Novel Fluoroindenoisoquinoline Non-Camptothecin Topoisomerase I Inhibitors

Contrary to other anticancer targets, topoisomerase I (TOP1) is targeted by only one chemical class of FDA-approved drugs: topotecan and irinotecan, the derivatives of the plant alkaloid, camptothecin. The indenoisoquinolines LMP400, LMP744, and LMP776 are novel noncamptothecin TOP1 inhibitors in clinical trial, which overcome the limitations of camptothecins. To further improve metabolic stability, their methoxy groups have been replaced by fluorine, as in the fluoroindenoisoquinolines NSC 781517 (LMP517), NSC 779135 (LMP135), and NSC 779134 (LMP134). We tested the induction and stability of TOP1 cleavage complexes (TOP1cc), and the induction and persistence of DNA damage measured by histone H2AX phosphorylation (γH2AX) compared with their parent compounds LMP744 and LMP776 in leukemia CCRF-CEM and colon carcinoma HCT116 cells. The fluoroindenoisoquinolines induced TOP1cc and γH2AX at nanomolar concentrations, and at higher levels than the parent indenoisoquinolines. The fluoroindenoisoquinoline LMP135 showed greater antitumor activity than topotecan in small-cell lung cancer cell H82 xenografts. It was also more potent than topotecan in the NCI-60 cancer cell line panel. Bioinformatics tools (http://discover.nci.nih.gov/cellminercdb) were used to investigate the following: (i) the correlations of fluoroindenoisoquinolines activity with other drugs, and (ii) genomic determinants of response in the NCI-60. The activity of the fluoroindenoisoquinolines was mostly correlated with camptothecin derivatives and the parent indenoisoquinolines, consistent with TOP1 targeting. Genomic analyses and activity assays in CCRF-CEM SLFN11–deleted cells showed that SLFN11 expression is a dominant determinant of response to LMP135. This study shows the potential value of the fluoroindenoisoquinolines for further development as novel anticancer agents targeting TOP1. Mol Cancer Ther; 17(8); 1694–704. ©2018 AACR.


Introduction
Topoisomerases relax DNA supercoils and resolve DNA knots and intertwining (catenanes) generated during DNA replication, transcription, repair, and chromatin remodeling (1). To do so, they induce transient breaks (topoisomerase cleavage complexes; TOPcc) in the DNA backbone that allow one of the strands to rotate around the other (TOP1) or the passage of another DNA segment through the break (TOP2). TOP1 enzymes (TOP1 and TOP1MT) induce single-strand breaks and TOP2 enzymes (TOP2a and TOP2b) induce double-strand breaks (1). Immediately after DNA relaxation or strand passage, topoisomerases ligate the broken DNA, restoring the DNA sequence and ensuring DNA integrity. Among DNA-targeted therapies, topotecan and irinotecan are widely used and effective in a broad range of tumors. They selectively trap TOP1 cleavage complexes (TOP1cc), blocking the religation of the TOP1-linked DNA single-strand breaks. Collisions with DNA replication forks convert the singlestrand breaks into double-strand break lesions that ultimately lead to cell death and anticancer activity (2,3).
The only chemical classes of FDA-approved TOP1 inhibitors are irinotecan and topotecan, which are water-soluble derivatives of the plant alkaloid, camptothecin (CPT). Irinotecan is widely prescribed for colorectal and gastric cancers, and topotecan for ovarian, cervical, and small-cell lung cancers. Yet, CPT derivatives suffer from well-documented limitations: (i) chemical instability of their E-ring, leading to the inactive carboxylate derivatives that bind tightly to serum albumin (4-6); (ii) rapid elimination due to short plasma half-life; (iii) highly reversible DNA damage due to their rapid dissociation from the TOP1cc (5, 7); (iv) active efflux from multidrug-resistant cancer cells expressing ABC transporters (ABCG2 and ABCB1); (v) gastrointestinal toxicity (including severe diarrhea for irinotecan); and (vi) dose-limiting bone marrow toxicity (6,8,9). On the basis of phenotypic COMPARE analysis across the 60 cancer cell lines of the NCI (NCI-60; refs. 10-13), we identified a novel chemical class of TOP1 inhibitors: the indenoisoquinolines (10,14). Lead optimization led to highly potent and specific TOP1cc-targeted indenoisoquinolines that overcome most of the limitations of the camptothecins. Indeed, the indenoisoquinolines are chemically stable, form less reversible TOP1cc, have long plasma half-life, are not substrates for the ABC drug efflux transporters and do not produce diarrhea (6,9,15). Three indenoisoquinolines, LMP400 (indotecan, NSC724998), LMP776 (indimitecan, NSC725776), and LMP744 (MJ-III-65, NSC706744) have entered clinical trials. LMP400 recently completed Phase I with demonstrable target engagement (15), and is now poised for phase II clinical trials. LMP776 is finishing phase I and LMP744 is beginning phase I clinical trials. Because the methoxy groups of LMP400, LMP776, and LMP744 ( Fig. 1) are likely to be metabolized in vivo by O-dealkylation catalyzed chiefly by hepatic cytochrome P450 enzymes (16)(17)(18), a common metabolic process that, for example, plays a significant role in human metabolism of etoposide and teniposide (19), the aim of the this study was to determine whether replacing the methoxy groups by fluorine would generate novel fluoroindenoisoquinolines retaining potent TOP1cc-targeting and antiproliferative activity in cells, thereby making them worthy of development as anticancer agents. Our recent studies have shown that addition of fluorine in position 3 yields fluoroindenoisoquinolines that retain potent activity on TOP1cc in biochemical assays with recombinant human TOP1 (17,18,20).

Cell lines and reagents
Human leukemia CCRF-CEM and colon carcinoma HCT116 cell lines were cultured in RPMI medium (Gibco) supplemented with 10% FBS. HCT116 cell lines were provided by the NCI-Frederick Division of Cancer Treatment and Diagnosis tumor/cell line repository. CCRF-CEM cells were obtained from the NCI Developmental Therapeutics Program. The SLFN11deleted CCRF-CEM cells by CRISPR/Cas9 have been recently described (24,28). H82 cells were obtained from the NCI repository. All cell lines were kept for 45 days maximum after thawing and tested for mycoplasma with MycoAlert Mycoplasma Detection Kit (Lonza). CPT, topotecan, LMP400, LMP744, and LMP776 were provided by the NCI Drug Developmental Therapeutics Program. LMP517, LMP135, and LMP134 were synthesized in the Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University (West Lafayette, IN;refs. 20,29).

In vitro TOP1cc trapping
A 3 0 -[ 32 P]-labeled 117-bp DNA substrate oligonucleotide was prepared as described previously (30). Radiolabeled DNA was incubated with recombinant human TOP1 in 20 mL reaction buffer (10 mmol/L Tris-HCl, pH 7.5, 50 mmol/L KCl, 5 mmol/L MgCl 2 , 0.1 mmol/L EDTA, and 15 mg/mL BSA) at 30 C for 20 minutes in the presence of the indicated drug concentrations. Reactions were terminated by adding SDS (0.5% final concentration) followed by the addition of two volumes of loading dye (80% formamide, 10 mmol/L sodium hydroxide, 1 mmol/L sodium EDTA, 0.1% xylene cyanol, and 0.1% bromophenol blue). Aliquots of reaction mixtures were subjected to 16% denaturing PAGE. Gels were dried and visualized by using Phos-phorImager and Image Quant software (Molecular Dynamics).

DNA-protein cross-links (DPC)
Alkaline elution was used to assess cellular TOP1cc by measuring DPC as described (4,(31)(32)(33). Before alkaline elution and drug treatments, genomic DNA of CCRF-CEM cells was radiolabeled with 0.02 mCi/mL [ 3 H]-thymidine for one to two doubling times at 37 C and chased in nonradioactive medium overnight. Cells were treated with the indicated concentrations of LMP compounds or CPT for 1 hour before scraping and alkaline elution. For reversal experiments, the cells were cultured in drug-free medium for the indicated times. Radioactivity in all fractions was measured with a liquid scintillation analyzer (PerkinElmer Life and Analytical Sciences), and DPC frequency, which reflects TOP1cc was determined as published (4,(31)(32)(33).

ICE bioassay
TOP1cc were detected using the immunocomplex of enzyme bioassay (34)(35)(36). In brief, treated or untreated CCRF-CEM cells were pelleted and immediately lysed in 1% sarkosyl. After homogenization with a Dounce homogenizer, cell lysates were gently layered on CsCl step gradients and centrifuged at 165,000Âg for 20 hours at 20 C. Half-milliliter fractions were collected, diluted with an equal volume of 25 mmol/L sodium phosphate buffer, pH 6.5, and applied to Immobilon-P membranes (Millipore Corporation) through a slot-blot vacuum manifold as described (36). TOP1cc were detected using the C21 TOP1 mouse antibody (BD Pharmingen catalog number 556597). Cellular DNA damage measured by gH2AX staining HCT116 cells were plated at 5 Â 10 4 cells per well in 4-well chamber slides, incubated for 24 hours and then treated with drugs. Cells were washed with PBS, fixed and permeabilized with absolute ethanol overnight at 4 C, blocked for 1 hour in PBS-BSA 8% at room temperature, and incubated for 2 hours at room temperature with primary antibody. 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). Images were captured with a confocal microscope (Nikon PCM2000). The primary antibody for histone gH2AX was mouse monoclonal (Millipore # 05-636: anti-gH2AX Ser139 antibody, clone JBW301) and the secondary antibody was a donkey anti-mouse conjugated to Alexa Fluor 568 from Molecular Probes. Signals were quantified using ImageJ program. An area was selected (same size) and the signal was measured in each nucleus, giving the intensity in arbitrary unit.

Cellular viability assays
CCRF-CEM parental and CCRF-CEM SLFN11-KO cells were seeded at 5,000 cells per well in 96-well white plates (#6005680 Perkin Elmer Life Sciences), and exposed to the indicated concentrations of topotecan, LMP517, LMP135, or LMP134 for 72 hours in triplicate. Cellular viability was determined using ATPlite 1-step kits (PerkinElmer). ATP levels of untreated cells were defined as 100%. Percentage viability was defined as: [(ATP in treated cells)/(ATP in untreated cells)]Â100 (28).
In vivo study in nude mice Athymic nude mice (nu/nu, female, 20-25 gram, 8-12 weeks old) from Charles River were transplanted with 5 million H82 human small-cell lung cancer cells. 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 topotecan and vehicle arm, and 6 mice for LMP135 were used. The animals were treated with either topotecan (1.5 mg/kg) administered intraperitoneally or with LMP135 (20 mg/kg) administered by intravenous push via tail vein once a day for 5 consecutive days. LMP135 was dissolved in 10 mmol/L citric acid and 5% dextrose. Topotecan was diluted in sterile water. The three axes (millimeters) of tumors were measured with a caliper to calculate tumor volume. Measurement was 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. Animal studies were approved by the Animal Care and Use Committee of the NCI-Frederick, and all animal care was in accordance with institutional guidelines.

Pharmacokinetic study
Blood for pharmacokinetic measurements of each compound was obtained at 5 minutes, 30 minutes, 1 hour, 2 hours, and 8 hours postdose via cardiac puncture on euthanized mice and drawn into tubes containing sodium heparin. All animal handling was conducted in accordance with NIH Animal Care and Usage Committee (ACUC) regulations. Plasma was obtained through centrifuging the blood for 10 minutes at 1500Âg, then transferred to cryovials and stored at À80 C until bioanalysis. Each time point was performed in triplicate.
On the day of measurement, 50 mL of thawed plasma was mixed with 3Âvolume of acetonitrile, mixed to precipitate plasma proteins, and centrifuged. Ten microliters of the resulting supernatant were injected onto a Waters ACQUITY UPLC BEH C18 column (2.1 Â 50 mm, 1.7 mm) and the analytes of interest (LMP134, LMP135, and LMP517) were chromatographically separated using a gradient elution. The column eluent was directed into a triple quadrupole mass spectrometer operated in the positive ion mode and the analytes were detected on the basis of their unique mass fragmentation of m/z 384.0 ! 309.5 (LMP134), m/z 418.1! 350.1 (LMP135), and m/z 411.1 ! 350.1 (LMP517). The assay had a calibration range of 50-25,000 ng/mL and was validated according to FDA guidelines (https://www.fda. gov/downloads/drugs /guidances/mcm368107.pdf).
Pharmacokinetic parameters were estimated using noncompartmental analysis and the area under the plasma concentration versus time curve (AUC) was estimated using the linear trapezoidal rule with adjustments for destructive sample and the Bailer method for estimating variance (PMID 3221328; 7724473). The maximum plasma concentration (C MAX ) was calculated as the highest average (n ¼ 3) plasma concentration. The elimination rate (K EL ) was the slope of the line through the terminal natural log-transformed average plasma concentrations; half-life (t 1/2 ) was calculated as natural log 2/K EL . Clearance (CL) was calculated as dose/AUC; volume of distribution was calculated as CL/K EL . All parameter estimates were calculated using Phoenix WinNonlin v7.0 (Certara Pharsight).

Fluoroindenoisoquinolines efficiently trap TOP1cc in biochemical assays
To determine the potency of three fluoroindenoisoquinolines (LMP517, LMP135, and LMP134), we performed TOP1cc trapping assays in the presence of recombinant human TOP1 and DNA. All three compounds induced TOP1cc and showed comparable patterns of cleavage sites. Those cleavage patterns were comparable with the first-generation indenoisoquinolines, LMP400, LMP744, and LMP776 (Fig. 1B). LMP517 and LMP135 were the most potent TOP1cc inducers (Fig. 1B). Like its closest analogue LMP744 (14), LMP517 inhibited TOP1cc formation at high concentration of 100 mmol/L, an effect which has been related to the ability of the LMP744 side chain to enable DNA intercalation at high concentration (14). These results demonstrate that derivatives with replacement of the 2,3-dimethoxy groups of LMP744 and LMP776 with a single 3-fluoro substituent, as in LMP135 and LMP517, retain potent TOP1cc trapping activity in vitro.

Fluoroindenoisoquinolines efficiently inhibit cellular TOP1
To assess the trapping of TOP1cc in cells by LMP517, LMP134, and LMP135 in cells, two methods were used: ICE bioassay and alkaline elution (4,(31)(32)(33)(34)(35)(36). For the ICE bioassay, CCRF-CEM cells and HCT116 cells were treated for 1 hour with 1 mmol/L of the three fluoroindenoisoquinolines in parallel with CPT and LMP744. Cell lysates were fractionated by cesium chloride gradient and TOP1cc were measured in the DNA-containing fractions (fractions 5-8; Fig. 2A). TOP1cc were detected in the treated but not in the untreated cells, demonstrating that all three fluoroindenoisoquinolines induced cellular TOP1cc.
The alkaline elution assay quantifies TOP1cc as DNA-protein crosslinks (DPC; ref. 4). CCRF-CEM cells were treated for 1 hour with a range of drug concentrations of LMP517, LMP134, and LMP135, and in parallel with similar concentrations of CPT, LMP744 (closest analogue of LMP517) and LMP776 (closest analogue of LMP135). All drugs induced DPC in a concentration-dependent manner (Fig. 2B). LMP517 was the most potent and compared favorably with LMP744. Also, the three fluoroindenoisoquinolines induced 1.5 to 4 times more DPC than CPT, LMP744, and LMP776 at 1 mmol/L (Fig. 2B). Together, the results of the ICE bioassay and alkaline elution experiments demonstrate that the fluoroindenoisoquinolines effectively target TOP1cc at nanomolar concentration in cells.

Fluoroindenoisoquinolines induce cellular DNA damage
Phosphorylation of histone H2AX on serine 139 (gH2AX) is a chromatin modification induced when DNA double-strand breaks are formed following DNA damage (37). Because CPT, topotecan, and the indenoisoquinolines LMP400, LMP776, and LMP744 induce this phosphorylation (15,33,38,39), we examined gH2AX signal in cells treated with LMP517, LMP135 or LMP134. Figure 3A  Experiments were also performed to determine drug concentration dependency. LMP517 was more potent than topotecan at 1 mmol/L (Fig. 3C), and LMP135 was more potent than LMP744 and LMP134 at their lowest concentration (0.1 mmol/L) with respect to gH2AX induction (Fig. 3D). All drugs induce DNA damage at submicromolar concentrations.
These results demonstrate that the three fluoroindenoisoquinolines rapidly induce DNA damage response and that LMP517 and LMP135 are superior to the other drugs tested with respect to gH2AX induction.
Persistent TOP1cc and DNA damage in response to the fluoroindenoisoquinolines Next, we examined the reversal kinetics of the TOP1cc and gH2AX induced by the fluoroindenoisoquinolines using alkaline elution and immunofluorescence microscopy (Fig. 4). CCRF-CEM cells were treated for 1 hour with 1 mmol/L CPT, LMP744, LMP776, LMP517, LMP134, or LMP135. After drug removal, cells were grown in fresh drug-free medium for an additional hour. DPC induced by LMP517 and LMP135 remained at 75% and 65% of initial response, respectively, at 1 hour after drug removal (Fig. 4A). In contrast, 2% of the DPC induced by CPT and 15% of the LMP134-induced DPCs remained at 1 hour after drug removal (about 50% for the indenoisoquinolines LMP744 and LMP776).
For DNA damage detection using gH2AX, HCT116 cells were treated with 1 mmol/L of drug for 1 hour and then grown in fresh drug-free medium for 6 hours (Fig. 4B). gH2AX signal was reversible for all drugs. The DNA damage signals elicited by LMP517 and LMP135 were more persistent than with the other drugs tested with approximately 80% of gH2AX signal remaining after drug removal. In contrast, about 35% of the topotecan-, 45% of LMP744-, and 40% of LMP134-induced gH2AX signals remained 6 hours after drug removal (Fig. 4B). These results demonstrate the persistence of the TOP1cc and DNA damage response (gH2AX) induced by LMP517 and LMP135.

Fluoroindenoisoquinolines behave like TOP1 inhibitors in the NCI-60 panel, and LMP135 is the most potent compound
To determine where the new fluoroindenoisoquinolines stand with respect to their antiproliferative activity in comparison with validated TOP1 inhibitors, we used the NCI-60 phenotypic data that monitor drug growth-inhibitory 50% concentrations (GI 50 ; refs. 11-13). Figure 5A represents the distribution of GI 50 values of LMP517, LMP135, and LMP134 in comparison with topotecan, LMP744, and LMP776 in the NCI-60 cell lines. Figure  all the fluoroindenoisoquinolines showed better potency than topotecan (258 nmol/L for topotecan) and the parent indenoisoquinolines (312 nmol/L for LMP744). Next, the COMPARE analysis (13) of CellMiner (11) was used to test whether the fluoroindenoisoquinolines exhibit a TOP1 inhibitor phenotype across the NCI-60. All fluoroindenoisoquinolines were most highly correlated with camptothecin and its derivatives and other indenoisoquinolines across the NCI-60 drug database, which encompasses over 20,000 publicly available compounds (see Supplementary Tables S1 and S2 for the complete dataset; ref. 11). Figure 5B shows the high potency of the most potent cellular inhibitor, LMP135, in comparison with topotecan (10-fold lower GI 50 than topotecan), and its highly significant correlation with topotecan across the NCI-60 (correlation r ¼ 0.88; P < 0.001). These analyses demonstrate that the fluoroindenoisoquinolines behave as prototypical TOP1 inhibitors across the NCI-60 cancer cell lines, and that they are more potent than topotecan.
LMP135 shows better tumor response than topotecan in H82 small-cell lung cancer xenografts As topotecan is FDA-approved for small-cell lung cancer (SCLC), we used SCLC H82 cell line as xenograft model to test the antitumor activity and tolerability of LMP135 in comparison with topotecan. Topotecan MTD was 1.5 mg/kg and LMP135 was 20 mg/kg. Nude athymic mice were inoculated with H82 cells. After tumors reached 100 mm 3 , the mice were treated with either topotecan at its MTD (1.5 mg/kg; ref. 38) or with LMP135 at its MTD (20 mg/kg). A single cycle of 5 days of LMP135 was sufficient to inhibit tumor growth with low toxicity (13% body-weight loss; Fig. 5C and D; Supplementary Fig. S1). LMP135 inhibited tumor growth until day 10, while topotecan inhibited tumor growth only until day 7, leading to a quicker growth. Mouse body weight loss was about 13% for LMP135 and 5% for topotecan ( Supplementary Fig. S1). These results show that LMP135 is a potentially promising molecule for development as an anticancer drug.
We also determined the pharmacokinetics of the fluoroindenoisoquinolines. The results in Supplementary Table S3 show a half-life of 2.5 hours and a volume of distribution of 452 mL for LMP135, which compares favorably with its parent compound LMP776 (T1/2 ¼ 1.4 hours and Vz ¼ 60 mL; ref. 40).
Fluoroindenoisoquinolines are selectively active in SLFN11proficient cells NCI-60 CellMiner (11,41; available at http://discover.nci.nih. gov/cellminercdb) was used to determine the genomic determinants of response to the fluoroindenoisoquinolines. Using the "Multivariate analyses" tool ( Fig. 6A), we found that the top genomic determinant was SLFN11 (Schlafen 11; see also Supplementary Table S4 for the complete dataset), which is a recently established determinant of response to TOP1 inhibitors, PARP inhibitors, hydroxyurea, gemcitabine, and other widely used therapeutic agents targeting DNA replication (21-24, 26, 28, 42). Figure 6A also shows the lack of correlation between the activity of LMP135 and two genes encoding most prominent drug efflux transporters, ABCB1 (encoding P-glycoprotein PgP) and ABCG2 [encoding MXR (mitoxantrone resistance protein) and known to confer resistance to topotecan; ref . 21]. The correlation between sensitivity to LMP135 and SLFN11 expression in the NCI-60 cell lines was highly significant (Fig. 6B). Yet there are several outliers that are highly sensitive to LMP135 in spite of low SLFN11 expression, indicating that additional genomic determinants influence the response of cancer cells to LMP135.
To determine the causality between SLFN11 and the activity of the fluoroindenoisoquinolines, we tested the response to the fluoroindenoisoquinolines in SLFN11-proficient CCRF-CEM and isogenic SLFN11-deleted CCRF-CEM cells (Fig. 6C; Supplementary Fig. S2; refs. 24, 28). Cells were grown in the presence of increasing concentrations of LMP517, LMP134, and LMP135 for 72 hours for viability analysis and topotecan was used as a control (21,22). The viability results show that lack of SLFN11 expression confers high resistance to all the drugs. The IC 90 of LMP135 was not reached in SLFN11-deleted cells under conditions where it was about 25 nmol/L in CCRF-CEM parental cells (Fig. 6C)  A, Quantitative analysis of TOP1cc after drug removal in CCRF-CEM cells. Cells were treated with 1 mmol/L CPT, LMP744, LMP776, LMP517, LMP134, or LMP135 for 1 hour at 37 C. Drugs were removed, and cells were grown for another hour in drug-free medium (R1). TOP1cc were determined as DPC by alkaline elution. Percent reversal for each drug is indicated in the text. B, Quantitative analysis of gH2AX persistence 6 hours after drug removal. HCT116 cells were treated for 1 hour with 1 mmol/L drug concentration, and either fixed at that point (1 hour) or grown for an additional 6 hours in drug-free medium (R). gH2AX signal intensities of 50 cells (each individual cell represented as a dot/circle) were measured by ImageJ. Same sized area signal was quantified in each nucleus. Percentage reversal for each drug is indicated in the text. Points below the horizontal dotted lines are gH2AX signals within background signal for untreated cells.
were 120 nmol/L and 55 nmol/L, respectively. In SLFN11-deleted cells, they were much higher (640 nmol/L and 300 nmol/L, respectively, Supplementary Fig. S2). These results demonstrate that SLFN11 could be a useful determinant of response for the clinical development of the fluoroindenoisoquinolines.

Discussion
Our main conclusions are: (i) the fluoroindenoisoquinolines represent a novel chemical class of potent TOP1 inhibitors; (ii) the fluoroindenoisoquinoline LMP135 exerts better antitumor activity compared with topotecan in the small-cell lung cancer H82 xenograft model; (iii) SLFN11 expression is a prominent determinant of response to the fluoroindenoisoquinolines, and lack of SLFN11 expression predicts resistance; and (iv) histone gH2AX can be used as a clinical biomarker of response (target engagement; refs. 15,43) for the fluoroindenoisoquinolines.
It is legitimate to develop noncamptothecin TOP1 inhibitors because of the established limitations of camptothecins and because TOP1 is a validated target for a broad range of cancers including colorectal, gastric, ovarian, lung, endometrial, and pediatric cancers (9,44). The indenoisoquinolines overcome most of the limitations of the camptothecins (see Introduction) and three derivatives are in Phase 1-2 clinical trials (LMP400, indotecan; LMP776, indimitecan and LMP744). The fluoroindenoisoquinoline derivatives (20,29) presented are potent and selective TOP1 inhibitors as they induce TOP1cc at nanomolar concentrations in the presence of recombinant human TOP1 enzymes (Fig. 1) and in cells (Figs. 2 and 4). They also induce DNA damage (gH2AX; Figs. 3 and 4), and their cellular activity is highly correlated with the established TOP1 inhibitors CPT and topotecan in the NCI-60 (Fig. 5) Comparison of the NCI-60 responses to the indicated TOP1 inhibitors. Each dot represents the growth-inhibitory 50% concentration (GI 50 ) of a given NCI-60 cell line for the indicated drug. Data were obtained from the NCI Developmental Therapeutics Program (DTP). Average GI 50 for all 60 cell lines is indicated below the graph. LMP744 and LMP776 were not tested under 10 nmol/L and LMP135 under 5 nmol/L, and in those cell lines in which the GI 50 was less than the minimum concentration tested, the GI 50 values were recorded as 10 and 5 nmol/L, respectively. The "averages" in these cases therefore represent mean-graph midpoint (MGM) values rather than true GI 50 averages. B, Comparison of the sensitivity of the NCI-60 cells to LMP135 versus topotecan. GI 50 of each drug for each cell line is indicated as well as equipotency curve. LMP135 is approximately 10-fold more potent than topotecan. C and D, Antitumor activity of LMP135 (20 mg/kg) compared with topotecan (1.5 mg/kg) represented as tumor volume (C) or as Kaplan-Meyer curves (D). Bars, SDs (n ¼ 10 for vehicle and topotecan; n ¼ 6 for LMP135). and 5), persistent TOP1cc and gH2AX response (Fig. 4) and lack of O-demethylation (29). The fluorine substituent is likely to affect their distribution and pharmacokinetics. LMP135 showed higher potency than its parent compound LMP776 (see Fig. 1A), and is also more potent than topotecan in the NCI-60 and in tumor model (Fig. 5). Hence, LMP135 might be worthy of consideration for further development as a novel anticancer agent including its use for tumor-targeted delivery using antibody-drug conjugates (ADC), liposomes, or PEGylation. Consistent with previous observations with camptothecins (21-23, 26, 28, 43), SLFN11 expression was the most significant genomic determinant of response to the fluoroindenoisoquinolines ( Fig. 6; Supplementary Table S3). Lack of SLFN11 was also demonstrated as a causal determinant of resistance to the indenoisoquinolines, as well as topotecan (21,22,42,43). Hence, it should be important to evaluate SLFN11 in clinical samples from patients treated with TOP1 inhibitors to establish the prognostic value of SLFN11 and its role in tumor resistance when tumors suppress its expression (23,42,45,46). IHC studies are ongoing to complement the genomic determination of SLFN11 expression, and it appears logical to use SLFN11 expression as a predictive biomarker for the development of the indenoisoquinolines. In addition, our study suggests that gH2AX could be applied to determine target engagement in response to the fluoroindenoisoquinolines in patient samples (15,(37)(38)(39).

Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.

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.

Acknowledgments
Our studies are supported by the Center for Cancer Research, the Intramural Program of the NCI (Z01-BC006161 to Y. Pommier). We wish to thank the Developmental Therapeutics Program (DTP) of the NCI for the cellular testing of the indenoisoquinolines in the NCI-60, and the CellMiner Genomics and Pharmacology Facility (CGPF) of the Developmental Therapeutics Branch, CCR, NCI (to William Reinhold) for the development of pharmacogenomic tools (http://discover.nci.nih.gov/cellminer) for our analyses of the NCI-60. The xenograft studies were supported and performed by the NCI Drug Development Collaborative (DDC). This project has been funded in whole or in part with Federal funds from the NCI and NIH, under contract no. HHSN261200800001E (to Y. Pommier). 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