D-Ring-Modified Analogues of Luotonin A with Reduced Planarity: Design, Synthesis, and Evaluation of Their Topoisomerase Inhibition-Associated Cytotoxicity

A- and D-ring-modified luotonin-inspired heterocycles have been synthesized and were evaluated for their activity against the viability of four cancer cell lines in vitro, namely, MCF7, HCT116, JURKAT, and NCI-H460. The analysis of results indicated that two of the synthesized derivatives displayed good inhibition against the growth of the human colon cancer HCT116 cell line, with potencies lower than but in the same order of magnitude as camptothecin (CPT). These two luotonin analogues also showed an activity similar to that of the highly potent alkaloid CPT as inhibitors of topoisomerase I and also inhibited topoisomerase II. These results show that complete planarity is not a strict requirement for topoisomerase inhibition by luotonin-related compounds, paving the way to the design of analogues with improved solubility.


Introduction
Cancer and related diseases are mainly caused by a number of genetic, environmental, and lifestyle-associated factors and the disease is characterized by a shift in the controlled mechanisms that govern cell proliferation and differentiation, and normal physiological activity [1]. Cancerous diseases are life-threatening and continue to be the leading causes of death, surpassing cardiovascular disease. In 2012, there were 14.1 million new cases of cancer worldwide and 8.2 million cancer-related deaths, and it is estimated that about 32.6 million patients have survived after 5 years of a cancer diagnosis [2]. erefore, the development of novel, more effective anticancer agents with good pharmacokinetic profiles remains an important and challenging goal in medicinal chemistry [3]. e NIH (National Institutes of Health) database describes a large number of anticancer compounds, organized according to their targets and mechanisms of action [4]. Polycyclic nitrogen heterocycles are interesting pharmacophores in this regard as they can interfere with the functioning of several DNA-associated enzymes, including the topoisomerases, in a process that is normally initiated by compound intercalation between the base pairs of double-stranded DNA [5].
e main drugs that have topoisomerase I as their target are camptothecins (CPTs) (Figure 1) but they suffer from several limitations in spite of being in clinical use. In the first place, the δ-lactone moiety in the E-ring of these compounds is highly sensitive to hydrolysis [20]. Furthermore, CPT has very poor solubility in water, which led to the introduction of basic substituents into topotecan and irinotecan, in order to allow the compounds to be formulated as salts. Dual inhibition of topoisomerases I and II by a single agent leads to compounds that are able to target a larger population of cells in comparison with selective inhibitors and increase antitumor activity and has been proposed as a worthwhile goal in the discovery of anticancer agents [21].
Luotonin A, an alkaloid from Peganum nigelastrum, can be viewed as a hydrolysis-stable CPT analogue. While both compounds share the same mechanism of action, namely, stabilization of the topoisomerase I− DNA complex, luotonin shows a much lower level of activity. e importance of the development of more potent luotonin A analogues has prompted much synthetic work [22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37], and these efforts have furnished several types of structural variations of the reference alkaloid. Among them, due to limitations in synthetic methodology, the least explored one is the manipulation of the D-ring of luotonin; in this regard, only two examples of D-ring-contracted luotonin analogues have been described [26,37].

Crystal Structure Preparation.
e starting point for the docking study was the crystal structure of the topoisomerase I-DNA covalent complex bound to topotecan (PDB code 1K4T) [38], the only clinically used camptothecin derivative that has been crystallized in its site of action. A molecule of polyethyleneglycol, an atom of Hg, and the lactone and hydroxyl acid forms of topotecan were manually removed from the crystal. Water molecules were then deleted using the Dockprep tool of Chimera [39], which was also employed to add hydrogen atoms to the enzyme and DNA molecules, repair truncated side chains via the Dunbrack rotamer library [40], and assign Gasteiger charges using the AMBER ff12SB force field [41]. From the resulting structure, AutoDock Tools 1.5.6 was used to generate the input file for docking [42], and the docking site for this study was defined as a box with dimensions 15 × 15 × 20Å. e centroid of this box was calculated using the coordinates of the closed lactone form of topotecan (x � 21.377, y � − 4.068, z � 28.192).

Ligand Preparation.
e ligands were prepared by ab initio energy minimization with Spartan'10 (3-21G level). Hydrogen atoms were then added, and the root of the torsion tree was detected. As in the previous case, the input file for the docking experiments was created with AutoDock Tools 1.5.6.

Docking Studies.
We employed AutoDock Vina [43] for the docking studies, using the parameters described above. For the validation of the protocol, we used the closed lactone form of topotecan as the ligand for one of the experiments and compared its most stable docking pose with the crystal structure, obtaining a RMSD value of 0.492Å. [44]. Melting points of all the compounds were measured by means of open capillary tubes and are uncorrected. 1 H, 13 C, and two-dimensional nuclear magnetic resonance spectra were recorded on a Jeol Instruments spectrophotometer (400 and 500 MHz) in CDCl 3 using TMS as the internal standard. Chemical shifts are given in ppm (δ-scale), and the coupling constant values are given in Hertz. Mass spectra were recorded on a Quattro Premier triple quadrupole mass spectrometer equipped with an electrospray ionization source (Z-spray) coupled with an Acquity UPLC system. Elemental analyses were performed on a PerkinElmer 2400 Series II elemental CHNS analyzer. (2). e procedure was adapted from the literature [45]. A solution of (±)-phenylalanine (1) (2.0 g, 12.12 mmol) in HBr (10 mL) was heated to 40°C. Formaldehyde (1.8 mL, 48.86 mmol) was then added dropwise. e reaction mixture was further heated to 70-80°C for 3 h, during which a white precipitate formed. e reaction mixture was cooled to room temperature, and the precipitate was filtered under vacuum and washed with cold ethanol. e white solid product thus obtained was left under vacuum to dry. Yield � 90% (1.79 g) and melting point � 295-300°C (decomposes).
e reaction mixture, after being stirred at room temperature for 12 h, was filtered under an inert gas atmosphere through celite with dry CH 2 Cl 2 as the eluent. e collected filtrate was concentrated to 20 mL, and at − 78°C, BF 3 ·O(C 2 H 5 ) 2 (0.110 mL, 3.75 mmol) was added. en, the reaction mixture was allowed to warm to room temperature and was stirred further for 2 days. e reaction mixture was  extracted with CH 2 Cl 2 (3 × 30 mL), and the obtained residue was purified by flash column chromatography. [6,7]indolizino [1,2- 13

Cell Cultures.
Four different cell lines, namely, MCF7 (human breast cancer), HCT116 (human colon cancer), NCI-H460 (human non-small-cell lung cancer), and JURKAT (human T-cell lymphoma) cell lines, were obtained from NCCS, Pune, India. All the cell lines were cultured in DMEM (Dulbecco's modified Eagle's medium), except JURKAT, which was cultured in RPMI (Roswell Park Memorial Institute) media supplemented with 10% FBS (fetal bovine serum) (Himedia, India), containing penicillin (100 I.U/mL) and streptomycin (100 μg/mL) at 37°C in a humidified atmosphere of 5% CO 2 in air. For compound testing, the adherent cells were trypsinized, counted, and seeded in 96-well plates for viability studies at a density of 20 × 10 3 cells per well, where they were left to adhere overnight before the experiments with our compounds. e compounds under assay were not added until the wells were observed to show at least 80% confluence [46].

Cytotoxicity Assay.
e cytotoxic behaviour of compounds 8a-i on adherent cells was determined by the MTT (i.e., 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) reduction assay. Prior to treatment (12-14 h), the cells were seeded in a 96-well plate at a density of 20 × 10 3 cells per well. e cell media contained in the wells were discarded, and the cells were treated with compounds 8a-i at 125 μM concentration. is concentration was chosen because it guaranteed all compounds to show at least some cytotoxicity to allow comparison.

Topoisomerase I Mediated DNA Relaxation Assay.
e inhibitory effect of compounds 8f and 8h on the activity of human topoisomerase I was assayed by means of a DNA relaxation assay. To this end, a final reaction volume of 30 μL containing the reaction buffer, the pGEM-5Zf(+) plasmid (Promega) in its native supercoiled form (200 ng), and the compounds under assay (the reference drug CPT and compounds 8f and 8h) were diluted in 1 mL DMSO to achieve their IC 50 concentrations, i.e., 40 μM, 68 μM, and 57 μM, respectively, and 1 unit of human topoisomerase I (Inspiralis) was incubated at 37°C for 30 min. e reaction corresponding to the native supercoiled form was carried out from a mixture containing the reaction buffer and plasmid (lane 1, Figure 2(a)) and that corresponding to the fully relaxed form was carried out from the reaction buffer, plasmid, and human topoisomerase I (lane 2, Figure 2(a)). For the remaining three experiments, the compounds assayed were also included (lanes [3][4][5]. Following incubation, the reactions were stopped by the addition of a mixture of the buffer employed for the subsequent electrophoresis and 20% SDS. e reaction products were loaded onto 1% agarose gel (GLPL), and an electrophoresis experiment was run at 4°C for 3 h at 100 V in Tris-borate-EDTA buffer. e gels were stained with ethidium bromide from Sigma, using the manufacturer's protocol. en, they were photographed under 254 nm UV light using a gel doc (UVitech) system, and the bands were analyzed using the inbuilt software. e inhibitory activities of the compounds under assay were calculated using the ratio of intensities of supercoiled to unwinded bands and normalized to the activity of CPT.

Statistical Analysis.
All the in vitro results given here were calculated as mean ± SD of at least three replicate analyses. e statistical analysis of the results was based on Student's t-test. Statistical significance was set at p < 0.05 (represented by * ), and p < 0.01 (represented by * * ) was considered a highly significant value. ese calculations were performed with the aid of Graphpad Prism software (version 6).

Library Design.
e aforementioned limitations in existing topoisomerase I inhibitors prompted us to study the synthesis and biological evaluation of D-ring-modified luotonin analogues. Because the low solubility of CPT derivatives can be ascribed to the high stability of their crystalline lattices in the solid state, which is due to strong intermolecular interactions between their planar polycyclic frameworks, we decided to modify the polycyclic system by replacing the D-ring pyrimidinone unit with a piperidine unit, as shown in Figure 3. is structural modification leads to increased three-dimensionality and hence is expected to be associated with the disruption of crystal packing via decreased planarity, which is a well-established, if relatively unexplored, strategy for increasing the solubility of organic compounds [48]. Furthermore, the proposed modification has the advantage of allowing the easy formation of salts at the strongly basic N-12 cyclic tertiary amine nitrogen.
In order to address the potential negative effects of diminished planarity on the efficiency of the DNA intercalation process, we carried out docking studies of the unsubstituted tetrahydrobenzo [6,7]indolizino [1,2-b]quinoline framework (compound 8a) onto the DNA-topoisomerase I covalent complex, which is generated in the course of the enzyme catalytic cycle and is the target of the CPT and luotonin. is study was carried out on the crystal structure of topotecan bound to the same complex, which is available in the Protein Data Bank (PDB 1K4T) [38]. Figure 4 shows the results for the docking of 8a compared with those of luotonin and shows that in spite of the loss of planarity, 8a maintains the interaction with the Arg-364 residue, with a shorter binding distance than luotonin, together with π− π stacking interactions (Figures 4(a) and 4(b)). While the binding free energy was somewhat lower than that of luotonin (− 10.2 vs. -11.4 kcal/mol), we hoped to compensate for this difference by the introduction of substituents allowing more efficient binding modes. For instance, the presence of a 2-methyl substituent (compound 8e) hampers the usual binding but the overall energy is slightly improved (− 10.9 kcal/mol) even though the interaction with Arg-364 is lost because two very efficient π-cation interactions are now possible (Figure 4(c)). On the other hand, the 2-chloro derivative 8h retains the binding mode of the parent compound because it allows a halogen bond with Glu-356 and a weak cation-dipole interaction with T-1(C�O), leading again to an improved binding energy (− 10.9 kcal/mol) (Figure 4(d)).

Synthesis.
e synthesis of luotonin analogues 8 was scheduled via a Pictet-Spengler/N-propargylation/intramolecular Povarov sequence of reactions. First, the compound 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (2) was prepared in an excellent yield (90%) from (±)-phenylalanine (1) via Pictet-Spengler condensation. e reaction of compound 2 with thionyl chloride afforded the carboxylate derivative (3) [45], which upon N-alkylation with propargyl bromide furnished compound 4. An excellent yield of compound 5 was achieved through sodium borohydride reduction of 4. Swern oxidation of compound 5 employing DMSO-oxalyl chloride furnished the carbaldehyde 6 in an excellent yield. Finally, the reaction of carbaldehyde 6 with substituted arylamines in the presence of 4Å molecular sieves afforded the target luotonin analogues 8 in moderate yields through intramolecular Povarov reaction (Scheme1). e synthetic part of this work has been published recently as a preliminary communication [44]. e structural elucidation of compounds 4-6 and 8 was performed by NMR studies. A reasonable mechanism for the formation of 8 involves the reaction of aldehyde 6 with arylamine to furnish arylimine 7 and a formal [4 + 2] intramolecular cycloaddition followed by a 1,3-hydrogen shift and air oxidation steps to furnish 8.

Biological Studies.
A cytotoxicity study of the luotonin analogues 8a-i was performed in order to examine their anticancer activity in four human cancer cell lines, namely, MCF7 (breast cancer), HCT116 (colon cancer), JURKAT (Tcell leukemia), and NCI-H460 (non-small-cell lung cancer), with CPT as a positive control. e synthesized compounds were first tested at a concentration of 125 μM for two different incubation times of 24 and 48 h, and the results are shown in Figure 5. In general, the activity order of the tested compounds against four different cancer cells lines is MCF7 < NCI-H460 < JURKAT < HCT116. e loss in cell viability was generally higher with an increased incubation period, i.e., upon moving from 24 to 48 h. Following this initial exploration, all cells were treated with increasing concentrations of the compounds, and the obtained concentration values for 50% cell death (IC 50 ) are shown in Table 1 and also in Figure 6 for the specific case of the HCT116 cell line, which showed the highest sensitivity. It is interesting to note that in this cell line, two of the compounds (8f and 8h) were slightly more active than the CPT positive control, which represents a very good level of activity for a luotonin analogue. For this reason, these two compounds were selected for further mechanistic studies. We next investigated the toxicity of compounds 8a-i on the nontumoral cell line NCM60 (normal human colon mucosal epithelial cell line). For all compounds, no loss in the viability of cells was observed when tested for both time periods of 24 and 48 h in the concentration range of 25-125 μM, thereby indicating the nontoxic nature of our compounds towards nontarget cells. e rationale for the choice of this particular cell line was the fact that it corresponds to colon cells, as HCT116. Furthermore, one of the major dose-limiting toxic effects found for CPT derivatives in clinical use, such as irinotecan, is the frequent appearance of severe diarrhea [49], and thus, the study of the NCM60 cells provides useful toxicological information.
In order to study the mechanism of action of our compounds, their inhibitory activity against purified human topoisomerase I was investigated by means of a relaxation assay that was performed by simultaneous incubation of supercoiled plasmid, pGEM-5Zf (+) DNA, enzyme, and the two more active compounds (8f at 68 μM and 8h at 57 μM). e inhibitory activity of the compounds was compared to that of standard topoisomerase I poison CPT, the same reference used in the cell-based cytotoxicity assays.
An agarose gel showing the separation of DNA topoisomers after incubation of the plasmid in the presence of human topoisomerase I and the inhibitors under assay is shown in Figure 2(a). From the figure, it is clear that the compounds 8f and 8h (lanes 4 and 5) showed very good activity that was comparable to that of the CPTcontrol (lane 3). A quantitation graphic is given in Figure 2(b), again showing activity similar to that of CPT for both compounds, which represents an unusually high potency for a luotonin derivative.
As mentioned above, dual inhibition of topoisomerases I and II is an increasingly attractive goal in the design of anticancer agents. Furthermore, batracyclin, a fused quinazoline derivative with some structural resemblance to luotonin, has been characterized as a dual inhibitor of DNA topoisomerases I and II [50]. For these reasons, we also investigated the effect of compounds 8f and 8h on the decatenation of kDNA by human topoisomerase II. e appearance of kDNA monomers was monitored in the reaction in order to determine whether an open circular decatenated kDNA form or a closed circular decatenated kDNA form had developed. If Topo II retains its normal function, catenated kDNA would disappear and bands for open circular intermediates and closed circular decatenated kDNA would appear. In contrast, the absence of these bands indicates inhibition of the enzyme. As shown in Figure 7, compounds 8f and 8h at their IC 50 concentrations were found to inhibit Topo II. erefore, these compounds can be viewed as dual inhibitors of both topoisomerase enzymes. Compounds that target Topo II can be divided into two classes, namely, poisons and catalytic inhibitors [51,52]. e first group includes most drugs in clinical use and leads to an increase in the levels of Topo II-DNA covalent complexes. A second class of compounds inhibits topo II activity but does not increase the levels of enzyme-DNA complexes.  However, the precise mechanism by which our compounds act on their target cannot be presently established.

Conclusions
A boron trifluoride diethyl etherate-catalyzed intramolecular hetero-Diels-Alder reaction provided convenient access to novel tetrahydrobenzo [6,7]indolizino [1,2-b]quinolone derivatives, designed as nonplanar analogues of the natural topoisomerase I inhibitor luotonin A by means of computational studies. Some of these compounds displayed good cytotoxicity against four different human cancer cell lines (MCF7, HCT116, JURKAT, and NCI-H460). In particular, compounds bearing 3-CH3 (8f ) and 2-Cl (8h) substituents have a good cytotoxic effect on the HCT116 cancer cell line. Compounds 8f and 8h were also potent topoisomerase I and II inhibitors. ese results indicate that complete planarity is not a strict requirement for topoisomerase inhibition by luotonin-related compounds, paving the way to the design of analogues with improved aqueous solubility.

Data Availability
e data used to support the findings of this study are included within the supplementary information file.