Abstract
Abstract 4427
Imatinib and other tyrosine kinase inhibitors (TKIs) are the standard first-line treatment for patients with chronic myeloid leukemia (CML) and have dramatically improved disease outcome. However, a proportion of patients fails to achieve optimal molecular response (major molecular response, MMR, or better) to TKI treatment and has a marked residual disease even after years of therapy start. The reasons for suboptimal imatinib response are unclear and these patients might benefit from treatment with more potent and broad spectrum 2nd generation TKIs (dasatinib and nilotinib) already from the diagnosis to prevent disease progression. By using a phosphoproteomic approach, we aimed to identify predictive markers for imatinib failure or suboptimal response, which could be used to guide treatment selection already at the time of diagnosis.
The study consisted of 9 CML patients in chronic phase with optimal and 10 patients with suboptimal response to imatinib. Patients with suboptimal response had complete cytogenetic response, but no MMR after 18 months of therapy, whereas all patients with optimal response were at least in MMR at 18 months. 7 of 19 patients belonged to intermediate Sokal risk group (5/10 in suboptimal and 2/9 in optimal group) and 12 patients to low Sokal risk group. One additional failure patient without any cytogenetic response to any of the available TKIs was also analyzed (patient had M351T imatinib resistant mutation). Bone marrow mononuclear cells (BM MNC) were collected at the time of diagnosis prior to any TKI therapy and were analyzed for 46 kinase phosphorylation sites central in the pathogenesis of CML (R&D Human Phospho-Kinase Array Kit). BM MNCs from 3 healthy controls and the K562 Ph+ cell line were used as controls. Quantitative analyses of the phosphoproteins were performed by measurement of array spot densities with CellProfiler2.0.
In K562 cell line, the protein phosphorylation was most prominent in pSTAT5b Y699 and only modest in pSTAT5a Y694. In diagnostic BM samples, pSTAT5b and pSTAT5a levels showed marked variability within individual patients. Src Y419 (P<0.01), c-Jun S63 (P<0.01), Yes Y426 (P<0.05), STAT3 Y705 (P<0.05), and AMPKα2 T172 (P<0.05) phosphorylation levels were significantly higher in CML patients compared to healthy controls.
Patient with the M351T mutation without any cytogenetic response to any TKI revealed highly active state of HSP27 S78/S82, which was not detected in any other patient or in healthy control, revealing a candidate therapy target for this individual patient
When median phosphorylation ratios of suboptimal vs. optimal patients’ diagnostic samples were compared, STAT5b Y699 (ratio 2.3), PLCγ-1 Y783 (ratio 1.9), STAT5a Y694 (ratio 1.6) and Pyk2 Y402 (ratio 1.6) levels were increased in suboptimal group. Patients in intermediate Sokal risk group had significantly higher phosphorylation levels of ERK1/2 T202/Y204 T185/Y187 (P<0.01), CREB S133 (P<0.05), Akt S473 (P<0.05), JNK T183/Y185 T221/Y223 (P<0.05) and Lck Y394 (P<0.05).
In conclusion, chronic phase CML patients had a distinct phosphoprotein pattern in BM, which differed from healthy controls and from the K562 cell line. Our phospohoproteomic approach revealed the importance of novel signaling pathways not previously discovered in CML such as the anti-apoptic HSP27 pathway. Furthermore, phosphoprotein pattern of suboptimally responding patients differed already at the time of diagnosis from optimally responding patients. In particular, high levels of phosphorylated STAT5b may herald a suboptimal imatinib response and warrant choosing a 2nd generation TKI as the first-line of therapy.
Porkka:Bristol-Myers Squibb: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. Mustjoki:Bristol-Myers Squibb: Honoraria; Novartis: Honoraria.
Author notes
Asterisk with author names denotes non-ASH members.