexpress the proteins Schlafen 12 (SLFN12) and PDE3A. When these two proteins are bound together by molecules called velcrins, it leads to a cytotoxic response by stimulating the RNase activity of SLFN12. In this phase I first-in-human study, we administered BAY 2666605, a selective and potent molecular glue that binds SLFN12 and PDE3A together, to patients with solid tumors. Dose escalation was limited to five treated patients due to severe thrombocytopenia. This effect was unexpected based on preclinical studies. Based on ex vivo studies performed on hematopoietic stem cells, the thrombocytopenia was likely due to inhibition of megakaryocyte maturation. In the future, a new cancer cell–specific and PDE3A-independent approach to activating SLFN12 RNase will be required to leverage the anticancer potential of SLFN12 for patient benefit.

Introduction

BAY 2666605 is a selective and potent molecular glue that belongs to a family of small molecules called velcrins (1, 2). Velcrins induce complex formation between the phosphodiesterase PDE3A and the protein Schlafen 12 (SLFN12), leading to a cytotoxic response in cancer cells that coexpress both proteins (3–5). Based on recent studies, the underlying mechanism for BAY 2666605-induced killing of cancer cells is independent of PDE3A enzymatic inhibition (1). Instead, PDE3A binding stimulates the RNase activity of SLFN12, resulting in cleavage of the specific SLFN12 substrate, tRNA–Leu–TAA. The cleavage of tRNA–Leu–TAA causes ribosomal pausing, inhibition of global protein synthesis, and cancer cell death (6).

Preclinically, BAY 2666605 showed nanomolar antiproliferative activity, and treatment of cancer cells with BAY 2666605 resulted in downregulation of the antiapoptotic protein MCL-1 in vitro (1). In vivo, treatment with BAY 2666605 demonstrated antitumor efficacy in various cancers with PDE3A and SLFN12 coexpression: in melanoma and glioblastoma xenograft models, as well as in melanoma, sarcoma, and ovarian cancer patient-derived xenograft models (1). PDE3A and SLFN12 coexpression was required for the antiproliferative activity of BAY 2666605 in vitro. Indeed, PDE3A- and SLFN12-negative patient-derived xenograft models did not respond to BAY 2666605 treatment (1). Based on nonclinical data generated using NanoString nCounter on commercially available formalin-fixed paraffin-embedded (FFPE) tumor tissues, detectable coexpression of SLFN12 and PDE3A is observed in 84% of melanoma samples (N = 43), 78% of sarcoma samples (N = 32; 14 nonrhabdomyosarcoma soft tissue sarcoma and 18 soft-tissue sarcoma samples), 84% of ovarian cancer samples (N = 44), and 60% of glioblastoma multiforme/anaplastic astrocytoma (GBM/AA) tumors (N = 73; ref. 1).

Here, we report the results of the first-in-human phase I study of BAY 2666605 (ClinicalTrials.gov identifier: NCT04809805), characterizing its safety, tolerability, and pharmacokinetics (PK) in patients with advanced solid tumors that were prospectively found to coexpress PDE3A and SLFN12, including metastatic melanoma, advanced recurrent sarcoma, and epithelial ovarian cancer (EOC).

Patients and Methods

Eligibility

In the dose-escalation phase of the trial, patients ≥ 18 years with melanoma, recurrent/advanced sarcoma, and EOC who tested positive for expression of both SLFN12 and PDE3A during study prescreening and who had radiologic progression after standard-of-care therapies were eligible. Further inclusion criteria were as follows: Eastern Cooperative Oncology Group performance status ≤ 2, hemoglobin ≥ 9.0 g/dL, absolute neutrophil count ≥ 1.5 × 10⁹/L, platelet count ≥ 100 × 10⁹/L, and adequate renal and hepatic function.

The major exclusion criteria included history of cardiac disease (coronary artery disease, angina pectoris and myocardial infarction, congestive heart failure, ejection fraction < 50%), clinically relevant rhythm abnormalities, uncontrolled arterial hypertension; active infection; history of hemorrhage, bleeding disorder or platelet function abnormalities; arterial or venous thromboembolic events within 6 months or deep vein thrombosis within 3 months before start of BAY 2666605, respectively, requiring anticoagulant/platelet antiaggregant therapy; major surgery within 4 weeks prior to the first dose; history of gastrointestinal ulcerations or perforation; central nervous system metastases (for the dose escalation only); anticancer therapy ≤ 4 weeks and radiotherapy ≤ 2 weeks of starting therapy; patients requiring therapy with anticoagulant/platelet antiaggregant/antiplatelet/antithrombin inhibitors; transfusion of blood products within 14 days or administration of colony stimulating factors within 4 weeks prior to the first dose of study intervention; and use of strong CYP3A inhibitors or inducers ≤ 14 days of starting therapy or during therapy.

Study design

This phase I study was conducted at three centers in the United States in accordance with Good Clinical Practice guidelines, provided by the International Conference on Harmonization and principles of the Declaration of Helsinki. The institutional review boards or ethics committees at each participating site approved the study, and all patients provided written informed consent before enrollment into prescreening consisting of SLFN12 and PDE3A expression testing on archival tumor tissue.

Patients were administered BAY 2666605 orally continuously in 28-day cycles in two different schedules, once-daily dosing (schedule 1) if the half-life was ≤72 hours or a loading dose (LD) once followed by once-daily maintenance (schedule 2) if the half-life was >72 hours, and until any of the discontinuation criteria were met (disease progression, unacceptable toxicity, or patient or investigator decision to end therapy). The first patient in each of the dose-escalation cohorts was to participate in a PK run-in, receiving a single dose within 7 to 14 days of the start of cycle 1. The planned escalation doses were 5, 10, 20, 40, 80, 160, and 320 mg (as once daily in schedule 1, as LD in schedule 2, and as the PK run-in dose) as well as 0.25, 0.5, 1, 2, 4, 8, and 16 mg for the once-daily maintenance in schedule 2 (Supplementary Table S1). Dose escalation was guided by Bayesian adaptive modeling, targeting a dose-limiting toxicity (DLT) rate of 25% (7). The DLT assessment period was 1 cycle (28 days).

Protocol-defined DLTs were the following: absolute neutrophil count < 0.5 × 10⁹/L for ≥7 days, febrile neutropenia, platelet count < 25 × 10⁹/L of any duration or platelet count < 50 × 10⁹/L associated with grade ≥ 3 bleeding, life-threatening anemia, any grade ≥ 3 nonhematologic toxicity (excluding optimally or sub-optimally treated nausea, vomiting, or diarrhea grade ≥ 3 resolving within 72 hours), any grade ≥ 2 hemorrhage, left ventricular ejection fraction (LVEF) decline to <40% or ≥20% drop from baseline LVEF, any grade ≥ 2 cardiac arrhythmia, and any study drug–related death. A specific GBM/AA cohort was planned to start at one dose level below the maximum tolerated dose (MTD), as well as an expansion cohort planned to enroll patients with metastatic melanoma if the MTD was established.

Study assessments

Adverse events (AE) were monitored throughout the study, and toxicities were assessed using the NCI Common Toxicity Criteria for grading AEs (version 5.0; ref. 8) and Medical Dictionary for Regulatory Activities for AE terminology Version 25.1. Physical examinations and vital signs were done at screening, during the PK run-in and on days 1, 2, 8, 9, 15, 16, 22, and 23 of cycle 1 and at every weekly visit thereafter. Blood samples for hematology and serum chemistry were taken at screening, PK run-in, on days 1, 8, 15, and 22 of cycle 1 and at every weekly visit thereafter. Cardiac markers such as troponin I or T, B-type natriuretic peptide, creatinine kinase (CK), and CK-MB were assessed pre-dose on PK run-in day 1, on days 1, 2, 8, 9, 15, 16, 22, and 23 of cycle 1 as well as on PK run-in day 2 and 2 hours post-dose on PK run-in day 1, day 1, and day 15. Triplicate echocardiograms (ECG) were performed at screening, PK run-in, on days 1, 2, 8, 9, 15, 16, 22, and 23 of cycle 1, and at every weekly visit thereafter, whereas 24-hour Holter ECGs were performed at screening, PK run-in, and days 1, 8, 15, and 22 of cycle 1 only. Cardiac function (echocardiogram) was monitored at screening and on day 22 or within 7 days before day 1 of cycle 2. Solid tumor assessments were done by CT scan or MRI at the end of cycle 2 and every two cycles thereafter. Tumor response and progression were evaluated based on the response evaluation criteria in non-GBM/AA solid tumors (RECIST 1.1; ref. 9).

Plasma samples for PK assessments were collected at pre-dose, 0.5, 1, 2, 4, 6, 8, and 12 hours after oral administration on days 1 and 15 (in addition, also at 24, 48, 72, 96, and 168 hours after the oral administration during PK run-in). PK parameters were assessed by non-compartmental methods and expressed as geometric means. According to the study protocol, for the first patient, the decision for the dosing schedule to be applied was made based on the t_1/2 calculated from PK samples up to and including 48 hours post-dose during the PK run-in. For the subsequent PK run-in patients, the decision for the dosing schedule to be applied was made based on the t_1/2 calculated from PK samples up to and including 168 hours post-dose of the previous patients.

Prescreening was performed on FFPE archival tumor tissue in order to only enroll patients with positive expression of both PDE3A and SLFN12. For that purpose, a customized RNA-based NanoString nCounter panel for quantification of SLFN12 and PDE3A was validated and run in a College of American Pathologists/Clinical Laboratory Improvement Amendments–certified laboratory. For each patient sample, raw SLFN12 and PDE3A counts were normalized using positive control and housekeeping gene counts. To further normalize data between runs, an expression ratio count was calculated for both SLFN12 and PDE3A by dividing the normalized counts of the patient samples with the normalized counts of a calibrator sample (universal human reference RNA, Agilent, # 750500). The ratio counts of SLFN12 and PDE3A were ranked as positive or negative expression based on predefined cut-off values.

Pharmacodynamic biomarker analysis was also planned for patients with melanoma for which fresh paired biopsies were to be collected.

Statistical methods

This was a dose-escalation trial designed to determine the MTD of BAY 2666605 and to characterize its PK, treatment-emergent AEs (TEAE) and DLTs. The safety analysis set included all patients who received ≥1 BAY 2666605 dose. The MTD-evaluable population included all patients who completed cycle 1 or who experienced a DLT in cycle 1. A stepwise dose-selection design based on a modified toxicity probability interval keyboard (mTPI2) design was used to guide dose-escalation decisions and to identify the MTD (7). PK parameters, biomarker SLFN12/PDE3A coexpression, and best tumor response were summarized or listed. Response for patients was assessed by investigators per RECIST 1.1 (9).

Comparison of SLFN12 and PDE3A expression levels between tumor samples and normal platelets using NanoString nCounter and IHC

To evaluate expression levels of SLFN12 and PDE3A in different tumor types, a set of archived FFPE tumor samples from GBM/AA, malignant melanoma, sarcoma, and EOC were analyzed. Furthermore, fresh preparations and FFPE pellets of platelets from healthy donors were tested. Platelets were isolated from platelet rich plasma by centrifugation using standard methods. All samples were procured from commercial providers under appropriate informed consent. RNA expression testing was performed after RNA extraction from FFPE sections using a research use only version of the NanoString (nCounter) assay used in the clinical study setting. PDE3A protein expression was analyzed by IHC using an anti-PDE3A antibody (Bethyl Laboratories, Bethyl Cat. # A302-740A, RRID: AB_10634214) on a Ventana Discovery autostainer using OptiView DAB chemistry. IHC samples were scored by a trained pathologist for PDE3A staining pattern (H-score readout).

Evaluation of BAY 2666605 ex vivo effects on megakaryocytes

The ex vivo effects of BAY 2666605 on human megakaryocytes (MK) were studied using two different validated assays at ReachBio Research Labs. In a semi-solid, collagen-based assay, colony-forming units of MK progenitors were measured and counted 14 days after adding test compounds by immunocytochemical staining for the MK marker CD41. In a liquid assay, human bone marrow–derived CD34⁺ cells in the presence of 50 ng/mL recombinant human stem cell factor and 50 ng/mL recombinant human thrombopoietin were incubated with test compounds over a period of 14 days. Stains for CD41⁺ (MK) and CD42b⁺ CD41⁺ (platelets) were carried out at various time points. Flow cytometry analysis was then performed using a cytoFLEX flow cytometer. In both assays, five compounds were tested at different concentrations: BAY 2666605, anagrelide, 3-OH anagrelide, trequinsin, and amrinone in parallel with a 5-fluorouracil internal positive control.

Data availability

Availability of the data underlying this publication will be determined according to Bayer’s commitment to the European Federation of Pharmaceutical Industries and Associations/Pharmaceutical Research and Manufacturers of America “Principles for Responsible Clinical Trial Data Sharing.” This pertains to scope, time point, and process of data access.

As such, Bayer commits to sharing, upon request from qualified scientific and medical researchers, patient-level clinical trial data, study-level clinical trial data, and protocols from clinical trials in patients for medicines and indications approved in the United States and European Union, as necessary for conducting legitimate research. This applies to data on new medicines and indications that have been approved by the US and EU regulatory agencies on or after January 1, 2014.

Interested researchers can use www.clinicalstudydatarequest.com to request access to anonymized patient-level data and supporting documents from clinical studies to conduct further research that can help advance medical science or improve patient care. Information on the Bayer criteria for listing studies and other relevant information is provided in the study sponsor’s section of this portal.

Data access will be granted to anonymized patient-level data, protocols, and clinical study reports after approval by an independent scientific review panel. Bayer is not involved in the decisions made by the independent review panel. Bayer will take all necessary measures to ensure that patient privacy is safeguarded.

Results

Patient characteristics

From April 2021 to May 2022, 47 patients were enrolled—of whom 42 were prescreening study failures (n = 40) or main trial screening failures (n = 2), with the remaining five patients receiving ≥1 BAY 2666605 dose. The reasons for the 40 prescreening study failures, either initially, just after study entry, or during the long time interval between prescreening and main screening were the following: not meeting entry criteria at time of treatment consent (n = 17), SLFN12/PDE3A test failure (n = 11: negative test in seven patients or invalid tumor tissue in four patients), death (n = 3), subject’s or physician’s decision (n = 5), and study termination by sponsor (n = 4). The two main trial screening failures were due to not meeting entry criteria.

Baseline disease and treatment characteristics are listed in Table 1. The median age was 58 years (range, 47–67 years). Four of the five patients were women. All patients had advanced disease (stage IV), and most of them had received less than four prior lines of systemic anticancer therapy. Five patients were treated in four different cohorts: one patient at schedule 1, dose level 1 with BAY 2666605 at 5 mg once daily, one patient at schedule 2, dose level 1 (5 mg LD followed by 0.25 mg maintenance once daily), one patient at schedule 2 dose, level 2 (10 mg LD followed by 0.5 mg maintenance once daily), and two patients at schedule 2, dose level 3 (20 mg LD followed by 1 mg maintenance once daily). The median number of cycles received was 2 (range 1–2).

Table 1.

Baseline patient characteristics (N = 5).

Characteristic	Total (n = 5)
Age, years
Median	58
Range	47–67
Sex, n (%)
Female	4 (80)
Male	1 (20)
Lines of prior systemic therapies, n (%)
1–2 therapies	2 (40)
3–4 therapies	3 (60)
Tumor type, n (%)
Soft tissue sarcoma	3 (60)
Malignant melanoma	1 (20)
EOC	1 (20)
Solid tumors, TNM stage at study entry, n (%)
Stage IV	5 (100)
Baseline value of ECOG PS
1	5 (100)

Abbreviations: EOC, epithelial ovarian cancer including fallopian tube and primary peritoneal cancer (in this specific patient, fallopian tube cancer); ECOG PS, Eastern Cooperative Oncology Group performance status; TNM, tumor–node–metastasis classification of malignant tumors.

Safety profile and patient disposition

The most common TEAE observed in the study was thrombocytopenia, reported early during treatment around cycle 1 day 15, in 3/5 (60%) of treated patients (Fig. 1; Supplementary Table S2), which was assessed as being related to BAY 2666605 in all three patients. The highest Common Terminology Criteria for Adverse Events (CTCAE) grade reported for thrombocytopenia was grade 4 in one patient. These events were nonserious, did not lead to bleeding nor necessitate platelet transfusion, and spontaneously recovered after drug interruption. One patient had grade 3 anemia, which was assessed as drug-related, but all other TEAEs were grade 1 in severity (Supplementary Table S2). No changes in white blood cells, cardiac markers, LVEF, triplicate ECGs, and 24-hour Holter during cycle 1 were identified.

Figure 1.

Individual time courses of platelet counts (10⁹/L). CXDX, cycle x, day x; EOT, end of treatment; FUP, follow-up; Maint, maintenance dose; SCRN, screening; UNS, unscheduled; LLOQ, lower limit of quantitation.

One patient, a 51-year-old woman with metastatic soft tissue sarcoma with pulmonary metastases, treated in schedule 2, dose level 3 (20 mg LD, 1 mg once-daily maintenance), experienced grade 3 pneumonitis, which was deemed a serious DLT [hospitalization and supportive therapy with oxygen, steroids, and antibiotics (cefepime) were necessary] and led to the interruption of BAY 2666605. On the same day of the occurrence of the pneumonitis, grade 3 thrombocytopenia was noted [the patient had low platelet counts (grade 1) at baseline, 110 × 10⁹/L]. The pneumonitis improved with the supportive therapy, and the patient was discharged on steroids. The study drug was permanently discontinued due to progressive disease.

Overall, the MTD of BAY 2666605 was not determined. The planned highest dose on any schedule was not tested because it was considered unlikely that a therapeutic window would be identified and therefore a positive benefit/risk ratio achieved. There were no dose reductions.

Of the five treated patients, two had a dose interruption in cycle 1: one patient, in schedule 1/dose level 1 due to thrombocytopenia and one patient, in schedule 2/dose level 3 due to pneumonitis, respectively, and the latter was also deemed a DLT as described above (Table 2). BAY 2666605 was not resumed due to these AEs (both assessed as study drug related) and was permanently discontinued due to progressive disease or physician decision. Most patients (80%) withdrew due to disease progression (Table 2). No deaths occurred during the study in the five treated patients.

Table 2.

Dose cohorts.

Patient/cancer type	Dose/cohort	PK run-in (Y/N)	Dose modification in cycle 1 (Y/N)	Reason for dose modification (drug interruption)	DLT (Y/N)	Reason for study drug discontinuation	Duration of treatment (days)
1 Sarcoma	5 mg once daily/schedule 1	Y	Y	AE: platelet count decreased grade 3 (and progressed to grade 4)	N	PD: radiologic assessment	35
2 Sarcoma	5 mg LD (0.25 mg once-daily maintenance)/schedule 2	Y	N/A		N	Physician decision	69
3 Melanoma	10 mg LD (0.5 mg once-daily maintenance)/schedule 2	Y	N/A		N	PD: radiologic assessment	63
4 Sarcoma	20 mg LD (1 mg once-daily maintenance)/schedule 2	N	Y	AE: pneumonitis (DLT)(other AE: platelet count decreased grade 3)	Y	PD: radiologic assessment	15
5 Ovarian	20 mg LD (1 mg once-daily maintenance)/schedule 2	N	N/A	(No dose modification but AE: platelet count decreased grade 3)	N	PD: non-radiologic assessment	55

Abbreviations: N, no; N/A, not applicable; PD, progressive disease; Y, yes.

PK

BAY 2666605 was rapidly absorbed after oral administration, with a time to reach maximum concentrations in plasma (t_max) ranging from approximately 0.5 to 2 hours.

Figure 2 presents individual plasma concentrations of BAY 2666605 from the PK run-in after oral administration of a single dose of 5 or 10 mg BAY 2666605. The estimated t_1/2 was <72 hours based on PK data from day 1 and day 2 of the PK run-in, so the first treated patient continued to be dosed with schedule 1. A ∼2-fold accumulation for C_max and ∼5-fold accumulation for daily exposure [AUC_(0–24)] was observed on cycle 1 day 15 compared with cycle 1 day 1 in this patient. The estimated plasma terminal t_1/2 of BAY 2666605 based on data from the PK run-in, in which PK samples were collected over 168 hours, was prolonged with all estimates over 360 hours (>15 days; Table 3). Consequently, subsequent patients were put on schedule 2, using a LD on cycle 1 day 1 with a maintenance dose (MD) starting on cycle 1 day 2, to achieve steady-state PK earlier and to minimize accumulation. All patients using schedule 2 (LD + MD) showed similar trough concentrations during cycle 1 after a once-daily maintenance dose was administered starting on cycle 1 day 2, suggesting little to no accumulation and that the aim to reach steady state early was achieved. Given the small sample size overall, the PK results should be interpreted with caution.

Figure 2.

Individual concentrations of BAY 2666605 (μg/L) in plasma after a single-dose administration of BAY 2666605 at doses of 5 and 10 mg. Values below LLOQ are depicted by a vertical arrow. LLOQ, lower limit of quantification.

Table 3.

Individual PK parameters of BAY 2666605 in plasma (PK analysis set).

Schedule	Schedule 1: once daily	Schedule 2: LD on cycle 1 day 1 with once daily + MD starting on cycle 1 day 2
Dose	5 mg (n = 1)	5 mg LD + 0.25 mg MD (n = 1)	10 mg LD + 0.5 mg MD (n = 1)	20 mg LD + 1.0 mg MD (n = 2)a
PK run-in (single dose)
Main PK
Cmax (μg/L)	56.5	50.6	50.0	NAb
AUC(0–24) (μg·hour/L)	174	359	306	NAb
Additional PK
AUC(0–168) (μg·hour/L)	575	1,320	989	NAb
tmax (hour)	0.500	1.00	2.13	NAb
t1/2 (hour)	(511)c	(505)c	(367)c	NAb
tlast (hour)	167	169	168	NAb
Cycle 1 day 1
Main PK
Cmax (μg/L)	50.9	52.6	40.3	256 (246–265)
AUC(0‐24)md (μg·hour/L)	210	366	345	1,440 (1,380–1,490)
Additional PK
tmax (hour)	0.517	0.933	2.10	0.983 (0.950–1.02)
tlast (hour)	24.9	24.3	23.3	24.9 (24.6–25.2)
Cycle 1 day 15
Main PK
Cmax,md (μg/L)	109	11.8	15.2	53.5 (n = 1)
AUC(0–24)md (μg·hour/L)	1,070	237	301	902 (n = 1)
Additional PK
tmax,md (hour)	0.483	2.03	1.15	0.600 (n = 1)
tlast,md (hour)	24.3	23.1	22.9	24.1 (n = 1)

Abbreviations: AUC_(0–24), area under the plasma concentration vs. time curve from t = 0 to 24 hours; AUC_(0–24)md, AUC_(0–24) after multiple-dose administration; C_max, maximum drug concentration in plasma; C_max,md, C_max after multiple-dose administration; n, number of patients; NA, not available; t_1/2, half-life; t_last, time of the last data point > LLOQ; t_last,md, t_last after multiple-dose administration; t_max, time to reach C_max; t_max,md, t_max after multiple-dose administration.

^aMedian (range) is given for n = 2.

^bData not available as PK run-in was only performed on the first three patients evaluable for PK as per protocol.

^cHalf-life could not be reliably calculated (time range < 2 half-lives does not meet the validity criteria for half-life calculation).

Tumor response

There were no objective responses (per RECIST 1.1 response criteria) observed in any of the five treated patients (objective response rate 0%). The best overall response was stable disease in one patient, progressive disease in three patients, and not evaluable in one patient. Median progression-free survival and overall survival were not determined.

Expression levels of SLFN12 and PDE3A in archival tumor tissue

Expression levels of SLFN12 and PDE3A in archival tumor tissue that the patients provided at study entry (prescreening) could be measured in 39 patients (four invalid samples; Supplementary Table S3). The prevalence of SLFN12/PDE3A coexpression per tumor type was 67% (8/12) for melanoma, 100% (4/4) for EOC, and 87% (20/23) for sarcoma. The mean (range) SLFN12 and PDE3A ratios across the 39 patients were as follows: SLFN12 4.21 (0.56–11.23) and PDE3A 1.53 (0.05–7.95).

Pharmacodynamics

No paired tumor biopsies were collected, and no pharmacodynamics analysis was performed in this study.

SLFN12 and PDE3A expression levels in tumor samples and platelets

Commercial tumor and platelet samples were analyzed using the nCounter NanoString assay for RNA expression and by IHC. Whereas SLFN12 RNA expression was similar between tumor samples and platelets, PDE3A RNA levels seemed to be significantly higher in platelets compared with tumor samples in the relevant indications (melanoma, GBM/AA, EOC, and sarcoma; Fig. 3A and B). This result was further confirmed at the protein level, by IHC (Fig. 3C).

Figure 3.

nCounter gene expression analysis for SLFN12 (A) and PDE3A (B) in different tumor tissues and human platelets from healthy donors. C, IHC staining for PDE3A in malignant melanoma (1–4) and normal platelets (5–8) and pathology (H-) score of PDE3A staining (mean of four samples).

Evaluation of BAY 2666605 ex vivo effect on human platelet progenitors

The SLFN12–PDE3A complex inducers BAY 2666605, anagrelide, and 3-OH anagrelide reduced the number of MK progenitor total colony-forming unit colonies in collagen only at the highest doses, and IC₅₀ values could not be established (Supplementary Fig. S1); whereas trequinsin and amrinone, PDE3A inhibitors lacking complex forming activity, did not have any effect. These data suggest that thrombocytopenia observed in trial patients was not likely caused by cytotoxicity of BAY 2666605 on hematopoietic stem cells or early MK progenitors. Instead, a flow-based assay measuring more mature MK (CD41⁺ bright gate) revealed that BAY 2666605 inhibits MK maturation in a dose-dependent manner without significantly affecting the total number of MK (Supplementary Fig. S2). A similar dose-dependent trend was observed for platelets (CD41⁺ CD42⁺ gate); however, the number of platelets forming in this assay was very low, making the data too variable to draw conclusions (data not shown). Trequinsin and amrinone had no effect on MK maturation, suggesting that the mechanism of action causing thrombocytopenia is via PDE3A–SLFN12 complex formation. Although these data were collected from a single donor, they provide mechanistic insight into what was observed in patients at an in vitro dose calculated to be in the range of free drug in the plasma. Further studies to dissect the selective mechanism of action of velcrins on MK are warranted.

Discussion

In this phase I first-in-human study, once daily BAY 2666605, a PDE3A–SLFN12 complex inducer administered orally in patients with solid tumors, induced as the most frequent AE event thrombocytopenia CTCAE grade 3 in three out of the five treated patients (progressing to CTCAE grade 4 in one patient). Thrombocytopenia was not reported as serious (no hospitalization, no bleeding, and no transfusion) and was manageable and reversible after drug interruption. The single DLT observed was pneumonitis, although the occurrence of a late immune-related pneumonitis following a prior systemic anticancer therapy containing an anti–programmed cell death protein-1 mAb 3 months prior to study entry could not be excluded.

BAY 2666605 was optimized to have reduced PDE3A activity in comparison to approved PDE3A inhibitors (manuscript in preparation), while maintaining SLFN12–PDE3A complex formation (2, 10). However, because positive inotropy and chronotropy had been observed in telemetered rats in relationship with the PDE3A inhibiting effects of BAY 2666605 and due to the high expression of PDE3A in normal cardiovascular tissues reported in the literature and in publicly available databases, several measures were implemented in the study protocol to mitigate cardiovascular risks, such as sufficient margins for the starting dose, exclusion criteria for prior and current cardiovascular diseases, clinical cardiac assessments, cardiac biomarkers (troponin), triplicate ECGs, 24-hour Holters, and rigorous stopping rules. No cardiovascular treatment-emergent events were reported during the study, including around the C_max, which were expected to be possible due to the cardiovascular effects of inhibiting PDE3A. In addition, although inhibition of human ADP-induced platelet aggregation was observed in vitro with BAY 2666605 in line with PDE3A inhibitory activity, none of the treated patients presented with bleeding.

Due to the long half-life of BAY 2666605 (>360 hours), and subsequent accumulation identified from the PK samples of the first patient treated at the first dose level of 5 mg once-daily continuous schedule (schedule 1), coupled with the grade 3 to 4 thrombocytopenia that was observed for the first time, and as planned per study protocol, the dosing schedule was changed from fixed dose once daily to a LD followed by maintenance dosing (schedule 2) in order to accelerate time to steady state systemic exposure. This adaptive measure was implemented in the clinical study protocol from the start of the trial, due to uncertainties in predicting human PK based on large differences observed in animal experiments across species. With schedule 2, the decrease in platelet count reoccurred in the two patients treated at dose level 3 (20 mg LD and 1 mg once-daily MD), confirming the thrombocytopenic effect of BAY 2666605 and its close link with systemic exposure.

Anagrelide, initially developed as a PDE3A inhibitor, was granted approval by the FDA in March 1997 for the treatment of essential thrombocythemia (11, 12). However, the mechanism of action by which anagrelide elicits its platelet-lowering effect via the inhibition of MK maturation has not been completely elucidated (13–20). Recently, it was discovered that although PDE3A is essential for anagrelide-induced cancer cell growth inhibition, the interaction between PDE3A and SLFN12, promoted by anagrelide, is required to induce cancer cell death (21). This confirms that anagrelide acts as a molecular glue similar to other velcrin compounds, such as DNMDP (3), 17-β-estradiol (22), and nauclefine (23), which bind to the catalytic pocket of PDE3A (24) and induce SLFN12–PDE3A complex formation.

An earlier study reported that anagrelide-induced MK cell death was blocked by cotreatment with the PDE3A inhibitor trequinsin, strongly suggesting that the platelet-lowering effect observed in patients directly derives from the PDE3A–SLFN12 interaction (24). In our ex vivo experiment with human platelet progenitors, we could confirm that BAY 2666605, anagrelide, and 3-OH anagrelide all arrested MK maturation and that BAY 2666605 had the greatest dose-dependent impact. Moreover, high expression of PDE3A and moderate-to-high expression of SLFN12 were seen in platelet samples of nine healthy donors. Altogether, these data confirmed the on-target BAY 2666605-induced thrombocytopenia, which was not identified in animal studies. There is no mouse ortholog of human SLFN12, and although monkeys express SLFN12, the sequence of the PDE3A interacting region of monkey SLFN12 is different than that of human SLFN12. In this study, no evidence of SLFN12 in complex with PDE3A was found in monkeys (data not shown).

No objective response was reported in the five treated patients, with the best response observed being stable disease. The archival tumor tissue of these patients coexpressed SLFN12 and PDE3A, as required at study entry. The expression levels of SLFN12 and PDE3A that would predict activity of BAY 2666605 could not be determined in humans. However, although coexpression of SLFN12 and PDE3A is required for efficacy and the levels of both proteins correlate with the efficacy of PDE3A–SLFN12 complex inducers, in cell lines, it seems that the expression level of PDE3A is the major determinant in the degree of response (25). A higher PDE3A expression was observed in platelets in comparison to tumor tissue, which may explain the predominant thrombocytopenia. Due to the high probability of not achieving a therapeutic window or a positive benefit/risk for the patients, the study was terminated during the dose-escalation phase. In the future, a new cancer cell–specific and PDE3A-independent approach to activating SLFN12 RNase will be required to leverage the anticancer potential of SLFN12 for patient benefit.

Supplementary Material

Supplementary Data 1

Supplemantary Figures 1 and 2; Supplementary Tables 1, 2, and 3.

Authors’ Disclosures

K.P. Papadopoulos reports other support from Bayer during the conduct of the study, as well as other support from AbbVie, Amgen, AnHeart, AstraZeneca, Bicycle Therapeutics, BioNTech, CytomX, Daiichi Sankyo, Incyte, Genmab, Merck, F-star, Mirati Therapeutics, Tempest Therapeutics, Sensei Therapeutics, Storm Therapeutics, Monte Rosa Therapeutics, Debiopharm Group, Treadwell Therapeutics, Eli Lilly and Company, Pfizer, Kezar Life Sciences, Pharmamar, Revolution Medicine, Frontier Medicines, and Fog Pharma outside the submitted work. M. McKean reports grants from Bayer during the conduct of the study, as well as grants from Aadi Biosciences, Alpine Immune Sciences, Arcus Biosciences, Arvinas, Ascentage Pharma Group, ASCO, Astellas, Aulos Bioscience, Bayer, Bicycle Therapeutics, BioMed Valley Discoveries, BioNTech, Boehringer Ingelheim, C4 Therapeutics, Dragonfly Therapeutics, EMD Serono, Epizyme, Erasca, Exelixis, Foghorn Therapeutics, G1 Therapeutics, Genentech/Roche, Gilead Sciences, GlaxoSmithKline, IconoVir Bio, IDEAYA Biosciences, Ikena Oncology, ImmVira Pharma, Infinity Pharmaceuticals, Jacobio Pharmaceuticals, Jazz Pharmaceutical, Kechow Pharma, Kezar Life Sciences, Kinnate BioPharma, Krystal Biotech, MedImmune, Mereo BioPharma, Metabomed, NBE Therapeutics, Nektar, Novartis, NucMito Pharmaceuticals, OncoC4, Oncorus, OnKure, PACT Pharma, Plexxikon, Poseida, Prelude Therapeutics, Pyramid Biosciences, Remix Therapeutics, Sapience Therapeutics, Scholar Rock, SeaGen, Synthrox, Takeda Pharmaceuticals, Teneobio, Tempest Therapeutics, Tizona Therapeutics, Tmunity Therapeutics, TopAlliance Biosciences, and Xilio; grants and other support from Bristol Myers Squibb, Daiichi Sankyo, Moderna, Pfizer, and Regeneron; and other support from AbbVie, Castle Biosciences, IQVIA, Merck, Pierre Fabre, and Revolution Medicine outside the submitted work. S. Goldoni reports being an employee of Bayer AG and owns company shares. I. Genvresse reports other support from Bayer AG during the conduct of the study, as well as holding Bayer shares. R. Li reports other support from Bayer US during the conduct of the study, as well as other support from Bayer US outside the submitted work, and reports being an employee of Pharmaceutical Division of Bayer US. G. Wilkinson reports being an employee and shareholder of Bayer AG. T.A. Yap reports other support from University of Texas MD Anderson Cancer Center, where T.A. Yap is a Vice President, Head of Clinical Development in the Therapeutics Discovery Division, which has a commercial interest in DDR and other inhibitors (IACS30380/ART0380 was licensed to Artios), and Seagen; personal fees from AbbVie, Acrivon, Adagene, Almac, Aduro, Amgen Inc., Amphista, Astex, Athena, Atrin, Avenzo, Avoro, Axiom, Baptist Health Systems, BioCity Pharma, Bloom Burton, Boxer, BridGene Biosciences, Bristol Myers Squibb, C4 Therapeutics, Calithera, Cancer Research UK, Carrick Therapeutics, Circle Pharma, Cybrexa, Daiichi Sankyo, Dark Blue Therapeutics, Debiopharm, Diffusion, Duke Street Bio, 858 Therapeutics, EcoR1 Capital, Ellipses Pharma, Entos, FoRx Therapeutics AG, Genesis Therapeutics, Genmab, Glenmark, GLG, Globe Life Sciences, Grey Wolf Therapeutics, GSK, Guidepoint, IDEAYA Biosciences, Idience, Ignyta, I-Mab, ImmuneSensor, Impact Therapeutics, Institut Gustave Roussy, Intellisphere, Janssen, Joint Scientific Committee for Phase I Trials in Hong Kong, Kyn, Lumanity, MEI Pharma, Mereo, Merit, Monte Rosa Therapeutics, Natera, Nested Therapeutics, Nexys, Nimbus, Novocure, Odyssey Therapeutics, OHSU, OncoSec, Ono Pharma, Onxeo, PanAngium Therapeutics, Pegascy, PER, Piper Sandler, Prelude Therapeutics, Prolynx, Protai Bio, Radiopharma Theranostics, resTORbio, Schrodinger, Servier, Synthis Therapeutics, TCG Crossover, TD2, Terremoto Biosciences, Tessellate Bio, Theragnostics, Terns Pharmaceuticals, Thryv Therapeutics, TOLREMO, Tome, Trevarx Biomedical, Varian, Veeva, Versant, Vibliome, Voronoi Inc., Xinthera, Zai Labs, and ZielBio; grants and personal fees from Artios, AstraZeneca, Bayer, BeiGene, Blueprint, Clovis, EMD Serono, F-Star, Kyowa Kirin, Merck, Pfizer, Pliant Therapeutics, Repare, Roche, Ryvu Therapeutics, SAKK, Sanofi, Synnovation, and Tango; and grants from BioNTech, Bristol Myers Squibb, Boundless Bio, Constellation, CPRIT, Cyteir, Department of Defense, Eli Lilly and Company, Exelixis, Forbius, GlaxoSmithKline, Genentech, Gilead, Golfers against Cancer, Haihe, IDEAYA Biosciences, Insilico Medicine, Ionis, Ipsen, Jounce, Karyopharm, KSQ, Mirati, Novartis, NIH/NCI, Prelude, Ribon Therapeutics, Regeneron, Rubius, Scholar Rock, Seattle Genetics, Tesaro, V Foundation, Vivace, Zenith, and Zentalis outside the submitted work. No disclosures were reported by the other authors.

Authors’ Contributions

K.P. Papadopoulos: Conceptualization, data curation, investigation, writing–review and editing. M. McKean: Data curation, investigation, writing–review and editing. S. Goldoni: Conceptualization, data curation, investigation, methodology, writing–original draft, writing–review and editing. I. Genvresse: Conceptualization, data curation, supervision, methodology, writing–original draft, writing–review and editing. M.F. Garrido: Conceptualization, data curation, supervision, investigation, methodology, writing–original draft, writing–review and editing. R. Li: Conceptualization, data curation, formal analysis, supervision, methodology, writing–original draft, writing–review and editing. G. Wilkinson: Conceptualization, data curation, supervision, investigation, methodology, writing–original draft, writing–review and editing. C. Kneip: Conceptualization, data curation, supervision, investigation, methodology, writing–original draft, writing–review and editing. T.A. Yap: Conceptualization, data curation, investigation, writing–review and editing.