Safety, Tolerability, Pharmacokinetics (PK) and Preliminary Antitumor Activity of the Cyclin-Dependent Kinase-9 (CDK9) Inhibitor AZD4573 in Relapsed/Refractory Hematological Malignancies: A Phase 1 First-in-Human Study

Introduction: CDK9, a serine/threonine kinase, regulates transcription elongation by phosphorylating RNA polymerase II at serine 2 (pSer2). Transient inhibition of CDK9 modulates expression of genes with short-lived transcripts and labile proteins, providing a potential therapeutic opportunity in tumors dependent upon oncogenes fitting these criteria. AZD4573 is a highly potent and selective CDK9 inhibitor that rapidly and preferentially decreases BCL2 family anti-apoptotic proteins MCL-1, BFL-1 as well as the oncogene MYC, inducing apoptosis in a broad range of human hematological malignant cell lines. AZD4573 monotherapy has shown antitumor effects associated with transient CDK9 inhibition in leukemia and lymphoma preclinical models. This multicenter, nonrandomized, open-label, phase 1 study (NCT03263637) was conducted to assess safety, tolerability, PK, pharmacodynamics (PD) and preliminary antitumor activity of AZD4573 in patients (pts) with relapsed or refractory hematological malignancies. Methods: The study comprised two parallel dose-escalation arms using a 3+3 design. Arm A included pts with diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma, or multiple myeloma. Arm B included pts with acute myeloid/lymphocytic leukemia, chronic myelomonocytic/lymphocytic leukemia, or hairy cell leukemia. Pts must have received at least 2 prior lines of therapy for the current histology. In each arm, pts received AZD4573 IV on an 8-week cycle using an intra-patient ramp up starting at 6 mg and increasing in 3 mg increments to a target dose of 18 mg (cohort 1) or 12 mg (cohort 2). The initial dosing regimen was 2 days on/12 days off but was changed to once weekly (QW) in cohort 1. Pts in Cohort 3 received AZD4573 9 mg QW on a 4-week cycle including intra-patient ramp up (3 mg, 6 mg, 9 mg). The primary objectives were to assess safety and to determine the biologically effective dose/maximum tolerated dose (MTD). Adverse events (AE) were assessed using CTCAE v4.03. Secondary/exploratory endpoints included PK, PD, and antitumor efficacy. Results: Overall, 44 pts were treated and evaluated for safety and efficacy: 22 each in Arms A and B. Median age was 69.0 years (range, 26.0 - 84.0), 65.9% were male, and median number of prior lines of therapy was 3 (range, 1-11). At data cutoff (Sept 30, 2021), all pts had discontinued treatment, mainly due to disease progression (47.7%) or AEs (22.7%). All pts had treatment-emergent adverse events (TEAEs) which were serious in 88.6%. The most common TEAEs were diarrhea (59.1%), pyrexia (52.3%) and nausea (50.0%). The most common TEAE that led to discontinuation was pneumonia (4.5%). Treatment-related AEs (TRAEs) occurred in 93.2% of pts and were serious in 59.1%, but none led to death. The most common TRAEs were diarrhea (50%), nausea (47.7%) and tumor lysis syndrome (TLS, 40.9%: laboratory TLS, 38.6%; clinical TLS, 6.8%) (Table 1). Elevated liver enzymes were transient and resolved spontaneously. These increases were assumed to be mainly due to down-modulation of hepatic transporter proteins and reduced enzyme clearance rather than direct hepatocellular cytotoxicity (based on clinical, in-vitro and in-silico assessments). Regarding MTD, the 18 mg dose was not tolerated in either arm (clinical TLS and acute kidney injury in Arm A; hypotension and liver injury in Arm B). AZD4573 12 mg was tolerated in Arm A, but not tolerated in Arm B due to clinical TLS (n=2). The recommended Phase 2 dose (RP2D) was 12 mg QW in Arm A and 9 mg QW in Arm B. AZD4573 showed dose-proportional increases in AUC and Cmax, with moderate PK variability (~30-60% CV), and a t1/2 of ~5 hrs. Furthermore, AZD4573 led to dose-dependent reduction in pSer2 (>75%) and MCL-1 (>70%) in peripheral blood as evidence of target engagement and mechanism of action. Median duration of exposure was 10.1 wks in Arm A (range, 1.0-130.1) and 5.2 wks in Arm B (range, 1.0-35.7). In Arm A, there was 1 complete and 1 partial response in 17 DLBCL pts. Median overall survival was not reached in Arm A and was 8.8 months in Arm B, with 7 (31.8%) deaths in each. Conclusion: AZD4573 had manageable safety and a PK profile suitable for QW dosing in a broad range of hematologic malignancies including lymphoma and leukemia. The reduction in pSer2 and MCL-1 supports the RP2D of 12 mg for lymphoma and 9 mg for leukemia. A combination study of AZD4573 and acalabrutinib (NCT04630756) is ongoing to further examine the DLBCL response signal seen in this study. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal