The Cytotoxic Activity of Diiron Bis-Cyclopentadienyl Complexes with Bridging C₃-Ligands

Abstract

Diiron bis-cyclopentadienyl bis-carbonyl cationic complexes with a bridging vinyliminium ligand, [Fe₂Cp₂(CO)(μ-CO){μ-η¹:η³-C³(R′)C²HC¹NMe(R″)}]CF₃SO₃ (R = Xyl = 2,6-C₆H₃Me₂, R′ = Ph, R″ = H, 2a; R = Xyl, R′ = R″ = Me, 2b; R = R′ = Me, R″ = H, 2c; R = Me, R′ = 2-naphthyl, R″ = H, 2d; R = Me, R′ = R″ = Ph, 2e), are easily available from commercial chemicals, robust in aqueous media and exert a variable in vitro cytotoxicity against cancer cell lines depending on the nature of the substituents on the vinyliminium ligand. The anticancer activity is, at least in part, associated to fragmentation reactions, leading to iron oxidation and active neutral and well-defined monoiron species. We report an innovative synthetic procedure for the preparation of 2a,c,d, and a facile method to access the monoiron derivative of 2a, i.e., [FeCp(CO){C¹(NMeXyl)C²HC³(Ph)C(O)}] (3a). According to IC₅₀ analyses at different times of incubation of the complexes, 3a is significantly faster in inhibiting cell viability compared to its diiron precursor 2a. The neutral complexes [Fe₂Cp₂(CO)(μ-CO){μ-k¹N:k¹C:k¹C-C³(R′)C²(Se)C¹(NMe₂)C⁴(CO₂Y)C⁵(CO₂Y)}] (R′ = Y = Me, 4a; R′ = Pr, Y = ^tBu, 4b; R′ = Y = Et, 4c) are obtained via the two-step modification of the vinyliminium moiety and comprise a bridging selenophene-decorated alkylidene ligand. The antiproliferative activity exhibited by 4a-c is moderate but comparable on the ovarian cancer cell line A2780 and the corresponding cisplatin resistant cell line, A2780cisR. Complexes 4a-c in aqueous solutions undergo progressive release of the alkylidene ligand as a functionalized selenophene, this process being slower in cell culture medium. Since the released selenophenes SeC¹{C(O)R′}C²(NMe₂)C³(CO₂Y)C⁴(CO₂Y) (R′ = Y = Me, 5a; R′ = Pr, Y = ^tBu, 5b) are substantially not cytotoxic, it is presumable that the activity of 4a-c is largely ascribable to the {Fe₂Cp₂(CO)₂} scaffold.

1. Introduction

There is a current great interest in developing new anticancer drugs based on transition metals [1,2,3,4], due to the peculiar properties provided by the presence of the metal center (e.g., redox activity and broad possibility of geometries, stereochemical configurations and kinetic behaviors) [5] and the opportunity to overcome the drawbacks associated to the use of platinum compounds currently employed in clinical treatments against several types of tumors (in particular, severe side effects and acquired resistance towards the drug) [6,7].

In this regard, iron is an appealing choice as a metal element, being easily available, implicated in diverse biological processes in the human body, essentially nontoxic in many forms and displaying a rich redox chemistry [8,9]. Monometallic iron complexes with one or two (ferrocenes) cyclopentadienyl ligands (η⁵-C₅H₅ = Cp) have been widely investigated for their cytotoxicity [10,11,12], however these species may suffer from a limited structural variability and the absence of water solubility [13,14], which is a highly desirable characteristic for a drug candidate.

Some of us have been involved for longtime in the organometallic chemistry of diiron bis-cyclopentadienyl complexes: with a notable relevance to the natural hydrogenase enzymes, the cooperative diiron core is a key feature, in that it enables reactivity patterns on the bridging site which are otherwise not accessible on homologous monoiron complexes, and thus leads to the stereo- and regio-selective stepwise construction of functionalized organic fragments [15,16]. The co-presence of Cp (electron donor) and CO (electron acceptor) ligands confers a substantial robustness to the complexes [17], and the fine control of the structural variability makes available a range of compounds with suitable physico-chemical properties for potential biological applications (e.g., appreciable water solubility, amphiphilicity).

We recently reported the first studies on the antiproliferative activity of di-organoiron complexes [18,19], and in particular we found that vinyliminium complexes 2 are substantially water-stable and exhibit a variable in vitro antiproliferative activity depending on the substituents on N, C¹ and C² [20,21,22]; 2 are usually prepared by means of a straightforward four-step procedure from the commercial [Fe₂Cp₂(CO)₄], via the intermediacy of aminocarbyne complexes 1 (Scheme 1, path a) [20,23]. Remarkably, IC₅₀ values in the nanomolar range have been recognized on A2780 and A2780cisR cells for some complexes of the series 2. The action of 2 seems multimodal, including the interaction with biomolecules and the interference with cellular redox processes, and in particular the compounds possess the ability to trigger the production of Reactive Oxygen Species (ROS) inside the tumor cells [24,25,26]. In principle, two processes might contribute to this phenomenon: (1) a slow, extensive degradation of the diiron scaffold, with the released iron(I) centers rapidly converting into iron(III) oxide [20]; (2) a monoelectron reduction of the cationic complexes, promoting a fragmentation affording iron(0) and the monoiron complexes 3, which are themselves cytotoxic through the production of ROS and display a significant selectivity with reference to the noncancerous HEK-293 cell line [27]. The optimal laboratory preparation of 3 consists in the reaction of the precursors 2 with an amine (Scheme 1, path b). The neutral diiron complexes 4a-c are accessible by means of a two-step synthesis consisting in the incorporation of a selenium atom within the vinyliminium moiety, followed by cyclization reaction with internal alkynes bearing electron withdrawing substituents (Scheme 1c) [28]. Complexes 4a-c are based on a {Fe₂Cp₂(CO)₂} core and comprise an additional bridging ligand which is an alkylidene derivatized with an unusually functionalized selenophene. Complexes of type 4 slowly break up into their constituting pieces in aqueous solution in the presence of air, i.e., cyclopentadiene, carbon monoxide, the iron centers (converting into Fe^III oxides) and the tetra-substituted selenophenes 5 (Scheme 1d) [29]. A very rare example of straightforward dissociation of a bridging Fischer alkylidene under mild conditions [30]. This approach has been proposed as a metal-assisted synthesis of organic molecules (5), otherwise not accessible following conventional routes. However, the expected, progressive release of the selenophene in physiological media, the related redox process involving the two irons and the fact that selenium may play some role against cancer [31,32,33,34,35], render 4 potentially intriguing from a biological perspective. Herein, we present an extension of our studies, including novel synthetic routes and the assessment of the cytotoxicity of various compounds under different conditions, with the primary aim of shedding light on the anticancer potential of diiron bis-cyclopentadienyl complexes and their mode of action.

2. Results and Discussion

First, we developed a new and convenient method for the synthesis of the diiron vinyliminium complexes 2a,c,d from the parent aminocarbyne complexes 1a-b (Scheme 2, path a). The displacement of the carbonyl ligand by means of trimethylamine-N-oxide was carried out in methanol solution in the presence of the alkyne, and led to isolation of 2a,c,d in good to high yields, following alumina chromatography. Complexes 2b,e were instead obtained according to the literature [20,36]. Using the improved synthetic procedure (Scheme 2), which is performed in one-pot without the needing of anhydrous solvents, the yields of 2a,c,d are similar to those obtained by means of the two-step synthesis [20,37].

Complexes 2a,c,d were unambiguously identified by means of their IR and ¹H NMR spectra (Figures S1–S3), being fully consistent with those reported previously (Figures S1–S3). The IR spectra of 2a-e (in CH₂Cl₂) share a common pattern with three bands, related to the terminal carbonyl ligand, the bridging carbonyl and the iminium function, respectively. For instance, in the case of 2c, such bands occur at 1990, 1806 and 1685 cm⁻¹, in the order given. According to ¹H NMR spectroscopy, complex 2a exists as a 2:1 mixture of E and Z isomers related to the nitrogen substituents, due to hampered rotation around the C¹-N axis, and both isomers share a common cis geometry of the Cp rings. The E/Z isomeric ratio resembles that one recognized following the former procedure. On the other hand, the NMR spectra of 2c,d display a single set of resonances, ascribable to the cis products. The D₂O solubility values of 2a-e were previously reported [20], ranging in between 4.3 × 10⁻⁴ (2e) and 1.4 × 10⁻² (2c) molar.

The vinyliminium complexes 2a-e were assessed for their cytotoxic activity against the human lung A549 and human pancreatic BxPC-3 cancer cell lines following 72 h incubation. The results are shown in

Table 1. IC₅₀ values (μM) determined for iron complexes and cisplatin on cell lines after 72 h exposure, unless otherwise specified. Values are given as the mean ± SD. Italic: previously reported data from ref. [20]. ⁽*⁾ Exposure time: 24 h.

Comp.	A2780	A2780cisR	A549	BxPC-3
2a	0.50 ± 0.06	1.2 ± 0.2	43 ± 4	11.5 ± 1.4
2a	5.3 ± 1.2 (*)	25 ± 2 (*)
2b	0.90 ± 0.06	4.0 ± 0.3	>100	>100
2c	35 ± 3	86 ± 7	>100	>100
2d	2.9 ± 0.2	9.3 ± 0.7	>100	92 ± 2
2e	1.8 ± 0.2	7.5 ± 0.3	>100	>100
3a	16 ± 2	26 ± 3
3a	37 ± 2 (*)	86 ± 4 (*)
4a	25.9 ± 1.6	29 ± 3
4b	37 ± 3	42 ± 6
4c	39 ± 2	50 ± 6
5a	64 ± 4	>100
5b	>100	>100
cisplatin	19.0 ± 1.1 (*)	71 ± 3 (*)
cisplatin	0.40 ± 0.07	26 ± 3	9.5 ± 0.6	4.6 ± 0.4

, together with previously reported IC₅₀ values concerning A2780 and A2780cisR cell lines [20]. Thus, complexes 2b-e are substantially inactive against the newly investigated cell lines, whereas 2a displays a significant cytotoxicity, with the IC₅₀ obtained on the BxPC-3 line not so far from that of the reference drug cisplatin under the same conditions.

As explained in the Introduction, the activity of 2a-e may be imputable, at least in part, to redox processes associated to either a slow degradation of the complexes or their monoelectron reduction, the latter in some cases affording monoiron derivatives as organometallic products. In this regard, during an experiment aimed to test the stability of 2a in dimethylsulfoxide at 37 °C, we serendipitously discovered a clean conversion into the monoiron species 3a and cyclopentadiene (Scheme 2, path b). Complex 3a, which is almost insoluble in water, was identified based on a comparison of IR and ¹H NMR data with those available in the literature [27]. The IR spectrum, in dichloromethane solution, exhibits three characteristic bands due to the carbonyl ligand (1919 cm⁻¹), the acyl group (1612 cm⁻¹) and the aminoalkylidene moiety (1571 cm⁻¹). The ¹H NMR spectrum (in the DMSO-d₆ reaction medium; Figure S4) displays a single set of resonances ascribable to the E-configuration of the nitrogen substituents, with the alkenic CH proton resonating at 6.78 ppm. With reference to the biological action, although the conversion of 2a into 3a does not occur in water, it cannot be ruled out that the fragmentation of complexes of type 2 leading to 3 takes place in a physiological environment to some extent not exclusively through a redox mechanism (see Introduction) [20,38].

In order to outline a possible correlation between the activity of 2a and that of 3a, we investigated the influence of time on the IC₅₀ values on A2780 and A2780cisR cells. The ratio between the IC₅₀ values after 24 and 72 h of incubation, respectively, are 11 (A2780) and 21 (A2780cisR) for 2a and 2.3 (A2780) and 3.3 (A2780cisR) for 3a. This set of data highlights that the action of the diiron 2a is much slower than that of the monoiron 3a which is, however, less cytotoxic; it may be concluded that 2a is better internalized than 3a inside cells, and the kinetic of its antiproliferative activity is correlated to fragmentation routes (including that leading to 3a).

Relevant to the cytotoxicity trials, the stability of 4a-c was monitored at variable times in DMSO-d₆/D₂O mixtures, maintained at 37 °C (see Experimental and

Table 2. Fraction (%) of diiron complex (measured as released cyclopentadiene, two CpH units being released per molecule [19]) degraded in aqueous solutions stored at 37 °C after variable times (hours).

Comp.	DMSO-d6/D2O				DMSO-d6/DMEM-d
	6 h	24 h	48 h	72 h	6 h	24 h	72 h
4a	3	13	27	70	5	9	100
4b	25	71	100	100	13	49	100
4c	17	60	88	100	9	20	100

for details). According to ¹H NMR spectroscopy, the degradation rate of the complexes follows the trend 4a < 4c < 4b, however extensive degradation occurs after 72 h in all cases (Table 2). The same trend was observed in analogous DMSO-d₆/deuterated cell culture medium (DMEM-d) mixtures after 24 h, albeit with a slower process concerning especially 4c and 4b. The diiron complexes 4a-c exhibit a moderate antiproliferative activity after 72 h incubation against the A2780 cells (Table 1); it is noteworthy that such activity is similar to that detected for the same compounds on the cisplatin resistant A2780cisR cell line. On the other hand, the cationic vinyliminium complexes 2 are much less cytotoxic towards the A2780cisR cells with respect to the A2780 ones. Interestingly, the selenophenes 5a-b are significantly less cytotoxic than the corresponding precursors 4a-b (Table 1); this fact outlines that the capability of the investigated di-organoiron complexes to kill cancer cells may be mainly ascribable to the interference with the cell redox balance promoted by the iron centers, with a possible secondary contribution offered by the bridging organic moiety. In addition, carbon monoxide might play some adjuvant role, and investigations are in due course to give insight into this feature. It should be mentioned that a range of transition metal carbonyl complexes have been widely studied for the therapeutic properties associated to the controlled release of the carbonyl ligand(s) [39].

3. Conclusions

The search for new transition metal anticancer drugs, overcoming the limitations associated to the administration of platinum chemotherapics, is an intensively explored field of research, and, in this regard, iron is a convenient choice as a metal element essentially due to its limited intrinsic toxicity. Various iron compounds have been investigated for their anticancer potential, however studies on diiron complexes are still in their infancy, despite several advantages provided by the presence of two cooperating metal centers. Following our recent description of the promising cytotoxic activity of cationic diiron complexes with a bridging vinyliminium ligand, here we report additional studies on a selection of such category of complexes. One compound of the series, containing a phenyl group on the C₃ chain, emerged as an effective cytotoxic agent against a panel of cancer cell lines. Since the mode of action seems ascribable, inter alia, to a slow fragmentation occurring in cells and affording a monoiron vinyl-aminoalkylidene derivative, we monitored the antiproliferative activities of the diiron phenyl precursor and of its monoiron counterpart at different times of incubation. The results support the hypothesis of the fragmentation reaction as one of the possible mechanisms of anticancer action. Then, analyses on neutral diiron complexes containing a selenophene-decorated ligand evidence a scarce influence of the organic moiety on the cytotoxicity, which is presumably best correlated to the redox activity of the iron centers and, eventually, to CO release. In summary, diiron organometallic complexes constitute an arsenal of robust compounds with ideal properties for a drug candidate (e.g., easy availability and straightforward synthesis, appreciable water solubility, water stability and amphiphilicity), and the wide opportunity for structural variability offered by the diiron core encourages further studies in the direction to find out optimal combinations of ligands and substituents for the development of potential pharmaceuticals deserving more advanced biological evaluation.

4. Experimental

4.1. Materials and Methods

Solvents and organic reactants were purchased from Merck or TCI Europe. Compounds 1a [40], 2b,e [20,36], 3a [27], 4a-c [29] and 5a-b [29] were synthesized according to the respective literature procedures; once obtained, all the products were stored in air. Chromatography separations were carried out on columns of deactivated alumina (Merck, 4% w/w water), unless otherwise specified. Infrared spectra of solutions were recorded on a Perkin Elmer Spectrum 100 FT-IR spectrometer with a CaF₂ liquid transmission cell (2300–1500 cm⁻¹ range) and processed with Spectragryph software [41]. NMR spectra were recorded on a Bruker Avance II DRX400 instrument equipped with a BBFO broadband probe. Chemical shifts (expressed in parts per million) are referenced to the residual solvent peaks (¹H) [42]. NMR signals due to a second isomeric form (where it has been possible to detect them) are italicized. Elemental analyses were performed on a Vario MICRO cube instrument (Elementar).

4.2. Synthesis of [Fe₂Cp₂(CO)(μ-CO){μ-η¹:η³-C³(R′)C²(R″)C¹NMe(R)}]CF₃SO₃, (2a,c,d)

General procedure: A red solution of 1a-b (ca. 0.15 mmol) and Me₃NO (ca. 0.30 mmol) in MeOH (5 mL) was treated with the appropriate alkyne (ca. 0.45 mmol), in an open system allowing the release of gaseous products (Me₃N and CO₂). The mixture was heated at 45 °C for ca. 4 h. Conversion was checked by IR spectroscopy and volatiles were removed under vacuum. The residue was dissolved in CH₂Cl₂ and moved on top of an alumina column (neutral, Brockmann activity I; h 6, d 2.3 cm). Impurities were eluted with CH₂Cl₂ and THF, then a band was collected using MeCN. The eluate was dried under vacuum and the resulting solid was triturated in Et₂O. The suspension was filtered; the solid was washed with Et₂O and dried under vacuum.

[Fe₂Cp₂(CO)(μ-CO){μ-η¹:η³-C³(Ph)C²(H)C¹NMe(Xyl)}]CF₃SO₃ (2a) [20,37]. From 1a and phenylacetylene. Brown solid, yield 70%. Anal. Calcd. for C₃₁H₂₈Fe₂F₃NO₅S: C, 53.55; H, 4.06; N, 2.01. Found: C, 53.42; H, 4.13; N, 1.94. IR (CH₂Cl₂): ῦ/cm⁻¹ = 2000vs (CO), 1819s (µ-CO), 1628m (C²C¹N). ¹H NMR (acetone-d₆): δ/ppm = 7.90 (d, ³J_HH ≈ 7.4 Hz), 7.63–7.43 (m, 5 H, Ph); 7.42–7.34 (m), 7.25 (t), 7.17 (d), 7.08 (d) (³J_HH ≈ 7.5 Hz, 3 H, C₆H₃Me₂); 5.67, 5.33 (s, 5 H, Cp); 5.38, 5.07 (s, 5 H, Cp); 5.13, 4.28 (s, 1 H, C²H); 4.42, 3.74 (s, 3 H, NMe); 2.60, 2.33, 2.08, 1.88 (s, 6 H, C₆H₃Me₂). Cis-E/cis-Z ratio ca. 2:1.

[Fe₂Cp₂(CO)(μ-CO){μ-η¹:η³-C³(Me)C²(H)C¹NMe₂}]CF₃SO₃ (2c) [20,23]. From 1b and propyne (THF solution ca. 1 mol/L). Brown solid, yield 90%. Anal. Calcd. for C₁₉H₂₀F₃Fe₂NO₅S: C, 42.02; H, 3.71; N, 2.58; S, 5.90. Found: C, 42.11; H, 3.61; N, 2.68; S, 5.80. IR (CH₂Cl₂): ῦ/cm⁻¹ = 1990vs (CO), 1806s (μ-CO), 1685m (C²C¹N). ¹H NMR (acetone-d₆): δ/ppm = 5.50 (s, 5 H, Cp); 5.18 (s, 5 H, Cp); 4.59 (s, 1 H, C²H); 3.95, 3.94 (s, 6 H, NMe + C³Me), 3.33 (s, 3 H, NMe). Cis isomer.

[Fe₂Cp₂(CO)(μ-CO){μ-η¹:η³-C³(2-naphthyl)C²(H)C¹NMe₂}]CF₃SO₃(2d) [20]. From 1b and 1-ethynylnaphthalene. Dark-green solid, yield 88%. Anal. Calcd. for C₂₈H₂₄F₃Fe₂NO₅S: C, 51.32; H, 3.69; N, 2.14; S, 4.89. Found: C, 51.20; H, 3.77; N, 2.03; S, 4.96. IR (CH₂Cl₂): ῦ/cm⁻¹ = 1988vs (CO), 1809s (μ-CO), 1687m (C²C¹N). ¹H NMR (acetone-d₆): δ/ppm = 8.22 (d, J = 7.2 Hz), 8.07 (d, J = 7.7 Hz), 8.00 (t, J = 7.0 Hz), 7.74 (t, J = 7.7 Hz), 7.65–7.53 (m) (7 H, C₉H₇); 5.54 (s, 5 H, Cp), 5.10 (s, 5 H, Cp), 4.79 (s, 1 H, C²H); 4.08, 3.51 (s, 6 H, NMe₂). Cis isomer.

4.3. Stability in DMSO/Water Mixtures

4.3.1. Formation of [FeCp(CO){C¹(NMeXyl)C²HC³(Ph)C(O)}], 3a, in DMSO-d₆

Under a N₂ atmosphere, compound 2a (ca. 4 mg) was dissolved in DMSO-d₆ (0.9 mL). The orange-brown solution was analysed by ¹H NMR then heated at 37 °C for 72 h. Next, the mixture was filtered over celite and ¹H NMR spectrum was recorded. ¹H NMR analysis revealed a quantitative conversion into a mixture of 3a and cyclopentadiene (1:1 molar ratio). 2a. ¹H NMR (DMSO-d₆): δ/ppm = 7.82 (d), 7.58 (t), 7.32 (d), 7.22 (t), 7.16 (d), 7.05 (d) (³J_HH ≈ 7.6 Hz, 3 H, C₆H₃Me₂); 7.50–7.43, 7.40–7.34 (m, 5 H, Ph); 5.61, 5.33, 5.28, 4.96 (s, 10 H, Cp); 5.11, 4.10 (s, 1 H, C²H); 4.25, 3.57 (s, 3 H, NMe); 2.46*, 2.22, 1.97, 1.77 (s, 6H, C₆H₃Me₂). *Over DMSO-d₆ peak. Cis-E/cis-Z ratio ca. 7:1. 3a. ¹H NMR (DMSO-d₆): δ/ppm = 7.40–7.13 (m, 8 H, C₆H₃Me₂ + Ph); 6.78 (s, 1H, C²H); 4.70 (s, 5 H Cp); 3.82 (s, 3 H, NMe); 2.22, 2.07 (s, 6 H, C₆H₃Me₂). CpH. ¹H NMR (DMSO-d₆): δ/ppm = 6.56 (m, 2 H), 6.47 (m, 2 H), 2.95 (m, 2 H). The mixture was extracted with hexane (2 × 10 mL), then the volatiles were removed under vacuum. Subsequent IR spectrum evidenced the presence of 3a as unique carbonyl species. IR (CH₂Cl₂): ῦ/cm⁻¹ = 1919vs (CO), 1612m (CO_acyl), 1571m (C¹N).

4.3.2. Stability in DMSO-d₆/D₂O

The selected diiron complex (ca. 8 mg) was dissolved in 5 mL of a mixture of DMSO-d₆ and D₂O (4:1 v/v), and the resulting solution was thermostated at 37 °C. An aliquot of the solution was withdrawn after 7, 24, 48 and 72 h, filtrated through a short celite pad and analyzed by ¹H NMR (Table 2).

4.3.3. Stability in DMSO-d₆/DMEM-d Cell Culture Medium

The selected diiron complex (ca. 8 mg) was dissolved in 5 mL of a mixture of DMSO-d₆ and DMEM-d cell culture medium (4:1 v/v), and the resulting solution was thermostated at 37 °C. An aliquot of the solution was withdrawn after 6, 24 and 72 h, filtrated through a short celite pad and analyzed by ¹H NMR (Table 2).

4.4. In Vitro Cytotoxicity Investigation

4.4.1. Cell Lines

In vitro cytotoxicity investigations were carried out by using human ovarian carcinoma cisplatin-sensitive A2780 (ECACC93112519), human ovarian carcinoma cisplatin-resistant A2780cisR (ECACC 93112517), human lung carcinoma A549 (CCL-185) and human pancreas adenocarcinoma BxPC-3 cell lines (CRL-1687). A2780 and A2780cisR were purchased from the European Collection of Authenticated Cell Cultures (ECACC), and A549 and BxPC-3 cell lines from the American Type Culture Collection (ATCC).

The cell lines were propagated as indicated by the supplier in RPMI 1640 (Merck-A2780, A2780cisR and BxPC-3) and Ham’s F12 medium (Merck-A549) containing 2 mM of L-glutamine (Merck), 1% of penicillin/streptomycin solution (Merck-10,000 U ml⁻¹: 10 mg ml⁻¹), 10% of fetal bovine serum (Merck-FBS) and antimycotic. The acquired resistance of A2780cisR cells was maintained by routine supplementation of media with 1 μM of cisplatin and the BxPC-3 medium was also supplemented with 1% of sodium pyruvate (Merck). The cells were maintained under standard tissue culture conditions of 37 °C and 5% atmosphere of CO₂.

4.4.2. Cytotoxicity Assay

Prior to incubation with the selected complex for 72 h, A2780, A2780cisR, BxPC-3 and A549 cells were seeded in 96 wells tissue culture polystyrene plates at a concentration of 3 × 10³, 6 × 10³, 7 × 10³ and 3 × 10³ cells per well, respectively, and allowed to proliferate for 24 h at 37 °C in a 5% atmosphere of CO₂. For the assay at 24 h, A2780 and A2780cisR cells were seeded at a concentration of 1 × 10⁴ and 1.5 × 10⁴ cells per well, respectively. Stock solutions of compounds were prepared in DMSO and were sequentially diluted in medium (final DMSO concentration of 0.5%) to achieve the final desired concentration (0–100 µM). Cells incubated with cisplatin (0−100 μM) were used as positive control. Cells were treated with the different concentration of compounds for 24 or 72 h at 37 °C in a 5% atmosphere of CO₂. At the end of the incubation time, cell viability was assessed by means of WST-1 tetrazolium salt reagent (Roche). Briefly, cells were incubated for 4 h with WST-1 reagent diluted 1:10, at 37 °C and 5% CO₂. Measurements of formazan dye absorbance, which directly correlates with the number of viable cells, were carried out with a micro-plate reader (Biorad, Milan) at 450 nm, using 655 nm as reference wavelength. The 50% inhibitory concentration of tested compound (IC₅₀) refers to concentration at which 50% of cell death is observed with respect to the control. All the in vitro biological tests were performed in triplicate. Concentration effect curves were generated by nonlinear regression curves (GraphPad Prism) and the data are reported as mean ± standard deviation.