Background: Targeting the stroma-AML cell interaction is a new dimension in anti-leukemia therapy. FLT3-ITD AML cells are known to derive protection from bone marrow stroma through up-regulation of resistance pathways such as MEK and JAK2. Pacritinib is a dual FLT3 and JAK2 inhibitor with equipotent activity for each target and is currently offered as stand-alone therapy for relapsed AML patients with FLT3-ITD mutant disease enrolled in the MRC AML 17 trial.

Aim: We set up this study to determine if pacritinib as a single agent is sufficient to abrogate the stroma mediated resistance of AML, and to investigate leukemic signalling in stroma-adherent AML cells.

Results: Pacritinib efficacy was assessed in 62 primary AML samples. FLT3-ITD samples were significantly more sensitive to pacritinib compared to FLT3 wild type AML samples after 48 hours exposure in vitro (mean IC50 92 nM vs 292 nM [range ITD 11-290 nM, range WT 19-2600 nm], p=0.01). Pacritinib induced apoptosis in AML cells but was not lethal to normal CD34+ bone marrow cells within the same dose range. FLT3-ITD AML cells cultured on MS5 stroma retained sensitivity to pacritinib after 7 days of co-culture (p<0.01, n=6).There was no significant effect on stromal cells. In this study, an MS5 stromal co-culture model was set up to enable interrogation of FLT3-ITD mutated primary AML cells that are adherent and non-adherent to stroma. PCR analyses performed on CD45+ adherent AML cells confirmed that they carried FLT3-ITD. Flow analyses demonstrated that the reduction of stroma-adherent AML cells was pacritinib dose dependent. Following 7 days of treatment and after drug wash-out, prior pacritinib exposure resulted in sustained suppression of leukemic outgrowth from stroma-adherent FLT3-ITD AML cells at 14 days of co-culture, compared to untreated cells (p<0.01, n=6). To establish if pacritinib could impact upon longer-term leukemic stromal adherence and survival in FLT3-ITD AML, we stained leukemic cells with Aldefluor and flow sorted the CD34+ 38- ALDH intermediate fraction, which has been shown to possess in vivo engrafting ability and to correlate with clinical relapse. Sorted cells were plated onto MS5 stroma and cobblestone area forming cell (CAFC) capacity determined as a function of pacritinib dose. Pacritinib reduced CAFC frequency at 5 weeks (from 500 nm, n=3).

To investigate leukemic cell signalling, FLT3-ITD mutant AML cells were cultured on stroma in the presence of pacritinib for up to 24 hours, and intracellular flow analyses performed on adherent and non-adherent fractions. Control stroma-free cultures demonstrated significant p-STAT5 knockdown from 1 hour of drug exposure. In the presence of stroma, non-adherent AML cells did not show this reduction in p-STAT5. P-STAT5 down-regulation in adherent AML cells was delayed (to 18 hours), but exhibited dose dependency (p<0.05 at 500 nm pacritinib, p=0.0005 at 1 μM; n=6). Interrogation of JAK2 signalling showed up-regulation of p-JAK2 in adherent AML cells in the presence of pacritinib from 3 hours in 2 of 5 samples. There was no abrogation of AKT or ERK signalling in stroma-adherent or non-adherent fractions and this was confirmed by immuno-blotting. Interestingly, pacritinib was observed to decrease non-phosphorylated beta catenin levels in stroma-adherent AML cells and this was not seen in non-adherent AML cells (p=0.029, n=5).

Short-term drug response Cell Glo luminescent assays reflected results from intracellular signalling experiments. Significant synergy was seen when pacritinib was combined with a MEK inhibitor (PD0325901) with accompanying down-regulation of ERK (mean CI 0.31 [range 0.19-0.44] at 5:1 ratio, n=20).

Conclusion: These results demonstrate that pacritinib can suppress leukemic outgrowth from stroma adherent cells in FLT3-ITD driven AML in medium-term (14 days) and long-term (5 weeks) assays. However, short-term intracellular signalling profiles suggest that resistance pathways can remain active in some cases and may require AKT/ERK targeting for their eradication. Good synergy is seen with MEK inhibition, making this an attractive combination for clinical evaluation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.