Synthesis, Characterization and Biological Investigation of the Platinum(IV) Tolfenamato Prodrug–Resolving Cisplatin-Resistance in Ovarian Carcinoma Cell Lines

The research on the anticancer potential of platinum(IV) complexes represents one strategy to circumvent the deficits of approved platinum(II) drugs. Regarding the role of inflammation during carcinogenesis, the effects of non-steroidal anti-inflammatory drug (NSAID) ligands on the cytotoxicity of platinum(IV) complexes is of special interest. The synthesis of cisplatin- and oxaliplatin-based platinum(IV) complexes with four different NSAID ligands is presented in this work. Nine platinum(IV) complexes were synthesized and characterized by use of nuclear magnetic resonance (NMR) spectroscopy (1H, 13C, 195Pt, 19F), high-resolution mass spectrometry, and elemental analysis. The cytotoxic activity of eight compounds was evaluated for two isogenic pairs of cisplatin-sensitive and -resistant ovarian carcinoma cell lines. Platinum(IV) fenamato complexes with a cisplatin core showed especially high in vitro cytotoxicity against the tested cell lines. The most promising complex, 7, was further analyzed for its stability in different buffer solutions and behavior in cell cycle and cell death experiments. Compound 7 induces a strong cytostatic effect and cell line-dependent early apoptotic or late necrotic cell death processes. Gene expression analysis suggests that compound 7 acts through a stress-response pathway integrating p21, CHOP, and ATF3.

However, the conjugation of FEL and the fenamates to a platinum(IV) complex and their effect on the cytotoxic activity has not been investigated. Thus, the synthesis and characterization of novel platinum(IV) complexes of MEF, FLU and TOLF as well as monoand disubstituted oxaliplatin-based platinum(IV) complexes of FEL (Scheme 1) were shown. The cytotoxicity of the free NSAIDs and the complexes was tested on isogenic pairs of cisplatin-sensitive and -resistant ovarian carcinoma cell lines. Cell death and cell cycle experiments were performed with the most promising platinum(IV) complex 7 containing TOLF, which was further analyzed for its stability in different buffer solutions. The ability of the (reduced) complex 7 to inhibit COX was examined in vitro. Finally, the influence of 7 on gene expression in cancer cells was examined.
The epithelial ovarian cancer (EOC) cell lines used in this study allow analysis of resistance mechanisms and are of clinical relevance because the majority of ovarian cancer patients eventually develop resistance to platinum compounds [72,73] contributing to a low survival rate below 40% [74].

Synthesis of the Active Esters 1-4
Due to the low reactivity of carboxylic acids, the NSAIDs were activated by N,Ndicyclohexylcarbodiimide (DCC) and 1-hydroxypyrrolidine-2,5-dione (NHS) in chloroform to form active esters with succinimidyl moiety as a leaving group (Scheme 1) [75].

Synthesis of the Active Esters 1-4
Due to the low reactivity of carboxylic acids, the NSAIDs were activated by N,Ndicyclohexylcarbodiimide (DCC) and 1-hydroxypyrrolidine-2,5-dione (NHS) in chloroform to form active esters with succinimidyl moiety as a leaving group (Scheme 1) [75]. The byproduct dicyclohexylurea is insoluble in chloroform and can be filtered off. The active esters 1-4 were purified by column chromatography (chloroform/cyclohexane) and were characterized by nuclear magnetic resonance (NMR) spectroscopy, mass spectrometry, and elemental analysis. The structures of the ligands are shown in Scheme 1. The active ester formation was monitored by 1 H and 13 C{ 1 H} NMR spectroscopy ( Figures S1-S8). The methylene protons of the active ester's succinimidyl moiety appear at 2.84-2.95 ppm, apart from other signals. In the case of ligand 4, besides the phenol protons in the aromatic region, the methylene protons at 3.99 ppm are noticeable. The amine protons of ligands 1-3, as derivatives of anthranilic acid, are shifted to a lower field around 8-9 ppm. The methyl protons of ligands 1 and 3 are detected as singlets at 2.33 and 2.15 ppm (ligand 1, Figure S1) and 2.29 ppm (ligand 3, Figure S5).

Synthesis of the Platinum(IV) Complexes 5-13
Cisplatin was synthesized in several steps from K 2 PtCl 4 according to literature [77]. The oxidation of cisplatin and oxaliplatin to O1 and O2 was performed with hydrogen peroxide (Supplementary Material, page S3) [36,78]. The ensuing reactions with the active esters 1-4 proceeded in dimethyl sulfoxide (DMSO) at 50-65 • C to the monosubstituted platinum(IV) complexes 5-11 [36]. Further modification of complex 11 for the synthesis of disubstituted platinum(IV) complexes was applied by reactions with acid anhydrides in dimethyl formamide (DMF) at room temperature [76]. The acetato group as a non-bioactive ligand [15][16][17] was chosen to enhance the solubility of complex 12 compared to the hydroxido ligand, while phenyl butyric anhydride was chosen as histone deacetylase inhibitor to increase the cytotoxicity of the platinum(IV) complex (13) [41,[79][80][81]. All compounds were purified by column chromatography (chloroform/methanol) and were characterized by NMR spectroscopy, mass spectrometry, and elemental analysis (see Section 3.2 and Supplementary Material, Figures S9-S26). Complexes 5-10 with axial fenamato ligands were obtained in yields from 5 to 15%, depending on the platinum(II) core and the fenamate. Higher yields were achieved using FEL as an axial ligand (33-38% for complexes 11-13). The 195 Pt chemical shifts vary depending on the coordination sphere from around 1020 ppm for cisplatin-based platinum(IV) complexes (5-7), 1400 ppm for monosubstituted oxaliplatin-based platinum(IV) complexes (8)(9)(10)(11), to 1600 ppm for disubstituted complexes 12 and 13. The characteristic amine protons of complexes 5-10 are detected at around 9 ppm as singlet, while the methyl groups of complexes 5, 7, 8, and 10 are shown at around 2 ppm. The resonance signal at 3.57 ppm is assigned to methylene protons of the platinum(IV) complex with axial FEL (11). The further reactions with acetic and phenyl butyric anhydride (12 and 13) lead to splitting of the methylene protons into doublets at 3.64 and 3.65 ppm.

Anticancer Activity
The cytotoxicity of the platinum(IV) complexes 5-8 and 10-13, as well as NSAIDs, was tested on the ovarian carcinoma cell lines A2780par and SKOV3par as well as their cisplatin-resistant analogues A2780cis and SKOV3cis [82]. Complex 9 was not tested due to its nominal yield. Cells were seeded on 96-well plates and were incubated for 24 h in medium. Cisplatin and oxaliplatin were examined as referential platinum(II) complexes. Both compounds were dissolved in 0.9% NaCl solution right before the experiment. The platinum(IV) complexes were dissolved and serially diluted in DMSO, finally diluted in Roswell Park Memorial Institute (RPMI) 1640 medium to reach concentrations from 0.1 to 100 µM, and added to the cells. Despite the cytotoxic effect of DMSO on cells, it is usually used as a solvent in biological tests [83]. To prevent damage to cells, the concentration of DMSO in cell culture experiments was limited to 0.1%. Additional triplets of wells with medium and medium with 0.1% DMSO functioned as control. Half-maximal inhibitory concentration (IC 50 ) values were determined by 2-(4,5-Dimethyl-1,3-thiazol-2-yl)-3,5-diphenyl-3H-1,2λ 5 ,3,4-tetrazol-2-ylium bromide (MTT) assay after an exposure of 48 h from at least three independent experiments ( Table 1). The metabolic activity of  50 values of the complexes 5-7 against 8 and 10, the oxaliplatin-based platinum(IV) complexes are less cytotoxic than the complexes with a cisplatin-core (Table 1). Especially, complexes 5 and 7 feature high antiproliferative activity in the selected cancer cell lines, and they are even more active than cisplatin. Remarkably, 5 and 7 have a higher effect on cisplatin-resistant cell lines (5: resistance factor (RF) 0.9 for A2780 and RF 0.3 for SKOV3; 7: RF 0.5 for A2780 and RF 0.3 for SKOV3, Table 1). The mono-substituted platinum(IV) complex with axial FEL 11 shows low cytotoxicity in A2780 cell lines, whereas the IC 50 values in SKOV3 cell lines are higher than the examined concentration range. Further substitution of 11 with acetic anhydride improves the antiproliferative activity of complex 12 in A2780 cell lines, potentially because of the increased lipophilic character of the acetate group compared to the hydroxido ligand (Table 1) [86][87][88]. Considerably decreased IC 50 values and thus higher activity are achieved by the reaction of 11 with phenyl butyric anhydride (complex 13, Table 1). The phenyl butyrato ligand enhances the lipophilicity and additionally functions as histone deacetylase inhibitor inducing cell death and cell cycle arrest in cancer cells [89,90]. Compound 13, which consists of the cytotoxic oxaliplatin-core, and two different biologically active ligands, belongs to the group of platinum(IV) complexes with triple-action. Hereby, cancer cells are attacked in three different ways, which improves the cytotoxicity compared to the mono-substituted oxaliplatin(IV) complexes (Table 1).
Due to its promising antiproliferative activity in the tested cancer cell lines (also slightly higher than that of complex 5), complex 7 was selected for further biological experiments. Besides a high cytotoxicity against tumor cells, chemotherapeutics should have a low activity against normal cells potentially resulting in reduced side-effects and the application of higher concentrations in vivo. Thus, we tested 7 on primary fibroblast cultures. Compound 7 showed a high cytotoxic activity with an IC 50 value of 0.68 +/− 0.03 µM. The selectivity index (SI = mean IC 50 Fibroblasts /IC 50 EOC ) was 1.47, whereas cisplatin showed an IC 50 of 8.26 +/− 1.1 µM and an SI of 0.58. Compound 7 is slightly more specific for EOC compared to cisplatin. Similar SI were observed for an oxaliplatin-based platinum(IV) prodrug with naproxen [37]. However, additional experiments are necessary to evaluate the selectivity, to detect side effects, and to estimate a clinically useful concentration. These experiments may consist of 3D-co-culture or mouse models [91,92].

Investigations on Stability Behavior
To analyze the stability behavior of complex 7, UV-Vis spectroscopic experiments were performed at 37 • C in phophate-buffered saline (PBS), 120 mM NaCl (extracellular concentration), 12 mM NaCl (intracellular concentration), and DMSO, as well as bovine serum albumin (BSA) protein and salmon sperm DNA over time ( Figure S27). Complex 7 shows overall stability in PBS, 12 mM NaCl, BSA, and DMSO over 23 h. In 120 mM NaCl solution, the absorption decreases over time without precipitation of 7, indicating slow degradation. Slight changes in the absorption are observed in DNA solution, where the absorption maxima around 280 nm are marginally blue shifted.
The reaction of 7 with ten-fold excess of ascorbic acid in aqueous phosphate buffer (5 mM, pH 7.4) was analyzed by ultra-high performance liquid chromatography highresolution mass spectrometry (UHPLC-HRMS). After 5 min, over 40% of the complex is degraded ( Figure S28). The further decomposition of 7 proceeds within 5 h. Simultaneously, the free ligand TOLF and cisplatin are formed progressively during the reaction, confirming the reduction of platinum(IV).

In Vitro Inhibition of COX Activity
The main principle of platinum(IV) prodrugs is based on the intracellular reduction of platinum(IV) resulting in the release of the biologically active axial ligands and generation of the original platinum(II) drug [1,3,15]. The influence of complex 7 and the reduced compound 7 on COX activity was analyzed in vitro. Whereas compound 7 could not inhibit COX-1 activity, the free ligand (TOLF) and the reduced complex 7 inhibited COX-1 activity significantly by 84.4% and 47.4%, respectively ( Figure S29; p < 0.005). Thus, the ligand TOLF must be released after reduction to platinum(II) to inhibit COX. However, it cannot be excluded that the conditions of this in vitro assay introduce bias. Albeit ascorbic acid alone does not inhibit COX activity ( Figure S29) and compound 7 was reduced before the addition to the COX reaction, another interaction of ascorbic acid with the assay system or compound 7 may cause the COX inhibition. Additionally, the used assay seems to be suboptimal because high concentrations (80 µM) were required to inhibit COX-1, and no influence on COX-2 activity was observed at this concentration ( Figure S29). This may be caused by suboptimal buffer conditions or incubation times to reach a high inhibition of COX. Similarly, Khoury et al. recently reported a COX-2 inhibitory effect of aspirin with this commercial kit at high concentrations (700 µM) only [93]. Moreover, in this study the authors observed no strong correlation between the cytotoxicity of platinum(IV) compounds and their lipophilicity or COX-2 inhibition strength [93]. Albeit we can prove the successful COX-1 inhibition by the reduced compound 7 ( Figure S29), IC 50 data from the combination treatment cisplatin + TOLF (Table 1) do not show an increased cytotoxicity compared to the cisplatin treatment alone. Thus, the high cytotoxicity of 7 is potentially not directly related to COX inhibition by TOLF. A COX-independent high cytotoxicity was similarily observed by Ravera et al. analyzing platinum(IV) prodrugs with ketoprofen and naproxen [35]. Thus, the tumor cell-specific effects seem to be independent from COX inhibition. However, COX inhibition has anti-inflammatory effects in vivo that may influence tumor properties.

Cell Death and Cell Cycle Distribution Analyses
Cell cycle and cell death experiments were performed by flow-cytometry assisted by propidium iodide (PI) staining of DNA. Since PI can only enter dead cells and binds stoichiometrically to nucleic acids, the measured emission informs about the number of dead cells and is proportional to the DNA content, thus indicating the cell cycle phase of individual cells [94,95]. The percentage of cell death indicates the fraction of dead cells in the population. As assumed, the percentage of dead cells shows a concentrationdependent increase upon treatment with complex 7 and significant differences compared to untreated cells ( Figure 1A, p < 0.05). SKOV3 cell lines are more strongly influenced by cell death induction and show between 12.1% and 34.1% dead cells. In contrast, A2780 cell lines show lower cell death between 3.8% and 6.2% albeit, treated with similarly effective concentrations as determined by the MTT assay (IC 50 , two-fold IC 50 ). However, the number of PI-negative, single-cell events by flow-cytometry is vigorously decreased after treatment, indicating decreased proliferation induced by cytostatic effects of complex 7 specifically in A2780 ( Figure 1B, p < 0.05). Differences in the cell death or number of alive cells between parental and cisplatin-resistant cells were not significant, whereas cisplatin treatment induced significantly higher cell death in parental compared to resistant cells ( Figure S30, p < 0.01). Thus, complex 7 activity is not affected by cisplatin-resistance mechanisms. Figure 2 shows the cell cycle phases depending on the treatment with compound 7 for each cell line. The majority of untreated cells stays in G1 phase, which decreases continuously with higher concentrations of complex 7. The treatment with at least 0.2 µM compound 7 leads to a significant cell cycle arrest in G2/M in all cell cultures and in the S phase except for A2780par ( Figure 2). Complex 7 influences the cell cycle distribution more strongly in SKOV3 compared to A2780. This may cause the observed higher cell death in SKOV3 vs. A2780 (Figure 1). concentrations as determined by the MTT assay (IC50, two-fold IC50). However, the number of PI-negative, single-cell events by flow-cytometry is vigorously decreased after treatment, indicating decreased proliferation induced by cytostatic effects of complex 7 specifically in A2780 ( Figure 1B, p < 0.05). Differences in the cell death or number of alive cells between parental and cisplatin-resistant cells were not significant, whereas cisplatin treatment induced significantly higher cell death in parental compared to resistant cells ( Figure  S30, p < 0.01). Thus, complex 7 activity is not affected by cisplatin-resistance mechanisms.   phase except for A2780par ( Figure 2). Complex 7 influences the cell cycle distribution more strongly in SKOV3 compared to A2780. This may cause the observed higher cell death in SKOV3 vs. A2780 (Figure 1).

Investigation of Apoptosis/Necrosis
To analyze induced cell death, apoptosis, and necrosis were measured in control and compound 7-treated cell lines using a commercial kit (see Section 3.7). This assay detects apoptotic and necrotic processes by live-cell real-time measurements of luminescence and fluorescence signals reflecting the presence of phosphatidylserine on the cell membrane surface or the accessibility of genomic DNA for a cell-impermeable dye, respectively.
The influence of complex 7 (3 µ M) on cell death of A2780 and SKOV3 cells (both parental and cisplatin-resistant) was analyzed for 70 h (Figure 3). All tested cell lines show cell death by apoptosis followed by secondary necrosis, and no general differences between parental and cisplatin-resistant cells were detected. Higher levels of (early) apoptosis for SKOV3 compared to A2780 cells resemble flow-cytometry-based cell death analyses ( Figure 1). A2780 cells show high levels of necrosis at later time points not analyzed by flow-cytometry. Resistant SKOV3 cells seem to respond earlier and with higher apoptosis induction compared to parental cells, validating the activity of complex 7 against cisplatin-resistant cells (RF < 1; Table 1). The exact cause for the decline of the luminescence signal after 24 h to 30 h is unknown. The commercial system uses a time-released luciferase substrate that must be converted by a cellular esterase from alive cells. Thus, a

Investigation of Apoptosis/Necrosis
To analyze induced cell death, apoptosis, and necrosis were measured in control and compound 7-treated cell lines using a commercial kit (see Section 3.7). This assay detects apoptotic and necrotic processes by live-cell real-time measurements of luminescence and fluorescence signals reflecting the presence of phosphatidylserine on the cell membrane surface or the accessibility of genomic DNA for a cell-impermeable dye, respectively.
The influence of complex 7 (3 µM) on cell death of A2780 and SKOV3 cells (both parental and cisplatin-resistant) was analyzed for 70 h (Figure 3). All tested cell lines show cell death by apoptosis followed by secondary necrosis, and no general differences between parental and cisplatin-resistant cells were detected. Higher levels of (early) apoptosis for SKOV3 compared to A2780 cells resemble flow-cytometry-based cell death analyses (Figure 1). A2780 cells show high levels of necrosis at later time points not analyzed by flow-cytometry. Resistant SKOV3 cells seem to respond earlier and with higher apoptosis 9 of 20 induction compared to parental cells, validating the activity of complex 7 against cisplatinresistant cells (RF < 1; Table 1). The exact cause for the decline of the luminescence signal after 24 h to 30 h is unknown. The commercial system uses a time-released luciferase substrate that must be converted by a cellular esterase from alive cells. Thus, a reduction of the cell number by treatment may led to a decreased conversion of the substrate.

Gene Expression Analyses
The expression of specific genes after drug treatment can give information about its mechanism of action [96,97]. Thus, we measured gene expression of the cell cycle inhibitor p21, of the stress response factor and apoptosis inductor CCAAT/enhancer-binding proteins homologous protein (CHOP, also known as DNA damage inducible transcript 3, DDIT3), and of the activating transcription factor 3 (ATF3) that is involved in various stress responses and a potential COX-independent target of NSAID [98] by real-time polymerase chain reaction (PCR). As shown in Figure 4, complex 7 induces a strong p21 expression both in cisplatin-sensitive and -resistant A2780 cells, whereas cisplatin at IC50 of sensitive cells induced a lower p21 expression in resistant cells. Cisplatin treatment of resistant cells with IC50 concentration induced p21 to identical levels as complex 7 ( Figure  4). The similar p21 overexpression under complex 7 treatment agrees with the observed identical changes of the cell cycle distribution between sensitive and resistant cells after complex 7 treatment (Figure 2). Thus, complex 7 can induce a p21-dependent cell cycle arrest both in cisplatin-sensitive and resistant cells at concentrations ≤1 μM. In addition, complex 7 promotes an upregulation of the apoptosis inducer CHOP that is overexpressed by various genotoxic agents in sensitive cells [99]. The stronger upregulation in SKOV3 compared to A2780 cells resembles the higher levels of apoptosis and cell death in this cell line (Figures 2 and 4). CHOP can be induced during endoplasmatic reticulum stress (ER stress) by ATF4 [100]. However, an increased splicing of Xbp1 indicative of unfolded protein response-induced ER stress [101] was not observed under complex 7 treatment, pointing to genotoxic or direct NSAID-based activation of the ER stress pathway. COX inhibitors can induce ATF3, an ER stress pathway gene mediating apoptosis in colorectal cancer cells and ferroptosis in gastric cancer cells [102][103][104]. Thus, complex 7 may activate the ER stress pathway similarly in ovarian cancer cells. As shown in Figure 4, ATF3 is upregulated by complex 7 both in cisplatin-sensitive and resistant A2780 and SKOV3 cells. This upregulation is increased in resistant cells compared to cisplatin treatment. Cisplatin induces less ATF3 in resistant cells, confirming data from gastric cancer where ATF3 expression mediates cisplatin sensitivity [104]. Moreover, ATF3 induction by compound 7 is concentration-dependent, whereas this effect is not seen for cisplatin at the analyzed concentrations ( Figure 4). Altogether, these experiments point to the contribution of the ER stress

Gene Expression Analyses
The expression of specific genes after drug treatment can give information about its mechanism of action [96,97]. Thus, we measured gene expression of the cell cycle inhibitor p21, of the stress response factor and apoptosis inductor CCAAT/enhancerbinding proteins homologous protein (CHOP, also known as DNA damage inducible transcript 3, DDIT3), and of the activating transcription factor 3 (ATF3) that is involved in various stress responses and a potential COX-independent target of NSAID [98] by realtime polymerase chain reaction (PCR). As shown in Figure 4, complex 7 induces a strong p21 expression both in cisplatin-sensitive and -resistant A2780 cells, whereas cisplatin at IC 50 of sensitive cells induced a lower p21 expression in resistant cells. Cisplatin treatment of resistant cells with IC 50 concentration induced p21 to identical levels as complex 7 (Figure 4). The similar p21 overexpression under complex 7 treatment agrees with the observed identical changes of the cell cycle distribution between sensitive and resistant cells after complex 7 treatment (Figure 2). Thus, complex 7 can induce a p21-dependent cell cycle arrest both in cisplatin-sensitive and resistant cells at concentrations ≤1 µM. In addition, complex 7 promotes an upregulation of the apoptosis inducer CHOP that is overexpressed by various genotoxic agents in sensitive cells [99]. The stronger upregulation in SKOV3 compared to A2780 cells resembles the higher levels of apoptosis and cell death in this cell line (Figures 2 and 4). CHOP can be induced during endoplasmatic reticulum stress (ER stress) by ATF4 [100]. However, an increased splicing of Xbp1 indicative of unfolded protein response-induced ER stress [101] was not observed under complex 7 treatment, pointing to genotoxic or direct NSAID-based activation of the ER stress pathway. COX inhibitors can induce ATF3, an ER stress pathway gene mediating apoptosis in colorectal cancer cells and ferroptosis in gastric cancer cells [102][103][104]. Thus, complex 7 may activate the ER stress pathway similarly in ovarian cancer cells. As shown in Figure 4, ATF3 is upregulated by complex 7 both in cisplatin-sensitive and resistant A2780 and SKOV3 cells. This upregulation is increased in resistant cells compared to cisplatin treatment. Cisplatin induces less ATF3 in resistant cells, confirming data from gastric cancer where ATF3 expression mediates cisplatin sensitivity [104]. Moreover, ATF3 induction by compound 7 is concentration-dependent, whereas this effect is not seen for cisplatin at the analyzed concentrations ( Figure 4). Altogether, these experiments point to the contribution of the ER stress pathway to the biologic activity of complex 7 and to the potential strategy to target this pathway for resolving platinum-resistance. However, additional experiments are necessary to (i) enable a statistical evaluation, which is not possible with the presently available data of n = 2 independent treatments and (ii) to determine the exact mechanism for the activation of the ER stress pathway by compound 7. REVIEW 10 pathway for resolving platinum-resistance. However, additional experiments are n sary to (i) enable a statistical evaluation, which is not possible with the presently ava data of n = 2 independent treatments and (ii) to determine the exact mechanism fo activation of the ER stress pathway by compound 7.

Materials and Techniques
All reactions were carried out under atmospheric conditions. The chemicals an vents used were commercially available and used without further purification. K and oxaliplatin were obtained from Umicore AG & Co. KG (Hanau-Wolfgang, Germ Chemicals were commercially available (TCI, Eschborn, Germany; abcr, Karlsruhe many; Acros Organics, Niderau, Germany; Carl Roth, Karlsruhe, Germany). Solve technical grade were distilled prior to their use. Silica gel (0.063-0.2 mm) was use column chromatography, and thin-layer chromatography (TLC) was carried out TLC aluminum sheets from Merck (Silica gel 60 F254). 1 H NMR, 13 Table S1). Elemental analyses wer formed with a Leco CHNS-932 device (Leco, Mönchengladbach, Germany). UV-Vis troscopic experiments were monitored with a JASCO UV/VIS V-760-ST spectropho ter (JASCO, Pfungstadt, Germany). Absorption spectra were measured from 2 800 nm with 1 nm steps and a scan speed of 400 nm/min. The measurements were malized to the respective buffer. Cell death and cell cycle analyses were performed a BD Accuri TM C6 Plus flow-cytometer (BD Franklin Lakes, NJ, USA). Apoptosis/ne was tested with the RealTime-Glo TM Annexin V Apoptosis and Necrosis Assay (Prom Walldorf, Germany). COX inhibition was examined by COX (ovine/human) Inh Screening Assay by Cayman Chemical (Ann Arbor, MI, USA). Real-time PCR experim were run on a Rotorgene cycler (Qiagen, Hilden, Germany).

Materials and Techniques
All reactions were carried out under atmospheric conditions. The chemicals and solvents used were commercially available and used without further purification. K 2 PtCl 4 and oxaliplatin were obtained from Umicore AG & Co. KG (Hanau-Wolfgang, Germany). Chemicals were commercially available (TCI, Eschborn, Germany; abcr, Karlsruhe, Germany; Acros Organics, Niderau, Germany; Carl Roth, Karlsruhe, Germany). Solvents of technical grade were distilled prior to their use. Silica gel (0.063-0.2 mm) was used for column chromatography, and thin-layer chromatography (TLC) was carried out using TLC aluminum sheets from Merck (Silica gel 60 F 254 ). 1 H NMR, 13 Table S1). Elemental analyses were performed with a Leco CHNS-932 device (Leco, Mönchengladbach, Germany). UV-Vis spectroscopic experiments were monitored with a JASCO UV/VIS V-760-ST spectrophotometer (JASCO, Pfungstadt, Germany). Absorption spectra were measured from 240 to 800 nm with 1 nm steps and a scan speed of 400 nm/min. The measurements were normalized to the respective buffer. Cell death and cell cycle analyses were performed with a BD Accuri TM C6 Plus flow-cytometer (BD Franklin Lakes, NJ, USA). Apoptosis/necrosis was tested with the RealTime-Glo TM Annexin V Apoptosis and Necrosis Assay (Promega, Walldorf, Germany). COX inhibition was examined by COX (ovine/human) Inhibitor Screening Assay by Cayman Chemical (Ann Arbor, MI, USA). Real-time PCR experiments were run on a Rotorgene cycler (Qiagen, Hilden, Germany).
General procedure 1 for the activation of carboxylic acids: Carboxylic acid (1 equiv.), DCC (1.1 equiv.), and NHS (1 equiv.) were dissolved in chloroform and stirred for 2-3 h at room temperature. After filtration, the filtrate was evaporated under reduced pressure, and the raw product was purified by column chromatography (chloroform/cyclohexane).

Stability Studies
Complex 7 was dissolved in 2% DMF in aqueous phosphate buffer (5 mM, pH 7.4) to a concentration of 0.02 mM and was reacted with a 0.2 mM solution of ascorbic acid in aqueous phosphate buffer in a 1:1 mixture. The reaction mixture was incubated at 37 • C while samples for UHPLC-HRMS were repeatedly analyzed over 12 h.

Cell Culture Conditions
Ovarian carcinoma cell lines A2780 and SKOV3, as well as their cisplatin-resistant cell lines A2780cis and SKOV3cis were cultivated in RPMI 1640 medium, supplemented with 10% fetal calf serum, 100 U/mL penicillin, and 100 µg/mL streptomycin (Life Technologies, Darmstadt, Germany) in an incubator (37 • C, 5% CO 2 , 90% humidity). For biological experiments, cells were counted, seeded on different well plates, and incubated overnight at 37 • C to enable cell attachment. Platinum-resistant A2780 and SKOV3 cells were established as described [82] by repeated rounds of 3-day incubations with increasing amounts of cisplatin starting with 0.1 µM. The concentration was doubled after 3 incubations interrupted by recovery phases with normal medium. Cells that survived the third round of 12.8 µM cisplatin were defined as resistant cultures. Resistant cultures were not steadily exposed to cisplatin for maintaining the resistant phenotype avoiding the accumulation of additional (epi-)genetic changes caused by cisplatin. However, early cryo stocks were used and IC 50 values of resistant cultures were stable over time.

Determination of IC 50 Values
Cells were seeded on 96-well plates (5000 cells per well for SKOV3 cell lines, 10,000 cells per well for A2780 cell lines). The platinum(IV) complexes 5-8 and 10-13 and the free NSAIDs were dissolved in DMSO, serially diluted in DMSO, and afterwards diluted in RPMI 1640 medium to receive the concentrations from 0.1 to 100 µM. Cisplatin and oxaliplatin were dissolved in 0.9% NaCl solution right before the experiment and diluted serially in RPMI 1640 medium. The samples were added in 200 µL per well in triplicates and were incubated at 37 • C, 5% CO 2 for 48 h. Viable cells were determined by MTT (Promega, Walldorf, Germany) assays (incubation time 2 h). After subtraction of blank MTT values, relative values were compared to the mean of the medium controls. Non-linear regression analyses using the Hill-slope were accomplished with GraphPad 5.0 software (Dotmatics, Boston, MA, USA).

Analyses with Flow-Cytometry
Cells were seeded on 6-well plates (150,000 cells per well). Complex 7 was dissolved in DMSO and diluted in RPMI 1640 medium to receive concentrations from 0.1 to 2 µM. The samples were added in 3 mL per well and incubated at 37 • C, 5% CO 2 for 48 h. For cell death experiments: Adherent and suspended cells were collected and stained with 1 µg/mL propidium iodide (PI) solution in PBS directly before measurement. For cell cycle experiments: Cells were trypsinized, washed, and fixed in 70% ice-cold ethanol at −20 • C for at least 3 h. After two washing steps with PBS, cells were incubated for 45 min at 4 • C with RNase buffer and stained with 50 µg/mL PI directly before measurement. Flow cytometry measurements analyzed at least 10,000 single cell events.

Real-Time-Glo TM Annexin V Apoptosis and Necrosis Assay (Promega, Catalog no. JA1011)
The assay was performed according to the manufacturer's instructions (Promega, Walldorf, Germany). Cells were seeded on a 96-well plate (5000 cells per well for SKOV3 cell lines, 10,000 cells per well for A2780 cell lines). Complex 7 was dissolved in DMSO and diluted in RPMI 1640 medium to reach the stated concentrations, which were added in 100 µL. Untreated cells in RPMI 1640 medium served as control. Next, 100 µL of the 2X detection reagent (containing 12 mL RPMI 1640 medium, 24 µL Annexin NanoBiT ® substrate, 24 µL CaCl 2 , 24 µL necrosis detection reagent, 24 µL Annexin V-SmBiT and 24 µL Annexin V-LgBiT) were added per well. The plate was shaken for 30 s, and both luminescence and fluorescence measurements were conducted at various time points over 70 h with a Tecan plate reader M200pro. The excitation wavelength was set at 485 nm, the emission wavelength was set to 530 nm.

COX (Ovine/Human) Inhibitor Screening Assay (Cayman Chemical, Item no. 560131)
According to the instructions of the manufacturer (Cayman Chemical, Ann Arbor, MI, USA), the assay was composed of three main parts: COX reactions, standard preparation, and enzyme-linked immunosorbent assay (ELISA). Complex 7 (dissolved in DMSO) and cisplatin (dissolved in 0.9% NaCl solution) were examined for their ability to inhibit COX-1 and COX-2. Both compounds were diluted in RPMI 1640 medium to reach stated concentrations.
Standard preparation: The prostaglandin screening standard was serially diluted in ELISA buffer.
ELISA: The format of the 96-well plate was adopted from the protocol: blank wells, non-specific binding wells (NSB, 100 µL ELISA buffer), maximum binding wells (B 0 , 50 µL ELISA buffer), prostaglandin screening standard (50 µL of each concentration in duplicate), background samples (50 µL in duplicate), COX-1/2 100% initial activity samples (50 µL in duplicate), COX-1/2 inhibitor samples (50 µL in duplicate), and an empty total activity well. Then, 50 µL prostaglandin screening acetylcholinesterase (AChE) tracer was added to all wells except for the blank and total activity wells. Next, 50 µL prostaglandin screening ELISA antiserum was added to each well except for the blank, NSB, and total activity wells. The plate, covered with plastic film, was incubated for 18 h at room temperature on an orbital shaker. The plate was emptied, and each well was rinsed with wash buffer five times. Next, 200 µL of Ellman's reagent in ultrapure water was added to each well, and the AChE tracer (5 µL) was added to the total activity well. The plate, covered with plastic film, was developed for 30 min in the dark at room temperature on an orbital shaker. The absorbance was measured at 410 nm with a Tecan plate reader.
The real-time PCR experiments were run on a Rotorgene cycler (Qiagen, Hilden, Germany). Reactions were performed using the FastStartUniversal SybrGreen Mastermix (Roche Diagnostics, Mannheim, Germany) containing forward and reverse primers (10 pmol each) and cDNA equivalent to 25 ng RNA. Primer-specific data are listed in Table 2. The PCR steps were as follows: initial denaturation and hot start activation at 98 • C for 10 min followed by 40 cycles of denaturation phase at 98 • C for 15 s, primer-specific annealing for 20 s at Ta, and elongation at 72 • C for 40 s. Subsequently, the melting temperature of the PCR product was determined to ensure specificity. Relative target gene expression was normalized to the expression of two housekeeping genes (GAPDH, HPRT) and calculated relative to untreated controls.

Conclusions
This work describes the synthesis and characterization of nine novel cisplatin-and oxaliplatin-based platinum(IV) complexes with different axial NSAID ligands. Whereas the free NSAIDs do not show relevant anticancer activity in the tested ovarian carcinoma cell lines, the corresponding platinum(IV) complexes containing the NSAIDs have significantly increased cytotoxicity. Specifically, the cisplatin-based complexes (5-7) exhibited a high cytotoxicity against both sensitive and resistant cell lines. Compared to the mono-substituted platinum(IV) complexes based on oxaliplatin, the triple-action complex 13 shows improved anticancer activity in all cell lines.
Due to the highest measured cytotoxicity, complex 7 was selected for further biological experiments. These experiments suggest that compound 7 acts independently of cisplatinresistance mechanisms and causes similar molecular effects in cisplatin-sensitive and -resistant cell lines. Compound 7 induces a strong cytostatic effect and cell-line-dependent early apoptotic or late necrotic cell death processes in SKOV3 and A2780, respectively. Gene expression changes under compound 7 treatment agree with these effects and may point to the contribution of ATF3-mediated stress response to the biologic activity.