Rearrangements of the MLL Gene Are Influenced by DNA Secondary Structure, Potentially Mediated by Topoisomerase II Binding

Rearrangements of the MLL gene are found in approximately 75–80% of both infant acute leukemia (IAL) and therapy related myeloid leukemia (tAML), the latter linked to prior exposure to Topoisomerase II inhibitors. In order to study the mechanism whereby such rearrangements are induced, a fine mapping screen was performed using Inverse PCR to analyze a MLL location at the intron 11/exon 12 border that is known to be a hot spot for the induction of DNA breaks and rearrangements. Human lymphoblastoid TK6 cells were exposed to the pro-apoptotic agent anti-CD95 and estrogen as both the apoptotic process and birth control formulations have been associated with the development of MLL-linked leukemia. The latter association was determined as part of an epidemiological analysis of IAL. In addition, blood samples from patients receiving Topoisomerase II inhibitor therapy, subsequent to a diagnosis of breast cancer or lymphoma, were examined in the same manner as such drugs are strongly associated with the etiology of tAML. From an initial screen of cells treated in-vitro, 13/27 of all rearrangement break points (translocations/insertions) were located within a 3 bp tract at the 5′ edge of a 10 bp palindrome that defined a potential 101 bp stem-loop with its 3′ palindrome partner. Analysis of blood samples taken up to a year after treatment initiation from lymphoma and breast cancer patients, showed a similar distribution of translocations with 10/20 events restricted to the same location on the 5′ of the palindrome. Further examination of the putative stem-loop structure showed a high stringency Topoisomerase II consensus sequence binding site at the geometric midpoint of the proposed stem-loop. In order to link the rearrangements observed with DNA cleavage events, TK6 cells exposed to anti-CD95 antibody were also screened for DNA breaks within the same region of MLL. Here, LM-PCR products spanning the region of interest were extracted from DNA gels and the location of cleavage determined by cloning and sequencing. From four to twenty four hours after anti-CD95 exposure, analysis of the breaks induced showed 24/37 (65%) were located at the base of the proposed stem loop associated with the DNA palindrome. 14/37 (38%) of breaks were found within a 40 bp tract at the 5′ side and 10/37 (27%) within a 20 bp tract at the 3′ side of the palindrome base. In each case, the hot spots for cleavage identified by LM-PCR included the palindromic sequences. Thus, unlike the 5′ restriction of MLL rearrangements, DNA fragmentation occurred at both sides of the proposed stem loop base, indirectly supporting the creation of such a structure in-vivo that is subject to local attack at the stem base. Palindrome association may be driven by the bending Topoisomerase II exerts when bound at the midpoint between each palindrome, such that each half of the proposed palindrome is brought into physical contact. This effect would be accentuated by poisoning of Topoisomerase II, subsequent to binding of drugs such as etoposide that stabilize the cleavable complex. We propose a model for a sub-group of MLL rearrangements that utilize Topoisomerase II mediated stabilization, either natively or subsequent to chemical poisoning, to assist in the creation of stable DNA structures that are permissive for DNA fragmentation. Further, the conversion of DNA double strand breaks at these sites to detectable rearrangements may be influenced by 5′ to 3′ DNA processing functions, such as transcription or replication.