Gold(I)-Triphenylphosphine Complexes with Hypoxanthine-Derived Ligands: In Vitro Evaluations of Anticancer and Anti-Inflammatory Activities

A series of gold(I) complexes involving triphenylphosphine (PPh3) and one N-donor ligand derived from deprotonated mono- or disubstituted hypoxanthine (HLn) of the general composition [Au(Ln)(PPh3)] (1–9) is reported. The complexes were thoroughly characterized, including multinuclear high resolution NMR spectroscopy as well as single crystal X-ray analysis (for complexes 1 and 3). The complexes were screened for their in vitro cytotoxicity against human cancer cell lines MCF7 (breast carcinoma), HOS (osteosarcoma) and THP-1 (monocytic leukaemia), which identified the complexes 4–6 as the most promising representatives, who antiproliferative activity was further tested against A549 (lung adenocarcinoma), G-361 (melanoma), HeLa (cervical cancer), A2780 (ovarian carcinoma), A2780R (ovarian carcinoma resistant to cisplatin), 22Rv1 (prostate cancer) cell lines. Complexes 4–6 showed a significantly higher in vitro anticancer effect against the employed cancer cells, except for G-361, as compared with the commercially used anticancer drug cisplatin, with IC50 ≈ 1–30 µM. Anti-inflammatory activity was evaluated in vitro by the assessment of the ability of the complexes to modulate secretion of the pro-inflammatory cytokines, i.e. tumour necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), in the lipopolysaccharide-activated macrophage-like THP-1 cell model. The results of this study identified the complexes as auspicious anti-inflammatory agents with similar or better activity as compared with the clinically applied gold-based antiarthritic drug Auranofin. In an effort to explore the possible mechanisms responsible for the biological effect, the products of interactions of selected complexes with sulfur-containing biomolecules (L-cysteine and reduced glutathione) were studied by means of the mass-spectrometry study.


Introduction
Gold-based medication was used for a wide range of ailments already in the distant history of ancient China 2500 BC [1,2]. Thus, it is quite extraordinary that even current clinical practice still recognizes chrysotherapy (chrysos, gold in Greek), i.e. the treatment of diseases by the administration of gold-based therapeutic agents, as an integral part of modern medicine [3,4]. During the last eighty years of clinical use of gold-based metallotherapeutics, two major groups of complexes have been introduced. The first one is represented by the parenteral preparations, containing relatively simple gold(I)-complexes, such as sodium aurothiomalate (Myochrysin, sodium ((2-carboxy-1carboxylatoethyl)thiolato)gold(I)) and aurothioglucose (Solganol, {(2 S,3R,4 S,5 S,6R)-3,4,5-trihydroxy-6-(hydroxymethyl)-oxane-2-thiolato}gold(I)) [5], and the second one comprises the orally administered drug Auranofin (Ridaura) [6][7][8] (triethylphosphine-(2,3,4,6-tetra-O-acetyl-1-D-thiopyranosato-S)gold(I)) (See Fig-ure 1). Nowadays, the gold-based metallotherapeutics do not represent antiarthritic agents of the first choice, since new highly potent and more specific anti-rheumatic compounds have been developed (i.e. biopharmaceuticals, novel non-steroidal antiinflammatory drugs, NSAIDs, or corticosteroids), however, they are still indicated in specific conditions, including moderately to severely active rheumatoid arthritis [8,9]. Although the clinical application of Auranofin has declined in recent years, there are several attempts for repurposing it in other significant indications, such as cytoprotective agents and in the treatment of HIV, severe microbial infections and cancer [1,2,4,8,10].
Several Auranofin inspired gold(I)-complexes have been studied for significant anti-inflammatory, and/or antitumor activities. Among linear gold(I) phosphine complexes, those involving ligands as e.g. dithiocarbamates [1], sulfanylpropenoates [11], naphthalimide derivatives [12], imidazole, pyrazole [13,14] as well as purine derivatives [15,16] have been investigated. It has been established that the gold(I)-phosphine moiety is responsible for the actual interactions [3,17] with the target sites of the biological molecules, i.e. mostly the selanyl-groups of enzymes [1,18], while the other, weaker bonded ligand influences the kinetic profile of the compounds. This principle has been employed by our group in several recently published works, describing highly antitumor active as well as anti-inflammatory potent gold(I) complexes of the type [Au(L)(PPh 3 )] involving various N6-benzyladenine derivatives (L) [15,16].
In this work, we report a series of gold(I) complexes of the general formula [Au(L n )(PPh 3 )], where HL n represents variously mono-and disubstituted derivatives of hypoxanthine (See Figure 2). The rationale for the selection of these N-donor ligands is based on the fact that variously substituted derivatives of hypoxanthine have been identified as promising inhibitors of diverse essential enzymes, as cyclin-dependent kinases [19,20] and O6-alkylguanine-DNA-alkyltransferase [21,22], and therefore in combination with the {Au-PPh 3 } moiety could bring additional means of influencing the cellular metabolism both on the level of cell division, and modulation of cellular responses to the inflammatory stimuli.
The complexes were fully characterized, including high resolution NMR spectroscopy and single crystal X-ray analysis (for complexes 1 and 3), and were evaluated for their cytotoxicity against a panel of nine human cancer cell lines (MCF7, HOS, THP-1, A549, G361, HeLa, A2780, A2780R (cisplatin resistant), and 22Rv1, and also for their in vitro anti-inflammatory effect on the model of LPS-stimulated human monocytic leukaemia (THP-1) macrophage-like cell line and the mechanisms of interaction of the complexes with sulfur-containing biomolecules (i.e. L-cysteine and reduced glutathione) were studied by means of mass spectrometry. Moreover, this work fits within the focus of our long term study dedicated to research and development of new transition metal complexes showing various types of biological effects, represented dominantly by the anticancer {Pt(II) [23], Pd(II) [24], Cu(II) [25,26] or Ru(III) [27]}, anti-inflammatory {Au(I/III)} [15,16,28], antidiabetic and cytoprotective {Cu(II) [29]} activities.
A RealTime Ready Cell Lysis Kit (Roche, Mannheim, Germany) served for isolation of RNA from cells and Transcriptor Universal cDNA Master (Roche, Mannheim, Germany) was used for reverse transcription of RNA to cDNA. Specific primers and probes (Gene Expression assays) for polymerase chain reaction  (PCR) were obtained from Applied Biosystems (Foster City, CA, USA). The following assays were chosen for the quantification of gene expression: Hs00174128_m1 for tumour necrosis factor-a (TNF-a), Hs01555410_m1 for interleukin-1b (IL-1b), and 4326315E for b-actin, which served as an internal control of gene expression. Quantitative PCR (qPCR) was performed with Fast Start Universal Probe Master (Roche, Mannheim, Germany). Instant ELISA Kits (eBioscience, Vienna, Austria) were used to evaluate the production of TNF-a and IL-1b by the enzyme linked immunosorbent assay (ELISA). The Immun-Blot PVDF (polyvinylidene fluoride) membrane 0.2 mm (Bio-Rad, Hercules, CA, USA) and albumin bovine fraction V (BSA) (Serva, Heidelberg, Germany) were used for Westernblot. Murine monoclonal anti-IkB-a (Cell Signaling, Danvers, MA, USA), murine monoclonal anti-b-actin (Abcam, Cambridge, UK) and goat polyclonal antimouse IgG (with the conjugated peroxidase) antibodies (Sigma-Aldrich, Saint Louis, MO, USA) were applied for immunodetection. The activity of the conjugated peroxidase was detected by an Opti-4CN Substrate Kit (Bio-Rad, Hercules, CA, USA).

Chemistry
Gold(I) complexes of the general formula [Au(L n )(PPh 3 )] (1-9), where L n stands for a deprotonated form of a hypoxanthinederived compound, were prepared by a slight modification of the previously reported procedure for the synthesis of similar   [Au(L)(PPh 3 )] complexes, in which HL stands for 6-benzylaminopurine derivatives [15,16]. Concretely, the reactions of the acetone solutions of [AuCl(PPh 3 )] (1 mmol in 10 mL) were mixed with the corresponding hypoxanthine derivative (HL n ; 1 mmol in 20 mL), and subsequently, 1 mL of 1 M NaOH was added into the reaction mixture. The mixture was stirred for 3 h, after which precipitated NaCl was filtered off. The colourless filtrate was evaporated to dryness and the residue was collected. Recrystallization was carried out from acetone. In the case when the residue was of gel-like consistency, 20 mL of diethyl ether was added and the suspension was sonicated to obtain a powder product. Crystals of complexes 1 and 3, suitable for single crystal X-ray analysis, were prepared by a slow diffusion of n-hexane into a saturated solution of the appropriate complex in toluene. The purity and composition of the obtained gold(I) complexes 1-9 were established based on the methods of elemental analysis, electrosprayionization mass spectrometry (ESI-MS), 1 H and 13 C NMR, and FT-IR spectroscopies, and thermal analysis (TG/DTA) (characterization data are given in File S1). The molecular and crystal structures were determined by single crystal X-ray analysis for complexes 1 and 3.

Physical Measurements
Elemental analyses (C, H, N) were carried out on a Flash 2000 CHNO-S Analyser (Thermo Scientific, USA). FT-IR spectra were measured on a Nexus 670 FT-IR spectrometer (ThermoNicolet, USA) by the ATR technique in the 200-4000 cm 21 range. 1 H, 13 C and 31 P NMR spectra were measured in DMF-d 7 on a Varian 400 MHz NMR spectrometer at 300 K using tetramethylsilane (SiMe 4 ) (for 1 H and 13 C spectra) and 85% H 3 PO 4 (for 31 P) as an internal reference standard. Mass spectra of methanol solutions of complexes were recorded on an LCQ Fleet Ion-Trap mass spectrometer using the positive mode electrospray ionization (Thermo Scientific, USA). Thermogravimetric (TG) and differential thermal (DTA) analyses were performed on a thermal analyser Exstar TG/DTA 6200 (Seiko Instruments Inc., Japan) in dynamic air conditions (50 mL min 21 ) between room temperature (ca 25uC) and 950uC (gradient 2.5uC min 21 ). Single crystal X-ray analyses were performed on an Xcalibur2 diffractometer equipped with a CCD detector Sapphire2 (Oxford Diffraction Ltd., Oxford, UK) using MoKa radiation (monochromator Enhance, Oxford Diffraction Ltd.), and using the v-scan technique at 120 K. Data collection, data reduction and cell parameter refinements were performed by the CRYSALIS software package [32]. The molecular structures were solved by direct methods and all non-hydrogen atoms were refined anisotropically on F 2 using the full-matrix least-squares procedure (SHELX-97) [33]. H-atoms were found in difference maps and refined by using the riding model with C2 H = 0.95, 0.98 and 0.99 Å , with U iso (H) = 1.2U eq (CH, CH 2 ) and 1.5U eq (CH 3 ). The highest residual peaks of 2.623 e Å -3 and 1.792 e Å -3 were located 0.85 Å from Au1 (for complex 1), and 0.88 Å from Au1 (for complex 3), respectively. Molecular graphics were drawn and additional structural parameters were interpreted in DIAMOND [34] and Mercury [35].

Maintenance and Preparation of Macrophages
The human monocytic leukaemia cell line THP-1 (ECACC, Salisbury, UK) was used for the in vitro anti-inflammatory activity evaluation. The cells were cultivated at 37uC in the RPMI 1640 medium supplemented with 2 mM L-glutamine, 10% FBS, 100 U/mL of penicillin and 100 mg/mL of streptomycin in humidified atmosphere containing 5% CO 2 . Stabilized cells (3 rd -15 th passage) were split into microtitration plates to get the concentration of 500 000 cells/mL and the differentiation to macrophages was induced by addition of phorbol myristate acetate (PMA) dissolved in DMSO at the final concentration of 50 ng/ml and the cells were incubated for 24 h. Unlike monocytes, differentiated macrophages tend to adhere to the bottom of the cultivation plates. For the subsequent 24 h, the cells were incubated with the fresh complete RPMI medium, i.e. containing antibiotics and FBS, without PMA. Then, the medium was aspirated, and the cells were washed with PBS and cultivated in the serum-free RPMI 1640 medium for next 24 h. These prepared macrophages were used for the detection of inflammatory response.

In Vitro Cytotoxicity Testing
Cytotoxicity in THP-1 cells was determined by the WST-1 assay. The THP-1 cells (floating monocytes, 500 000 cells/mL) were incubated in 100 mL of the serum-free RPMI 1640 medium and seeded into 96-well plates in triplicate at 37uC. Measurements were taken 24 h after the treatment with the tested compounds dissolved in 0.1% DMSO in the concentration range of 0.16- 10 mM. Viability was determined by the WST-1 test according to the manufacturer's manual. The amount of created formazan (correlating with the number of metabolically active cells in the culture) was calculated as a percentage of the control cells, which were treated only with 0.1% DMSO and was set-up as 100%. The cytotoxic IC 50 values of the tested compounds were calculated from the obtained data. The WST-1 assay was performed spectrophotometrically at 440 nm (FLUOstar Omega, BMG Labtech, Germany).

Drug Treatment and Induction of Inflammatory Response
Differentiated macrophages were pretreated with 300 nM solutions of the complexes 1-9, HL 1-9 , AuCl, [AuCl(PPh 3 )], PPh 3 and Auranofin dissolved in DMSO (the final DMSO concentration was 0.1%) and with 0.1% DMSO solution (vehicle) for 1 h; the given concentrations of the tested compounds lack the cytotoxic effect (cell viability .94%). The inflammatory response was triggered by addition of 1.0 mg/mL LPS dissolved in water to the pretreated macrophages, control cells were without the LPS treatment.

RNA Isolation and Gene Expression Evaluation
For the evaluation of the expression of TNF-a, IL-1b, and bactin mRNA, the total RNA was isolated directly from the THP-1 cells in cultivation plates using a RealTime Ready Cell Lysis Kit, according to the manufacturer's instructions.
The gene expression was quantified by two-step reversetranscription quantitative (real-time) PCR (RT-qPCR). The reverse transcription step was performed by Transcriptor Universal cDNA Master using cell lysate as a template. The reaction consists of 3 steps: (1) primer annealing, 29uC for 10 min; (2) reverse transcription, 55uC for 10 min; and (3) transcriptase inactivation, 85uC for 5 min. FastStart Universal Probe Master and Gene Expression assays were used for qPCR. These assays contain specific primers and TaqMan probes that bind to an exonexon junction to avoid DNA contamination. The parameters for the qPCR work were adjusted according to the manufacturer's recommendations as follows: 50uC for 2 min, then 95uC for 10 min, followed by 40 cycles at 95uC for 15 s and 60uC for 1 min. The results were normalized to the amount of ROX reference dye, and the change in gene expression was determined by the DDC T method. Transcription of the control cells was set as 1 and other experimental groups were multiples of this value.

Evaluation of Cytokine Secretion by ELISA
Macrophages, which were pretreated with the tested compounds for 1 h, were incubated with LPS for next 24 h. After this period, the medium was collected and the concentration of TNF-a and IL-1b was measured by an Instant ELISA kit according to the manufactures' manual.

Determination of IkB Degradation by Western Blot
Macrophage-like THP-1 cells were pretreated with the tested compounds and stimulated by LPS as described above. Thirty minutes after the addition of LPS, the medium was aspirated and cells were washed by cold PBS. Subsequently, the cells were collected using the lysis buffer [50 mM Tris-HCl pH 7.5, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 270 mM sucrose, 0.1% (v/v) Triton X-100, and cOmplete Protease Inhibitor Cocktail (Roche, Germany)] and scraper. The lysis of cells was facilitated by short (ca. 30 s) incubation in the ultrasonic water bath. The protein concentration was determined according to Bradford's method. For protein separation, 30 mg of protein was loaded onto 12% polyacrylamide gel. Then, they were electrophoretically transferred on the PVDF membranes, which were subsequently blocked by 5% BSA dissolved in TBST buffer [150 mM NaCl, 10 mM Tris base pH 7.5, 0.1% (v/v) Tween-20]. The membranes were incubated with the primary anti-IkB-a antibody at the concentration of 1:500, or with the primary anti-bactin at the concentration of 1:5000 at 4uC for 16 h. After washing, the secondary anti-mouse IgG antibody diluted 1:2000 was applied on the membranes and incubated for 1 h at laboratory temperature (,22uC). The amount of the bound secondary antibody was detected colorimetrically by an Opti-4CN kit according to the manufacturer's manual.

Statistical Evaluation
The cytotoxicity data were expressed as the percentage of viability, when 100% represent the treatment with vehicle (DMF or DMSO). The experiments were conducted in triplicate using cells from different passages. The IC 50 values were calculated from viability curves. The results are presented as arithmetic mean 6 standard error of the mean (SE). The significance of the differences between the results was evaluated by the ANOVA analysis with p,0.05 considered to be significant (QC Expert 3.2, Statistical software, TriloByte Ltd., Pardubice, CZ).
The statistically significant differences between individual groups during anti-inflammatory activity testing were assessed by the one-way ANOVA test, followed by Tukey's post-hoc test for multiple comparisons. GraphPad Prism 5.02 (GraphPad Software Inc., San Diego, CA, USA) was used for the analysis.

Interactions with L-Cysteine and Reduced Glutathione Analyzed by Mass Spectrometry
The interaction experiments between the selected representative complexes 1 and 6, bearing the same ethoxy-substitution on the C6 atom, but differing in the substituents in the C2 position, and the mixture of physiological levels of cysteine and glutathione were performed on a Thermo Scientific LTQ Fleet Ion-Trap mass spectrometer, using the positive ionization mode. The reaction system contained the physiological concentrations of L-cysteine (290 mM) and reduced glutathione (6 mM) [36] and the tested complex (20 mM) in the methanol:water (1:1, v/v) mixture. The reference system was comprised of the solution of complex (20 mM) in the methanol:water (1:1, v/v) mixture. The flow injection analysis (FIA) method was used to introduce the reaction system (5 mL spikes) into the mass spectrometer, while pure acetonitrile was used as a mobile phase. The ESI-source was set up as follows: source voltage was 4.5 kV, the vaporizer temperature was 160uC, the capillary temperature was 275uC, the sheath gas flow rate was 30 L/min, and auxiliary gas flow rate was 10 L/min. The system was calibrated according to the manufacturer specifications and no further tuning was needed.

General Properties of the Au(I) complexes
This work reports on the preparation, thorough characterization, and in vitro cytotoxic and anti-inflammatory activities of a series of gold(I)-triphenylphosphine complexes [Au(L 1-9 )(PPh 3 )] involving mono-or disubstituted derivatives of hypoxanthine  (HL [1][2][3][4][5][6][7][8][9] ). The complexes were synthesized by a modification of a previously published method [15,16], i.e. by a reaction of the precursor complex [AuCl(PPh 3 )] with an equimolar amount of the corresponding organic molecule HL [1][2][3][4][5][6][7][8][9] in acetone with the addition of 1 M NaOH. The prepared complexes 1-9 are very well soluble in acetone, alcoholic solvents, toluene, DMSO and DMF, and partially soluble in water at laboratory temperature. The complexes were prepared non-solvated, as evidenced by the results of simultaneous TG/DTA analysis (see Figure S3 in File S1). The complexes were further characterized as chemical individuals by elemental analysis, 1 H, 13 C and 31 P NMR spectroscopy, FT-IR spectroscopy and ESI-MS. Single crystals suitable for X-ray analysis were prepared in the cases of 1 and 3, thus providing information about the molecular and crystal structures of the complexes (discussed in detail below). 1 H and 13 C NMR spectra were measured for all the presented complexes. The obtained results were very beneficial for the confirmation of the purity and composition of 1-9. The observed signals unambiguously proved the presence of both hypoxanthine and phosphine derivatives in the structures of the Au(I) complexes as well as gave information about the coordination mode of these ligands. The most conclusive evidence concerning the identification of the donor atom of HL 1-9 was obtained from the 13 C NMR spectra by comparing the chemical shifts in the uncoordinated and coordinated hypoxanthine derivative. The resulting differences, calculated as coordination shifts; Dd = d complex -d ligand , ppm; Table 1), showed that the greatest changes occurred for the signals corresponding to the carbons C4 and C8, shifted by 2.02-4.82 ppm, and 3.57-6.89 ppm, respectively. This observation indirectly pointed to N9 as the coordination site, as the two mentioned carbons lie in the direct vicinity to this nitrogen. The set of the most intensive signals at around 130 ppm, characteristic of all the spectra of 1-9, was assigned to the carbon atoms of triphenylphosphine. Accordingly, the 1 H NMR spectra of all the nine Au(I) complexes showed very strong multiplet signals at around 7.70 ppm with the relative integral intensity well corresponding to the calculated value of 15 hydrogens of PPh 3 . Additionally, the C8 H signal of the corresponding HL 1-9 ligand was shifted by 0.08-0.27 ppm upfield in the spectra of the complexes with respect to the spectra of the uncoordinated HL 1-9 , corresponding with the results following from 13 C NMR spectroscopy about N9 being the site of coordination to gold. What should be also pointed out is the absence of the signal assignable to the proton N9 H in the spectra of 1-9, which well agrees with the presence of deprotonated hypoxanthine-based ligands in the studied complexes symbolized as L 1-9 . All the 31 P NMR spectra of the complexes were characteristic of the presence of one singlet at 31.67-33.13 ppm (see Figure S5 in File S1), which is significantly shifted as compared to the signal of free PPh 3 (-5.96 ppm), which confirms the coordination of PPh 3 to gold(I) atom through phosphorus, thus forming the Au-P bond. In summary, the presence and chemical shift of all the detected signals in 1 H, 13 C and 31 P NMR spectra, as well as the relative intensity of the peaks in the proton spectra, indicate that the gold(I) atom in 1-9 is coordinated by one deprotonated hypoxanthinederived compound (L 1-9 ) binding via N9 and one PPh 3 molecule. These conclusions are in good agreement with the results following from single crystal X-ray analysis of complexes 1 and 3 discussed below.
Mass spectra measured in the positive ionization mode also indirectly confirmed the composition of the presented complexes, as all the spectra of 1-9 contained the [M+H] + molecular peaks. The presence of the hypoxanthine-based derivatives was demonstrated by the observed peaks corresponding to the adducts of [HL n +Na/K] + (for details of ESI+MS characterization, see File S1).
Further properties of the presented Au(I) complexes were studied by FT-IR spectroscopy. The courses of the spectra of 1-9 in the mid-IR region were qualitatively very similar to those of the uncoordinated HL 1-9 . The very intensive peaks at 1600-1578 cm 21 could be assigned to the n(C -… N) ring stretching vibrations characteristic of heterocyclic molecules with nitrogen ring atoms. Further assigned characteristic vibrations could be found at 3057-3048 cm 21 for v(C-H) ar , 2986-2889 cm 21 for v(C-H) aliph , and also 1480-1436 cm 21 for n(C -… C) ring . The typical vibrations of the ether functional group found in the O6substituted hypoxanthine derivatives were detected at 1339-1272 cm 21 for the aromatic C6-O stretch and at 1118-1087 cm 21 for the O-C10 stretch. The medium to strong intensity bands found between 800-700 cm 21 , and at ca. 690 cm 21 can be attributed to the out-of-plane C-H bending vibrations, and ring bending, respectively, determining the presence of an aromatic ring in all the complexes 1-9 [37,38]. In the far spectra of the Au(I) complexes, new bands were identified as compared with the spectra of free HL [1][2][3][4][5][6][7][8][9] . These maxima at 506-492 cm 21 , and 329-321 cm 21 could be assigned to v(Au-N), and v(Au-P) stretching vibrations, respectively, although it should be noted that triphenylphosphine alone has intensive vibrations in this region of IR spectra, so these bands could be overlapped [15,16,37,39].

Molecular and Crystal Structures of [Au(L 1 )(PPh 3 )](1) and [Au(L 3 )(PPh 3 )] (3)
The crystallization method of slow n-hexane diffusion into the saturated toluene solution of the complexes allowed the preparation of single crystals suitable for X-ray analysis of two complexes, [Au(L 1 )(PPh 3   tures are depicted in Figure 3, and 4, respectively. Crystal data and structure refinement parameters (Table S1 in File S1), selected bond lengths and angles (Table S2 in File S1), as well as noncovalent interaction parameters (Tables S3 and S4 in File S1) are given in File S1. In both complexes 1 and 3, the central gold(I) atom is twocoordinated in the distorted linear arrangement of the {NP} donor set formed by one N-donor deprotonated hypoxanthine derivative and the electroneutral molecule of triphenylphosphine bonded via phosphorus. Comparing the coordination bonds in 1 and 3, both bond lengths are mutually comparable between the complexes, as they were found to be 2.046(3), and 2.042(3) Å for Au-N; and 2.2344(9), and 2.2356(8) Å for Au-P, respectively. On the other hand, it is evident that the Au-P bonds are significantly longer than the Au-N bonds, which is most likely connected with the bulky phenyl moieties surrounding the phosphorus donor atom in PPh 3 . This bond distance difference is not uncommon in similar complexes. The search in the Cambridge Structural Database (CSD ver. 5.35, February 2014 update) [40] within the deposited 113 mononuclear compounds involving the N-Au-PPh 3 motif with  Tables S3 and S4 in File S1, Figures S1 and S2 in File S1.

In Vitro Cytotoxicity
For the evaluation of in vitro cytotoxicity, the presented complexes were tested by the MTT assay and the results were compared with the clinically applied metallodrug cisplatin, used as the reference standard, as well as with the starting compounds, i.e. hypoxanthine derivatives HL 1-9 , and gold-containing inorganic compounds AuCl or HAuCl 4 (applied in the concentration range of 0.01-50.0 mM, unless their solubility was found to be lower). The first phase of testing involved screening of all the prepared complexes 1-9 against two human cancer cell lines, breast adenocarcinoma (MCF7) and osteosarcoma (HOS). The starting compounds were found inactive up to 50 mM or to the limiting concentration of their solubility against these cells. Then, this testing showed a varied effect of the tested complexes on the viability of the selected cancer cell lines. Complex 7 showed to be inactive up to the tested concentration range on both cell lines (IC 50 .50 mM). Complexes 8 and 9 were found to be cytotoxic inactive against the MCF7 cells in the used concentration ranges given by their solubility, therefore their cytotoxicity can be evaluated as .25 mM. Complex 2 was moderately active on both cell lines, however, the IC 50 values, i.e. 27.160.2 mM (MCF7) and 30.062.1 mM (HOS), were higher than those for cisplatin (in the case of MCF7, significantly higher, ANOVA, p,0.05). This pointed to the fact that the n-butyl group as the R1 substituent showed to be the least beneficial for the resulting cytotoxic activity against these cell lines within the tested series of compounds (n-Bu present in 2 and 7). The best results were observed for the complexes 4-6, all of which possessed significantly better cytotoxicity than cisplatin (ANOVA, p,0.05) against both cell lines, concretely the IC 50 values were found to be 2-5-times lower. The most promising data were found for complex 4 against the MCF7 cell line, i.e. IC 50 (Table 2).
Before using the THP-1 cell line for testing of the anti-phlogistic effect, cytotoxicity of complexes 1-9 and uncoordinated hypoxanthine derivatives HL 1-9 was also tested on this cell line ( Table 2). The IC 50 values for 1-7 were between 1.03-2.15 mM, which are significantly lower than for the MCF7 and HOS cell lines. Interestingly, 8 has the IC 50 only 5.360.2 mM and 9 does not even have any cytotoxic effect up to the concentration of 10 mM, whereas their structural analogues without chlorine at C2 (3 and 4) have the IC 50 values equal to 1.8760.09 mM, and 2.1560.19 mM, respectively. In all the cases, cytotoxicity was lower (i.e. the IC 50 values were higher; 1.03-5.28 mM) than for commercially used drug Auranofin (IC 50   ligand, showing remarkable cytotoxicity towards leiomyosarcoma cells with the IC 50 values in the comparable range as 1-7 (0.7-2.1 mM) have been reported [41]. Complexes 1-8 additionally demonstrate the hormesis effect (higher viability; in this case up to 168%) around the concentration of 0.3 mM.
Based on the promising results for complexes 4-6 against all the three above mentioned cell lines, further testing was performed for these representatives on a panel of human cancer cell lines involving lung carcinoma (A549), malignant melanoma (G361), cervix epitheloid carcinoma (HeLa), ovarian carcinoma (A2780), ovarian carcinoma resistant to cisplatin (A2780R) and prostate carcinoma (22Rv1). The testing again revealed that the starting compounds (HL n and AuCl, HAuCl 4 ) were not toxic in the tested concentration range, only a weak antiproliferative effect was observed for HAuCl 4 on the G361 cell line (38. 160.8 mM). The IC 50 values determined for 4-6 showed that viability of all the employed cell lines was reduced by the tested complexes comparably or better than by the metallodrug cisplatin (Table 3, Figure 5). Complexes 5 and 6 were significantly more in vitro antitumour active (ANOVA, p,0.05) against all the cell lines as compared with cisplatin, which is valid also for complex 4 with the exception of the G361 cells, against which its activity was found comparable with cisplatin (3.960.5 mM, and 5.360.2 mM, respectively). Focusing on individual cell lines, the lowest antiproliferative effect of 4-6 was observed for A549 and HeLa, which in the testing did not respond to the treatment by cisplatin up to 50 mM. The IC 50 values for the tested Au(I) complexes showed moderate activity in the micromolar range (<15-20 mM). On the other hand, low micromolar IC 50 values (<3-5 mM) resulted from the cytotoxicity testing against all the other cancer cell lines, i.e. G361, 22Rv1, A2780 as well as A2780R. It should be also pointed out that complexes 4-6 showed to be ca. 3-times (A2780), 5-times (A2780R) and 7-times (22Rv1) more effective than cisplatin. Moreover, the evaluation of cytotoxicity of 4-6 on both A2780 cell lines sensitive and resistant to cisplatin allowed the calculation of the resistance factors (RF), i.e. the ratio of IC 50 (A2780R)/IC 50 (A2780). The RF values are equal to 1.19 (4), 1.14 (5), 1.21 (6) and 2.25 (cisplatin), which unambiguously shows that 4-6 are able to circumvent the acquired resistance of cancer cells to cisplatin. It can be additionally pointed out that the tested complexes are comparably active on both A2780 and A2780R cell lines, as there is no statistically significant difference in their cytotoxic activity against these cancer cells (ANOVA, p, 0.001).

In Vitro Anti-Inflammatory Activity
To evaluate the anti-inflammatory effect of the presented complexes in vitro, their ability to diminish the production of proinflammatory cytokines TNF-a and IL-1b in LPS-stimulated macrophages derived from the THP-1 cell line was determined. The results of this study showed that complexes 1-7 significantly decreased secretion of TNF-a in the LSP-stimulated cells, while complexes 8 and 9 exhibited no effect, as the extent of their influence on TNF-a secretion was comparable to the vehicle alone (0.1% DMSO) ( Figure 6A). It is interesting to note, that the comparison of complexes 1-4 and their analogues 6-9 differing only in the substitution at C2 (hydrogen, and chlorine, respectively) showed that in all the cases, the complexes 1-4 were more effective in the TNF-a level attenuation. Moreover, all the active complexes (1-7) had comparable activity as the commercial reference drug Auranofin, but importantly they possessed a lower cytotoxic effect in the employed THP-1 cell line. To elucidate the role of each constituent part of the complexes (N-donor ligand, PPh 3 or gold(I) species) in diminishing TNF-a secretion, free hypoxanthine derivatives (HL 1-9 ), [AuCl(PPh 3 )], PPh 3 , and AuCl were also tested. It was found that none of these starting materials showed the anticipated effect. Moreover, the compound HL 7 significantly increased the production of TNF-a. These results indicated that in order to show the anti-inflammatory effect, the corresponding ligand has to be bonded into the Au(I)-complex. This observation is in agreement with the previous studies, where no anti-phlogistic effect was observed for free and uncoordinated purine-derived compounds applied as ligands in Au(III)-complexes [28].
The extent of the influence of complexes 1-9 on IL-1b secretion was very similar as for TNF-a ( Figure 6B), but it should be noted that IL-1b production was affected slightly more than TNF-a. The same observation was also made in previous studies [15,43], which showed that the tested gold(I) complexes preferentially inhibit IL-1b production as compared to TNF-a. The results proved that similarly to TNF-a, compounds 1-7 significantly decreased the level of IL-1b, whereas complex 9 showed again no effect. The only difference as compared to the above-described TNF-a production influence study was found for complex 8, which was inactive in the case of TNF-a secretion, but it significantly reduced the production of IL-1b. However, its inhibitory activity was still lower as compared with the activity of 1-7. Testing of the constituent parts of the complexes as well as of starting materials revealed the element necessary for diminishing the production of IL-1b. The free HL n molecules were again found to be inactive, with the exception of HL 6 which reduced the level of the cytokine significantly as compared with the vehicle-treated cells (p,0.05). On the other hand, unlike in the case of TNF-a, AuCl significantly decreased the level of this pro-inflammatory cytokine (p,0.01) as well as the [AuCl(PPh 3 )] complex, which even curtailed its level as effectively as complexes 1-9. These results indicate that the presence of the {Au-PPh 3 } moiety in the complexes is crucial for the reduction of IL-1b secretion, but the coordinated N-donor organic ligand can dramatically change the activity of such a complex as it is visible on the example of the complexes 8 and 9 with low or no anti-inflammatory activity.
To determine whether the production of the pro-inflammatory cytokines TNF-a and IL-1b is regulated by a post-translation or post-transcription mechanism in the presence of the Au(I) complexes, the level of the corresponding mRNA was determined. The obtained results showed that selected representative complexes 2 and 7 were able to decrease the transcription of these cytokines after the LPS stimulation ( Figure 7). This finding indicates that the target site for the Au(I) complexes should be up-stream the transcription, which is in accordance with previous studies [15,44].
The transcription of the cytokines TNF-a and IL-1b is controlled by transcription factor NF-kB. This protein is kept in cytoplasm by its inhibitor IkB. After the activation of the signalling pathway, this inhibitor is degraded [45]. Gold-containing compounds are known for their ability to inhibit this signalling pathway [46]. The results from the herein presented study are in agreement with the suggested mechanism. Complexes 2 and 7, as well as Auranofin, were able to attenuate the degradation of IkB nonsignificantly ( Figure S4 in File S1). This blocking of the NF-kB signalling pathway leads to lower transcription of TNF-a and IL-1b genes and thus, decreases their secretion.
Interactions with a Mixture of Cysteine and Glutathione Analysed by ESI-MS As soft Lewis acids, Au(I) species prefer the formation of strong coordination bonds with soft Lewis base ligands, i.e. thiolate or selenolate ions, or phosphine derivatives, while the latter ones form the most stable bonds. It is a known fact, that Au(I) complexes behave likewise, as they bind to sulfanyl-containing substances, such as amino acid cysteine (Cys) or small proteins, such as glutathione (GSH), and with high molecular weight proteins (e.g. albumin or globulins [47]), in the biologically relevant environments (i.e. blood or serum) by the ligand exchange mechanism. The complexes formed this way can be considered as transport intermediates. The exchange of N-ligands for S-ligands occurs relatively quickly (within 20 minutes when interacting with albumin and globulins in the blood [48]), while the P-ligand exchange proceeds much more slowly. It seems that in this mechanism the cooperative effects of adjacent thiolato or selenolato ligands in the neighbourhood of the interaction site play an important role. In connection with the above mentioned, the ligand exchange is interpreted as one of the molecular mechanisms of incorporation of gold into the active site of selenium-containing flavoreductases, such as thioredoxin reductase [49]. In the scope of our work, we strived to uncover the molecular behaviour of selected complexes 1 and 6 (applied in the concentration of 20 mM, which approximately corresponds to the highest therapeutic blood levels of gold during chrysotherapy [50]) in biologically relevant conditions using a mixture of cysteine (at the 290 mM concentration) and reduced glutathione (at the 6 mM concentration) [36].
Based on the results of the ESI-MS experiments, we confirmed that both complexes 1 and 6 react with the used sulfhydrylcontaining substances in time-independent manner by the ligandexchange mechanism based on the substitution of the N-ligand (L n ) by the cysteine or glutathione molecule. This mechanism was confirmed by the emergence of ions at 1224.24 m/z, corresponding to the [(Au-PPh 3 ) 2 +GSH] + intermediate, and an ion at 1355.25 m/ z, corresponding to the ionic species [Cys+(Au-PPh 3 ) 2 +PPh 3 + CH 3 OH+Na] + (see Figure 8). Unlike our previous reports on the biological activities of Au(I) complexes [15,16], in this interaction experiments we observed both the intermediates involving cysteine and glutathione molecules. This might indicate that the herein reported compounds are much more susceptible to the sulfurcontaining molecules with several donor atoms, like glutathione.
In concordance with the above mentioned suggestion as well as in accordance with the previously reported behaviour of some Au(I) complexes in water-containing solutions [16], the mass spectra of the reacting systems involving sulfur-containing molecules and also the reference solutions of complexes revealed a considerable instability of the complexes demonstrated by the appearance of the intensive ion at 721.

Conclusions
A series of gold(I) complexes of the general composition [Au(L n )(PPh 3 )] (1-9), involving a combination of N-donor (HL n ) and P-donor triphenylphosphine ligands, was prepared and thoroughly characterized. The in vitro cytotoxicity results against a panel of nine human cancer cell lines (MCF7, HOS, A549, HeLa, A2780, A2780R, 22Rv1, G-361 and THP-1) revealed the complexes as more anticancer active than cisplatin, with the best IC 50 < 1-5 mM. Moreover, the calculated resistance factors, i.e. the ratio of the IC 50 values found for the cisplatin resistant and sensitive cell lines (A2780R)/IC 50 (A2780) showed that the complexes are able to circumvent the acquired resistance of cancer cells to cisplatin, as the resistance factor values are equal to 1.19 (4), 1.14 (5), 1.21 (6) as compared to the value of 2.25 for cisplatin. Further, the complexes 1-9 were evaluated for their in vitro anti-inflammatory activity on the model of the LPSactivated THP-1 monocytes. It was found out that the complexes 1-7 exhibited the ability to influence the cell-cycle of THP-1 cells resulting in the hormetic effect, and they were also able to significantly attenuate the production of pro-inflammatory cytokines TNF-a and IL-1b at comparable levels as gold(I)-based reference drug Auranofin, but with lower toxicity than this metallodrug. Based on the results of the ESI-MS experiments may be concluded that representative complexes 1 and 6 react with the sulfur-containing substances (cysteine and reduced glutathione) in time-independent manner by the ligand-exchange mechanism based on the substitution of the N-ligand (L n ) by the cysteine or glutathione molecule.