Stroma-Based Activation of pSTAT3^Y705 Confers Resistance to FLT3 Inhibitors in FLT3 ITD-Positive AML

Acute myeloid leukemia (AML) is an aggressive hematopoietic neoplasm that carries the worst prognosis among the hematologic malignancies. Up to 30% of AML patients exhibit activating mutations in FLT3 tyrosine kinase. FLT3 internal tandem duplications (ITDs) comprise ~70% of these mutations and are associated with a poor prognosis. Most patients treated with a single-agent FLT3 tyrosine kinase inhibitor (TKI) relapse within months due to secondary mutations in the FLT3 tyrosine kinase domain (TKD). Results from trials of FLT3 TKIs in AML reveal that leukemic blasts are more easily cleared from peripheral blood than from bone marrow (BM), suggesting that the BM microenvironment promotes survival of AML cells, including leukemia initiating cells, despite inhibition of FLT3. In this conceptual framework, extrinsic factors allow AML cells to survive TKI exposure until AML cell-intrinsic resistance is conferred by FLT3 TKD mutations, leading to clinical relapse. Here, we investigated the role of the BM microenvironment in protection of FLT3⁺AML cells from treatment with AC220 (quizartinib), a clinically available FLT3 TKI.

To investigate the potential of the BM microenvironment to mediate TKI resistance in AML, we cultured FLT3-ITD⁺ AML cell lines, including MOLM-13, MOLM-14 and MV411, and the CML cell line, K562 (control; FLT3 wild-type), with graded concentrations of AC220 under the following conditions: (i) in regular medium (RM), (ii) in direct contact (DC) with human HS-5 BM stromal cells, or (iii) in HS-5 conditioned medium (CM). Cell proliferation and apoptosis assays revealed that, in RM,AC220 reduced proliferation and increased apoptosis of MOLM-13, MOLM-14 and MV411 cells, but had no effect on K562 cells. DC greatly reduced the effects of AC220 in all three FLT3-ITD⁺ AML cell lines, with comparable results observed between DC and CM. To confirm these data using primary cells, CD34⁺ blasts from a patient with newly diagnosed FLT3-ITD⁺ AML were similarly cultured in RM versus CM ± AC220. Consistent with results in cell lines, CM rescued primary AML cells from AC220-mediated cell death. These data indicate that soluble factors from the BM environment protect FLT3-ITD⁺ cells from the effects of FLT3 inhibition.

Our lab and others have demonstrated that HS-5 DC and CM activate STAT3 in chronic myeloid leukemia, which mediates resistance to BCR-ABL1 TKIs (Bewry et al. Mol Cancer Ther 2008, Traer et al. Leukemia 2012, Eiring et al. Leukemia 2015). To interrogate the role of STAT3 in BM-mediated protection of AML cells from FLT3 inhibition, all cell lines were assessed for pSTAT3^Y705 and total STAT3 by immunoblot analysis under each culture condition. In FLT3-ITD⁺ AML cells grown in RM, pSTAT3^Y705 was undetectable, irrespective of AC220 dose. In contrast, pSTAT5^Y694 was readily detected at steady state and suppressed by AC220. AML cells cultured in HS-5 DC or in HS-5 CM exhibited strong upregulation of pSTAT3^Y705 that was unaffected by AC220, suggesting that soluble factor(s) promote STAT3 activation in AML. pSTAT5^Y694, on the other hand, was slightly elevated by HS-5 DC or CM, but remained under control of FLT3 kinase activity. In order to mechanistically implicate STAT3 activation in stroma-based protection, we used a retroviral shRNA construct to knockdown STAT3 (shSTAT3) compared to an empty vector control (LMP) in MOLM-14 cells. STAT3 knockdown (~70%) was confirmed by qRT-PCR and immunoblot analyses. Cells containing shSTAT3 and LMP were cultured for 72 hours in RM or CM ± AC220, followed by analysis using MTS assays. As expected, CM increased the IC₅₀ of AC220 from 1.37 nM to 6.24 nM in LMP-expressing cells (n=3). In contrast, shSTAT3 reduced the IC₅₀ of AC220 from 6.24 nM to 2.87 nM (n=3) in CM, with minimal effects in RM. Similarly, pharmacologic inhibition of STAT3 using the novel STAT3 inhibitor, BP-5-087 (Eiring et al. Leukemia 2015), reduced the IC₅₀ of AC220 from 10.07 nM to 5.91 nM in CM. Analogous experiments in additional FLT3-ITD⁺cell lines and primary AML cells, using shSTAT3, dominant-negative STAT3 constructs and BP5-087 are ongoing.

Our data suggest that STAT3 is a critical signaling node in FLT3-independent TKI resistance mediated by the BM microenvironment. Therapeutic strategies designed to combine FLT3 and STAT3 inhibition may inhibit the survival of leukemic cells in the BM niche, thereby preventing subsequent clinical relapse conferred by TKD mutations.