Luminescent cyclometalated platinum(II) complexes containing N-heterocyclic carbene ligands with potent in vitro and in vivo anti-cancer properties accumulate in cytoplasmic structures of cancer cells

Abstract

Contrary to most platinum-based anti-cancer agents which target DNA, coordination of N-heterocyclic carbene (NHC) ligands to cyclometalated platinum(II) complexes confers these luminescent complexes to other cellular target(s). The strong Pt–C_carbene bond(s) renders the platinum(II) complexes to display unique photophysical properties and enhanced stability against biological reduction and ligand exchange reactions. The platinum complexes described in this work are highly cytotoxic and display high specificity to cancerous cells. Among them, [(C^N^N)Pt^II(N,N′-ⁿBu₂NHC)]PF₆ (1a, where HC^N^N = 6-phenyl-2,2′-bipyridine) with a lipophilic carbon chain on the carbene ligand induces apoptosis in cancer cells, demonstrates an enhancing synergistic effect with cisplatin in vitro, and displays potent in vivo activities using nude mice models. As this complex is strongly emissive, its cellular localization can be traced using emission microscopy. In contrast to common platinum-based anti-cancer agents, 1a does not accumulate in the vicinity of DNA but preferentially accumulates in cytoplasmic structures including sites where active survivin, an inhibitor of apoptosis (IAP), is located. In vitro, 1a significantly inhibits the expression of survivin, activates poly(ADP-ribose) polymerase (PARP) and induces apoptosis in cancer cells. Given the ease of structural modification of NHC ligand to alter the overall biological activities, these [(C^N^N)Pt^II(NHC)]⁺ complexes having unique photophysical properties provide an entry to a new class of potential anti-cancer drug leads.

---

Introduction

Since the reports on the preparation and isolation of metal N-heterocyclic carbene (NHC) complexes derived from azolium compounds respectively in 1962 and in 1991,¹ various NHC complexes including that of platinum, palladium, gold, rhodium, silver and copper have been reported to display extraordinary practical applicability especially in the area of catalysis.² Metal complexes of NHC ligands are reminiscent of metal-phosphine complexes,³ but they offer additional advantages including high thermal stability and stability towards air and moisture,⁴ and the coordination of carbene ligand to metal ion can be achieved under mild experimental conditions.⁵ Additionally, functionality and lipophilicity of the carbene ligands could be systematically modified. In literature, studies on metal-NHC complexes have mainly been focused on their catalytic properties. Yet, their biological activities and therapeutic properties, particularly that of platinum-NHC complexes,^6a remain largely unexplored.⁶

The discovery of cisplatin has stimulated worldwide efforts to develop new anti-cancer platinum complexes.⁷ Cisplatin (Platinol), carboplatin (Paraplatin), and recently oxaliplatin have received worldwide approval for clinical uses.⁸ However, clinical studies revealed that the elevated level of biological reductant glutathione (GSH) in some cancer cells could deactivate these platinum complexes subsequently leading to the generation of drug resistance.^9a To circumvent this problem, it is of paramount importance to identify new compounds which are stable and display different modes of anti-cancer action. Cyclometalated platinum(II) complexes containing tridentate π-conjugated organic ligands have been receiving a surge of interest for their application in light-emitting devices and chemical sensors.¹⁰ These complexes display rich and diverse photoluminescent properties that are sensitively affected by local medium. The planar motifs of these platinum(II) complexes could insert between two adjacent DNA base pairs through non-covalent ligand–ligand π–π stacking interactions, thus rendering them as DNA metallointercalators instead of covalently cross-linking to the DNA base pairs.¹¹ Extensive studies have revealed that various platinum(II) intercalators display promising in vitro and in vivo anti-cancer activities.¹² To synergize the advantages of the NHC ligands (ease in structural modification, capability of forming stable metal complexes and their relative non-toxic nature) and of the cyclometalated platinum(II) complexes (luminescent properties and DNA intercalation), we developed [(C^N^N)Pt^II(NHC)]⁺ complexes which display intriguing photoluminescent properties as well as anti-cancer activities.

Results

Synthesis, characterization and photophysical properties

The [(C^N^N)Pt^II(NHC)]⁺ complexes (Fig. 1) were prepared from precursors P1–P10 having a dynamic range of lipophilicity (Fig. S1 and Table S1, Electronic Supporting Information†).¹³ These complexes were characterized by FAB-MS, ¹H NMR and ¹³C NMR spectroscopies, and elemental analyses. They display a good solubility (>10 mg mL⁻¹) in dichloromethane (CH₂Cl₂), methanol (CH₃OH), acetonitrile (CH₃CN), and dimethyl sulfoxide (DMSO), but are less soluble (<0.5 mg mL⁻¹) in water. They are stable at room temperature for at least one month as confirmed by NMR spectroscopy. Using 1a as an example, we examined the stability of these complexes against reduction/substitution by glutathione (GSH) by means of UV-visible spectrophotometry. We found that treating 1a with GSH at 2 mM in PBS/DMSO (19 : 1) did not cause significant spectral changes (<5%) upon standing the solution for 72 h (Fig. S2, ESI†). In contrast, cisplatin has been reported to be highly unstable in the presence of glutathione.⁹

Fig. 1 [(C^N^N)Pt^II(NHC)]⁺ complexes.

X-Ray crystallography

Crystals of 1a, 1c and 1f (Fig. 2) were obtained by slow diffusion of diethyl ether into CH₃CN solutions, and crystallographic data of 1a, 1c and 1f (Table S2, ESI) and bond angles/distances (Table S3, ESI) are given in the Supporting Information.† A distorted square-planar geometry is revealed from the trans C–Pt–N angles of 160.5(3)° in 1a and of 159.5(3)° in 1c. The NHC ligands in both 1a and 1c are perpendicular to the [(C^N^N)Pt^II]⁺ plane with an angle of 93.0°. Intermolecular Pt–Pt distances are 8.87 and 8.65 Å for 1a and 1c, respectively. Complex 1f has two NHC planes at torsion angles of 103.17° and 107.81° from the C^N^N planes. This complex displays intramolecular interaction between the two [(C^N^N)Pt^II]⁺ planes as revealed by the Pt–Pt distance of 3.536 Å and the π–π separation between two [(C^N^N)Pt^II]⁺ planes of 3.435 and 3.495 Å (Fig. S3, ESI†). The Pt–C_carbene distances in 1a and 1c are 1.984(7) and 2.016(9) Å, respectively, while 1f has two Pt–C_carbene distances of 1.968(11) and 1.974(12) Å.

Fig. 2 Perspective views of the X-ray crystal structures of the cations of 1a (upper left), 1c (upper right), and 1f (bottom).

Variable-temperature ¹H NMR spectroscopic measurements

As revealed from X-ray crystallography, the NHC ligand spatially restrains the mononuclear [(C^N^N)Pt^II(NHC)]⁺ complexes from intermolecular π–π and Pt–Pt interactions. These interactions are crucial to the photophysical properties and notably the luminescent properties of these platinum(II) complexes. The ¹H-NMR spectra of 1a (Fig. S4, ESI) and 1c (Fig. S5, ESI) in CD₃CN at temperature between 233 K and 333 K are given in the Supporting Information.† In brief, the spectra reveal no significant changes in both structure and intermolecular interaction(s). Conversely, the ¹H-NMR spectrum of 1f in CD₃CN is temperature-dependent (Fig. S6, ESI†). For an example, there is a cluster of NMR signals at δ = 6.50 at 333 K, whereas these signals are resolved into three sets of doublet when the temperature is lowered to 233 K.

Absorption and emission spectroscopy

The UV-visible spectral data of [(C^N^N)Pt^II(NHC)]⁺ complexes are listed in Table 1. The absorption spectra of the mononuclear complexes 1a–1d (Fig. 3, upper) show intense intraligand transitions (IL) in the 245–375 nm spectral region, and singlet metal-to-ligand charge transfer (¹MLCT) transitions Pt(5d) → C^N^N (π^*) at 375–400 nm.^14a When comparing these absorption spectra with that of [(C^N^N)Pt^IICl], substitution of the Cl⁻ by NHC ligand leads to blue-shift of ¹MLCT transition(s). For complexes 1b, 2a and 3a, all of which have the same N,N′-ⁿPr₂NHC ligand, the absorption spectra of 2a and 3a are red shifted from that of 1b (Fig. S7, ESI†). It should be noted that although 1e is a binuclear complex, its absorption spectrum (Fig. 3, lower) is similar to those of the mononuclear complexes 1a–1d. However, the absorption spectrum of the binuclear complex 1f, which has a shorter bridging ligand, is red shifted from that of 1a–1d at λ > 370 nm.

Table 1 UV-visible absorption data of 1a–1f, 2a, 2b, 3a and 3b (2 × 10⁻⁵ mol dm⁻³ in CH₃CN)

Complex	λ abs/nm (ε/dm^3 mol^−1 cm^−1)
1a	270 (60300), 337 (12700), 354 (11600), 368 (5120), 392 (2730), 430 (1520), 485 (1030)
1b	255 (49500), 266 (52900), 316 (17200), 330 (19000), 337 (20700), 353 (19100), 369 (9060), 392 (3720), 484 (330)
1c	251 (38600), 255 (39800), 266 (43300), 312 (13000), 335 (15000), 352 (13700), 394 (2600), 429 (1250), 485 (750)
1d	252 (34900), 255 (35900), 266 (37600), 312 (11800), 337 (13600), 352 (12600), 392 (2480), 429 (1220), 485 (640)
1e	254 (49600), 265 (48400), 334 (18200), 354 (15800), 384 (5750), 418 (3540), 485 (1060)
1f	255 (38100), 319 (16300), 331 (16600), 355 (9480), 364 (5570), 380 (6210), 420 (3530), 464 (1680), 485 (670)
2a	245 (30500), 287 (36000), 358 (14700), 425 (2260), 471 (1290)
2b	283 (33300), 322 (16700), 356 (10300), 420 (2850), 487 (1330)
3a	241 (20100), 264 (19400), 287 (21900), 319 (14800), 355 (8100), 395 (2090), 433 (870)
3b	246 (52700), 287 (40300), 320 (25400), 355 (13900), 376 (5850)

---

Fig. 3 UV-Visible absorption spectra of 1a–1d (upper) and 1e–1f (lower, 2 × 10⁻⁵ mol dm⁻³) in CH₃CN at 298 K.

Using 1c as an example, solvatochromism of the ¹MLCT absorption band of the mononuclear [(C^N^N)Pt^II(NHC)]⁺ complexes at around 350 nm in different solvents is observed, and details of the spectral data are given in the Supporting Information (Table S4 and Fig. S8†).

The mononuclear complexes, 1a–1d, 2a and 3a, are emissive in solid state and in degassed CH₃CN with emission λ_max at 545 to 546 nm, emission lifetime (τ₀) of 0.6 to 2.9 μs, and emission quantum yield (Φ) of 0.087 to 0.23 (Table 2). Complexes 1a (Fig. 4) and 1b–1d show well-resolved vibronic structured emission bands (λ_ex = 340 nm) with emission λ_max being insensitive to the complex concentration. For both the mononuclear complexes 1a and 1c, there is no significant change in the emission energies with complex concentration in the range of 1 × 10⁻⁶ to 1 × 10⁻⁴ mol dm⁻³ (Fig. S9, ESI†). There is also no significant change in emission energy upon changing the solvent polarity, while both τ₀ and Φ decrease as the polarity of the solvent increases (non-emissive in DMF).

Table 2 Emission data of 1a–1f, 2a, 2b, 3a and 3b

Complex	Degassed solution (CH3CN; 298 K) λem [nm] (τ0 [μs]); Φ (quantum yield)^a	Solid state emission (298 K) λem [nm] (τ0 [μs]) ^b	Solid state emission (77 K) λem [nm] (τ0 [μs]) ^b	Glassy emission (77 K) λem [nm] (τ0 [μs]) ^c
a Complexes 1a–1f, 2a, 2b, 3a and 3b were excited at 340 nm. b The emission data were measured by excitation at 350 nm. c The glassy emission data were measured in a concentration of 2 × 10^−5 mol dm^−3 in a MeOH–EtOH–DMF mixture (5 : 5 : 1, v/v).
1a	545 (1.2); 0.23	511, 538 (max, 0.7), 569	524 (max, 7.4), 560	516 (max, 75), 552, 595
1b	546 (1.1); 0.19	527, 553 (max, 0.7), 590	542 (max, 5.0), 581	515 (max, 71), 552, 590
1c	545 (0.9); 0.12	510 (max, 0.6), 542, 582	520 (max, 6.2), 540, 552, 590	515 (max, 64), 551, 595
1d	546 (0.6); 0.087	596 (max, 0.6), 670	564 (max, 5.7), 607, 669	514 (max, 74), 552, 591
1e	546 (0.8); 0.051	574 (1.4)	535 (max, 7.5), 578	456, 487, 516 (max, 59), 549
1f	619 (1.3); 0.056	592 (2.1)	572, 615 (max, 8.2)	525, 617 (max, 251)
2a	546 (1.2); 0.13	557 (1.9)	548 (max, 6.9), 568	521 (max, 112), 557, 602
2b	542, 610 (max, 0.8), 660; 0.031	620 (0.9)	637 (4.6)	511 (max, 232), 539, 620
3a	546 (2.9); 0.11	554 (2.1)	540 (max, 5.8), 580	518 (max, 108),554, 596
3b	540 (max, 0.6), 631, 663; 0.032	605 (0.8)	485 (max, 3.9), 566, 670	513 (max, 101), 547. 583

---

Fig. 4 Emission spectra of 1a in CH₃CN at 298 K and 77 K (2 × 10⁻⁵ mol dm⁻³), λ_ex = 340 nm (normalized intensities).

The binuclear complexes 1e, 1f, 2b and 3b are emissive both in solid state and in degassed CH₃CN. (Table 2). At room temperature, the emission λ_max of complexes 1f, 2b and 3b in solutions, all of which have a C₁ spacer (methyl linker) on the bridging carbene ligand, are significantly red-shifted compared to that of the mononuclear complexes 1b, 2a and 3a, respectively. In contrast, complex 1e which has a C₃ spacer (propyl linker) displays photophysical properties similar to that of the mononuclear counterparts in the context of emission energies and τ₀ in solutions. In solid state, 1f shows a broad emission band at 594 nm with a lifetime of 2.1 μs at room temperature. Upon cooling to 77 K, the band width of the emission decreases and the emission is resolved into two peak maxima at 572 nm and 615 nm. The emission λ_max of 1f (Fig. S10, ESI†) in both solid state (λ_em = 592 nm) and in CH₃CN (λ_em = 619 nm) are significantly red-shifted from its mononuclear counterpart 1b (solid: λ_em = 553 nm, CH₃CN: λ_em = 546 nm, Fig. S11, ESI†). Regarding the solvent effect, there is no significant change in the emission energy upon changing the solvent polarity from CH₃CN to CH₃OH, acetone, CH₂Cl₂ and tetrahydrofuran (Fig. S12, ESI†), but the emission lifetime at 617 nm varies from 1.6 μs to 0.2 μs and emission quantum yield decreases from 0.074 to 0.014 when the solvent is changed from CH₂Cl₂ to CH₃OH. Such solvatochromic behavior of the metal-metal-to-ligand charge transfer excited state (³MMLCT) emissions of the dinuclear Pt^II complexes is similar to that of the ³MLCT emission of the mononuclear complex 1c.^14b The emission of 3b in CH₃CN is at an energy similar to that of 3a (3a: 546 nm and 3b: 540 nm; Table 2).

The emissions of 1a, 1f, 2a–2b and 3a–3b in frozen CH₃CN solutions at 77 K have been studied. The emissions of the complexes 1a (Fig. 4), 2a (Fig. S13, ESI†) and 3a (Fig. S14, ESI†) in 77 K CH₃CN solutions are insensitive to the complex concentration from 10⁻⁶ to 10⁻⁴ mol dm⁻³. Vibronically structured emission bands with λ_em 527–570 nm and vibrational spacing of ca. 1200 cm⁻¹ attributed to the skeletal stretching of the tridentate HC^N^N ligand were recorded. The emission λ_max of the binuclear complex 1f in 77 K CH₃CN solution is at 634 nm which is red-shifted from that at 298 K (Fig. S15, ESI†); another binuclear complex 2b also displays red-shifted emission band at 625 nm (cf., RT: 610 nm) in frozen CH₃CN (Fig. S16, ESI†). The emission properties of all of the [(C^N^N)Pt^II(NHC)]⁺ complexes in glassy solutions (MeOH–EtOH–DMF = 5 : 5 : 1) at 77 K have also been studied (Table 1). The emission of 1f is sensitive to the complex concentration (10⁻⁶–10⁻⁴ mol dm⁻³, Fig. S17, ESI†). This complex displays red-shifted emission band with λ_max at 621 nm, 617 nm and 612 nm in glassy solution at complex concentration of 10⁻⁴ mol dm⁻³, 10⁻⁵ mol dm⁻³ and 10⁻⁶ mol dm⁻³, respectively.

Cytotoxicity tests (MTT, (3-(4,5-dimethylthiazol-2-yl)-2,5-tetrazolium bromide) Assay)

Since the discovery of the clinically-used cisplatin for anti-cancer treatment, a variety of platinum(II) complexes have been identified to display promising anti-cancer activities.⁷ In this study, we examined the in vitro anti-cancer properties of [(C^N^N)Pt^II(NHC)]⁺ towards human cancer cell lines of cervical epithelioid carcinoma (HeLa), hepatocellular carcinoma (HepG2), and nasopharyngeal carcinoma (SUNE1), and a normal cell derived human lung fibroblast cell line (CCD-19Lu) by means of MTT assay.¹⁵ Using 1a as an example, the IC₅₀ values (dose required to inhibit 50% cellular growth) were determined from the dose-dependence of surviving cancer cells after cellular exposure to 1a for 48 h (Fig. S18, ESI†). Complex1awas found to display promising anti-cancer activity towards these three cancer cell lines (IC₅₀ = 0.057 – 0.77 μM), and specifically displays higher cytotoxic potency towards HeLa cells. Notably, this complex is less cytotoxic to the normal-derived human cell line (CCD-19Lu) with IC₅₀value of 11.6 μM, which is 232 folds higher than that towards HeLa cells. For comparison, the cytotoxicity of cisplatin was also examined. We found that 1a displays ∼300-fold higher potency toward HeLa than cisplatin. Cisplatin is also less cytotoxic toward CCD-19Lu cells, as its IC₅₀ value was found to be >100 μM. In addition, 1a was found to display a synergistic anti-cancer effect with cisplatin. With the co-incubation of cisplatin at fixed concentrations of 0.5 and 5 μM, the IC₅₀ value of 1a against HeLa cells shifted from 0.057 μM, to 0.035 and 0.021 μM, respectively.

The cytotoxicity and IC₅₀ values of other [(C^N^N)Pt^II(NHC)]⁺ complexes including 1b–1f were determined in a similar manner, and the results are summarized in Table 3. The [(C^N^N)Pt^II(NHC)]⁺ complexes are more cytotoxic than cisplatin; the IC₅₀ of 1c and 1d with NHC ligand having two N-CH₂CH₃ and N-CH₃ groups, respectively, are at least 31 and 45 fold higher than that of cisplatin towards HeLa. Although the two dinuclear complexes 1e and 1f both show higher potency than cisplatin, they are less cytotoxic than the mononuclear complexes 1a–1d (Table 3). Besides 1a, the cytotoxicities of other platinum complexes towards normal human lung fibroblast cell line of CCD-19Lu were also examined. All of these complexes were found to display lower cytotoxicity to this cell line.

Table 3 Cytotoxic IC₅₀ values (μM) of 1a–1f in human carcinoma cell lines of HeLa, HepG2, SUNE1 and in a normal derived human lung fibroblast cell line (CCD-19Lu)

Complex	HeLa	HepG2	SUNE1	CCD-19Lu
1a	0.057	0.77	0.14	11.6
1b	0.052	1.1	0.16	4.3
1c	0.48	1.3	0.32	2.1
1d	0.33	0.31	0.51	5.7
1e	3.9	7.1	5.6	27
1f	8.0	9.4	6.4	40
Cisplatin	15	15	2.4	>100

---

Using human hepatocellular carcinoma cell line (HepG2) as a model, complexes 1b, 2a and 3a were chosen for examining the effect of ligand variation on the cytotoxicity via increasing the lipophilicity of the tridentate C^N^N ligand. We found that 2a (IC₅₀ = 0.49 μM) and 3a (IC₅₀ = 0.27 μM) were more cytotoxic towards HepG2 cells when compared with 1b (IC₅₀ = 1.1 μM).

In addition to the cytotoxicity evaluation, the anti-angiogenic and anti-metastatic activities of 1a have been examined.¹⁶ By means of the tube-formation assay on MS1 cells and wound-healing assay on HeLa cells, there was no apparent anti-angiogenic (Fig. S19, ESI†) and anti-metastatic (Fig. S20, ESI†) activities of 1a at its sub-toxic concentration.

In vivo anti-cancer study

In addition to the promising in vitro anti-cancer activity, the in vivo anti-cancer activity of 1a was examined. The experiments were conducted at PearL Materia Medica Development (Shenzhen) Limited and performed with approval from the Committee on the Use of Live Animals for Teaching and Research. NCI-H460 non-small lung carcinoma cells were first implanted into nude mice. When the tumors were approximately 50 mm³ in size, nude mice were randomly separated into four groups to receive treatment of twice-a-week intraperitoneal injection of 20% PET vehicle control (20% PET = 12% polyethylene glycol 400; 6% ethanol; 2% Tween 20; 80% phosphate-buffered saline), 1a at 1 mg kg⁻¹, 1a at 3 mg kg⁻¹ or the positive control cyclophosphamide at 30 mg kg⁻¹. After 28 days, the mice were sacrificed and the tumors were taken out and their weights were measured. Results showed that injection of 1a at 3 mg kg⁻¹ and cyclophosphamide at 30 mg kg⁻¹ significantly inhibited the NCI-H460 tumor growth by 55 ± 11% (0.78 ± 0.19 g) and 56 ± 6% (0.76 ± 0.10 g), respectively (Fig. 5), whereas treatment of 1a at 1 mg kg⁻¹ was not effective. Importantly, 1a did not cause death of mice, and regular body-weight measurement showed that mice receiving 1a (32.6 ± 5.1 g) had no significant weight loss compared to that in the vehicle control group (34.5 ± 4.7 g).

Fig. 5 (A) Statistical representation of the tumor weights and (B) photographs showing the reduction of NCI-H460 tumor in size upon treatment of 1a in nude mice models.

Localization in cancer cells

In view of the favorable emission properties of the cyclometalated [(C^N^N)Pt^II(NHC)]⁺ complexes, HeLa cells treated with 1a (1 μM) for 1 h were imaged by fluorescence microscope (Fig. 6, left column).¹⁷ To ascertain the site of cellular localization, the 1a-treated cells were co-stained with other known fluorescent dyes including the mitotracker^TM (for mitochondria and cytoplasmic structure), Hoechst 3342 (for DNA) and lysotracker^TM (for lysosomes) (Fig. 6, middle column). We found that majority of 1a can be co-localized with mitotracker (Fig. 6, right column). In contrast, 1a did not co-localize with the DNA binder Hoechst 33342 and the lysosome binder lysotracker^TM.

Fig. 6 Fluorescent microscopic examination of 1a in the same batch of HeLa cells either co-incubated with mitotracker^TM (top), Hoechst 33342 (middle) or lysotracker^TM (bottom), showing that the majority of 1a can be co-localized with mitotracker.

Inhibition of survivin and activations of caspases and poly(ADP-ribose) polymerase

Survivin, an inhibitor of apoptosis (IAP), is selectively expressed in most human cancers and is associated with tumor progression in patients.¹⁸ Importantly, survivin expression has been reported at low or non-detectable levels in normal tissue, rendering this protein to serve as an important target for anti-cancer treatment.¹⁹ As described in the previous section, 1a was found to accumulate in the cytoplasmic structure of cancer cells where activated survivin is located. By means of a survivin enzyme immunometric assay,²⁰ treatment of 1a in HeLa cells time- (24 and 48 h) and dose- (0.1 and 1 μM) dependently inhibited the survivin expression (Fig. 7A). We further confirmed the expression of survivin by Western blotting (Fig. 7B). In concomitance with the survivin inhibition, 1a was able to activate caspase-3, and could activate poly(ADP-ribose) polymerase (PARP) in HeLa cells in a time- and dose-depending manner (Fig. 7C).

Fig. 7 Inhibition of the (A) activity and (B) expression of survivin, and (C) activation of caspase 3 and PARP-1 in HeLa cells treated with 1a.

Interaction with double-stranded DNA

Platination of DNA is widely believed to account for the anti-cancer effect of cisplatin as well as various platinum(II) complexes. Yet in our study, we found that 1a preferentially accumulates in the cytoplasmic structure of the cells instead of covalently binding to nuclear DNA (Fig. 6). In cell-free conditions, 1a was found to display poor binding affinity to double stranded DNA. The interaction between 1a and calf-thymus DNA (ctDNA) in PBS/DMSO (19 : 1) solution was examined by means of UV-visible absorption titration experiment. Isosbestic spectral changes and hyperchromicity (∼15%) were observed for the electronic transitions of the [(C^N^N)Pt^II]⁺ moiety at 280–320 nm upon addition of ctDNA to a solution of 1a (Fig. S21, ESI†). The binding constant (K_b) of 1a toward ctDNA was determined, from the plot of [ctDNA]/Δε_apversus [ctDNA], to be 4.8 × 10³ dm³ mol⁻¹.²¹

A gel-mobility-shift assay was employed to examine the intercalating property of 1a.²² A 100-bp DNA ladder treated with 1a or ethidium bromide (DNA intercalator) was resolved by agarose-gel electrophoresis (Fig. S22, ESI†). Only samples that contained EB (lanes B and C) or 1a (lanes D–G) exhibited a tailing effect, which can be accounted for by the elongation of DNA resulting from the intercalation of EB or 1a with DNA. In contrast, DNA treated with vehicle control (lanes A and H) did not reveal the apparent tailing effect. By viscosity analysis,²² we further confirmed that 1a could act as a DNA metallointercalator. Addition of either EB or 1a increases the viscosity of the DNA by increasing its hydrodynamic length (Fig. S23, ESI†). In contrast, Hoechst 33342 (minor groove binder) and the vehicle control failed to lengthen the DNA and did not cause any changes in DNA viscosity.

In addition to the interaction with double stranded DNA, the inhibitory activity of 1a on topoisomerase I (TopoI), a DNA binding protein which catalyzes topological changes in DNA by the formation of DNA strand breaks, was examined.²³ TopoI induced formation of nicked or relaxed supercoiled forms of tertiary DNA structure which could be inhibited in the presence of TopoI poison camptothecin (CPT, Fig. S24, ESI†). However, co-incubation of 1a at concentrations up to 1 μM with TopoI could not inhibit the protein activity since both the nicked- and relaxed forms of the supercoiled DNA were still detected.

Discussion

In this work, the cyclometalated platinum(II) complexes bearing N-heterocyclic carbene (NHC) ligand [(C^N^N)Pt^II(NHC)]⁺ are lipophilic cations having a range of lipophilicity attributed to the different substituents on NHC or [C^N^N]⁻ ligands. X-ray crystal structures revealed that the Pt–C_carbene distances of [(C^N^N)Pt^II(NHC)]⁺ complexes are shorter than that of Pt–L distances in related [Pt^II(C^N^N)L]ⁿ⁺ complexes (where L = halide, phosphine or amine).^14,24,26 The strongly coordinating tridentate C^N^N and carbene ligands account for the stability of the platinum(II) complexes against GSH reduction/substitution in solutions. The stability under physiological conditions could be a critical factor for new drug design since elevated cellular GSH level has been implicated in cisplatin-resistant cancer cells presumably through sequestration of cisplatin.⁹

Variable temperature-¹H NMR experiments showed that there is no close intermolecular interaction between the [(C^N^N)Pt^II(NHC)]⁺ cations at temperature from 233 K to 333 K. As revealed from the X-ray crystal structures of 1a and 1c, the orientation of the NHC ligand disfavors the approach of two [(C^N^N)Pt^II]⁺ planes in close proximity, accounting for the absence of intermolecular Pt–Pt and π–π interactions. Thus, these mononuclear complexes have a low tendency to aggregate in solutions and this is important as aggregation of the platinum(II) complexes could impede the entrance to cancerous cells. In contrast, the short bridging carbene ligand of 1f confines the two [(C^N^N)Pt^II]⁺ planes in close proximity rendering intramolecular Pt–Pt interactions feasible. Together with the changes in chemical shifts of ¹H NMR signals and emission properties at various temperature, we reckon that 1f displays fluxional intramolecular conformational change²⁵ with much closer intramolecular contact between two [(C^N^N)Pt^II]⁺ planes at low temperature, thus accounting for the different emission behaviors when compared to the mononuclear complexes 1a and 1c.

At λ < 400 nm, the absorption spectra of the [(C^N^N)Pt^II(NHC)]⁺ complexes are dominated by intense ¹IL transitions; there are less intense absorption bands with ε values of 300–3900 dm³ mol⁻¹cm⁻¹ in the 400–500 nm spectral region. It should be noted that [Pt^II(C^N^N)Cl], the parental precursor for the [(C^N^N)Pt^II(NHC)]⁺ complexes, displays a much lower molar absorption coefficient in the spectral region of 400–500 nm.²⁴ With reference to previous work on binuclear platinum(II) complexes bearing π-conjugated C- and N-donor ligands having metal–metal and ligand–ligand interactions,^24,26 the absorption bands of 1f and 2b at 410–490 nm are tentatively assigned to metal-metal-to-ligand charge transfer ¹MMLCT ¹[d_σ* → σ (π^*)] and intraligand ¹[σ*(π) → σ (π^*)] transitions.^24,26 Although 1e is a binuclear complex, its absorption spectrum is similar to those of the mononuclear complexes 1a–1d. Thus, the longer carbon chain of the bridging carbene ligand in 1e renders larger spatial separation between two [(C^N^N)Pt^II]⁺ units resulting in the latter behaving as two independent non-interacting cations.

With reference to previous work on degassed CH₃CN solutions of [(C^N^N)Pt^II(PPh₃)]⁺,²⁴ the structureless emission bands of 1a–1d at 545 nm (RT) are assigned to excited states with mixed ³MLCT and ³IL characters. Their emission quantum yields of 0.087 to 0.23 are higher than that of [(C^N^N)Pt^II(PPh₃)]⁺ (cf., Φ = 0.062) in CH₃CN under similar conditions. The emission energies of 1f and 2b are significantly red-shifted from their corresponding mononuclear complexes 1b and 2a, revealing that the intramolecular Pt–Pt and π–π interactions in 1f and 2b account for the low lying ³MMLCT excited states.²⁴ On the other hand, for 1f and 2b in 77 K-frozen CH₃CN solutions and at higher concentrations (>10⁻⁴ mol dm⁻³), the observed red-shifted emissions are tentatively ascribed to excimeric intraligand excited states²⁷ arising from weak π-stacking interactions between the C^N^N ligands.

A strategy in the design of anti-cancer agents is to view the [(C^N^N)Pt^II(NHC)]⁺ complexes as planar lipophilic cations. The use of planar π-conjugated organic cations to target mitochondria in cancer treatment has previously been proposed.^28a As compared to the clinically-used platinum drugs cisplatin, carboplatin and oxaliplatin, the [(C^N^N)Pt^II(NHC)]⁺ complexes are stabilized by the tridentate C^N^N and NHC ligands, rendering them to have a much higher kinetic stability in physiological conditions.

We have examined the cytotoxicity of the [(C^N^N)Pt^II(NHC)]⁺ complexes toward a panel of cancer cell lines by means of MTT assay.¹⁵ These complexes, notably 1a, were found to exhibit potent in vitro anti-cancer activities and display higher potencies than the clinically-used cisplatin. Complex 1a showed high specificity to fast-growing HeLa cells as revealed from the smallest cytotoxic IC₅₀ value toward this cell line. This complex also displays synergistic effect with cisplatin as indicated from the change in the IC₅₀ value in the presence of a fixed concentration of cisplatin, implying that 1a and cisplatin exert different anti-cancer mechanism(s). Complex 1a is relatively less cytotoxic to the human cell line derived by normal lung fibroblast cells (CCD-19Lu), rendering this complex to have safety therapeutic windows (at ranges of dose which is cytotoxic to cancer cells only). Furthermore, by means of tube-formation and wound-healing assays, we found that 1a displays no apparent anti-angiogenic and anti-metastatic activity at its sub-cytotoxic concentrations, suggesting that the anti-cancer activity of 1a is dominated by its cytotoxic property. When comparing the cytotoxicity values of 1a–1d, we found that lengthening the hydrophobic carbon chain of NHC ligand (e.g., 1a and 1b) confers higher lipophilicity to the complex thus facilitating the cellular uptake of the metal complex by cancer cells.^28b The in vivo anti-cancer activity of 1a was examined. Complex1acould significantly inhibit tumor growth in vivo with no significant reduction in body weight of the examined mice, and no induction of acute toxicity to the nude mice. These promising in vivo data warrant evaluation of the anti-cancer efficacy of 1a.

Due to their strong emission properties, the localization of the [(C^N^N)Pt^II(NHC)]⁺ complexes in cancer cells can be traced using emission microscopy. Using 1a as an example, we found that this complex mainly co-localized with mitotracker, revealing that 1a could accumulate in the cytoplasmic structures. In contrast, this complex did not co-localize with the known fluorescent DNA binder and we found that DNA (using calf thymus DNA as an example) does not quench the emission of 1a (data not shown). Thus, we reckon that DNA is unlikely to be its primary target. In a cell-free absorption-titration experiment, 1a only weakly interacts with calf-thymus DNA (ctDNA) and is at least ten times less favorable in binding to ctDNA compared to the reported [(terpy)Pt^II(L)]⁺ (where terpy = 2,2′6′,2′′-terpyridine)²⁹ or [(C^N^C)Au^III(L)]⁺ (where HC^N^CH = 2,6-diphenylpyridine).³⁰ Although 1a weakly interacts with ctDNA, gel-mobility-shift and viscosity assays showed that this complex lengthens the DNA ladders revealing its DNA intercalative property. In addition to DNA, 1a displayed no apparent inhibitory effect on the activity of a DNA-related protein topoisomerase I (TopoI).

Survivin, a member of the inhibitors of apoptosis (IAP) protein family, is highly expressed in carcinoma cells but rarely present in normal tissue.¹⁸ Active survivins are dominated in cytoplasmic structures. Recent reports revealed that an imidazolium-based lipophilic cation (YM155) displayed marked anti-tumor activity in vivo as a novel survivin suppressant.^20,31 Since [(C^N^N)Pt^II(NHC)]⁺ complexes are lipophilic cations and we found that 1a accumulates in the cytoplasmic structure of the cancer cells, the effect of 1a on survivin expression has therefore been examined by means of a survivin enzyme immunometric assay. For the first time in the literature, we have demonstrated that a metal complex can suppress the expression of survivin dose- and time-dependently. In concomitance with the survivin inhibition, 1a activates caspase-3 and poly(ADP-ribose) polymerase (PARP) whereas these activations are signaling events of apoptosis.³² Apart from the inhibition of survivin, the possibility of 1a to trigger damage in mitochondria and/or mitochondrial DNA, and hence apoptosis could not be ruled out.

Conclusion

In the search for new anti-cancer drug leads with better efficacies and pharmacological profiles, a new class of cyclometalated platinum(II) complexes containing NHC ligands have been synthesized and characterized. These complexes are strongly emissive and stable under physiological conditions. Among the complexes examined, 1a is the most potent anti-cancer agent with cytotoxic activity higher than that of the clinically-used cisplatin under in vitro conditions and could significantly inhibit tumor growth in the nude mice model. Complex 1a preferentially accumulates in cytoplasmic structures, and could act as a survivin suppressant in triggering activation of poly(ADP-ribose) polymerase (PARP) and hence apoptosis in cancer cells. It is conceived that this class of anti-cancer platinum(II) complexes [(C^N^N)Pt^II(NHC)]⁺ may complement the clinically used cisplatin by offering new modes of mechanism.

Acknowledgements

We acknowledge support from the ITF-Tier 2 project (ITS/134/09FP) administrated by Innovation and Technology Commission (HKSAR, China), and the Areas of Excellence Program (AoE/P-10/01) administrated by University Grants Council (HKSAR, China). We thank Drs J.-S. Huang and C.-N. Lok for their helpful discussion to this project.

Notes and references

1. (a) A. J. Arduengo III, M. Kline, J. C. Calabrese and F. Davidson, J. Am. Chem. Soc., 1991, 113, 9704; (b) H. W. Wanzlick, Angew. Chem., Int. Ed. Engl., 1962, 1, 75.
2. See for selected reviews and examples: (a) J. C. Y. Lin, R. T. W. Huang, C. S. Lee, A. Bhattacharyya, W. S. Hwang and I. J. B. Lin, Chem. Rev., 2009, 109, 3561; (b) D. McGuinness, Dalton Trans., 2009, 6915; (c) D. Pugh and A. A. Danopoulos, Coord. Chem. Rev., 2007, 251, 610; (d) E. A. B. Kantchev, C. J. O'Brien and M. G. Organ, Angew. Chem., Int. Ed., 2007, 46, 2768; (e) M. S. Sanford, J. A. Love and R. H. Grubbs, J. Am. Chem. Soc., 2001, 123, 6543.
3. (a) F. E. Hahn and M. C. Jahnke, Angew. Chem., Int. Ed., 2008, 47, 3122; (b) K. J. Cavell and D. S. McGuinness, Coord. Chem. Rev., 2004, 248, 671; (c) C. M. Crudden and D. P. Allen, Coord. Chem. Rev., 2004, 248, 2247.
4. (a) J. M. Praetorius and C. M. Crudden, Dalton Trans., 2008, 4079; (b) F. E. Hahn, Angew. Chem., Int. Ed., 2006, 45, 1348; (c) C. M. Crudden and D. P. Allen, Coord. Chem. Rev., 2004, 248, 2247; (d) W. A. Herrmann, Angew. Chem., Int. Ed., 2002, 41, 1290.
5. See for selected examples: (a) A. L.-F. Chow, M.-H. So, W. Lu, N. Zhu and C.-M. Che, Chem. Asian J., 2010 DOI:10.1002/asia.201000499; (b) Y. Unger, A. Zeller, S. Ahrens and T. Strassner, Chem. Commun., 2008, 3263; (c) C. Nieto-Oberhuber, S. López and A. M. Echavarren, J. Am. Chem. Soc., 2005, 127, 6178; (d) H. M. J. Wang and I. J. B. Lin, Organometallics, 1998, 17, 972.
6. See for selected examples: (a) M. Skander, P. Retailleau, B. Bourrie, L. Schio, P. Mailliet and A. Marinetti, J. Med. Chem., 2010, 53, 2146; (b) J. J. Yan, A. L.-F. Chow, C.-H. Leung, R. W.-Y. Sun, D.-L. Ma and C.-M. Che, Chem. Commun., 2010, 46, 3893; (c) K. M. Hindi, M. J. Panzner, C. A. Tessier, C. L. Cannon and W. J. Youngs, Chem. Rev., 2009, 109, 3859; (d) M.-L. Teyssot, A.-S. Jarrousse, M. Manin, A. Chevry, S. Roche, F. Norre, C. Beaudoin, L. Morel, D. Boyer, R. Mahiou and A. Gautier, Dalton Trans., 2009, 6894; (e) P. J. Barnard and S. J. Berners-Price, Coord. Chem. Rev., 2007, 251, 1889.
7. (a) C.-M. Che and F.-M. Siu, Curr. Opin. Chem. Biol., 2010, 14, 255; (b) P. C. A. Bruijnincx and P. J. Sadler, Curr. Opin. Chem. Biol., 2008, 12, 197; (c) R. W.-Y. Sun, D.-L. Ma, E. L.-M. Wong and C.-M. Che, Dalton Trans., 2007, 4884; (d) D. Wang and S. J. Lippard, Nat. Rev. Drug Discovery, 2005, 4, 307; (e) B. Rosenberg, L. Vancamp, J. E. Trosko and V. H. Mansour, Nature, 1969, 222, 385.
8. (a) L. Kelland, Nat. Rev. Cancer, 2007, 7, 573; (b) M. Galanski, M. A. Jakupec and B. K. Keppler, Curr. Med. Chem., 2005, 12, 2075.
9. (a) C. Meijer, N. H. Mulder, H. Timmer-Bosscha, W. J. Sluiter, G. J. Meersma and E. G. E. de Vries, Cancer Res., 1992, 52, 6885; (b) O. Heudi, H. Brisset, A. Cailleux and P. Allain, Int. J. Clin. Pharmacol. Ther., 2001, 39, 344.
10. See for selected examples: (a) P. Wang, C.-H. Leung, D.-L. Ma, S.-C. Yan and C.-M. Che, Chem.–Eur. J., 2010, 16, 6900; (b) D.-L. Ma, C.-M. Che and S.-C. Yan, J. Am. Chem. Soc., 2009, 131, 1835; (c) S. C. F. Kui, I. H. T. Sham, C. C. C. Cheung, C.-W. Ma, B. Yan, N. Zhu, C.-M. Che and W.-F. Fu, Chem.–Eur. J., 2007, 13, 417; (d) S.-W. Lai and C.-M. Che, Top. Curr. Chem., 2004, 241, 27; (e) W. Lu, B.-X. Mi, M. C. W. Chan, Z. Hui, C.-M. Che, N. Zhu and S.-T. Lee, J. Am. Chem. Soc., 2004, 126, 4958; (f) K.-H. Wong, M. C. W. Chan and C.-M. Che, Chem.–Eur. J., 1999, 5, 2845.
11. (a) A. H. J. Wang, J. Nathans, G. Vandermarel, J. H. Vanboom and A. Rich, Nature, 1978, 276, 471; (b) K. W. Jennette, S. J. Lippard, G. A. Vassilia and W. R. Bauer, Proc. Natl. Acad. Sci. U. S. A., 1974, 71, 3839.
12. See for selected examples: (a) D. M. Fisher, R. R. Fenton and J. R. Aldrich-Wright, Chem. Commun., 2008, 5613; (b) D.-L. Ma and C.-M. Che, Chem.–Eur. J., 2003, 9, 6133; (c) C.-M. Che, M. Yang, K.-H. Wong, H.-L. Chan and W. Lam, Chem.–Eur. J., 1999, 5, 3350.
13. I. V. Tetko, Drug Discovery Today, 2005, 10, 1497.
14. (a) T.-C. Cheung, K.-K. Cheung, S.-M. Peng and C.-M. Che, J. Chem. Soc., Dalton Trans., 1996, 1645; (b) J. A. Bailey, V. M. Miskowski and H. B. Gray, Inorg. Chem., 1993, 32, 369.
15. T. Mosmann, J. Immunol. Methods, 1983, 65, 55.
16. See for an example: R. W.-Y. Sun, M. F.-Y. Ng, E. L.-M. Wong, J. Zhang, S. S.-Y. Chui, L. Shek, T.-C. Lau and C.-M. Che, Dalton Trans., 2009, 10712.
17. See for examples: (a) R. W.-Y. Sun, C. K.-L. Li, D.-L. Ma, J. J. Yan, C.-N. Lok, C.-H. Leung, N. Zhu and C.-M. Che, Chem.–Eur. J., 2010, 16, 3097; (b) L.-K. Sy, S.-C. Yan, C.-N. Lok, R. Y. K. Man and C.-M. Che, Cancer Res., 2008, 68, 10229.
18. D. C. Altieri, Adv. Cancer Res., 2003, 88, 31.
19. (a) M. J. Duffy, N. O'Donovan, D. J. Brennan, W. M. Gallagher and B. M. Ryan, Cancer Lett., 2007, 249, 49; (b) D. C. Altieri, Nat. Rev. Cancer, 2003, 3, 46.
20. See for an example: T. Nakahara, M. Takeuchi, I. Kinoyama, T. Minematsu, K. Shirasuna, A. Matsuhisa, A. Kita, F. Tominaga, K. Yamanaka, M. Kudoh and M. Sasamata, Cancer Res., 2007, 67, 8014.
21. C. V. Kumar and E. H. Asuncion, J. Am. Chem. Soc., 1993, 115, 8547.
22. K. E. Reinert, Biochim. Biophy. Acta, 1973, 319, 135.
23. J. C. Wang, Nat. Rev. Mol. Cell Biol., 2002, 3, 430.
24. S.-W. Lai, M. C. W. Chan, T.-C. Cheung, S.-M. Peng and C.-M. Che, Inorg. Chem., 1999, 38, 4046.
25. V. W.-W. Yam, P. K.-Y. Yeung, L.-P. Chan, W.-M. Kwok, D. L. Phillips, K.-L. Yu, R. W.-K. Wong, H. Yan and Q.-J. Meng, Organometallics, 1998, 17, 2590.
26. (a) P. Shao and W. Sun, Inorg. Chem., 2007, 46, 8603; (b) V. W.-W. Yam, K. H.-Y. Chan, K. M.-C. Wong and B. W.-K. Chu, Angew. Chem., Int. Ed., 2006, 45, 6169; (c) W. Lu, M. C. W. Chan, N. Zhu, C.-M. Che, C. Li and Z. Hui, J. Am. Chem. Soc., 2004, 126, 7639; (d) W. Lu, N. Zhu and C.-M. Che, Chem. Commun., 2002, 900; (e) S.-W. Lai, H.-W. Lam, W. Lu, K.-K. Cheung and C.-M. Che, Organometallics, 2002, 21, 226.
27. (a) D. Kim and J.-L. Bredas, J. Am. Chem. Soc., 2009, 131, 11371; (b) S. Develay, O. Blackburn, A. L. Thompson and J. A. G. Williams, Inorg. Chem., 2008, 47, 11129; (c) J. Kalinowski, M. Cocchi, D. Virgili, V. Fattori and J. A. G. Williams, Adv. Mater., 2007, 19, 4000; (d) S.-W. Lai, M. C. W. Chan, K.-K. Cheung and C.-M. Che, Inorg. Chem., 1999, 38, 4262.
28. (a) J. S. Modica-Napolitano and J. R. Aprille, Adv. Drug Delivery Rev., 2001, 49, 63; (b) M. J. McKeage, S. J. Berners-Price, P. Galettis, R. J. Bowen, W. Brouwer, L. Ding, L. Zhuang and B. C. Baguley, Cancer Chemother. Pharmacol., 2000, 46, 343.
29. D.-L. Ma, T. Y.-T. Shum, F. Zhang, C.-M. Che and M. Yang, Chem. Commun., 2005, 4675.
30. C. K.-L. Li, R. W.-Y. Sun, S. C.-F. Kui, N. Zhu and C.-M. Che, Chem.–Eur. J., 2006, 12, 5253.
31. T. Iwasa, I. Okamoto, K. Takezawa, K. Yamanaka, T. Nakahara, A. Kita, H. Koutoku, M. Sasamata, E. Hatashita, Y. Yamada, K. Kuwata, M. Fukuoka and K. Nakagawa, Br. J. Cancer, 2010, 103, 36.
32. V. Schreiber, F. Dantzer, J.-C. Ame and G. de Murcia, Nat. Rev. Mol. Cell Biol., 2006, 7, 517.

---

Footnotes

† Electronic supplementary information (ESI) available: Detailed synthetic procedure and characterization of the precursors and platinum(II) complexes; experimental procedures for emission and lifetime measurements, single crystal X-ray diffraction, cell culture, cytotoxicity evaluation, tube-formation assay, wound-healing assay, in vivo anti-cancer study, detection of cellular localization of 1a, survivin enzyme immunometric assay, western blotting, absorption titration, gel-mobility-shift assay, viscosity measurement and topoisomerase I assay. Fig. S1, Schematic drawings of the precursors; Fig. S2, UV-vis spectrum of 1a; Fig. S3, molecular diagram of 1f; Fig. S4, VT-NMR spectroscopic measurement of 1a; Fig. S5, VT-NMR spectroscopic measurement of 1c; Fig. S6, VT-NMR spectroscopic measurement of 1a; Fig. S7, UV-vis spectra of 1b, 2a and 3a; Fig. S8, UV-vis spectra of 1c; Fig. S9, degassed fluid emission of 1a and 1c; Fig. S10, emission spectra of 1f; Fig. S11, emission spectra of 1b; Fig. S12, emission spectra of 1f in different solvents; Fig. S13, excitation and emission spectra of 2a; Fig. S14, excitation and emission spectra of 3a; Fig. S15, excitation and emission spectra of 1f; Fig. S16, excitation and emission spectra of 2b; Fig. S17, emission spectra of 1f at different concentrations; Fig. S18, cytotoxicity profiles of 1a; Fig. S19, tube-formation assay of 1a; Fig. S20, wound-healing assay of 1a; Fig. S21, absorption titration of 1a with ctDNA; Fig. S22, gel-mobility shift assay of 1a; Fig. S23, viscosity measurement of 1a; Fig. S24, topoisomerase I assay of 1a; Table S1; calculated lipophilicity of the carbene ligands; Table S2, crystal data of 1a, 1c and 1f; Table S3, selected bond distances and bond angles of 1a, 1c and 1f; Table S4, UV-vis absorption and emission data of 1c. CCDC reference numbers 756500, 756501 and 756623. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c0sc00593b

‡ These authors contributed equally to this work.

---