The activity of therapeutic molecular cluster Ag5 is dependent on oxygen level and HIF-1 mediated signalling

Abstract

Regions of hypoxia occur in most solid tumours and are known to significantly impact therapy response and patient prognosis. Ag5 is a recently reported silver molecular cluster which inhibits both glutathione and thioredoxin signalling therefore limiting cellular antioxidant capacity. Ag5 treatment significantly reduces cell viability in a range of cancer cell lines with little to no impact on non-transformed cells. Characterisation of redox homeostasis in hypoxia demonstrated an increase in reactive oxygen species and glutathione albeit with different kinetics. Significant Ag5-mediated loss of viability was observed in a range of hypoxic conditions which mimic the tumour microenvironment however, this effect was reduced compared to normoxic conditions. Reduced sensitivity to Ag5 in hypoxia was attributed to HIF-1 mediated signalling to reduce PDH via PDK1/3 activity and changes in mitochondrial oxygen availability. Importantly, the addition of Ag5 significantly increased radiation-induced cell death in hypoxic conditions associated with radioresistance. Together, these data demonstrate Ag5 is a potent and cancer specific agent which could be used effectively in combination with radiotherapy.

Keywords: Ag5; HIF-1; Hypoxia; Radiation; Redox.

Gradients of oxygen exist within solid tumours, including well-oxygenated (pink) areas and those that are increasingly hypoxic (blue). Ag5 inhibits both GSH and Trx therefore impacting the main cellular anti-oxidant pathways. Ag5 exposure has little to no effect on normal cells while leading to loss of viability in a range of cancer cell lines. Sensitivity to Ag5 in hypoxia can be rescued through loss of HIF-1a. Hypoxic cells are radiosensitised by the addition of Ag5. CAC – colorectal adenocarcinoma, CCC- clear cell carcinoma, NSCLC – non small cell lung carcinoma, OAC- oesophageal adenocarcinoma. Figure created with BioRender.Com, license number WK25R9XLMQ.

Fig. 1

Ag5 is toxic to cancer cells but not non-transformed cells. A. Lung cancer lines (A549, H460) and non-transformed lung lines (MRC5, HFL-1) were treated with Ag5 (0, 0.5, 1, 1.25, 1.5 μM) for 1 h and analysed via MTT 20 h later. B. A549 and MRC5 cells were treated with Ag5 (0, 0.25, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.25, 1.5, 2 μM) for 1 h and analysed via MTT 20 h later. C. A549 and H460 cells were treated with the control compound, Ag+ (0, 0.5, 1, 1.25, 1.5 μM) for 1 h, followed by an MTT assay 20 h later. D. Levels of ROS (MitoSOX) were determined in untreated A549 and MRC5 cells. E. Representative images from part D. Scale bar = 20 μm F. A panel of oesophageal cancer cell lines (OE21, OE33, SKGT4, FLO1) were treated with Ag5 (0, 0.5, 1, 1.25, 1.5 μM) for 1 h and an MTT assay was carried out after 20 h G. A549 cells were treated with Ag5 (1 μM) for 1 h followed by annexin-V and 7-AAD assay for apoptosis after 2 and 6 h H. A549 cells were treated with Ag5 (1 μM) for 1 h, followed by a lipid peroxidation assay 6 and 24 h after drug treatment. Erastin (10 μM) was used as a positive control and NEM (200 μM) to verify that the fluorescence recorded was due to lipid peroxidation. Data shown in A, B, C, D, E, F and G are n = 3. Data shown in part H is n = 3 except for the Erastin + NEM condition which is n = 2. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate, except for the IF). Data presented as mean + SEM. Statistical testing was done using a two-way ANOVA test for the cell viability assays or a t-test for the lipid peroxidation assay. **p < 0.01, ***p < 0.001, ****p < 0.0001, ns = non-significant.

Fig. 2

Changes to redox in hypoxic conditions. A. A549 cells were exposed to 21, 2 and < 0.1 % O₂ (6 h) followed by staining for total ROS (CellROX). H₂O₂ (25 μM, 3 h) was used as a positive control to induce ROS. Representative images are shown. Scale bar = 12 μm B., Quantification of total ROS fold change in A549 cells stained using the CellROX assay in 21, 2, <0.1 % O₂ (6 h). H₂O₂ (25 μM) was added as a control that induces ROS. C. A549 cells were exposed to 21, 2 and < 0.1 % O₂ (6 h) followed by staining for mitochondrial ROS (MitoSOX). MitoPQ (10 μM, 6 h) was used as positive control. Representative images are shown. Scale bar = 12 μm. D. Quantification of mitochondrial ROS fold change in A549 cells stained using the MitoSOX assay in 21, 2 and < 0.1 % O₂ (6 h). MitoPQ (10 μM) was added as a control that induces mitochondrial ROS. E. The fold change in TrxR activity was determined in A549 cells exposed to 21, 2 or <0.1 % O₂ for 6 h F. A549 cells were exposed to the O₂ levels shown (6 h) and GSH levels were determined (GSH GLO assay). In each case NEM (200 μM, 6 h) was included to verify that the luminescence recorded was due to GSH. G. A549 cells were exposed to the O₂ levels shown (6 h) and GSH levels determined using FL-1. In each case NEM (200 μM, 6 h) was included to verify that the fluorescence recorded was due to GSH. H. A549 cells were exposed to 2 % O₂ (0, 0.5, 2, 6, 12 h) followed by staining for mitochondrial ROS (MitoSOX). I. A549 cells were exposed to 2 % O₂ (0, 0.5, 2, 6, 12 h) followed by determination of GSH levels using FL-1. J. The data shown in parts H and I are plotted together to illustrate the differential kinetics of hypoxia induced mitochondrial ROS levels (MitoSOX) and GSH levels (FL-1) in A549 cells exposed to 2 % O₂ (0, 0.5, 2, 6, 12 h). Data shown in A, B, C, D, E, F, G, H and I are n = 3/4. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate, except for the IF). Data presented as mean + SEM. Statistical testing was done using an unpaired t-test. *p < 0.05, **p < 0.01, ns = non-significant.

Fig. 3

Hypoxic cells are less sensitive to Ag5. A. Schematic representation of cell viability assays with Ag5 with pre-exposure to hypoxic conditions. Cells were placed in hypoxia for 6 h and then treated with Ag5 for 1 h B. A549 cells were exposed to 21, 2, 0.5 and < 0.1 % O₂ (6 h) before treatment with or without Ag5 (1 μM) for 1 h. After 20 h samples were harvested for western blotting of HIF-1α and β-actin as a loading control. C. A549 cells were exposed to 21, 2, 0.5 and < 0.1 % O₂ (6 h) and then treated with Ag5 (0, 0.5, 1, 1.25, 1.5 μM) for 1 h, followed by an MTT assay 20 h later. D. A549 cells were exposed to 21, 2, 0.5 and < 0.1 % O₂ (6 h) and then treated with Ag5 (0, 0.5, 1, 1.25, 1.5 μM) for 1 h, followed by a colony survival assay. E. Representative images of the colonies formed in part D. F. A549 cells were exposed to 21 or 2 % O2 (4 h) followed by Ag5 treatment (5 min) and staining for mitochondrial ROS (MitoSOX). G. A549 cells were exposed to 21 and 2 % O₂ (4 h) and then treated with Ag5 (0, 1, 2, 3 μM) for 15 min, followed by western blotting for PRDX3, with GAPDH as a loading control. H. Quantification of the percentage of oxidised/reduced PRDX3 in A549 cells after exposure to 21 and 2 % O₂ (4 h) followed by Ag5 treatment (15 min) and western blotting. I. Schematic representation of cell viability assays with Ag5 with no pre-exposure to hypoxic conditions. Cells were exposed to Ag5 for 1 h J. A549 cells were treated with Ag5 (1 μM) and then immediately exposed to 21, 2, 0.5 and < 0.1 % O₂. Media was changed after 1 h and western blotting was carried out 20 h later for HIF-1α with β-actin as a loading control. K. A549 cells were treated with Ag5 (0, 0.5, 1, 1.25, 1.5 μM) and then immediately exposed to 21, 2, 0.5 and < 0.1 % O₂. Media was changed after 1 h and an MTT assay was carried out 20 h later. Data shown in C, D, E, F, G, H and K are n = 3/4. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate). Data presented as mean + SEM. Statistical testing was done using a two-way ANOVA test with each bar compared to the 21 % O₂ counterpart. *p ≤ 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns = non-significant.

Fig. 4

Sensitivity to Ag5 is HIF-1α dependent. A. RKO and RKO^{HIF−1α−/−} cells were treated with Ag5 (0, 0.5, 0.75, 1, 1.25 μM) for 1 h in 21 % O₂ followed by an MTT assay 20 h later. B. RKO and RKO^{Hif−1α−/−} cells were exposed to 21 % and <0.1 % O₂ (8 h) followed by western blotting for HIF-1α, with β-actin as a loading control. C. RKO and RKO^{HIF−1α−/−} cells were exposed to 2 % O₂ (6 h), followed by treatment with Ag5 (0, 0.5, 0.75, 1, 1.25 μM) for 1 h and then an MTT assay 20 h later D. Untreated RCC4 and RCC4^VHL−/− cells were western blotted for VHL, HIF-1⍺ and a loading control (β-actin). E. RCC4 and RCC4^VHL−/− cells were treated with Ag5 (0, 0.5, 0.75, 1, 1.25 μM) for 1 h in 21 % O₂ followed by an MTT assay 20 h later. Data shown in A, C and E are n = 3. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate). Data presented as mean + SEM. Statistical testing was done using a two-way ANOVA test. *p < 0.05, **p < 0.01, ****p < 0.0001, ns = non-significant.

Fig. 5

The role of HIF-1 in redox. A. RKO and RKO^{HIF−1α−/−} cells were exposed to 21 or 2 % O₂ (6 h) followed by staining for total ROS (CellROX). Menadione (100 μM, 6 h) was used as a positive control to induce ROS. Scale bar = 20 μm B. Quantification of total ROS in RKO and RKO^{HIF−1α−/−} cells stained using the CellROX assay in 21 and 2 % O₂ (6 h). As a control, menadione (100 μM) was added to induce ROS. C. RKO and RKO^{HIF−1α−/−} cells were exposed to 21 or 2 % O₂ (6 h) followed by staining for mitochondrial ROS (MitoSOX). Menadione (100 μM, 6 h) was used as a positive control to induce ROS. Scale bar = 20 μm. D. Quantification of mitochondrial ROS in RKO and RKO^{HIF−1α−/−} cells stained using the MitoSOX assay in 21 and 2 % O₂ (6 h). As a control, menadione (100 μM) was added to induce ROS. E. RKO and RKO^{Hif−1α−/−} cells were exposed to the O₂ levels shown (6 h) and GSH levels determined (FL-1). In each case NEM (200 μm, 6 h) was also included to verify that the fluorescence recorded was due to GSH. Data shown in A, B, C, D and E are n = 3. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate, except for the IF). Data presented as mean + SEM. Statistical testing was done using an unpaired t-test. *p < 0.05, **p < 0.01, ***p < 0.001 ns= non-significant.

Fig. 6

Sensitivity to Ag5 in hypoxia can be restored by altering cellular redox. A. A549 cells were either untreated or treated with FCCP (6 h), Ag5 (1 h) or both FCCP (6 h) and Ag5 (1 h) followed by an MTT assay 20 h later. B. Levels of PDK1, PDK3 and a loading control (β-actin) are shown in untreated RKO and RKO^shPDK1/3 cells. C. RKO and RKO^shPDK1/3 cells were exposed to 2 % O₂ (0, 0.5, 1, 2, 3, 6 h) followed by staining for mitochondrial ROS (MitoSOX). D. RKO and RKO^shPDK1/3 cells were treated with Ag5 (0, 0.25, 0.5, 0.75, 1, 1.25 μM) for 1 h in normoxia (21 % O₂) followed by an MTT assay 20 h later. E. RKO and RKO^shPDK1/3 cells were exposed to hypoxia (2 % O₂) (6 h), followed by treatment with Ag5 (0, 0.5, 0.75, 1, 1.25 μM) for 1 h followed by an MTT assay 20 h later. F. Schematic representation of the experimental set up for combining Ag5 with radiation in hypoxia (not to scale). Importantly, radiation (indicated with black lightning strike) was delivered in hypoxic conditions i.e. without reoxygenation. Colonies were then allowed to form in normoxic (21 % O₂) conditions. G. A549 cells were exposed to <0.1 % O₂ (4 h), followed by treatment with Ag5 (1 μM) for 1 h. During Ag5 treatment, cells were irradiated (0, 2, 4 Gy) in hypoxic conditions. Cells were harvested after the 1 h treatment and western blotted for HIF-1α, with β-actin as a loading control. H. A549 cells were treated as in part E followed by return to normoxic conditions (21 % O₂), and colony survival assay. Data shown in A, C, D, E and H are n = 3. Black dots on the graphs shown represent biological repeats (each of which was carried out in triplicate). Data presented as mean + SEM. Statistical testing was done using a two-way ANOVA test for cell viability and unpaired t-test for colony survival. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 ns= non-significant.