Novel Nickel(II), Palladium(II), and Platinum(II) Complexes with O,S Bidendate Cinnamic Acid Ester Derivatives: An In Vitro Cytotoxic Comparison to Ruthenium(II) and Osmium(II) Analogues

(1) Background: Since the discovery of cisplatin’s cytotoxic properties, platinum(II) compounds have attracted much interest in the field of anticancer drug development. Over the last few years, classical structure–activity relationships (SAR) have been broken by some promising new compounds based on platinum or other metals. We focus on the synthesis and characterization of 17 different complexes with β-hydroxydithiocinnamic acid esters as O,S bidendate ligands for nickel(II), palladium(II), and platinum(II) complexes. (2) Methods: The bidendate compounds were synthesized and characterized using classical methods including NMR spectroscopy, MS spectrometry, elemental analysis, and X-ray crystallography, and their cytotoxic potential was assessed using in vitro cell culture assays. Data were compared with other recently reported platinum(II), ruthenium(II), and osmium(II) complexes based on the same main ligand system. (3) Results: SAR analyses regarding the metal ion (M), and the alkyl-chain position (P) and length (L), revealed the following order of the effect strength for in vitro activity: M > P > L. The highest activities have Pd complexes and ortho-substituted compounds. Specific palladium(II) complexes show lower IC50 values compared to cisplatin, are able to elude cisplatin resistance mechanisms, and show a higher cancer cell specificity. (4) Conclusion: A promising new palladium(II) candidate (Pd3) should be evaluated in further studies using in vivo model systems, and the identified SARs may help to target platinum-resistant tumors.


Introduction
Cisplatin was first synthesized by M. Peyrone in 1845, and its anticancer properties were discovered accidentally by B. Rosenberg and coworkers in 1965 [1]. Rosenberg s discovery led to the approval of the drug by the FDA in 1979 [2]. The proposed mechanism of action involves binding to its main target (DNA) through the specific DNA-base guanine, and the formation of intra-and inter-strand adducts. Adducts lead to distortions of the helical DNA structure, DNA damage, the disturbance of DNA replication and transcription, and the activation of several intracellular signal pathways potentially inducing apoptosis [3][4][5][6][7]. Cisplatin-based therapy is limited by its toxic side effects, the low selectivity of the drug, and resistance mechanisms [2,5,8,9]. Therefore, soon after cisplatin's development, compounds may exhibit both DNA-damaging activity and also a DNA-independent mode of action, e.g., reactive oxygen species (ROS) induction [41,42].
Applying the well-accepted approach of designing potential metal-based anticancer drugs with SARs other than cisplatin, we report on new platinum(II), palladium(II), and nickel(II) complexes with β-hydroxydithiocinnamic acid esters as bidendate O,S-chelating ligands. Our aim was to determine and compare the activity of these non-classical complexes and to identify the most suitable β-hydroxydithiocinnamic acid ester as ligand. We previously reported on the synthesis of this group of compounds in general [43][44][45][46][47][48], and, in the present work, add new insights into their cytotoxic activity, as well as their characteristics, including molecular structures and stability determinations. Recently, a novel mixed platinum(II) complex with an O,S-chelating ligand and the general formula [Pt(PPh 3 ) 2 (L-O,S)]PF 6 (L-O,S = N,N-morpholine-N -benzoylthiourea) has been synthetized and tested, and proves to be active against tumor cells. The choice of the β-hydroxydithiocinnamic acid ester as the ligand system is based on our described promising results for ruthenium(II) and osmium(II) complexes also bearing this O,S-bidendate ligand [49,50]. Figure 1 shows an overview of the compounds discussed in this work. Both platinumsensitive and -resistant epithelial ovarian cancer (EOC) cell lines were chosen as models for the in vitro comparison of the compounds' cytotoxic effects. EOC is a leading cause of death in women with gynecologic cancer (approx. 220,000 new cases annually worldwide) [51]. Standard care comprises cytoreductive surgery, combined with chemotherapy using a platinum-based regimen in combination with other cytotoxic drugs, plus molecularly targeted strategies for maintenance therapy. While EOC is, in the majority of cases, a platinum-sensitive disease, eventually the majority of patients will relapse and develop a platinum resistance. Platinum resistance is the main challenge to a long-lasting successful therapeutic effect, thus contributing to the low five-year survival rate of approximately 40% [51].
Nickel is another metal that can form complexes with organic ligands, but is not as well studied as Pt, Pd, Ru, or Cu compounds for cytotoxic activity. Studies focusing on the comparison of nickel(II) complexes and their platinum, palladium, and copper analogues show acceptable but no outstanding cytotoxic activity for nickel complexes [39,40]. Nickel may still have some pharmacological properties which are useful for anticancer drug design because many classes of metalloproteins exhibit nickel-ions [40]. Additionally, Ni compounds may exhibit both DNA-damaging activity and also a DNA-independent mode of action, e.g., reactive oxygen species (ROS) induction [41,42].
Applying the well-accepted approach of designing potential metal-based anticancer drugs with SARs other than cisplatin, we report on new platinum(II), palladium(II), and nickel(II) complexes with β-hydroxydithiocinnamic acid esters as bidendate O,S-chelating ligands. Our aim was to determine and compare the activity of these non-classical complexes and to identify the most suitable β-hydroxydithiocinnamic acid ester as ligand. We previously reported on the synthesis of this group of compounds in general [43][44][45][46][47][48], and, in the present work, add new insights into their cytotoxic activity, as well as their characteristics, including molecular structures and stability determinations. Recently, a novel mixed platinum(II) complex with an O,S-chelating ligand and the general formula [Pt(PPh3)2(L-O,S)]PF6 (L-O,S = N,N-morpholine-N′-benzoylthiourea) has been synthetized and tested, and proves to be active against tumor cells. The choice of the β-hydroxydithiocinnamic acid ester as the ligand system is based on our described promising results for ruthenium(II) and osmium(II) complexes also bearing this O,S-bidendate ligand [49,50]. Figure 1 shows an overview of the compounds discussed in this work. Both platinum-sensitive and -resistant epithelial ovarian cancer (EOC) cell lines were chosen as models for the in vitro comparison of the compounds' cytotoxic effects. EOC is a leading cause of death in women with gynecologic cancer (approx. 220,000 new cases annually worldwide) [51]. Standard care comprises cytoreductive surgery, combined with chemotherapy using a platinum-based regimen in combination with other cytotoxic drugs, plus molecularly targeted strategies for maintenance therapy. While EOC is, in the majority of cases, a platinum-sensitive disease, eventually the majority of patients will relapse and develop a platinum resistance. Platinum resistance is the main challenge to a long-lasting successful therapeutic effect, thus contributing to the low five-year survival rate of approximately 40% [51].  Generally, characterization with 1 H and 13 C{ 1 H} NMR spectroscopy shows results comparable to those published previously for similar Ni, Pd, and Pt complexes [46]. Table 1 displays compound M2 as an example, showing four chosen signals in the same range for the three metal complexes (M = Ni, Pd, Pt). Compared to L2, a high-field shift is observable for 13 C signal 2, due to the complexation of the metal via the thiocarbonyl carbon, and the resulting shield of the carbon atom. Complexation results in a low-field shift for 13 C signal 4, as the oxygen atom exhibits a σ-donor character. Interesting changes are observable for the methine proton, signal 1. Whereas Ni2, Pd2, Pt2, and PtDMSO2 show a shift to higher ppm values compared to L2, the opposite is shown for Ru2 and Os2 [18,50]. This is potentially caused by the better donor ability of the cymene ligand. The chemical structures of Ru2, Os2, and Pt2 are shown in the Supplementary Materials ( Figure S1). Compared to PtDMSO2, a platinum(II) complex with one L2 as a bidendate ligand, DMSO, and a labile chloride ligand, the 1 H methine signals for Ni2, Pd2, and Pt2 are shifted around 0.2 ppm up to higher field [18].
The mass spectra for all nickel, palladium, and platinum complexes show molecular ions including the unique isotope pattern for Ni, Pd, or Pt, as well as fragments originating from α-cleavages specific to the β-hydroxydithiocinnamic acid esters, as described previously [18].
With the help of 1 H NMR spectroscopy, the stability of the complexes was studied. We did not detect any degradation for Ni and Pd complexes, and only minor degradation for Pt complexes. Experiments were carried out at room temperature using DMSO-d6 or dichloromethane as solvents, and at 37 °C in DMSO-d6, showing the same results. Examples (37 °C, DMSO-d6, 48 h measurements) are shown in the Supplementary Materials ( Figure S2).  [18] or [50] with permission of the Royal Society of Chemistry and from the authors, respectively.
Generally, characterization with 1 H and 13 C{ 1 H} NMR spectroscopy shows results comparable to those published previously for similar Ni, Pd, and Pt complexes [46]. Table 1 displays compound M2 as an example, showing four chosen signals in the same range for the three metal complexes (M = Ni, Pd, Pt). Compared to L2, a high-field shift is observable for 13 C signal 2, due to the complexation of the metal via the thiocarbonyl carbon, and the resulting shield of the carbon atom. Complexation results in a low-field shift for 13 C signal 4, as the oxygen atom exhibits a σ-donor character. Interesting changes are observable for the methine proton, signal 1. Whereas Ni2, Pd2, Pt2, and PtDMSO2 show a shift to higher ppm values compared to L2, the opposite is shown for Ru2 and Os2 [18,50]. This is potentially caused by the better donor ability of the cymene ligand. The chemical structures of Ru2, Os2, and Pt2 are shown in the Supplementary Materials ( Figure S1). Compared to PtDMSO2, a platinum(II) complex with one L2 as a bidendate ligand, DMSO, and a labile chloride ligand, the 1 H methine signals for Ni2, Pd2, and Pt2 are shifted around 0.2 ppm up to higher field [18].
The mass spectra for all nickel, palladium, and platinum complexes show molecular ions including the unique isotope pattern for Ni, Pd, or Pt, as well as fragments originating from α-cleavages specific to the β-hydroxydithiocinnamic acid esters, as described previously [18].
With the help of 1 H NMR spectroscopy, the stability of the complexes was studied. We did not detect any degradation for Ni and Pd complexes, and only minor degradation for Pt complexes. Experiments were carried out at room temperature using DMSO-d 6 or dichloromethane as solvents, and at 37 • C in DMSO-d 6 , showing the same results. Examples (37 • C, DMSO-d 6 , 48 h measurements) are shown in the Supplementary Materials ( Figure S2).
In addition, we carried out stability measurements for Ni3, Pd3, and Pt3 (100µM) using UV-VIS spectroscopy in different buffers at room temperature (Suppl. Figure S3). Measurements in 100% DMSO confirmed the stability determined by NMR. In the analyzed buffers (10% DMSO with: PBS, 120 mM NaCl, 12 mM NaCl), the compounds precipitated within the analyzed time span (11 h), resulting in a general decrease in absorbance. However, we did not observe strong changes in the spectra that would point to a decomposition of the compounds. In both NaCl buffers there was a slight absorbance increase at higher wavelengths, and this effect was stronger in Pd3 than in Pt3. Interestingly, the compounds did not precipitate in 10% DMSO supplemented with bovine serum albumin (BSA), potentially due to protein binding. This effect could prevent precipitation in the cell culture medium (supplemented with fetal calf serum), and may result in a steady release of the compounds over time.

Molecular Structures
Nickel(II) complexes Ni1, Ni3, Ni4, and Ni6, as well as palladium(II) complex Pd1, were characterized by means of single crystal X-ray structure determination. Figure 2 and Table 2 show the molecular structures and characteristics of Ni1 and Pd1. Data for the other nickel(II) complexes and further bond lengths and angles, as well as data for PtDMSO8, are shown in the Supplementary Materials ( Figures S4 and S5, Tables S1 and S2).  [18] or [50] with permission of the Royal Society of Chemistry and from the authors, respectively.
In addition, we carried out stability measurements for Ni3, Pd3, and Pt3 (100µM) using UV-VIS spectroscopy in different buffers at room temperature (Suppl. Figure S3). Measurements in 100% DMSO confirmed the stability determined by NMR. In the analyzed buffers (10% DMSO with: PBS, 120mM NaCl, 12mM NaCl), the compounds precipitated within the analyzed time span (11 h), resulting in a general decrease in absorbance. However, we did not observe strong changes in the spectra that would point to a decomposition of the compounds. In both NaCl buffers there was a slight absorbance increase at higher wavelengths, and this effect was stronger in Pd3 than in Pt3. Interestingly, the compounds did not precipitate in 10% DMSO supplemented with bovine serum albumin (BSA), potentially due to protein binding. This effect could prevent precipitation in the cell culture medium (supplemented with fetal calf serum), and may result in a steady release of the compounds over time.

Molecular Structures
Nickel(II) complexes Ni1, Ni3, Ni4, and Ni6, as well as palladium(II) complex Pd1, were characterized by means of single crystal X-ray structure determination. Figure 2 and Table 2 show the molecular structures and characteristics of Ni1 and Pd1. Data for the other nickel(II) complexes and further bond lengths and angles, as well as data for PtDMSO8, are shown in the Supplementary Materials ( Figure S4 and S5, Table S1 and S2). The bond lengths and angles of the nickel(II) and palladium(II) complexes are in good agreement with previously reported values [18,46]. As the structures are quite symmetric, all bond lengths and angles are in the same range for both β-hydroxydithiocinnamic acid esters cis-coordinated around the square-planar metal(II) center. Therefore, only one value is chosen for each discussion.  The bond lengths and angles of the nickel(II) and palladium(II) complexes are in good agreement with previously reported values [18,46]. As the structures are quite symmetric, all bond lengths and angles are in the same range for both β-hydroxydithiocinnamic acid esters cis-coordinated around the square-planar metal(II) center. Therefore, only one value is chosen for each discussion.

Biological Behavior
A further aim of this study was to characterize all metal complexes for their cytotoxic properties in vitro, and to determine structure-activity relationships. Therefore, all compounds were tested against a panel of cancer cell lines with different sensitivity to cisplatin: the ovarian cancer cell lines SKOV3/SKOV3cis and A2780/A2780cis [52,53], and lung cancer cell line A549. Selected compounds were additionally tested against non-cancerous cells: keratinocytes, fibroblasts, and MCF10A. Due to the low solubility of the new compounds in water, DMSO was used as a solvent. The toxic influence of DMSO was determined earlier, and experiments were carried out with 0.5 % DMSO in cell culture media and used as a reference in MTT assays (see Section 3) [18]. The conditions of these experiments were the same as for PtDMSO, Ru(II), and Os(II), and have been published [18,49,50]. Thus, alongside comparisons of different metals and different substitution patterns of the ligands described in this paper, comparisons to the other already published systems are possible. Additionally, the IC 50 values of the different β-hydroxydithiocinnamic acid ester ligands have previously been evaluated [50].
All IC 50 values for the 17 metal(II) complexes, as well as the reference cisplatin (CDDP), are displayed in Table 4, and an exemplary dose-response curve is shown in Supplementary  Figure S6 50 values for each cell line shows the highest activity for compound Pd3, which is more active than cisplatin on four of the five cell lines, resulting in the lowest mean IC 50 value ( Figure 3A). In Table 4, all IC 50 values lower than IC 50 of the reference cisplatin are marked in red, highlighting that palladium complexes specifically exhibit high activity. In addition to the single IC 50 values (Table 4), we calculated compound-specific or metal-specific mean values for all or specific groups of cell lines (Figures 3 and 4). The high variance of mean values caused by the heterogeneity of cell-line-specific sensitivity prevents significant differences. However, this may reflect the clinical situation if no predictive biomarker is available, and comparisons of metal-specific mean values provide information about the general effects of different metal ions. Figure 3 depicts all compounds, ordered by increasing mean IC 50 values. It can be concluded that Pd3, Pd4, and Pd1 are the most active compounds included in this study, followed by Ni1 and Pt4. Pd compounds exhibit a mean IC 50 value (all six compounds, all five cell lines) that is lower compared to the mean IC 50 for cisplatin (all five cell lines) pointing to a generally high cytotoxic activity of these complexes ( Figure 4A). Moreover, it is shown in Figure 4B that the mean IC 50 value on both resistant cell lines for Ni and Pd complexes are lower than that of cisplatin. These compounds (i.e., Pd) act specifically on the resistant cell lines, and may be an alternative option for cisplatin resistant tumors in anticancer therapy. This points to another mode of action for both Pd and Ni compounds (for discussion, see below). Both the high cytotoxic activity of Pd compounds against cisplatin-resistant cell lines and their non-superior activity against cisplatin-sensitive cell lines have previously been described [54][55][56][57][58]. Moreover, a multinuclear Pd(II) complex can potentially improve the treatment of cancer stem cells that are more resistant to platinum [59].   The most active compounds for each metal, Pd3, Ni1, and Pt4, have been tested against non-cancerous cells' keratinocytes, fibroblasts, and MCF10a to evaluate their toxicity in general (Table 5). It is known that cisplatin shows toxic side effects by interacting with normal proliferating cells. This is proven by our experiments, which show low IC50 values for CDDP on these cells. Specifically, the most active metal(II) complexes included in this work do not attack those cells, and it can be concluded that these complexes may show a higher selectivity for cancer cells. The importance of both the detected high activity (against cisplatin resistant cells) and the high selectivity are potentially affected by the limitations of this study. First, these data have to be validated in vivo to make conclusions about clinical benefit and use. Secondly, the unknown stability in biological systems/buffers limits our knowledge about the active species. Although we measured a high stability of the metal complexes in DMSO-d6 over 48h by NMR spectroscopy (Suppl. Figure S2), earlier data of Pt-complexes with the O,S-bidendate ligand, DMSO, and chloride showed a reduced stability in biological buffer solutions [18]. Thus, As mentioned above (see Introduction), Huq and coworkers reported some general structure-activity relationships (SARs) for palladium(II) complexes [21] in 2016. They proposed a higher activity for ortho-substituted phenyl rings. The top five compounds (regarding the mean IC 50  as most active. Moreover, calculating the percentage of activity relative to the mean of specific substance groups (e.g., relative percentage of activity of Ni compounds with orthosubstituted methoxy, or ethoxy ligand relative to the mean IC 50 of all Ni compounds with methoxy or ethoxy ligands, respectively) proves the higher activity of ortho-substituted complexes ( Figure 3B). Metal complexes with ortho-substitution are significantly more active than para-or meta-substituted ones. This difference in activity must be related to the behavior of the complexes because the ligands themselves have similar activities ( Figure 3B). Thus, the presented data support the results regarding SARs for Pd complexes from Huq et al. [21] and point to similar relationships for other metal compounds (e.g., Ni, Pt). Regarding the length of the chain (methoxy vs. ethoxy group), there is no clear correlation seen for the metal complexes in general. However, Pt complexes with longer alkyl chains (ethoxy group) are significantly more active than the complexes with a methoxy residual ( Figure 3C). The same effect is seen for the β-hydroxydithiocinnamic acid esters where compounds 4-6 (ethoxy group) are significantly more active than 1-3 (methoxy group; Figures 3C and 4C). Nevertheless, the activity of specific compounds is affected by the combination of all characteristics. The effect strength seems to be metal ion > substitution position > alkyl-chain length (Figure 3), and the most promising candidate compared to cisplatin is Pd3, which bears a para-methoxy group at the phenyl ring.   The most active compounds for each metal, Pd3, Ni1, and Pt4, have been tested against non-cancerous cells' keratinocytes, fibroblasts, and MCF10a to evaluate their toxicity in general (Table 5). It is known that cisplatin shows toxic side effects by interacting with normal proliferating cells. This is proven by our experiments, which show low IC 50 values for CDDP on these cells. Specifically, the most active metal(II) complexes included in this work do not attack those cells, and it can be concluded that these complexes may show a higher selectivity for cancer cells. The importance of both the detected high activity (against cisplatin resistant cells) and the high selectivity are potentially affected by the limitations of this study. First, these data have to be validated in vivo to make conclusions about clinical benefit and use. Secondly, the unknown stability in biological systems/buffers limits our knowledge about the active species. Although we measured a high stability of the metal complexes in DMSO-d6 over 48h by NMR spectroscopy (Suppl. Figure S2), earlier data of Pt-complexes with the O,S-bidendate ligand, DMSO, and chloride showed a reduced stability in biological buffer solutions [18]. Thus, we cannot exclude speciation processes and a certain contribution of specific degradation/speciation products to the biologic activity. Nevertheless, this may not affect the main results. For compound 2 (meta-OMe), a comparison of the β-hydroxydithiocinnamic acid ester (L2), the nickel(II), palladium(II), and platinum(II) complexes of this work (Ni2, Pd2, Pt2), the previously reported platinum(II) complex with one O,S-bidendate ligand, DMSO, and chloride as additional ligands (Ptdmso2), and the corresponding ruthenium(II) and osmium(II) complexes, could be conducted (structures of PtDMSO2, Ru2, and Os2 areshown in Figure S1) [18,50]. Table 6 and Supplementary Figure S7 show the IC 50 values for all 2 compounds, as well as the reference cisplatin on the five cell lines. All metal(II) compounds show lower IC 50 values than the free β-hydroxydithiocinnamic acid ester L2. In general, the ligands themselves exhibit a lower activity than the complexes ( Figure 4C). Compounds Os2 and Pd2 show the best results and lower IC 50 values than cisplatin. For the platinum(II) complex, it can be concluded that it exhibits a lower activity not superior to the reference. However, the data show that the resistance factors for all substances are lower than for cisplatin, proving that the β-hydroxydithiocinnamic acid esters and their metal complexes are able to elude the cisplatin resistance mechanisms of ovarian cancer cell lines in vitro. Moreover, Ni2, Pd2, and Os2 are even more active compared to cisplatin in SKOV3cis, whereas the two most active compounds in A2780cis are Os2 and Pd2, showing lower IC 50 values than the reference substance. Thus, the presented data show that complexes of metals with β-hydroxydithiocinnamic acid ester are not affected by the resistance mechanisms of cisplatin and the compounds likely have another mode of action. Ru(II) complexes exert their activity by impacting the DNA itself, and the mitochondrial activity, autophagy pathway, ROS generation, and ROS mediated apoptosis [23,24,60]. Our analyses point to a DNA-independent induction of cell death, potentially mediated by protein interactions [49,50]. Similarly, the mode of action of Os, Pd, and Ni compounds is described as both DNA-directed and DNA-independent (e.g., ER-stress induction, protein targeting, ROS generation) [28,35,[37][38][39]41,42]. In addition, Ru(II) and Os(II) compound antitumor activity and specificity can be increased by redox modulators [60,61]. However, both Ru(II) and Os(II) compounds can also act independent of ROS by inhibiting protein synthesis [62,63]. Specifically, it can be suggested that the DNA-independent mechanisms are responsible for the high activity against cisplatin resistant cells. These modes of action should be analyzed in detail to improve the treatment of cancer patients. Moreover, metal-based nanoparticles, photoactivated chemotherapy, catalytic active compounds, or complexes with bioactive ligands may lead to new therapeutics and improved outcomes for cancer patients [64][65][66][67][68].

Materials and Techniques
For NMR spectroscopy, a Bruker Avance 200 MHz, 400 MHz, or 600 MHz spectrometer was used. Chemical shifts referring to SiMe 4 are stated in ppm. Mass spectra were measured with a Finnigan SSQ 710 single quadrupole mass spectrometer operating with the direct electron ionoization at 70 eV. Elemental composition was detected with a Leco CHNS-932 apparatus. For column chromatography, Silica gel 60 (0.015-0.040 mm) was used, and TLC was performed using Merck TLC aluminum sheets (Silica gel 60 F 254 ). Chemicals were ordered from Aldrich, Acros, or Fisher Scientific, and were used without additional purification. Prior to use, all solvents were dried and distilled according to standard procedures.

Stability Determinations
NMR spectra were measured on a Bruker Avance 400 MHz system. Substances were solved in DMSO-d 6 or CD 2 Cl 2 and measured directly at 37 • C or room temperature for 72 h. NS = 128 scans, t = 709 s/2891 seconds break, 72 measurements.
For UV-VIS spectroscopy, a JASCO UV-VIS V-760 spectrometer was used. Spectra were measured between 240 and 800 nm at 1 nm steps with a scan speed of 400 nm/min. Compound measurements at 100 µM concentration were normalized to the respective buffer.

Biological Assays
Cell cultures were kept under standard conditions (5 % CO 2 , 37 • C, 90% humidity) in an RPMI medium with 10% FCS, 100 µg/mL streptomycin, and 100 U/mL penicillin (Life Technologies, Darmstadt, Germany). Reference cisplatin (Sigma, Taufkirchen, Germany) was dissolved freshly in 0.9% NaCl solution at a concentration of 1 mg/mL, and diluted appropriately. Described metal(II) complexes and their ligands were dissolved in DMSO. Platinum-resistant A2780 and SKOV3 cells were established as described [18]. IC 50 values were determined using the CellTiter96 non-radioactive proliferation assay (MTT assay, Promega, Walldorf, Germany). A total of 5000 cells were allowed to attach per well of 96-well plates for 24 h and treated for 48 h with different concentrations of the substances (ligands tests: 0, 1, 10, 50, 100, 500, 1000 µm) and for cisplatin and metal complexes from 0 to 100 µM (0.1, 1, 5, 10, 50, 100 µM). Each measurement was performed in triplicate and repeated three times. The amount of metabolic active cells was quantified using the MTT assay. Relative values compared to the mean of medium controls were calculated after background subtraction. Non-linear regression analyses were conducted in GraphPad 5.0 software using the Hill slope.

Conclusions
Overall, we report on 17 novel metal complexes with O,S ligands, and include a comparison with previously reported results on other platinum(II) molecules, as well as ruthenium(II) and osmium(II) counterparts [18,49,50]. The bidendate compounds were characterized using classical methods, including NMR spectroscopy, MS spectrometry, elemental analysis, and some molecular structures. Stability determinations show stable compounds in DMSO for palladium(II) and nickel(II) complexes; molecular structures show a cis-geometry for all square-planar measured metal(II) complexes. The comparison of NMR spectra and molecular structures shows both characteristic changes after complexation of the β-hydroxydithiocinnamic acid esters to the metal(II) center, resulting in an elongation of the -C-S bonds of the thiocarbonyl groups, and a shortening of the -C-O-bonds. SAR analyses regarding the metal ion (M), the alkyl-chain position (P), and the length (L) revealed the following order of effect strength for in vitro activity: M > P > L. In general, the highest activities have Pd complexes and ortho-substituted compounds. The analysis of IC 50 values shows promising results for the palladium(II) complexes, as some of them show lower values compared to cisplatin and are able to elude cisplatin resistance mechanisms in ovarian cancer cell lines. Therefore, the most active compound Pd3 will be further investigated in vivo.