Disruption Of Ikaros In Patients’ Chronic Phase CML Cells Induces The Features Of Accelerated Phase Disease

Accelerated phase (AP) CML is marked by a rising leucocyte count in the face of maintenance treatment, exaggerated basophilia and accumulating blasts. However, AP is typically short-lived and rapidly succeeded by blast crisis. Genetic alterations resulting in loss of IKAROS activity are associated with the development of both BCR-ABL1+ myeloid and lymphoid leukemias, as well as the related BCR-ABL1-negative MPNs. However, how loss of IKAROS functionality contributes to the malignant features of these leukemias is not clear. Here we demonstrate that direct suppression of IKAROS activity in CD34+ cells from chronic phase (CP) CML patients leads to their rapid acquisition of features of AP without causing progression to acute leukemia.

To suppress IKAROS activity in CD34+ CP CML cells, we transduced samples from 4 patients with a lentiviral GFP vector encoding a dominant-negative isoform of IKAROS (dnIK6), or an empty control YFP vector, and compared their progeny generated in cultures containing mouse stromal cells engineered to produce human SCF, IL-3, FLT3L and G-CSF. In all cases, we found dnIK6 transduction caused a rapid (evident within 2 weeks) and sustained (for at least 6 weeks) increase in total cell output (4-20x higher than YFP+ controls at 4-6 weeks, p<0.01). More detailed analysis showed that both populations of transduced cells produced similar numbers of monocytes and neutrophils and the increased cell output from the dnIK6-transduced cells was largely due to a consistent 50-200x increase in basophil production (with occasional increased eosinophils also), quantified using stained cytospins, and confirmed by flow cytometry (presence of CD34-33+117-14-15-HLA-DR-25+9+ cells). Similarly, comparison by qRT-PCR of dnIK6 versus control nonadherent cells present after 2-4 weeks showed 10-50-fold lower MPO, 2-9-fold higher HDC and 20-65-fold higher PRG2 transcripts. Notably, the cloning efficiency of the initially transduced control and dnIK6-transduced CD34+ CML cells in both semi-solid media and suspension culture was not significantly different. However, the proportion of dying/dead mature (nonadherent) cells (as measured by propidium iodide staining) derived from control cells was ∼30x higher than from the dnIK6-transduced cells (4.9±1.0% vs 0.15±0.03%, p=0.002). This suggested a dnIK6-augmented survival of terminally differentiating cells causing an accumulation of their mature progeny. In addition, assessment of 6-week cultures revealed a strikingly higher number of dnIK6-derived CD34+ cells and colony-forming cells as compared to controls (49±15-fold and 19±6-fold higher, respectively).

To interrogate the molecular mechanisms underlying the dnIK6 effects on cell output, we compared the levels of activated signaling proteins by intracellular flow cytometry. This showed increased activation of STAT5 (80±14% higher median fluorescent intensity (MFI), p<0.001), MAPK (61±11% higher MFI, p=0.01) and β-catenin (44±9% higher MFI, p=0.02) in the CD34+ progeny of dnIK6-transduced CP CML cells. In addition, qRT-PCR analysis showed transcripts encoding key repressors of JAK-STAT (SOCS2, SH2B3) and β-catenin signaling (including GSK3B, AXIN1 and CK1) were suppressed in dnIK6-transduced CD34+ cells by comparison to CD34+ derivatives of control-transduced cells from the same patients. After a total of 12 weeks in culture, the control-transduced CML cells had largely disappeared, whereas the population produced by the dnIK6-transduced cells continued to expand, with a mean cumulative cell output after 12 weeks of 6x10⁶ cells from 10⁴ starting CD34⁺ cells (vs 2x10⁵ from controls) along with a marked increase in CD34+ cells, rising from 1-3% at week 6, to 12-60% at week 12. However, plating of these CD34+ cells in semi-solid media yielded basophil or monocyte but not blast colonies, demonstrating that differentiation was not strongly impaired.

In summary, IKAROS disruption in patients’ CP CD34+ CML cells recapitulates cardinal features of AP disease, including an increased cell output associated with an enhanced production of basophils and an eventual accumulation of primitive cells. We further identify dnIK6-mediated suppression of SOCS2 and the β-catenin destruction complex, with enhanced activation of STAT5 and β-catenin in progenitor cells, as a likely key mechanism contributing to the altered biology of AP cells.