Glycomimetic Antagonist of E-Selectin, GMI-1271, Enhances Therapeutic Activity of the Hypomethylating Agent, 5-Azacitidine, in the KG1 Model of AML

Components of the bone marrow stroma afford a permissive environment for the malignant progression of AML. These stromal components include E-selectin, an adhesion molecule expressed in the bone marrow vasculature. Through the recognition and binding of E-selectin by specific ligands on AML blasts, survival pathways are activated within the blasts leading to the development of/increase in chemoresistance. GMI-1271 is a novel E-selectin antagonist that disrupts this process, down-regulating survival pathways and enhancing chemotherapy response with improved survival in mouse xenograft and syngeneic AML tumor models. GMI-1271 is currently in clinical trials for the treatment of AML in combination with intensive chemotherapy; the FDA has granted Breakthrough Therapy Designation to GMI-1271 for the treatment of adults with relapse/refractory AML. Azacitidine (5-AZC) is a DNA-methyltransferase inhibiting cytosine nucleoside analog that at low doses induces DNA hypomethylation and transcriptional activation, while at higher doses is directly cytotoxic to neoplastic cells including AML blasts. Vidaza (5-AZC, Celgene) is approved in Europe for the treatment of limited populations with AML. In these studies we evaluated the anti-tumor activity of GMI-1271 in combination with 5-AZC in the KG1 AML tumor model to identify if there was potential for therapeutic benefit of the combination.

Seven days post i.v. injection of KG1 cells, female NSG mice (n=10/group) were treated with saline; GMI-1271 (40 mg/kg IP QDx14) alone; 5-AZC (5 mg/kg IP Q3Dx5) alone, or the combination of GMI-1271 and 5-AZC. All treatments were well tolerated. The median survival time (MST) of mice treated with 5-AZC was 88 days and statistically different (P<0.005) to groups treated with saline (MST=69.5 days) or GMI-1271 alone (MST=69 days). At study conclusion (Day 110 post tumor injection), 20% of mice treated with 5-AZC remained alive; as expected all mice treated with saline or GMI-1271 alone succumbed to progressive tumor growth. Importantly, the therapeutic activity of 5-AZC was significantly enhanced when combined with GMI-1271 (MST>110 days, P=0.0217 compared to 5-AZC alone) with 70% of mice surviving to study conclusion. These results indicate that E-selectin/AML blast interaction in the KG1 model protects from the anti-tumor activity of 5-AZC and that GMI-1271 can attenuate this protection.

To investigate the nature of the observed in vivo activity of GM-1271 and 5-AZC in the KG1 model, in vitro assays of cytotoxicity and E-selectin ligand expression in KG1 and the KG1a variant were assessed. The IC50 of 5-AZC alone or in combination with a saturating dose of GMI-1271 on KG1 cells was 518 and 528 nM, respectively. The IC50 of 5-AZC alone or in combination with GMI-1271 on KG1a cells was 637 and 711 nM, respectively. These results imply that the enhanced anti-tumor activity observed in vivo was not due to a simple shift in the cytotoxicity profile of 5-AZC. KGI and KG1a cells were subsequently cultured for 96 h in a noncytotoxic concentration (100 nM) of 5-AZC and the reactivity of the cells to HECA-452, a surrogate marker of E-selectin ligand, or binding E-selectin was determined by flow cytometry. Treatment with 5-AZC resulted in an increased reactivity of KG1 and KG1a cells to HECA-452 (29 and 27%, respectively compared to nontreated cells) and an increased binding of E-selectin to KG1 cells (32% compared to nontreated cells). Collectively, these results demonstrate that noncytotoxic concentrations of 5-AZC lead to an increased expression of E-selectin ligands on KG1 and KG1a cells.

In summary, these studies show that even though the hypomethylating agent 5-AZC improves outcomes with AML, therapeutic potential could be improved. In this regard, 5-AZC at noncytotoxic concentrations promotes the increased expression of E-selectin ligands on AML blasts which may dampen 5-AZC activity. However, attenuation or disruption of E-selectin ligand/E-selectin interactions with GMI-1271 in combination with 5-AZC overcomes this tumor survival pathway with the potential to maximize therapeutic outcome. This work has relevance to patients with AML who do not achieve optimal or long-term response with 5-AZC and other hypomethylating agents. Clinical assessment of GMI-1271 with hypomethylating agents in AML is planned.