Design, Optimization, and Pre-Clinical Evaluation of Direct, Mechanism-Based STAT3 Inhibitors for Treating Myeloid Disorders

We have identified STAT3 as a convergence point for oncogenic signaling in tyrosine kinase inhibitor (TKI)-resistant chronic myeloid leukemia (CML) lacking BCR-ABL1 kinase domain mutations. In addition, we found that STAT3 activity contributes to disease in other myeloid disorders, including acute myeloid leukemia (AML) and myeloproliferative neoplasms (MPNs). Utilizing TKI-resistant CML as a model system, we identified BP-5-087 as a small molecule inhibitor of STAT3 that reduces STAT3 phosphorylation and nuclear transactivation (Eiring et al. Leukemia, 2014).

Binding of BP-5-087 to the STAT3 SH2 domain was initially assessed using fluorescence polarization (FP) assays and high-resolution computational docking simulations. To further validate the binding motif of BP-5-087, we conducted time-resolved electrospray ionization mass spectrometry/hydrogen-deuterium exchange experiments. Fold-change in deuterium uptake was analyzed for 68 STAT3 peptides representing 71% sequence coverage, and mapped onto the crystal structure of STAT3. This analysis precisely defined the binding epitope for BP-5-087 within the STAT3 SH2 domain.

We next tested the effects of BP-5-087 in several myeloid malignancies using relevant disease models. (i) CML stem and progenitor cells from TKI-resistant patients without kinase domain mutations were treated with BP-5-087 ex vivo, using short-term liquid culture, clonogenic and LTC-IC assays. BP-5-087 treatment significantly reduced colony formation by CML stem and progenitor cells (p<0.01), with no effect on normal human CD34⁺ cord blood (CB) cells. (ii) Similarly, BP-5-087 also increased apoptosis and reduced viability (p<0.05) of primary AML blasts treated ex vivo with BP-5-087 for 72 hours in liquid culture. (iii) CD34⁺ cells from patients with myelofibrosis were also treated with BP-5-087 in clonogenic assays, and similar to CML, BP-5-087 reduced myeloid colony formation, although to a lesser extent.

The in vivo activity of BP-5-087 was next evaluated in a murine model of JAK2 V617F-induced MPN. Briefly, Balb/c bone marrow was transduced with JAK2 V617F-GFP, followed by injection into lethally irradiated recipients. After disease induction, mice were treated with BP-5-087 (25 mg/kg) by once-daily oral gavage. No toxicities were observed after 40 days of treatment in BP-5-087-treated mice. While BP-5-087 did not significantly reduce the percentage of GFP⁺ cells, there was a 41% reduction of spleen weight in BP-5-087-treated mice compared to vehicle-treated controls (p<0.05). Post study analysis revealed BP-5-087 plasma concentrations <1 μM, suggesting that insufficient bioavailability contributed to the modest in vivo effects.

To advance the lead optimization of our STAT3 inhibitor series, we instituted a comprehensive screening cascade. We first developed a computational model (quantitative structure-activity relationship, QSAR) to guide and prioritize selection of new inhibitor candidates for synthesis. Compounds are initially ranked using a methanethiosulfonate (MTS)-based cell viability assay in a TKI-resistant, STAT3-dependent CML cell line (AR230^R). Inhibition of STAT3 is confirmed using a cell-based STAT3 reporter assay and an in vitro FP-based binding assay. Optimization of potency is balanced by the goals of reducing molecular weight (MW) and calculated LogP (cLogP) compared to BP-5-087 (MW: 694.8; cLogP: 7.3). Compounds with improvements in these categories are then subject to toxicity testing utilizing clonogenic assays with CD34⁺ CB cells. Non-toxic compounds are evaluated for their pharmacokinetic profile in Balb/c mice and tested for activity in primary samples from CML, AML and MPN patients. These activities have directed us to a lead compound, AM-1-124, which displays significant improvements in potency, MW, cLogP, and in vivo half-life compared to BP-5-087. AM-1-124 had minimal effects in the CB toxicity assay and induced apoptosis in primary AML patient samples at 2-fold lower concentrations than BP-5-087. With AM-1-124 as our current lead compound, we are continuing our iterative evaluation of novel STAT3 inhibitors utilizing our screening cascade. Design and testing of optimized, orally active inhibitors will enable further evaluation of STAT3 as a target in animal models of myeloid leukemia and will justify the clinical development of these compounds for patients in need of new targeted therapies.