Nonrandom chromosomal translocations are a common finding in the malignant cells of patients with acute and chronic leukemia. In general, these translocations lead to 1 of 2 molecular consequences. One class of translocation leads to the production of a fusion gene, such as the BCR-ABL oncoprotein associated with chronic myelogenous leukemia (CML); a second class of translocations leads to the unscheduled expression of a “proto-oncogene,” which is often a transcription factor involved in normal developmental regulation. This unscheduled expression of proto-oncogenes is most often caused by translocations that juxtapose regulatory regions of antigen receptor genes (immunoglobulin or T-cell receptor genes) and commonly involves transcription factors such as MYC, LMO2, or SCL (TAL1). Given that this second class of translocation involves antigen receptor genes, it is not surprising that translocations that activate proto-oncogenes are most commonly seen in lymphoid rather than myeloid leukemias.
However, the same genes that are activated via chromosomal translocation can also be activated via unknown mechanisms. This is true for both human and mouse T-cell acute lymphoblastic leukemia (T-ALL).1,2 For instance, gene profiling studies have indicated that SCL and LMO1 or LMO2 are expressed in almost 50% of T-ALL patients. Furthermore, MSH2-deficient mice, which frequently develop T-ALL, activate both SCL and LMO2. A “smoking gun,” in the form of a chromosomal rearrangement that juxtaposes an antigen receptor gene regulatory region with the proto-oncogene, can be identified in some of the T-ALL patients. However, in many patients—perhaps more than half—no such rearrangement can be found. There are at least 2 models that can be developed to explain this finding. The regulatory region of the proto-oncogene may have undergone a subtle mutation, leading to unscheduled, inappropriate activation of the proto-oncogene. Alternatively, a transcription factor that controls expression of the proto-oncogene may be inappropriately activated, leading to unscheduled expression of the proto-oncogene. These 2 models lead to a testable hypothesis. In the first case, expression of the proto-oncogene should be monoallelic; in the second case, expression of the proto-oncogene should be biallelic.
In this issue, Ferrando and colleagues (page 1909) report their study of allele-specific expression of SCL, LMO2, and HOX11. As anticipated, in each case where there was a smoking gun, in the form of a t(10;14) translocation for HOX11 or a SIL-SCL fusion (Tal1d) for SCL, expression of the proto-oncogene was monoallelic. However, about half of the patients who had activated SCL or LMO2 showed biallelic expression, suggesting that inappropriate activation of an upstream transcription regulatory molecule could be the cause of proto-oncogene activation. Moreover, they demonstrated that SCL and LMO2 were normally expressed in the most primitive double-negative (DN1, DN2, and DN3) thymocytes and down-regulated to undetectable levels as thymocytes matured to the double-positive (DP) and single-positive (SP) stages. Therefore, it is feasible that aberrant expression of these genes is caused by a lack of normal down-regulation in the maturing thymocytes, perhaps due to the absence of a normal silencing molecule or the inappropriate persistence of a positive transcription regulator. Taken together, these findings suggest that genes critical for the development of T-ALL (“genes of interest”) might lie upstream of SCL or LMO2 in a transcriptional cascade.
