Leukemogenesis involves sequential acquisition of mutations in several genes. Various chromosome translocations occur early in leukemogenesis and may be initiating events, but are not sufficient to cause leukemia. The most common translocation in childhood ALL is a cryptic t(12;21) that produces a TEL-AML1 fusion protein. These and other translocations often occur in utero, with full-fledged leukemia developing years later following additional mutations in cells from the translocation-positive preleukemic clone.
Why do some childhood ALLs relapse? Several possibilities, not mutually exclusive, exist. Prevailing theories include outgrowth of cells that contain additional mutations and re-emergence of dormant cells from the original clone. It is also possible that the “original” leukemia was cured, but a “new” leukemia developed because new cooperating mutations occurred in the ancestral translocation-positive preleukemic clone.
In this issue, Konrad and colleagues (page 3635) examined this question by using antigen-receptor gene rearrangements andTEL deletions that occur in concert withTEL-AML1 fusion to track cells from different stages in leukemogenesis. They showed previously that some cases of late relapse are not re-emergence of the fully leukemic clone but are due to outgrowth of a sibling clone that shares the same ancestralTEL-AML1 fusion gene but contains different cooperatingTEL deletions. They now confirm and extend their earlier findings. Identical TEL-AML1 genomic fusion sequences in diagnostic and relapse specimens prove that they were derived from the same ancestral clone. But the presence of different antigen-receptor rearrangements showed that these were sibling clones that had developed along divergent pathways. The clone dominant at diagnosis was undetectable at relapse, suggesting that it had been cured. Remarkably, the clone dominant at relapse was present at very low (0.1-0.01%) levels at initial diagnosis. The relapse clones disappeared slowly at diagnosis but responded rapidly to therapy at relapse, demonstrating different biology at different stages of disease. Left unanswered is the critical question of whether the rare relapse clone cells present at initial diagnosis were fully leukemic at that time or were still in the preleukemic or almost leukemic phase.
These observations provide fascinating insights into leukemia biology and how childhood ALL is, or is not, cured. Important observations such as these always raise additional questions. For cure to occur, is it critical to eliminate every cell in the leukemic clone? What about cells from the ancestral preleukemic clone; must they be eradicated? It has never been clear why prolonged low-intensity maintenance chemotherapy is so important in childhood ALL. It isn't needed for Burkitt leukemia, a much more aggressive disease. Perhaps maintenance therapy treats the ancestral preleukemic clone. We know that chemotherapy treatment can cause leukemia. Many children with ALL who are cured today could also be cured with less therapy. Could this overtreatment increase the risk of relapse by exposing preleukemic cells to potentially mutagenic agents? Secondary mutations provide a tool to address these questions. We will learn more in the coming years; some of what we learn may surprise us.
