Arene Ruthenium(II) Complexes Bearing the κ-P or κ-P,κ-S Ph2P(CH2)3SPh Ligand

Neutral [Ru(η6-arene)Cl2{Ph2P(CH2)3SPh-κP}] (arene = benzene, indane, 1,2,3,4-tetrahydronaphthalene: 2a, 2c and 2d) and cationic [Ru(η6-arene)Cl(Ph2P(CH2)3SPh-κP,κS)]X complexes (arene = mesitylene, 1,4-dihydronaphthalene; X = Cl: 3b, 3e; arene = benzene, mesitylene, indane, 1,2,3,4-tetrahydronaphthalene, and 1,4-dihydronaphthalene; X = PF6: 4a–4e) complexes were prepared and characterized by elemental analysis, IR, 1H, 13C and 31P NMR spectroscopy and also by single-crystal X-ray diffraction analyses. The stability of the complexes has been investigated in DMSO. Complexes have been assessed for their cytotoxic activity against 518A2, 8505C, A253, MCF-7 and SW480 cell lines. Generally, complexes exhibited activity in the lower micromolar range; moreover, they are found to be more active than cisplatin. For the most active ruthenium(II) complex, 4b, bearing mesitylene as ligand, the mechanism of action against 8505C cisplatin resistant cell line was determined. Complex 4b induced apoptosis accompanied by caspase activation.


Introduction
One of today's most clinically used antitumor drug cisplatin was synthesized in 1845 by M. Peyrone. However, the structure remained unknown for the next 50 years [1,2]. A. Werner deducted the square planar structure, and the cisplatin was distinguished from the trans analog. Cisplatin was approved in 1978 as an antitumor agent for testicular and ovarian cancers [3][4][5]. A major disadvantage of cisplatin are its strong side effects due to its nephrotoxicity, neurotoxicity and ototoxicity [6,7]. This led to the development for alternative drugs (carboplatin, oxaliplatin, etc.) [8]. However, the side effects of the general high toxic properties of the platinum compounds were not suppressed.
The organoruthenium(II/III) compounds seem to be particularly suitable because of their lower general toxicity in comparison with cisplatin, as well as their ability to utilize iron pathways in the body [21,22]. Keppler, Sadler and Dyson gave significant contribution in the field of ruthenium-based anticancer drugs [23][24][25]. For some ruthenium compounds, it was shown that they express a very good cytotoxic activity; importantly, particular compounds also particular compounds also possess an antimetastatic activity [26,27]. In some cases, they can overcome the resistance of cancer cells while the ruthenium complexes hardly affect normal cells [17]. For certain cancer lines, it has already been demonstrated that their resistance to an organic drug can be overcome by complexing to ruthenium [28]. A problem of the first anticancer active ruthenium(III)-based compound, fac-[Ru(NH3)3Cl3] ( Figure  1A), is the low solubility [29]. Subsequently, water-soluble compounds such as the NAMI-A ( Figure 1B) were synthesized [26]. NAMI-A shows inhibition of the formation of metastases in the lung independently of the cytostatic activity without attacking the actual tumor. Certain properties, such as faster ligand (aqua) exchange of the ruthenium in the oxidation state +2 versus +3, suggest that it is more suitable for reactions in biological systems [30,31]. It was also shown that the oxidation state +2 is stabilized through π-bonded arene ligands [32]. Existing problems such as side effects, solubility, and resistances remain in part. Several arene ruthenium(II) complexes exhibited both in vitro and in vivo promising anticancer activity. Such complexes were active in vitro in the range of 6-300 μM against human cancer cell lines ( Figure 1C) [32,33]. Up to now, there are barely a few cytotoxic active ruthenium(II) complexes bearing phosphorus ligands (type D-F; Figure 1) [34][35][36]. Complex [Ru(η 6 -p-cymene)Cl2(pta)] (pta = 1,3,5-triaza-7-phosphaadamantane; Figure 1D) relived almost no anticancer activity, but on the other hand a promising antimetastatic activity [37]. Our group has explored neutral arene ruthenium(II) and iridium(III) complexes having κP-and κP,κS-coordinated ω-diphenylphosphino-functionalized alkyl phenyl sulfide, sulfoxide, and sulfone ligands (type F, Figure 1) on their anticancer activity [34,35,[38][39][40]. All complexes were found very active, importantly particular complexes showed in vitro cytotoxicities equal or higher than cisplatin.
Here, we describe the synthesis and characterization of various neutral (2a, 2c and 2d) and cationic arene ruthenium(II) complexes (4a-4e, 3b, 3e) with κPand κP,κS coordinated, respectively, 3-diphenylphosphino-functionalized propyl phenyl sulfide ligand Ph 2 P(CH 2 ) 3 SPh. Solvolysis as well as their cytotoxic activity, especially the influence of the arene ligands, were explored. Furthermore, on the most active compound the mechanism of action against 8505C tumour cell line was elucidated.

Molecular Structure and Chemical Properties of the Arene Ruthenium(II) Complexes
Ruthenium(II) complexes 2a-4e were characterized by microanalysis, IR and NMR ( 1 H, 13 C, 31 P) spectroscopies and purity was determined with elemental analysis. Singlecrystal X-ray structure analyses were performed for 3b, 4d, and 4e.

Molecular Structure and Chemical Properties of the Arene Ruthenium(II) Complexes
Ruthenium(II) complexes 2a-4e were characterized by microanalysis, IR and NMR ( 1 H, 13 C, 31 P) spectroscopies and purity was determined with elemental analysis. Singlecrystal X-ray structure analyses were performed for 3b, 4d, and 4e.

Infrared Spectroscopic Data
The IR spectra of the ruthenium(II) complexes showed characteristic bands around 290 cm -1 , which arise from Ru-Cl vibrations, while the bands found at 250 cm -1 are characteristic for bridging chlorido ligands in the dimers. These two bands are used for dis- tinguishment between bridging and terminal chlorido ligands in appropriate ruthenium(II) arene complexes and can be easily used to determine the structure [41][42][43][44]. The X-ray crystallography for compounds 3b, 4d and 4e confirmed assigned bands to be consistent with the assumed terminal Ru-Cl vibration. Most studies make no use of the fingerprint region and just the standard range for IR is mentioned and observed. Absorptions at around 680 cm -1 could be assigned to P-C vibrations of the ligand [45]. C=C-and C-H bands are found at 1400 and 1600 cm -1 as well as 3000 cm -1 , respectively and are in the expected ranges for ruthenium(II) complexes [34,35]. The dominant band at 742-748 cm -1 derived from thioether S-C parts of the prepared complexes [46].

NMR Data
The NMR spectra confirm the constitution of the complexes and all signals were found in the expected range with correct intensities in the 1 H NMR spectra. Thus, in the 1 H NMR spectra (Figure 3), the resonances of the coordinating aromatic moiety (arene ligand) in the neutral complexes are found within the expected chemical shift range (5-6 ppm) but slightly upfield in comparison to the appropriate ruthenium(II) dimers. In the case of complexes bearing more complex aromatic system than benzene, the corresponding proton resonances are found at expected values. The resonances of the hydrogen atoms from the propyl chain of the Ph 2 P(CH 2 ) 3 SPh ligand appeared in the range of 1 ppm to 3 ppm. The hydrogen atoms of the phenyl moieties from Ph 2 P(CH 2 ) 3 SPh 2 are resonating between 7 to 8 ppm.
guishment between bridging and terminal chlorido ligands in appropriate ruthenium(II) arene complexes and can be easily used to determine the structure [41][42][43][44]. The X-ray crystallography for compounds 3b, 4d and 4e confirmed assigned bands to be consistent with the assumed terminal Ru-Cl vibration. Most studies make no use of the fingerprint region and just the standard range for IR is mentioned and observed. Absorptions at around 680 cm -1 could be assigned to P-C vibrations of the ligand [45]. C=C-and C-H bands are found at 1400 and 1600 cm -1 as well as 3000 cm -1 , respectively and are in the expected ranges for ruthenium(II) complexes [34,35]. The dominant band at 742-748 cm -1 derived from thioether S-C parts of the prepared complexes [46].

NMR Data
The NMR spectra confirm the constitution of the complexes and all signals were found in the expected range with correct intensities in the 1 H NMR spectra. Thus, in the 1 H NMR spectra (Figure 3), the resonances of the coordinating aromatic moiety (arene ligand) in the neutral complexes are found within the expected chemical shift range (5-6 ppm) but slightly upfield in comparison to the appropriate ruthenium(II) dimers. In the case of complexes bearing more complex aromatic system than benzene, the corresponding proton resonances are found at expected values. The resonances of the hydrogen atoms from the propyl chain of the Ph2P(CH2)3SPh ligand appeared in the range of 1 ppm to 3 ppm. The hydrogen atoms of the phenyl moieties from Ph2P(CH2)3SPh2 are resonating between 7 to 8 ppm.
As seen in Figure 2, in the 1 H NMR spectra of the cationic complexes, additional resonances in comparison to neutral ones could be identified. In the range between 1 and 4.2 ppm, the protons of the propyl chain [Ph2P(CH2)3SPh] can be found. For complex 4a the protons of the coordinated aromatic show the same resonance as for 2a. A similar behavior was observed for the compound 4e. The other complexes (4b-4d) showed a splitting of these resonances of the aromatic systems. The phenyl residues of the Ph2P(CH2)3SPh are slightly shifted in comparison to neutral complexes (7.4 to 8.2 ppm). The 13 C NMR spectra show the same expected results as in 1 H NMR spectra. By means of C,H-COSY NMR spectroscopy, appropriate assignment of the resonances was possible (see Figures S1-S10).  As seen in Figure 2, in the 1 H NMR spectra of the cationic complexes, additional resonances in comparison to neutral ones could be identified. In the range between 1 and 4.2 ppm, the protons of the propyl chain [Ph 2 P(CH 2 ) 3 SPh] can be found. For complex 4a the protons of the coordinated aromatic show the same resonance as for 2a. A similar behavior was observed for the compound 4e. The other complexes (4b-4d) showed a splitting of these resonances of the aromatic systems. The phenyl residues of the Ph 2 P(CH 2 ) 3 SPh are slightly shifted in comparison to neutral complexes (7.4 to 8.2 ppm). The 13 C NMR spectra show the same expected results as in 1 H NMR spectra. By means of C,H-COSY NMR spectroscopy, appropriate assignment of the resonances was possible (see Figures S1-S10).
Singlets were found in the 31 P NMR spectra of Ph 2 P(CH 2 ) 3 SPh 2 ligand. Neutral ruthenium(II) complexes (2a, 2c and 2d) [Ru(η 6 -arene)Cl{Ph 2 P(CH 2 ) 3 SPh-κP}] showed chemically induced shift upfield in the 31 P NMR spectra (ca. 45 ppm). However, the formation of six-membered ruthenacycles (3b, 3e and 4a-4e) resulted in downfield shifts of up to 7.2 ppm in comparison to neutral complexes and upfield ca. 38 ppm in comparison to the phosphorous resonance in the free ligand. In the 31 P NMR spectra of cationic 4a-4e, besides the resonances resulting from the coordinated Ph 2 P(CH 2 ) 3 SPh 2 ligand, a septet at -144 ppm was observed for the PF 6 anion.

Stability of Complexes in DMSO
The synthesized ruthenium(II) complexes were investigated for in vitro antitumor activity (vide infra), and hence, stability of ruthenium(II) complexes in DMSO was investigated, since DMSO was used as solubilizing agent. Earlier, Gasser et al. studied the behavior of [Ru(η 6 -arene)Cl 2 (L)] complexes (L = N-heterocyclic ligands) in DMSO [47].
All synthesized ruthenium(II) complexes are stable and storable for several weeks in air. As expected, decomposition reactions occur faster in solution than in solid state. On the basis of a solution color change (orange/red → brown/black), degradation is already visible after storage at room temperature for more than four weeks. Subsequently, such behavior is documented with 1 H and 31 P NMR spectroscopies. For the neutral and cationic ruthenium(II) complexes, a similar behavior in DMSO was observed. The investigation 4a stability in DMSO, as an example, over 72 h is presented in Figure 4. The 1 H NMR spectra over time clearly indicate that 4a degrades to low extend for investigated period of time. Thus, for the in vitro studies 4a is acting on the cells.
In the 31 P NMR spectra of cationic 4a-4e, besides the resonances resulting from the coordinated Ph2P(CH2)3SPh2 ligand, a septet at -144 ppm was observed for the PF6 -anion.

Stability of Complexes in DMSO
The synthesized ruthenium(II) complexes were investigated for in vitro antitumor activity (vide infra), and hence, stability of ruthenium(II) complexes in DMSO was investigated, since DMSO was used as solubilizing agent. Earlier, Gasser et al. studied the behavior of [Ru(η 6 -arene)Cl2(L)] complexes (L = N-heterocyclic ligands) in DMSO [47].
All synthesized ruthenium(II) complexes are stable and storable for several weeks in air. As expected, decomposition reactions occur faster in solution than in solid state. On the basis of a solution color change (orange/red → brown/black), degradation is already visible after storage at room temperature for more than four weeks. Subsequently, such behavior is documented with 1 H and 31 P NMR spectroscopies. For the neutral and cationic ruthenium(II) complexes, a similar behavior in DMSO was observed. The investigation 4a stability in DMSO, as an example, over 72 h is presented in Figure 4. The 1 H NMR spectra over time clearly indicate that 4a degrades to low extend for investigated period of time. Thus, for the in vitro studies 4a is acting on the cells.  Within 72 h, there were appearances of new chemical shifts with low intensity detectable in the aromatic region, ascribed to free Ph 2 P(CH 2 ) 3 SPh, in both 1 H and 31 P NMR spectra. However, after 72 h for neutral and cationic ruthenium(II) complexes, much clearer appearance of degradation products could be identified. Apart from the chemical shifts belonging to the free Ph 2 P(CH 2 ) 3 SPh ligand observed in the 1 H NMR, consequently, the new resonances are also noted in the 31 P NMR spectra. The decomposition reactions also occur in different solvents. For instance, degradation in chloroform (it might be due to chlorination, often via a radical pathway) is much faster than in DMSO. Moreover, significant decompositions in chloroform could be observed, for example, after only 12 h for the complex 4a. To summarize, the decomposition of ruthenium(II) complexes in DMSO occurs only after days.
Within 72 h, there were appearances of new chemical shifts with low intensity detectable in the aromatic region, ascribed to free Ph2P(CH2)3SPh, in both 1 H and 31 P NMR spectra. However, after 72 h for neutral and cationic ruthenium(II) complexes, much clearer appearance of degradation products could be identified. Apart from the chemical shifts belonging to the free Ph2P(CH2)3SPh ligand observed in the 1 H NMR, consequently, the new resonances are also noted in the 31 P NMR spectra. The decomposition reactions also occur in different solvents. For instance, degradation in chloroform (it might be due to chlorination, often via a radical pathway) is much faster than in DMSO. Moreover, significant decompositions in chloroform could be observed, for example, after only 12 h for the complex 4a. To summarize, the decomposition of ruthenium(II) complexes in DMSO occurs only after days.
For further analysis, 8505C anaplastic thyroid carcinoma, resistant to chemotherapy, was selected. To define the cause of decreased number of viable cells in cultures exposed to IC 50 dose of 4b, presence of apoptotic as well as necrotic cells was estimated by Annexin V-FITC/PI staining. As could be seen in Figure 6a, cultivation in the presence of 4b elevated percentage of early apoptotic cells, marked as Ann + /PI − . In addition, occurrence of late apoptotic, double positive cells (Ann + /PI + ), was found in cultures exposed to 4b.
For further analysis, 8505C anaplastic thyroid carcinoma, resistant to chemotherapy, was selected. To define the cause of decreased number of viable cells in cultures exposed to IC50 dose of 4b, presence of apoptotic as well as necrotic cells was estimated by Annexin V-FITC/PI staining. As could be seen in Figure 6a, cultivation in the presence of 4b elevated percentage of early apoptotic cells, marked as Ann + /PI − . In addition, occurrence of late apoptotic, double positive cells (Ann + /PI + ), was found in cultures exposed to 4b.  These cells are rather secondary than primary necrotic, having in mind that apoptotic cells in culture must necrotize at the end point. Obtained results indicated that 4b induced apoptosis of 8505C. Subsequently, apostat staining showed that apoptosis triggered by the investigated drug in 8505C cells was accompanied by caspase activation (Figure 6b). While autophagy often follows the apoptosis as a cell attempt to repair damage, but also, under some circumstances, the same process could mediate cell removal herein. The amount of autophagosomes in cytoplasm of 8505C cells treated with 4b was quantified using supravital dye acridine orange (AO). Flow-cytometric analysis (Figure 6c) revealed slightly elevated fluorescence upon the treatment with 4b., thus pointing out the irrelevance of autophagic process to the drugs antitumor action. The same mode of action was found for complexes 2f and 4f [40]; however, the compound described herein is superior to those already published since an IC 50 dose of 4b causes similar effect as double IC 50 doses of complexes 2f and 4f.

Crystallography
Data for X-ray diffraction analyses of single crystals of 3b·H 2 O, 4d, and 4e were collected on an Rigaku Oxford Gemini S diffractometer at 110K using Mo-Kα radiation (λ = 0.71073 Å, graphite monochromator, CrysAlis Pro Version 1.171.36.28). Absorption corrections were applied multiscanning with the SCALE3 ABSPACK algorithm (T min /T max : 0.89/1.00, 3b·H 2 O; 0.84/1.00 4d; 0.98/1.00, 4e), respectively, of the CrysAlisPro software package. The structures were solved with direct methods using SHELXS-2013 and refined using full-matrix least-square routines against F 2 with SHELXL-2013 [49]. All nonhydrogen atoms were refined with anisotropic displacement parameters and hydrogen atoms with isotropic ones. Carbon-bonded hydrogen atoms were placed in calculated positions according to the riding model. The hydrogen atom positions of the water molecule of 3b·H2O were taken from difference Fourier maps and refined with DFIX and DANG constraints. CCDC 1907326-1907328 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif (accessed on 29 December 2020).
The cell lines 518A2, 8505C, A253, MCF-7 and SW480 were routinely maintained as monolayers in nutrient medium (RPMI-1640 supplemented with 10% FCS, 2 mM Lglutamine, 0.01% sodium pyruvate and 1% penicillin/streptomycin) at 37 • C in a humidified atmosphere with 5% CO 2 . Stock solutions of investigated compounds were prepared in DMSO at a concentration of 20 mM, filtered through Millipore filter, 0.22 µm, before use, and diluted by nutrient medium to various working concentrations. After standard trypsinization, cells were seeded at 2.5 × 10 3 cells/well in 96-well plates for viability determination and 1.5 × 10 5 cells/well in 6-well plate for flow cytometry.

Determination of Cell Viability by Sulphorhodamine Assay (SRB)
The viability of adherent viable cells was measured by SRB assay [48]. Cells were exposed to a wide range of doses of the drugs for 96 h and then fixed with 10% of TCA for 2 h at 4 • C. After fixation, cells were washed in distilled water, stained with 0.4% SRB solution 30 min at RT, washed, and dried overnight. The dye was dissolved in 10 mM TRIS buffer, and the absorbance was measured at 540 nm with the reference wavelength at 640 nm. IC50 values, defined as the concentrations of the compound at which 50% of cell inhibition occurs ± SD were calculated using four-parameter logistic function and presented as mean from three independent experiments.

AnnexinV-FITC/PI, AO Staining and Caspase Detection
Cells were exposed to IC 50 dose of 4b for 72 h. After trypsinization cells were stained with AnnV-FITC/PI (Biotium, Hayward, CA, USA) or apostat according to the manufacturer's instruction. Alternatively, cells were stained with solution of 100 µM AO 15 min at 37 • C. Cells were analyzed with CyFlow ® Space Partec with Partec FloMax ® software.

Conclusions
In this work, the synthesis of various neutral and cationic ruthenium complexes of the general formulae [Ru(η 6 -arene)Cl 2 {Ph 2 P(CH 2 ) 3 SPh-κP}] and [Ru(η 6 -arene)Cl{Ph 2 P(CH 2 ) 3 SPh-κP,κS}]X (arene = benzene, mesitylene, indane, thn, and 1,4-dialin; X = Cl − or PF 6 − ), respectively, was established. Complexes were characterized by IR and multinuclear NMR spectroscopy. Moreover, crystal structures of 3b, 4d, and 4e complexes were obtained and confirmed proposed structures. The stability of the complexes in DMSO, thus possibility of DMSO to replace ligands, was investigated using NMR spectroscopy. Solvolysis is considerably hindered within the first 3 days, therefore the applied ruthenium(II) complexes in in vitro investigations did not suffer with the DMSO substitution in the stock solution.
The cytotoxicity of all arene ruthenium(II) complexes was determined in five human cancer cell lines (518A2, 8505C, A253, MCF-7 and SW480). All ruthenium(II) complexes demonstrated high cytotoxic potential with the IC 50 values down to the low micromolar range. Selected cationic ruthenium(II) complex bearing the mesytil moiety (4b) was found to induce apoptosis in 8505C cisplatin resistant cell line. This process was associated with caspase activation. Taken together, herein are synthesized ruthenium(II) complexes with strong anticancer potential, whose mechanism of action is based on the caspase triggered apoptosis, thus encouraging future development of this promising ruthenium(II) complexes.