Bicyclic Basic Merbarone Analogues as Antiproliferative Agents

Pyrimido-pyrimidine derivatives have been developed as rigid merbarone analogues. In a previous study, these compounds showed potent antiproliferative activity and efficiently inhibited topoisomerase IIα. To further extend the structure–activity relationships on pyrimido-pyrimidines, a novel series of analogues was synthesized by a two-step procedure. Analogues 3–6 bear small alky groups at positions 1 and 3 of the pyrimido-pyrimidine scaffold whereas at position 6a (4-chloro)phenyl substituent was inserted. The basic side chains introduced at position 7 were selected on the basis of the previously developed structure–activity relationships. The antiproliferative activity of the novel compounds proved to be affected by both the nature of the basic side chain and the substituents on the pyrimido-pyrimidine moiety. Derivatives 5d and 5e were identified as the most promising molecules still showing reduced antiproliferative activity in comparison with the previously prepared pyrimido-pyrimidine analogues. In topoisomerase IIα-5d docking complex, the ligand would poorly interact with the enzyme and assume a different orientation in comparison with 1d bioactive conformation.

Merbarone ( Figure 1) is a thio-barbituric catalytic TopoII inhibitor that blocks the proliferation of several cancer cell lines [12]. In clinical trials, merbarone showed nephrotoxicity issues and poor efficacy potentially ascribable to its high ionizability (and therefore poor bioavailability) at physiological pH [10,13,14]. With the aim of preparing merbarone analogues endowed with improved pharmacodynamic and pharmacokinetic properties, we previously synthesized compounds 1 (Figure 1) as conformationally constrained analogues in which the H-bonded pseudo-bicyclic structure of merbarone has been converted into a pyrimido-pyrimidine moiety [15,16]. In cell-based assays, derivatives 1 showed constrained analogues in which the H-bonded pseudo-bicyclic structure of merbarone has been converted into a pyrimido-pyrimidine moiety [15,16]. In cell-based assays, derivatives 1 showed interesting antiproliferative properties and the nature of the substituent at position 7 of the pyrimido-pyrimidine scaffold was found to deeply affect the biological activity. In particular, compounds 1a-e ( Figure 1) were identified as the most promising compounds with (sub)micromolar IC50 values against various cell lines (namely, MT-4, HeLa and MCF-7). Furthermore, selected members of this series were able to inhibit TopoIIα enzyme [16]. Owing to the interest in developing novel merbarone analogues [17,18] and in order to extend the structure activity relationships (SARs) of compounds 1, derivatives 3-6 ( Figure 2) were designed and synthesized. The novel derivatives were substituted at position 7 with the most promising side chain identified in the previous studies [15,16], and bear at positions 1 and 3 small alkyl chains (i.e., Me or Et groups). Moreover, the effect on activity of the 4-chloro substitution on the phenyl ring at position 6 was evaluated.

Results and Discussion
The synthesis of compounds 3-6 was carried out on the basis of previously developed procedures [15,16]. Briefly, the condensation of malononitrile with two equivalents of methyl or ethyl isothiocyanate in basic conditions led to the formation of the pyrimido derivatives 7 and 8 (Scheme 1). Differently from what was observed using phenyl isothiocyanate [15,16], neither the use of different bases (i.e., NaH, t-BuOK) or the use of three equivalents of alkyl heterocumulenes led to the isolation of the pyrimido-pyrimidine analogue. Derivatives 7 and 8 were then reacted, in a one-pot condensation, with phenyl or 4-chlorophenyl isothiocyanate and iodomethane to afford 9-12 (Scheme 1). Finally, the S-methylated compounds were converted into the final derivatives 3-6 by reacting with the proper amine (1.1. equivalents, Scheme 1). The different reactivity of 9-12 intermediates towards amines a-e required different reaction conditions and the desired compounds were obtained in moderate to good yields (Table 1). Owing to the interest in developing novel merbarone analogues [17,18] and in order to extend the structure activity relationships (SARs) of compounds 1, derivatives 3-6 ( Figure 2) were designed and synthesized. The novel derivatives were substituted at position 7 with the most promising side chain identified in the previous studies [15,16], and bear at positions 1 and 3 small alkyl chains (i.e., Me or Et groups). Moreover, the effect on activity of the 4-chloro substitution on the phenyl ring at position 6 was evaluated. constrained analogues in which the H-bonded pseudo-bicyclic structure of merbarone has been converted into a pyrimido-pyrimidine moiety [15,16]. In cell-based assays, derivatives 1 showed interesting antiproliferative properties and the nature of the substituent at position 7 of the pyrimido-pyrimidine scaffold was found to deeply affect the biological activity. In particular, compounds 1a-e (Figure 1) were identified as the  most promising compounds with (sub)micromolar IC50 values against various cell lines  (namely, MT-4, HeLa and MCF-7). Furthermore, selected members of this series were able to inhibit TopoIIα enzyme [16]. Owing to the interest in developing novel merbarone analogues [17,18] and in order to extend the structure activity relationships (SARs) of compounds 1, derivatives 3-6 ( Figure 2) were designed and synthesized. The novel derivatives were substituted at position 7 with the most promising side chain identified in the previous studies [15,16], and bear at positions 1 and 3 small alkyl chains (i.e., Me or Et groups). Moreover, the effect on activity of the 4-chloro substitution on the phenyl ring at position 6 was evaluated.

Results and Discussion
The synthesis of compounds 3-6 was carried out on the basis of previously developed procedures [15,16]. Briefly, the condensation of malononitrile with two equivalents of methyl or ethyl isothiocyanate in basic conditions led to the formation of the pyrimido derivatives 7 and 8 (Scheme 1). Differently from what was observed using phenyl isothiocyanate [15,16], neither the use of different bases (i.e., NaH, t-BuOK) or the use of three equivalents of alkyl heterocumulenes led to the isolation of the pyrimido-pyrimidine analogue. Derivatives 7 and 8 were then reacted, in a one-pot condensation, with phenyl or 4-chlorophenyl isothiocyanate and iodomethane to afford 9-12 (Scheme 1). Finally, the S-methylated compounds were converted into the final derivatives 3-6 by reacting with the proper amine (1.1. equivalents, Scheme 1). The different reactivity of 9-12 intermediates towards amines a-e required different reaction conditions and the desired compounds were obtained in moderate to good yields (Table 1).

Results and Discussion
The synthesis of compounds 3-6 was carried out on the basis of previously developed procedures [15,16]. Briefly, the condensation of malononitrile with two equivalents of methyl or ethyl isothiocyanate in basic conditions led to the formation of the pyrimido derivatives 7 and 8 (Scheme 1). Differently from what was observed using phenyl isothiocyanate [15,16], neither the use of different bases (i.e., NaH, t-BuOK) or the use of three equivalents of alkyl heterocumulenes led to the isolation of the pyrimido-pyrimidine analogue. Derivatives 7 and 8 were then reacted, in a one-pot condensation, with phenyl or 4-chlorophenyl isothiocyanate and iodomethane to afford 9-12 (Scheme 1). Finally, the S-methylated compounds were converted into the final derivatives 3-6 by reacting with the proper amine (1.1. equivalents, Scheme 1). The different reactivity of 9-12 intermediates towards amines a-e required different reaction conditions and the desired compounds were obtained in moderate to good yields ( Table 1).
The antiproliferative properties of compounds 3-6 were assessed in SKOV-3 and MCF-7 cell lines by MTT assay ( Table 2). The derivatives were assayed at a fixed concentration (10 µM) and compound 1d (Figure 1) was considered as reference compound. Overall, the tested derivatives proved to be able to inhibit the proliferation of SKOV-3 and/or MCF-7 cells less efficiently than the reference compound, highlighting that the replacement of the N-phenyl rings at positions 1 and 3 with small alkyl chains was detrimental for the antiproliferative activity. Compounds 3c,d, 5d,e and 6e showed significant cytotoxicity (i.e., growth inhibition percentage greater than 30%) against SKOV-3 cells, derivative 5d being the most active molecule ( Table 2). Within the 1,3-dimethyl substituted compounds, the 4-chloro substitution of the phenyl ring led to the loss of antiproliferative activity (compare 3c with 4c and 3d with 4d, Table 2). Conversely, within the 1,3-diethyl substituted series (derivatives 5 and 6), the effects of the phenyl substitution on activity are related to the nature of the basic chain. Thus, the 3-(4-methylpiperazin-1-yl)propyl derivatives 5e (N-phenyl) and 6e (N-4-chlorophenyl) showed similar antiproliferative properties whereas the 4-chloro substitution on the phenyl ring of the 3-(dibutyl-amino)propyl compound 5d led to the loss of activity (compare 5d with 6d, Table 2). Compounds 3c,d, 5d and 6e also kept their antiproliferative activity (even though reduced) against MCF-7 cells whose proliferation proved to be more affected than that of SKOV-3 cells by derivatives 5e and 6d (Table 2).
Moreover, to further investigate the ability of compound 5d to induce apoptotic cell death in cultured cell lines, as already demonstrated for other TopoIIα inhibitors [19], we performed the Muse™ Annexin V & Dead Cell Assay using the SKOV-3 ovarian adenocarcinoma cell line. As shown in Figure 3, the 48 h treatment with 5d caused an approximately 30% increase of apoptotic cells when compared to the control (DMSO treated). The obtained profile showed the bigger increase in early apoptotic cells (20%) while the late apoptotic cells showed an increase of about 8%.
The antiproliferative properties of compounds 3-6 were assessed in SKOV-3 and MCF-7 cell lines by MTT assay ( Table 2). The derivatives were assayed at a fixed concentration (10 µM) and compound 1d (Figure 1) was considered as reference compound. Overall, the tested derivatives proved to be able to inhibit the proliferation of SKOV-3 and/or MCF-7 cells less efficiently than the reference compound, highlighting that the replacement of the N-phenyl rings at positions 1 and 3 with small alkyl chains was detrimental for the antiproliferative activity. Compounds 3c,d, 5d,e and 6e showed significant cytotoxicity (i.e., growth inhibition percentage greater than 30%) against SKOV-3 cells, derivative 5d being the most active molecule ( Table 2). Within the 1,3-dimethyl substituted compounds, the 4-chloro substitution of the phenyl ring led to the loss of antiproliferative activity (compare 3c with 4c and 3d with 4d, Table 2). Conversely, within the 1,3diethyl substituted series (derivatives 5 and 6), the effects of the phenyl substitution on activity are related to the nature of the basic chain. Thus, the 3-(4-methylpiperazin-1yl)propyl derivatives 5e (N-phenyl) and 6e (N-4-chlorophenyl) showed similar antiproliferative properties whereas the 4-chloro substitution on the phenyl ring of the 3-(dibutylamino)propyl compound 5d led to the loss of activity (compare 5d with 6d, Table 2). Com-Scheme 1. Chemical synthesis of derivatives 3-6.  Moreover, to further investigate the ability of compound 5d to induce apoptotic cell death in cultured cell lines, as already demonstrated for other TopoIIα inhibitors [19], we performed the Muse™ Annexin V & Dead Cell Assay using the SKOV-3 ovarian adenocarcinoma cell line. As shown in Figure 3, the 48 h treatment with 5d caused an approximately 30% increase of apoptotic cells when compared to the control (DMSO treated). The  The TopoIIα inhibitors 1 [16] are chemically related to the apoptosis inducer 5d. On this basis, to assess whether TopoIIα could represent a biological target for 5d, a docking simulation (Autodock 4.2) [20] was carried out considering 1d as template molecule. As the ionization state of compounds could affect the interaction with the biological target [21], the pKa values of compounds 1d and 5d were calculated (ACD labs software). Both ligands showed a pKa value of 9.60, thus highlighting that, at physiological pH, the tertiary amine group of the two molecules would be in a protonated form. According to the docking simulation, the ternary complex TopoIIα-DNA-5d would be stabilized by a hydrogen bond involving the ligand NH group at position 7 and the phosphate function in the DNA strain. Additional contacts would occur between: (i) the thiocarbonyl group at position 2 and the side chain of His578 and (ii) the 3-(dibutyl-amino)propyl substituent and the side chain of Asp541 and Arg487 ( Figure 4A,B). The calculated Ki value of the docking pose was 9.55 mM, thus foreseeing a weak interaction between the ligand and the TopoIIa enzyme. Furthermore, 5d would assume a different binding orientation in comparison with that calculated for the tri-phenyl substituted derivative 1d ( Figure 4C) which showed a calculated Ki value of 772 µM. The TopoIIα-DNA-1d complex would be stabilized mainly by a hydrogen bond between the NH group at position 7 and the side chain of Asp541. Furthermore, a polar interaction between the protonated tertiary amine group of 1d and the DNA phospho-diesteric group would occur. Additional Van der Waals interactions would involve the phenyl ring at position 6 and Asp543 and Lys614 side chains whereas the dibutyl moiety would be in contact with DG3, Glu461, His758 and Leu616 ( Figure 4D).  The TopoIIα inhibitors 1 [16] are chemically related to the apoptosis inducer 5d. On this basis, to assess whether TopoIIα could represent a biological target for 5d, a docking simulation (Autodock 4.2) [20] was carried out considering 1d as template molecule. As the ionization state of compounds could affect the interaction with the biological target [21], the pKa values of compounds 1d and 5d were calculated (ACD labs software). Both ligands showed a pKa value of 9.60, thus highlighting that, at physiological pH, the tertiary amine group of the two molecules would be in a protonated form. According to the docking simulation, the ternary complex TopoIIα-DNA-5d would be stabilized by a hydrogen bond involving the ligand NH group at position 7 and the phosphate function in the DNA strain. Additional contacts would occur between: (i) the thiocarbonyl group at position 2 and the side chain of His578 and (ii) the 3-(dibutyl-amino)propyl substituent and the side chain of Asp541 and Arg487 ( Figure 4A,B). The calculated Ki value of the docking pose was 9.55 mM, thus foreseeing a weak interaction between the ligand and the TopoIIa enzyme. Furthermore, 5d would assume a different binding orientation in comparison with that calculated for the tri-phenyl substituted derivative 1d ( Figure 4C) which showed a calculated Ki value of 772 µM. The TopoIIα-DNA-1d complex would be stabilized mainly by a hydrogen bond between the NH group at position 7 and the side chain of Asp541. Furthermore, a polar interaction between the protonated tertiary amine group of 1d and the DNA phospho-diesteric group would occur. Additional Van der Waals interactions would involve the phenyl ring at position 6 and Asp543 and Lys614 side chains whereas the dibutyl moiety would be in contact with DG3, Glu461, His758 and Leu616 ( Figure 4D).

Chemistry
All building blocks used are commercially available. Amines (a-e, Scheme 1), isothiocyanates, malononitrile and 60% sodium hydride dispersion in mineral oil were pur-

Chemistry
All building blocks used are commercially available. Amines (a-e, Scheme 1), isothiocyanates, malononitrile and 60% sodium hydride dispersion in mineral oil were purchased from Chiminord and Aldrich Chemical (Milan, Italy). Solvents were reagent grade. DMF was dried on molecular sieves (5 Å 1/16" inch pellets). Unless otherwise stated, all commercial reagents were used without further purification. Organic solutions were dried over anhydrous sodium sulphate. A thin layer chromatography (TLC) system for routine monitoring the course of reactions and confirming the purity of analytical samples employed aluminum-backed silica gel plates (Merck DC-Alufolien Kieselgel 60 F 254 ): CHCl 3 was used as developing solvent, and detection of spots was made by UV light and/or by iodine vapors. Yields were not optimized. Melting points were determined on a Fisher-Johns apparatus and are uncorrected. IR spectra were recorded on a Perkin Elmer 398 spectrometer as KBr discs. NMR spectra were recorded on a Varian Gemini 200 or Bruker Avance DPX 300 Spectrometer or JEOL JNM-ECZR instrument. Chemical shifts were reported in δ (ppm) units relative to the internal standard tetramethyl-silane, and the splitting patterns were described as follows: s (singlet), t (triplet) and m (multiplet). The first order values reported for coupling constants J were given in Hz. Elemental analyses were performed by an EA1110 Elemental Analyser (Fison-Instruments, Milan, Italy); all compounds were analyzed for C, H, N and S and the analytical results were within ± 0.4% of the theoretical values. For mass spectra collection, each sample was dissolved in DMSO (final concentration: 10 mM) and after further dilution in acetonitrile (final concentration 100 nM), it was analyzed by flow injection mass spectrometry (FIA-MS). Briefly, five microliters of sample were injected into an eluent flow containing 0.1% formic acid in acetonitrile, generated by a Vanquish UHPLC system (Thermo Fisher Scientific, San Jose, CA, USA). The flow rate was 100 µL/min. The eluent was directly sent to a Q-Exactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) equipped with a heated electrospray ion source (HESI-II). Prior to each series of acquisitions, the mass spectrometer was externally calibrated with Positive Ion Calibration Solution (Thermo Fisher Scientific, San Jose, CA, USA). The following operating parameters were applied: resolution 35,000; sheath and auxiliary gas flow rate were 35 and 10 respectively; spray voltage 3.5 kV; capillary temperature 250 • C; S-lens RF level 100. The autogain control (AGC) was optimized at 1e6 with a maximum injection time (maxIT) of 250 ms. Full scan data were processed with Xcalibur version 4.1 (Thermo Fisher Scientific, San Jose, CA, USA). High resolution mass spectra, ranging from 100 to 600 m/z, were acquired in positive ion mode and the identity of each analyte was confirmed by comparing the experimental data with both their theoretical molecular weight (+H + ) and their expected isotopic pattern. IR, NMR and mass spectra are reported in Supplementary Materials, Figures S1-S88.

General Procedure for the Synthesis of Compounds 7 and 8
To an ice-cooled, dry DMF solution (30 mL) of malononitrile (760 mg, 10 mmol) and the appropriate isothiocyanate (20 mmol), 60% sodium hydride dispersion in mineral oil (870 mg, 20 mmol) was added. The reaction mixture was stirred at room temperature for 15h. After addition of cool water (50 mL) and 2N HCl (pH = 0), a solid precipitated. The crude solid was collected by filtration and purified by crystallization from EtOH/DCM mixture.

General Procedure for the Synthesis of Compounds 3-6
To a dry DMF (10 mL) solution of the suitable derivative 9-12 (1 mmol), the appropriate amine (1.2 mmol) was added. The reaction mixture was stirred at variable time and temperature as detailed in Table 1. After the addition of water (10 mL), a solid separated. The crude precipitate was collected by filtration and purified by crystallization from DCM/MeOH mixture. EuroClone, Milan, Italy) and incubated at 37 • C in 5% CO 2 in a humidified environment. Briefly, the neoplastic cells were plated in 96 well plates at an adequate number to reach 80-85% of confluence at the end of the assay. 16 h after plating, compounds were dissolved in DMSO to give a 10 mM stock solution, diluted in growth medium and added at a final concentration of 10 µM. After 48 h of incubation, 30 µL of MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide) at a concentration of 2 mg/mL in PBS, were added in each well. Then, after further 4 h of incubation, the sur-natant was removed and 100 µL/well of DMSO were used to dissolve the Formazan precipitate that can be found in vital cells. After 20 min, the results were read at 570 nm by means of a spectrophotometer. The results are expressed as percentage of the control samples in which the cells were incubated with the same amount of DMSO but without compounds. The assays were repeated three times. In each set, every single compound was tested six times. Means and standard deviations were calculated.

Annexin V & Dead Cell Assay
Muse™ Annexin V & Dead Cell Assay was performed following the manufacturer instruction (Millipore, Billerica, MA, USA). Briefly, cultured cells were treated for 48 h with 2.5 µM compound 5d (final concentration in growth medium). Once dissociated with trypsin, the cells were resuspended in DMEM containing 10% FBS, 2 mM Glutamine and 1% pen-strep to achieve a final cell concentration of 1 × 10 5 -1 × 10 7 cell/mL. Then, 100 µL of the cell suspension were mixed with 100 µL of Muse™ Annexin V & Dead Cell reagent. After a 20 min incubation in the dark, the sample was read using the Muse™ Cell Analyser. The control sample was treated in the same way as the 5d compound treated sample.

Data Availability Statement:
The data presented in this study are available in this article.
Acknowledgments: C.R., A.P. and M.P. acknowledge Italian Ministry of Health for the support. C.R. acknowledge the support from the COST Action CA17104 STRATAGEM "New diagnostic and therapeutic tools against multidrug resistant tumors". The authors acknowledge Riccardo Raggio for 13 C NMR spectra collection.