The ALL1 gene (also called MLL, HRX, or Htrx1) at the cytogenetic band 11q23 is consistently altered by chromosome rearrangements in acute leukemias (ALs) of early infancy, in ALs developed after exposure to topoisomerase (topo) II-inhibitory drugs, and in a small subset of de novo ALs in children and adults. Because exposure to natural or medicinal substances blocking topo II during pregnancy have been proposed as etiological agents for infant leukemia, we have compared the distribution of ALL1 gene breakpoints in infant leukemias with an altered ALL1 gene configuration to those in secondary leukemia associated with prior exposure to topo II targeting drugs and in reference to the major topo consensus binding site in exon 9. ALL1 gene breakpoint distribution was determined by Southern blot hybridization and/or reverse transcription-PCR of the ALL1/AF4 fusion cDNA in 70 patients. Using restriction enzyme analysis, the 8.3-kb ALL1 breakpoint cluster region was divided in a centromeric portion of 3.5 kb (region A) and telomeric portion of a 4.8 kb (region B). ALL1 breakpoint were located in region A in 8 of 28 (28.5%) cases of infant ALs, 16 of 24 (66%) cases of de novo ALs, and 0 of 5 cases of therapy-related (TR) ALs. Conversely, ALL1 breakpoints in region B were detected in 20 of 28 (71.5%) cases of infant AL, 8 of 24 (33%) cases of de novo AL, and 5 of 5 (100%) cases of TR AL (P = 0.002). These results were confirmed by direct sequencing of the ALL1/AF4 fusion transcript in 30 cases (19 infants and 11 child and adult de novo cases). The analysis of ALL1/AF4 junction types showed that children and adults with de novo leukemia had ALL1 breakpoints in intron 6 (9 cases) or intron 7 (2 cases), whereas breakpoints in infant cases were mainly located in intron 8 (14 cases) and less frequently in intron 6 (4 cases) and intron 7 (1 case). The difference in ALL1 breakpoint location between infant and noninfant AL patients with ALL1/AF4 fusion was statistically significant (P = 0.00005). These data demonstrated that infant and TR ALs share a similar biased clustering of ALL1 gene breakpoints, which supports the possibility that topo II inhibitors may also operate in utero and play a crucial role in the etiology of infant leukemia.