Alantolactone promotes ER stress-mediated apoptosis by inhibition of TrxR1 in triple-negative breast cancer cell lines and in a mouse model

3.1 ATL effectively induces cell death in human TNBC cells

Previous studies have reported that ATL can suppress cell viability in several tumour types. Recent evidence suggests that ATL (Figure 1A) can inhibit human breast cancer cell line, MDA-MB-231 cell growth. To better understand the effect of ATL on TNBC cell viability, we treated three TNBC cell lines, MDA-MB-231， BT-549 and MDA-MB-468, under increasing concentration of ATL (10, 20, 30 μmol/L) for 24 h. The cell viability was assessed by MTT assay. As shown in Figure 1B, ATL decreased the viability of MDA-MB-231, BT-549 and MDA-MB-468 cells in a dose-dependent manner, with IC₅₀ values of 14.22, 12.15 and 14.8 μmol/L respectively.

We assessed whether ATL-induced suppression of cells viability was associated with the increased apoptosis in cancer cells. Based on cell viability data, we selected concentration of 10, 20 and 30 μmol/L ATLs for the Annexin V/Propidium Iodide (PI) analyses. Our results showed that ATL induced significant dose-dependent increased apoptotic death in TNBC cells (Figure 1C,D). Next, we characterized the nucleic morphological change by Hoechst staining. The control cells displayed normal, round nuclei with faint and even fluorescence, while the cells treated with ATL (30 μmol/L) exhibited highly and condensed fluorescent nuclei, a characteristic morphology of apoptotic cells (Figure S1A). To be more rigorous, we treated MDA-MB-231, BT-549 and MDA-MB-468 cells with ATL and examined the levels of apoptosis-related proteins by western blotting. Treatment of TNBC cells with ATL decreased Bcl-2 (Figure 1E) while increased the levels of Bax. As caspase-3 is a crucial component of the apoptotic machinery, and the cleavage of caspase-3 is a central event in the process of apoptosis, we further examined the expression of cleaved caspase-3. As shown in Figure 1E, the cleaved forms of caspase-3 were increased in TNBC cells exposed to ATL. These data suggest that ATL could induce apoptosis of TNBC cells.

3.2 ATL effectively inhibits proliferation and colony formation of human TNBC cells

We next assessed the effect of ATL on the TNBC cell growth. Propidium iodide staining of cells revealed remarkable accumulation of cells in the G2/M cell cycle phase in a concentration-dependent manner, with a decrease in the population of cells in G1 and S phase, following exposure to ATL (Figure 2A,B). Western blotting analysis of G2/M phase-related proteins also showed that obviously reduced expression levels of CyclinB1 and Cdc2 after treatment of cells with ATL (Figure 2C).

Furthermore, colony formation assay also showed that ATL significantly prevented colony formation of TNBC cells at 5 and 10 μmol/L levels (Figure 2D).

3.3 ATL increases ROS levels in human TNBC cells

Reactive oxygen species accumulation has been reported to induce apoptosis in several cancer cells.27, 28 To examine whether apoptosis and reduced viability in TNBC cells following ATL treatment are involved in increased ROS levels, we further treated cells using 2',7'-Dichlorodihydrofluorescein diacetate (DCFH-DA), which is rapidly oxidized by ROS to produce fluorescence 2',7'-dichlorofluorescein (DCF) inside of cells. Treating MDA-MB-231 cells with ATL induced significant dose-dependent increased fluorescence intensity compared with those of non-treated cells (Figure 3A-C). Next, we checked if ATL-induced ROS is repressed by antioxidant. Pre-treatment of MDA-MB-231 cells with N-acetyl cysteine (ROS scavenger; N-acetyl-L-cysteine (NAC), 1 mmol/L) significantly reduced ATL-induced increases in DCF fluorescence as expected (Figure 3D,E).

We next wanted to check whether the negative growth signals from ATL can be attenuated by reduction of ROS levels. Pre-treatment of cells with NAC (1 mmol/L) was able to attenuate ATL-induced apoptosis (Figure 3F; Figure S1A,B) and cell cycle arrest (Figure 3G; Figure S1C). As expected, NAC prevented ALT-induced alterations of apoptosis-related proteins, Bcl-2, Bax and cleaved caspase-3; and G2/M-phase proteins, cyclinB1 and CDC2 (Figure 3H). Furthermore, pre-treatment of cells with NAC significantly attenuate the inhibition of colony formation (Figure 3I) caused by ATL. These finding suggests that the anti-tumour effect of ATL, at least partly, depends on increasing ROS levels in human TNBC cells.

3.4 ROS-dependent ER stress pathway contributes to ATL lethality in TNBC cells

Increased ROS levels and imbalance of the intracellular redox status increase unfolded proteins in the ER and induce ER stress response.29 Endoplasmic reticulum stress or unfolded protein response (UPR) induces protein kinase R (PKR)-like ER kinase (PERK)-mediated phosphorylation of eIF2a. Phosphorylated-eIF2α allows preferential translation of activating transcription factor 4 (ATF4), while blocking cap-dependent protein translation. In the ER stress pathway, ATF4 is a key transcription factor which mediates the induction of the pro-death transcriptional regulator DNA damage inducible transcript 3 (CHOP). Thus, we tested the expressions of these ER stress-related proteins in TNBC cells following ATL treatment, and examined whether alternation of these proteins related to the ATL-mediated ROS generation. The time course result shows that ATL (30 μmol/L) could significantly induce the phosphorylation of eIF2α, protein levels of ATF4 and CHOP in MDA-MB-231 cells. And the expression levels of these ER stress-related proteins reached the peak about 9 h following treatment (Figure 4A). We further noted that dose-dependent activation of ER stress by ATL treatment in TNBC cells (Figure 4B; Figure S2A). Moreover, pre-treatment of cells with NAC (1 mmol/L) markedly attenuated the induction of these ER stress related proteins (Figure 4C; Figure S2B). We next investigated the effect of ATL on the morphology of ER in MDA-MB-231 cells. Electron microscopy showed that a 9 h ATL challenge is conspicuous to cause ER swelling, while pre-treatment with NAC totally reversed this morphological alteration (Figure 4D). Electron microscopy indeed revealed swollen mitochondria with disrupted cristae in cells exposed to ATL (Figure S3A). To determine ATL treatment indeed causes ER stress and UPR response in TNBC cells, we evaluated XBP1S and ATF6 expression in TNBC cells after treatment with ATL. The IRE1α-XBP1S pathway has been reported to enhance or suppress cancer progression in different contexts.30, 31 On the other hand, ATF6α-dependent UPR has shown cytoprotective functions leading to oncogenic roles in tumourigenesis.32, 33 We found that ATL treatment markedly increased XBP1S expression (Figure S2C), but decreased ATF6 expression in MDA-MB-231 cells (Figure S2D) indicating that ER stress and a UPR response are indeed activated after treatment with ATL. Next, knockdown analysis was performed to further investigate that sustained ER stress indeed contributes to the ALT-induced cell death. Knockdown of ATF4 by siRNA markedly reduced ATF4 expression in MDA-MB-231 cells (Figure 4E), accordingly we observed that reduced apoptotic cells in ATF4 knockdown cells treated with ALT (Figure 4F,G). These findings indicate that ALT-induced cell death is at least in part a result of sustained ER stress.

3.5 TRXR1 is up-regulated in clinical TNBC cancer tissues and inactivated by ATL

Thioredoxin reductase 1 overexpression has been reported in several malignancies. Overexpression of TrxR1 was associated with aggressive tumour growth and poor survival. In addition, it has been reported that TrxR system contributes to tumour cell resistance to oxidative stress.18 We firstly examined TrxR1 mRNA expression in TNBC. GSE59590 dataset showed that TrxR1 mRNA expression was significantly up-regulated in patients with TNBC compared to the patients without TNBC in breast cancer specimens (Figure 5A). The Kaplan-Meier survival analysis using 3021 available breast cancer patients from the Kaplan-Meier Plotter showed that high expression of TrxR1 was significantly associated with poor prognosis of breast cancer patients (Figure 5B). We further collected 20 breast cancer specimens from the patients with TNBC and checked TrxR1 levels in biopsy specimens by using immunochemical staining (Figure 5C-D). The clinical information of patients was listed in the Table S1. Our results showed that TNBC specimens displayed significantly increased TrxR1 immunoreactivity compared with the normal adjacent breast tissue from the same patient. Of note, 85% TNBC tissue specimens showed moderate to strong TrxR1 expression; however, only 5% normal adjacent tissues present moderate to strong TrxR1 expression. These results indicate that TrxR1 expression is significantly up-regulated in TNBC tissues. Further, we tested the TrxR1 enzyme activity by using the 5,5-dithio-bis-(2-nitrobenzoic acid) (DTNB) assay. As shown in Figure 5E and F, the activities of TrxR1 in tumours were significantly up-regulated compared to the corresponding normal breast specimens. All together, these findings indicate that TrxR1 might play pivotal functions in TNBC carcinogenesis.

We next want to know whether TrxR1 is a target of ATL in TNBC cells. A recent study showed that ATL inhibits the recombinant TrxR1 in HeLa cells.34 To investigate the structural mechanism of ATL binding to the TrxR1 protein, we performed a molecular simulation of ATL-TrxR1 complex using AutoDock. Our result showed that ATL not only can insert into the C-terminal active site of TrxR1 but also form a strong covalent bond (Figure 5G). It has been reported that the redox motif containing Cys-497, Sec-498 plays a vital role in enzyme inactivation, thus competitive inhibition to these residues could significantly desensitize the enzyme.35, 36 During the docking process, the alkenyl in ATL was detected as Michael acceptor to form a hard covalent bond with Cys-497 while the cyclohexane part inserted into the hydrophobic pocket. Thus occupying the redox active centre may block the nature enzymatic recognition. This docking study suggests that TrxR1 is the potential target of ATL and blocking the critical residues in redox centre could inhibit its enzyme activity. We further tested the direct inhibitory effects of ATL on TrxR1 enzyme activity by using DTNB assay. When lysates prepared from MDA-MB-231 cells were incubated with various concentrations of ATL for 2 h, the DTNB reducing activity of TrxR1 decreased in a dose-dependent manner (Figure 5H). We confirmed these results by measuring TrxR1 enzyme activity in pre-reduced recombinant human TrxR1 (Figure 5I). We next examined the importance of TrxR1 in the cytotoxic effect of ATL. Knockdown of TrxR1 by two different TrxR1 siRNAs markedly reduced TrxR1 expression in MDA-MB-231 cells (Figure 5J). The TrxR1 knockdown by either siRNA1 or siRNA2 resulted in increased apoptosis and ROS levels in MDA-MB-231 cells induced by ATL (Figure 5K; Figure S4A,B). These findings indicate that ATL inactivates TrxR1, consequently increased cellular ROS in human TNBC cells and TrxR1 plays critical functions in ATL-induced cell apoptosis in TNBC.

3.6 ATL inhibits the growth of MDA-MB-231 cell xenograft in vivo, accompanied with decreased TRXR1 activity and increased ROS level

We found that ATL effectively induces apoptosis and inhibits proliferation in TNBC cells through ROS-dependent ER-stress pathway by decreasing TrxR1 activity. To determine the impact of ATL in vivo, we produced MDA-MB-231 xenografts in nude mice. We treated mice with ATL at 15 or 30 mg/kg for 20 days. NAC (0.5 g/L) was administered in the drinking water for a group of mice which are treated with 30 mg/kg ATL for the same days. The high dose (30 mg/kg) of ATL decreased tumour volume in mice, while the NAC-treated group was able to diminish the ATL-induced tumour growth inhibition (Figure 6A,B). Without significant changes in body weights were observed in ATL-treated mice (Figure 6C). Haematoxylin and eosin staining analyses of vital organs (heart, liver and kidney) did not show any histological change that would indicate ATL toxic effects (Figure 6D). Furthermore, we observed that ATL treatment increased the levels of lipid peroxidation product malondialdehyde (MDA) in tumour tissues (Figure 6E) indicating increased ROS production. In addition, NAC significantly abolished the increasing MDA level induced by ATL (Figure 6E). We next stained the tumour tissues with ROS-sensitive Dihydroethidium (DHE) and DCFH-DA. DHE forms a red fluorescent product upon reaction with ROS and intercalates with DNA. Microscopic examination revealed that increased DHE and DCFH-DA fluorescence signals in tumour tissues from mice with ATL treatment, indicating increased ROS levels in mice with ATL treatment (Figure 6F). NAC treatment reversed the ATL-induced increase in ROS levels (Figure 6F). Western blot analyses of the tumour tissues showed that ATL treatment increased the levels of ATF4, CHOP (Figure 6G) indicating activation of the ER-stress pathway in vivo. As expected, NAC prevented ALT-induced alterations of ER-stress related proteins (Figure 6G). We next assessed ATL-induced apoptosis by cleaved caspase-3 levels through both Western blot analyses and immunohistochemistry. We found that cleaved caspase-3 levels were increased in tumour tissues following treatment with ATL (Figure 6G,H), while cell proliferation (ki-67 levels) was decreased (Figure 6H). Importantly, NAC treatment reversed the ALT-induced alterations of cleaved caspase 3 and ki-67 levels (Figure 6G-H). More importantly, ATL also inhibited TrxR1 activity in MDA-MB-231 xenografts (Figure 6I). These results are consistent with the in vitro studies and show that ATL inhibits tumour growth by suppressing TrxR1 activity, subsequently elevates ROS production and induces ER-stress-related apoptosis.