Mithplatins: Mithramycin SA-Pt(II) Complex Conjugates for the Treatment of Platinum Resistant Ovarian Cancers

Abstract

DNA coordinating platinum (Pt) containing compounds cisplatin and carboplatin have been used for the treatment of ovarian cancer therapy for four decades. However, recurrent Pt-resistant cancers are a major cause of mortality. To combat Pt-resistant ovarian cancers, we designed and synthesized a conjugate of an anticancer drug mithramycin with a reactive Pt(II) bearing moiety, which we termed mithplatin. The conjugates displayed both the Mg²⁺-dependent noncovalent DNA binding characteristic of mithramycin and the covalent crosslinking to DNA of the Pt. The conjugate was three times as potent as cisplatin against ovarian cancer cells. The DNA lesions caused by the conjugate led to the generation of DNA double-strand breaks, as also observed with cisplatin. Nevertheless, the conjugate was highly active against both Pt-sensitive and Pt-resistant ovarian cancer cells. This study paves the way to developing mithplatins to combat Pt-resistant ovarian cancers.

Graphical Abstract

The authors describe bifunctional DNA targeting hybrid molecules that consist of a mithramycin core and a Pt complex, connected by a flexible linker. The mithramycin portion binds DNA non-covalently, whereas the Pt forms mono-adducts, intra- or interstrand DNA crosslinks. The molecules inhibit ovarian cancer cells overcoming their resistance to cisplatin.

Introduction

With nearly 14,000 annual deaths, ovarian cancer is the fifth deadliest malignancy in the US among women, following lung, breast, colon and pancreatic cancers.^[1] The primary chemotherapy treatment after ovarian cancer surgery is a combination of paclitaxel, a tubulin targeting agent, and carboplatin, a platinum (Pt) containing drug that forms DNA monoadducts and crosslinks.^[2] However, despite initial success, 80% of advanced stage cancers recur with a poor prognosis.^[3] The prognosis is especially poor for Pt-resistant cancers, where only 15% of patients respond to another therapy.^[4] Therefore, novel therapeutics overcoming Pt resistance could significantly lower ovarian cancer mortality.^{[3c, 5]}

Mithramycin (MTM) is a DNA binding bacterial natural product with antibiotic and anticancer activities. Unlike Pt, MTM binds DNA noncovalently in the minor groove.^[6] MTM had been clinically used to treat testicular, breast and bone cancers and Paget’s bone disease,^[7] but it was discontinued due to a very narrow therapeutic window and severe side effects, such as liver toxicity and bleeding.^[8] However, MTM was recently discovered as a highly potent agent against Ewing sarcoma,^[9] a bone cancer affecting children and young adults. A follow-up phase I/II trial with MTM as a single agent against Ewing sarcoma was terminated due to poor pharmacokinetic properties and liver toxicity.^[10] To overcome these obstacles, we launched a derivatization program that yielded new 3-side chain analogues of MTM showing promising pharmacological properties and significantly increased selectivity against Ewing sarcoma cell lines compared to other cell lines.^[11] MTM and its derivatives also show good activity against ovarian cancer cell lines, partly due to their inhibitory effect on the transcription factor Sp1, which is closely linked to ovarian tumorigenesis and platinum resistance.^{[11c, 12]} In this study, we describe a 3-side chain conjugate of MTM with a Pt containing moiety (termed MPt1), which has a dual mechanism of action of MTM and Pt drugs, and which shows improved potency over cisplatin against ovarian cancer cell lines. This potency is only slightly affected by the Pt resistance status of the cells.

Results

Design of MTM-Pt conjugates

MTM (1; Schcme 1) forms a Mg²⁺-coordinated dimer, which binds DNA noncovalently at X(G/C)(G/C)X sequences in the minor groove.^[13] The tricyclic chromophore and the sugar side chains are directly involved in the dimerization and DNA binding, whereas the 3-side chain protrudes into the solvent, and it can be modified without destabilizing the complex with DNA.^[14] Because the Pt in the Pt drugs coordinates the N⁷ of the guanines in the major groove of the DNA,^[15] we reasoned that linking Pt with MTM may be beneficial to targeting the Pt to G/C-rich DNA sequences, where the Pt would have improved chances of forming a cytotoxic interstrand DNA crosslink (ICL). Furthermore, because MTM is a dimer, DNA sites, where formation of two symmetrical crosslinks are possible, would be preferentially targeted by such MTM-Pt conjugates, which we termed mithplatins. Repair of such extremely bulky DNA lesions may pose a serious obstacle. Additionally, the presence of the MTM moiety could prevent efflux of by the mechanisms that contribute to the resistance to conventional Pt drugs, such as cisplatin or carboplatin. With this rationale in mind, we designed a conjugate of an MTM derivative, mithramycin SA (MTM SA; 2; Scheme 1), and a Pt (II) complex, connecting the Pt complex to the 3-side chain of MTM SA by an aliphatic linker (3; Scheme 1). If two moieties of a drug interact with two biological target sites at the same time, both affinity and specificity can be increased synergistically. We reasoned that binding of the MTM moiety to its cognate site in the minor groove of DNA may occur simultaneously with formation of monoadducts or crosslinks by the Pt atoms in the major groove for some DNA sequences. An interstrand DNA crosslink is a highly distorting lesion that involves local base pair unstacking and may not require a long linker for such dual interaction to occur. Our hypothesis was that the designed dual-action agents, consisting of a combination of a mithramycin core moiety and a suitably linked Pt(II) complex would form very stable and bulky doubly crosslinked DNA adducts that would not be effectively repaired by the DNA repair mechanisms involved in resistance to conventional Pt drugs. Furthermore, even if the simultaneous DNA binding by the two moieties of the conjugate is not possible, such a conjugate is not likely to be a substrate of the efflux pumps that function to eliminate the conventional Pt drugs from the cells, as another major contributor to the Pt resistance.

Scheme 1.

Chemical structures of mithramycin (MTM; 1), mithramycin SA (MTM SA; 2), and a model MTM SA-Pt (II) complex conjugate (MPt1; 3)

Synthesis of mithplatin

MTM SA (compound 2 in Scheme 1) was discovered as a combinatorial biosynthetic analogue of MTM 1 produced by Streptomyces argillaceus upon insertional inactivation of the mtmW gene.^[16] Its free carboxylic acid group in the 3-side chain allows selective derivatization of MTM SA 2 without availing protecting groups. To overcome low yields of the MTM SA producer S. argillaceus M7W1 and cumbersome isolation procedures to separate MTM SA from two major congeners produced by the same strain, MTM SK and MTMS DK, we developed a chemical method using sodium periodate/potassium permanganate to obtain MTM SA from commercially available MTM, optimized to yield ~ 80% MTM SA 2. The conditions were optimized to oxidatively cleave the 3-side chain to completion without cleaving the two 1,2-diol moieties of the terminal sugar D-olivose (6-side chain) and D-mycarose (2-side chain). This method was successful both on a small (several mg) and a larger (>1 g) scale, with yields between 55 and 60%.

The Pt(II) complex was synthesized following established procedures, and linked via a PyBOP-mediated condensation process, which we had optimized previously to synthesize MTM SA analogues with amino acid containing 3-side chains.^[11a] The Pt(II) complex containing 2-picolylamine ligand was attached to a three-carbon alkyl chain linker derived from 1,3-diaminpropane (Scheme 3).

Scheme 3. Synthesis of MPt1 using a side chain derived from 3-amino propanol. A 60:40 mixture of interconverting diastereomers was obtained.

Reagents and conditions: a) Boc₂O (1.1 eq), Et₃N (2 eq), THF, 0 °C to rt, 12h, 98%; b) PPh₃ (2 eq), CBr₄ (1.5 eq), THF, 0 °C to rt, 74%; c) 2-Picolylamine (2 eq), CH₃CN, 60 °C, 12h, 65%; d) K₂[PtCl₄] (1.5 eq), DMF- Water (2:1), 50 °C, 55%; e) TFA (10 eq),DCM, 0 °C to rt, 98%; f) 2 (0.33eq), PYBOP(2 eq), Et₃N, DMF, 0 °C, 45 min, 30%.

We began our synthesis of mithplatin (MPt1; 3) by preparing the alkyl chain inker. N-Boc protection of 3- amino propanol 4 with di-tert-butyl dicarbonate and Et₃N yielded known compound 5 in quantitative fashion, which was further subjected to Appel reaction conditions resulting in bromide 6 in 74% yield.^[17],[18] In the next step, bromide 6 was treated with excess 2-picolylamine to give mixture of mono and di N-alkylated product. Chromatographic purification of crude reaction product give compound 7 in 65% yield. Further complexation of 7 with potassium tetrachloroplatinate yielded Pt(II) complex 8 in 85% yield. Complex formation was evident from ¹HNMR spectrum of 8. ¹⁹⁵Pt- NMR further confirmed the metal complexation. Trifluoracetic acid-mediated N-Boc deprotection of 8 yielded intermediate 9, which was used without purification in the next step. PyBOP mediated coupling reaction between 9 and MTM SA 2 was carried out in DMF solvent at 0°C. Reaction was monitored with HPLC-MS using combination of mass and UV absorbance to identify elution peaks corresponding to desired products. HPLC-MS profile indicated presence of isomeric mixture in 60:40 ratio. HPLC was performed to isolate individual isomer. Each individual isomer was converted back to original isomeric mixture during freeze drying of compounds (see. supporting info). MPt1 was obtained in 30% isolated yield as yellow powder. NMR and HRMS were performed for structure conformation.

Two DNA binding modes of MPt1

MPt1 was designed to bind DNA non-covalently by its MTM moiety and form Pt-mediated DNA monoadducts or crosslinks. We tested these two DNA binding functions of MPt1. To observe the formation of a Mg2+-coordinated dimer, which is a functional form of MTM^{[6a, 13, 19]} and, presumably, MPt1, where the Mg²⁺-coordinating chromophore of MTM is preserved, we examined the intrinsic fluorescence of MPt1 and MTM as a control in the presence and in the absence of Mg²⁺. Both MPt1 and MTM exhibited a 4-fold decrease in their intrinsic fluorescence (Fig. 1, data at 0 μM DNA) upon addition of Mg²⁺, signifying Mg²⁺-mediated dimerization observed previously for MTM^{[14, 20]}. To measure the non-covalent MTM-like DNA binding of MPt1, we monitored the increase in intrinsic MPt1 fluorescence upon titrating double-stranded (ds) palindromic DNA oligomer AATGAGCTCATT containing a single MTM binding site into MPt1 held at the constant concentration of 5 μM (in dimers) (Fig. 1). MTM was used as a control, also at 5 μM in dimers. A dramatic increase in MTM fluorescence upon its binding DNA has been used previously for quantitative measurements of the binding affinity and sequence specificity of MTM, its analogues and complexes with different divalent metal ions.^{[14, 21]} Here, in agreement with the previous studies, in the presence of Mg2+ we observed a ~7-fold increase in MTM fluorescence upon DNA titration and an equilibrium binding constant Kd of < 1 μM (only upper bound could be measured; Fig. 1, filled squares). In contrast, in the absence of Mg²⁺ no DNA binding was observed (Fig. 1, open squares). Similarly, in the presence of Mg²⁺, MPt1 displayed a 5-fold increase in fluorescence similar to the previously reported fluorescence enhancement of Mg²⁺-coordinated MTM dimer with increasing DNA concentration (Fig. 1, filled circles).^[14] In the absence of Mg²⁺, the fluorescence signal did not change upon titrating DNA indicating the absence of DNA binding (Fig. 1, open circles). These results indicated that MPt1, like MTM, binds DNA with high affinity as a Mg²⁺-coordinated dimer. The equilibrium constant for the (MPt1)2-Mg²⁺-DNA binding obtained from these data was Kd = 1.8 ± 0.7 μM, comparable to the low- μM affinity of MTM.^[14]

Figure 1.

Equilibrium binding isotherms for MPt1-DNA complex formation in the presence (filled circles) and in the absence (open circles) of Mg2+. MTM was used as a control, in the presence (filled squares) and in the absence (open squares) of 5 μM Mg2+. The solid and dashed curves are the best-fit 1:1 binding isotherms for the MPt1 and MTM data in the presence of Mg2+, respectively, for Kd = 1.8 ± 0.7 μM (MPt1) and Kd < 1 μM (MTM).

Next, we examined whether MPt1 could form DNA adducts or crosslinks. We incubated MPt1 with fluorescently labeled 21-mer ds DNA in the absence of a divalent metal ion and used cisplatin as a positive control. The DNA sequence (Figure 2) was designed to form Pt monoadducts, intra- and interstrand crosslinks (ICLs). The results, visualized on the denaturing polyacrylamide gel (Fig, 2) showed that MPt1 formed ICLs as well as intrastand crosslinks and/or monoadducts, demonstrating that the Pt of MPt1 crosslinked DNA, as designed.

Figure 2.

A 16% denaturing polyacrylamide gel showing DNA and its reaction products with MPt1 and cisplatin. The DNA sequence is shown on the bottom, where 6-carboxyfluorescein is denoted as 6FAM.

Anti-cancer activity of MPt1 against ovarian cancer cells in vitro

We assessed the cytotoxic activity of MPt1 in four human ovarian cancer cell lines: Caov-3, OVCAR3, UWB1.289 and UWB1-PlatR (Figure. 3). Cisplatin and MTM were used as a reference. UWB1-PlatR cells are a derivative line of UWB1.289 with acquired platinum resistance, developed via repeated exposures to carboplatin, as described in Experimental Section. Cisplatin IC₅₀ values varied by 3-fold across the cell lines. The UWB1-PlatR cell line was significantly more resistant to cisplatin than the parental UWB1.289, with IC₅₀ values of 8.9 ± 3.0 μM [mean ± standard deviation (SD)] and 3.0 ± 1.4 μM, respectively (p ≤ 0.001) (Figure. 3A), resulting in a resistance factor of 3. Caov-3 cells were also less sensitive to cisplatin than UWB1.289 cells with mean cisplatin IC₅₀ of 6.6 ± 4.9 μM (p < 0.05). Mean cisplatin IC₅₀ in OVCAR3 cells was 5.3 ± 2.0 μM, lower than that in UWB1-PlatR cells (p < 0.05). Caov-3, OVCAR3, and UWB1-PlatR cell lines were significantly more sensitive to MPt1 than cisplatin, with mean IC₅₀ of 1.3 ± 1.2 μM, 0.8 ± 0.07 μM and 3.3 ± 1.6 μM, respectively (p ≤ 0.01), while UWB1.289 sensitivity to MPt1 was similar to that of cisplatin (Figure. 3A). The resistance factor for MPt1 among the UWB1.289 and UWB1-PlatR cells was 1.9, however, the MPt1 IC₅₀ values were not significantly different (p > 0.05). The IC₅₀ values for MTM were similar across all four cell lines (~70 nM), consistent with previous studies,^[22] regardless of cisplatin resistance (Figure. 3B). These observations demonstrated a broad activity of MPt1 across the cell lines tested, including cisplatin-resistant cell lines, in ~1 μM range.

Figure 3.

In vitro sensitivity of ovarian cancer cell lines to (A) MPt1, cisplatin and (B) mithramycin treatment. Scatter plots depict replicate and mean (line) IC₅₀ values from a minimum of three biological replicates. Differential sensitivity was determined using one-way analysis of variance (MPt1 and cisplatin; p < 0.0001) and Bonferroni’s multiple comparison tests (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001).

Anti-proliferative drug synergy was assessed for the combination of cisplatin and mithramycin in ovarian cancer cells. Caov-3, OVCAR3, UWB1.289 and UWB1-PlatR cells were treated for 72 hours with cisplatin and mithramycin as single agents and in combination. Cell viability was then measured and drug synergy was evaluated using the Bliss independence model. Results are summarized in Supplementary Table 1. Mean Bliss Synergy Scores ranged from −3.09 to 4.75, none of which were significantly different from 0, indicative of drug additivity in each cell line tested.

DNA damage following MPt1 treatment

The induction of DNA damage, specifically double-strand breaks, was measured in the same four ovarian cancer cells following exposure to MPt1, cisplatin and MTM. Immunofluorescent detection of phosphorylated histone H2AX (γH2AX) was used to identify double-strand DNA breaks after treating Caov-3, OVCAR3, UWB1.289 and UWB1-PlatR cells for 48 hours (Figure. 4A-D). Mean nuclear γH2AX signal intensity was measured and expressed as the fold-increase in mean signal over matched vehicle-treated control cells (Figure. 4E). Cisplatin treatment significantly induced double-strand DNA breaks in all cell lines compared to vehicle-treated (blank media) controls (p < 0.001). Mean γH2AX fold-increase over control ± SD was 5.8 ± 0.5 in UWB1.289 cells, 5.9 ± 1.6 in Caov-3 cells, 8.5 ± 1.7 in UWB1-PlatR cells and 11.6 ± 2.7 in OVCAR3 cells. Significant induction of double-strand DNA breaks, compared to vehicle-treated (1% ethanol) controls, was also observed following MPt1 treatment across all cell lines (p < 0.001). The fold-increase in mean γH2AX staining ± SD was 1.7 ± 0.3 in UWB1.289 cells, 2.5 ± 0.8 in UWB1-PlatR cells, 2.5 ± 0.5 in Caov-3 cells and 4.0 ± 1.0 in OVCAR3 cells. Surprisingly, MTM also significantly increased double-strand DNA breaks compared to vehicle-treated (0.1% DMSO) controls. The fold-increase in mean γH2AX staining ± SD was 3.1 ± 0.7 in Caov-3 cells (p < 0.001), 4.5 ± 1.0 in OVCAR3 cells (p < 0.001) and 1.5 ± 0.1 in UWB1.289 cells (p < 0.01). MTM treatment did not significantly induce γH2AX in UWB1-PlatR cells compared to controls.

Figure 4.

DNA double-strand break formation following MPt1, MTM and cisplatin treatment in human ovarian cancer cells. Representative γH2AX immunofluorescence images for (A) Caov-3, (B) OVCAR3, (C) UWB1.289 and (D) UWB1-PlatR cells treated with the indicated drug concentrations for 48 hours. In merged images, blue indicates Hoechst and red indicates γH2AX signal. Scale bar = 50 μm. (E) Mean γH2AX signal intensities (± SD) measured following 48 hours treatment and normalized to vehicle control. Statistical significance was determined using one-way analysis of variance (p < 0.001) and Bonferroni’s multiple comparison tests versus vehicle-treated controls (**p < 0.01; ***p < 0.001).

Caspase 3/7 activation following MPt1 treatment

Apoptosis was monitored in ovarian cancer cells via activation of caspases 3 and 7. Cells were treated for 24, 48 and 72 hours with 15 μM cisplatin, 15 μM MPt1 or 200 nM MTM. Caspase 3 and 7 activation was measured with fluorescence microscopy using CellEvent Caspase-3/7 Green Detection Reagent, a specific fluorogenic substrate activated upon caspase-mediated cleavage. Mean nuclear fluorescence was measured and normalized to vehicle-treated control cells (Figure. 5). MPt1 induced caspase 3/7 activation in all cell lines tested, although the timing and magnitude differed. At 24 hours post-treatment, OVCAR3 (p < 0.001), UWB1.289 (p < 0.05) and UWB1-PlatR (p < 0.05) cells showed significant induction of apoptosis. MPt1 activated caspase 3/7 at 48 hours in Caov-3 (p < 0.05), OVCAR3 (p < 0.001) and UWB1-PlatR (p < 0.01) cells and at 72 hours in OVCAR3 (p < 0.001), UWB1.289 (p < 0.001) and UWB1-PlatR cells (p < 0.001). Similarly, cisplatin induced apoptosis at 24, 48 and 72 hours in OVCAR3 cells (p < 0.001), at 72 hours in Caov-3 cells (p < 0.05) and at 24 hours in UWB1.289 cells (p < 0.05). Notably, a significant induction of caspase 3/7 activity was not observed for cisplatin-treated UWB1-PlatR cells at any of the time points tested. Caov-3 cells treated with 200 nM MTM showed higher caspase 3/7 activity at 72 hours (p < 0.05) and UWB1-PlatR cells had caspase activation at 24 (p < 0.001) and 72 hours (p < 0.01) post-treatment. MTM induced apoptosis in OVCAR3 cells at 48 (p < 0.001) and 72 hours (p < 0.001) but not in UWB1.289 cells at any of the time points tested.

Figure 5.

Mean CellEvent Caspase-3/7 Detection Reagent nuclear fluorescence normalized to vehicle control and expressed as fold-change (± SD) in Caov-3, OVCAR3, UWB1.289 and UWB1-PlatR ovarian cancer cells. Fluorescence was measured following 15 μM MPt1, 15 μM cisplatin or 200 nM MTM treatment for 24, 48 and 72 hours. Statistical significance was determined using two-way analysis of variance (p < 0.0001) and Bonferroni’s multiple comparison tests versus vehicle-treated controls (*p < 0.05; **p < 0.01; ***p < 0.001).

DISCUSSION AND CONCLUSIONS

With the goal of designing a Pt-based agent that could overcome Pt resistance in ovarian cancer, we designed a conjugate of a DNA binding natural product, MTM, with a Pt bearing reactive moiety.^[30] As a proof of principle, a prototypical conjugate (3; Scheme 1), which we called MPt1, was synthesized. To improve the production of MPt1, a chemical procedure was developed to obtain MTM SA from MTM, rather than isolating MTM SA as a shunt product by direct fermentation of a mutant bacterial strain.^[23] This new semisynthetic approach also stimulated our recent MTM-oxime based derivatization strategy.^[24] A similar oxidative cleavage was previously used by Preobrazhenskaya et al. for the 3-side-chain cleavage of olivomycin.^[25] However, olivomycin -in contrast to mithramycin-does not have any sugar moiety vulnerable to this oxidative cleavage procedure. Destruction of the terminal sugar moieties would render the mithramycin moiety of the conjugates almost inactive, since the D-mycarose sugar is particularly important for optimal the DNA-interaction of mithramycin and its analogues.^[13-14] The resulting procedure is a careful balance aiming at side chain cleavage without destroying the terminal sugar moieties. The PyBop-mediated amide coupling between MTM SA (2) and Pt complex 9 yielded MPt1 as interconverting mixture of isomers. ¹ H NMR analysis indicated that these two isomers were epimers. There are two types of chirality associated with MPt1: 1) stereogenic centers in the MTM part of the molecule and 2) a chirality in the Pt complex. The Pt complex of MPt1 contains a chelating ring that can adapt two nonplanar conformations (δ/λ). Interconversion between these enantiomeric conformations may be a reason for the observed interconverting epimer of MPt1.^[26]

The main design principle of MPt1 was to ensure that it was bifunctional: the MTM part would bind DNA noncovalently, analogously to MTM, whereas the Pt complex would crosslink to DNA, similarly to cisplatin. The DNA crosslinks was expected to include ICLs, the most cytotoxic Pt lesions. Indeed, we observed that MPt1 displayed both of these DNA binding modes. In agreement with the ICL formation by MPt1 in the test tube, MPt1 induced ds DNA breaks, an intermediate in repair of ICLs, in all cell lines tested. Because ICL/ds DNA breaks is an established mechanism of action of cisplatin, we concluded that MPt1 functioned as designed. It remains to be tested whether the two DNA binding modes can occur simultaneously for the same MPt1 molecule. We predict that such bidentate binding can happen, but it may need optimization of the length of the linker connecting the Pt complex to the MTM part.

Cisplatin resistance remains a major clinical challenge in the treatment of ovarian cancer as well as many other solid tumors.^[27] In MPt1, to overcome Pt resistance, the MTM part was meant to improve targeting the reactive Pt complex to G/C-rich DNA, to pose a significant block to DNA repair and to prevent efflux by cisplatin efflux pumps. Testing MPt1 against a number of ovarian cancer cell lines demonstrated broad anticancer activity and enhanced cytotoxicity over cisplatin in most cell lines tested, including the Pt-resistant UWB1-PlatR, in the 1-3 μM range. While the cells were also sensitive to MTM in the ~700 nM range, the clinical use of MTM was limited by severe adverse effects.^{[7b, 28]} This study gives us hope that mithplatins will retain the main benefit of MTM- a broad activity independent of cisplatin sensitivity, while maintaining or minimizing the potential for adverse effects characteristic of Pt drugs.

Experimental Section

Chemistry

All commercial reagents were used without further purification. Solvents were dried and distilled following the standard procedures. TLC was carried out on pre-coated plates (Merck silica gel 60, GF254), and the spots were visualized with UV, fluorescent light or by charring with phosphmolybdic acid hydrate (PMA). Column chromatography was performed on silica gel (230-400 mesh). ¹H and ¹³C NMR spectra for the compounds were recorded with Varian 400, 500 MHz and Bruker 600 AV4 spectrometers. ¹⁹⁵Pt was performed using Varian 400 spectrometer using. ¹H and ¹³C chemical shifts are reported in ppm downfield of tetramethylsilane and referenced to residual solvent peak (CHCl₃; δ_H = 7.26 and δ_C = 77.23, d₄-MeOH; δ_H= 3.31 and δ_C = 49.1). ₁₉₅Ptchemical shifts are described against the standard K₂[PtCl₄] for which the chemical shift in D₂O is at −1628.0 ppm. Multiplicities are reported using the following abbreviations: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad resonance, ap = apparent. Mass spectra were taken on Simazu QTOF mass spectrometer. The phrase ‘usual work up’ or ‘worked up in usual manner’ refers to washing of the organic phase with water (2 × 1/4 the volume of organic phase) and brine (1 × 1/4 the volume of organic phase), and drying (anhydrous Na₂SO₄), filtration, and concentration under reduced pressure. Yields referred to isolated yields after purification.

General Procedure A: N- Boc protection of amines

To a stirred solution of Amine (equiv.) in dry THF (1.5mL/mmol) were added Boc₂O (1.1 equiv.) and Et₃N (2 equiv.) at 0 °C under Ar atmosphere, and the reaction mixture was stirred for 12h at room temperature. After adding water, the resultant mixture was evaporated to half volume and extracted three times with Dichloromethane. The combined organic layers were washed with brine, dried over Na₂SO₄, filtered, and evaporated under reduced pressure. The residue was purified by silica gel flash column chromatography to afford N-Boc-protected derivative.

General Procedure B: N-Alkylation 2- Picolylamine

To the stirred solution of bromide (1 equiv.) in acetonitrile (5mL/ mmol) 2-Picolylamine (2 equiv.) of was added at rt. After stirring for 1h, the reaction mixture was heated at 60 °C overnight. Upon complete consumption of stating material the reaction mixture was cooled to room and quenched with addition of saturated solution of K₂CO₃. Solvent removed under reduced pressure. The residue was dissolved in DCM and worked up in the usual manner. Column chromatography with silica gel using 5% methanol in dichloromethane as eluent provided the desired product.

General Procedure C: Synthesis Pt (II) Complexes

The solution of potassium tetrachloroplatinate (1.5 equiv) in DMF (3 mL/mmol) and distilled water (3 mL/mmol) were added to a solution of chelating amine ligand (1 equiv.) in DMF (3 mL/mmol). The resulting mixture was stirred in the dark for 24h at 50 °C. Then, the solvent was evaporated, and the residue was stirred vigorously in a saturated aqueous potassium chloride solution (5 mL) for 20 min. The resulting precipitate was filtered by vacuum filtration, washed with water (3×10 mL), and dried in a desiccator over phosphorus pentoxide for 1 day. All complexes were prepared as precipitated off-white powder.

General Procedure D: N-Boc-deprotection

To stirred solution of N- Boc- protected Pt (II) complex in CH₂Cl₂(10mL/mol) was added trifluoroacetic acid (10 equiv.) at 0 °C then the reaction mixture was stirred at 25°C for 12h. CH₂Cl₂ was evaporated and replaced by anhydrous toluene which was then evaporated to azeotrope an excess tifluoroacetic acid. This operation was repeated three times to yield an off-white solid which was dried in vacuo. The free amine was only characterized by LCMS and directly used without further purification in the PyBop coupling reaction with MTM SA 2.

General Procedure E: PyBop mediated coupling reaction

To a stirred solution of MTM SA 2 (1 equiv) in dry DMF were added PyBop (1.5 equiv), free amine (3.0 equiv) and excess triethyl amine (adjusted to pH 8). The reaction mixture was stirred at room temperature under argon atmosphere until the disappearance of MTMSA (2). It was quenched by adding saturated NaCl solution and extracted with n-BuOH. The organic fraction was collected and concentrated under reduced pressure and purified by HPLC to obtain pure MTMSA analogues.

Oxidative cleavage of mithramycin to obtain MTMSA

Solid pure mithramycin MTM (1.0, 0.92 mmole, 1.0 equiv.) was slowly added to a stirred solution of sodium periodate (0.275g, 1.3 mmole, 1.4 equiv.) dissolved in 5.2 mL (4.0 mL/ mmol) of water and 4.0 mL (3.0 mL/ mmol) of acetone taken in a round-bottomed flask under N₂ balloon at 5°C (ice-bath). The resulting yellow solution was stirred for 1h. Then potassium permanganate (0.030g, 0.185mmole, 0.2 equiv.) dissolved in 1.0 mL water (5.0 mL/ mmol) was added potion to the reaction mixture simultaneously with equal volume of 1.0 mL of acetone (5.0 mL/ mmol) over a period of 3h. After stirring was continued for 12h at 5°C, additional sodium periodate (0.075g, 0.375 mmol, 0.4 equiv.) was added and stirring was continued for another 6h. After complete consumption of starting material MTM (reaction was monitored using LCMS technique) 5 mL water was added, and aqueous phase was extracted with n-butanol (4x25mL). Combined organic phase was collected and concentrated under vacuo.

Obtained blackish gummy liquid was diluted with 5.0 mL of toluene and azeotropic distillation was performed under reduced pressure. This process was repeated for 2-3 times until a yellow solid material appears. The crude mass was dissolved in methanol (50 mL) and final purification was done by semi-prep HPLC with reversed phase C18 column (25%-100% CH₃CN-H₂O gradient/0.1% formic acid, 34 mins, 254/280 nm, at a flow rate of 3.0 mL/min).

Desired peak was collected between 16.5- and 17.5-min. The isolated combined solution was dried in vacuo and yielded the MTMSA (0.520 g, 0.51 mmol, 55%) as yellow solid.

The recorded ¹H, ¹³C data were in good agreement with the literature.¹ It confirms compound MTMSA prepared by this chemical oxidation method is identical with the previously reported MTMSA from the mutant strain M7W1.^{[12a, 16]}

¹H NMR (500 MHz, Methanol-d₄) δ 8.43 (s, 1H), 6.62 (s, 1H), 6.45 (d, J = 1.5 Hz, 1H), 5.24 (dd, J = 9.8, 2.2 Hz, 1H), 5.17 (dd, J = 9.6, 2.1 Hz, 1H), 4.95 – 4.89 (m, 2H), 4.83 – 4.80 (m, 2H), 4.72 – 4.70 (m, 1H), 4.67 (d, J = 9.5 Hz, 1H), 4.42 (d, J = 11.7 Hz, 1H), 4.11 (d, J = 1.9 Hz, 1H), 3.79 (ddq, J = 16.9, 8.7, 5.7, 5.3 Hz, 3H), 3.70 (d, J = 2.7 Hz, 1H), 3.67 (d, J = 4.4 Hz, 1H), 3.55 (ddt, J = 11.6, 8.8, 6.0 Hz, 3H), 3.43 (s, 3H), 3.36 – 3.31 (m, 2H), 3.16 – 3.06 (m, 3H), 2.97 – 2.90 (m, 3H), 2.74 – 2.65 (m, 2H), 2.52 – 2.37 (m, 3H), 2.19 (ddd, J = 12.4, 4.9, 1.9 Hz, 2H), 2.15 (s, 3H), 1.92 (dd, J = 13.6, 2.2 Hz, 1H), 1.61 – 1.52 (m, 3H), 1.40 (d, J = 6.1 Hz, 3H), 1.31 (d, J = 6.2 Hz, 6H), 1.27 – 1.22 (m, 6H), 1.20 (d, J = 6.4 Hz, 3H). ¹³C NMR (100 MHz, CD₃OD) δ 202.94, 175.03, 163.70, 159.08, 155.38, 138.34, 135.51, 116.52, 110.66, 109.99, 107.74, 100.75, 100.54, 98.57, 98.43, 97.47, 96.33, 79.53, 79.30, 77.78, 76.75, 76.69, 76.46, 75.91, 75.17, 74.81, 72.19, 71.98, 71.84, 70.62, 70.59, 70.44, 70.32, 69.02, 58.04, 43.88, 43.80, 39.28, 36.76, 36.45, 31.69, 27.08, 25.79, 17.28, 17.19, 17.00, 16.71, 15.57, 7.00.

Intermediate 7

Compound 7 (65%, 0.864g) was prepared as yellowish gummy liquid by N-alkylation of 2- Picolylamine (1.08g, 10.07 mmol) with bromide 6 (1.2g, 5.03 mmol) in acetonitrile (10 mL) following general procedure B.

¹H NMR (400 MHz, Chloroform-d) δ 8.47 (d, J = 4.9 Hz, 1H), 7.56 (td, J = 7.7, 1.8 Hz, 1H), 7.24 (d, J = 7.8 Hz, 1H), 7.08 (dd, J = 7.5, 4.8, 1H), 5.32 (s, 1H), 3.83 (s, 2H), 3.14 (q, J = 6.3 Hz, 2H), 2.65 (t, J = 6.7 Hz, 2H), 1.63 (m, 2H), 1.36 (s, 9H). ¹³C NMR (100 MHz, Chloroform-d) δ 159.35, 156.12, 149.21, 136.48, 122.30, 121.98, 78.82, 77.48, 77.16, 76.84, 54.91, 47.10, 38.93, 29.72, 28.42.). HRMS (TOF MS ES−) m/z calcd for C₁₄H₂N₃O₂ [M-H]− 265.1790 found 265.1780.

Intermediate 8

Compound 11 (55%, 0.86g) was prepared as off-white solid by complexation with potassium tetra chloroplatinate (1.60, 4.41 mmol) with pyridine derivative 7 (0.780, 2.94 mmol) in DMF- H₂O (24 mL) following general procedure C

¹H NMR (500 MHz, Methanol-d4) δ 9.08 (d, J = 5.8 Hz, 1H), 8.08 (td, J = 7.8, 1.6 Hz, 1H), 7.63 (d, J = 7.9 Hz, 1H), 7.49 – 7.29 (m, 1H), 6.90 – 6.75 (m, 1H), 4.53 (dd, J = 16.3, 5.7 Hz, 1H), 4.13 (dd, J = 16.3, 2.8 Hz, 1H), 3.11 (m, 2H), 2.89 (dt, J = 10.6, 5.6 Hz, 2H), 2.17 – 2.01 (m, 1H), 1.89 – 1.76 (m, 1H), 1.40 (s, 9H). ¹³C NMR (150 MHz, DMSO-d6) δ 163.36, 154.99, 146.77, 138.45, 123.83, 121.80, 76.96, 59.62, 51.93, 36.67, 27.63, 27.09. ¹⁹⁵Pt-NMR (85.82 MHz, Methanol-d4) δ: −2147.48. HRMS (TOF MS ES−) m/z calcd for C₁₄H₂₃Cl₂N₃O₂Pt [M-H]− 530.0815, found 530.0734

MPt1 (3)

MPt1 (3.6 mg, 25%) was prepared as a yellow solid from MTMSA (2) (10 mg, 0.0097 mmol) using PyBop (10 mg,0.0194), amine (0.0125mg, 0.0291mmol) and DMF (0.2 mL), following the general procedure D described above

¹H NMR (600 MHz, Methanol-d₄) δ = 9.09 (bs, 0.4H), 8.98 (d, J = 5.7 Hz, 0.5 H), 8.06 (bs, 0.5H), 7.99 (bs, 0.5H), 7.59 (bs, 1H), 7.41 (bs, 0.6H), 7.13-7.10 (overlap, m, 1.33 H), 6.80 (d, J = 6 Hz, 1H), 5.36– 5.34 (overlap, m, 1H), 5.12 – 5.09 (m, 1H), 4.99 (dd, J = 9.8, 2.1 Hz, 1H), 4.84(bs, s, 1H), 4.78 (d, J = 9.5 Hz, 1H), 4.69 (d, J = 8.6 Hz, 1H), 4.62 – 4.56 (m, 1H), 4.55 (d, J = 11.8 Hz, 0.5H), 4.50 (d, J = 16.2 Hz, 0.6H), 4.22 (dd, J = 36.9, 16.0 Hz, 1H), 4.12 (d, J = 4.4 Hz, 1H), 3.90 – 3.88 (m, 2H), 3.78-3.76 (m, 1H), 3.71 – 3.65 (m, 3H), 3.64 – 3.59 (m, 1H), 3.58 – 3.55 (m, 1H), 3.52 (bs, 3H), 3.49 – 3.46 (m, 1H), 3.44 – 3.39 (m, 1H), 3.38-3.33 (m, 1H), 3.21 – 3.11 (m, 2H), 3.06 (p, J = 8.9, 8.4 Hz, 2H), 2.99 (s, 1H), 2.96 – 2.92 (m, 2H), 2.87(bs, 1H), 2.72 – 2.66 (m, 1H), 2.62 – 2.59 (m, 1H), 2.47- 2.42 (m, 2H), 2.32 (bs, 1H), 2.22 – 2.15 (m, 2H), 2.13 (s, 3H), 2.03-1.90 (m,3H), 1.82-174 (m, 2H), 1.62-154 (m, 3H), 1.40 – 1.21 (m, 17H). ¹³C NMR (150 MHz, Methanol-d₄) δ = 204.3, 174.5, 165.0, 164.8, 159.6, 156.5, 149.3, 140.2, 139.8, 135.0, 128.8, 126.67, 125.5, 123.6, 119.3, 111.5, 109.1, 108.36, 102.3, 102.1, 100.2, 99.9, 99.0, 97.3, 81.2, 80.8, 78.1, 77.9, 77.4, 76.6, 76.5, 73.7, 73.6, 73.5, 72.1, 71.9, 71.8, 70.5, 62.3, 60.5, 60.4, 54.3, 54.1 45.2, 44.9, 40.7, 38.3, 38.2, 37.8, 33.2, 33.1, 30.75, 29.10, 28.97, 28.24, 28.12, 27.3, 18.9, 18.8, 18.3, 17.1, 8.64. ¹⁹⁵Pt-NMR (85.82 MHz, Methanol-d4) δ: −2072.6. HRMS (TOF MS ES−) m/z calcd for C₅₈H₈₃Cl₂N₃O₂₂Pt, [M-H]⁻ 1438.4415, found 1438.4511.

Cell culture media and materials

Ovarian cancer cell lines NIH:OVCAR-3 (OVCAR3) (ATCC^® HTB-161^™), Caov-3 (ATCC^® HTB-75^™) and UWB1.289 (ATCC^® CRL-2945^™) were purchased directly from ATCC and absence of mycoplasma was confirmed independently. All cells were maintained as subconfluent monolayers at 37 °C, 5% CO₂. OVCAR3 cells were grown in high glucose (4500 mg/L) RPMI-1640 (ATCC) with 0.01 mg/mL bovine insulin (MilliporeSigma) and 20% fetal bovine serum (FBS) (MilliporeSigma). Caov-3 cells were grown in modified Dulbecco’s Modified Eagle’s Medium (DMEM) (ATCC) containing 4500 mg/L glucose, 4 mM L-glutamine, 1 mM sodium pyruvate, 1500 mg/L sodium bicarbonate supplemented with 10% FBS. UWB1.289 cells were grown in 1:1 (v:v) mixture of RPMI-1640 (with 2 mM L-glutamine) (Lonza) and mammary epithelial growth medium (MEGM) (PromoCell) and supplemented with 3% FBS. Platinum-resistant UWB1.289 cells (UWB1-PlatR) were established from parental cells by incubating in growth media containing 5 μM carboplatin, in recurrent 48 hour cycles. Following each 48 hour treatment, cells were allowed to recover in platinum-free growth media for 2 to 3 weeks, until typical proliferation resumes, and culture vessel becomes repopulated. A total of 6 cycles were used to establish UWB1-PlatR cells and the stability of the resistant phenotype was verified for a minimum of 8 passages following 6^th treatment cycle.

The DNA binding assay.

The DNA oligomers used in this study were purchased from Integrated DNA Technologies. A DNA oligomer 5’-AATGAGCTCATT-3’ was annealed with a complementary oligomer, as described previously.^[14] In the titration MPt1 and MTM (added from DMSO stocks) were held at the constant concentration of 5 μM in dimers and the DNA concentration was varied in the range of 0-15 μM. The DNA binding assay was carried out in 96-well plates in 40 mM Hepes pH 7.0, 100 mM NaCl and 5 mM MgCl₂ (or in without 5 mM MgCl₂, as specified) at 21 °C. The fluorescence of MPt1 and MTM was measured and analyzed to obtain the values for the equilibrium binding constant K_d and the 1:1 isotherm fit to the data as we described previously.^[14]

The DNA crosslinking assay.

A DNA oligomer 5’-6FAM-CCGGAAGCA AGCAAGGAAGTG-3’ bearing a 5’-6-carboxyfluorescein (6FAM) label and a complementary unlabeled oligomer were annealed as described previously.^[14] This ds DNA was incubated at 2 μM with MPt1 or cisplatin at either 4 μM or 10 μM in 10 μL reactions in 1 mM Na cacodylate, pH 6.0 for 16 hours at 37 °C. The reaction was terminated by adding 10 μL of the 2× loading solution (0.05% bromophenol blue, 0.05% xylene cyanol, 95% formamide, 20 mM EDTA), and heated at 95 °C for 5 min and transferred to ice. 10 μL of each sample was loaded onto the pre-ran 7 M urea-16% polyacrylamide gel (18 cm × 24 cm) and continued running in 0.5× TBE (44.5 mM Tris, 44.5 mM Borate Acid and 1 mM EDTA) at 20 W constant power for 1.5 hours. The gel was scanned directly on Typhoon FLA 9000 (GE Healthcare) in the fluorescence mode.

The IC₅₀ assay

Caov-3, OVCAR3, UWB1.289 and UWB1-PlatR ovarian cancer cells were seeded in white-walled 96-well microplates at 5x10³ cells per well in 100 μL growth media and incubated for 24 hours at 37 °C, 5% CO₂ to allow cells to attach. The growth media was then removed and replaced with fresh media containing serially diluted compounds or blank media for untreated controls. Within an experiment, each drug concentration was tested in duplicate. MPt1 concentrations ranged from 0.5 nM to 3.8 μM, cisplatin from 0.2 μM to 50 μM; mithramycin from 0.5 nM to 500 nM. Cells were incubated with compound at 37°C, 5% CO₂ for 72 hours. Cell viability was then assessed using the CellTiter-Glo 2.0 viability assay (Promega), with luminescence measured using a Varioskan LUX multimode microplate reader (ThermoFisher Scientific). Luminescence was used to determine % viability of drug-treated relative to untreated control cells. Dose-response curves were fit to the data (four parameter log-logistic model) and IC₅₀ values were calculated with R statistical software (version 3.6.1), package drc (version 3.0).^[29] Each drug was analyzed with each cell line using data from a minimum of three independent experiments. Statistical analyses were performed using GraphPad Prism (version 5.01). One-way analysis of variance and Bonferroni’s multiple comparison tests were used to assess differences in cisplatin, MPt1 and MTM IC₅₀ across cells lines and differences in cisplatin versus MPt1 IC₅₀ within each cell line.

Drug synergy analysis

Ovarian cancer cells (5x10³ cells per well) were seeded in white-walled 96-well microplates and incubated overnight at 37 °C, 5% CO₂. Five serially diluted concentrations of cisplatin (50, 16.7, 5.6, 1.9, 0.6, 0 μM) and mithramycin (500, 100, 20, 4, 0.8, 0 nM) were tested alone and in combination using a 6x6 matrix design. Cells were treated for 72 hours, and cell viability was measured using CellTiter-Glo 2.0 viability assay (Promega) and a Varioskan LUX multimode microplate reader (ThermoFisher Scientific). The percentage of viable cells relative to vehicle treated control cells at each of the drug combination levels was used to assess drug synergy using the Bliss independence model within the synergyfinder package (version 1.10) in R [PMID: 29344898].

DNA double-strand break assay

Ovarian cancer cells (Caov-3, OVCAR3, UWB1.289, UWB1-PlatR) were seeded 5x10³ cells per well of black-walled, clear-bottom 96-well plates (ThermoFisher Scientific) in 100μL complete growth media. Cells were allowed to attach overnight at 37° C with 5% CO₂. Media was then removed and replaced with media containing 15 μM MPt1, 0.2 μM mithramycin, 15 μM cisplatin (positive control), 1% ethanol, 0.1% DMSO (v:v) in complete media or blank complete media (vehicle controls). Cells were incubated another 48 hours at 37° C with 5% CO₂. Following treatment, cells were fixed for 15 minutes at room temperature with 4% paraformaldehyde, permeabilized for 15 minutes at room temperature with 0.25% Triton X-100 and blocked for 1 hour at room temperature with 0.1% (w/v) bovine serum albumin in D-PBS. Nuclei were labeled with Hoechst 33258 and double-strand DNA breaks were assessed using the HCS DNA Damage Kit (Invitrogen), with a primary antibody targeting phosphorylated histone H2AX (γH2AX) and an AlexaFluor 555-conjugated secondary antibody. Cells were imaged with the CellInsight CX7 High Content Analysis Platform (ThermoFisher Scientific) and nuclear γH2AX was quantified using HCS Studio software (ThermoFisher Scientific). Signal intensities were normalized to vehicle-treated control and expressed as fold-change relative to control. Signal differences were analyzed using GraphPad Prism (version 5.01). One-way analysis of variance and Bonferroni’s multiple comparison tests were used to determine if treatment-induced fold-change in γH2AX signal within each cell line was significantly different from vehicle-treated control cells.

Caspase 3/7 assay

Ovarian cancer cells (Caov-3, OVCAR3, UWB1.289, UWB1-PlatR) were seeded 5x10³ cells per well of black-walled, clear-bottom 96-well plates (ThermoFisher Scientific) in 100μL complete growth media and allowed to attach overnight. Media was removed and replaced with media containing 15 μM MPt1, 0.2 μM mithramycin, 15 μM cisplatin, or vehicle control. Cells were incubated another 24, 48 or 72 hours at 37°C with 5% CO₂. Following treatment, cells were incubated for 45 minutes at 37°C with 5 μM CellEvent Caspase-3/7 Green Detection Reagent (Invitrogen) and 1.5 μg/mL Hoechst 33342 in D-PBS containing 5% FBS. Reagent removed and cells washed with D-PBS before fixing for 15 minutes at room temperature with 4% paraformaldehyde (ThermoFisher Scientific). Fixative removed and replaced with 100 μL D-PBS. Cells were imaged with the CellInsight CX7 High Content Analysis Platform (ThermoFisher Scientific) and nuclear fluorescence (Ex 502nm/Em 530nm) was quantified using HCS Studio software (ThermoFisher Scientific). Signal intensities were normalized to vehicle-treated control and expressed as fold-change relative to control. Signal differences were analyzed using GraphPad Prism (version 5.01). At each time point, two-way analysis of variance and Bonferroni’s multiple comparison tests were used to determine if treatment-induced fold-changes in fluorescence for each cell line was significantly different from vehicle-treated controls.

Scheme 2. Oxidative cleavage procedure to synthesize MTM SA from mithramycin.

Reagents and conditions: NaIO₄ (1.8 eq), KMnO₄ (0.2 eq), acetone-water (1:1.3), 5 °C, 22h