Cardiac glycosides inhibit cancer through Na/K-ATPase-dependent cell death induction

Abstract

Successful drug repurposing relies on the understanding of molecular mechanisms of the target compound. Cardiac glycosides have demonstrated potent anticancer activities; however, the pharmacological mechanisms underlying their anticancer effects remained elusive, which has restricted their further development in cancer treatment. A bottleneck is the lack of comprehensive understanding about genes and signaling pathways that are altered at the early stage of drug treatment, which is key to understand how they inhibit cancer. To address this issue, we first investigated the anticancer effects of a panel of 68 naturally isolated cardiac glycosides. Our results illustrate critical structure activity relationship of these compounds on cancer cell survival. We confirmed the anticancer effect of cardiac glycoside in mouse tumor xenografts. Through RNA sequencing, quantitative PCR and immunoblotting, we show that cardiac glycoside first activated autophagy and then induced apoptosis. Further activating autophagy by rapamycin or inhibiting apoptosis by caspase inhibitor mitigated cardiac glycoside-induced cell death, whereas inhibiting autophagy by RNA interference-mediated depletion of critical autophagy genes enhanced cell death. While depletion of Na/K-ATPase, the protein target of cardiac glycosides, by RNA interference inhibited both autophagy activation and apoptosis induction by cardiac glycoside, expression of human, but not rodent Na/K-ATPase, increased cell sensitivity to cardiac glycoside. In conclusion, our analyses reveal sequential activation of autophagy and apoptosis during early stages of cardiac glycoside treatment and indicate the importance of Na/K-ATPase in their anticancer effects.

Keywords: Anticancer; Apoptosis; Autophagy; Cardiac glycosides; Na/K-ATPase.

Fig. 1.

Anticancer effects of cardiac glycosides. (A) Structural backbone of 68 cardiac glycosides studied herein. (B) SAR of survival inhibition rate by cardiac glycosides in U2OS cells. 68 natural cardiac glycosides were tested and presented as the Box-Whisker plot showing 1st and 3rd quartile, based on the effects of the substitution groups. The more inhibition, the stronger the anticancer effect it has. (C) Structure of ANTO2. (D) IC₅₀ of ANTO2 in A549 cells representing average and standard deviation from five replicates. (E) A549 cells were treated with 500 nM ANTO2 for indicated times and cell cycle profile was analyzed by FACS. (F) Clonogenic survival assay of A549 cells treated with 500 nM ANTO2 for indicated times. Data represent average and SEM from 3 replicates. (G) Effect of ANTO2 on A549 tumor growth in nude mice. Relative tumor growth is obtained by normalizing the tumor volume to that at day 1. Data represent mean and standard deviation from 10 mice per group. Two-tailed t-student test was used to determine the statistical significance. * P < 0.001 between control and ANTO2 group.

Fig. 2.

Transcriptional profiling of U2OS cells after ANTO2 treatment. U2OS cells were treated with 500 nM ANTO2 in duplicate for 2 and 8 h, and RNA seq was performed. Differentially expressed genes were analyzed. Data criteria for selection are PPEE ≤ 0.05 and absolute value of log2 (fold change) ≥ 1. Scatter plot of all expressed genes in control versus ANTO2 treatment 2 h (A) or 8 h (B). Blue and orange dots represent down- and up-regulated genes, respectively. Brown dots represent non-changed genes. Hierarchical cluster analysis of gene expressed with more than 5 supported reads in total across all samples (C) or in control and ANTO2 8h group (D) with scaled and FPKM normalized expression values analyzed by the javaTreeview software. Each column represents an experimental condition whereas each row represents a gene. Expression differences are indicated by colors with red to green representing up- and down-regulated genes, respectively. Data shown are from duplicated samples. Functional annotation of most up-regulated (E) and down-regulated (F) gene pathways using DAVID GOTERM_CC_ALL category. Log2 values were used to cluster all the candidate genes using R pheatmap (v0.7.7) with Euclidean distance and the complete linkage method based on FPKM normalized expression profile for each sample.

Fig. 3.

Differentially expressed genes regulated by ANTO2. (A, B) Violin plots of pathways significantly regulated in ANTO2 8 h group by KEGG. Log2 values were used to cluster all candidate genes (2681 genes) using R pheatmap (v0.7.7) with Euclidean distance and the complete linkage method based on FPKM normalized expression profile for each sample. The p-value is calculated using the wilcox test. (C) U2OS cells were treated with 500 nM ANTO2 for the indicated times and gene expression levels were measured by qPCR. Data represent average and standard deviation from 5 replicates.

Fig. 4.

Cardiac glycoside activates autophagy and apoptosis. (A) U2OS cells were treated with 500 nM ANTO2 for the indicated times, and expression levels of indicated proteins were assessed by specific antibodies. (B) The band intensity of each blot was quantitated by the Image J software and normalized to that in control group. Data represent average and standard deviation from three replicates. *P < 0.05 is considered as significant.

Fig. 5.

Time course effect of ANTO2 on autophagy and apoptosis in A549 cells. (A) A549 cells were treated with indicated concentrations of ANTO2 for 12 h, and protein levels were analyzed by indicated specific antibodies. Arrow indicates cleaved PARP. Short and long exposures for LC3B are shown. (B) The relative protein ratio was quantitated and shown as average and standard deviation from three replicates. *P < 0.005 between control and treated groups.

Fig. 6.

Dose-dependent response of A549 cells to ANTO2. (A) A549 cells were treated with 500 nM ANTO2 for indicated h, and protein levels were analyzed by indicated specific antibodies. Arrow indicates cleaved PARP. (B) The relative protein level ratio was quantitated and shown as average and standard deviation from three replicates. *P < 0.005 between control and treated groups.

Fig. 7.

Induction of autophagy and apoptosis by ANTO2. (A) A549 cells were treated with 500 nM ANTO2 for 0, 2, 4 and 12 h, stained with anti-LC3B antibodies and visualized under fluorescence microscopy. Representative images are shown. Bar is 10μm. (B) Fluorescence intensity of LC3B in each cell from A was quantitated by the Image J software from the indicated number of cells. (C, D) A549 cells were treated as indicated and cell death rate was measured by the trypan blue exclusion assay. Data represent average and standard deviation from 5 replicates. *P < 0.001 between control and treated groups.

Fig. 8.

Effect of ANTO2 on H1650 and HOP62 lung cancer cells. H1650 (A) or HOP62 (B) cells were treated with 500 nM ANTO2 for 0, 2, 4 and 8 h, and gene expression levels were assessed by qPCR. Data represent average and standard deviation from five replicates. (C) HOP62 cells were treated with 500 nM ANTO2 for the indicated times and protein expression levels were measured by specific antibodies. (D, E) The relative protein level ratio was quantitated and shown as average and standard deviation from three replicates. *P < 0.005 between control and treated groups.

Fig. 9.

Roles of autophagy and apoptosis in ANTO2-induced cell death. (A) A549 control, ULK1 or BECN1 depleted cells were treated with DMSO, 10 μM Z-VAD-FMK or 500 nM ANTO2 for indicated hours, and protein expression was analyzed using specific antibodies. (B) The relative protein level ratio was quantitated and shown as average and standard deviation from three replicates. (C) Quantitation of dead cell population in A by trypan blue assay. Data represent average and standard deviation from 5 replicates. *P < 0.005 between control and treated groups.

Fig. 10.

The protective role of autophagy in ANTO2-induced cell death. (A) A549 cells were treated with 500 nM ANTO2, 200 nM rapamycin (Rapa) or both for indicated times and protein expression was assessed by specific antibodies. (B) The relative protein ratio was quantitated and shown as average and standard deviation from three replicates. (C) Quantitation of dead cell population in A by trypan blue assay. Data represent average and standard deviation from 5 replicates. *P < 0.005 between ANTO2 and ANTO2 + Rapa groups.

Fig. 11.

Roles of Na/K-ATPase in the effects of cardiac glycosides. (A) SAR of 68 natural cardiac glycosides on the enzymatic activity of Na/K-ATPase in vitro using ouabain as the model compound. Relative enzyme activity is presented in log10 scale as the Box-Whisker plot showing 1^st and 3^rd quartile based on the effects of the substitution groups of cardiac glycosides. The lower the enzymatic activity, the more inhibition by the compound. (B) A549 cells were transfected with siRNA control or targeting the α1 subunit of the Na/K-ATPase for 48 h, treated with 500 nM ANTO2 for indicated times and protein expression was analyzed by western blotting using specific antibodies. (C) The band intensity of each lane in B was quantitated and normalized to that of the control group. Data represent average and standard deviation from 3 replicates. (D) Cell death rate of cells from B. Data represent average and standard deviation from at least 5 replicates. *P < 0.05 is considered significant.

Fig. 12.

Human but not rodent Na/K-ATPase is responsive to cardiac glycosides. (A) A549 cells or mouse embryonic fibroblasts (MEFs) were treated with 500 nM ANTO2 for indicated times and protein expression was examined. (B) The band intensity of each lane in A was quantitated and normalized to that of the control group. Data represent average and standard deviation from 3 replicates. (C) Cell death rate of A549 and MEF cells after 12 h ANTO2 treatment in A was analyzed by trypan blue staining. *P < 0.005. (D) MEF parental cells or those expressing either the wild type human Na/K-ATPase α1 subunit or a rat α1 mutant (D376E) that does not bind cardiac glycoside were treated with 500 nM ANTO2 or not for 12 h, and the dead cell rate was counted by trypan blue assay. Data represent average and standard deviation from 5 replicates. *P < 0.05 was considered significant.