2.1. Synthesis and Characterization of Ligands
Ligands
2–
4 were synthesized following the synthetic route shown in
Scheme 2 from 1,10-phenanthroline-5,6-dione (
1). The decarbonylation of
1 under basic conditions affords 4,5-diazafluoren-9-one (
3) [
25,
26]. Both compounds (
1,
3) have ketone groups in their structure that can undergo aldol addition reactions. Compounds
1 and
3 react with acetone in the presence of alumina as a catalyst [
27]. The conversion of
1 into
2 is nearly quantitative, whereas
3 and
4 reach an equilibrium. The first conversion can be understood in terms of product solubility as compound
2 is poorly soluble in acetone and shifts the reaction to the right. Compound
4 has a greater solubility in acetone and it does not precipitate in reaction media. In addition, in
1, the two adjacent ketone groups can act as an electron-withdrawing group to each other, thus, they might be more electrophilic than the carbonyl group in
3.
Ligand
3 was characterized by
1H-NMR [
26]. Ligands
2 and
4 were characterized by NMR (
1H and
13C{
1H}), ESI-TOF mass spectrometry, and additionally, their structure has been confirmed by single-crystal X-ray diffraction (see
Figures S32 and S33, Supplementary Materials). The
1H-NMR of compound
2 confirms that the symmetry of the molecule is lost, and, consequently, all the hydrogens of the two pyridine moieties are inequivalent and the hydrogens of the methylene group are diastereotopic (see
Figure S3, Supplementary Materials). The aldol reaction on
1 generates a chiral center, thus, compound
2 is obtained as a racemic mixture. In the case of
4, the hydrogen signals of the bipyridine moiety are equivalent, as well as the hydrogens of the methylene group (see
Figure S6, Supplementary Materials).
2.2. Synthesis and Characterization of Ag(I) Complexes
The reaction of the compounds
1–
4 with AgNO
3 in a 2:1 mol ratio yielded complexes
1Ag–
4Ag (
Scheme 3). All the complexes have been characterized by ESI-TOF mass spectrometry and NMR spectroscopy (see
Section 3 Materials and Methods).
The MS data confirm the proposed [Ag(Ligand)
2]
+ stoichiometry. The presence of the two isotopes of nearly equal abundance of silver,
107Ag (51.8%) and
109Ag (48.2%), are observed in the MS spectra of the complexes. The characteristic isotopic peak doublet with almost equal intensity in the ESI-TOF mass spectra of the complexes
1Ag–
4Ag agrees with the proposed mononuclear species (see
Figure S28, Supplementary Materials for ESI-TOF mass spectrum of complex
2Ag).
The
1H-NMR spectra for the silver complexes
1Ag–
4Ag are different from those of the free ligands.
Figure 3 shows the
1H-NMR spectra of both
4 and the corresponding complex
4Ag registered in DMSO-
d6. The chemical shifts of the bipyridine hydrogens of the ligands appear deshielded after coordination to the metal center, especially H
b and H
c. This change in the chemical shift is also observed in the hydrogen of the hydroxyl group of
2Ag and
4Ag. The NMR data for
1Ag and
3Ag agree with those previously reported for [Ag(1,10-phenanthroline-5,6-dione)
2]PF
6 [
22] and [Ag(4,5-diazafluoren-9-one)
2]NO
3 [
28], respectively.
Given that in the synthesis of complex 2Ag, the starting material ligand 2 is a racemate, the product obtained is a mixture of diastereoisomers since each enantiomer of 2 has the same capability of coordinating to silver. The distance between the aromatic hydrogens and the chiral center does not allow for differentiation of the signals corresponding to the different expected diastereoisomers (25% of the complex (R,R), 25% of its enantiomer (S,S), and 50% of the (R,S) meso complex).
The X-ray diffraction structure of
1Ag (
Figure 4) confirms the expected coordination of two units of the ligand.
The four Ag-N bond lengths (Å) for
1Ag are listed in
Table 4. Although the crystal structure of the analogous [Ag(1,10-phenanthroline-5,6-dione)
2]PF
6 was already reported [
22], in our case, the asymmetric unit of
1Ag also has one molecule of H
2O and one of MeOH (
Figure 4; for structural details see
Figure S43 and Table S3, Supplementary Materials). In both cases, the cationic complexes display a distorted tetrahedral geometry, and similar bond lengths and angles have been found.
Two different polymorphs were obtained for
3Ag. In crystal structure A depicted in
Figure 5) the NO
3− acts as a ligand and binds to the metal center through one of the oxygens, and the adopted geometry is distorted square pyramidal. In addition, the asymmetrical unit has two slightly different molecules (see
Figure S45 and Table S4, Supplementary Materials), although only one is depicted in
Figure 5 for clarity purposes.
In crystal structure B (
Figure 5, right), the NO
3− is coordinated to silver in a distorted trigonal bipyramidal structural environment. A similar structure to the latter has already been reported [
29]. Selected bond lengths (Å) and angles (deg) for the two different polymorphs are collected in
Table 5.
2.3. Synthesis and Characterization of Pt(II) Complexes
The reaction of the compounds
1–
4 with [Pt(C
6H
5)
2(THF)
2] in a 1:1 mol ratio yielded complexes
1Pt–
4Pt (
Scheme 4). All the complexes have been characterized by ESI-TOF mass spectrometry and NMR spectroscopy (see
Section 3, Materials and Methods).
The
1H-NMR spectra for the Pt(II) complexes confirm the coordination of the ligands to the metal center.
Figure 6 shows the
1H-NMR spectra of
1 (spectrum a) and the corresponding complex
1Pt (spectrum b). The chemical shifts of the hydrogens H
b and H
c of the ligands appear deshielded after coordination, whereas for H
a its chemical shift decreases. The X-ray structures of
1Pt and
3Pt (
Figure 7) allow us to justify this fact since the H
a hydrogens point toward the center of the pentafluorophenyl rings, which causes a shielding of the signal. Of particular interest are the
1H-NMR spectra of complexes
1Pt and
2Pt (see
Figure 6, spectra b and c), in which the hydrogens closest to the nitrogens H
a (and H
a* for
2Pt) have signals with a broad base, which correspond to the
195Pt satellites, confirming the coordination through the N atoms.
The
19F-NMR spectra give information about the structural symmetry of the
N,N-ligand. In complex
1Pt, the two pentafluorophenyl rings are equivalent, and a set of signals is observed that corresponds to the F
o, F
m, and F
p (
Figure 7, spectrum a). The four F in the
ortho position are the most deshielded ones, their multiplicity is a multiplet and the characteristic
195Pt satellites are observed. The four F in the
meta position correspond to the most shielded signal, multiplets with a broad base due to the
195Pt satellites. The two F in the
para position give a triplet and no
195Pt satellite is observed. In complex
2Pt, the asymmetry of the
N,N-ligand makes the two C
6F
5 inequivalent, and, consequently, two sets of signals are observed—one for each aryl group (
Figure 7, spectrum b). Thus, the F
p gives two different signals, two triplets with close chemical shifts. The signals of F
o are overlapped and the same is observed for the F
m. The complex
4Pt 19F-NMR spectrum shows a different scenario (
Figure 7, spectrum c): the two pentafluorophenyl rings are equivalent, which justifies the fact that only one signal is observed in
19F-NMR for the F
p. However, as the coordination plane is not a plane of symmetry, the two F
o and the two F
m of each fluorinated aryl are chemically inequivalent, giving rise to two signals.
Regarding chirality, it is necessary to highlight that complex 2Pt is a racemate.
The X-ray diffraction structures of
1Pt and
3Pt are depicted in
Figure 8, left and right, respectively. Selected bond lengths (Å) and angles (deg) are collected in
Table 6. In both structures, the fluorinated rings are almost perpendicular to the coordination plane, with angles of 69.84(16)° and 64.26(14)° for
1Pt and 58.05(15)° and 75.28(16)° for
3Pt, forcing the aromatic hydrogen H
a of the ligand to point toward the center of the fluorinated rings.
In both cases, Pt-N and Pt-Cipso distances are similar to those found for [PtPf
2(bipy)] [
30].
In
1Pt, Pt-N distances are longer than those found for [PtCl
2(1,10-phenanthroline-5,6-dione)] as a consequence of the higher trans influence of the C
6F
5 group than the Cl ligand [
31]. Elongation of the distances is also observed in
3Pt compared to those observed for [PtCl
2(4,5-diazafluoren-9-one)] [
32].
2.4. Stability of the Ag(I) and Pt(II) Complexes in DMSO
The solution behavior of the Ag(I) and Pt(II) complexes was studied by
1H-NMR spectroscopy. The complexes were dissolved in DMSO-
d6 in a 5 mM concentration and spectra were recorded immediately, after 24 h, and after 48 h standing in the dark at room temperature. The
1H-NMR spectra of all Ag(I) complexes remained unmodified over 48 h and no free ligand was observed. These complexes are stable in solution and no coordination of DMSO to Ag(I) was observed.
Figures S29 and S30 (Supplementary Materials) show the aromatic region of the
1H-NMR spectra of the free ligand
2 vs.
2Ag and
4 vs.
4Ag, respectively, registered at room temperature in DMSO-
d6 at different times. The Pt(II) complexes
1Pt,
2Pt, and
4Pt are stable in DMSO solution, no free ligand was observed in the experiment conditions.
Figure S31 (Supplementary Materials) shows the
1H-NMR spectra of the free ligand
4 vs
. 4Pt at the tested times. The scenario for complex
3Pt was much more complicated since a solvolysis reaction was observed in the experiment.
The
1H-NMR of
3Pt (see
Figure S32, Supplementary Materials)—registered immediately after dissolving the compound in DMSO-
d6—indicated the presence of two species. After 24 h, a third compound, which corresponded to the free ligand
3, was detected. After 48 h, no new chemical species were observed. The
19F-NMR spectra were crucial to identifying the components of the solvolysis reaction (see
Figure S33, Supplementary Materials), especially the F signal in the
para position. In the
19F-NMR spectra of complex
3Pt immediately after dissolution in DMSO-
d6 (
Figure 9, spectrum a), three triplets were observed in the region of F
p signals. Two of these signals have the same integration values, which indicates they might be signals of the same compound and suggests that this compound is asymmetrical. In the
19F-NMR spectra of the complex 24 h after dissolution in DMSO-
d6 (
Figure 9, spectrum b), a new signal arises that corresponds to [Pt(C
6F
5)
2(DMSO)
2]. The complex [Pt(C
6F
5)
2(DMSO)
2] can be synthesized in situ in the NMR sample tube by dissolving [Pt(C
6F
5)
2(THF)
2] in DMSO-
d6 (
Figure 9, spectrum c). The detection of this compound is in agreement with the detection of free ligand
3 in the
1H-NMR.
The species proposed to be present in the solvolysis mixture of
3Pt are depicted in
Figure 10. The detection of free ligand
3 and [Pt(C
6F
5)
2(DMSO)
2] indicates that DMSO can displace the two coordination positions of ligand
3 in complex
3Pt. The asymmetric species detected should be the product in which DMSO has substituted only one coordination position (structure
3Pt-DMSO,
Figure 10). Lastly, the other species detected immediately after dissolution might correspond to complex
3Pt. After 48 h of dissolution of complex
3Pt 5 mM in DMSO, the concentrations of the species are 0.65 mM for compound
3Pt-DMSO, 1.55 mM for
3Pt, 2.8 mM for ligand
3, and 2.8 mM for [Pt(C
6F
5)
2(DMSO)
2] (calculated by
1H and
19F-NMR signal integration).
2.5. Stability of the Ag(I) and Pt(II) Complexes in DMSO-d6:H2O
To study the stability of the complexes in conditions closer to physiological ones, the complexes were dissolved in DMSO-
d6:H
2O (3:1) in a 2.5 mM concentration, and spectra were recorded immediately, after 24 h, and after 48 h standing in the dark at 37 °C. Complexes
2Ag,
3Ag,
4Ag,
2Pt, and
4Pt remained unmodified over 48 h, and no free ligand was observed.
Figures S34 and S35 (Supplementary Materials) show the aromatic region of the
1H-NMR spectra of the free ligand
4 vs.
4Ag and
2 vs.
2Pt, respectively, under these experimental conditions. Complex
3Pt, as analogously observed in DMSO, experiences solvolysis.
Figures S36 and S37 (Supplementary Materials) show the NMR spectra for this experiment. Under this scenario, at least five species are detected by NMR, which include complex
3Pt, ligand
3, [Pt(C
6F
5)
2(solvent)
2] (solvent: DMSO, and/or H
2O), and the other two species are proposed to be
3Pt-DMSO and
3Pt-H2O (the two possible species in which DMSO or water have substituted only one coordination position).
Complexes
1Ag and
1Pt show a new scenario. Ligand
1 in DMSO:H
2O (3:1) shows two species in the
1H-NMR spectrum (see
Figure 11), whereas in pure DMSO, only one species is detected. This indicates that the ligand reacts with water. The water solution behavior of ligand
1 has already been studied [
27] and it was found that the hydration reaction takes place at a pH near neutrality; thus, we hypothesize that in the DMSO:H
2O mixture, ligand
1 is in equilibrium with its hydrated form
1-H2O (see
Figure 11).
For complex
1Ag, the
1H-NMR spectra for these experimental conditions (see
Figure S38, Supplementary Materials) indicate that both ligand
1 and its hydration form (
1-H2O) are coordinated to Ag, and neither of the two free ligands are observed; thus, equilibrium between complex
1Ag and the hydrated form
1Ag-H2O occurs. The NMR data indicate that
1Ag is the predominant species. For complex
1Pt, the NMR data prove that no decoordination has occurred since neither [Pt(C
6F
5)
2(solvent)
2] nor free ligands
1 or
1-H2O are detected, and that at least three species coexist. The NMR data are complicated and it is not possible to be certain of the structure of these species; however, based on the hydration reaction that takes place for the ligand
1 and complex
1Ag, it is reasonable to propose an equilibrium of
1Pt, its hydrated form
1Pt-H2O, and the double-hydrated form
1Pt-2H2O (
Figure 12). The hydration of a complex with 1,10-phenanthroline-5,6-dione as a ligand has been previously reported for [Ru(phen)
2(1,10-phenanthroline-5,6-dione)]
2+, in fact, the authors indicate that the hydrated form is the biologically active form of the complex [
33].
It is also important to highlight that no hydrolysis (C6F5H) nor reductive elimination product (C12F10) are observed in any of the stability tests in DMSO:H2O of the Pt(II) complexes, indicating that the pentafluorophenyl ligand does not act as a leaving group or it is kinetically inert.
2.6. In Vitro Cytotoxicity
The cytotoxicity of the ligands and the complexes was investigated in vitro by testing the antiproliferative activities against the A-549 cell line (carcinomic human alveolar basal epithelial cells). The cell counting kit (CCK8) assay was employed to assess the growth inhibition, and the cell proliferation inhibitory activities of the compounds are shown in
Figure 13 and
Table 7 as pIC
50 and IC
50 values, respectively. NH
4NO
3 was tested as a negative control for nitrate counterion evaluation. No cytotoxicity was observed (
Table 7, entry 14), which proves that nitrate does not contribute to the activity of silver(I) complexes. Cisplatin was tested as a reference compound.
Figure 13 shows that phenanthroline-type ligands displayed cytotoxic activity against A-549 cells with IC
50 values in the low micromolar range (
Table 7, entries 1 and 2). The aldol addition of acetone to compound
1 does not affect its antiproliferative activity since both compounds
1 and
2 have similar IC
50 values. The 4,5-diazafluorene-type ligands, compounds
3 and
4, are not toxic on A-549 cells (
Table 7, entries 3 and 4). The coordination with silver led to a slight improvement in the activity of complex
1Ag compared to the ligand. The IC
50 values observed for ligand
1 and complex
1Ag are similar to those reported for compound
1 and [Ag(1,10-phenanthroline-5,6-dione)
2]ClO
4 in other cell lines—human kidney adenocarcinoma (A-498) and human hepatocellular carcinoma (Hep-G2) [
34]. Complex
2Ag maintains the same activity as the ligand. Compounds
1Ag and
2Ag are slightly more potent than the structurally similar complex [Ag(phen)
2]AcO (IC
50 = 5.4 μM in A-549) [
35]. In the case of the non-active ligands, their coordination with silver gave life to active compounds, complexes
3Ag and
4Ag, with
3Ag being much more toxic with an IC
50 of 6.25 ± 1.15 μM. Complex
3Ag has already been described as a DNA intercalator [
29], with a DNA affinity bigger than for the free ligand [
28]. With these in hand, we can observe that the coordination with silver(I) led to an improvement or maintenance in activity. Among the silver complexes, complex
1Ag is the most toxic with an IC
50 of 0.67 ± 0.30 μM.
Regarding the Pt complexes, the coordination of phenanthroline-type ligands to Pt—complexes
1Pt and
2Pt—led to complexes less active than the ligands but still more potent than [PtCl
2(phen)] (with an IC
50 in the range of 40–80 μM in A-549 cells) [
36]. The coordination with Pt of the 4,5-diazafluorene-type ligands surprisingly led to active complexes, complexes
3Pt and
4Pt, with IC
50 values below 3 μM. It is important to highlight that all the Pt complexes are active, which reinforces the use of pentafluorophenyl as a viable alternative to classical leaving groups in Pt(II) bioactive complexes. Complex
3Pt is the most toxic Pt complex, with an IC
50 of 0.94 ± 0.71 μM. Still, this IC
50 value should be compared with caution to others because the cells were exposed to the solvolysis product of
3Pt with DMSO and H
2O; thus, it was a mixture of complexes
3Pt,
3Pt-DMSO,
3Pt-H2O, [Pt(C
6F
5)
2(solvent)
2], and ligand
3 (see
Section 2.4 and
Section 2.5). Among the components of this mixture, ligand
3 does not contribute to the cytotoxic activity since its IC
50 is greater than 50 μM (
Table 7, entry 3). In addition, it is noteworthy to remember that ligand
1 and complexes
1Ag and
1Pt experience hydration in water media, therefore, cells are exposed to an equilibrium of the compounds in ketone form and their hydrated counterparts. Furthermore, it is necessary to mention the compounds with chirality—compounds
2,
2Ag, and
2Pt. The IC
50 reported for
2 and
2Pt correspond to the racemic mixture and the value reported for
2Ag to the diastereomeric mixture, and since these three determinations show cytotoxic activity, enantiomeric and diastereomeric separations might be of interest for future research. Finally, a remarkable observation is that eight of the compounds—
1,
2,
1Ag,
2Ag,
3Ag,
1Pt,
3Pt, and
4Pt—are more toxic than the reference cisplatin.
In the second step of the biological evaluation, those compounds that showed antiproliferative activity in the A-549 cells were then tested on the MRC-5 non-malignant lung fibroblasts for studying selective toxicity. New candidates for antitumoral therapy should have low toxicity to non-cancerous cells (high IC
50 values) but high cytotoxicity (low IC
50 values) to cancerous cells. To study the selectivity of the compounds synthesized, the selectivity index (SI) was calculated following Equation (1).
Those compounds with an SI greater than three are considered highly selective for that cancer cell line [
37]. For compounds that are non-cytotoxic to MRC-5 cells (IC
50 > 50 μM), they are assigned an SI greater than the value calculated assuming an IC
50 = 50 μM. Regarding the ligands, none of the two active ligands (compounds
1 and
2) are selective, on the contrary, they have proven to be more toxic to MRC-5 cells (
Table 7, entries 1 and 2). Among silver complexes, only complex
3Ag has an acceptable SI of 3.2 (
Table 7, entry 7). For Pt complexes, only complex
1Pt has proven to be selective, with the greatest SI of all compounds tested (SI > 6.1,
Table 7, entry 9). It is worth mentioning that the clinically approved antitumoral drug cisplatin did not exhibit enough selectivity for the lung cancerous cell line (
Table 7, entry 13), thus, the selective profile of compounds
3Ag and
1Pt are noteworthy.