Synthesis, Characterization, and Cytotoxicity of the First Oxaliplatin Pt(IV) Derivative Having a TSPO Ligand in the Axial Position

The first Pt(IV) derivative of oxaliplatin carrying a ligand for TSPO (the 18-kDa mitochondrial translocator protein) has been developed. The expression of the translocator protein in the brain and liver of healthy humans is usually low, oppositely to steroid-synthesizing and rapidly proliferating tissues, where TSPO is much more abundant. The novel Pt(IV) complex, cis,trans,cis-[Pt(ethanedioato)Cl{2-(2-(4-(6,8-dichloro-3-(2-(dipropylamino)-2-oxoethyl)imidazo[1,2-a]pyridin-2-yl)phenoxy)acetate)-ethanolato}(1R,2R-DACH)] (DACH = diaminocyclohexane), has been fully characterized by spectroscopic and spectrometric techniques and tested in vitro against human MCF7 breast carcinoma, U87 glioblastoma, and LoVo colon adenocarcinoma cell lines. In addition, affinity for TSPO (IC50 = 18.64 nM), cellular uptake (ca. 2 times greater than that of oxaliplatin in LoVo cancer cells, after 24 h treatment), and perturbation of cell cycle progression were investigated. Although the new compound was less active than oxaliplatin and did not exploit a synergistic proapoptotic effect due to the presence of the TSPO ligand, it appears to be promising in a receptor-mediated drug targeting context towards TSPO-overexpressing tumors, in particular colorectal cancer (IC50 = 2.31 μM after 72 h treatment).


Introduction
The discovery of the antitumor activity of cisplatin, cis-diamminedichloridoplatinum(II) (Figure 1) [1,2], was a corner stone that triggered the interest in the development of platinum(II)/(IV) complexes in oncology. However, only a few of these complexes-carboplatin and oxaliplatin ( Figure 1)-have been approved worldwide for the use in the clinics [3][4][5].

Introduction
The discovery of the antitumor activity of cisplatin, cis-diamminedichloridoplatinum(II) (Figure 1) [1,2], was a corner stone that triggered the interest in the development of platinum(II)/(IV) complexes in oncology. However, only a few of these complexes-carboplatin and oxaliplatin ( Figure 1)-have been approved worldwide for the use in the clinics [3][4][5].  N,N-dipropylacetamide was designed as a biomarker for the diagnosis of neuroinflammation, neurodegeneration, and tumor progression [26].
Pursuing our interest for tissue-specific targeting of metal complexes, in recent papers we have synthesized platinum(II) compounds with ligands specific for TSPO, such as 2- [6,8- The resulting cisplatin-like compound, cis-[PtCl 2 (TZ6)] (Figure 2), has been shown to possess affinity and selectivity for the TSPO comparable to those of TZ6. In solvents with low dielectric constants, we also observed the formation of a dimeric aggregate formed through non-covalent intermolecular interactions of the planar aromatic cycles of the ligands. This finding further supports the potential intercalating ability of cis-[PtCl 2 (TZ6)] toward DNA [27]. The poor aqueous solubility of cis-[PtCl 2 (TZ6)], which is typical of platinum(II) complexes with bidentate aromatic ligands [28], compelled us to pursue the synthesis of Pt compounds with TSPO ligands endowed with enhanced water solubility. To this end, in a subsequent work we synthesized two new Pt compounds structurally analogous to picoplatin and differing for the anionic ligands, cis-[PtI 2 (NH 3 [29]. Both complexes contained the imidazopyridinic ligand [2-(4-chlorophenyl)-8-aminoimidazo-[1,2-a]pyridin-3-yl]-N,N-di-n-propylacetamide and were endowed with high affinity and selectivity for TSPO [30]. selectivity for the TSPO comparable to those of TZ6. In solvents with low dielectric constants, we also observed the formation of a dimeric aggregate formed through non-covalent intermolecular interactions of the planar aromatic cycles of the ligands. This finding further supports the potential intercalating ability of cis-[PtCl2(TZ6)] toward DNA [27]. The poor aqueous solubility of cis-[PtCl2(TZ6)], which is typical of platinum(II) complexes with bidentate aromatic ligands [28], compelled us to pursue the synthesis of Pt compounds with TSPO ligands endowed with enhanced water solubility. To this end, in a subsequent work we synthesized two new Pt compounds structurally analogous to picoplatin and differing for the anionic ligands,  [29]. Both complexes contained the imidazopyridinic ligand [2-(4-chlorophenyl)-8-aminoimidazo-[1,2-a]pyridin-3-yl]-N,N-di-n-propylacetamide and were endowed with high affinity and selectivity for TSPO [30].  In addition, the two compounds were massively accumulated in glioma cells (10-to 100-fold enhanced accumulation) and were capable of inducing apoptosis similar to cisplatin [30].

cis-[PtI
In this work we have developed the first Pt(IV) complex containing a TSPO ligand in axial position (the two axial positions are those perpendicular to the square-planar coordination plane of the platinum(II) precursor).
The design of a Pt(IV) complex with a TSPO ligand in the axial position was motivated by the reasons mentioned earlier in this paragraph: the targeting of the tumor site and the potential synergistic proapoptotic effect caused by the TSPO ligand released by intracellular reduction of the Pt(IV) conjugate. We selected the oxaliplatin derivative cis,trans,cis-[Pt(ethanedioato) Cl(2-hydroxyethanolato)(1R,2R-DACH)] (DACH = diaminocyclohexane) as Pt(IV) precursor In addition, the two compounds were massively accumulated in glioma cells (10-to 100-fold enhanced accumulation) and were capable of inducing apoptosis similar to cisplatin [30].
In this work we have developed the first Pt(IV) complex containing a TSPO ligand in axial position (the two axial positions are those perpendicular to the square-planar coordination plane of the platinum(II) precursor).
The design of a Pt(IV) complex with a TSPO ligand in the axial position was motivated by the reasons mentioned earlier in this paragraph: the targeting of the tumor site and the potential synergistic proapoptotic effect caused by the TSPO ligand released by intracellular reduction of the Pt(IV) conjugate. We selected the oxaliplatin derivative cis,trans,cis-[Pt(ethanedioato)Cl(2-hydroxyethanolato) (1R,2R-DACH)] (DACH = diaminocyclohexane) as Pt(IV) precursor (compound 1 in Figure 3) [31]. The choice of this monofunctional Pt(IV) derivative of oxaliplatin was guided by several considerations: (i) it contains a single reactive group (the terminal hydroxyl moiety of 2-hydroxyethanolato), thus preventing the formation of different condensation products (mono-and di-substituted) in the conjugation with biovectors; (ii) the presence of an axial chloride makes the reduction potential less negative (i.e., "easier" reduction from a thermodynamic point of view) and increases the reduction rate [12,32,33]; (iii) the 2-hydroxyethanolato residue should act as a spacer increasing the distance between the conjugated TSPO ligand and the Pt center, allowing an easier interaction with the receptor. (compound 1 in Figure 3) [31]. The choice of this monofunctional Pt(IV) derivative of oxaliplatin was guided by several considerations: (i) it contains a single reactive group (the terminal hydroxyl moiety of 2-hydroxyethanolato), thus preventing the formation of different condensation products (mono-and di-substituted) in the conjugation with biovectors; (ii) the presence of an axial chloride makes the reduction potential less negative (i.e., "easier" reduction from a thermodynamic point of view) and increases the reduction rate [12,32,33]; (iii) the 2-hydroxyethanolato residue should act as a spacer increasing the distance between the conjugated TSPO ligand and the Pt center, allowing an easier interaction with the receptor.  Hence, the Pt(IV) precursor was tethered to a potent TSPO ligand characterized by a 2-phenyl-imidazo[1,2-a]pyridine acetamide structure and containing a terminal carboxylic group useful for its further conjugation, 2-(4-(6,8-dichloro-3-(2-(dipropylamino)-2-oxoethyl)imidazo [1,2-a]pyridin-2-yl)phenoxy)acetic acid (compound 2 in Figure 3) [34]. In this way we  Figure 3).
The novel Pt(IV) complex has been fully characterized by spectroscopic and spectrometric techniques and tested in vitro against human MCF7 breast carcinoma, human U87 glioblastoma, and human LoVo colon adenocarcinoma cell lines. In addition, the affinity of 3 for TSPO was assessed by receptor binding assays, measuring its ability to displace [ 3 H]-1-(2-chlorophenyl)-Nmethyl-N-(1-methylpropyl)-3-isoquinolinecarboxamide ([ 3 H]-PK 11195). Cellular uptake experiments were performed in order to correlate the cytotoxicity of 3 with its cellular uptake ability. Finally, cell cycle analysis was performed to evaluate the capability of 3 to perturb the LoVo cells cycle progression.

Synthesis and Characterization of 3
Although platinum(IV) complexes were identified by Rosenberg as having anticancer activity at the same time of the discovery of that of cisplatin [1], their clinical efficacy has been tested only recently. The properties of Pt(IV) complexes are substantially different from those of Pt(II) species since six-coordinate octahedral platinum(IV) complexes have a saturated coordination sphere that renders them less prone to substitution reactions than platinum(II) complexes. In addition, the possibility to coordinate two additional ligands in the axial positions enables tuning of the chemical properties and conjugation of cancer-targeting ligands. The preliminar intracellular reduction to Pt(II), with simultaneous release of the two axial ligands, is necessary for Pt(IV) complexes to exert their antitumor activity. Thus, Pt(IV) complexes are generally defined as prodrugs of platinum-based drugs. In addition, the intracellular reduction of Pt(IV) complexes releases the axial ligands that can be themselves biologically active. The Pt(IV) precursor is indeed a dual-threating pharmaceutical agent, which combines two biologically active components into a single molecule [12,14]. In specifically designed Pt(IV) complexes, the Pt-core and the released axial ligands may have different intracellular targets; moreover, the axial ligands may exhibit a targeting property towards substrates that are specifically overexpressed in tumor cells. This is the rationale that prompted us to prepare a Pt(IV) complex with a TSPO ligand in the axial position.
Most of the Pt(IV) complexes present in the literature are prepared via oxidation of predesigned Pt(II) compounds with chlorine [35][36][37] or hydrogen peroxide [38], leading to Pt(IV) complexes with symmetric axial ligands. However, more recently, unsymmetric Pt(IV) complexes have also been synthesized [39,40].
For the preparation of the unsymmetric Pt(IV) derivative of oxaliplatin with just one TSPO ligand in axial position, we first treated oxaliplatin ( Figure 3) with N-chlorosuccinimide (NCS)-a source of "positive chlorine"-in ethane-1,2-diol, obtaining the asymmetric compound 1, as previously reported [31,41]. This synthetic strategy is based on the observation that the oxidative addition to Pt(II) complexes is generally assisted by the solvent coordinating opposite to the attacking "positive chlorine" [37,42,43]. In the present case, by using ethane-1,2-diol as solvent, this latter coordinates trans to chlorine [41]. The Pt(IV) complex 1 was then conjugated with 2 [34] forming the ester complex cis,trans,cis With reference to the coupling reaction, the best result was obtained using EDC and HOBt as coupling reagents in the presence of TEA in DMF; a similar synthetic approach was used for the conjugation of the glycolic monomer of PEG with PLGA [44].
Compound 3 was characterized by elemental analysis, ESI-MS, and NMR spectroscopy. In the ESI-MS spectrum a peak at m/z = 976.33, corresponding to [3 + Na] + , was evident (data not shown). The experimental isotopic pattern of this peak was in good agreement with the theoretical one.
The NMR characterization of the compound started from the assignment of methyl protons 12 and 15 (triplets falling at 0.82 and 0.87 ppm) of the dipropylacetamidic chains (see Figure 3 for numbering of protons), that show TOCSY cross-peaks ( Figure 4) with methylenic protons 11 and 14 (overlapping multiplets at 1.51 and 1.61 ppm). These signals show TOCSY cross-peaks with two signals, overlapping with the signal of the solvent, assigned to methylenes 10 and 13 falling at 3.34 and 3.25 ppm. The two singlets falling at 4.25 and 4.83 ppm were assigned to the methylene groups 9 and 22, respectively. This assignment was confirmed by a NOESY 2D NMR experiment that shows a NOESY cross-peak between the methylene group 9 and the imidazopyridine proton 5 located at 8.57 ppm (cross-peak A in Figure 5), and an additional NOESY cross-peak of the methylene 22 with the protons 17 and 19 of the phenoxy ring located at 7.03 ppm (cross-peak B in Figure 5). Proton 7 of the imidazopyridine ring was assigned to the singlet at δ = 7.65 ppm while the doublet located at 7.55 ppm, integrating for two protons, was assigned to protons 16 and 20 of the phenoxy ring ( Figure 4, TOCSY cross-peak with protons 17 and 19). With reference to the hydroxyethanolato spacer, the multiplet at 4.20 ppm was assigned to the methylenic protons 23. This signal shows a TOCSY cross-peak ( Figure 4) with the multiplet located at 3.23 ppm (overlapping with the signal of water) assigned to methylene 24. The 1R,2R-cychlohexanediamine is in a chair conformation and, owing to the axial asymmetry of the Pt complex, different signals are expected for both the axial and the equatorial protons. Therefore, the multiplet at 1.06 ppm was assigned to the axial protons 28 and 29, which showed TOCSY cross-peaks with the multiplet at 1.43-1.50 ppm (assigned to the axial protons 27 and 30 and to the equatorial protons 28 and 29 of 1R,2R-DACH), the multiplet at 2.01 ppm (attributed to equatorial protons 27 and 30), and two signals overlapping with the solvent (assigned to the methynic protons 25 and 26). All the chemical shifts found for the methylenic and methynic protons of DACH are in good agreement with those reported for similar asymmetric Pt(IV) octahedral complexes having a chlorido and glycolic moiety in the axial positions [41]. The asterisks indicate residual solvent peaks.
This assignment was confirmed by a NOESY 2D NMR experiment that shows a NOESY cross-peak between the methylene group 9 and the imidazopyridine proton 5 located at 8.57 ppm (cross-peak A in Figure 5), and an additional NOESY cross-peak of the methylene 22 with the protons 17 and 19 of the phenoxy ring located at 7.03 ppm (cross-peak B in Figure 5).
Proton 7 of the imidazopyridine ring was assigned to the singlet at δ = 7.65 ppm while the doublet located at 7.55 ppm, integrating for two protons, was assigned to protons 16 and 20 of the phenoxy ring ( Figure 4, TOCSY cross-peak with protons 17 and 19). With reference to the hydroxyethanolato spacer, the multiplet at 4.20 ppm was assigned to the methylenic protons 23. This signal shows a TOCSY cross-peak ( Figure 4) with the multiplet located at 3.23 ppm (overlapping with the signal of water) assigned to methylene 24. The 1R,2R-cychlohexanediamine is in a chair conformation and, owing to the axial asymmetry of the Pt complex, different signals are expected for both the axial and the equatorial protons. Therefore, the multiplet at 1.06 ppm was assigned to the axial protons 28 and 29, which showed TOCSY cross-peaks with the multiplet at 1.43-1.50 ppm (assigned to the axial protons 27 and 30 and to the equatorial protons 28 and 29 of 1R,2R-DACH), the multiplet at 2.01 ppm (attributed to equatorial protons 27 and 30), and two signals overlapping with the solvent (assigned to the methynic protons 25 and 26). All the chemical shifts found for the methylenic and methynic protons of DACH are in good agreement with those reported for similar asymmetric Pt(IV) octahedral complexes having a chlorido and glycolic moiety in the axial positions [41]. The [ 1 H-195 Pt]-HSQC 2D NMR spectrum recorded in DMSO-d6 is reported in Figure 6. The spectrum shows four cross peaks falling at 6.83/887.2, 7.21/887.2, 7.67/887.2, and 8.02/887.2 ppm ( 1 H/ 195 Pt) due to the coupling of 195 Pt with the four magnetically non-equivalent NH2 protons of DACH. In addition, a cross-peak located at 3.23/887.2 ppm is assigned to the coupling between 195 Pt and the CH2 of the glycolic linker. The 195 Pt chemical shift is in good agreement with those reported for similar asymmetric Pt(IV) octahedral complexes with a PtClN2O3 core [41].The assignment of 13 C signals has been accomplished by a [ 1 H-13 C]-HSQC 2D NMR spectrum ( Figure 7 and Table 1). The chemical shifts found for 1R,2R-DACH and the glycolic linker are in good agreement with those reported for similar Pt(IV) complexes [41]. In particular, the methylene groups of DACH fall at 23.  As far as the amine groups are concerned, it is possible to observe (Figure 4, top) the presence of four different signals for the NH 2 protons of coordinated DACH located at 6.83, 7.21, 7.67 and 8.02 ppm (only two signals are observed in the case of complexes with equal axial ligands).
The [ 1 H-195 Pt]-HSQC 2D NMR spectrum recorded in DMSO-d 6 is reported in Figure 6. The spectrum shows four cross peaks falling at 6.83/887.2, 7.21/887.2, 7.67/887.2, and 8.02/887.2 ppm ( 1 H/ 195 Pt) due to the coupling of 195 Pt with the four magnetically non-equivalent NH 2 protons of DACH. In addition, a cross-peak located at 3.23/887.2 ppm is assigned to the coupling between 195 Pt and the CH 2 of the glycolic linker. The 195 Pt chemical shift is in good agreement with those reported for similar asymmetric Pt(IV) octahedral complexes with a PtClN 2 O 3 core [41]. The assignment of 13 C signals has been accomplished by a [ 1 H-13 C]-HSQC 2D NMR spectrum (Figure 7 and Table 1). The chemical shifts found for 1R,2R-DACH and the glycolic linker are in good agreement with those reported for similar Pt(IV) complexes [41]. In particular, the methylene groups of DACH fall at 23.5 and 30.1 ppm, two values very similar to those reported for the product of esterification between BOC-L-alanine and cis,trans,cis-[Pt(cyclobutane-1,1 1 -dicarboxylate)Cl(2-hydroxyethanolato)(1R,2R-DACH)] (23.5, 30.1 and 30.7 ppm) [41]. The cross peak at 2.54/61.5 ppm ( 1 H/ 13 C) was assigned to the methynic group of DACH while the cross-peaks at 3.23/65.5 and 4.20/64.3 ppm ( 1 H/ 13 C) were assigned, respectively, to C24 and C23 of the glycolic moiety (66.0 and 64.5 ppm in the ester compound between BOC-L-alanine and cis,trans,cis-[Pt(cyclobutane-1,1 1 -dicarboxylate)Cl(2-hydroxyethanolato)(1R,2R-DACH)] mentioned above) [41]. The cross-peaks located at 122.3 and 123.5 ppm ( 13 C) were assigned to C5 and C7 of the imidazopyridine ring, respectively. The cross peaks falling at 7.03/114. 5      The asterisks indicate residual solvent peaks.

Stability of 3
The stability of 3 was investigated by 1 H-NMR spectroscopy in DMSO-d 6 /D 2 O (5:95, v/v) at pH 7.4 (2 mM phosphate buffer) and 37˝C. A portion of the spectra, recorded at different times, is reported in Figure 8

Stability of 3
The stability of 3 was investigated by 1 H-NMR spectroscopy in DMSO-d6/D2O (5:95, v/v) at pH 7.4 (2 mM phosphate buffer) and 37 °C. A portion of the spectra, recorded at different times, is reported in Figure 8

Stability of 3
The stability of 3 was investigated by 1 H-NMR spectroscopy in DMSO-d6/D2O (5:95, v/v) at pH 7.4 (2 mM phosphate buffer) and 37 °C. A portion of the spectra, recorded at different times, is reported in Figure 8     An HPLC analysis of the sample was performed after 108 h incubation at 37˝C. Two peaks were visible in the chromatogram (not shown) having retention times of 8.7 and 21.81 min, which are comparable to those found for the Pt(IV) precursor 1 and the free TSPO ligand 2, respectively. Therefore, the HPLC investigation confirmed that 3 undergoes a slow hydrolysis of the ester bond in physiological-like conditions. No reduction of the Pt(IV) complex occurred under the conditions used in this experiment.

Biological Assays
The affinity towards TSPO of complex 3 and of the free ligand 2 was evaluated by testing their ability to prevent binding to the rat cerebral cortex of the selective TSPO ligand [ 3 H]-PK 11195. The results, expressed as inhibitory concentration (IC 50 ), are listed in Table 2. Compound 3 showed high affinity for the TSPO receptor (IC 50 of 18.64 nM), although it was lower than that found for the free ligand 2 (IC 50 of 2.12 nM). The free ligand 2, in turn, was more affine than the reference ligand PK 11195 (IC 50 of 4.27 nM). These results indicate that the choice of a terminal carboxylic residue on the free ligand 2 for conjugation to a Pt(IV) complex was correct and does not significantly alter the affinity of 2 for the TSPO. The cytotoxic activity of complex 3 and of its precursors was assessed by the MTT in vitro assay performed in human MCF7 breast carcinoma, human U87 glioblastoma, and human LoVo colon adenocarcinoma cell lines. The cytotoxicity of the compounds, expressed as IC 50 values after 72 h incubation (except in two cases where it is expressed as the % of cell viability at the maximum concentration tested of 100 µM) are reported in Table 3. Table 3. Concentration inducing 50% cell survival inhibition (IC 50 ) (except in two cases where is given the % of cell viability at the maximum concentration of 100 µM) after 72 h treatment. Compound 2 showed only a marginal cytotoxic effect at 100 µM concentration (the highest concentration used in the investigation) in the case of MCF7 and LoVo cells and a weak activity against U87 cell line (IC 50 = 70.2 µM). Moreover, we could assess that compound 2 causes cell death through induction of apoptosis as evidenced by the formation of a sub-G0/G1 cell population (cell cycle analysis of LoVo cell line, see following discussion). These results are in line with previous results indicating that TSPO ligands, constructed on the imidazopyridinic scaffold, are endowed with good proapoptotic activity [30] although, in some cases, high concentrations are needed to achieve cytotoxic and proapoptotic effects, as also demonstrated for PK 11195 [45]. In particular, Kugler et al. [45] evidenced a peculiar property of TSPO ligands: being inert when no challenge is present but counteracting programmed cell death when lethal agents are present. They are therefore recognized to be potentially useful for the treatment of brain trauma and neurodegenerative brain diseases [45]. In the present case it is difficult to say if the low activity of compound 3 is due to the TSPO ligand 2 preventing cell death induction by oxaliplatin or to the intrinsic lower activity in cellular experiments of Pt(IV) species as compared to their Pt(II) counterparts. In this regard, further experiments will be carried out by giving the ligand 2 and the oxaliplatin derivative 1 simultaneously but as individual components.

Cell
In all three cell lines compound 3 had activity comparable to that of the Pt(IV) precursor complex 1 (IC 50 ratios of 1.7, 1.8, and 0.92 for MCF7, U87, and LoVo cell lines, respectively). In the above mentioned cell lines both Pt(IV) compounds were less active than oxaliplatin. This latter result is not surprising since Pt(IV) compounds usually show higher IC 50 values in in vitro investigation with respect to the related Pt(II) compounds since they require a preliminary intracellular reduction process before exerting their cytotoxic effect [46].
As already reported [8], oxaliplatin is a chemotherapeutic drug mainly used for the treatment of colorectal cancer, therefore, we extended our investigation only to LoVo cells. First we measured the intracellular accumulation of platinum by ICP-MS. LoVo cells were exposed to IC 50 concentrations of compounds 3, 1, and oxaliplatin for a short (4 h) and a long (24 h) period. Interestingly, the results (Table 4) indicate that for 24 h incubation the uptake of compound 3 is greater than that of 1 and that the latter is greater than that of oxaliplatin (with a ratio of 2 between the uptake of 3 and that of oxaliplatin). Interestingly, for the shorter time (4 h) the Pt uptake follows the order 3 > oxaliplatin > 1 indicating that factors other than passive diffusion might play a role at short time exposure. Most likely, organic cation transporters (OCT) may facilitate the uptake of oxaliplatin [47] at short time exposure and then, after saturation, the passive diffusion of the most lipophilic compounds becomes prominent. There does not appear to be a direct correlation between Pt uptake and cytotoxic effect (72 h) but, in our opinion, this is an intrinsic limitation of cellular testing where Pt(IV) compounds, which require reductive activation, generally result in less cytotoxicity than Pt(II) species and the directing role of TSPO ligands might remain hidden. In this case a more reliable answer can come from in vivo experiments which are being planned.
The capability of the new compounds 1 and 3 to perturb the cell cycle progression of LoVo cells at their IC 50 concentrations was investigated by FCM analysis and compared to that of oxaliplatin and of the TSPO ligand 2 (Figure 9). After 24 h treatment with the test compounds, cells exhibited only a slight modification of the cell cycle distribution. As expected, oxaliplatin delayed the S-phase and started to accumulate cells in G2/M-phase and G0/G1 phase [48]. Compounds 3 and 1 induced no notable changes compared to oxaliplatin. The TSPO ligand 2 induced a sub-G0/G1 cell population after 24 h, consistent with the formation of apoptotic cells as previously demonstrated for this class of compounds [30].
Compound 3 did not induce apoptosis as the free TSPO ligand 2 did, however, this is because the concentration of compound 2 released after intracellular hydrolysis of compound 3 is not sufficient to exert apoptosis (note that compound 2 was administered at a much higher dose due to its lower cytotoxicity). and of the TSPO ligand 2 (Figure 9). After 24 h treatment with the test compounds, cells exhibited only a slight modification of the cell cycle distribution. As expected, oxaliplatin delayed the S-phase and started to accumulate cells in G2/M-phase and G0/G1 phase [48]. Compounds 3 and 1 induced no notable changes compared to oxaliplatin. The TSPO ligand 2 induced a sub-G0/G1 cell population after 24 h, consistent with the formation of apoptotic cells as previously demonstrated for this class of compounds [30].

Stability of Compound 3
The stability of compound 3 in a physiological-like solution was investigated by NMR and HPLC. Because of the poor water solubility of 3, the complex was previously dissolved in DMSO-d 6 and then diluted with deuterated aqueous phosphate buffered saline solution (2 mM, pH 7.4) obtaining a final concentration of 52 µM in DMSO-d 6 /D 2 O (5:95, v/v). The sample was incubated at 37˝C in the dark and 1 H-NMR spectra recorded at different time intervals.

Membrane Preparation
Membranes from C6 rat glioma cells were prepared as described by Piccinonna et al. [51] with minor modifications. In brief, cells scraped in PBS and harvested, were then homogenized with a Brinkman Polytron (Kinematica AG, Münstertäler, Germany). After centrifugation at 37,000ˆg for 30 min at 4˝C the resulting pellet was resuspended in ice-cold 10 mM PBS and stored at´80˝C until use.

Receptor Binding Assays
The ability of complex 3 and the precursor TSPO ligand 2 to bind with high affinity to TSPO was assessed by in vitro receptor binding assays performed as described in our previous works [49,52]. In brief, a suspension of C6 rat glioma cells membranes (100 µg) in PBS was treated with 0.7 nM [ 3 H]-PK11195 (a selective ligand for TSPO) and compound 2 or 3 at different concentrations. After an incubation time of 90 min at 25˝C, the samples were rapidly filtered through Whatman GF/C glass microfiber filters and the filters washed with 3ˆ1 mL of ice-cold PBS. For the determination of the nonspecific binding the compound PK 11195 (10 µM) was used. A specific binding equal to 90% was determined under these experimental conditions.

Cell Proliferation Assay
Determination of cell growth inhibition was performed on MCF7, U87 and LoVo cell lines, using the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazoliumbromide (MTT) assay. The determination of the % of cell viability after treatment with the test compounds, was performed as described by Denora et al. [34]. Briefly, cells were dispensed on 96 microtiter plates at a density of 5000 cells/well and, after overnight incubation, were exposed for 72 h to solutions of the compounds with concentration in the range 0.01-100 µM. After addition of MTT (10 µL of 0.5% w/v in PBS) and incubation for an additional 3 h at 37˝C, the cells were lysed with 100 µL of DMSO/EtOH 1:1 (v/v) solution, dispensed in each well. A PerkinElmer 2030 multilabel reader Victor TM X3 (manufactured for WALLAC Oy, Turku, Finland by PerkinElmer Singapore was used for the absorbance determination at 570 nm.

Cell Cycle Analysis
Cells were dispensed in 60 mm tissue culture dishes at a density of 500,000 cells/dish and after 24 h were treated with compounds 1, 3, or oxaliplatin at their IC 50 concentrations or with compound 2 at 100 µM for an additional 24 h. Then, the experimental procedure described in our previous work was performed [53].

Cellular Uptake
A Varian 820-MS ICP mass spectrometer (Varian Italia, Cernusco sul Naviglio (MI), Italy) was used to perform the inductively coupled plasma mass spectrometry (ICP-MS) analyses for the determination of the platinum cellular uptake in LoVo colon cancer cells. Cells were seeded in 60 mm tissue culture dishes at a density of 500,000 cells/dish and incubated at 37˝C in a humidified atmosphere with 5% CO 2 . After 1 day, the culture medium was replaced with 3 mL of medium containing complex 3, the precursor 1, or oxaliplatin at their IC 50 concentrations, and incubated for 4 and 24 h. After the incubation period, the cell monolayer was washed twice with ice-cold PBS and then digested with 2 mL of an HNO 3 (67%)/H 2 O 2 (30%), 1:1 (v/v), solution for 4 h at 60˝C in a stove. The platinum content was determined by ICP-MS.

Conclusions
The first Pt(IV) complex containing a TSPO ligand in the axial position has been synthesized, fully characterized, and its cytotoxic activity tested in vitro. The TSPO ligand in the axial position of the conjugate 3 maintains the high affinity for the TSPO receptor in C6 rat glioma cells at nanomolar level. Therefore, the new compound appears to be very promising in a receptor-mediated drug targeting context towards TSPO-overexpressing tumors, in particular colorectal cancer.
Indeed, the best cytotoxic effect of 3 was observed in LoVo colon cancer cells. However, the cytotoxicity of 3 was 5-fold lower than that of oxaliplatin, which is a good result considering the different oxidation state of the metal in 3 and in oxaliplatin.
Remarkably, compound 3, which in MCF7 and U87 cell lines was less active than 1, proved to be more active than 1 in LoVo cells. Since TSPO is overexpressed in colorectal cancer [19], an additional increase of potency of 3 with respect to 1 could, in principle, derive from the presence of a TSPO ligand in axial position.
We expect a clearer answer about the potential of these combination compounds from in vivo testing, where the required reductive activation of the Pt(IV) precursor, and the hydrolysis necessary for the release of the TSPO ligand might not represent a handicap as it does for in vitro cellular testing.