Relation of Metal-Binding Property and Selective Toxicity of 8-Hydroxyquinoline Derived Mannich Bases Targeting Multidrug Resistant Cancer Cells

Simple Summary Effective treatment of cancer is often limited by the resistance of cancer cells to chemotherapy. A well-described mechanism supporting multidrug resistance (MDR) relies on the efflux of toxic drugs from cancer cells, mediated by P-glycoprotein (Pgp). Circumventing Pgp-mediated resistance is expected to make a significant contribution to improved therapy of malignancies. Interestingly, MDR cells exhibit paradoxical hypersensitivity towards a diverse set of anticancer chelators. In this study we explore the relation of chemical and structural properties influencing metal binding and toxicity of a set of 8-hydroxyquinoline derivatives to reveal key characteristics governing “MDR-selective” activity. We find that subtle changes in the stability and redox activity of the biologically relevant metal complexes significantly influence MDR-selective toxicity. Our results underline the importance of chelation in MDR-selective toxicity, suggesting that the collateral sensitivity of MDR cells may be targeted by preferential iron deprivation or the formation of redox-active copper(II) complexes. Abstract Resistance to chemotherapeutic agents is a major obstacle in cancer treatment. A recently proposed strategy is to target the collateral sensitivity of multidrug resistant (MDR) cancer. Paradoxically, the toxicity of certain metal chelating agents is increased, rather than decreased, by the function of P-glycoprotein (Pgp), which is known to confer resistance by effluxing chemotherapeutic compounds from cancer cells. We have recently characterized and compared the solution’s chemical properties including ligand protonation and the metal binding properties of a set of structurally related 8-hydroxyquinoline derived Mannich bases. Here we characterize the impact of the solution stability and redox activity of their iron(III) and copper(II) complexes on MDR-selective toxicity. Our results show that the MDR-selective anticancer activity of the studied 8-hydroxyquinoline derived Mannich bases is associated with the iron deprivation of MDR cells and the preferential formation of redox-active copper(II) complexes, which undergo intracellular redox-cycling to induce oxidative stress.


8-Hydroxyquinoline derived Mannich bases used in this study
Figure S1: Compounds applied in the study. Compounds that were investigated previously with regard to their iron(III) and copper(II) binding abilities, are framed with boxes. Chlorination in R5 is indicated by blue circles, lack of chelating quinolinium nitrogen by grey circles.

Characterization of employed cell line panel
Expression of Pgp in the applied cell line panel was confirmed by Westernblot. Furthermore, Expression of BCRP (ABC-G2) and MRP-1 (ABC-C1) was excluded by Westernblot. Functionality of Pgp is shown by Calcein-AM assay. The characterization of the panel is shown in Figure S2.

Synthesis and Characterization of non-chelating derivatives
Compounds NC-2, NC-3 and NC-4 without chelating moieties were synthesized by the modified Mannich-reaction starting from 1-naphthol and the respective amines in the presence of paraformaldehyde. In case of NC-4, the ring-closed [1,3]naphthoxazine was obtained as an intermediate and cleaved under acidic conditions. Products were obtained in moderate to good yields.

Instrumentation, detailed description and characterization of compounds
Chemicals used for synthesis were at least of reagent grade quality, obtained from commercial suppliers and used without further purification. Solvents were used as received or dried. All new compounds whose biological activity was evaluated in this work have a purity ≥98% as confirmed by NMR and/or elemental analysis.

2-(2-Fluorobenzyl-aminomethyl-1-naphthol hydrochloride (NC-4).
Hydrolytic cleavage of the intermediate (0.3 g, 1.02 mmol) was performed by refluxing a suspension in 70 mL aq. HCl (20 %) for 2.5 h. Upon evaporation of the solvent, the crude product was crystallized from EtOAc (25 mL) and recrystallized with EtOH (30 mL). The final product was obtained in a yield of 78% (0.253 g);. 1 H NMR (DMSO-d6, Figure   Anal. calcd. for C18H15FN2O (294.32): C,73.45;H,5.14;N,9.52. Found: C,74.02;H,5.18;N,9.48.   Table X were plotted against the phenolic-OH pK a (A) and quinolinium nitrogen pK a (B) of Q-1 (black circles), Q-2 (blue triangles), Q-3 (green diamonds) and Q-4 (purple squares) (57), as well as for Cl-Q-1 (grey circles), Cl-Q-2 (light blue triangles), Cl-Q-3 (light green diamonds) and de-Cl-Q-4 (magenta squares). Even though pK a values have an impact on lipophilicity (logD 7.4 Table S1: Toxicity of ligands and in situ preformed complexes in MES-SA and MES-SA/Dx5 cells in the absence and presence of 1 µM TQ. IC 50 values and standard deviation of at least three independent experiments are given in µM calculated as ligand equivalents. Selectivity-ratios are given as the ratio of IC 50 values obtained in Pgp negative divided by the Pgp positive cells. In order to avoid misinterpretation due to different amount of ligand present in the respective complexes, concentrations are expressed as ligand equivalents. (Therefore at the 1:3 metal-to-ligand ratio, the IC 50 values refer to "1/3 [ML 3 ]".) While complexation with iron(III) seems to slightly decrease toxicity of the ligand, complexation with copper(II) increases the efficiency. Even though with pre-formed complexes similar trends could be observed as in the co-incubation experiments, these trends were much less pronounced in case of the pre-mixed complexes.   Figure S21: Impact of Pgp on toxicity of ligands (black, Q-1 (A), Q-2 (B), Q-3 (C) and Q-4 (D)) and complexes at different metal-to-ligand ratios illustrated by different colors: Fe:L = 1:3 (orange), 1:2 (light red), 1:1 (bordeaux), Cu:L = 1:2 (green), 1:1 (blue) and L (black). The toxic activity (pIC 50 values) is compared in Pgp negative (MES-SA, x-axis) and Pgp positive (Dx5, y-axis) cells (A-D). While having no effect on the toxicity of ligand and complexes of Q-1 (E) and Q-2 (F), the Pgp inhibitor TQ decreases the activity of MDR selective ligands and complexes (Q-3 (G), Q-4 (H)). Expectedly, toxicity in Pgp negative MES-SA cells is not altered by co-administration of TQ (black, Q-1 (I), Q-2 (J), Q-3 (K) and Q-4 (L)).  Table S3: Characterization of intracellular ROS production with the DCFDA assay for P-gp negative MES-SA and P-gp positive MES-SA/Dx5 cells. Fold changes in fluorescence are shown after a 2 h with the investigated 8-hydroxyquinoline ligands. Data S-refer to Figure 7 and allow additional comparison of R5 unsubstituted and chloro-substituted derivatives of Q-1 to Q-4 in a concentration range from 6.25µM to 50µM. Fold change of fluorescence obtained in the presence of 5 mM NAC is given in brackets.