BCR‐ABL1 tyrosine kinase activity at diagnosis, as determined via the pCrkL/CrkL ratio, is predictive of clinical outcome in chronic myeloid leukaemia

Imatinib has become first-line therapy in chronic myeloid leukaemia (CML), but recent studies suggest that more than one-third of patients fail imatinib over 2–3 years. (de Lavallade et al, 2008; Lucas et al, 2008) It would be clinically helpful to prospectively identify patients unlikely to achieve a complete cytogenetic response (CCRe) following imatinib treatment. Prognostic scoring methods, such as the Sokal score, activity of the imatinib uptake transporter, hOCT1 (human organic cation transporter 1) (Wang et al, 2008) and BCR-ABL1 transcript type (Lucas et al, 2009) have been shown to correlate with clinical outcome. However, none of these parameters are sufficiently powerful to reliably prospectively predict patients destined to fare poorly. The phosphorylation status of CrkL has been identified as a surrogate marker of BCR-ABL1 tyrosine kinase activity (White et al, 2005), because BCR-ABL1 tyrosine kinase activity cannot be reliability detected in CML patient samples (Patel et al, 2007). White et al (2005) showed, by Western blotting, that changes in the level of CrkL phosphorylation after 1 month of imatinib treatment correlated with clinical outcome. Flow cytometry has the advantage that it requires fewer patient cells (∼103 cells) than Western blotting, produces quantitative values, and results can be obtained rapidly. Additionally this methodology could be utilized more easily at other laboratories. A representative FACS histogram of a patient sample is shown in Fig 1. Patient characteristics and optimization of pCrKL and CrKL detection by flow cytometry. (A) Summary of patients characteristics. (B) Representative FACS plot showing pCrKL and CrKL (C) in a newly diagnosed CML patient and changes following in vitro treatment with 5 μmol/l imatinib for 24 h. Our study demonstrated that measuring pCrkL levels alone did not correlate with clinical outcome (Fig 2A). Hundred per cent of patients with a pre-treatment pCrkL/CrkL ratio <25 achieved CCRe following 12 months of imatinib treatment (Fig 2B). Of these, 75% of patients achieved their CCRe within 6 months. No patient whose pCrkL/CrkL ratio was greater than 25 achieved CCRe following 12 months of imatinib treatment (P = 0·0003). pCrKL, CrKL and pCrKL/CrKL ratio level at diagnosis and clinical outcome. FACS assessment of pCrKL (A) and pCrKL/CrKL ratio (B) in untreated newly diagnosed CML patients, stratified by clinical outcome following 12 months of imatinib treatment. (C) pCrKL/CrKL ratio at diagnosis and clinical outcome at 12 months for imatinib- and nilotinib-treated patients. Patients with a pCrKL/CrKL ratio below 25 responded to imatinib treatment. Six out of 14 imatinib-treated patients failed to achieve a CCRe following imatinib treatment. Three switched to the second generation tyrosine kinase inhibitor nilotinib and the other three remain on imatinib. No BCR-ABL1 kinase domain mutations were detected in patients who failed to achieve a CCRe by 12 months. The remaining eight patients, with a diagnostic pCrkL/CrkL ratio below 25 and a maximum follow up of 2 years, remain in CCRe. During this study six patients commenced nilotinib treatment at initial diagnosis. Their pCrkL/CrkL MFI ratios at diagnosis ranged from 7·5 to 47·3 Fig 2C. Following 12 months of treatment all achieved a complete cytogenetic response. Patients with a pCrkL/CrkL MFI ratio greater than 25 who were treated with nilotinib achieved a CCRe compared to 0% of imatinib-treated patients (P = 0·03). This is probably because nilotinib is more potent BCR-ABL1 tyrosine kinase inhibitor (O’Hare et al, 2005). These data suggest that patients with high BCR-ABL1 tyrosine kinase activity (pCrkL/CrkL MFI ratio >25) have a low probability of responding to imatinib, but may nevertheless achieve a complete cytogenetic response on nilotinib treatment. All nilotinib-treated patients remained in CCRe with continued treatment. In addition, serial monitoring of the trends in the pCrkL/CrkL ratio following commencement of imatinib treatment had no predictive value. We analysed the diagnostic pCrkL/CrkL ratio data for both imatinib- and nilotinib-treated patients in relation to other clinical parameters such as white blood cell count at diagnosis, Sokal score and hOCT1 (for imatinib-treated patients only), (Davies et al, 2009) and found no correlation. This suggests that the pCrkL/CrkL ratio is an independent biomarker of clinical outcome for imatinib treated patients. It is interesting to speculate that the response to imatinib therapy is related to BCR-ABL1 tyrosine kinase activity present at diagnosis. A fixed dose of imatinib may therefore only be able to suppress the lower kinase activity i.e. pCrkL/CrkL ratio <25; while nilotinib, which is known to be a more potent BCR-ABL1 kinase inhibitor than imatinib (Kantarjian et al, 2007), may suppress the BCR-ABL1 tyrosine kinase activity to a greater proportional extent. This may explain the higher incidence of CCRe following treatment with nilotinib. In conclusion, the pCrkL/CrkL ratio prior to commencement of treatment is a potent predictor of clinical outcome in patients treated with imatinib. Its role as a predictive biomarker of response to tyrosine kinase inhibitor therapy merits further investigation in larger series, though it is essential that this is performed on freshly isolated cells prior to any treatment. The authors would like to thank the University of Liverpool Biobank staff Joanna Middleton, Melanie Oates and Alison Hayes for their help and support during this project. CML, RJH and REC designed the study and wrote the manuscript. CML and AG performed the laboratory work for this study. KK and SW were responsible for taking patient samples and providing clinical information. REC was the principal investigator. The authors reported no potential conflicts of interest.