Nucleolytic Cleavage of the Mixed Lineage Leukemia Breakpoint Cluster Region during Apoptosis*

VP-16 (etoposide) has recently been shown to induce topoisomerase II (TOP2)-mediated DNA cleavage within the mixed lineage leukemia (MLL) breakpoint cluster region (bcr), suggesting a role of TOP2 in MLL gene rearrangement. In our current studies, we have compared the induction of DNA cleavage within the MLL bcr in different cell lines after treatment with various anticancer drugs. All anticancer drugs tested including VP-16 (a TOP2-directed drug), camptothecin (a topoisomerase I-directed drug), 5-fluorouracil and methotrexate (antimetabolites), and vinblastine (a microtubule inhibitor) induced the same site-specific cleavage within the MLL bcr. This cleavage was shown to be nuclease-mediated but not TOP2-mediated by the following observations: 1) drug-induced cleavage within the MLL bcr was not protein-linked; 2) unlike TOP2-mediated cleavage, drug-induced DNA cleavage within the MLL bcr was kinetically slow and coincided with the formation of the apoptotic nucleosomal DNA ladder; 3) drug-induced cleavage within the MLL bcr was unaffected in cells with reduced nuclear TOP2; and 4) drug-induced cleavage within the MLL bcr was abolished by the caspase inhibitor, Z-Asp(OCH 3 )-Glu(OCH 3 )-Val- Asp(OCH 3 )-FMK. The possibility that an apoptotic nu-

known as the breakpoint cluster region (bcr) (1,5,12,13). Surprisingly, even the de novo AML have their breakpoints mapped within the same bcr. However, the majority of the t-AML and de novo AML seem to have their breakpoints fall at the telomeric half and the centromeric half of the bcr, respectively (1).
The MLL bcr was found to contain eight Alu repeat elements (10,12), several DNA topoisomerase II (TOP2) consensus sites (12,14), and two scaffold attachment regions (1). Initially, both Alu repeat-mediated recombination and V(D)J recombination (15,16) have been suggested to be involved in the rearrangement of the MLL gene. However, more recent studies have suggested the involvement of TOP2 in mediating MLL gene translocation based on the following obervations: 1) t-AML with the characteristic MLL gene rearrangement is repeatedly associated with the prior use of TOP2-directed drugs (poisons) (2,4,5,(17)(18)(19); 2) several TOP2 consensus sites are found within the bcr (12,14), and furthermore, in vitro cleavage using purified TOP2 has mapped TOP2-mediated cleavage sites close to some of the known breakpoints (20); and 3) Strick et al. (21) have shown that, upon treatment with TOP2-directed drugs, the MLL gene is reversibly cleaved within the bcr.
Although cleavage of the MLL bcr is suggested to be mediated by TOP2 (21), it has also been suggested to be a generalized cellular response to apoptotic stimuli (22). To determine the enzyme(s) responsible for DNA cleavage, we have characterized site-specific cleavage of the MLL bcr in cell lines treated with different anticancer drugs. We showed that all anticancer drugs induced prominent DNA cleavage within the MLL bcr. However, the slow onset of the double-stranded break within the MLL bcr coincided kinetically with the appearance of the apoptotic nucleosomal DNA ladder rather than the formation of TOP2 cleavable complexes. These DNA double-stranded breaks are not protein-linked, suggesting that a nuclease rather than a topoisomerase is involved. Furthermore, this drug-induced DNA cleavage within the MLL bcr was abolished by co-treatment with the caspase inhibitor, Z-DEVD-FMK. In the aggregate, our results support the notion that the slow onset of the site-specific cleavage within the MLL bcr is probably due to the activation of an apoptotic nuclease.
PCR Amplification of the MLL Gene Probe-A two-round PCR of 25 cycles each was performed to amplify the last 338 nucleotides of the MLL bcr to be used as the probe in Southern hybridization. Primers MLL7254 5Ј-AGGTGTGGTGGTGGGCACCTGTAGT-3Ј and MLL8342 5Ј-GGATCCACAGCTCTTACAGCGAACACAC-3Ј were used in the first 25 cycles of amplification in a total volume of 50 l. Two l of the PCR product was then used as template for the second 25 cycles of amplification using primers MLL8005 5Ј-CCCTGAGTGCCTGGGACCAAAC-TAC-3Ј and MLL8342. The PCR product was then ligated with the SmaI-linearized vector, pBluescript KS. The resulting plasmid, pKS-MLLp, harboring the 3Ј-most 338 nucleotides of the bcr, corresponding to nucleotides 8005-8342 of the MLL bcr (Genbank TM accession number U04737).
MLL bcr Cleavage Assay and Southern Analysis-The cells growing in log phase were treated with various drugs as described in the figure legends. At each time point, the cells were lysed with lysis buffer containing 1% SDS, 20 mM EDTA, and 10 mM Tris-HCl, pH 8. The samples were incubated at 50°C overnight in the presence or the absence of 400 g/ml proteinase K. Upon extraction with phenol/chloroform/isoamyl alcohol, DNA was treated with RNase A at 37°C for 1 h, and extraction was repeated. Upon ethanol precipitation, the DNA samples were digested with BamHI to produce the 8.3-kb bcr fragment and then analyzed on 1% agarose gel. Southern transfer was performed using the alkaline transfer method (23). The MLL probe for Southern hybridization was generated by PCR amplification using primers MLL8005 and MLL8342 and the pKS-MLLp template. This probe hybridizes to the 3Ј-most 338 nucleotides of the bcr. Overnight hybridization was performed at 65°C in 1 mM EDTA, 7% SDS, and 0.25 M Na 2 HPO 4 , pH 7.2. The filter was washed once in 1 mM EDTA, 5% SDS, and 20 mM Na 2 HPO 4 , pH 7.2, and twice in 1 mM EDTA, 1% SDS, and 20 mM Na 2 HPO 4 , pH 7.2, prior to autoradiography.

RESULTS
Induction of Chromosomal DNA Breakage within the MLL bcr by Anticancer Drugs-Cleavage within the MLL bcr in cells treated with various anticancer drugs was determined by using an indirect end labeling method (see Fig. 1A for the probe used in Southern hybridization). As shown in Fig. 1 (Fig. 1, see the right panels in B-D), a number of anticancer drugs including VP-16, CPT, vinblastine, 5-fluorouracil, and methotrexate induced DNA cleavage within the MLL bcr (the 8.3-kb bcr fragment is indicated by arrow a, and the 1.5-kb DNA fragment is indicated by arrow b). The appearance of the 1.5-kb DNA fragment is indicative of DNA cleavage within the MLL bcr (Fig. 1A). The cells treated with various anticancer drugs were also monitored for the formation of apoptotic nucleosomal DNA ladder ( Fig. 1, left panels in B-D). In each case, the appearance of the 1.5-kb DNA fragment is paralleled by the formation of nucleosomal DNA ladders (Fig. 1).
Cleavage within the MLL bcr Is Kinetically Distinct from the Trapping of TOP2-DNA Covalent Complexes-The time course of the formation of the 1.5-kb DNA fragment was monitored in various cells treated with the TOP2-directed anticancer drug, VP-16. As shown in Fig. 2, the formation of the 1.5-kb DNA fragment was kinetically slow. No visible cleavage was observed within 1 h of VP-16 treatment. The formation of the 1.5-kb DNA fragment was observed only after 2-3 h of VP-16 treatment (Fig. 2, A-C). The slow formation of the 1.5-kb DNA fragment contrasted with the rapid formation of TOP2-DNA covalent complexes, which was monitored by a band depletion assay (Fig. 2D). Within 10 min of VP-16 treatment, the band intensities of TOP2␣ and TOP2␤ were reduced more than 50%. Prolonged treatment up to 120 min did not result in further change of the band intensities, consistent with the previous results that the formation of TOP2-DNA covalent complexes is rapid (25,26). These results suggest that the formation of the 1.5-kb DNA fragment may not be the direct result of the formation of TOP2-DNA covalent complexes.
Drug-induced Slow Cleavage within the MLL bcr Is Not Mediated by TOP2--TOP2-DNA covalent (cleavable) complexes are characterized by both reversibility at higher temperatures such as 65°C (27) and protein-DNA cross-links (25,26). To test whether this drug-induced slow DNA cleavage within the MLL bcr is mediated by TOP2, we have examined these two unique properties associated with the formation of TOP2-DNA covalent complexes. To test the reversibility of DNA cleavage within the MLL bcr, BV173 cells were treated with 100 M of VP-16 for various times. At each time point, the cells were either lysed immediately or incubated at 65°C for 10 min prior to lysis. As shown in Fig. 3A, site-specific DNA cleavage within the MLL bcr, as evidenced by the formation of the 1.5-kb cleavage product (see bands indicated by arrow b), was induced at 3 and 6 h of VP-16 treatment. Brief heating to 65°C for 10 min prior to lysis had no detectable effect on the formation of the 1.5-kb DNA fragment (compare lanes 3 and 4 with lanes 10 and 11, respectively), suggesting that DNA cleavage within the MLL bcr is not mediated by TOP2.
Another characteristic of TOP2-mediated DNA cleavage is the formation of covalent TOP2-DNA cross-links (25,26). We have performed the following experiment to test the possible covalent association of protein with the cleaved DNA fragments. BV173 cells treated with VP-16 were lysed at various times. The cell lysates were incubated at 50°C overnight in the presence or the absence of proteinase K. Extraction with phenol/chloroform/isoamyl alcohol was then performed, and genomic DNA was extracted. Without proteinase K digestion, TOP2-mediated DNA cleavage products, which are proteinassociated, are expected to be lost (in the interphase) during extraction. With proteinase K digestion, TOP2-mediated DNA products are expected to be recovered in the aqueous phase during extraction. As shown in Fig. 3A, DNA cleavage within the MLL bcr, as evidenced by the formation of the 1.5-kb DNA fragment, was induced at 3 and 6 h of VP-16 treatment (lanes  3 and 4). The formation of this DNA cleavage product was not affected by proteinase K digestion (Fig. 3A, lanes 7 and 8), suggesting that the fragment was not protein-linked.
The absence of protein-DNA cross-links in the DNA cleavage products cannot completely rule out the possibility of TOP2mediated DNA cleavage within the MLL bcr. TOP2 can be degraded by the ubiquitin/26 S proteasome pathway in cells treated with TOP2-directed drugs such as VP-16. 2 To test whether TOP2 had been degraded by the ubiquitin/26 S proteasome pathway, the proteasome inhibitor, MG132, was used to block this degradative pathway. The cells were pretreated with 1 M of MG132 for 1 h followed by co-treatment with 100 M of VP-16. DNA was extracted in the presence and the absence of proteinase K as described above and analyzed by Southern blotting and indirect end labeling. As shown in Fig.  3B, the formation of the 1.5-kb DNA fragment was not affected by MG132 with or without proteinase K treatment (compare lanes 2 and 3 with lanes 5 and 6, respectively). Again, this result is consistent with the notion that DNA cleavage within the MLL bcr is mediated by a nuclease but not TOP2. Another stringent test for the possible involvement of TOP2 in DNA cleavage within the MLL bcr is the use of TOP2deficient cell lines. HL-60/MX2 is a cell line derived from HL-60 after selection for resistance to mitoxantrone (28,29). It has undetectable TOP2␤ and a reduced level of nuclear TOP2␣ (30). 3 As shown in Fig. 4, the time course of the formation of the 1.5-kb cleavage product was about the same in HL-60 and HL-60/MX2 cells treated with CPT, suggesting that TOP2 is unlikely to be responsible for the formation of the 1.5-kb cleavage product. As a control, the formation of the 1.5-kb cleavage product was shown to be significantly delayed in HL-60/MX2 cells as compared with HL-60 cells treated with TOP2-directed drug, VP-16 (Fig. 4), reflecting the reduction of TOP2-mediated DNA damage in HL-60/MX2 cells. These results suggest that TOP2-mediated DNA damage can lead to the formation of the 1.5-kb cleavage product but that TOP2 is not directly responsible for this slow cleavage event.
Cleavage of the MLL bcr Is Mediated by an Apoptotic Nuclease-K562 cells have been reported to exhibit delayed induction of apoptosis by various agents including VP-16 as compared with HL-60 cells (31). As shown in Fig. 5A, the nucleosomal DNA ladders, which are characteristics of apoptotic cell death, were prominently induced in HL-60 but not K562 cells after 4 h of treatment with either VP-16 or CPT. The  1, 5, and 9). At each time point, the cells were lysed immediately (lanes 1-8) or heated at 65°C for 10 min prior to lysis (lanes 9 -11). The cell lysates were incubated at 50°C overnight in the presence (lanes 1-4 and 9 -11) or the absence (lanes 5-8) of proteinase K. DNA samples were then prepared for Southern hybridization as described in the legend of Fig. 1. B, BV173 cells were pretreated with 1 M of MG132 for 1 h followed by co-treatment with 100 M of VP-16 for 2 h (lanes 2 and 5) or 3 h (lanes 3 and 6) or with Me 2 SO for 3 h (lanes 1 and 4). The cell lysates were treated with (lanes 1-3) or without (lanes 4 -6) proteinase K overnight at 50°C. DNA samples were then prepared for Southern hybridization. Arrows a and b as well as lanes D (Me 2 SO) have been described in the Fig. 1 legend. formation of these nucleosomal DNA ladders has been attributed to the activation of the apoptotic nuclease CAD (32,33). Our studies have suggested that DNA cleavage within the MLL bcr is not TOP2-mediated but correlated with the appearance of the nucleosomal DNA ladders. Consequently, we have tested whether the formation of the 1.5-kb DNA cleavage product is correlated with the appearance of the nucleosomal DNA ladders in K562 and HL-60 cells. As shown in Fig. 5B, the 1.5-kb DNA cleavage product was prominently induced in HL-60 but not K562 cells after 4 h of treatment with either VP-16 or CPT (compare lanes 2 and 3 with lanes 5 and 6, respectively), suggesting that the DNA cleavage within the MLL bcr may be mediated by an apoptotic nuclease. It should be noted that the cellular TOP2 levels in both HL-60 and K562 cells have been shown to be similar (31). In addition, the amounts of VP-16induced TOP2-linked DNA breaks in both HL60 and K562 cells are comparable (31).
The most likely candidate nuclease is the CAD that has been shown to be responsible for the nucleosomal DNA ladder formation (32,33). To test this possibility, Z-DEVD-FMK, a known cell-permeable and irreversible caspase 3 inhibitor, which also inhibits caspases 6, 7, 8, and 10 (34, 35) was used. BV173 cells were pretreated for 1 h with 50 M of Z-DEVD-FMK, followed by co-treatment with VP-16 for 3 h. As shown in Fig. 6, the formation of the nucleosomal DNA ladder induced by VP-16 was completely abolished by Z-DEVD-FMK (Fig. 6A,  compare lanes 2 and 4). The formation of the 1.5-kb cleavage product, which is indicative of DNA breakage within the MLL bcr, was concomitantly abolished upon Z-DEVD-FMK treatment (Fig. 6B, compare lanes 2 and 4). DISCUSSION Previous studies have suggested that TOP2 may be responsible for the MLL gene rearrangement in patients with t-AML. This is primarily based on the observation of Strick et al. (21) that TOP2 is responsible for DNA cleavage within the MLL bcr in cells treated with TOP2-directed drugs. Our current studies, on the other hand, have indicated that an apoptotic nuclease is involved in DNA cleavage within the MLL bcr despite that both studies are conducted using the same cell line (BV173) and the same TOP2 poison, VP-16. Our results, which argue against TOP2 being the major "nuclease" for site-specific cleavage within the MLL bcr, are summarized as follows: 1) DNA cleavage within the MLL bcr is observed in cells treated with not only the TOP2 inhibitor VP-16 but also other anticancer drugs including a topoisomerase I inhibitor CPT, antimetabolites such as 5-fluorouracil and methotrexate, and a microtubule inhibitor vinblastine; 2) site-specific DNA cleavage within the MLL bcr is kinetically slower than the formation of TOP2 cleavable complexes; 3) unlike TOP2-mediated DNA cleavage, site-specific cleavage within the MLL bcr is not associated with covalently bound protein, and the possibility that the covalently bound protein is removed by the ubiquitin/26 S proteasome pathway is also ruled out by the use of the 26 S proteasome inhibitor, MG132; 4) unlike TOP2-mediated DNA cleavage, which is highly reversible, site-specific cleavage within the MLL bcr cannot be reversed by incubation at 65°C (27); and 5) site-specific DNA cleavage within the MLL bcr is unaffected by nuclear TOP2 levels. We have shown that this cleavage occurs with about the same efficiency in both HL-60 and its TOP2-deficient variant HL-60/MX2 (28). Similar results have been obtained previously in another TOP2 mutant cell line, CEM/VM-1-5 treated with cytosine arabinoside (22). Although our results argue against TOP2 being responsible for site-specific cleavage within the MLL bcr in transformed cells, we cannot rule out the possibility that TOP2 is the nuclease responsible for site-specific DNA cleavage within the MLL bcr during transformation of hematopoietic stem cells giving rise to t-AML.
Our results have also provided evidence supporting the involvement of an apoptotic nuclease in the direct cleavage of the MLL bcr: 1) the appearance of the nucleosomal DNA ladder, which is characteristic of apoptotic cell death, is correlated with the appearance of the 1.5-kb cleavage product, which is indicative of site-specific DNA cleavage within the MLL bcr, and the co-appearance of the apoptotic nucleosomal DNA ladder and the 1.5-kb cleavage product is demonstrated in a variety of conditions including the use of different anticancer drugs and different cell lines; and 2) the caspase inhibitor, Z-DEVD-FMK, was shown to inhibit the formation of the 1.5-kb DNA cleavage product. In the apoptosis process, a family of cysteine proteases, the caspases, are activated to cleave a group of cellular proteins (reviewed in Ref. 36) including ICAD. Cleavage of ICAD allows the activation of the nuclease CAD (32). CAD has been well established as the enzyme responsible for cleaving the chromatin leading to the formation of high molecular weight and nucleosomal DNA fragments (32,33). It seems possible that CAD may be responsible for site-specific DNA cleavage within the MLL bcr during high molecular weight DNA fragmentation and more prominently during nucleosomal DNA fragmentation.
It is difficult to reconcile the differences between our results and those obtained by Strick et al. (21). Nevertheless, it is possible that VP-16 may induce a low level of reversible TOP2 cleavable complexes within the MLL bcr. This low level of TOP2 cleavable complexes may have been detected by Strick et al. (21) but missed in our current studies. This initial signal may be followed by activation of the apoptotic nuclease CAD, which leads to more extensive and irreversible cleavage within the MLL bcr, which was detected in the current studies but somehow missed by Strick et al. (21). Indeed, previous studies have demonstrated that VP-16 induces TOP2-mediated DNA cleavage at an early time followed by nucleolytic degradation of chromosomal DNA into both high molecular weight and nucleosomal DNA fragments (37).
The relevance of the site-specific DNA cleavage within the MLL bcr to MLL gene rearrangement is unclear. However, the location of the cleavage site lies within the telomeric half of the bcr, which is known to be the region undergoing MLL gene rearrangement in t-AML patients (1). It seems possible that a specific region within the telomeric half of the bcr is highly sensitive to nucleases/topoisomerases such as CAD and TOP2. Depending on the specific conditions, either CAD or TOP2 may be primarily responsible for site-specific cleavage within this region of the bcr. Previous studies have demonstrated that TOP2 can be activated to cleave DNA in the presence of TOP2directed anticancer drugs as well as under certain cellular stresses such as oxidative stress, thiol stress, and acidic pH (37,38). 4 A low level of TOP2-mediated DNA cleavage may be sufficient to induce apoptotic cell death resulting in extensive CAD-mediated DNA cleavage within the same region of the bcr. Other apoptotic stimuli (e.g. CPT and other anticancer drugs) may activate CAD to cleave the bcr by mechanisms not involving TOP2. Whether CAD-and TOP2-mediated cleavage within the MLL bcr may result in DNA rearrangements is unknown. However, it has been demonstrated that TOP2-mediated DNA cleavage is highly efficient in mediating DNA sequence rearrangements (39 -43). DNA double-stranded breaks resulting from CAD-mediated DNA cleavage are also expected to lead to DNA sequence rearrangements via either homologous recombination or nonhomologous end joining (44 -49). Although the majority of patients with the MLL gene translocation have previously been treated with TOP2-directed drugs, there are cases where MLL gene translocations were observed without prior treatment with TOP2-targeting drugs (20,50). Thus, nuclease-mediated cleavage of the MLL bcr could be relevant in MLL gene rearrangement. Clearly, further studies are necessary to establish the roles of the enzymes involved in site-specific DNA cleavage within the MLL bcr and MLL gene rearrangements.