Chromosome band 11q23 is involved in reciprocal chromosome translocations in 5–10% of children and adults with acute lymphoblastic leukaemia (ALL) or acute myelogeneous leukaemia (AML). Nearly all these rearrangements affect the ALL-1 gene, also termed MLL1, HRX and HTRX, located at 11q23. The translocations occur within an 8.3kb breakpoint cluster region of the gene and result in the replacement of ALL-1-coding sequence 3' to the breakpoint by the coding sequence of the translocation partner (Gu et al, 1992; Tkachuk et al, 1992). In a second and less frequent type of rearrangement, ALL-1 undergoes partial tandem duplication (self-fusion), resulting in the production of a larger protein (Schichman et al, 1995). The most common ALL-1 translocations are t(4; 11), t(9; 11), t(11; 19), t(6; 11) and t(10; 11), accounting for 40, 27, 12, 5 and 5% of cases, respectively. However, the total number of different loci participating in ALL-1 translocations is amazingly high and has already exceeded 50 (Huret et al, 2001).
There is an association between particular chromosomal translocations involving ALL-1 and subtypes of leukaemias. Thus, t(4; 11) nearly always occurs in ALL, t(9; 11) is mostly associated with AML, as are t(11; 19)(q23;p13.1), t(6;11), t(10;11) and the partial duplications. t(11; 19)(q23;p13.3) occurs in both AML and ALL. ALL1-associated ALLs are characterised as CD10−, CD19+ and by B-cell precursor phenotype (pro B), with a high frequency of myeloid-associated (CD15 and/or CD65) antigen expression. Hence, they are also designated biphenotypic or mixed-lineage leukaemia. The AMLs are usually of the myelomonocytic (FAB-M4) or monoblastic (FAB-M5) morphological subtypes and sometimes express cell-surface lymphoid markers. The prognosis of patients with 11q23 abnormalities is dismal. Age is an important prognostic factor. Thus, less than 25% of infants (<1 year) and adults older than 40 or 50 years, with t(4; 11) or t(9; 11), are cured. Patients of intermediate age, in particular 2–9 or 1–14 years old, have significantly better prognosis (Johansson et al, 1998; Swansbury et al, 1998; Pui et al, 2002).
The hallmark of ALL1-associated leukaemias is their epidemiology (reviewed in Greaves, 1999; Biondi et al, 2000). They predominate infant acute leukaemia (80% of ALL and 65% of AML), and account for the majority of therapy-related (secondary) leukaemias developing in 5–15% of primary cancer patients who had received intensive chemotherapy, including epipodophyllotoxins (VP16) and anthracyclines (doxorubicin) (Pui and Relling, 2000). These drugs are inhibitors of topoisomerase II, and act by stabilising double-strand DNA breaks generated by this enzyme. In both infant and secondary leukaemias, the latency period is remarkably brief. In infants, ALL-1 rearrangements can be already detected in utero and the average age at diagnosis is 6 months, and in secondary leukaemias, latency averages around 18–30 months (reviewed in Greaves, 1999; Pui and Relling, 2000). The very brief time between the initiating event and clinical disease, in conjunction with the high rate of concordance in identical twins with infant leukaemia, points to the effectiveness of ALL-1 fusion proteins in driving the initial clone into frank leukaemia. Based on the findings from therapy-related leukaemia, it was hypothesised (Ross et al, 1994) that transplacental exposure of the foetus to natural substances that inhibit topoisomerase II (such as flavonoids in foods and drinks) might play a role in infant leukaemia. Such exposure will induce accumulation of DNA's free ends from within the breakpoint cluster region of ALL-1, and this would eventually result in 11q23 translocations. The results of recent biochemical experiments (Strick et al, 2000) are consistent with this hypothesis. A structural feature shared by many topoisomerase II inhibitors, including anticancer drugs, is the quinone moiety. The metabolism of quinones is controlled by the enzyme NQO1 that converts toxic benzoquinones to hydroxyquinones. A polymorphism in nucleotide 609 of the enzyme substitutes a proline into serine and consequently inactivates the enzyme. Comparison of paediatric leukaemias with and without ALL-1 rearrangement indicated significant bias in the former for NQO1 genotypes conferring low or no enzymatic activity (Wiemels et al, 1999; Smith et al, 2002), suggesting that inactivating NQO1 polymorphism increases the risk for ALL-1-associated leukaemias.