A Conformationally Restricted Gold(III) Complex Elicits Antiproliferative Activity in Cancer Cells

Abstract

Diamine ligands are effective structural scaffolds for tuning the reactivity of transition metal complexes for catalytic, materials, and phosphorescent applications, and have been leveraged for biological use. In this work, we report the synthesis and characterization of a novel class of cyclometalated [ĈN] gold(III) complexes bearing secondary diamines including a norbornane backbone: (2R,3S)-N²,N³-dibenzylbicyclo[2.2.1]heptane-2,3-diamine, or cyclohexane backbone: (1R,2R)-N¹,N²-dibenzylcyclohexane-1,2-diamine. X-ray crystallography confirms square planar geometry and chirality at nitrogen. The electronic character of the conformationally restricted norbornane backbone influences electrochemical behavior with redox potentials of -0.8 – -1.1 V, atypical for Au(III) complexes. These compounds demonstrate promising anticancer activity, particularly, complex 1, which bears a benzylpyridine organogold framework and supported by the bicyclic conformationally restricted diamino norbornane shows good potency in A2780 cells. We further show that cellular response to 1 evoke ROS production and does not induce mitochondrial dysfunction. This class of complexes provides significant stability and reactivity for different applications in protein modification, catalysis, and therapeutics.

Introduction

The resurgence of gold chemistry over the past two decades has contributed to the development of synthetic methods,^1–4 catalytic transformations,^{5, 6} materials for electronics,^{7, 8} reagents for bioorthogonal reactions,⁹ protein modifications,^{10, 11} and therapeutic agents.^12–17 In the context of drug discovery, great impetus is derived from auranofin, an FDA approved drug for treating rheumatoid arthritis.¹⁸ Although Au(III) complexes are high-valent and present opportunities for ligand tuning due to having more coordination sites than conventional linear Au(I), the instability of Au(III) has been its Achilles’ heel. Ongoing efforts to stabilize the Au(III) center is crucial to accessing complexes for improved performance or activity.^19–22 Multidentate ligands and cyclometalation to form Au – C bonds are common strategies deployed to resolve this bottleneck associated with Au(III) instability.^23–30 In addition to the aforementioned strategies, strong σ-donor ligands such as bidentate phosphorus and nitrogen chelates play a dominant role in generating stable Au(III) complexes.^{24, 31–34}

There is extensive literature describing the utility of organogold Au(III) complexes in catalysis,^35–39 material science,^40–43 sensors,^44–46 and anticancer activity.^47–56 Although organometallic gold(III) complexes bearing N^N bidentate ligands with sp²-hybridized nitrogen have been recently reported,^57–60 bidentate ligands bearing sp³-hybridized primary or secondary amines are underexplored, yet hold tremendous potential for accessing stable but distinct scaffolds.²⁴ This has implications on their reactivity and consequent applications. For example, reports using chiral primary DACH ligands as ancillary ligands of cyclometalated Au(III) give rise to unique structural scaffolds dictated by the starting cyclometalated Au(III) framework.²⁴ It is important to note that chiral Au(III) compounds have garnered immense interest lately, but more is needed in the utility of chiral diamine ligands.^{24, 37, 61–63} Other water-soluble amine ligands, including biguanide derivatives with good physiological stability and anticancer activity, have also been reported.^{32, 64, 65} The use of secondary amines as ancillary ligands has the potential to impart metal-center chirality, high stability, and good reactivity for different applications including bioconjugation, medicinal chemistry, catalysis, and material science.^{66, 67}

In this work, we aimed at preparing stable [ĈN]Au(III) complexes bearing uncommon N^NH chelates derived from chiral secondary amines to generate unprecedented monocationic [ĈN]-Au(III)RN^NHR derivatives. The secondary diamine ligands used in this work possess a norbornane (bicyclic) backbone or cyclohexane backbone to influence stability and reactivity due to conformational constraints imposed by these ligands. The norbornene diamines were synthesized via the recently reported 1,2-diamination of alkenes.⁶⁸ Inspiration is gained from platinum-based anticancer complexes bearing chiral amine ligands that show differential activity against cancer.^69–72 Considerations regarding the structure and stereoelectronic properties of chiral diamines are helpful in the design of new Au(III) diamine complexes. Moreover, the bicyclic property of diamine ligands can modulate steric hindrance, stereoelectronic, non-planarity, and lipophilic properties toward exciting, new applications.⁷³ The chirality and the optical properties of these complexes present new opportunities for catalysis, medicinal inorganic chemistry, and materials science.⁷⁰ In this work, the synthesis of four chiral Au(III) complexes bearing secondary diamine ligands are reported along with their complete spectroscopic characterization (¹H-NMR, ¹³C-NMR), and their purity was ascertained by HPLC/ESI-MS. The structures of these compounds were elucidated by X-ray crystallography.

Results and Discussion

Norbornane-2,3-diamines are bicyclic, conformationally restricted compounds due to decreased freedom of intramolecular rotations or motions around chemical bonds. The amino groups can be cis (usually exo, exo) or trans (exo, endo). Preparation of cis diamines has previously been achieved by the addition of N₂O₄ to norbornenes followed by reduction⁷⁴ or by double Curtius rearrangements of carboxylic acids;^{73, 75} the preparation of trans diamines has been achieved by nucleophilic attack of azides on aziridines followed by reduction.^{76, 77} In our work, we prepared cis-2,3-norbornanediamine via 1,3-dipolar cycloaddition between norbornene and an alkyl azide followed by N-alkylation and reduction.⁶⁸ We rationalized that structural modification to the [ĈN]-cyclometalated Au(III) framework with N-donor ancillary ligands that are conformationally restricted provide unique reactivity and stability. Diversification of cyclometalated Au(III) complexes bearing sterically or geometrically constrained ligands is attractive to exploring gold-macromolecular interactions as well as unveil new biological targets and reactivity for catalysis and material application. Cyclometalated (ĈN) Au(III) systems supported with σ-donor diamino ligands are known to be stable.^{35, 78} We previously reported Au(III) complexes bearing the primary diamine ligand, R,R-DACH and their stability and antiproliferative action investigated.²⁴ Thus, we hypothesized that the use of secondary amines would enhance complex stability. Here, we report the use of two different secondary diamine ligands that chelate Au(III) centers to form cationic complexes of the type, [Au(III)RN^NHR]⁺. The goal was to assess the impact of steric and electronic properties in the norbornane- or cyclohexyl-derived ligands on complex stability and activity. To achieve monocationic Au(III) complexes, the reaction of [ĈN]-cyclometalated Au(III) and diaminonorbornane was performed in an equimolar DCM/MeOH solution in the presence of a strong nucleophilic base, NaOtBu at room temperature for 16 h. The base activates the diamine ligand for coordination, deprotonation of the diamine ligand imparts a cationic state on the overall Au(III) complex. When Au(III) starting material with benzylpyridine was allowed to react with (1R,2R)-N¹,N²-dibenzylcyclohexane-1,2-diamine, we optimized conditions to use Ag₂O instead of NaOtBu to minimize gold reduction and maximize yield. The diamine ligands used formed a coordination bond (Au–N) with the Au(III) center trans to the pyridine N(sp²). The other nitrogen in the diamine ligand formed a covalent bond (Au–NH) with the Au(III) center and displays chirality at the nitrogen as shown by the X-ray crystal structure (Figure 3). All the complexes were obtained in good synthetic yields and displayed excellent stability when open to air and moisture.

Figure 3.

X-ray crystal structure of complex 1, 2, and 3, ellipsoids drawn at 50%. Hydrogen atoms and solvents were omitted for clarity.

Purification of the complexes were by normal phase silica-gel flash chromatography using Teledyne Combi-Flash to afford yellow colored powder for complexes 1, 2, and 4. The use of the silver-mediated reaction to afford complex 3 was purified by filtration through alumina pad and the dichloromethane filtrate was precipitated with hexane. We found that complexes tolerated the purification conditions well with no observable degradation or reduction to elemental gold on the solid phase silica or alumina. Purity of the compounds were ascertained by HPLC.

Complexes 1–4 were characterized by standard techniques including UV-Vis, NMR spectroscopy, ESI-MS, and cyclic voltammetry. The proton signals from the norbornane or cyclohexyl backbone were well resolved and located between 1.5 ppm and 4.5 ppm as expected. The aromatic protons from the cyclometalated framework and the benzyl units on the diamine ancillary ligand were at characteristic downfield regions of the H-NMR spectra.^{24, 79, 80} Electronic absorption profiles of 1–4 was measured in RPMI by UV-Vis spectroscopy. All complexes display a characteristic high energy band at 250 nm attributed to charge transfer behavior. Except for complex 4, a red -shifted, low energy peak in the range of 350 – 400 nm (λmax) was observed for 1, 2, and 3 (Figure 1). Typically, high energy absorption bands are often associated with π - π electronic transitions MCLT (metal-to-ligand charge-transfer) or LMCT (ligand-to-metal charge transfer).⁸¹ Oscillator strength and molecular orbital (MO) contributions obtained from TDDFT theoretical calculations of other cyclometalated Au(III) compounds support the electronic transitions observed for 1 – 3.^82–84 It is not clear as to why the low-energy bands were not observed for complex 4, perhaps may be associated with stability in RPMI.

Figure 1.

UV-Vis absorption spectrum of complexes 1–4 in biologically relevant media RPMI at room temperature, final concentration of the complexes was 50 μM.

The electrochemical behavior of 2 and 3 were examined by cyclic voltammetry in DMSO using tetrabutylammonium hexafluorophosphate [N(Bu)₄PF₆] as supporting electrolyte at a Pt electrode and Ag/AgCl as a reference electrode. We found that 2 bearing a norbornane ligand show a major reduction event at about -0.8 V vs Ag/AgCl electrode (Figure 2). In the case of 3, which bear the cyclohexyl-derived ligand, the reduction event was at -1.1 V vs Ag/AgCl. The free diamino norbornane ligand did not show redox activity, which may suggest that the observed reduction and oxidation events are Au(III) complex related. It is possible the conformational restriction of the norbornane ligand imparts significant electrochemical stability to the Au(III) center in complex 2, compared to the cyclohexyl-derived ligand in complex 3. The result demonstrates the influence of restricted bicyclic ligands in tuning redox behavior of cationic organogold complexes with potential benefits for stability under physiological conditions. Evidence for disproportionation in DMSO and reduction to metallic gold was not observed.

Figure 2.

Cyclic voltammogram recorded at a platinum electrode in DMSO solution of 10 mM 2 or 3 with 0.10 M N(Bu)₄PF₆ supporting electrolyte at a scan rate of 0.1 V/s using Ag/AgCl reference electrode at room temperature.

Using single-crystal X-ray diffraction, the structures of 1, 2, and 3 were elucidated (Figure 3). Single crystals of 1, 2, and 3 were grown by vapor diffusion as follows: (1) vapor diffusion of hexane into a concentrated CHCl₃ solution; (2) vapor diffusion of diethyl ether into a concentrated dichloromethane solution; and (3) vapor diffusion of hexane into a concentrated dichloromethane solution. In complexes 1 and 2, all the bond lengths around Au(III) have similar lengths except that of Au-N (trans to pyridinium nitrogen) bond, which were relatively shorter at 1.995 and 1.992 respectively. It is possible that the electronic contribution by the diamino norbornane influences bond strength. The structures of the complexes show square planar geometry around the gold center. The triflate counter ion of complex 2 was from the diamino norbornane synthesis via the azide – alkene cycloaddition.⁶⁸ The complex bearing the cyclohexyl-derived diamine, 3, showed that all four bonds around the Au(III) center possess the same bond length (Table 1). Based on the X-ray crystal structures, the diamine ligands maintain an R’-S’ configuration in the norbornane Au(III) complex and S’-S’ configuration for the cyclohexyl derivative.

Table 1.

Selected interatomic distances (Å) and bond angles (°) from the X-ray crystal structures of 1, 2, and 3.

Bond length (Å)
	1	2	3
Au1 - N1	2.077	2.0746	2.019
Au1 - N2	2.141	2.1617	2.027
Au1 - N3	1.995	1.992	2.107
Au1 - C1	2.027	2.014	2.137
Angle (°)
N3 - Au1 - Cl	101.12	98.74	90
N3 - Au1 - N1	173.11	171.23	177.39
C1 - Au1 - N1	85.66	88.42	87.4
N3 - Au1 - N2	76.68	75.61	83.73
C1 - Au1 - N2	177.8	174.05	173.36
N1 - Au1 - N2	96.54	97.38	98.86

Intrigued by the optical and electrochemical properties of the conformationally restricted complexes 1 and 2, we subjected 1 to solution stability studies using RPMI media. We monitored the absorption signal of 1 by HPLC-ESI-MS over a 20-hour period (Figure 4a). Briefly, a stock concentration of complex 1 was dissolved in RPMI and at each time, aliquots of 100 μL were taken and diluted with 100 μL methanol for LC-MS analysis. Complex 1 eluted at a retention time of 8 min (A₂₆₀ nm) with a m/z = 670.2, which remained unaltered throughout the study. Formation of other peaks was not observed, indicative of a soluble and perhaps physiologically stable complex. Encouraged by the stability of 1 in RPMI we conducted a more rigorous physiological stability using the well-known biological reductant, L-glutathione (L-GSH) and serum albumin, which is the most abundant protein in whole blood of most mammals including humans. First, we incubated complex 1 with L-GSH in 1:10 ratio and using UV spectrophotometry monitored the absorption bands (Figure S35). Changes to the absorption peak at 280 nm and the formation of a peak at 570 nm was observed by the 1 h time point, which remained unaltered till about 56 h. To further characterize this phenomenon, we used LC-ESI-MS to elucidate any potential speciation (Figure 4b) and found that upon incubation of 1 with L-GSH, a reductively eliminated product of the arylpyridine covalently linked to L-GSH through a C-S bond (1-GSH, m/z 475) forms together with intact complex 1 (m/z 670.2). It must be noted that cyclometalated Au(III) complexes of the ĈN archetype can undergo reductive elimination to form C-S arylated products induced by thiol nucleophiles such as cysteines and proteins^{4, 85–88} or direct C-N reductive elimination from Au(III).⁸⁹ Whereas the C-S transformation has been a useful bioorthogonal strategy in biological systems and important for target engagement by gold compounds, a careful balance of reactivity and stability is needed for therapeutic development making ligand tuning an important descriptor in gold-based probe/drug discovery. Interestingly, a significant amount of 1 remains intact in solution over the reaction period of 24 h. This suggests that 1 may maintain pharmacological activity and further demonstrates the utility of this class of compounds in the biological setting. Second, we studied the interaction of 1 with bovine serum albumin (BSA) in a reaction by UV spectrophotometry (Figure S35), and LC-ESI-MS (Figure 4c). Whereas no major alterations were observed in the UV absorption profile over a 56 h reaction duration, the peak corresponding to complex 1 based on the retention time and m/z, although intact, diminishes over time but is observable at 24 h.

Figure 4.

Stability studies. A. HPLC trace (A₂₆₀ nm) following the incubation of 1 with RPMI medium for 20 h. At respective time points reaction was aliquoted (100 μL) and diluted with methanol (100 μL) prior to HPLC runs. Blank run of RPMI-methanol is included to account for the baseline. B. Formation of 1-GSH by the chemical reaction of 1 and L-GSH as observed by LC-MS and monitored over 24 h period. C. HPLC trace (A₂₆₀ nm) following the incubation of 1 in methanol and BSA in PBS for 24 h.

The antiproliferative potential of 1–4 was investigated in a panel of breast cancer cells (MDA-MB-231, MDA-MB-468, MCF7) and an ovarian cancer cell line (A2780) by MTT assay (Table 2). The rationale for these cell lines was to evaluate the response of different cancer subtypes and respective metabolic status to 1–4. In general, the conformationally restricted Au(III) complexes 1 and 2 were more potent with IC₅₀ of 1.6 μM and 4.2 μM in glycolytic A2780 ovarian cancer cells respectively. In the triple negative breast cancer (TNBC) cell line, MDA-MB-231 which is dependent on both oxidative phosphorylation (OXPHOS) and glycolysis,⁹⁰ the IC₅₀ was 7.3 μM and 6.2 μM for 1 and 2 respectively. The recorded IC₅₀ values extrapolated from dose-response curves for the cyclohexyl counterpart, 3 was a 2–4-fold higher in the breast cancer cells but maintained activity in A2780 cells. Whereas compound 4 displayed activity across all cell lines, it was the least active in A2780 cells with an IC₅₀ of 5.2 μM. It became clear that the OXPHOS dependent TNBC cell line,⁹¹ MDA-MB-468 was not responsive to 1–4. This observation may likely offer insights into the potential mechanism of action of these complexes, eliminating the mitochondria as the major target.

Table 2:

Cell viability of 1–4 expressed as IC₅₀ values (μM) in a panel of cancer cells after 72 h treatment. Gold compounds were freshly prepared in DMSO and used immediately. DMSO concentration was < 1%.

Complex	IC50 (μM) MDA-MB 231 (breast)	IC50 (μM) MDA-MB 468 (breast)	IC50 (μM) MCF7 (breast)	IC50 (μM) A2780 (ovarian)
1	7.3 ± 0.04	28.8 ± 0.13	14.6 ± 0.04	1.6 ± 0.13
2	6.2 ± 0.02	13.5 ± 0.12	14.4 ± 0.03	4.2 ± 0.04
3	17.7 ± 0.03	97.1 ± 0.24	28.6 ± 0.03	4.7 ± 0.07
4	4.8 ± 0.03	12.7 ± 0.11	13.7 ± 0.03	5.2 ± 0.04

In the context of SAR, we investigated the intracellular accumulation of 1 – 4 by measuring the gold content in A2780 cells (Figure 5). Briefly, A2780 cells were treated with the complexes at 5 μM or 10 μM for 18 h respectively, and GFAAS was used for gold quantification after washing the cells and digestion. In the two treatment conditions used, it was clear that 2 was most taken up and 3 the least taken up in A2780 cells. The triflate counter ion in 2 may potentiate physicochemical properties useful for enhanced cellular uptake. Moreover, the combination of benzylpyridine and dibenzylcyclohexane-1,2-diamine ligand in 3 is not favorable for increased cellular uptake. This explains the relatively poor potency of 3 among the compound series evaluated. At a treatment concentration of 5 μM, the cellular uptake of 1 and 2 were comparable. Conceivably, the high uptake of 2 dictates the promising antiproliferative activity observed across multiple cell lines. From our studies, we notice that the conformationally restricted norbornane ligands enhance electrochemical and solution stability of the overall Au(III) complex and demonstrate improved cellular uptake and anticancer potency in cells. Importantly, the use of benzoylpyridine for cyclometallation as in 2 and 4 contributes to uptake and antiproliferative action in cancer cells.

Figure 5.

Cellular uptake. Intracellular accumulation of 1–4 in A2780 cancer cells. Cells were treated with 5 μM (left) or 10 μM (right) for 18 h.

We used a fluorescence assisted cell sorting (FACS) study to assess the apoptotic potential of 1 in ovarian cells. A2780 cells were treated with 1 at 5 μM along with Annexin-V-FITC and propidium iodide (Figure 6). The experiment showed that cells treated with 1 induced 15 % apoptosis at 5 μM in A2780 cells after 24 h of exposure when compared to control cells. The apoptotic phenotype was further verified by examining critical markers of apoptosis via immunoblotting (Figure 6d). The accumulation of cleaved PARP and depleted PARP in a concentration-dependent manner by 24 h supports an apoptotic mechanism. However, based on the minimal alteration to cytochrome c expression, it is possible that apoptosis occurs via a cytochrome c-independent mechanism.⁹²

Figure 6.

Dual staining flow cytometry experiment showing apoptosis at 5 μM of 1 in A2780 cells for 24 h. Propidium iodide and FITC-labeled Annexin V was used to detect apoptotic populations (a) control and (b) complex 1 treated. (c) Bar chart representation of the apoptotic populations extrapolated from a and b. Data analyzed by two-tailed unpaired Student’s t test (*p < 0.05). (d) Expression levels of apoptotic proteins in A2780 following a 24 h treatment of 1 at indicated concentrations.

As demonstrated through electrochemical studies, the complexes under investigation are redox- active species; therefore, we conducted flow cytometry studies to examine whether the redox activity of 1 produces total, intracellular ROS (Figure 7a, b). In A2780 cells, cellular ROS increased in a dose-dependent manner whereas N-acetyl cysteine (NAC) co-treatment largely reversed the effect. Mitochondrial ROS was undetectable using the Mitosox assay in a flow cytometry experiment (Figure 7c, d) measured under similar conditions as the total ROS study. The elevation of cellular ROS in A2780 by 1 is a critical driver of cell death.⁹³ Gold complexes that evoke ROS-dependent mechanisms often exert thioredoxin inhibition as a probable cause of ROS accumulation leading to oxidative stress. Taken together, these studies detail total ROS specificity induced by 1 resulting in cell death. Further studies beyond the scope of this report are needed to characterize the molecular target(s) of this class of compounds.

Figure 7.

(a) Total intracellular ROS levels induced by 1 using DCFDA flow cytometry assay. (b) Bar chart representation of data extrapolated from a. (c) mitochondrial ROS measured by MitoSOX flow cytometry (PE channel) assay. (d) Bar chart representation of data extrapolated from c. Cells were exposed to complex 1 for 2 h prior to assay. Data are plotted as the mean ± SD. (n =3). Data were analyzed by ordinary one-way ANOVA followed by Dunnett’s multiple comparison test and two-tailed unpaired Student’s t test (** p < 0.01, * p < 0.05, n.s. = not significant).

Initial cell viability studies showed that OXPHOS-dependent cells do not respond well to complexes 1–4. To assess the effect of conformationally restricted compounds on mitochondrial-related toxicity as induced by other gold(III) agents,⁹⁴ we evaluated the effect of 1 on mitochondrial function and glycolysis⁹⁵ in A2780 cells. Using the Seahorse extracellular flux analyzer (Seahorse XF96)⁹⁶ A2780 cells was subjected to the cell mito stress test that measures mitochondrial respiration and concomitant glycolytic effects (Figure 8a, b). Briefly, cells were pretreated with 1 for 12 h followed by mitochondrial complex inhibitors (oligomycin, rotenone, antimycin A) or uncoupler (carbonyl cyanide-4-(trifluoro- methoxy) phenylhydrazone (FCCP) to measure oxygen consumption rate (OCR), extracellular acidification rate (ECAR), and other extrapolated parameters such as basal respiration, maximal respiration, and proton leak that characterizes mitochondrial stress.⁹⁷ We found that the measure of mitochondrial respiration, which is OCR was not significantly affected by the pretreatment of 1 in A2780 (Figure 8a). Further, extrapolated mitochondrial parameters including basal respiration and maximal respiration remained unaltered (Figure S 38).⁹⁸ Additionally, spare reserve capacity (SRC) is a robust functional read out to evaluate mitochondrial reserve in non-transformed and cancer cells.⁹⁹ SRC in A2780 decreased following the treatment of 1 at 10 μM. Coupling efficiency in response to 1 followed a similar pattern as the ATP parameter, with no changes at higher concentrations in comparison to control cells, pointing to unaltered mitochondrial metabolism by 1 in A2780 cancer cells. ECAR, which is often a measure of glycolytic performance was activated by 1 at 10 μM in A2780 (Figure 8b) with a dose-dependent increase in non-mitochondrial oxygen consumption rate, further confirming our hypothesis that this class of conformationally restricted gold(III) adopt a mechanism-of-action independent of mitochondrial modulation. Functional studies using mitochondrial membrane potential (MMP) showed that 1 does not depolarize MMP in TMRE-based flow cytometry studies (Figure 8c, d). Overall, these studies verify that 1 does not induce mitochondrial dysfunction in cancer cells.

Figure 8.

(a) OCR from cell mito stress study using Seahorse XF96. A2780 cancer cells were pretreated with 1 (12 h) and various inhibitors of ETC were added at indicated time points. (b) ECAR from cell mito stress study using Seahorse XF96. A2780 cancer cells were pretreated with 1 (12 h) and various inhibitors of ETC were added at indicated time points. (c) Mitochondria membrane potential measured by flow cytometry analysis of TMRE fluorescence intensity. (d) Bar chart of MMP extrapolated from (c). Data are plotted as the mean ± SD. (n = 3). Data were analyzed by ordinary one-way ANOVA followed by Dunnett’s multiple comparison test (**** p < 0.0001, n.s. = not significant).

Conclusion

In summary, a new cyclometalated gold(III) scaffold supported by conformationally restricted bicyclic norbornane ligands have been synthesized. The characterization of these four chiral complexes reveals monomeric gold(III) species with chirality of the diamine ligand coordinated to the gold center. Interestingly, these reactions proceed by ligand substitution of the chlorido ligands on the starting cyclometalated gold(III) scaffolds. The spatial arrangement of ligand surrounding the gold center was confirmed by X-ray crystallography. The unique chirality at the nitrogen center covalently bonded to the gold (NH-Au) present additional complexity to the chirality of the overall complex, which requires further investigation. It is worth mentioning that these complexes are air and moisture stable. This new structural class of gold(III) compounds provides a platform for new synthetic materials that will be beneficial to catalysis, biological probe development, therapeutics, and material science.

The stability of complex 1 in RPMI media, and the relative stability in the presence of a biological reductant or protein points to the biological relevance of these scaffolds. In vitro, all four complexes were more cytotoxic in glycolysis-dependent cell lines compared to cells that are highly dependent on oxidative phosphorylation. Unlike other gold anti-cancer agents previously developed in our laboratory and others, the initial cell viability studies prompted a potentially different mechanism of action independent of mitochondrial function. An increase in apoptotic population was observed after treatment with 1, an effect we attribute to increased oxidative stress induced in A2780. We also noted that the induction of ROS by 1 was in cellular locations other than the mitochondria since no significant increase in mitochondrial ROS was observed.

Investigating the mitochondrial bioenergetics of A2780 cells pretreated with 1 showed perturbation of the glycolysis rate of the cells while oxidative consumption rate remained relatively unchanged. Taken together, tuning the organogold center with conformationally restricted bicyclic diamines exert a different biological consequence independent of mitochondrial dysfunction. Further investigation beyond the scope of this report is, however, required to fully characterize the mechanism of anticancer action of this new class of compounds.

The SAR studies further outline the impact of conformationally restricted norbornane ligands on electrochemical and solution stability of the overall Au(III) complex and potentiates cellular uptake and anticancer potency in cells as demonstrated by 1. Notably, the use of benzoylpyridine for cyclometalation as in 2 and 4 influences uptake and antiproliferative action in cancer cells. Overall, conformationally restricted ligands may pave the way towards a more diverse class of biologically relevant Au(III) complexes.

Scheme 1:

Synthetic procedure for complexes 1–4.