The Response of FLT3/ITD AML to FLT3 Inhibition: Apoptosis of Peripheral Blood Blasts and Differentiation of Bone Marrow Blasts

Abstract 943

FLT3/ITD mutations are present in approximately 23% of adult AML cases. FLT3/ITD therefore presents an attractive therapeutic target for tyrosine kinase inhibitors. To date, the clinical responses to FLT3 inhibitors have been primarily limited to clearance of peripheral blood blasts, while bone marrow blasts remain largely unaffected. AC220 (quizartinib), a novel FLT3 inhibitor with 10–50 times more potency in vivo than other FLT3 inhibitors, has had a high response rate in an exploratory subset of patients in an ongoing clinical trial for relapsed and refractory FLT3/ITD AML. A high proportion of FLT3/ITD AML patients treated with AC220 displayed the typical rapid clearance of peripheral blasts seen with other FLT3 inhibitors. Strikingly, however, (as reported at last year's ASH meeting), the bone marrow blasts underwent terminal differentiation in response to AC220 treatment over the course of 4–8 weeks, as established by the presence of the same FLT3/ITD mutation in mature circulating neutrophils isolated after treatment with AC220. This differentiation was frequently accompanied by a syndrome of fever, skin nodules resembling Sweet's syndrome, and pulmonary nodules. To study the mechanisms associated with this phenomenon, we developed an in vitro model of FLT3 inhibitor-induced differentiation. In order to more closely reproduce the bone marrow microenvironment in vitro, we prepared human stromal cell layers derived from healthy bone marrow donors and co-cultured them with Molm14 cells (FLT3/ITD cell line), as well as with primary AML blasts from FLT3/ITD AML patients in the presence and absence of AC220. In the absence of co-culture with stroma, both the Molm14 cell line and the primary samples underwent rapid apoptosis in the presence of increasing concentrations of AC220, similar to the response of peripheral blood blasts in AC220 -treated patients. However, when grown in co-culture with stroma, Molm14 cells underwent relatively rapid (24–48 hours) morphologic differentiation, as evidenced by the development of characteristic multi-lobed nuclei, and by reduction of nitroblue tetrazolium (NBT) after stimulation with endotoxin (respiratory burst activity). Similar data were observed in Molm14 cells treated with the FLT3 inhibitor sorafenib, indicating that this differentiation is likely a class effect of FLT3 inhibitors. We have also used our model system to study differentiation in primary AML patient blasts. Pre-treatment blasts were obtained from a patient enrolled on the AC220 trial (who exhibited terminal differentiation of marrow blasts in response to AC220), and differentiation was induced in vitro by co-culturing the blasts with stroma in the presence of AC220. After five days of drug exposure during stromal co-culture, the blasts remained viable (assessed by trypan blue exclusion), but morphologically began to resemble myelocytes, and demonstrated reduction of NBT in response to endotoxin. Flow cytometry analysis of these blasts after eight days of AC220 exposure demonstrated increased side scatter, expression of CD15, and loss of CD117 and CD34, all indicative of maturation. Blasts cultured in the absence of stroma displayed decreased P-FLT3, P-STAT5, and P-ERK in response to AC220. In contrast, blasts exposed to AC220 on stroma demonstrated continued activity of P-ERK despite loss of P-FLT3 and P-STAT5. CEBPalpha protein levels were significantly higher in blasts on stroma, but decreased rapidly over the first 48 hours of culture. These findings confirm that a FLT3/ITD mutation can be sufficient to induce a differentiation block, and, in the bone marrow microenvironment, FLT3 inhibition does not induce immediate apotosis, but rather differentiation. The continued activity of P-ERK (presumably by parallel signaling pathways activated by stromal cells) may be responsible for the survival of the blasts in the setting of FLT3 inhibition, with CEBPalpha inducing a differentiation program in this context. We conclude, therefore, that the response of a FLT3/ITD AML patient to potent FLT3 inhibition consists of rapid apoptosis of blasts in the peripheral blood, and, in most cases a slower process of terminal differentiation in bone marrow blasts. Furthermore, our in vitro model will allow us to explore the precise mechanisms whereby differentiation is blocked in FLT3/ITD AML, and how this can be overcome with FLT3 inhibition and other targeted therapies.