Commentary: topoisomerases as targets in acute leukemia: I, II, I plus II or none?

DNA topoisomerases maintain DNA topology and are essential in DNA replication, supercoiling, transcription, and chromatin remodeling. Type I topoisomerase enzymes cleave one strand of DNA and form a transient covalent complex with either the 3′ or 5′ end of that DNA molecule (so-called ‘cleavable complex’). These enzymes are the targets for the topoisomerase poisons topotecan and irinotecan. Type II DNA topoisomerases create double strand breaks that allow for catenation/decatenation of DNA molecules. These enzymes are targets for the epipidophyllotoxins etoposide and teniposide, m-amsacrine, and at higher concentrations doxorubicin, daunorubicin, and mitoxantrone (reviewed in [1], [2]).

The clinical utility of some topoisomerase inhibitors are firmly established: daunorubicin is a major drug in AML therapy, while irinotecan in combination with 5-fluorouracil and leukovorin has improved the survival of patients with metastatic colon cancer. The latter combination is undergoing clinical trials as a front-line therapy for a number of solid tumors [3]. Etoposide has been widely used for solid and hematologic malignancies. Etoposide (and in pediatric protocols teniposide) are incorporated into treatment algorithms for acute lymphoblastic leukemia and aggressive non-Hodgkins lymphomas. Etoposide is used in three-drug induction for acute myeloid leukemia, for stem cell mobilization and in preparative regimens for stem cell transplantation. While etoposide is clearly active against AML, the superiority of etoposide-containing regimens in AML has not been established in randomized trials.

Based on early in vitro data demonstrating marked activity of the topoisomerase I inhibitor camptothecin against a variety of hematologic malignancies, the leukemia therapeutics community welcomed the availability of topotecan as a potential cytotoxic agent for the treatment of leukemia with a novel molecular target. Topotecan underwent Phase I studies in acute leukemia as an intravenous continuous infusion and daily intravenous bolus; the drug was clearly cytoreductive, but few complete remissions were achieved among typical multiply relapsed Phase I patient populations [4], [5], [6]. A 20% response rate was seen in patients with MDS and chronic myelomonocytic leukemia treated with a 5-day continuous infusion of topotecan; however, toxicity and treatment-related mortality were substantial (∼20%) [7]. A Phase II ‘window’ study in untreated high risk adult acute lymphoblastic leukemia patients was terminated early because of lack of sufficient single agent activity [8]. Obvious empiric combinations were studied: ara-C plus topotecan induced remissions in AML and MDS patients, but such remissions were short-lived [9], [10]. Sequential use of alkylators and topotecan were attempted based on models which showed an increase in replicative forks following alkylator therapy [11]. Minor activity of cyclophosphamide plus topotecan was demonstrated in Phase I trials; a trial of carboplatin plus topotecan is ongoing, led by Mayo Clinic investigators [11]. Cyclophosphamide plus ara-C and topotecan was studied in 63 patients with refractory or relapsed acute leukemia [12]. Seventeen percent of AML patients achieved complete remission; this regimen is undergoing further Phase II investigation in relapsed AML by the Eastern Cooperative Oncology Group. To date, topotecan has not made a major impact on resistant myeloid malignancies as a single agent or in empiric and biologically rational combinations.

Another potentially rational combination involved sequential use of topo I and II inhibitors. These combinations derived from the fact that the two topoisomerases are involved in different catalytic processes. In vitro studies suggested that the simultaneous addition of topo I and topo I poisons might be antagonistic [13]. However, sequential addition appeared additive or synergistic for cell kill; maximum cell kill was achieved when etoposide was given following topotecan [14]. One potential explanation for this synergistic interaction was the up-regulation of topo II following exposure to a topo I inhibitor.

The well-designed and executed study in this issue by Cooper and colleagues from Case Western Reserve attempted to translate this observation clinically. Dose escalation of a 72h continuous infusion of topotecan was followed by a fixed dose of etoposide at 100mg/m2 for 5 days to patients with relapsed or refractory acute leukemia (22 AML, 6 ALL, 1 CML blast crisis). While several patients achieved bone marrow aplasia, only two (of 31 total treated) achieved complete response. The protocol was terminated early due to minimal observed efficacy. A prior Phase I trial of this combination was performed by the National Cancer Institute of Canada (NCIC). Topotecan was given as a 5-day continuous infusion followed by IV bolus etoposide on days 6-8. One remission was achieved in 10 patients treated [15]. Several Phase I studies of sequential topotecan and etoposide were performed in patients with solid tumors. Although isolated responses were identified, toxicity, primarily hematologic, was substantial [16], [17], [18], [19].

In the current study, in contrast to expectations, topo II activity increased in only 3 of 25 patients following the administration of topotecan. Moreover, patients who developed bone marrow response were not more likely to have higher topo I and II activity prior to treatment compared to those who did not respond. In the prior NCIC study, topo II levels increased in peripheral blood blasts during the first 72h of topotecan infusion but returned to near baseline by day 5; topo II levels decreased in bone marrow blasts by day 5 compared to pretreatment [15]. In the solid tumor studies, topo II was upregulated in only one of six tumors sequentially biopsied during sequential therapy [19]. In previous studies in acute leukemia, response to topo poisons was not correlated with cellular topo activity [5], [6], [8], [20], [21], [22]. While the identification of molecular markers which correlate with clinical outcomes and resistance is a laudatory goal, these studies are necessarily underpowered to detect such correlations due to the extremely low meaningful clinical response rate. Many patients in these studies develop clinically insignificant transient tumor clearance with immediate leukemic regrowth. For biological variables to be clinically useful, they should predict clinically meaningful endpoints such as sustained complete remission or survival. To detect such correlations, significant numbers of such responses need to be observed, a challenge in any Phase I trial.

While much has been learned from the three-dimensional structure of topoisomerases, the exact mechanism by which they affect DNA movement subsequent to DNA cleavage is still unclear. The mechanism by which topo poisons inhibit these enzymes is also unknown. It has been suggested that camptothecins inhibit the re-ligation step in topo I-mediated catalysis, interact with DNA and topo I simultaneously, intercalate into DNA and lead to ‘base flipping’ or DNA bending which interferes with the DNA re-ligation step [23], [24], [25]. The large number of potential molecular mechanisms complicates analysis of the action of topo inhibitors.

Two important questions regarding topo poisons remain to be fully clarified: the mechanism of cell kill, and the mechanisms of resistance. Collisions between advancing DNA replication forks and topo I cleavable complexes may lead to induction of apoptosis in S-phase. Double strand breaks seem to be important molecular intermediates in the initiation of cell death but the subsequent signal transduction steps leading to cell death are not clear. In primary samples of human acute leukemia cells, the topotecan concentration required to stabilize 50% of cellular topo I in topo I-DNA complexes ranged from 3 to greater than 100μM (median, 30μM), compared to 4μM required in cell lines [26]. This result suggested that the formation of topo-I DNA adducts may be diminished in relapsed acute leukemia cells.

While these drugs are highly selective for their targets, other processes that they inhibit are likely rather pleotropic. Based on their interference with basic cellular processes, many potentially lethal combinations of cellular events could occur; a dominant molecular pathway mediating the anti-tumor effect of these drugs has not been found. This lack of basic understanding hampers the further development of rational drug combinations. Targeting multiple essential processes simultaneously may lead to synergistic interactions; however, it is most important that the agents have non-cross-reactive mechanisms of resistance. Early data suggested that unlike etoposide, topotecan was not subject to drug efflux by MDR1; however, this is probably not true [27], [28], [29]. Moreover, as with many malignancies, AML cells may resist apoptosis induction by drugs which induce cell death through many pathways [30], [31].

Why is a drug like topotecan, which inhibits an essential cellular enzyme, so ineffective in cell killing? Many mechanisms of resistance to topo I inhibitors are likely; targeting these resistance mechanisms may prove more fruitful than trying to increase cell kill by targeting essentially the same molecular principle twice (with topo I/II combinations). Combination regimens work best before resistant clones dominate tumors; thus, therapeutic window approaches in less heavily pre-treated patients and strategies to bypass resistance mechanisms (particularly pathways which lead to global inhibition of apoptosis) may prove fruitful. The paucity of effective regimens for patients with relapsed and refractory acute leukemia (indeed the majority of patients with acute leukemia) mandates the urgent development of such strategies.