Combined CXCR-4 Inhibition with Novel Agent GPC-100 (burixafor) and Beta 2 Adrenergic Receptor Blockade Enhances Cytarabine Response for Acute Myeloid Leukemia Blasts on Stroma

Introduction: Outcomes in adverse risk and relapsed/refractory (R/R) acute myeloid leukemia (AML) remain poor. The aberrant bone marrow (BM) microenvironment is believed to contribute to treatment resistance. C-X-C chemokine receptor type 4 (CXCR4; CD184) is a G protein-coupled receptor for CXCL12 (SDF-1α) that mediates migration to and retention of normal and malignant cells in the BM. The CXCR4 inhibitor plerixafor improves yields of mobilized normal hematopoietic stem cells (HSCs) in combination with filgrastim (G-CSF). Clinical trials of CXCR4 inhibitors have been conducted to mobilize AML out of the protected BM niche, including plerixafor with 7+3 or MEC, BL-8040 with cytarabine (araC), LY2510924 with idarubicin/araC, and ulocuplumab (human IgG4 antibody) with MEC. For the ulocuplumab trial NCT01120457, at the study site of the authors of this abstract, CR/CRi correlated with CXCR4 expression >10% (Becker et al. ASH 2014).

The objectives of this preclinical study were to identify the patient population most likely to benefit from CXCR4 inhibition and explore the role of combined blockade of CXCR4 and the beta-2 adrenergic receptor. The combination of propranolol, a non-selective beta-blocker, and a new CXCR4 inhibitor, GPC-100 (burixafor), has been shown to augment mobilization of HSCs in a pre-clinical study (Sukhtankar et al. PLoS One 2023).

Methods: AML patient blood and BM samples (N=124) were obtained with informed consent on an IRB approved protocol. Multicolor flow cytometry was performed with sequential gating on blasts or leukemia stem cells (LSCs), then analysis for CXCR4 and beta-2 adrenergic receptor (ADRB2) expression. Clinical characteristics including age, gender, race/ethnicity, new diagnosis vs relapse, ELN2022 risk, cytogenetics, FLT3/NPM1 mutation status, CR1 duration, and overall survival were obtained. Statistical correlations were performed for CXCR4 expression (blasts vs. LSCs, % expression, MFI) and the clinical features. RNAseq was performed for patient samples with high, intermediate, or low expression of CXCR4. An in vitro high throughput drug screening viability assay was performed for primary AML cells on stroma. GraphPad Prism9 was used for statistical analysis, and Mann Whitney test (two groups) and ordinary one-way ANOVA (>2 groups) used for comparison.

Results: The range of CXCR4 expression by primary AML blasts is 0.05 to 99.7% (median 13.1%; mean 16.2%) (N=109) and by AML LSCs is 0% to 99% (median 2.9%, mean 15.2%) (N=35). The expression by blasts and LSCs for each patient exhibited tight correlation, R²=0.88 p<0.0001). There was higher % expression of CXCR4 by both blasts (p=0.003) and LSCs (p=0.002) in samples from newly diagnosed patients versus relapse. Patient samples positive for both FLT3-ITD and NPM1 exhibit higher CXCR4 % expression compared to single FLT3-ITD or NPM1 mutated or non-mutated. For a subset of 20 patient samples, analyses for both CXCR4 and ADRB2 expression were performed. ADRB2 expression ranged from 0.4 to 20.1%, and there was high expression (35.9 to 98.7%) of CXCR4 by the ADRB2 expressing subpopulation. RNAseq revealed upregulation of TP53, downregulation of CD58 (LFA-3) and RARA, and similar patterns of gene expression by heatmaps for high CXCR4 vs. low CXCR4 expressing samples as defined by flow cytometry. In addition, high throughput drug screening of AML on stroma, but not in suspension or on CXCL12 coated plates, revealed that combination of the CXCR4 inhibitor GPC-100 and beta blocker propranolol, with araC, increased drug sensitivity (reduced IC50 by ≥4 to >10 fold) as compared to araC alone for AML cells on HS-5 human stromal cell line or autologous patient mesenchymal stromal cells. Retrospective review of patient data from the clinical trial of ulocuplumab in AML (NCT01120457) revealed that 8/13 patients (76.9%) on non-selective beta blockers with ADRB2 blocker activity achieved CR/CRi as compared to 5/13 (35.7%) for those on beta1 selective blockers or 18/48 (37.5%) for those not on beta blockers, p 0.11 (ns) for the non-selective group vs. others.

Conclusions: These studies support further investigation of whether simultaneous blockade of CXCR4 and ADRB2 may potentiate chemotherapy response in AML, perhaps by disrupting microenvironment mediated chemotherapy protection. Patients with new diagnosis AML may be more susceptible to this approach than R/R AML due to higher CXCR4 expression by both blasts and LSCs.