Abstract

The ability to predict which CML patients respond poorly to imatinib may facilitate a more aggressive approach to therapy in these patients. We have used western blotting to assess the phosphotyrosine content of the adaptor protein Crkl (p-Crkl) -a major substrate for BCR-ABL. PB was collected from 57 newly diagnosed CML patients prior to the initiation of therapy and dose dependent inhibition of Crkl phosphorylation, in the presence of imatinib, was measured following incubation of primary cells for 2 hours with varying doses of imatinib. Using this assay we were able to determine that the median IC50 for imatinib was 0.6 uM (R 0.375–1.8uM). For statistical analysis IC50’s were then grouped into low and high relative to the median, and compared to molecular response achieved at both 3 and 12 months. Using Log Rank Survival Analysis there was a significant difference in the probability of achieving a 2 log depletion of BCR-ABL (measured by quantitative real-time Q-PCR) by 3 months between the two groups (Low IC50 30% n=31, High IC50 4% n=26; p=0.023 n=57). In addition the probability of achieving a 3 log reduction of BCR-ABL at 12 months (MMR, major molecular response) was significantly different (Low IC50 47%, High IC50 23%, p=0.034 ). We therefore assessed whether the IC50 was a surrogate measure of the Sokal Score or was an independent predictor of response. When we examined the median Sokal scores of both the low and high IC50 groups we found that there was no significant difference between the two groups (Low IC50 = 1.0, High IC50 = 0.9 p=>0.05 ) suggesting that IC50 was not a surrogate measure of Sokal. Log Rank survival analysis revealed that while Sokal Score did not predict for a 2 log depletion at 3 months (Low (43% n=32) vs Intermediate (21%n=25) vs High (21% n=29) p>0.05 ) it was predictive of a greater probability of MMR at 12 months (Low 53% vs Intermediate 42% vs High 21% p=0.044 ). The combined effect of these parameters was assessed:

Probability of achieving a Major Molecular Response.

AllLow SokalIntermediate SokalHigh Sokal
Note: Not all Sokal Scores available at the time of analysis. 
Low IC50 47% n=30 67% n=9 50% n=6 17% n=12 
High IC50 23% n=23 20% n=10 22% n=9 0% n=4 
p-value 0.034 0.037 >0.05 >0.05 
AllLow SokalIntermediate SokalHigh Sokal
Note: Not all Sokal Scores available at the time of analysis. 
Low IC50 47% n=30 67% n=9 50% n=6 17% n=12 
High IC50 23% n=23 20% n=10 22% n=9 0% n=4 
p-value 0.034 0.037 >0.05 >0.05 

Evidence presented here suggests that the IC50 level measured in blood cells from CML patients prior to imatinib therapy provides a statistically significant indicator of response at both early and late timepoints. Further to this, the combined measures of Sokal Score and IC50 provide a powerful tool for prediction of molecular response, which clearly delineates those patients groups who have a low probability of achieving a major molecular response. Preliminary investigations into the components of both Sokal and IC50 which may be contributing to these findings have highlighted a group of patients (n=7) who have a high blast percentage (5 to 21%) and low IC50. Unlike the majority of patients with low IC50 these patients all fail to achieve MMR. When these patients are removed from the low IC50 group the probability of MMR in the remaining patients is 71%. Further investigations into underlying mechansims for the inter-patient variation in IC50 are currently being undertaken.

Author notes

Corresponding author