The Indenoisoquinoline LMP517: A Novel Antitumor Agent Targeting both TOP1 and TOP2

The camptothecin derivatives topoisomerase I (TOP1) inhibitors, irinotecan and topotecan, are FDA approved for the treatment of colorectal, ovarian, lung and breast cancers. Because of the chemical instability of camptothecins, short plasma half-life, drug efflux by the multidrug-resistance ABC transporters, and the severe diarrhea produced by irinotecan, indenoisoquinoline TOP1 inhibitors (LMP400, LMP776, and LMP744), which overcome these limitations, have been developed and are in clinical development. Further modifications of the indenoisoquinolines led to the fluoroindenoisoquinolines, one of which, LMP517, is the focus of this study. LMP517 showed better antitumor activity than its parent compound LMP744 against H82 (small cell lung cancer) xenografts. Genetic analyses in DT40 cells showed a dual TOP1 and TOP2 signature with selectivity of LMP517 for DNA repair-deficient tyrosyl DNA phosphodiesterase 2 (TDP2)- and Ku70-knockout cells. RADAR assays revealed that LMP517, and to a lesser extent LMP744, induce TOP2 cleavage complexes (TOP2cc) in addition to TOP1ccs. Histone γH2AX detection showed that, unlike classical TOP1 inhibitors, LMP517 targets cells independently of their position in the cell cycle. Our study establishes LMP517 as a dual TOP1 and TOP2 inhibitor with therapeutic potential.


Introduction
Topoisomerases play critical roles in genome organization and stability. They remove supercoils and DNA intertwining induced by essential DNA processes including replication, transcription, and chromatin remodeling (1)(2)(3)(4). Topoisomerase I (TOP1) induces single-strand breaks, allowing the DNA to rotate on itself, while topoisomerases II (TOP2a and TOP2b) induce DNA double-strand breaks (DSB), allowing a DNA helix to pass through another. Both classes of enzyme act by forming transient catalytic intermediates called topoisomerase cleavage complexes (TOPcc). Rejoining of the DNA strand(s) restores DNA integrity. Unresolved TOPccs lead to protein-linked DNA single-or double-strand breaks that are lethal as cells undergo replication or transcription (2). For this reason, TOP1 and TOP2 inhibitors are widely used for cancer treatment. TOP1ccs are targeted by the clinical camptothecin (CPT) derivatives topotecan and irinotecan, and CPT derivatives are the only chemical class of TOP1 inhibitor approved by the FDA (5). TOP2ccs are targeted by etoposide, doxorubicin (as well as other anthracycline derivatives), and mitoxantrone (4,6). Although TOP1 and TOP2 inhibitors are highly specific for their respective enzyme targets (TOP1 and TOP2, respectively), they share the same mechanism of action (7); they poison TOPccs by impeding DNA resealing as the drugs stack with the base pairs flanking the break made by the topoisomerases while making amino acid hydrogen bonds with the topoisomerases. This mode of inhibition, termed interfacial inhibition, was first discovered for the topoisomerase inhibitors and has been generalized to other natural products targeting macromolecular interfaces (7).
Following the trapping of TOPccs by topoisomerase inhibitors, cells must repair the irreversible topoisomerase-linked DNA breaks induced when replication and/or transcription machineries collide with the TOPccs. Tyrosyl DNA phosphodiesterases 1 and 2 (TDP1 and TDP2) remove the topoisomerase polypeptides covalently linked to the ends of the breaks (8)(9)(10); and the associated DSBs are repaired by two main pathways: homologous recombination (HR) and nonhomologous end joining (NHEJ). HR primarily repairs TOP1ccs during S-phase when replicated DNA is available as repair template (11)(12)(13)(14). In contrast, NHEJ relies on Ku70/80 and DNA-dependent protein kinase complexes, which directly joins the broken ends in nonreplicative cells. NHEJ is the primary repair pathway for TOP2ccs following their processing by TDP2 (8,14,15). Notably, inactivation of NHEJ by knocking-out Ku70 in chicken lymphoblastoid DT40 cells increases resistance to CPTs (14). Indicating that toxic replication intermediates are generated by NHEJ in response to TOP1ccs while NHEJ effectively repairs TOP2ccs (14,15).
The indenoisoquinolines are new TOP1 inhibitors that have been developed to overcome the limitations of CPTs (5,16). Three indenoisoquinolines derivatives are in phase I/phase II clinical trials, LMP400 (indotecan), LMP776 (indimitecan) and LMP744 (5,17,18). Unlike the CPTs, the indenoisoquinolines are chemically stable, induce persistent TOP1ccs, are not substrates of drug efflux by ABC transporters and have an extended plasma half-life (17,18). Moreover, unlike irinotecan, they do not cause severe diarrhea (5,17,18). We recently reported the anticancer activities of second-generation indenoisoquinolines, the fluoroindenoisoquinolines, in which the addition of a fluorine at position 3 replaces the methoxy groups at position 2, 3 of the clinical indenoisoquinolines ( Fig. 1A; refs. [19][20][21]. We showed that among these fluoroindenoisoquinolines, LMP517 generated the most DNA damage, which correlated with its potency for trapping TOP1cc (19). The aim of this study was to identify the molecular mechanisms explaining the enhanced LMP517 potency and to compare it with its parent clinical derivative LMP744.

Survival assays
DT40 cells were seeded at 5,000 cells per well in 96-well white plates (catalog no. 6005680, PerkinElmer Life Sciences) and exposed to the indicated concentrations of LMP744, LMP517, etoposide, or topotecan for 72 hours, performed at least three times with in-experiment triplicates. Cellular viability was determined using ATPlite 1-step kit (PerkinElmer). ATP levels of untreated cells were defined as 100%. Percent viability was defined as: [(ATP signal in treated cells)/(ATP signal in untreated cells)] Â 100.

Mouse antitumor experiments
Athymic nude mice (nu/nu, female, 20-25 g, 8-12-week-old) from Charles River, were transplanted with 5 million H82 human small-cell lung cancer cells (SCLC). When the tumor volume reached between 100 and 125 mm 3 , the animals were randomized into treatment groups based on tumor volume and body weights using the StudyLog software. Ten mice for LMP744, vehicle arm and for LMP517 were used. The animals were treated with either LMP744 (10 mg/kg) administered intravenous (i.v.) or with LMP517 (10 mg/kg) administered by i.v. push via tail vein once a day for 5 consecutive days once (1 cycle) or twice (2 cycles). LMP744 and LMP517 were dissolved in 10 mmol/L citric acid, 5% dextrose. The three axes (millimeters) of tumors were measured with a caliper to calculate tumor volume. Measurements were made every 3 or 4 days. Maximum allowable weight loss tolerated of 20% was never reached. Mice were euthanized if tumors presented necrosis or exceeded 20 mm in diameter.
Frederick National Laboratory is accredited by AAALAC International and follows the Public Health Service Policy for the Care and Use of Laboratory Animals. Animal care was provided in accordance with the procedures outlined in the "Guide for Care and Use of Laboratory Animals" (National Research Council; 2011; National Academies Press; Washington, DC).
Histone gH2AX detection HCT116 and HCT116 FUCCI cells were plated at 50,000 cells per well in 4-well chamber slides, incubated for 72 hours and then treated with drugs. After treatment, slides were fixed in paraformaldehyde 4% for 10 minutes at room temperature then permeabilized with 0.02% Triton X-100 for 5 minutes at room temperature. After rinsing with PBS, the slides were blocked for 1 hour in PBS-BSA 8% at room temperature and incubated for 2 hours with primary antibody at room temperature (Abcam ab22551 mouse anti-gH2AX). After washing with PBS, cells were incubated with secondary antibody for 1 hour at room temperature, washed with PBS and mounted with Vectashield with DAPI (Vector Laboratories). The primary antibody for histone gH2AX was mouse monoclonal anti-gH2AX Ser140 antibody, clone JBW301 (Abcam: ab22551) and the secondary antibody was a chicken anti-mouse Alexa 647 (A21463 Invitrogen).

Results
The fluoroindenoisoquinoline LMP517 has an improved antitumoral activity over LMP744 To test the antitumoral activity of LMP517, we xenografted nude mice with the SCLC cell line H82 (19) and compared LMP517 with its parent indenoisoquinoline LMP744 in one or two treatment cycles (5 days of treatment per cycle). LMP744 and LMP517 (chemical structures in Fig. 1A) both had an maximum tolerated dose (MTD) of 10 mg/kg. At this concentration, only LMP517 induced a reduction in tumor growth (Fig. 1B) with marginal body weight lost (Supplementary Fig. S1). LMP517 treatment resulted in an average survival of 30 days for 1 cycle and 36 days for the 2 cycles protocol versus 19 days for both one cycle and two cycles of LMP744 (Fig. 1C). This result demonstrates the increased efficacy of the fluoroindenoisoquinoline LMP517 over LMP744 in this model system. The following experiments were performed to elucidate the molecular pharmacology of LMP517 and understand why it is more potent than its parent counterpart LMP744.
LMP517 displays a comparable phenotype with etoposide in isogenic TDP-and Ku70-knockout DT40 cells Cells have redundant pathways to excise unresolved covalent complexes between topoisomerases and DNA (2). The most specific pathways involve TDP1 and TDP2, which hydrolyze the covalent bond between the tyrosine of the topoisomerase polypeptides (TOP1 and TOP2, respectively) and the DNA backbone. TDP1 inactivation inhibits the repair of TOP1cc and increases sensitivity to CPT, and to a lesser extent to etoposide (26,27). Conversely, TDP2 inactivation increases TOP2cc accumulation and sensitizes cells to etoposide (9,28). Here we used DT40 TDP1-or TDP2deficient cells (referred as DT40 tdp1 and DT40 tdp2, respectively; refs. 27, 28) to determine the impact of TDP1 and TDP2 on the activity of LMP517.
As expected, DT40 tdp1 cells were hypersensitive to CPT with an IC 50 of 2 nmol/L versus 15 nmol/L for DT40 WT (Fig. 2A). DT40 tdp2 cells were also sensitive to CPT, consistent with the role of TDP2 as secondary repair pathway for TOP1ccs (9, 28) (IC 50 : 5 nmol/L for DT40 Tdp2; Fig. 2A). LMP744 showed a similar sensitivity pattern as CPT with an IC 50 of 6 nmol/L versus 25 nmol/L for DT40 WT cells and 15 nmol/L for DT40 tdp2 cells.
Repair of TOP2cc by TDP2 generates direct substrate for NHEJ (2,8,9,15) whereas NHEJ exerts the opposite effect in response to TOP1cc by initiating toxic NHEJ (14). We used this difference to further determine whether LMP517 could be acting as a TOP2 poison  ( Fig. 2B). DT40 Ku70-deficient cells showed the expected resistance to topotecan while being hypersensitive to etoposide, LMP517, and LMP744 (Fig. 2B). The phenotypes observed in TDP1-, TDP2-, and Ku-deficient DT40 cells suggest that LMP517 could act as a TOP2 poison.
LMP517 is a dual TOP1 and TOP2 inhibitor in biochemical assays and in HCT116 and TK6 cells To directly determine the effects of LMP517 on TOP1 and TOP2, we performed biochemical assays with negatively supercoiled plasmid pBR322 in the presence of recombinant TOP1 or TOP2, using CPT and etoposide as positive controls. As expected, both LMP517 and LMP744 trapped TOP1ccs (Fig. 3A). Indeed, when TOP1 was applied to supercoiled DNA in the presence of LMP744 or LMP517, we observed extensive accumulation of nicked DNA, as in the case of CPT (Fig. 3A, lane C; ref. 30). Notably, in the presence of TOP2, we observed both linearized and nicked DNA with LMP517 or LMP744, which are characteristic results obtained with trapping of TOP2 by etoposide (Fig. 3A, lane E at right; ref. 31). The induction of nicked and linear DNA species is consistent with TOP2 trapping by LMP517 and LMP744.
To test whether LMP517 also acts as a dual TOP1 and TOP2 inhibitor in cells, we treated human lymphoblast TK6 and colon carcinoma HCT116 cells for 1 hour with LMP517, CPT, etoposide, or LMP744 and detected TOP1cc and TOP2cc by Radar assay (32). Figure 3B shows that LMP517 induces both TOP2a and TOP2bccs in addition to TOP1ccs in both cell lines tested ( Fig. 3B and C). In HCT116, we also tested the indenoisoquinoline LMP744 and found that it also induced TOP2ccs, although to a lesser extent than LMP517 (Fig. 3C).  Together, these results demonstrate that LMP517 is a dual inhibitor of TOP1 and TOP2 (both TOP2a and TOP2b) in biochemical and cellular assays.
LMP517 induces histone gH2AX in both G 1 -and S-G 2 cell-cycle phase cells, consistent with its dual activity against TOP1 and TOP2 It is well established that TOP1ccs trapped by CPTs induce DNA damage primarily in replicative cells (S-phase; refs. 2, 33) while etoposide targets cells in all phases of the cell cycle by trapping both TOP2a and TOP2b (2,34). To test the induction of DNA damage by LMP517 and whether such damage is dependent on replication, we used HCT116 FUCCI cells, which express tagged cell-cycle reporter peptides, mAzami-Green-tagged N-terminus of geminin (green channel) or mKusabira-Orange2-tagged N-terminus of CDT1 (red channel). This dual labeling allows the detection of cells in the S/G 2 -or G 1 -phase, respectively (22).
Notably, cells treated with LMP517 did not show a bimodal gH2AX signal distribution (Fig. 4A). Analyses of the GFP and RFP signals and cell-cycle distribution showed that, like etoposide, LMP517 induced gH2AX signal in G 1 -phase cells (88% and 89% of total G 1 -phase cells, respectively; Fig. 4B, left). Only a small percentage of G 1 -phase cells showed gH2AX signal when treated with CPT or LMP744 (29% and 23% of total G 1 -phase cells, respectively; Fig. 4B, left). These results demonstrate that LMP517, like etoposide and contrary to CPT and LMP744, induces DNA damage in cells in G 1 -phase of the cell cycle, consistent with cellular damage induced by TOP2 trapping.
Schlafen 11, NHEJ, and BRCAness are determinants of response to LMP517 Schlafen 11 (SLFN11) is an established dominant determinant of response to both TOP1 and TOP2 inhibitors (37)(38)(39). We recently reported, in isogenic human leukemia CCRF-CEM cells, an increased resistance to LMP517 in the absence of SLFN11 expression (19). To determine the predictive value of SLFN11 expression in nonisogenic cells, we analyzed the activity of LMP517 across the NCI-60 cell line panel using CellMinerCDB (Fig. 5A; refs. 40, 41). A significant correlation was observed between the activity of LMP517 and SLNF11 expression. However, some cells seem to respond in the absence of SLFN11 and other, SLFN11-positive, were not hypersensitive to  LMP517, indicating other cellular determinants of response. Using the CellMinerCDB tools (http://discover.nci.nih.gov/cellminercdb; refs, 40, 41), we identified two other determinants of response: XRCC6 (KU70) copy number and Aprataxin and PNKP Like Factor (APLF) expression, which were both negatively correlated to LMP517 response (Supplementary Fig. S4A and S4B). These results are consistent with those shown in Fig. 2B, demonstrating that NHEJ plays a major role in LMP517 cytotoxicity. The predicted response with all three determinants of response (SLFN11, Ku70, and APLF) is represented in Fig. 5B. Figure 5C shows the correlation between SLFN11 expression, XRCC6 DNA copy number, and APLF expression with LMP517 using the CellMinerCDB "Multivariate analyses" tool (http://discover. nci.nih.gov/cellminercdb; ref. 41).
We also recently demonstrated a selectivity of the first-generation indenoisoquinolines for homologous recombination deficient (HRD; BRCA1, BRCA2, and PALB2 deficient) cells and their synergistic combination with the PARP inhibitor olaparib (11). To determine whether LMP517 was also selective for BRCA1-, BRCA2-, and PALB2-deficient cells, we tested isogenic DT40 cells inactivated for those genes. Hypersensitivity of DT40 cells lacking BRCA1, BRCA2, and PALB2 was observed (Fig. 5D). Next, we combined LMP517 with olaparib (Fig. 5E) and found additive effects but no synergism between the two drugs. This result is consistent with previous studies showing additive activities but no synergism between TOP2 inhibitors and PARP inhibitors (42), consistent with the conclusion that TOP2 is a prominent cellular target of LMP517.

Discussion
Our results demonstrate that the fluoroindenoisoquinoline LMP517 (19,20) displays improved antitumoral efficacy in murine H82 (SCLC) xenograft compared with its parent compound LMP744 (see Fig. 1), which was recently introduced in phase I clinical trials based on its antitumor activity in dog lymphomas (18). Hence, LMP517 has the potential to be developed as second-generation indenoisoquinoline inhibitor. Further studies are warranted to determine LMP517 potency against other models and determine its clinical potential.
Comparing LMP517 with its parent indenoisoquinoline LMP744 reveals that the addition of a fluorine and removal of the methoxy groups in the A-ring (see Fig. 1) increased the ability of LMP517 to target TOP2. In our recent study of LMP517, we also noticed the ability of LMP517 to intercalate into DNA at high concentration as it has been previously established for LMP744 (36). This intercalating effect of LMP517 and LMP744 may explain the inhibitory effect on TOP2 as several chemical classes of DNA intercalating agents effectively trap TOP2ccs (45)(46)(47) by stacking with the base pairs flanking the TOP2cc, including the clinically used anthracyclines (doxorubicin, daunorubicin, epirubicin, idarubicin), mitoxantrone, and amsacrine, as well as ellipticine derivatives (6,43). Indeed, DNA intercalation of drugs and ligands such as benzo[a]pyrene carcinogens at a TOP2 cleavage site can directly block the DNA religation by TOP2 (7,48,49).
Our results provide groundwork for the potential clinical development of LMP517 as a dual TOP1 and TOP2 inhibitor and potentially as a tumor-targeted delivery payload (5,50).

Disclosure of Potential Conflicts of Interest
Y. Pommier has ownership interest in a patent. No potential conflicts of interest were disclosed by the other authors.

Disclaimer
The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. government.