USE OF BCR/ABL GENOMIC BREAKPOINT FOR MRD MONITORING IN CHILDHOOD ALL AND ITS COMPARISON WITH OTHER STANDARD METHODS

ABSSUB-5210

Background: Among childhood ALL subtypes, patients harbouring BCR/ABL fusion belong to those with the worst prognosis. The role of minimal residual disease (MRD) in this subgroup is unclear; however, MRD is used in HSCT indication according to the worldwide protocol for the BCR/ABL+ ALL - EsPhALL. In general, the standard method for MRD monitoring in ALL is quantification of Ig/TCR rearrangements. Quantification of the fusion transcript is easier/cheaper technique; however, as we showed previously, the correlation between these two methods in some BCR/ABL+ patients is not satisfactory. We demonstrated presence of BCR/ABL within the non-lymphoid (Ig/TCR-negative) population and worse prognosis of such patients (Zaliova, 2009). Moreover, mRNA quantification might not always correspond to the number of positive cells (possible changes of expression under particular conditions or in various cell types).

Aims: We aimed to establish MRD monitoring based on the detection of the genomic (intronic) BCR/ABL fusion and the BCR/ABL DNA quantification.

Methods: We performed multiplex long distance DNA PCR to find genomic BCR/ABL breakpoint in 15 patients with minor BCR/ABL fusion. The patient-specific fusion sequence was used for MRD quantification by qPCR. Using this approach we measured 149 bone marrow (BM) and 64 peripheral blood (PB) samples from 12 patients. The results were compared with data from Ig/TCR and BCR/ABL transcript quantification. For correlation analysis double negative results were excluded.

Results: We found genomic BCR/ABL breakpoint in 14/15 (93%) patients. Analysis of MRD data in BM samples confirmed poor correlation between Ig/TCR and BCR/ABL mRNA quantification (R²=0.72; n=89). Moreover, we saw similar difference also when comparing the two DNA methods (Ig/TCR vs. BCR/ABL DNA, R²=0.68; n=92). The correlation between BCR/ABL mRNA and BCR/ABL DNA approach was more satisfactory (R²=0.83; n=98).  Correlation in PB samples was higher (R²=0.79, 0.80 and 0.89, respectively).

Summary/Conclusion: We believe that the BCR/ABL quantification at genomic level brings the most precise picture of leukaemic burden. Our analysis confirmed poor correlation between BCR/ABL and Ig/TCR data. As we showed previously, at least in some patients this discrepancy is caused by the presence of BCR/ABL in non-lymphoid cells.

In our cohort, correlation of the two BCR/ABL based methods was in general good. However, in some patients (4/12) we observed higher genomic BCR/ABL MRD compared to the transcript levels in consecutive samples suggesting either that treatment (possibly by imatinib) can in some patients influence the BCR/ABL expression or that BCR/ABL+ cells might differ in their responsiveness to treatment (cells with low BCR/ABL expression being more resistant). This might be important from the clinical point of view as these data suggest that not only Ig/TCR monitoring, but also BCR/ABL transcript quantification might underestimate the real MRD load.

We found better correlation of all methods in PB reflecting probably more uniform cell population released from BM. However, in significant number of paired samples the MRD levels were more than 1 log lower in PB compared to BM (48%, 38% and 27% using BCR/ABL mRNA, Ig/TCR and BCR/ABL DNA quantification, respectively) questioning the use of PB for standard MRD monitoring.

The most reliable method for useful and reasonable MRD detection in BCR/ABL+ ALL needs to be discussed within large therapeutic protocol and, importantly, with respect to clinical data.

Support: MH CZ–DRO Univ. Hospital Motol 00064203; UNCE 204012.

Keywords: ALL, BCR-ABL, MRD

http://www.ehaweb.org/