Relationship Between BCR/ABL Fusion Proteins and Leukemia Phenotype

To the Editor:

We read with interest the recent Editorial in Blood by Melo1 concerning the relationship between BCR/ABL variant transcripts and the leukemia phenotype determination. Regarding the chronic myeloid leukemia (CML), a strict correlation was suggested between at least three distinct clinico-hematologic entities and the type of BCR/ABL fusion protein produced by different BCR breakpoints (p210 CML, p190 CML, and p230 CNL). The last BCR/ABL rearrangement variant, with a c3a2 junction, was recently described in five patients with particular clinico-hematologic features proposed as neutrophilic-chronic myeloid leukemia (CML-N).2 

By revision of 10 cases from the literature, Melo1 emphasized a correlation between the BCR/ABL transcript b3a2 and the CML with thrombocythemic onset (or mimicking essential thrombocythemia [ET]). We wish to make some comments about this particular form of CML not extensively discussed in the editorial. Recently, by a careful revision of the literature, we identified 29 cases of CML with thrombocythemic onset3-6 and found 3 further cases (1 with b3a2, 1 with b2a2, and 1 with b3a2-b2a2) of 82 consecutive patients observed in our institution. Taken together, the data showed that BCR/ABL transcripts may be of b3a2, b2a2, or b3a2-b2a2 types, without a strong evidence of one specific transcript variant associated with the Ph⁺ ET phenotype. Although b3a2 is the most frequent among all variants detected in CML cases with thrombocythemic onset, it should be noted that this variant is also the most frequent variant in CML cases with classical features. Moreover, another fusion transcript has been identified in 2 cases of CML with thrombocythemic onset, designated c3a2, not mentioned by Melo.7,8 Interestingly, 3 of the 5 patients with c3a2 junction recently described as CML-N also showed a high platelet count at onset (1,020, 870, and 1,240 × 10⁹/L, respectively).2 

In short, we wish to emphasize that at least four different BCR/ABL junction variants (b3a2, b2a2, b3a2-b2a2, and c3a2) may be detected in thrombocythemic CML, although with various frequencies. We feel that the determination of a particular phenotype of CML, namely CML with thrombocythemic onset, might be related to additional genetic changes other than BCR/ABL fusion protein. We agree with Melo1 that much progress has been made in drawing “a map linking the DNA, RNA, and protein defects to the specific cells affected and to the clinical features in the patient” with CML. However, we think that much more caution is needed before assessing a close relationship between one specific molecular defect and a distinct clinico-hematologic manifestation of the disease, at least in CML patients with thrombocythemic features at onset. Further studies on larger series of cases are warranted to clarify whether c3a2 transcript variant is distinctive of a peculiar CML phenotype, ie, either CML-N or CML with thrombocythemic onset or both.