Solution studies of dinuclear polyamine-linked platinum-based antitumour complexes

Abstract

The aquation profiles of two novel dinuclear polyamine-linked, platinum-based antitumour complexes [{trans-PtCl(¹⁵NH₃)₂}₂{μ-(¹⁵NH₂(CH₂)₆¹⁵NH₂(CH₂)₆¹⁵NH₂)}]³⁺ (BBR3007, 1,1/t,t-6,6, 1) and [{trans-PtCl(¹⁵NH₃)₂}₂{μ-(¹⁵NH₂(CH₂)₆¹⁵NH₂(CH₂)₂¹⁵NH₂(CH₂)₆¹⁵NH₂)}]⁴⁺ (BBR3610, 1,1/t,t-6,2,6, 1′) have been probed using 2D [¹H, ¹⁵N] HSQC NMR spectroscopy. Reported herein are the rate constants for the hydrolysis of 1 and 1′, as well as the acid dissociation constants of the coordinated aqua ligands in their aquated derivatives. The aquation and anation rate constants for the single step aquation model in 15 mM NaClO₄ (pH 5.4) at 298 K are, for 1, k₁ = 7.2 ± 0.1 ×10⁻⁵ s⁻¹, k₋₁ = 0.096 ± 0.002 M⁻¹ s⁻¹ and, for 1′, k₁ = 4.0 ± 0.2 × 10⁻⁵ s⁻¹, k₋₁ = 1.4 ± 0.1 M⁻¹ s⁻¹. The effect of the linker backbone (Pt(tetra(m)mine vs. polyamine) was evaluated by comparison with previous data for the trinuclear complex [{trans-PtCl(NH₃)₂}₂(μ-trans-Pt(NH₃)₂{NH₂(CH₂)⁶NH₂}₂)]⁴⁺ (1,0,1/t,t,tor BBR3464). The pK₁ for 1,0,1/t,t,t(3.44) is closest to that of 1 (3.12), while the pronounced difference for 1′ (4.54), means that 1′ is the least aquated of the three complexes at equilibrium. pK_a values of 5.92 were calculated for the aquated forms of both 1 and 1′, which are 0.3 pKunits higher than for either 1,0,1/t,t,t, or the dinuclear 1,1/t,t. The higher pK_a values for both polyamine-linked compounds may be attributed to the formation of macrochelates between the central NH₂ groups and the {PtN₃O}coordination sphere of the aquated species.

Introduction

Polynuclear platinum complexes (PPCs) belong to a family of novel anticancer therapeutics that are designed based on the hypothesis that structures that are fundamentally different from the mononuclear platinum complex, cisplatin, should exhibit novel DNA binding and antitumour properties.¹ The dinuclear complex, [{trans-PtCl(NH₃)₂}₂(μ-NH₂(CH₂)₆NH₂)]²⁺ (1,1/t,tor BBR3005), and the trinuclear complex, [{trans-PtCl(NH₃)₂}₂(μ-trans-Pt(NH₃)₂{NH₂(CH₂)₆NH₂}₂)]⁴⁺ (1,0,1/t,t,tor BBR3464), Scheme 1, are two such examples of this class of anticancer therapeutics, and both exhibit significantly enhanced cytotoxicity compared to cisplatin and its derivatives.^1,2 Whilst cisplatin predominantly forms short-range intra- and inter-strand crosslinks, 1,1/t,t and 1,0,1/t,t,tboth contain one or more flexible hexanediamine linkers that allow for the formation of long-range intra- and inter-strand {Pt,Pt}DNA crosslinks.^3–6 Another distinguishing feature of both complexes is that they are positively charged (2⁺ and 4⁺, respectively). This feature has been shown to facilitate binding to DNA, through preassociative electrostatic and H-bonding interactions.⁵ Cytotoxicity and cellular accumulation have also been shown to be charge-dependent.^1,7 1,0,1/t,t,thas undergone Phase I⁸ and Phase II^9,10 clinical trials, the only Pt(_II) agent structurally distinct from cisplatin and its analogs to do so. The products of degradation in blood plasma can be replicated by reactions with sulfur-containing proteins such as glutathione.^11,12

Scheme 1.

Due to their biological relevance, namely in cell proliferation, polyamines have been exploited in drug design with several recent reviews highlighting advances in the field.^13,14 We have previously shown that incorporation of a linear polyamine such as spermidine or spermine into the basic framework, and replacement of the central tetraa(m)mine unit of PPCs, produces a series of dinuclear compounds with significant cytotoxic and antitumour activity compared to 1,0,1/t,t,t.^15–17 The complexes [{trans-PtCl(NH₃)₂}₂{μ-(NH₂(CH₂)₆NH₂(CH₂)₆NH₂)}]³⁺ (BBR3007, 1,1/t,t-6,6 or 1) and [{trans-PtCl(NH₃)₂}₂{μ-(NH₂(CH₂)₆NH₂(CH₂)₂NH₂(CH₂)₆NH₂)}]⁴⁺ (BBR3610/CTI3610, 1,1/t,t-6,2,6 or 1′) (Scheme 1) are two further examples. Complex 1′ is one of the most potent platinum complexes reported and is of special interest because it and it’s derivatives are potential “2nd-generation” PPCs for clinical trials.^18–21 The global DNA binding profile is similar to those of 1,0,1/t,t,t.²² The cytotoxicity of BBR3007 is lower than BBR3610 but this also depends on other factors besides DNA binding such as cellular accumulation.¹⁵

The substitution-labile chloride ligands of the terminal Pt am(m)ine coordination spheres in PPCs exchange readily for aqua ligands in aqueous medium, with an ensuing equilibrium between the aqua and hydroxo species (see Scheme 2).¹ 2D [¹H, ¹⁵N] HSQC NMR spectroscopy has been routinely used to study the aquation chemistry of ¹⁵N-labelled Pt am(m)ine complexes, and allows for all platinated species to be observed at micromolar concentrations.^23–29 It has been used also to monitor the mechanism and rates of formation of long-range interstrand crosslinks by PPCs in DNA.^3,5,6,30

Scheme 2.

Given the clinical relevance of 1′ and its derivatives, we have begun a comparison of its DNA binding profile with the “parent” clinical agent 1,0,1/t,t,t.³¹ In that study the role of the central NH² groups in the spermine-like linker of 1′ was also probed, by comparison with the spermidine-like derivative 1. Reported in this paper is a study of the solution behavior of both 1 and 1′, including measurement of the pK_a values of the coordinated water ligands in their aquated derivatives. These values are of interest because the aquated species is more reactive than its hydroxo counterpart in the context of DNA binding.

Results

Fully ¹⁵N-labelled 1 and 1′ were prepared by adaptations of published procedures,^15,32 involving synthesis of the polyamine linkers as ¹⁵N-derivatives, followed by coupling with two equivalents of monoactivated ¹⁵N-transplatin.

2D [¹H, ¹⁵N] HSQC NMR spectra were recorded over time to examine the spectral changes upon aquation, as previously reported.^26,27 Representative spectra are shown in Fig. 1, the collated ¹H and ¹⁵N chemical shifts are listed in Table 1 and the aquation profiles are shown in Fig. 2. The pK_a values of the coordinated aqua ligands were calculated from the pH profiles of Fig. 3.

Fig. 1.

2D [¹H,¹⁵N] HSQC NMR spectra of the Pt–¹⁵NH₃ and Pt–¹⁵NH₂ regions of 1 (a) and 1′ (b) in 15 mM NaClO₄ (pH 5.4) after reaching equilibrium at 298 K. The crosspeaks are assigned to the dichloro species, 1/1′ and the {PtN₃Cl}(2a/2a′) and {PtN₃O}(2b/2b′) moieties of the monoaqua monochloro species, 2/2′, respectively (Scheme 1). In both cases the peak at δ 4.02/− 67.0 ppm (labelled ’i’) is assumed to be a ¹⁵N tetraam(m)ine Pt impurity and is similar to that observed previously for ¹⁵N-1,1/t,t.²⁵ For 1′ the peak at δ 3.93/− 61.3 (labelled ^†) may be attributed to a polymeric ¹⁵N-labelled impurity with dangling amines induced by bis-substitution at the Pt centre, with similar structures previously observed.⁴⁴

Table 1.

¹H and ¹⁵N chemical shifts for the Pt–¹⁵NH₃ and Pt–¹⁵NH₂ groups of 1 and 1′ and their aquated derivatives^a

	1,1/t,t-6,6 (1) b, d		1,1/t,t-6,2,6 (1′) b,e		1,0,1/t,t,tc
Species	Pt–NH3	Pt–NH2	Pt–NH3	Pt–NH2	Pt–NH3	Pt–NH2
L/L′	δ1H/15N	δ1H/15N	δ1H/15N	δ1H/15N	δ1H/15N	δ1H/15N
Cl/Cl	3.85/− 64.5	4.98/− 47.0	3.85/− 64.5	4.98/− 47.0	3.84/− 64.7	4.97/− 46.9
H2O(OH)/Cl	3.85/− 64.5	4.98/− 47.0	3.85/− 64.5	4.98/− 47.0	3.84/− 64.7	4.97/− 46.9
H2O/Clf	4.08/− 62.1	5.15/− 65.3	4.09/− 62.0	5.16/− 65.3	4.09/− 61.9	5.12/− 65.1
OH/Clf	3.81/− 63.1	4.48/− 56.9	3.81/− 63.1	4.47/− 56.9	3.83/− 63.0	4.46/− 56.9

^a1H referenced internally to TSP and ¹⁵N referenced externally to ¹⁵NH₄Cl; δ in the ¹⁵N dimension is ± 0.2 ppm.

^bThis work.

^cFrom ref. 26

^dAt low pH ¹H/¹⁵N crosspeaks for the central ¹⁵NH₂ group of 1 and the aquated species 2 are visible at δ 7.88/−25.7 ppm and δ 7.10/30.2 ppm, respectively (see Fig. 4a).

^eAt low pH ¹H/¹⁵N crosspeaks for the central ¹⁵NH₂ group of 1′ and the aquated species 2′ are visible at δ 7.98/25.6 ppm and δ 7.10/31.1 ppm, respectively (see Fig. 4b).

^fValues derived from the pH titration curves (Fig. 3).

Fig. 2.

Plot of the time dependence of species in the aquation of 1 (a) and 1′ (b) in 15 mM NaClO₄ (5% D₂O/95% H₂O) at pH 5.4 and 298 K according to the single aquation model shown in Scheme 2. Key: open squares, the dichloro species 1/1′, open circles, the monoaqua monochloro species, 2/2′.

Fig. 3.

Plots of ¹H and ¹⁵N chemical shifts vs. pH for the Pt–¹⁵NH₃ groups of the monoaquated derivatives of 1 (a) and 1′ (b) in 15 mM Na phosphate.

Aquation behavior of 1 and 1′

Aquation studies for both 1 and 1′ were carried out in 15 mM NaClO₄ at pH 5.4 and 298 K, allowing direct comparison with previous aquation studies of 1,0,1/t,t,t.²⁶ At this pH, the crosspeaks in the Pt–¹⁵NH₃ region of the 2D [¹H, ¹⁵N] HSQC spectra are free from overlap, enabling the spectral changes corresponding to hydrolysis to be monitored over time. Complex 1 took 6 h to reach equilibrium, whilst 1′ achieved equilibrium after 2 h.

Fig. 1a shows a 2D [¹H, ¹⁵N] HSQC NMR spectrum of 1 after the sample had reached equilibrium. Two major crosspeaks in the Pt–¹⁵NH₃ region at δ 3.85/− 64.5 and 4.08/−62.2 ppm are assigned to the dichloro species, 1, and the {PtN₃O}end (2b) of the monoaqua monochloro species, 2, respectively. Corresponding crosspeaks in the Pt–¹⁵NH₂ region are observed at δ 4.98/−47.0 ppm (1) and 5.08/−64.2 ppm (2b). This change in the ¹⁵NH₂ chemical shift of Δδ = 17.2 ppm is consistent with the displacement of the trans chloro group with an aqua ligand and is similar to that reported previously for 1,0,1/t,t,t.²⁶

A 2D [¹H, ¹⁵N] HSQC NMR spectrum of the solution of 1′ after reaching equilibrium is shown in Fig. 1b. The ¹H/¹⁵N chemicals shifts of the Pt–¹⁵NH₃ groups are δ 3.85/−64.5 (1′) and 4.08/−62.2 ppm (2b′) and are identical to those observed for 1. In the Pt–¹⁵NH₂ region, the corresponding crosspeaks are at δ 4.98/−47.0 (1′) and 5.08/−64.3 ppm (2b′). The crosspeaks corresponding to the {PtN³Cl} moiety of the monoaqua monochloro species, 2a and 2a′, in both the Pt–¹⁵NH₃ and Pt–¹⁵NH₂ regions, are overlapped with those of the dichloro species (1 and 1′).

Table 1 compares the ¹H/¹⁵N chemical shifts of the Pt–¹⁵NH₃ and Pt–¹⁵NH₂ groups of 1, 1′ and their aquated derivatives, as well as the corresponding species for 1,0,1/t,t,t²⁶ and no significant differences are observed.

The proposed reaction pathway for the aquation of both 1 and 1′ is shown in Scheme 2, along with the model used in the analysis of the kinetic data, which assumes that both complexes undergo a single aquation step. The time dependent changes in concentration of the species detected during the aquation are shown in Fig. 2, along with the curves of best fit to the kinetic model. The rate constants and derived equilibrium constants are listed in Table 2, in comparison to those of 1,0,1/t,t,t.²⁶

Table 2.

Rate and equilibrium constants for the aquation of 1 and 1′ at 298 K in 15 mM NaClO₄ (pH 5.4) in comparison to 1,0,1/t,t,t^a

Parameter	1	1′	1,0,1/t,t,tb
k1 (10−5 s−1)	7.2 ± 0.2	4.0 ± 0.2	10.7 ± 0.1
k−1 (M−1 s−1)	0.096 ± 0.002	1.4 ± 0.1	0.294 ± 0.004
pK1	3.12 ± 0.02	4.54 ± 0.06	3.44 ± 0.04

^aValues derived using the single aquation model shown in Scheme 2.

^bFrom ref. 26

pK_a determination of the aquated derivatives of 1 and 1′

To determine the pK_a values of the coordinated aqua ligands, solutions of 1 and 1′ (in 15 mM Na phosphate), that had reached equilibrium over a 24 h period, were titrated across the pH range ~2–10. The pH dependent changes in the ¹H/¹⁵N chemical shifts of the Pt–¹⁵NH₃ groups of the terminal {PtN₃O}moieties of the aquated derivatives 2 and 2′, were measured by 2D [¹H, ¹⁵N] HSQC NMR spectroscopy (see Fig. 3). Identical pK_a values of 5.92 ± 0.02 were derived from the change in ¹H chemical shifts for the aquated derivatives of both compounds. The pH titration of 1 was also carried out in 100 mM NaClO₄ and an identical pK_a value was obtained. Interestingly, this value is 0.3 pK units higher than that of the aquated forms of both 1,1/t,t and 1,0,1/t,t,t(5.62) (measured in 100 mM NaClO₄).^26,27

Analysis of the central polyamine linker groups of 1 and 1′

The polyamine-linked complexes have the advantage that ¹⁵N-labelling of the central NH₂ groups allows interrogation by 2D [¹H, ¹⁵N] HSQC NMR methods and helps delineate the role of central charge in the overall profile of these agents. The central NH₂ groups, for both 1 and 1′, were only clearly visible in the [¹H,¹⁵N] HSQC NMR spectra at lower pH and there are interesting differences between the two compounds apparently related to the presence of the central ethylenediamine unit in 1′.

A ¹H NMR spectrum recorded for 1, at pH 5.4 in 15 mM NaClO₄ (Fig. S1a^†) shows clearly four resonances corresponding to the four inequivalent –CH₂– groups (1–6) in the alkyl chain (see Scheme 1). The chemical shifts of these resonances are listed in Table 3. The 1 and 6 –CH₂– groups show ¹⁵N-coupling (²J(¹H–¹⁵N) = 7.79 Hz) as a consequence of their proximity to the terminal Pt–¹⁵NH₂ and central ¹⁵NH₂ groups, respectively. For 1′, five ¹H NMR resonances are observed (Fig. S1b^†) corresponding to the inequivalent –CH₂– protons of the alkyl chain (1–7; see Scheme 1), and the chemical shifts (for 1–6) are similar to those observed for 1 (Table 3). Surprisingly, no ¹⁵N splitting is observed for the –(CH₂)₂– protons (7) which are bonded to the central ¹⁵NH₂ groups. For the ¹⁵N-labelled complex, these –CH₂– protons constitute the AA′ part of an A2XX′A2′ spin system, as a result of unequal ¹⁵N–¹H spin–spin coupling to the two ¹⁵N atoms. A similar spin-system is observed for the bidentate phosphines with –(CH₂)₂– backbones, such as Ph₂P(CH₂)₂PPh₂ (dppe), where a quasi-triplet is observed for these protons.³³ The ¹⁵N {¹H}DEPT NMR spectrum of 1, at pH 5.4, clearly shows a ¹⁵N resonance at δ 25.7 ppm assignable to the central –¹⁵NH₂– group in the polyamine chain (Fig. S2a^†), however for 1′, no ¹⁵N resonance correlating to the central –¹⁵NH₂(CH₂)₂¹⁵NH₂– was observed under these conditions (Fig. S2b^†).

Table 3.

¹H chemical shifts of the alkyl CH₂ groups of 1 and 1′^a

–CH2–	1	1′	1,0,1/t,t,tb
1c	2.69	2.68	2.68
2	1.68	1.70	1.68
3	1.39	1.39	1.38
4	1.39	1.39	1.38
5	1.69	1.70	1.68
6c	3.03	3.11	2.68
7	—	3.42	—

^aFor numbering scheme see Scheme 1; ¹H chemical shifts referenced to TSP.

^bFrom ref. 26

^c2J(¹H–¹⁵N) = 7.79 Hz.

Fig. 4a shows the central ¹⁵NH₂ region of the 2D [¹H, ¹⁵N] HSQC NMR spectrum of 1 at pH 2.5 (after allowing the sample to reach equilibrium). The major crosspeak (δ 7.88/25.7 ppm) is assigned to the central NH₂ group of the dichloro species, as it correlates with the resonance observed in the ¹H spectrum recorded of a freshly prepared sample (Fig. S1a^†) and the ¹⁵N{¹H} DEPT spectrum (Fig. S2a^†). The crosspeak at δ 7.10/30.2 ppm is assigned to the central NH₂ groups of the monoaqua monochloro species, 2, and the crosspeak at δ 7.46/12.0 ppm to the terminal NH₃⁺ group of a dangling primary amine, most likely arising from acidic cleavage of the Pt–NH₂R bond. As the pH is increased the crosspeaks diminish in intensity. At pH 4.5 the peak for 2 was no longer visible but that for 1 was still present, albeit weaker in intensity. The crosspeaks for 1 and the dangling amine species (d) were no longer visible above pH 5.7 and pH 5.2, respectively.

Fig. 4.

The central ¹⁵NH₂ region of the 2D [¹H, ¹⁵N] HSQC (top) and ¹H NMR spectra (bottom) of 1 (pH 2.5 in 100mM NaClO₄) (a) and 1′ (pH 2.1 in 15 mM Na phosphate) (b) showing the dichloro species, 1/1′, and the aquated derivatives, 2/2′. Crosspeaks/multiplets labelled ’d’ are assigned to Pt species with a dangling amine, most likely arising from acidic cleavage of the Pt-NH₂R bond (see text). The high relative intensity of these peaks may be explained by a slower exchange with bulk water for NH protons of the terminal 1° amine compared to those of the 2° amine in the linker of 1/1′.

Fig. 4b shows the 2D [¹H, ¹⁵N] HSQC NMR spectrum and the corresponding ¹H NMR spectrum of the central ¹⁵NH₂ group of 1′ at pH 2.1. The ¹H/¹⁵N crosspeak at δ 7.98/25.6 ppm is assigned to the dichloro species and is deshielded in the ¹H dimension by δ 0.1 ppm compared to the analogous crosspeak for 1 (see above). The ¹H/¹⁵N crosspeak at δ 7.10/31.1 ppm is assigned to the central NH₂ groups of the monoaqua monochloro species, 2′. The most intense ¹H/¹⁵N crosspeak observed at δ 7.50/12.0 ppm correlates with a doublet of triplets centered at δ 7.49 ppm in the ¹H NMR spectrum (splitting = 72 Hz) and is assigned, as for 1, to a Pt by-product with a dangling amine and shows the characteristic splitting pattern.¹² Further evidence for the assignment is a small resonance at δ 3.01 ppm in the ¹H NMR spectrum (labelled ′d′, Fig. S1b^†) assigned to the CH₂ group nearest the terminal NH₃⁺ moiety. In the [¹H,¹⁵N] HSQC NMR spectrum the ¹H/¹⁵N crosspeak corresponding to the central ¹⁵NH₂ group of the aquated species, 2′, was no longer visible above pH 3.5 whilst the analogous crosspeaks for the dichloro species 1′ and the dangling amine impurity (d) disappeared as the pH was raised to pH 4.1 and pH 4.9, respectively.

Discussion

2D [¹H, ¹⁵N] HSQC NMR spectroscopy is a powerful tool in the understanding of the aquation profile of ¹⁵N-labelled platinum am(m)ine anticancer complexes. The aquation and acid dissociation constants of two fully ¹⁵N-labelled novel platinum anticancer complexes, 1 and 1′, have been studied under conditions similar to those previously used for 1,0,1/t,t,t, allowing for a direct comparison.²⁶

There is literature precedence for the ¹⁵N chemical shifts of ¹⁵N-labelled free polyamines (such as putriscene, spermine and spermidine)³⁴ but no Pt-based anticancer complexes with ¹⁵N-labelled polyamine linkers have been investigated previously. An interesting aspect of the ¹⁵N-{¹H} and [¹H,¹⁵N] HSQC NMR spectra of the platinated polyamines is that the observation of resonances for the central NH₂ groups is pH-dependent. For 1, the ¹H/¹⁵N resonance for the central ¹⁵NH₂ group in the polyamine chain is visible below pH 5.7. The ¹⁵N chemical shift (δ 25 ppm) is in agreement with previous studies on related ¹⁵N-labelled free polyamines.^34–36 However for 1′, the ¹H/¹⁵N resonance for the central –¹⁵NH₂(CH₂)₂¹⁵NH₂– moiety is not observed above pH 4.1. Perrin showed that decreasing the number of –CH₂– groups between the two terminal amine groups of diamines (such as H₂N(CH₂)₂NH₂, H₂N(CH₂)₃NH₂, H₂N(CH₂)₄NH₂), increased the difference between the two acid dissociation constants pK_a1 and pK_a2.³⁷ For ethylenediamine pK_a1 is 9.9–10.2 and pK_a2 is 6.8–7.5. It is not expected that the pK_a of the central linker will change dramatically upon platination.³⁸ Discrete ¹⁵N chemical shifts for various polyamines (spermine, spermidine, thermospermine and thermine)^34,35 have been observed over the pH range 5.9–12.5. In all cases, however, the –¹⁵NH₂– groups are separated by at least three –CH₂– groups. The absence of the corresponding resonances for the ethylenediamine moiety of 1′ under the experimental conditions may be attributed to the close proximity of the two –NH₂– groups. At pH > 4.1 ¹H exchange between the two ¹⁵N atoms may broaden the signals out beyond detection. This phenomenon is not observed in the case of 1 as there is only one central amine group; the central NH₂ resonance is clearly observed.

The aquation studies of 1 and 1′ reveal interesting differences compared to 1,0,1/t,t,t. For the forward reaction (Pt–Cl → Pt–OH₂), the rate constants for both polyamine-linked complexes are lower than that of 1,0,1/t,t,t, but more significant for 1′ (2.7-fold lower) than 1 (1.5-fold lower). Similarly, the rate constant for the anation reaction (Pt–OH₂ → Pt–Cl) is significantly higher for 1′ (1.4 M⁻¹ s⁻¹) compared to both 1 (0.096 M⁻¹ s⁻¹) and 1,0,1/t,t,t(0.294 M⁻¹ s⁻¹). Thus, the dichloro species of 1′ is more favoured at equilibrium in comparison to either 1 or 1,0,1/t,t,t. Based on these results it is clear that the central NH₂ groups have an influence over the kinetics of the aquation of the terminal {PtN₃Cl} group, whereas it might have been expected that the central linker (be it platinum or amine) is far enough away from the terminal {PtNH₃Cl} groups so as not to influence them.

A further point of interest is that the pK_a values of aquated derivatives of both 1 and 1′ (5.92) are 0.3 pKunits higher than that of 1,0,1/t,t,t or 1,1/t,t. Hence at physiological pH for both complexes there will be a greater proportion of the more reactive aquated species (and less of the less reactive hydroxo species) compared to 1,0,1/t,t,t. A possible explanation for this difference is the formation of a hydrogen bond between the central NH₂ group and the aqua ligand of the terminal {PtN₃O} moiety (see Fig. 5). The flexibility of the linker will help in this regard. Formation of such a macrochelate could explain also the large ¹H chemical shift changes of the central NH₂ groups upon aquation (Fig. 4). Notably, the ¹H/¹⁵N shifts of the aquated species of both 1 and 1′ in phosphate and perchlorate buffers are similar (Fig. 4) and the pK_a value of the aquated derivatives are also the same in the different electrolytes. In studies of the interaction of the related 1,1/c,c with PO₄³, a macrochelate species with phosphate bridging between the two Pt groups was characterized.²⁷ The distance between the two terminal Pt groups in 1,1/c,cis identical to the distance between the terminal Pt groups and the central NH₂ groups of 1 and 1′. Further, a unique glutathione-bridged moiety producing a novel 11-membered chelate ring was also observed with 1,1/c,c.³⁹ The nature of the macrochelate is however different in the present case. Formation of the phosphate-bridged macrochelate is facilitated by the cis geometry of the Pt–NH₃ groups in 1,1/c,c. Both 1 and 1′, have transPt–NH₃ groups and hence the electrolyte does not play a role in the macrochelate formation.

Fig. 5.

Proposed macrochelate structure with a possible H-bond between the terminal {PtN₃O} moiety and the central NH₂ groups of the polyamine linker (1′ has been used as the example). The structure was created using the program package Spartan ’08 v1.0.0 (Wavefunction, Inc., Irvine, California). The illustration was created using the program package VMD.⁴⁵

Conclusion

The two novel dinuclear polyamine-linked platinum antitumour complexes 1 and 1′ exhibit interesting aquation chemistry compared to 1,0,1/t,t,tand 1,1/t,t. Clear evidence for the effect on aquation of the central amine motifs has been found, suggesting novel solution behavior which merits further investigation. BBR3610 was designed to have the same distance between the Pt–Cl units as BBR3464, with the central ethylenediamine unit mimicking the Pt(tetraamine) unit of BBR3464 with respect to hydrogen-bonding and electrostatic interactions.¹⁶ The observation of macrochelate species in the present case suggests that there is significantly more conformational flexibility with the polyamine linker – or in essence the central Pt(tetraamine) unit in BBR3464 induces some conformational rigidity in the trinuclear species, which may be reflected in its’ reactions with biomolecules. The polyamine linkers influence both the kinetics of the aquation reactions and the pK_a values of the coordinated water ligands, compared to the trinuclear (central Pt(tetra(m)mine linker) case. The global DNA binding profile of BBR3610 is similar to that of BBR3464.²² Kinetic studies of substitution of BBR3610 by small molecules such as methionine and 5′-GMP,²⁰ also suggest a broadly similar profile in aqueous medium to that of BBR3464, with likelihood of similar degradation (metabolism) profiles for polyamine-bridged species in comparison to that of the trinuclear drug.^12,20 The use of less substitution-labile leaving groups²¹ or use of different geometries (1,1/c,c) by use of 1,2-dach carrier groups^22,40 are valid approaches to systematically alter the pharmacokinetic profile of the PPCs.

Experimental

Synthesis of compounds

[{trans-PtCl(¹⁵NH₃)₂}₂{μ-(¹⁵NH₂(CH₂)₆¹⁵NH₂(CH₂)₆¹⁵NH₂)}]³⁺(1) and [{trans-PtCl(¹⁵NH₃)₂}₂{μ-(¹⁵NH₂(CH₂)₆¹⁵NH₂(CH₂)₂-¹⁵NH₂(CH₂)₆¹⁵NH₂)}]⁴⁺ (1′). The fully ¹⁵N-labelled derivatives of 1 and 1′ were prepared by adaptations of published procedures.^15,32 The general strategy is to synthesize the fully ¹⁵N-labelled linkers with the central ¹⁵NH₂ amine groups blocked using the tert-butyloxycarbonyl (BOC) protecting group. The protected linker is then incorporated into the dinuclear compound by reaction with two equivalents of monoactivated transplatin (trans-[PtCl₂(¹⁵NH₃)₂]) – deprotection and removal of the central BOC protecting groups affords the desired fully ¹⁵N-labelled 1 and 1′.

NMR spectroscopy

The NMR spectra were recorded using a Bruker 600 MHz spectrometer (¹H, 599.92 MHz; ¹⁵N, 60.79 MHz). The ¹H shifts were referenced internally to TSP (sodium-3-trimethylsilyl-D₄-propionate) and the ¹⁵N shifts were referenced externally to ¹⁵NH₄Cl (1.0 M in 1.0 M HCl in 5% D₂O/95% H₂O) (δ = 0.0 ppm). 2D [¹H, ¹⁵N] HSQC NMR spectra were recorded using the pulse sequence of Stonehouse et al.⁴¹ 1D ¹⁵N {¹H} DEPT NMR spectra were recorded using a zgig pulse sequence with WALTZ decoupling⁴² optimized for ¹J(¹⁵N–¹H) = 72 Hz and a 90° pulse of 12.75 µs, with 256 transients recorded over a period of 1.5 h. Continuous wave (CW) decoupling in the ¹H NMR spectra was achieved by irradiating the desired resonances at a power level of 15 dB at the frequency of the Pt–¹⁵NH₃ group in the 1D ¹⁵N {¹H} DEPT NMR spectra.

pH measurements

The pH values were measured using a Shindengen ISFET (semiconductor) pH meter (pH Boy-KS723 (SU-26F) and calibrated against buffers of pH 4.0, 6.9 and 10.0. To avoid leaching of chloride ions, 5 µL samples were placed on the electrode and the pH recorded without returning the aliquot to the sample. Adjustments to the pH were carried out using 0.02, 0.05, 0.1 or 0.5 M solutions of either HClO₄ or NaOH in 5% D₂ O/95% H₂O.

Aquation studies

Solutions were prepared by dissolving 1 (0.80 mg, 0.85 µmol) or 1′ (0.35mg, 0.33 µmol) in NaClO₄ (478 µL, 15 mM) in 5%D₂O/95% H₂O with TSP (2 µL, 13.3 mM) to give initial concentrations of 1.80 mM and 0.69 mM of 1 and 1′, respectively. A series of ¹H and 2D [¹H, ¹⁵N] HSQC NMR spectra were recorded at 298 K, until equilibrium conditions were attained. For the 1D {¹H} ¹⁵N DEPT experiments, solutions of 1 (2.00 mg, 2.13 µmol) or 1′ (0.70 mg, 0.67 µmol) in Na phosphate (478 µL, 15 mM) were prepared in 5% D₂O/95% H₂O with TSP (2 µL, 13.3 mM), to give initial concentrations of 4.44 mM and 1.39 mM of 1 and 1′, respectively.

Data analysis

Concentrations of species were determined by measuring peak volumes in the Pt–¹⁵NH₃ region of the 2D [¹H, ¹⁵N] HSQC NMR spectra using the plug-in “2D NMR Analysis” developed for the program ImageJ.⁴³ Peak volumes were converted into concentrations relative to the initial concentration of the dichloro species of 1 or 1′. The concentrations of 1 and 2 (and 1′ and 2′) at each time point were derived based on the relative volumes of the Pt–¹⁵NH₃ crosspeaks after correcting for peak overlap, according to the single aquation model, as described previously.²⁶ Differential equations were used to fit to either first- or second-order rate equations. The aquation rate constants were determined using a non-linear optimization process using SCIENTIST (Version 2.0, MicroMath Inc.) with errors reported for one standard deviation.

pK_a determination of the aquated derivatives of 1 and 1′

Solutions of 1 and 1′ (in 15 mM Na phosphate), that had reached equilibrium over a 24 h period, were titrated across the pH range ~2–10. Adjustments in pH were performed using 0.02, 0.05, 0.1 or 0.5 M solutions of either HClO₄ or NaOH in 5% D₂O/95% H₂O. For each pH point, 1D ¹H and 2D[¹H, ¹⁵N] HSQC NMR spectra were recorded and changes in the ¹H and ¹⁵N chemical shifts of the Pt–¹⁵NH₃ groups of the {PtN₃O} moiety (2b and 2b′) were monitored.

Kaleidagraph (Synergy Software, Reading, PA) was used to analyze the pH titration data using eqn (1):

$$\delta =({\delta }_{\mathrm{A}}[{\mathrm{H}}^{+}]+{\delta }_{\mathrm{B}}{K}_{\mathrm{a}})/([{\mathrm{H}}^{+}]+K_{\mathrm{a}})$$ (1)

where K_a is the acid dissociation constant for the Pt–OH₂ group of the aquated species and δ_A and δ_B are the chemical shifts of the aqua and the hydroxo species, respectively. Fig. 3 shows the pH titration curves that have been fit to eqn (1) for the aquated derivatives of both 1 and 1′.