The role of imatinib (IMA) intracellular transport in overall resistance and treatment outcome prediction has been intensively studied since the first description of the main influx (hOCT-1) and efflux (ABCB1) transport proteins [1]. Several studies that correlate the pretreatment activity/mRNA expression of IMA transporters with subsequent therapy response have been reported to date [2–5]. However, different cell populations from patients who received various degrees of pretreatment were used for these analyses. Recent data have shown that the composition of the analyzed material at the time of testing (i.e., percentage of different cell types) has a critical impact on the resultant activity/mRNA expression of IMA transporters, and thus on the overall correlation and data interpretation [6–9]. Therefore, it is not surprising that the predictive value of pretreatment transporter characteristics remains controversial.

In this study, we investigated the predictive value of pretreatment mRNA expression levels of hOCT-1 and ABCB1 in different cell populations with regard to the response to therapy at 6 and 12 months of IMA therapy. Expression levels were assessed in peripheral blood (PB) leukocytes (LEU, n = 30), polymorphonuclear cells (PMNC, n = 23), and mononuclear cells (MNC, n = 21) of PB LEU obtained from 30 patients (Table 1) with de novo chronic myelogenous leukemia (CML). Moreover, the available bone marrow cells (BM) were also included and analyzed (BM, n = 11). The PB and BM samples were obtained, processed, and analyzed as previously described [6]. Responses to therapy were classified according the European LeukemiaNet 2009 (ELN) criteria [10]: responders show an optimal response at 6 months and 12 months, while non-responders reflect a suboptimal response or therapy failure. The assessed pretreatment expression levels were stratified into two groups according to the median—a low-mRNA expression group below the median and a high-mRNA expression group equal to or above the median. The statistical evaluation of the data obtained was performed using the Fisher’s exact tests and summarized in Table 2.

Table 1 Patient characteristics: white blood cell (WBC) count (109/L); age at diagnosis (years); Sex M (male), F (female); initial IMA dosage (mg/day), optimized IMA dosage (mg/day); Sokal score; Hasford score; therapy responses at 6 and 12 months of IMA therapy according the ELN 2009 criteria [10]
Full size table
Table 2 Pretreatment mRNA expression levels of hOCT-1 and ABCB1 in de novo CML patients and therapy responses at 6 and 12 months of IMA therapy according the ELN 2009 criteria [10]
Full size table

1 hOCT-1

The results obtained clearly indicate that there was no statistically significant relationship between pretreatment mRNA expression levels assessed in any selected cell populations and response to therapy at 6 and 12 months (all with p > 0.05). Moreover, at both monitored time points, there were a substantial proportion of patients who exhibited an optimal response regardless of low pretreatment hOCT-1 mRNA expression, which suggests that the therapy response at these time points is not directly related to these values. Similar results were obtained when non-responders were evaluated.

2 ABCB1

An analogous evaluation was performed for pretreatment ABCB1 mRNA expression levels. Similarly, no statistically significant relationship between the pretreatment ABCB1 values and therapy response was observed (Table 2).

In summary, our findings show no statistically significant relationship between the pretreatment levels of hOCT-1 or ABCB1 mRNA expression in different cell populations and therapy response at 6 and 12 months of IMA therapy. Despite the short observational period and small cohort of patients, our data indicate that there are a substantial proportion of patients who exhibit an optimal response regardless of their low pretreatment hOCT-1 mRNA expression or high ABCB1 expression, and vice versa (i.e., suboptimal response/therapy failure regardless of high hOCT-1 or low ABCB1 expression). Similar results were reported in a larger study, where substantial frequencies of optimal responses were also associated with low hOCT-1 pretreatment values [5], indicating some analogy between the assessment of mRNA expression and functional activity of these transport proteins and their predictive values. Regardless of the fact that some correlation and similarities between these two independent approaches have been reported [2, 6, 9], it must be noted here that this does not necessarily indicate that there is a correlation between the predictive values of expression of transport proteins and their activity.

In conclusion, our data focused on the predictive value of mRNA expression of transport proteins, clearly indicate that assessment of pretreatment hOCT-1 and ABCB1 mRNA expression is probably not suitable for implementation in routine clinical practice. The effect of individual cell types, tumor burden dependence, and other pre-analytical and analytical biases [6, 7] lead to the conclusion that the desired stratification of CML patients (using pretreatment hOCT-1 and ABCB1 mRNA levels) into responders/non-responders prior to IMA therapy is difficult to obtain. Therefore, we emphasize that complex predictors (i.e., Sokal or Euro score), even in the era of tyrosine kinase inhibitors, are still better tools for predicting treatment outcome than pretreatment transporter mRNA level.