Influence of geometric isomerism on the binding of platinum anticancer agents with phospholipids

Abstract

Reported herein is a detailed NMR and DFT study of the interaction of the ¹⁵N-labelled dinuclear platinum anticancer compound [{cis-PtCl(NH₃)₂}₂{μ-H₂N(CH₂)₆NH₂}]²⁺ (¹⁵N-1, 1,1/c,c) with 1,2-dihexanoyl-sn-glycero-3-phosphate (DHPA), as a comparison with an earlier study of the interaction of the same water-soluble phospholipid fragment with the geometric trans isomer (1,1/t,t). The reaction of ¹⁵N-1 with the sodium salt of DHPA was studied at 298 K, pH ~5.6, by [¹H,¹⁵N] HSQC 2D NMR spectroscopy. The NMR data, supported by DFT models, provide evidence that the monofunctional DHPA adduct of ¹⁵N-1 exists in two conformational forms, with different orientation of the (CH₂)₆ linker; one has an interaction between the unbound {PtN₃Cl} moiety and the coordinated DHPA molecule. Similarly, two bifunctional adduct conformers are identified, in which one has an interaction between the phosphate groups of the two bound DHPA molecules. When compared to the previously reported reactions of 1,1/t,t with DHPA, equilibrium conditions of the 1,1/c,c reaction are reached more slowly (120 h), similar to the reaction with phosphate. The rate constant for the first step of DHPA binding (k_L) is slightly lower (1.6 fold) for the cis- compared to the trans-isomer, whereas the rate constant for the reverse reaction is 4-fold lower, resulting in a much greater proportion of DHPA bound species at equilibrium.

Graphical abstract

A holistic approach to the molecular mechanism of platinum anti-cancer agents extends DNA interactions to membrane phospholipids relevant to cellular accumulation and delivery.

Introduction

Platinum chemistry has been in the forefront of elucidation of kinetics and mechanisms of substitution reactions in inorganic chemistry.¹ The favourable kinetic profiles (reactions with half-lives of hours and days) were perhaps very well suited for an era with less advanced technical instrumentation than exists today and most textbooks rely heavily on examples of platinum complex reactions to elucidate the principles of associative reactions and elaborate on theoretical and practical aspects of the kinetic trans effect as well as the thermodynamic trans influence. Dominating and underpinning many of these aspects is the concept of Pt(II) as a quintessential ‘soft’ acid with high affinity for soft bases such as sulphur nucleophiles. Nitrogen donors, especially of “aromatic” systems such as pyridine or delocalised systems of purines and pyrimidines can be considered as intermediate soft bases. This balance is manifested in many discussions of the biological activity of platinum complexes – comparing metabolism and deactivation (predominantly S-binding) and target activation/fixation (Pt-N binding on DNA). In either case, covalent binding is expected to produce essentially irreversible reactions with formation of strong Pt-S and Pt-N bonds. In contrast, systematic study of oxygen donors in biology is lacking, especially with respect to formation of Pt-O-donor species in reaction pathways associated with perhaps substitution-labile or transient species formed during metabolism, cell entry and even target binding. Monodentate O-donors are hard bases and Pt-O bonds are expected to be more substitution-labile compared to chloride, or N and S-donors (HSAB). Nevertheless, these interactions could be relevant in the biological milieu due to high local concentrations of common “weak” ligands such as carbonate and sulfate.

In previous work we pioneered the use of [¹H,¹⁵N] NMR spectroscopy to explore the aquation chemistry of cisplatin² and other platinum anticancer complexes³ including polynuclear platinum complexes (PPCs).^{4, 5} We have also carried out detailed studies exploring the kinetics of binding of PPCs to simple O-donor anions (phosphate, acetate and sulfate)^{6, 7} and showed how subtle changes to the platinum coordination sphere (trans vs cis isomers) influences the pK_a of the coordinated aqua ligands and the speciation profile.⁶

In recent work we have built on these fundamental studies to explore the binding of PPCs to neglected biological targets with O-donor ligands: Heparan Sulfate Proteoglycans (HSPGs) and phospholipids, which are important components of cell membranes and are involved in PPC cellular internalization.⁸ Our studies focus on developing a detailed understanding of reactions with simple model complexes, which can then be extrapolated to relevant biological targets. For example, we have investigated the interaction of the trinuclear (Triplatin or 1,0,1/t,t,t) and dinuclear (1,1/t,t) PPCs with three D-glucosamine residues containing varied O-sulfate and N-sulfate or N-acetyl substitutions, which represent monosaccharide fragments present within the repeating disaccharide sequences of cell surface heparan sulfate.⁹ We have studied also the interaction of these PPCs with 1,2-dihexanoyl-sn-glycero-3-phosphate (DHPA), a shorter, water-soluble model for the phospholipid DPPA (Chart 1).¹⁰ These latter studies showed that the monofunctional DHPA adduct of 1,1/t,t coordinated to a second DHPA molecule in two different orientations, leading to formation of two conformers of the bifunctional adduct. Triplatin, which contains a charged {PtN₄}²⁺ central linker, is capable of forming higher order structures through “phosphate clamp” interactions with two molecules of DHPA.

Chart 1.

Structures of the platinum complexes and phospholipids discussed in this work.

In the present paper, we study the effect of geometric isomerism on phospholipid binding by comparing the previously studied dinuclear trans isomer (1,1/t,t) to its cis isomer (1,1/c,c). The kinetics of 1,1/c,c-DHPA binding are also compared with previous studies on simple phosphate binding and with those reported for the binding of 1,0,1/t,t,t and 1,1/t,t with DHPA.¹⁰ The work highlights how the interplay of modern experimental ([¹H, ¹⁵N] NMR spectroscopy) and theoretical (DFT modelling) techniques allows prediction of interactions of platinum anticancer complexes with relevant biological targets by extrapolating studies with simple model systems.

Results and Discussion

NMR studies

The reaction of ¹⁵N-1 with DHPA (1:5 mol ratio) in 100 mM NaClO₄ reaction was followed for a period of 5 days in the approach to equilibrium. Representative [¹H,¹⁵N] HSQC NMR spectra are shown in Fig. 1 and the chemical shifts of the species observed during the reaction are listed in Table 1. The assignment of the species observed was made based on the previous reaction of 1,1/c,c with phosphate⁶ and comparison of the ¹H and ¹⁵N chemical shifts of the peaks in the cis-and trans-NH₃ and Pt−NH₂ regions.

Figure 1.

2D [¹H,¹⁵N] HSQC NMR (600 MHz) spectra at 298K of ¹⁵N-1 (1 mM) with DHPA (5 mM) in 100 mM NaClO₄ (a) after 36 min, (b) 17.4 h and (c) 117 h after mixing. The Pt-NH₂ region is shown as an inset.

Table 1.

¹H and ¹⁵N chemical shifts for species observed during the reaction of ¹⁵N-labelled 1,1/c,c (1) with DHPA in 100 mM NaClO₄ at 298 K.^a

Compound	Species	cis-NH3		Pt-NH2		trans-NH3
		δ 1H	δ 15N	δ 1H	δ 15N	δ 1H	δ 15N
1,1/c,c	Pt-Cl(A)b	3.76	−66.0	4.49	−44.2	4.29	−68.8
	Pt-H2O(B)	3.92	−63.9	4.61	−42.1	4.22	−84.2
	Pt-DHPA (Cm/Cb)b	3.86	−65.3	4.56	−43.5	4.21	−83.6
	Pt-DHPA(D)	3.90	−65.3	4.65	−42.8	4.26	−84.0
	Pt-PO4(E)c	3.86	−62.5	4.64	−41.5	4.17	−86.4

^a1H referenced to TSP (sodium 3-trimethylsilyl-[D4]-propionate); ¹⁵N referenced to ¹⁵NH₄Cl (external).

^bFor one conformer of the mono DHPA adduct the trans-NH₃ and Pt-NH₂ peaks of the {PtN₃Cl} unbound end are distinguishable from those of the parent chlorido complex at δ ¹H/¹⁵N = 4.32/−68.2 and 4.61/−42.8, respectively (see Fig. 1(b) and (c), peaks labelled C_m’).

^cProduct from the reaction of ¹⁵N-1 with a minor hydrolysis product of DHPA, which has ¹H,¹⁵N chemical shifts similar to that of the {PtON₃} unit of the chlorophosphato species formed on reaction of ¹⁵N-1 with phosphate at a similar pH;⁶ analogous minor products were observed in the reactions of 1,1/t,t and 1,0,1/t,t,t with the same DHPA sample.¹⁰

In the first spectrum recorded after 36 min (Fig. 1(a)), in addition to the peaks for ¹⁵N-1 (labelled A), there are three peaks assignable to the {PtN₃O} group of the monoaquated species (labelled B) with chemical shifts consistent with those obtained from pH titration curves at this pH (5.6).⁶

Monofunctional adduct formation.

Careful inspection of the [¹H,¹⁵N] HSQC NMR spectra provides evidence for two different conformational forms of the monofunctional DHPA adduct. ¹H,¹⁵N peaks assignable to the first formed adduct (1,1/c,c-MF-1) were just visible in the first spectrum and continued to increase in intensity over time. In contrast to the reaction with phosphate,⁶ in the trans-NH₃ region the peak for the {PtN₃Cl} group (labelled C_m’, Fig. 1(b)) is slightly deshielded in the ¹H dimension (Δδ 0.03) and distinguishable from that of the parent chlorido species, suggesting an interaction with the bound DHPA molecule. Notably the cis-geometry allows interactions between the two end {PtN₃Y} groups of 1,1/c,c that is not possible in the 1,1/t,t case.⁶ A strongly deshielded peak at δ 4.61/−42.8 (labelled C_m’, Fig. 1(b)), is assigned to the Pt-NH₂ protons cis to the unbound chlorido group, on the basis of the H-bonding interactions observed in the model (Fig. 2(a)), see below). In the second monofunctional adduct (1,1/c,c-MF-2) there is no interaction between the {PtN₃O} and {PtN₃Cl} groups (Fig, 2(b)) and hence the ¹H/¹⁵N peaks for the {PtN₃Cl} group will be coincident with that of the parent species.

Figure 2.

DFT optimized models of the DHPA bound adducts of 1,1/c,c (1). Two different orientations of the (CH₂)₆ linker lead to different conformers of the monofunctional (1,1/c,c-MF-1 (a) and (1,1/c,c-MF-2 (b)) and bifunctional (1,1/c,c-BF-1 (c) and 1,1/c,c-BF-2 (d)) adducts. In (a) and (c) there is an interaction between the two {PtN₃Y} coordination spheres. The models are largely consistent with the ¹H,¹⁵N peaks observed in the [¹H,¹⁵N] HSQC NMR spectra (Fig. 1). Expansions showing the details of the H-bonding interactions are shown in Fig. 3 and Fig. 4.

For both adducts, coordination of the DHPA molecule by the phosphate O atom places the cis-NH₃ in a similar hydrogen bonded environment to that previously reported for the Pt-NH₃ groups in 1,1/t,t,¹⁰ and this peak (δ 3.86/−65.3, labelled C_m) is similarly deshielded in both ¹H (Δδ = 0.10) and ¹⁵N (Δδ = 0.7) dimensions with respect to the parent chlorido species. The Pt−NH₂ group (δ 4.56/−43.5), which is similarly cis to the coordinated DHPA molecule, exhibits a similar deshielding with respect to the parent chlorido complex (¹H Δδ = 0.07, ¹⁵N Δδ = 0.7). The trans-NH₃ group (δ 4.21/−83.6) is noticeably shielded in the ¹H dimension (Δδ −0.08) with respect to the parent chlorido complex, whereas in the case of 1,1/t,t the ¹H shifts for the analogous Pt-NH₂ protons were almost identical in species A and C.¹⁰

By observing the change in intensity of trans-NH₃ peak for the unbound {PtN₃Cl} group (C_m’ Fig. 1) the mono-DHPA adduct (1,1/c,c-MF-1) reached a maximum intensity after about 9 h, and then decreased so that it represents only 2% of the total Pt species at equilibrium

Bifunctional adduct formation.

In the reaction of 1,1/c,c with phosphate,⁶ distinct peaks were observed in the trans-NH₃ region assignable to the phosphatoaqua species, attributable to an interaction between the coordinated aqua and phosphato groups which influences the ¹H chemical shift of the trans-NH₃ ligand. This species is an intermediate to the formation of the macrochelate phosphate bridged species which represented 25% of the total Pt species at equilibrium. No similar peaks were observed in the reaction with DHPA indicating that there is no interaction between the bound DHPA phosphate group and the unbound {PtN₃OH₂} moiety. However, analysis of the ¹H,⁵N peaks, in conjunction with the DFT models (below), provides evidence for formation of two distinct bifunctional DHPA adducts. ¹H,¹⁵N peaks assignable to 1,1/c,c-BF-1 (labelled D) were first visible after about 4 h and grow slowly accounting for 18% of the final total Pt species. There are significant changes in ¹H and ¹⁵N shifts indicative of interactions between the two bound DHPA molecules. The protons of the cis-NH₃ and Pt−NH₂ groups (both cis to the bound DHPA molecule) are more strongly deshielded (Δδ 0.04 (NH₃) and 0.09 (NH₂)); the protons of the trans-NH₃ group are similarly deshielded (Δδ 0.05), whereas a shielding (Δδ −0.4) is observed in the ¹⁵N dimension. For the second bifunctional adduct (1,1/c,c-BF-2), the ¹H and ¹⁵N chemical shifts of the Pt-NH₂ and cis-and trans-NH₃ groups of the bound {PtN₃O} groups (peaks labelled C_b in Figs. 1(b and c)) are indistinguishable from those in the mono-DHPA adduct(s) (C_m) indicating that there is no interaction between the two ends of the 1,1/c,c molecule. These peaks represent 64% of the total Pt species present at the end of the reaction.

DFT modelling

Two DFT optimized models for the monofunctional adduct of 1,1/c,c with DHPA were constructed and are shown in Fig. 2 (a and b). In the first model (1,1/c,c-MF-1; (Fig. 2(a)) the orientation of the (CH₂)₆ chain enables the unbound {PtN₃Cl} sphere to interact with the coordinated DHPA molecule. The model is analogous to previously constructed models of 1,1/c,c aquation products in phosphate, where the cis-geometry enabled interaction between the two Pt coordination spheres.⁶ Expansion of the H-bonding interactions are shown in Fig. 3 (a and b). For the {PtN₃O} moiety (Fig. 3(a)) the cis-NH₃ group has short H-bonds to the POCH₂ oxygen atom (distance = 2.36 Å) and the carbonyl oxygen from DHPA aliphatic chain b (distance = 2.00 Å), which also has a H-bond (distance = 2.20 Å) to the trans-NH₃ group. There are no H-bonds involving the Pt−NH₂ protons. For the {PtN₃Cl} moiety (Fig. 3(b)) a hydrogen bond between the trans-NH₃ group and the carbonyl oxygen of chain a of the coordinated DHPA (distance = 2.05 Å), is consistent with the observation of a distinct ¹H/¹⁵N peak (C_mʹ; Fig. 1(b)). The NH₂ group is H-bonded to both the same carbonyl oxygen (distance = 2.20 Å) and the phosphate oxygen of the coordinated DHPA (distance = 1.97 Å) and is consistent with the strongly deshielded ¹H,¹⁵N peak in the Pt-NH₂ region (C_mʹ;Fig 1(b)).

Figure 3.

Hydrogen bonding interactions in the DFT optimized DHPA bound adducts of 1,1/c,c (1) in which there is an interaction between the two {PtN₃Y} coordination spheres (Fig. 2). (a) and (b) the monofunctional adduct (1,1/c,c-MF-1; see Fig. 2(a)); (c) and (d) the bifunctional adduct (1,1/c,c-BF-1; see Fig. 2(c)).

For the second model (1,1/c,c-MF-2; Fig. 2(b)), the (CH₂)₆ chain has a similar elongated orientation to that in the 1,1/t,t adduct reported previously, so that there is no interaction between the two {PtN₃Y} coordination spheres and the am(m)ine protons form analogous H-bonds to the coordinated DHPA molecule (Fig. 4(a)).¹⁰ The cis-NH₃ group has a short H-bond to the phosphate oxygen atom (distance = 1.95 Å), whilst the NH₂ group (analogous to one of the cis-NH₃ groups in 1,1/t,t) has a short H-bond (distance = 1.97 Å) to the carbonyl oxygen from DHPA aliphatic chain a. The trans-NH₃ group is analogous to the NH₂ group in 1,1/t,t and similarly is not involved in any H-bonding interaction. These features are consistent with the ¹H shifts in the [¹H,¹⁵N] HSQC NMR spectra (Fig. 1) where peaks (C_m) in the cis-NH₃ and Pt−NH₂ regions are strongly deshielded, whereas the trans-NH₃ peak is slightly shielded compared to the parent chlorido species.

Figure 4.

Hydrogen bonding interactions in the DFT optimized DHPA bound adducts of 1,1/c,c (1) in which there is no interaction between the two {PtN₃Y} coordination spheres (Fig. 2). (a) and (b) the monofunctional adduct (1,1/c,c-MF-2; see Fig. 2(b)); (c) and (d) the bifunctional adduct (1,1/c,c-BF-2; see Fig. 2(d)).

The NMR data provide evidence for different conformers of the bifunctional adduct in which there is either an interaction, or no interaction between the two {PtN₃O} groups. The DFT optimized model (1,1/c,c-BF-1) shown in Fig. 2(c) (and Fig. 3(c and d)) has a hydrogen bonding interaction between the phosphate groups of the two bound DHPA molecules and is consistent with the strongly deshielded Pt-am(m)ine protons (peaks D in Fig. 1(c)). The positioning of the two Pt coordination spheres in close proximity enables a short H-bond (distance = 1.85 Å) between one of the Pt−NH₂ groups and the phosphate oxygen of the DHPA coordinated to the other Pt atom, and a second H-bond (distance = 2.08 Å) to the carbonyl oxygen on aliphatic chain a of the coordinated DHPA. The second NH₂ group has a single H-bond (distance = 2.44 Å) to the POCH₂ oxygen atom of the coordinated DHPA molecule. Both trans-NH₃ groups have short H-bonds (1.92 Å and 2.05 Å) to the carbonyl oxygen (of chain b or chain a, respectively) of the coordinated DHPA molecules. One of the cis-NH₃ groups has a single H-bond (distance = 2.23 Å) to the phosphate oxygen atom of the coordinated DHPA, whilst the other has a longer H-bond to phosphate oxygen (distance = 2.84 Å) and an additional H-bond (distance = 2.02 Å) to the carbonyl oxygen of chain b of the coordinated DHPA.

In the case where there is no interaction between the two {PtN₃Y} moieties, the H-bonding interactions between the Pt-a(m)mine units and the coordinated DHPA molecules are expected to be similar to that of the monofunctional adduct (coincident peaks C_m/C_b in Fig. 1). The DFT model of the bifunctional adduct (1,1/c,c-BF-2) shown in Fig. 2(d) (and Fig. 4(c and d)), is largely consistent with the NMR results, except there are short H-bonds between aliphatic chain carbonyl oxygen atoms and both trans-NH₃ groups. Whilst there are no similar H-bonds in the 1,1/c,c-MF-2 model (Fig. 2(b)), they are found in the 1,1/c,c-MF-1 model (see above) and thus it is reasonable to assume that NMR peaks for the different adducts will be strongly overlapped.

Kinetics

The time dependence of the concentrations of species in the reaction (1:5 ¹⁵N-1:DHPA) are shown in Fig. 5. Notable differences between the reaction of ¹⁵N-1 with 1,1/t,t are the much greater proportion of DHPA adduct formed in the present case and the slower attainment of equilibrium.¹⁰ For ¹⁵N-1 only 9% of the total Pt remained coordinated to chloride at equilibrium (120 h) compared to 35% for 1,1/t,t after 35 h.¹⁰

Figure 5.

Plot of the time dependence of species observed in the reaction of ¹⁵N-1 (1,1/c,c) with DHPA (1: 5 ratio) in 100 mM NaClO₄. The curves are computer best fits to the model shown in Scheme 1. The concentrations are based on the ¹H,¹⁵N peaks in the cis-NH₃ region (Fig. 1) and refer to the {PtN₃Y} end groups so that the initial concentration of A (Pt-Cl) is twice that of ¹⁵N-1. The rate constants are provided in Table 2. Key: A open squares; B open circles; C open triangles; D open diamonds; E asterisk; released chloride x.

To allow comparison with the reactions with 1,1/t,t, the same kinetic model (Scheme 1) that was previously reported¹⁰ was used for the reaction with ¹⁵N-1, with the concentrations of species based on the ¹H,¹⁵N peaks in the cis-NH₃ region, so that species C is derived from both monofunctional and bifunctional adducts. The curves are computer best fits and the rate constants are tabulated in Table 2.

Scheme 1.

Pathways for the reaction of 1,1/c,c (1) with DHPA, from analysis of the time dependence of peaks in the [¹H,¹⁵N] HSQC NMR spectra (Fig. 1). A simplified kinetic model was used for the equivalent {PtN₃Y} end groups in different species that were indistinguishable. The species “D” corresponds only to bifunctional adducts, whereas species “C” is derived from both monofuctional and bifunctional adducts (see text). “E” corresponds to a minor product from the reaction of ¹⁵N-1 with a minor hydrolysis product of DHPA¹⁰ and accounts for 9% of the total species at equilibrium (Fig. 5). The derived rate constants are shown in Table 2.

Table 2.

Rate and equilibrium constants for the reaction of 1 with DHPA in 100 mM NaClO₄ at 298 K, pH 5.6 in comparison to phosphate and DHPA reactions of 1,1/t,t and 1,0,1/t,t,t.^a

	1,1/c,c (1)		1,1/t,t
Parameter	DHPA	phosphateb	DHPAc	phosphateb
k1(10−5.s−1)	2.07 ± 0.05	2.18 ± 0.03	2.64 ± 0.10	2.49 ± 0.04
k-1 (M−1.s−1)	0.13 ± 0.01	0.145 ± 0.005	0.197 ± 0.020	0.40 ± 0.01
kL (M−1.s−1)	0.045 ± 0.002	0.0071 ± 0.0001	0.075 ± 0.010	0.0086 ± 0.0002
k-L(10−5.s−1)	0.15 ± 0.02	0.70 ± 0.02	5.68 ± 1.03	3.9 ± 0.1
ko (M−1.s−1)	0.0060 ± 0.0005		0.0035 ± 0.0009
k-o (10−5.s−1)	0.23 ± 0.04		0.95 ± 0.35
kBF (10−5.s−1)	2.48 ± 0.45		3.42 ± 0.32
k-BF(10−5.s−1)	8.41 ± 1.58		2.35 ± 0.26
pK1	3.81 ± 0.05	3.82 ± 0.02	3.87 ± 0.06	4.21 ± 0.02
pKL	−4.49 ± 0.07	−3.01 ± 0.02	−3.12 ± 0.14	−2.34 ± 0.02
pKo	−3.43 ± 0.11		−2.56 ± 0.32
pKBF	0.53 ± 0.17		−0.16 ± 0.09

^aThe rate constants for reaction with DHPA were derived from the kinetic model in Scheme 1.

^bValues for phosphate taken from ref.⁶ for the related simplified kinetic model.

^cValues for the DHPA binding of 1,1/t,t are taken from ref.¹⁰.

We have used previously [¹H,¹⁵N] HSQC NMR to investigate also the aquation of ¹⁵N-1 and 1,1/t,t in the presence of phosphate, under similar conditions (pH 5.3–5.9, 298K) to those used here in the reaction with DHPA.⁶ The rate and equilibrium constants for these reactions, obtained from the simplified kinetic model (analogous to that in Scheme 1) are listed also in Table 2.

For reactions in phosphate, equilibrium conditions for 1,1/t,t are established far more rapidly than for cisplatin,¹¹ due to the higher anation rate constant keeping the equilibrium more to the chlorido form. An important feature of the reaction of 1,1/t,t in phosphate is the high value for the rate constant for the reverse reaction (k_-L); comparison of pK₁ and pK_L in Table 2 demonstrate the clear preference for chloride over phosphate. For 1,1/t,t the equilibrium in phosphate was attained after 12 h and strongly favoured the dichlorido species. For 1,1/c,c, equilibrium conditions in the presence of phosphate are reached much more slowly (> 80 h) than for 1,1/t,t due to the slower (6-fold) aqua displacement of the bound phosphate – resulting in a greater proportion of phosphate-bound species at equilibrium.⁶

Comparison of the rate constants (Table 2) shows that the rate constant for the first step of DHPA binding (k_L) is slightly lower (1.6 fold) for the cis- compared to the trans-isomer, whereas the rate constants for the reverse reaction is 4-fold lower for 1,1/c,c. For 1,1/t,t the rate constant for the first step of DHPA binding (k_L) is about 8 times higher than that for phosphate binding, but the rate constants for the reverse reactions are quite similar so that equilibrium conditions are attained more slowly. On the hand for 1,1/c,c, k_L is 6-fold higher for DHPA compared to phosphate, whist k_-L is 5-fold lower, so that overall equilibrium conditions are attained after a similar time (120 h). The reduced lability of the bound phosphate was attributed to the ability to form a macrochelate-phosphate bridged species.⁶ Whilst a comparable species is not observed in the present case the formation of mono-(1,1/c,c-MF-1) and bifunctional (1,1/c,c-BF-1) DHPA adducts in which there is an interaction between the two {PtN₃Y} groups (a feature only accessible for the cis geometric isomer) provides an explanation for the slower release of the bound DHPA molecules.

Conclusions

These results confirm the relevance of phospholipid interactions for a full description of the biology of platinum antitumour agents. There are distinct differences between the pair of geometric 1,1/t,t and 1,1/c,c isomers, with equilibrium conditions being reached much more slowly for the latter. Further, the comparison of the forward and reverse rate constants show that in the 1,1/c,c case there is a much greater proportion of DHPA-bound species at equilibrium. These geometric differences may also extrapolate to the mononuclear cis and trans-[PtCl₂(NH₃)₂] pair. While significant work has been done on DNA interactions to explain the lack of meaningful anticancer activity of the trans isomer, much less comparative information is available for the mononuclear pair with other relevant biomolecules such as phospholipids. Since the 1,1/c,c compound contains the same geometrical configuration as cisplatin – the presence of cis-{Pt(NH₃)₂} coordination spheres -the interactions observed here may be extrapolated to possible weak interactions of cisplatin itself. As we have observed, the cellular localization and concentration of biomolecules such as phospholipids or heparan sulfate, which are predominantly oxygen-donors, does not preclude loose association of small molecules such as cisplatin, and these cellular entities may better describe the immediate fate of cisplatin, rather than existing as simple aquated cations.

Unlike the mononuclear pair, both the 1,1/t,t and 1,1/c,c isomers are antitumour active. Cellular accumulation in the murine leukaemia L1210/0 observes the relative order 1,0,1/t,t,t > 1,1/c,c > cisplatin > 1,1/t,t, but with overall similar levels of accumulation. In the resistant cell line L1210/DDP accumulation of cisplatin and the two dinuclear complexes, is diminished whereas accumulation of Triplatin is essentially equivalent.¹² Previous studies also indicated that the 1,1/c,c isomer was less collaterally sensitive to cisplatin (that is the 1,1/c,c was less able to overcome cisplatin resistance).¹² Phospholipids are one of the major components of cell membranes and diminished accumulation of the dinuclear compounds may reflect their sequestering by membrane components such as phospholipids. In this respect, the previously observed “non-covalent” interactions observed with the 4+ Triplatin compared to the dinuclear pair may contribute to the enhanced internalization and greater ability to overcome cisplatin resistance, given that reduced cellular accumulation is a significant contributor to cisplatin resistance.

The amphipathic nature of PPCs where lipophilic diamine chains link hydrophilic Pt-am(m)ine coordination spheres is remarkably complementary for strong interaction with the amphiphilic phospholipids. This complementarity is not available to the smaller cisplatin, which may engage in weak covalent phosphate interactions, but has no intrinsic lipophilic component. The effectiveness of the incorporation of cisplatin, carboplatin and oxaliplatin into liposomes to enhance biodistribution and lower toxicity may be diminished in the absence of this complementarity, which may affect release rates of the drug.^13–15 In general, the ability to systematically target membrane biomolecules suggest approaches to drug design that can play a significant role in dictating the overall biology of platinum drugs and have functional consequences contributing to the overall cytotoxic profile of platinum agents.

Experimental

Materials

1,2-dihexanoyl-sn-glycero-3-phosphate (monosodium salt) (DHPA) was purchased from Avanti. The nitrate salt of ¹⁵N labelled [{cis-PtCl(NH₃)₂}₂(μ-NH₂(CH₂)₆NH₂)]²⁺ (1,1/c,c; ¹⁵N-1) was prepared from cis-[PtCl₂(¹⁵NH₃)₂] and AgNO₃ using procedures similar to those previously published.¹⁶

Sample Preparation

15 mM stock solution of phospholipid was prepared by dissolving DHPA (10 mg, 0.026 mmol) in double distilled H₂O (1.71 mL) and adjustment of the pH to 5.4 using 0.1–1 M NaOH or HClO₄.

Reaction of 1,1/c,c (¹⁵N-1) with DHPA (1:5 ratio)

Freshly prepared solution of ¹⁵N-labelled 1,1/c,c (¹⁵N-1) (0.35 mg, 0.45 μmol) in H₂O (30 μL) was combined with 15 mM DHPA stock solution (160 μL), 200 mM NaClO₄ (240 μl), H₂O (54 μL), D₂O (24 μl) and 10 mM TSP (2 μL) to give a total volume of 480 μl and final concentrations of 1,1/c,c (1 mM) and DHPA (5 mM) in 100 mM NaClO₄. The starting pH was 5.6.

NMR Spectroscopy

The NMR spectra were recorded on a Bruker 600 MHz spectrometer (¹H, 600.1 MHz; ¹⁵N, 60.8 MHz; ³¹P, 242.9; ¹⁹⁵Pt, 129.0 MHz) fitted with a pulsed field gradient module and 5 mm triple resonance probehead. The ¹H NMR chemical shifts are internally referenced to TSP (sodium 3-trimethylsilyl-[D4]-propionate) and the ¹⁵N chemical shifts externally referenced to ¹⁵NH₄Cl (1.0 M in 1.0 M HCl in 5% D₂O in H₂O). The ¹H spectra were acquired with water suppression using the watergate 3–9-19 pulse sequence.^{17, 18} The two-dimensional [¹H,¹⁵N] heteronuclear single-quantum coherence (HSQC) NMR spectra optimized for ¹J(¹⁵N,¹H) = 72 Hz were recorded using standard Bruker phase sensitive HSQC pulse sequence.¹⁹ The ¹⁵N signals were decoupled by irradiating with the GARP-1 sequence at field strength of 6.9 kHz during the acquisition time. Typically for 1D ¹H spectra, 32 scans and 32 K points were acquired using a spectral width of 12 kHz and a relaxation delay of 2.5 s. For kinetics studies involving [¹H, ¹⁵N] HSQC NMR spectra, 4 transients were collected for 96 increments of t₁ (allowing spectra to be recorded on a suitable timescale for the observed reaction), with an acquisition time of 0.069 s, spectral widths of 6 kHz in f₂ (¹H) and 5.5 kHz in f₁ (¹⁵N). 2D spectra were completed in 14 minutes. The 2D spectra were processed using zero-filling up to the next power of 2 in both f₂ and f₁ dimensions.

NMR spectra were recorded at 298 K, and the sample was maintained at this temperature when not immersed in the NMR probe. NMR sample (including buffers, acids, etc.) was prepared so that there was a 5% D₂O / 95% H₂O concentration (for deuterium lock but with minimal loss of signal as a result of deuterium exchange). The pH of the solution was measured on a Shindengen pH Boy-P2 (su19A) pH meter and calibrated against pH buffers of pH 6.9 and 4.0. The volume of 5.0 μL of the solution was placed on the electrode surface and the pH recorded. These aliquots were not returned to the bulk solution (as the electrode leaches Cl^-). Adjustments in pH were made using 0.1 M and 1.0 M HClO₄ in 5% D₂O in H₂O, or 0.1 M and 1.0 M NaOH in 5% D₂O in H₂O.

Data Analysis

The kinetic analysis of the reactions was undertaken by measuring peak volumes in the cis-NH₃ region of the [¹H,¹⁵N] HSQC NMR spectra using the Bruker XWINNMR software package and calculating the relative concentrations of the various species at each time point, in the same manner as described previously for the reactions of 1,1/t,t and 1,1/c,c with phosphate.⁶ Peak volumes were determined using an identical vertical scale and threshold value. In order to simplify the system, we used a simplified model in which the species derived from ¹⁵N-1 are treated as mononuclear [cis-Pt(NH₃)₂(NH₂(CH₂)₆NH₂)Y]ⁿ⁺. In this case the total volume for each peak accounted for the total concentration of Pt-Cl (A), Pt-OH₂ (B) or ligated (DHPA adducts C or D) species (see Scheme 1). The peak volumes were normalized and the concentration of each species was determined for each spectrum based on the initial concentration of the parent complex. The data were then subjected to kinetic analysis. The appropriate differential equations were integrated numerically and rate constants determined by a non-linear optimization procedure using the program SCIENTIST (Version 2.01, Micromath Inc.), as we have described previously.⁶ The errors quoted represent one standard deviation. The kinetic models for all reactions are depicted and the rate and equilibrium constants are defined in Scheme 1.

DFT Calculations

All DFT computations were performed using the Gaussian 16 suite of programs²⁰ either on the Griffith University’s HPC cluster, Gowonda, or on clusters at the National Computing Infrastructure (NCI). The hybrid version of the parameter-free PBE functional, PBE0, which demonstrated excellent performance for geometry optimization of transition metal complexes, was used for these computations. Full geometry optimizations without any constraints were performed with 6–31G(2d,2p) for C, H, N, P, O and Hay and Wadt effective core potential (LanL2DZ) basis sets for Pt, with diffuse functions only on oxygen. All systems were solvated using the IEFPCM solvation model with water as solvent. NMR data suggested that 1,1/c,c forms two mono and two bifunctional adducts with DHPA. The starting structure for the first mono adduct (1,1/c,c-MF-1) was analogous to the previously reported aquaphosphato model of 1,1/c,c,⁶ whilst the starting structure for the second (1,1/c,c-MF-2) was constructed in a similar manner to that of the previously reported 1,1/t,t monofunctional adduct.¹⁰ For the first bifunctional adduct (1,1/c,c-BF-1), DHPA fragments at both ends in the starting structure were constructed in a cis orientation to each other. The starting structure for the second bifunctional adduct (1,1/c,c-BF-2) was constructed by introducing a DHPA molecule at the {PtN₃Cl} end of the DFT optimized structure of 1,1/c,c-MF-1, in a similar orientation to that of the coordinated DHPA molecule. All the systems were pre-optimized using the semi-empirical method, PM6, MOPAC²¹ and the resulting geometry was then used as the input for DFT calculations.