Final Analysis of the Dose Escalation, Expansion and Biomarker Correlations in the Ph I/II Trial BGBC003 with the Selective Oral AXL Inhibitor Bemcentinib (BGB324) in Relapsed/Refractory AML and MDS

Background

AXL has been established as an independent prognostic factor in AML. AXL also represents a novel immune checkpoint because it directly inhibits NK cells and suppresses antigen presentation by antigen-presenting cells.

Bemcentinib is a first-in-class, selective, orally bioavailable, AXL kinase inhibitor which is being evaluated as a therapy for solid tumours and myeloid malignancies in multiple phase II clinical trials.

Increased levels of the plasma soluble, shed form of the AXL receptor (sAXL) upon bemcentinib exposure have been found across the clinical trial programme.

The Phase I/II trial BGBC003 (NCT02488408) evaluates the anti-leukaemic effect of bemcentinib in patients with relapsed / refractory AML and int-2 and high-risk MDS.

Methods

Part A of the BGBC003 consisted of a dose escalation part per standard 3+3 design, followed by cohort expansion at the recommended phase 2 dose.

Plasma protein biomarker levels were measured using the DiscoveryMap v3.3 panel (Myriad RBM) at pre-dose and after one cycle of treatment.

Gene expression analysis was carried out on RNA extracted from BM-MNCs by qPCR using TaqMan.

The presence of phosphorylated AXL was measured using Western blotting.

The TCRß repertoire was quantified by NGS of DNA isolated from PBMNCs using an Illumina MiSeq sequencer. TCRß genes and the IGH repertoire were analysed with BIOMED2-TCRß-A and -B and BIOMED2-FR1/-FR3 primer pools, respectively. Using genomic DNA as template, the amplicons were tagged with Illumina adapters and indices in two consecutive PCR reactions. Demultiplexing and FastQ formated data output was generated by the MiSeq reporter. Analysis of TCRß and IGH data was performed on a Microsoft Cloud using our in-house analysis pipeline Pippa.

Results

In total, 36 patients were enrolled in part A of the study: 27 male (75%), median age of 72 years old (51 - 85), 3 patients with MDS (8%), median of 2 lines of prior therapy (0 - 6).

A loading dose of 400 mg on days one to three followed by 200 mg daily thereafter has been established as safe and recommended phase 2 dose.

In patients who received treatment for at least 21 days (n=28), an ORR (CR/Cri/PR) of 25% (7/28) was observed. Median DoR was 63 days.

In the subset of patients with available plasma measurements at screening (n=19), 12 showed low levels of plasma soluble shed AXL and an ORR of 58% (sAXL-low; 7 of 12 patients) was observed. No responses occurred in patients with high levels of sAXL at screening (sAXL-high; 0 out of 7). sAXL levels were correlated with exposure and one cycle of treatment with bemcentinib significantly increased sAXL protein levels, particularly in responders to treatment indicating on-target activity of the drug.

Median PFS from the day of randomisation until evidence of disease progression (PD) per bone marrow assessment or discontinuation due to PD whichever occurred earlier, was 62 days overall (95% CI: 18 - 106 d), 124 days in sAXL-low patients and 63 days in sAXL-high patients, respectively.

Adverse events were mostly gastrointestinal and mild or moderate in severity congruent with bemcentinib monotherapy being well-tolerated in this patient population. There were no grade 4 or 5 events.

We observed a diversification of the TCR repertoire in the bone marrow in 75% of evaluable patients (6/8). In the peripheral blood we detected increased TCR diversification in 44% of evaluable patients (4/9). These data indicate that the contribution of the Gas6-AXL axis to immunosuppression might be particularly pronounced in the bone marrow microenvironment. This notion is in concordance with earlier data indicating a specific upregulation of Gas6 by bone marrow stroma cells upon paracrine interaction with AML cells leading to locally increased Gas6 levels. We also observed diversification of the BCR, however less frequently (bone marrow 42%, peripheral blood 11%).

Conclusions

Bemcentinib is well tolerated in relapsed / refractory MDS and AML patients and exhibits anti-leukaemic activity. Furthermore, we observed a diversification of the TCR repertoire particularly in the bone marrow thereby supporting a role of AXL as immune checkpoint. Patient benefit was correlated with the level of the plasma marker soluble AXL at screening. This predictive biomarker candidate warrants further exploration in larger cohorts and has potential for development as companion diagnostics to select patient populations more likely to respond to bemcentinib treatment.