New Oxaliplatin-Pyrophosphato Analogs with Improved In Vitro Cytotoxicity

Two new Pt(II)-pyrophosphato complexes containing the carrier ligands cis-1,3-diaminocyclohexane (cis-1,3-DACH) and trans-1,2-diamine-4-cyclohexene (1,2-DACHEX), variants of the 1R,2R-diaminocyclohexane ligand present in the clinically used Pt-drug oxaliplatin, have been synthesized with the aim of developing new potential antitumor drugs with high bone tropism. The complexes are more stable at physiological pH than in acid conditions, with Na2[Pt(pyrophosphato)(cis-1,3-DACH)] (1) slightly more stable than [Pt(dihydrogenpyrophosphato)(1,2-DACHEX)] (2). The greater reactivity at acidic pH ensures a greater efficacy at the tumor site. Preliminary NMR studies indicate that 1 and 2 react slowly with 5’-GMP (used as a model of nucleic acids), releasing the pyrophosphate ligand and affording the bis 5’-GMP adduct. In vitro cytotoxicity assays performed against a panel of four human cancer cell lines have shown that both compounds are more active than oxaliplatin. Flow cytometry studies on HCT116 cells showed that the pyrophosphato compounds with the non-classical 1,3- and 1,4-diaminocyclohexane ligands (1 and 4) are the most capable to induce cells’ death by apoptosis and necrosis.


Introduction
Cisplatin is one of the most potent drugs used for the treatment of a variety of cancers, including ovarian, testicular, small-cell lung, and colorectal cancer [1][2][3][4]. Severe side effects, such as nephro-, neuro-, and oto-toxicity, limit its use [5,6]. In order to find more effective chemotherapeutics with reduced toxicity, many new platinum complexes have been synthesized and tested for their anticancer activity. However, only carboplatin and oxaliplatin have obtained the Food and Drug Administration (FDA) approval [7]. Despite the progress in antitumor therapy, the search for innovative pharmacological strategies that can limit, at least in part, some of the unsolved therapeutic problems, remains a very active field of research [8]. A strategy for reducing the systemic side effects of Pt-based drugs consists in the choice of carrier ligands that promote the specific accumulation of the drug in target organs or cells [9][10][11]. One remarkable example is represented by bisphosphonate ligands (BPs), that can deliver the therapeutic agent specifically to the bone tissue, and can be ascribed to the Drug Targeting and Delivery (DTD) approach, together with the several platinum(II) complexes with phosphonate ligands already investigated [12][13][14][15][16][17][18][19]. Another class of platinum compounds with bone tropism is represented by phosphaplatins, where the platinum ion is coordinated to two cisN-donors and to a single bidentate

Synthesis and NMR Characterization
The phosphaplatin derivative 3, also known as PT-112, is currently under clinical trials for the treatment of advanced solid tumors, either alone or in combination with avelumab (a programmed cell death, PD-1 inhibitor). Interestingly, complex 3 is capable of inducing an immunogenic form of cancer cell death [28]. Inspired by these results and pursuing our interest on the pharmacological effect of small modifications in the diamine carrier ligand (which already led to the synthesis of compound 4 containing the cis-1,4-DACH ligand), we have now prepared two more phosphaplatin analogues of 3 bearing the cis-1,3-DACH (1) and the 1,2-DACHEX carrier ligands (2).

Synthesis and NMR Characterization
The phosphaplatin derivative 3, also known as PT-112, is currently under clinical trials for the treatment of advanced solid tumors, either alone or in combination with avelumab (a programmed cell death, PD-1 inhibitor). Interestingly, complex 3 is capable of inducing an immunogenic form of cancer cell death [28]. Inspired by these results and pursuing our interest on the pharmacological effect of small modifications in the diamine carrier ligand (which already led to the synthesis of compound 4 containing the cis-1,4-DACH ligand), we have now prepared two more phosphaplatin analogues of 3 bearing the cis-1,3-DACH (1) and the 1,2-DACHEX carrier ligands (2). Na 2 [Pt(pyrophosphato)(cis-1,3-DACH)] (1) was prepared following a two-step reaction. The precursor complex [Pt(CO 3 )(cis-1,3-DACH)] was obtained by treating [PtI 2 (cis-1,3-DACH)] with Ag 2 CO 3 in water. The carbonate precursor was then treated with a slight defect of sodium pyrophosphate at pH = 8.0, affording the desired compound 1, which was obtained as a disodium salt by precipitation with acetone (Scheme 1).

Synthesis and NMR Characterization
The phosphaplatin derivative 3, also known as PT-112, is currently under clinical trials for the treatment of advanced solid tumors, either alone or in combination with avelumab (a programmed cell death, PD-1 inhibitor). Interestingly, complex 3 is capable of inducing an immunogenic form of cancer cell death [28]. Inspired by these results and pursuing our interest on the pharmacological effect of small modifications in the diamine carrier ligand (which already led to the synthesis of compound 4 containing the cis-1,4-DACH ligand), we have now prepared two more phosphaplatin analogues of 3 bearing the cis-1,3-DACH (1) and the 1,2-DACHEX carrier ligands (2).
Na2[Pt(pyrophosphato)(cis-1,3-DACH)] (1) was prepared following a two-step reaction. The precursor complex [Pt(CO3)(cis-1,3-DACH)] was obtained by treating [PtI2(cis-1,3-DACH)] with Ag2CO3 in water. The carbonate precursor was then treated with a slight defect of sodium pyrophosphate at pH = 8.0, affording the desired compound 1, which was obtained as a disodium salt by precipitation with acetone (Scheme 1). Compound 1 was characterized by elemental analysis, ESI-MS, and multinuclear 1D and 2D NMR. The ESI-MS spectrum showed the presence of a peak at m/z = 551.9617, corresponding to [1+Na] + , with the experimental isotopic pattern of the peak in good agreement with the theoretical one. Both the elemental analysis and the ESI-MS are in agreement with the formation of a complex having Pt and pyrophosphate in a 1:1 ratio. The 1 H-NMR characterization of compound 1 was carried on with the help of a COSY spectrum in D 2 O (Figure 2a). The cross-peak A correlates the signals at 4.53 and 1.85 ppm that were assigned to H 5ax and H 5eq , respectively. Of note is the huge de-shielding of H 5ax , which indicates that the cyclohexane ring is forced in a rigid chair conformation with H 5ax falling in the de-shielding cone of the Pt atom. The cross-peaks which correlate the signals at 2.73/1.88 (B) and at 2.73/1.56 (D) ppm were assigned to H 1/3 /H 2eq and H 1/3 /H 2ax , respectively. Finally, the signals at 2.73 and 1.74 ppm, which are correlated by cross-peak C, could be assigned to the methynic protons H 1/3 and H 4/6 , respectively. In the [ 1 H, 13 C] HSQC spectrum (Figure 2b), the cross-peaks ( 13 C/ 1 H) falling at 16.74/1.85 (F) and 16.74/4.53 (E) ppm confirmed the huge separation (2.68 ppm) between the H 5 axial and the H 5 equatorial protons. The cross-peak observed at 30.00/1.74 (D) ppm allowed the assignment of C 4/6 , while the cross-peaks falling at 36.13/1.88 (B) and 36.13/1.56 (C) ppm ( 13 C/ 1 H) belong to C 2 /H 2eq and C 2 /H 2ax , respectively. Finally, cross-peak A, resonating at 45.95/2.73 ppm, was assigned to C 1/3 . The 31 P NMR of 1 dissolved in D 2 O is reported in Figure 2c. A single peak with unresolved platinum satellites (platinum satellites usually broaden with an increase in molecular size and chemical shift anisotropy) [29,30] and falling at 1.58 ppm was assigned to the two phosphorous atoms of coordinated pyrophosphate ligand. This chemical shift is downfield if compared to that of free pyrophosphate (−6.35 ppm) measured in the same experimental conditions. The presence of a single peak in the 31 P NMR supports the formation of a symmetric complex with both phosphate groups coordinated to Pt through one deprotonated oxygen atom. The 195 Pt NMR of 1 dissolved in D 2 O is reported in Figure 2d, and shows a single peak falling at -1643.31 ppm. This signal is in the range typical for a Pt(II) atom in an O 2 N 2 coordination environment and is similar to those found in [Pt(dihydrogenpyrophosphato)(1R,2R-DACH)] [21,22] and [Pt(dihydrogenpyrophosphato)(cis-1,4-DACH)] [24].
Another interesting ligand that we have recently investigated is the trans-1,2-diamino-4-cyclohexene (1,2-DACHEX) [26,31]. With respect to 1R,2R-DACH, this ligand contains a double bond in the cyclohexane ring that greatly reduces its flexibility while extending its planarity. Similar to 1,2-DACH, 1,2-DACHEX can also be present in two asymmetric forms: 1R,2R and 1S,2S; however, due to the smaller puckering of 1,2-DACHEX with respect to the 1,2-DACH ring, we expect a smaller dependence of the overall stereochemistry of the complex upon the chirality of the diamine and used the racemic mixture. The 1,2-DACHEX analogue of oxaliplatin, [Pt(OXA)(1,2-DACHEX)] (OXA = oxalate), previously investigated, was able to overcome the oxaliplatin resistance in colorectal cancer cells and to induce the formation of lipid droplets in this cell line, thus suggesting a mechanism of action which includes severe metabolic stress. These results prompted us to prepare the 1,2-DACHEX analog of 3, [Pt(dihydrogenpyrophosphato)(1,2-DACHEX)] (2). Compound 2 was obtained by a one-pot synthesis in water, starting from [Pt(CO 3 )(DMSO) 2 ] (this latter was obtained by reacting [PtCl 2 (DMSO) 2 ] with Ag 2 CO 3 ). The carbonate precursor was treated, in water, with a slight defect of sodium pyrophosphate and then with 1 equivalent (0.121 mmol) of 1,2-DACHEX at pH = 8.0 (Supplementary Materials, Scheme S1). The precipitation of the desired product 2 was achieved by first lowering the pH of the mother solution to 1.0 with nitric acid and then adding acetone (the bis-protonated neutral form of 2 is insoluble in water/acetone). Another interesting ligand that we have recently investigated is the trans-1,2-diamino-4-cyclohexene (1,2-DACHEX) [26,31]. With respect to 1R,2R-DACH, this ligand contains a double bond in the cyclohexane ring that greatly reduces its flexibility while extending its planarity. Similar to 1,2-DACH, 1,2-DACHEX can also be present in two asymmetric forms: 1R,2R and 1S,2S; however, due to the smaller puckering of 1,2-DACHEX with respect to the 1,2-DACH ring, we expect a smaller dependence of the overall stereochemistry of the complex upon the chirality of the diamine and used the racemic mixture.     With the aim of investigating the mechanism of formation of complex 2, we recorded the 31 P NMR spectra of the reaction mixture in H2O/D2O (90:10) at different time intervals (Figure 4). The 31 P spectrum recorded after 2 h of reaction showed, apart from the signal of free pyrophosphate (marked with a full diamond in Figure 4), two doublets falling at −6.42 and −3.65 ppm that were confidently assigned to the two phosphorous atoms of a mono-coordinated pyrophosphate in the intermediate complex 2a (marked with a full star and a full circle in Figure 4). The coordination environment of this intermediate complex is completed by one DMSO ligand still bound to the Pt atom. After 2 h, we could also observe a small signal falling at 1.61 ppm (marked with a full triangle in Figure 3) belonging to complex 2. The 31 P NMR spectra recorded after 4 and 7 h revealed a decrease in In the [ 1 H, 13 C] HSQC spectrum (Figure 3b), the cross-peaks ( 13 C/ 1 H) observed at 127.76/5.45 (A) and 62.43/2.72 (B) ppm were assigned to C 4/5 /H 4/5 and C 1/2 /H 1/2 , respectively. Finally, the cross-peaks ( 13 C/ 1 H) falling at 35.43/2.56 (C) and 35.43/2.16 (D) ppm allowed the assignment of C 3/6 . The 31 P NMR of 2 in D 2 O (Figure 3c) exhibits a single peak with unresolved platinum satellites falling at 1.79 ppm and assigned to the phosphorous atoms of the coordinated pyrophosphate ligand. Additionally, in this case, the presence of a single peak in the 31 P NMR is evidence in support of the formation of a symmetric complex with both phosphate groups coordinated to Pt through one oxygen atom. The 195 Pt NMR of 2 in D 2 O (pH* = 9) is reported in Figure 3d. A single peak falling at −1766 ppm is in agreement with a Pt(II) atom in a O 2 N 2 coordination environment, as also found for complexes 1, 3 [21,22], and 4 [24].
With the aim of investigating the mechanism of formation of complex 2, we recorded the 31 P NMR spectra of the reaction mixture in H 2 O/D 2 O (90:10) at different time intervals ( Figure 4). The 31 P spectrum recorded after 2 h of reaction showed, apart from the signal of free pyrophosphate (marked with a full diamond in Figure 4), two doublets falling at −6.42 and −3.65 ppm that were confidently assigned to the two phosphorous atoms of a mono-coordinated pyrophosphate in the intermediate complex 2a (marked with a full star and a full circle in Figure 4). The coordination environment of this intermediate complex is completed by one DMSO ligand still bound to the Pt atom. After 2 h, we could also observe a small signal falling at 1.61 ppm (marked with a full triangle in Figure 3) belonging to complex 2. The 31 P NMR spectra recorded after 4 and 7 h revealed a decrease in intensity of the signal of the free pyrophosphate with the simultaneous increase of the signal at 1.61 ppm. The latter, after 18 h reaction, represents the only signal accounting for the total phosphorus and indicates the end of the reaction.  The synthetic procedure used for the synthesis of compound 2 could be applied, with some advantages, also to the synthesis of phosphaplatin 3, [Pt(dihydrogenpyrophosphato)(1R,2R-DACH)], firstly reported by Bose et al. [21] (Supplementary Materials, Scheme S2). Differently from Bose, we could prepare complex 3 from the carbonate precursor [Pt(CO3)(1R,2R-DACH)], that can be easily prepared in water solution in high yield and is very reactive towards more strongly coordinating ligands such as pyrophosphate. Thus, by using the carbonate precursor, which, in addition, is more soluble in water than the [PtCl2(1R,2R-DACH)] precursor used in Bose's procedure, we could reduce the reaction time and increase the yield.

NMR Experiments at Different pH Values and Calculation of pKa Values
Pyrophosphate, being a tetrabasic acid, has four ionization equilibria with pKa values corresponding to pKa1 = 0.85, pKa2 = 1.49, pKa3 = 5.77, and pKa4 = 8.22 [32]. The pH titration of complexes 1 and 2 was carried on by recording 31 P NMR spectra at different pH* values ( Figure 5). The titration curves plotting the 31 P chemical shift as a function of pH* exhibit changes in the slope (inflection points) corresponding to the various deprotonation steps (Scheme 2). As the pH* of the samples was increased from ca. 1.0 to 2.5, the 31 P resonance The synthetic procedure used for the synthesis of compound 2 could be applied, with some advantages, also to the synthesis of phosphaplatin 3, [Pt(dihydrogenpyrophosphato) (1R,2R-DACH)], firstly reported by Bose et al. [21] (Supplementary Materials, Scheme S2). Differently from Bose, we could prepare complex 3 from the carbonate precursor [Pt(CO 3 )(1R,2R-DACH)], that can be easily prepared in water solution in high yield and is very reactive towards more strongly coordinating ligands such as pyrophosphate. Thus, by using the carbonate precursor, which, in addition, is more soluble in water than the [PtCl 2 (1R,2R-DACH)] precursor used in Bose's procedure, we could reduce the reaction time and increase the yield.

NMR Experiments at Different pH Values and Calculation of pKa Values
Pyrophosphate, being a tetrabasic acid, has four ionization equilibria with pK a values corresponding to pK a1 = 0.85, pK a2 = 1.49, pK a3 = 5.77, and pK a4 = 8.22 [32]. The pH titration of complexes 1 and 2 was carried on by recording 31 P NMR spectra at different pH* values ( Figure 5). The titration curves plotting the 31 P chemical shift as a function of pH* exhibit changes in the slope (inflection points) corresponding to the various deprotonation steps (Scheme 2). As the pH* of the samples was increased from ca. 1.0 to 2.5, the 31 P resonance shifts to lower field with an inflection point (calculated by fitting the points in this pH range to the Henderson-Hasselbach equation) falling at pH* = 2.28 and 2.02 for complexes 1 and 2, respectively. Very likely, these pH* values correspond to the first deprotonation step of 1 and 2 (pK a1 in Scheme 2). By further increasing the pH from 2.5 to 6.0, we observed a further de-shielding of the phosphorous nuclei, with a second inflection point falling at pH* = 4.55 and 4.23 for compounds 1 and 2, respectively. These pH values correspond to the second deprotonation of the coordinated pyrophosphate (pK a2 in Scheme 2). Further increase of pH* to ca. 13 had no significant effect on the 31 P chemical shift. The pKa values obtained for 1 and 2 are comparable to those reported for 3 [21] and 4 [24], for which a similar biphasic profile, corresponding to two protonation/deprotonation steps, was observed (pK a1 = 2.06 and pK a2 = 4.40 for 3 and pK a1 = 2.03 and pK a2 = 4.47 for 4). 2). Further increase of pH* to ca. 13 had no significant effect on the 31 P chemical shift. The pKa values obtained for 1 and 2 are comparable to those reported for 3 [21] and 4 [24], for which a similar biphasic profile, corresponding to two protonation/deprotonation steps, was observed (pKa1 = 2.06 and pKa2 = 4.40 for 3 and pKa1 = 2.03 and pKa2 = 4.47 for 4).

Stability in Physiological Conditions
To investigate the stability of 1 and 2 in physiological-like conditions, their solutions in D2O containing HEPES buffer (50 mM, pH = 7.4) and NaCl (120 mM) were monitored at 37 °C by 31 P NMR spectroscopy. As can be seen from Figure 6a, at pH = 7.4 and 37 °C, compound 1 resulted to be very stable, and only after 672 h (28 days) could a small new signal Scheme 2. Protonation/deprotonation steps for 1 (a) and 2 (b) with corresponding pK a1 and pK a2 .

Stability in Physiological Conditions
To investigate the stability of 1 and 2 in physiological-like conditions, their solutions in D 2 O containing HEPES buffer (50 mM, pH = 7.4) and NaCl (120 mM) were monitored at 37 • C by 31 P NMR spectroscopy. As can be seen from Figure 6a, at pH = 7.4 and 37 • C, compound 1 resulted to be very stable, and only after 672 h (28 days) could a small new signal falling at −6.54 ppm and assignable to free pyrophosphate be observed. This result is in line with the data reported in the literature for the analogous complexes 3 [21] and 4 [24], which also were quite stable but not to the extent of compound 1. Compound 2 was significantly less stable than 1 and kept at pH = 7.4 and 37 • C, showed the appearance of a small 31 P NMR signal falling at -6.54 ppm, and corresponding to free pyrophosphate, already after 120 h (Figure 6b). However, overall, also compound 2 can be considered stable for the time necessary to perform the in vitro cytotoxicity tests (72 h, see following section). falling at −6.54 ppm and assignable to free pyrophosphate be observed. This result is in line with the data reported in the literature for the analogous complexes 3 [21] and 4 [24], which also were quite stable but not to the extent of compound 1. Compound 2 was significantly less stable than 1 and kept at pH = 7.4 and 37 °C, showed the appearance of a small 31 P NMR signal falling at -6.54 ppm, and corresponding to free pyrophosphate, already after 120 h ( Figure 6b). However, overall, also compound 2 can be considered stable for the time necessary to perform the in vitro cytotoxicity tests (72 h, see following section).

Stability in Acidic Environment
To simulate the acidic environment surrounding cancer tissues, the stability of 1 and 2 was also investigated at pH* = 6.5 and 37 °C by 31 P NMR spectroscopy. As can be seen from Figure 7a, compound 1 resulted to be still stable, and only after 168 h could a small 31 P NMR signal falling at −6.91 ppm and corresponding to free pyrophosphate be observed. Similar to what was observed at pH* = 7.4, also at acid pH, complex 2 resulted to be slightly less stable than complex 1, with the appearance of a small peak of free pyrophosphate already after 96 h (Figure 7b). The decrease in stability of both Pt-pyrophosphate complexes at acid pH suggests that they can release the pyrophosphate ligand, and form the pharmacologically active species, preferentially at the more acidic tumor site than in the blood stream or in the healthy tissues (a drug targeting and delivery advantage).

Stability in Acidic Environment
To simulate the acidic environment surrounding cancer tissues, the stability of 1 and 2 was also investigated at pH* = 6.5 and 37 • C by 31 P NMR spectroscopy. As can be seen from Figure 7a, compound 1 resulted to be still stable, and only after 168 h could a small 31 P NMR signal falling at −6.91 ppm and corresponding to free pyrophosphate be observed. Similar to what was observed at pH* = 7.4, also at acid pH, complex 2 resulted to be slightly less stable than complex 1, with the appearance of a small peak of free pyrophosphate already after 96 h (Figure 7b). The decrease in stability of both Pt-pyrophosphate complexes at acid pH suggests that they can release the pyrophosphate ligand, and form the pharmacologically active species, preferentially at the more acidic tumor site than in the blood stream or in the healthy tissues (a drug targeting and delivery advantage).
from Figure 7a, compound 1 resulted to be still stable, and only after 168 h could a small 31 P NMR signal falling at −6.91 ppm and corresponding to free pyrophosphate be observed. Similar to what was observed at pH* = 7.4, also at acid pH, complex 2 resulted to be slightly less stable than complex 1, with the appearance of a small peak of free pyrophosphate already after 96 h (Figure 7b). The decrease in stability of both Pt-pyrophosphate complexes at acid pH suggests that they can release the pyrophosphate ligand, and form the pharmacologically active species, preferentially at the more acidic tumor site than in the blood stream or in the healthy tissues (a drug targeting and delivery advantage).

Reaction with 5'-GMP
A peculiar feature of phosphaplatins is that, unlike cisplatin, their pharmacological activity could not depend upon reaction with DNA [23]. However, in a previous work, the pyrophosphate derivative 4 was shown to be able to bind DNA [25], and therefore it could be worth also investigating the reaction of compounds 1 and 2 with DNA. As a preliminary step, we monitored by 31 P and 1 H NMR the reaction of 1 and 2 with 5'-GMP (2.5 equivalents, 0.010 mmol) in 50 mM HEPES buffer (pH* = 7.4) and 120 mM NaCl at 37 • C. The 31 P spectrum, recorded soon after mixing of the reactants, showed only the peaks of free 5'-GMP (4.20 ppm) and of the starting substrates: 1.74 ppm for complex 1 (Figure 8) and 1.91 ppm for complex 2 (Figure 9a). In both cases, with time, it was possible to observe the appearance of a new peak at −6.49 ppm assigned to free pyrophosphate and a decrease in intensity of the signal belonging to the starting substrates. However, only in the case of compound 1 was a new peak at 4.43 ppm ( Figure 8) clearly evident, which could be confidently assigned to the bis-adduct [Pt(5'-GMP) 2 (cis-1,3-DACH)].
We argued that the absence of a separated signal downfield of that of free 5'-GMP (which could indicate the formation of the bis-adduct [Pt(5'-GMP) 2 (1,2-DACHEX)]) in the case of compound 2 could be due to the slight broadness of these signals, which could mask the separation between free and coordinated 5'-GMP (Figure 9a). Therefore, in the case of 2, the 31 P NMR spectrum was also recorded on a 500 MHz NMR instrument; as expected, a downfield peak falling at 4.27 ppm was observed (marked with an arrow in Figure 9b) that could be assigned to the coordinated 5'-GMP in the bis-adduct. The reaction of 2 with 5'-GMP was also monitored by 1 H NMR spectroscopy ( Figure 10). The spectrum recorded at time 0 showed the signals of the free nucleotide (e.g., the sharp singlet at 8.23 ppm belonging to proton H8 of guanine and marked with a full diamond in Figure 10) and the signals of the starting substrate (e.g., the vinylic protons of coordinated trans-1,2-diamine-4-cyclohexene marked with a full star in Figure 10). After 336 h, 3 new signals falling at 5.54, 8.66, and 8.80 ppm could be observed. The peak at 5.54 ppm (marked with a full triangle in Figure 10) is assigned to the vinylic protons of DACHEX in the bis-adduct [Pt(5'-GMP) 2 (1,2-DACHEX)], while the signals at 8.66 and 8.80 ppm (marked with a full circle in Figure 10) are assigned to the H8 of the two coordinated 5'-GMPs, which are not equivalent due to the absence of a plane of symmetry passing through the cyclohexene ring. The three new signals grew in intensity with time, while the signals associated to complex 2 decreased. of free 5'-GMP (4.20 ppm) and of the starting substrates: 1.74 ppm for complex 1 8) and 1.91 ppm for complex 2 (Figure 9a). In both cases, with time, it was pos observe the appearance of a new peak at −6.49 ppm assigned to free pyrophosph a decrease in intensity of the signal belonging to the starting substrates. However, the case of compound 1 was a new peak at 4.43 ppm (Figure 8) clearly evident, could be confidently assigned to the bis-adduct [Pt(5'-GMP)2(cis-1,3-DACH)]. We argued that the absence of a separated signal downfield of that of free 5 (which could indicate the formation of the bis-adduct [Pt(5'-GMP)2(1,2-DACHEX)] case of compound 2 could be due to the slight broadness of these signals, which mask the separation between free and coordinated 5'-GMP (Figure 9a). Therefore case of 2, the 31 P NMR spectrum was also recorded on a 500 MHz NMR instrum expected, a downfield peak falling at 4.27 ppm was observed (marked with an ar Figure 9b) that could be assigned to the coordinated 5'-GMP in the bis-adduct. Th tion of 2 with 5'-GMP was also monitored by 1 H NMR spectroscopy ( Figure 10). Th trum recorded at time 0 showed the signals of the free nucleotide (e.g., the sharp at 8.23 ppm belonging to proton H8 of guanine and marked with a full diamond in  Figure 10) are assigned to the H8 of the two coordinated 5'-GMPs, which are not equivalent due to the absence of a plane of symmetry passing through the cyclohexene ring. The three new signals grew in intensity with time, while the signals associated to complex 2 decreased.   The reaction of 1 and 2 with 5'-GMP was also performed at lower NaCl concentration (4 mM) in order to simulate the cytoplasmatic chloride concentration. The 31 P NMR spectra recorded at pH* = 7.4 and 4 mM NaCl (see ESI, Supplementary Materials, Figures S1 and S2) appeared to be comparable to those obtained in the presence of 120 mM NaCl for both complexes 1 and 2, indicating that the rate-determining step is the slow solvolysis of the pyrophosphate. These results are in line with the data already reported for 3 (no detectable binding to DNA after 7 days of incubation [20]) and 4 (slow reaction [24]).

Cytotoxicity Assays
The in vitro cytotoxicity of the Pt-pyrophosphato complexes (1, 2, 3, and 4) was evaluated on a panel of human cancer cell lines and compared to that of cisplatin and oxaliplatin. Cell lines representative of colon (HCT-116), prostate (PC3), breast (MDA-MB231), and cervical (OV2008) carcinomas were included. The cell viability was evaluated by means of the MTT test after 72 h of incubation with different concentrations of the tested compounds; the IC50 values, calculated from dose-survival curves, are plotted in Figure  11 and reported in Table S1, in the Supplementary Materials. The reaction of 1 and 2 with 5'-GMP was also performed at lower NaCl concentration (4 mM) in order to simulate the cytoplasmatic chloride concentration. The 31 P NMR spectra recorded at pH* = 7.4 and 4 mM NaCl (see ESI, Supplementary Materials, Figures S1 and S2) appeared to be comparable to those obtained in the presence of 120 mM NaCl for both complexes 1 and 2, indicating that the rate-determining step is the slow solvolysis of the pyrophosphate. These results are in line with the data already reported for 3 (no detectable binding to DNA after 7 days of incubation [20]) and 4 (slow reaction [24]).

Cytotoxicity Assays
The in vitro cytotoxicity of the Pt-pyrophosphato complexes (1, 2, 3, and 4)  As expected, cisplatin was found to be the most effective in all cell lines (average IC50 = 6 μM) due to the faster substitution of the two chlorido ligands if compared to that of the bidentate pyrophosphato ligand. From this point of view, oxaliplatin represents a better reference compound, with the bidentate oxalato ligand being more similar to the bidentate pyrophosphato ligand. Compound 1 (average IC50 = 23 μM) was found to be much more effective than all other Pt-pyrophosphato complexes (2, 3, and 4) and oxaliplatin (average IC50 = 80 μM for OXP), and to possess a discrete activity also toward the prostate cancer cell line PC3 (IC50 = 29 μM, up to 3 times greater than that of the other complexes all of which have IC50 = 100 μM). Compound 2 follows, which is more active than 3, 4, and oxaliplatin in all cell lines except PC3. Finally, compound 3 was more active than 4 and oxaliplatin in HCT-116 and OV2008 cell lines ( Figure 11).

Apoptosis Assay
In order to obtain information on tumor cells' death, the Annexin V/propidium iodide (PI) assay was used in flow cytometry analyses. Annexin V provides a very sensitive method for detecting cellular apoptosis, while PI is commonly used to detect necrotic or late apoptotic cells (when positive to Annexin too) [33]. The FITC Annexin V staining was carried out after 24 and 48 h of incubation of HCT116 cells with the Pt-pyrophosphato complexes (1)(2)(3)(4) or oxaliplatin/cisplatin (used as reference compounds) at the IC50 concentrations (Supplementary Materials, Table S1). The Flow Cytometry (FCM) dot plots of Annexin V/PI-positive HCT116 cells are reported in Figure 12. The two-dimensional dot plots can be divided into four quadrants, corresponding to: (i) viable cells, which are negative to both probes As expected, cisplatin was found to be the most effective in all cell lines (average IC 50 = 6 µM) due to the faster substitution of the two chlorido ligands if compared to that of the bidentate pyrophosphato ligand. From this point of view, oxaliplatin represents a better reference compound, with the bidentate oxalato ligand being more similar to the bidentate pyrophosphato ligand. Compound 1 (average IC 50 = 23 µM) was found to be much more effective than all other Pt-pyrophosphato complexes (2, 3, and 4) and oxaliplatin (average IC 50 = 80 µM for OXP), and to possess a discrete activity also toward the prostate cancer cell line PC3 (IC 50 = 29 µM, up to 3 times greater than that of the other complexes all of which have IC 50 = 100 µM). Compound 2 follows, which is more active than 3, 4, and oxaliplatin in all cell lines except PC3. Finally, compound 3 was more active than 4 and oxaliplatin in HCT-116 and OV2008 cell lines ( Figure 11).

Apoptosis Assay
In order to obtain information on tumor cells' death, the Annexin V/propidium iodide (PI) assay was used in flow cytometry analyses. Annexin V provides a very sensitive method for detecting cellular apoptosis, while PI is commonly used to detect necrotic or late apoptotic cells (when positive to Annexin too) [33]. The FITC Annexin V staining was carried out after 24 and 48 h of incubation of HCT116 cells with the Pt-pyrophosphato complexes (1)(2)(3)(4) or oxaliplatin/cisplatin (used as reference compounds) at the IC 50 concentrations (Supplementary Materials, Table S1). The Flow Cytometry (FCM) dot plots of Annexin V/PI-positive HCT116 cells are reported in Figure 12. The two-dimensional dot plots can be divided into four quadrants, corresponding to: (i) viable cells, which are nega-  The data reported in Figure 12 show that, after 24 h incubation, apoptosis was more evident with 1 and 4 than with 3 and 2. Moreover, only 1 and 4 induced necrosis. All compounds induced apoptosis to a lesser extent than oxaliplatin, but it must be taken into account that the IC50 concentration of the latter drug (72 μM) was much higher than that of the other compounds (see Supplementary Materials, Table S1). After 48 h of incubation, the induction of apoptosis increased in all treated cells, while necrosis increased mainly in the case of 1 and 4.

Reagents and Equipment
Commercial reagent-grade chemicals and solvents (Sigma, Milan, Italy) were used as received without further purification. 1 H-, 31 P-, and 195 Pt-NMR and [ 1 H-13 C]-HSQC spectra were recorded on Bruker Avance DPX 300 MHz (Bruker Italia S.r.L., Milano, Italy) and Agilent Technologies 500 MHz (Agilent Technologies Italia S.p.A.; Cernusco sul Naviglio, Milano, Italy) instruments. Standard pulse sequences were used for 1 H, 31 P{ 1 H}, and 195 Pt{ 1 H} (121.5, 242.9, and 64.5 MHz, respectively) one-dimensional spectra. 1 H chemical shifts were referenced using the internal residual peak of the solvent (D2O: 4.80 ppm), 31   The data reported in Figure 12 show that, after 24 h incubation, apoptosis was more evident with 1 and 4 than with 3 and 2. Moreover, only 1 and 4 induced necrosis. All compounds induced apoptosis to a lesser extent than oxaliplatin, but it must be taken into account that the IC 50 concentration of the latter drug (72 µM) was much higher than that of the other compounds (see Supplementary Materials, Table S1). After 48 h of incubation, the induction of apoptosis increased in all treated cells, while necrosis increased mainly in the case of 1 and 4.
J. Peters-Cancer Center Amsterdam. All necessary cell culture material was purchased from EuroClone (Pero, Milan, Italy). The MTT used for cytotoxicity studies was purchased from Sigma-Aldrich (Milan, Italy). The FITC Annexin V Apoptosis Detection Kit I was from BD PharmingenTM (San Jose, CA, USA, 556547). The complex [Pt(CO 3 )(DMSO) 2 ] was prepared according to an already reported procedure [35]. The characterization of the compound provided data in accordance with those reported in the literature.

[PtI 2 (1R,2R-DACH)]
This complex was prepared following Dhara's procedure [36].   31  This complex was prepared by a modification of an already reported procedure [21]. Briefly, tetrasodium pyrophosphate decahydrate (39.25 mg, 0.088 mmol) was dissolved in 5.0 mL of water at 55 • C and the pH of the solution was adjusted to 8.0 with 1.0 M HNO 3 . The solution was kept under magnetic stirring at 55 • C for 15 min and then treated with [Pt(CO 3 )(1R,2R-DACH)] (50 mg, 0.135 mmol). The reaction mixture was stirred at 55 • C for 3.5 h and then concentrated to a minimum volume (ca. 2 mL) by rotary evaporation. After lowering the pH of the concentrated solution to ca. 1-2 by addition of HNO 3 (1.0 M), acetone (100 mL) was added to induce the precipitation of the desired product as a yellow solid that was separated by filtration of the mother liquor and dried under vacuum. Obtained: 35.2 mg (82% yield). The characterization of 3 provided data in accordance with those reported in the literature [31].

[Pt(Dihydrogenpyrophosphato)(cis-1,4-DACH)] (4)
This complex was prepared according to an already reported procedure and its characterization provided data in accordance with those reported in the literature [24].

NMR Experiments at Different pHs and Calculation of pKa Values
Acidity constants for complexes 1 and 2 were determined by 31 P NMR chemical shift/pH titrations. Either compound 1 or 2 (0.0047 mmol) was dissolved in 0.550 mL of D 2 O and transferred into an NMR tube. The pH* of the sample was adjusted to the required value by addition of DClO 4 (10, 1.0, 0.5, 0.1, or 0.05 M) or NaOD (1.0, 0.5, 0.1, or 0.05 M) solutions, and the pH* value was measured by using a 3 mm diameter electrode for NMR tubes. No control of the ionic strength was performed. The pH titration curves were fitted to the Henderson−Hasselbalch equation using the program KaleidaGraph (version 3.5, Synergy Software, Reading, PA, USA) [37].

Stability in Physiological or Acidic Conditions
The stability of compounds 1 and 2 at 37 • C in buffered solutions was assessed by 31 P NMR spectroscopy. Each compound (~2 mg) was dissolved in 0.8 mL of D 2 O containing 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES) buffer (pH* = 7.4 or pH* = 6.5) and 120 mM NaCl. The resulting solutions were transferred into NMR tubes and maintained at 37 • C. 31 P NMR spectra were recorded from time to time over a period of one month.

Reactivity with 5 -GMP
Complex 1 (2 mg, 0.0041 mmol) or 2 (2 mg, 0.0041 mmol) was dissolved in 0.8 mL of D 2 O containing 50 mM HEPES buffer (pH* = 7.4) and 120 mM (to simulate the extracellular environment) or 4.0 mM (to simulate the intracellular environment) of NaCl. After addition of 4.1 mg (0.010 mmol) of guanosine 5 -monophosphate sodium salt (5 -GMP-Na·6H 2 O), the reaction mixture was transferred into an NMR tube that was maintained at 37 • C. The reactions were monitored up to one month by recording 1 H and 31 P NMR spectra at defined time intervals.

Cell Viability Assay
The number of living cells was evaluated by the 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay. Briefly, HCT116 human colon cancer cells, PC3 human prostate cancer cells, OV2008 human ovarian cancer cells, and MDA MB 231 human breast cancer cells were plated in 96-well plates at a density of~5000 cells/well. After one day of incubation at 37 • C in a humid atmosphere with 5% CO 2 , the culture medium was replaced with 100 µL of fresh medium (control cells) or medium containing different concentrations of the test compounds in the range 1.56-100 µM. Untreated cells were used as positive controls. After the incubation period of 72 h, 10 µL of a 0.5% (w/v) MTT/PBS solution was added to each well and the incubation was prolonged for a further 2 h. After this period, the medium was removed and replaced with 100 µL of DMSO. The absorbance of each well was measured by a microplate reader (Wallac Victor3, 1420 Multilabel Counter, Perkin-Elmer, Singapore). Each compound concentration was tested in triplicate, and results were presented as percentage of the control value. IC 50 values were calculated using non-linear regression in GraphPad Prism 5.01 (San Diego, CA, USA).

Annexin V/PI Apoptosis Assay
The Annexin V/PI assay was conducted as previously described [33] in order to investigate the mechanism of cell death induced by the various compounds on HCT116 cells. In particular, after 24 and 48 h of treatment with the compounds at concentrations equal to their IC 50 values calculated by the MTT assay previously described, cells were subjected to Annexin-V FITC/propidium iodide staining, which allowed detection of early and late apoptosis as Annexin V-positive cells, late apoptosis as Annexin V/PI-positive cells, and necrotic cells as PI-positive cells. The FITC Annexin V Apoptosis Detection Kit I was from BD PharmingenTM (San Jose, CA, USA, 556547) and the analysis was performed on an Attune NxT acoustic focusing cytometer (Life Technologies, Thermo Fischer Scientific, Waltham, MA, USA). Data were interpreted using the CytExpert software v. 1.2 provided by the manufacturer.

Conclusions
Two new Pt(II) derivatives (1 and 2) with a pyrophosphato ligand, that differ from those already reported in the literature (3 and 4) for the diamine carrier ligand, have been synthesized and fully characterized with the aim of investigating how slight modifications in the carrier ligand can impact the antitumor activity of this class of drugs characterized by high bone tropism. Compounds 1 and 2 were synthesized, starting from the carbonateprecursor [Pt(CO 3 )(diamine)]. This procedure could also be applied to the synthesis of 3 with better results. 31 P{ 1 H} NMR experiments at different pH values have allowed the determination of the acidity constants of complexes 1 and 2 and an estimation of their stability in physiological solutions. The complexes are more stable in physiological conditions than in acidic conditions (characteristic of the tumor sites) and compound 1 appeared to be slightly less reactive than compound 2.
NMR studies on the reactivity of 1 and 2 towards 5'-GMP (used as a model of nucleic acid) showed that both compounds react slowly with 5'-GMP, forming the bis-adduct, and that the rate does not appear to depend upon the NaCl concentration (120 mM to mimic the extracellular chloride concentration and 4 mM to mimic the intracellular concentration). Finally, in a preliminary pharmacological investigation, the in vitro cytotoxicity assays, performed against a panel of four human cancer cell lines, showed an activity order of 1 > 2 > 3 > 4 > oxaliplatin. Cisplatin, that resulted to be the most active of all compounds, cannot be taken as a reference since it contains monodentate chlorido ligands which are known to be much more hydrolysable than the bidentate pyrophosphate and oxalate. Interestingly, the cis-1,3-DACH diamine ligand appears to confer the best cytotoxic effect. Flow cytometry studies performed to investigate the mechanism of cell death induced by compounds on HCT116 cells showed that the Pt(II)-pyrophosphato compounds with the non-classical 1,3-and 1,4-diaminocyclohexane ligands (1 and 4) are the most capable to induce cell death by apoptosis. A future perspective will be a deeper pharmacological investigation of the new Pt(II)-pyrophosphato derivatives, focusing on their potential selectivity towards bone tumors as well as their embodiment in hydroxyapatite matrices to be used for the local treatment of bone tumors after implantation.
Supplementary Materials: The following are available online. Scheme S1: synthesis of compound 2. Scheme S2: synthesis of compound 3. Figure S1: 31 P NMR spectra at different time intervals of 1 in the presence of 5'-GMP. Figure S2

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.