HuR inhibition overcomes cFLIP-mediated doxorubicin resistance in triple-negative breast cancer

Abstract

Triple-negative breast cancer (TNBC) presents therapeutic challenges due to limited targeted treatment options and resistance to chemotherapy drugs, such as doxorubicin. This study investigated doxorubicin resistance mechanisms and a strategy to overcome it. A doxorubicin-resistant cell subline (231-DR) was developed from MDA-MB-231 TNBC cells, and enhanced expression of cellular FLICE-inhibitory protein (cFLIP) in 231-DR cells was identified as a potential driver of the resistance. Overexpression of cFLIP conferred resistance to doxorubicin-induced apoptosis, whereas siRNA-mediated cFLIP depletion induced apoptosis, particularly in 231-DR cells. Furthermore, the RNA-binding protein Hu antigen R (HuR) was found to regulate cFLIP expression. HuR Inhibition with KH-3 or RNA interference reduced cFLIP levels. Importantly, KH-3 sensitized TNBC cells to doxorubicin-induced apoptosis. In summary, this study delineates cFLIP’s role in mediating doxorubicin resistance and identifies HuR as a positive regulator of cFLIP, offering a novel therapeutic avenue against chemoresistance in TNBC by combining HuR inhibition with doxorubicin.

Introduction

Triple-negative breast cancer (TNBC) represents a distinct molecular subtype of breast cancer characterized by the absence of estrogen and progesterone receptors, as well as the lack of human epidermal growth factor receptor 2 (HER-2) overexpression^1,2. In the United States, TNBC accounts for 10% of all diagnosed breast cancers in women, with a five-year relative survival rate of 77%, compared to 90% for overall breast cancer³. Despite the emergence of targeted therapies, chemotherapy remains the primary treatment modality⁴. However, the efficacy of chemotherapy is greatly compromised by chemotherapeutic resistance, leading to more than 90% treatment failure in metastatic cancer⁵. Therefore, unraveling the mechanisms underlying chemoresistance and identifying strategies to overcome it are critical for improving the survival rate of TNBC patients.

Doxorubicin (DXR), a DNA topoisomerase II inhibitor, is a widely used chemotherapeutic agent for various types of cancers, including TNBC. DXR is an anthracycline antibiotic extracted from Streptomyces peucetius spp^6,7. The anticancer effects of DXR are achieved through multiple mechanisms, including induction of cell death, arrest of DNA repair, and generation of oxidative stress to cellular membranes^6,8. TNBC cells employ numerous mechanisms to develop resistance to chemotherapeutic drugs, including ABC transporters, cancer stem cells, hypoxia-induced chemoresistance, evasion of apoptosis, microRNAs, and high heterogeneity¹. In the context of TNBC, doxorubicin resistance has been associated with epigenetic alterations in histone deacetylase⁹, and the activity of ubiquitin specific protease 7 (USP7)¹⁰. Doxorubicin resistance in TNBC can be reversed by inhibiting key proteins that contribute to it, such as lysyl oxidase¹¹, thioredoxin-interacting protein (TXNIP)¹², and P-glycoprotein¹³. Strategies to enhance the efficacy of doxorubicin include combining it with other chemotherapy drugs, such as paclitaxel in esophageal squamous cancer cells¹⁴, or with other target therapy drugs, such as olaratumab in metastatic soft tissue sarcoma cells¹⁵, and gamitrinib in breast and prostate cancer cells¹⁶.

The cellular FLICE-like inhibitory protein (cFLIP) serves as a key regulator of apoptosis, necroptosis, and autophagy^17,18. The CFLAR gene in human cells encodes three major protein isoforms of cFLIP: the long form cFLIP_L, and two short forms (cFLIP_S and cFLIP_R)¹⁸. The cFLIP_R is specifically expressed in certain human cells such as human primary T cells¹⁸. The heterodimers formed between pro-Caspase-8 and cFLIP_L or cFLIPs inhibit the activation of Caspase-8, thus preventing Caspase-8-mediated apoptosis^17,18. Moreover, DXR is known to induce apoptotic cell death in tumor cells¹⁹. A study showed that resistance to doxorubicin-induced apoptotic cell death in lung cancer cells H460 is related to the inhibition of Caspase-3 and Caspase-8²⁰. Exogenous expression of cFLIP_L was associated with the delayed activation of Caspase-3 and -8 in H460 cells treated with doxorubicin²⁰. In our study, we established a DXR-resistant cell sub-line and found that cFLIP was overexpressed in these resistant cells. Therefore, we hypothesized that cFLIP may contribute to doxorubicin resistance in breast cancer.

The Hu antigen R (HuR) is a well-characterized RNA-binding protein, which binds to target mRNA, stabilizing the mRNA and/or enhancing the translation of encoded proteins most of the time^21,22. HuR target proteins have been implicated in multiple cancer hallmarks, including chemoresistance²³. While HuR is predominantly located in the nucleus²⁴, its hinge domain allows HuR to shuttle between the nucleus and cytoplasm²⁵. Increased cytoplasmic accumulation of HuR is associated with high-grade tumors, poor clinical outcomes in breast cancer²⁶, and resistance to tamoxifen in breast cancer cells²⁷ and adriamycin in ovarian cancer cells²⁸. Conversely, HuR overexpression is also associated with drug resistance. Overexpression of HuR resulted in chemoresistance in glioma cells²⁹ and colon cancer cells³⁰. Inhibiting HuR expression, translocation, or the HuR-RNA interaction can sensitize cancer cells to anticancer treatments. For instance, siRNA-mediated HuR knockdown sensitizes colorectal cancer cells to irradiation-induced apoptosis³¹. A small molecule HuR inhibitor, MS-444, which inhibits HuR cytoplasmic translocation, can suppress cell proliferation, and induce the expression of apoptotic genes in colorectal cancer cells³². Our group recently reported that KH-3, a small molecule inhibitor that disrupts the HuR-mRNA interaction, sensitized TNBC cells to docetaxel by inducing apoptotic cell death and suppressing proliferation³³.

One study suggested that HuR interacts with the mRNA encoding cFLIP in pancreatic cancer cells, and silencing of HuR attenuates the mRNA stability of cFLIP³⁴. However, the regulation of cFLIP by HuR regulation has not been previously investigated in the context of breast and prostate cancer. Despite these findings, the understanding of the relationship between HuR and cFLIP remains limited.

This study aims to unveil the mechanism and key protein(s) responsible for doxorubicin resistance in TNBC, as well as to identify strategies for enhancing the efficacy of doxorubicin by reversing the identified changes in protein levels. To achieve this, we established a doxorubicin-resistant cell subline (231-DR) derived from the human TNBC cell line MDA-MB-231. Firstly, the expression levels of multiple anti-apoptotic genes were measured to identify the potential proteins associated with doxorubicin resistance. Subsequently, the regulation of HuR on these genes was investigated and verified. Additionally, the combination of doxorubicin and the HuR inhibitor KH-3 was performed. The results showed that cFLIP was obviously increased in 231-DR cells and the exogenous expression of cFLIP attenuated the doxorubicin-induced apoptosis. Moreover, HuR inhibition downregulated the expression of cFLIP and overcame cFLIP-mediated doxorubicin resistance. Furthermore, the combination of KH-3 and doxorubicin synergistically suppressed cell proliferation and induced apoptotic cell death.

Results

cFLIP was overexpressed in doxorubicin-resistant 231-DR cells

To mimic clinically acquired resistance, the human TNBC cell line MDA-MB-231 was maintained with the presence of doxorubicin (DXR). The DXR concentration was increased gradually until a cell sub-line with heightened resistance to DXR (231-DR) was established. The 231-DR cells were cultured with 150 nM DXR, which was removed before the experiment. An MTT-based cytotoxicity assay evaluated the induced resistance to doxorubicin. The IC₅₀ of doxorubicin for 231-DR cells was 1.48 µM, over 20-fold higher than the IC₅₀ for the parental MDA-MB-231 cells (0.064 µM) (Fig. 1A). To further validate the effects of KH-3 on 231-DR cells, a clonogenic assay was conducted. The colony number decreased with the increased concentration of doxorubicin (Fig. 1B). The number of 231-DR colonies was smaller than that of the parental cells, possibly due to the longer doubling time, which was further increased by KH-3 treatment (Fig. 3E). And only colonies with over 50 cells were counted. The IC₅₀ of doxorubicin in inhibiting the relative plating efficiency was 32.52 nM in parental cells compared with 623.1 nM in 231-DR cells (Fig. 1C).

Fig. 1. cFLIP was overexpressed in doxorubicin-resistant 231-DR cells.

A Cell viability curves were determined by the MTT-based cytotoxicity assay of MDA-MB-231 cells, and 231-DR cells treated with doxorubicin for 4 days. B Representative images of colonies of cells treated by doxorubicin (left) and the colony numbers/per well (right) determined by the colony formation assay. MDA-MB-231 cells and 231-DR cells were treated with doxorubicin for 9 or 16 days, respectively. Two-way ANOVA test, **P < 0.01, ***P < 0.001, ****P < 0.0001. C Relative plating efficiency of cells treated with doxorubicin in (B). The relative plating efficiency was calculated based on the question: $\mathit{plating\; efficiency}=\frac{\mathit{Colony\; number}\left(\mathit{treat}\right)}{\mathit{Colony\; number}\left(\mathit{control}\right)}$. D RT-qPCR analysis on mRNA abundance in MDA-MB-231 cells and 231-DR cells. Two-way ANOVA test, *P < 0.05, ***P < 0.001 (n = 2). E Western blot analysis of whole cell lysates of MDA-MB-231 cells and 231-DR cells. GAPDH was used as the loading control. F Western blot analysis of whole cell lysates of MDA-MB-231 cells transiently transfected with control vector or vector containing cFLIP_L ORF. Cells were treated with the vehicle control (DMSO) or doxorubicin for 48 h post 24 h transfection. GAPDH was used as the loading control. G Cell viability of cells transiently transfected with cFLIP siRNA. Cells were harvested post 3 d transfection and the cell viability was determined by the AO/PI staining assay. The cell viability of cells transfected with siRNA with a non-targeting sequence was served as the negative control (siNC) to provide a baseline reference for the target-specific knockdown. Two-way ANOVA test, ***P < 0.001, ****P < 0.0001 (n = 2). All results are shown as mean ± SD.

Fig. 3. The HuR inhibitor KH-3 inhibited cell growth in 231-DR cells.

A The cell viability curves determined by the MTT-based cytotoxicity assay of MDA-MB-231 cells, and 231-DR cells treated with KH-3 for 4 days. B Representative images of colonies of cells treated by KH-3 (left) and the colony numbers/per well (right) determined by the colony formation assay. MDA-MB-231 cells and 231-DR cells were treated with KH-3 for 9 or 16 days, respectively. Two-way ANOVA test, *P < 0.05, **P < 0.01, ****P < 0.0001. C Relative plating efficiency of cells treated with KH-3 in (B). The relative plating efficiency was calculated as described before. D The percent of cell death. The cell viability was determined by the trypan blue solution staining assay. Cells were treated with DMSO (vehicle control), 5 μM or 10 μM KH-3. Two-way ANOVA test, *P < 0.05, **P < 0.01, ***P < 0.001. E Growth curves of viable cells. F Western blot analysis of whole cell lysates of MDA-MB-231 cells. Prior to KH-3 treatment for 24 h, cells were pretreated with Z-VAD-FMK for 1 h. GAPDH was used as the loading control. All results are shown as mean ± SD of three replicates. All experiments were performed for three repeats, similar results were found, and representative results were shown.

One major well-established doxorubicin-induced cell death mode is apoptosis, although other modes have been reported^35–37. It is reported that resistance to doxorubicin-induced apoptotic cell death is associated with the inhibition of Caspase-3 and Caspase-8 activation²⁰. Therefore, certain anti-apoptotic proteins may contribute to doxorubicin resistance in breast cancer cells. We compared the expression of multiple genes with anti-apoptotic activities via inhibiting the activation of caspase proteins. The mRNA level of CFLAR was increased in 231-DR cells compared to the parental 231 cells, along with a decrease in BIRC4 (encoding for XIAP) and BIRC5 (encoding for Survivin) mRNA levels (Fig. 1D). Similarly, protein analysis revealed a noticeable increase in cFLIP_L and cFLIP_S levels in 231-DR cells, but no significant increase was observed in XIAP and Survivin (Fig. 1E).

To test our hypothesis that cFLIP may contribute to doxorubicin resistance, we manipulated the expression of cFLIP. Transient transfection of DNA constructs containing the cFLIP_L ORF (encoding for the canonical cFLIP_L isoform) was performed to determine whether cFLIP overexpression rescued cells from doxorubicin-induced cell death. Both low- and high-dose doxorubicin treatments (1.0 µM and 1.5 µM) in MDA-MB-231 cells induced the activation of Caspase-8, Caspase-3, and the cleavage of PARP, which was attenuated by cFLIP overexpression (Fig.1F). We also used another human TNBC cell line, SUM159, to test whether cFLIP overexpression would reduce doxorubicin-induced cell death (Supplementary Fig. 1). Interestingly, the presence of doxorubicin reduced cFLIP level in the overexpression group nearly to the basal level, which is the possible reason of no obvious attenuation of doxorubicin-induced apoptosis markers. Reversely, we used RNA interference to knockdown cFLIP levels in both MDA-MB-231 and 231-DR cells. siRNA-mediated cFLIP silencing significantly reduced cell viability in both cell lines compared to the negative control siRNA (siNC), resulting in a more pronounced decrease in viability in 231-DR cells compared to MDA-MB-231cells (Fig.1G). Furthermore, the IC50 of doxorubicin was decreased by 50% in 231-DR cells with cFLIP knockdown compared to cells treated with the control siNC (Supplementary Fig. 2). These findings suggest that cFLIP overexpression confers doxorubicin resistance to parental MDA-MB-231 cells.

HuR bound to cFLIP and upregulated its expression

The RNA-binding protein HuR binds to target mRNA, stabilizes the mRNA, and/or enhances the translation of the encoded proteins most of the time^21,22. Both HuR overexpression^29,30 and increased HuR cytoplasmic accumulation^27,28 contribute to drug resistance. Therefore, total and cytoplasmic protein levels of HuR were measured in this study. The 231-DR cells showed a slightly enhanced cytoplasmic HuR protein level as compared to the parental cells, while no significant difference was observed in the whole-cell lysate (Fig. 2A). The search results from the Encyclopedia of RNA interactomes (ENCORI) database³⁸ (https://rnasysu.com/encori/) (Supplementary Table 5) indicated that HuR binds to the 3′ or 5′ UTR of mRNA transcripts of CFLAR, BIRC4, and BIRC5. The RNA levels of CFLAR and ELAVL1 show a positive correlation in breast invasive carcinoma (BRCA) tumor samples from the TCGA databases, as analyzed using GEPIA2³⁹ (Supplementary Fig. 3). The RNP-IP results confirmed the binding of HuR to the above mRNA transcripts (Fig. 2B), suggesting that increased cytoplasmic HuR, along with the subsequent elevation of encoded proteins from HuR target mRNA transcripts, may be critical for doxorubicin resistance.

Fig. 2. HuR bound to cFLIP and upregulated its expression.

A Western blot analysis of the whole-cell/cytoplasmic lysates of MDA-MB-231 cells and 231-TR cells. The α-Tubulin was used as the loading control. The Lamin A/C was used as the indicator of the presence of nuclear protein. For protein quantification, HuR was first normalized to its corresponding α-Tubulin followed by a second normalization to cytoplasmic HuR in 231 cells. B RNP-IP analysis of HuR bound RNAs in MDA-MB-231 cells. Cell lysates were subjected to RNP-IP followed by the RT-qPCR analysis. The enrichment of RNAs in HuR IP was compared to the control IgG IP. Two-way ANOVA test, **P < 0.01, ****P < 0.0001 (n = 2). C RT-qPCR analysis on the mRNA levels of CFLAR, BIRC4, and BIRC5 in MDA-MB-231 cells and 231-DR cells treated by KH-3 for 8 h, or 24 h. Two-way ANOVA test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (n = 2). D Western blot analysis of whole-cell lysates of MDA-MB-213 cells treated with 10 µM KH-3 from 16 h to 70 h. GAPDH was used as the loading control. E, F Western blot analysis of cell lysates of MDA-MB-231 cells (E) stably transduced doxycycline (DOX) inducible HuR shRNA or (F) transiently transfected with HuR siRNA. GAPDH was used as the loading control. G RT-qPCR analysis of the RNA levels of CFLAR or (H) Western blot analysis of the protein levels of HuR, cFLIP, and downstream targets in parental MDA-MB-231 cells, sgControl and two HuR KO clones treated with 10 µM KH-3 for 24 h. GAPDH was used as the loading control. Two-way ANOVA test, **P < 0.01, ***P < 0.001, ****P < 0.0001 (n = 2). I Western blot analysis of HeLa Tet-Off cells. Cells were maintained with the presence of G418 (100 µg/mL), hygromycin (100 µg/mL), and doxycycline (2 µg/mL) for 72 h prior to harvesting. To induce the overexpression of HuR, doxycycline was completely removed at time zero. Results are shown as mean ± SD of two replicates. All experiments were performed for three repeats, and the representative results were shown.

To reveal the potential roles of HuR in regulating the expression of CFLAR (cFLIP), BIRC4 (XIAP), and BIRC5 (Survivin), the small molecule HuR inhibitor KH-3, which inhibits the function of HuR through intervening HuR-mRNA interactions, was utilized as a tool compound to investigate the impact of KH-3 on the mRNA and protein levels of the above HuR targets. KH-3 treatment significantly downregulated the mRNA levels of CFLAR, BIRC4, and BIRC5 in MDA-MB-231 cells (Fig. 2C). The protein levels of cFLIP_L, XIAP, and Survivin were reduced after 16 or 24 h of treatment and recovered later (Fig. 2D).

To verify whether the decreased protein level of cFLIP was caused by the inhibition of HuR rather than potential unknown off-target effects of KH-3, shRNA, and siRNA targeting ELAVL1 (HuR) were utilized to manipulate HuR expression. Both the inducible shRNA and the transient siRNA system successfully silenced the HuR protein and downregulated the protein level of cFLIP_L in MDA-MB-231 cells (Fig. 2E, F). Although KH-3 treatment reduced the protein levels of both cFLIP_L and cFLIP_S (Fig. 2D), siRNA-mediated HuR knockdown only reduced the protein level of cFLIP_L (Fig. 2F). Therefore, this study primarily focused on the canonical cFLIP_L isoform.

Furthermore, to determine whether KH-3-induced loss of cFLIP was HuR dependent, the CRISPR/CAS9 approach was used to deplete the HuR protein level as described before²⁶. Parental cells and cells with control sgRNA (sgControl) exhibited significantly decreased mRNA levels of CFLAR after 24 h of KH-3 treatment (Fig. 2G). In contrast, two clones with HuR knockout (HuR KO1 and HuR KO2) did not show a decrease in CFLAR levels (Fig. 2G). Similarly, the protein level of cFLIP was depleted by 24 h of KH-3 treatment in sgControl cells, but only partially decreased in HuR KO1 and HuR KO2 cells (Fig. 2H). In addition, noticeably enhanced cleavage of Caspase-8 and PARP, downstream proteins of cFLIP-mediated apoptosis, was observed in sgControl cells treated with KH-3 compared to the vehicle control. However, in HuR KO1 and HuR KO2 cells treated with KH-3, the cleavage of Caspase-8 returned to the “baseline level” of sgControl cells treated with the vehicle control, and the cleavage of PARP was attenuated (Fig. 2H). In contrast, overexpression of HuR led to an increase in the protein level of cFLIP_L (Fig. 2I).

Taken together, our findings suggest that the inhibition of HuR downregulates cFLIP expression, the KH-3-induced loss of cFLIP is, at least in part HuR-dependent, and the cleavage of Caspase-8 and PARP is HuR-dependent. Conversely, cFLIP overexpression contributes to the resistance to doxorubicin-induced apoptotic cell death, suggesting that HuR inhibition could overcome cFLIP-mediated doxorubicin resistance.

The HuR inhibitor KH-3 inhibited cell growth in 231-DR cells

To test the hypothesis, we first validated the effects of HuR inhibition on doxorubicin-resistant 231-DR cells by utilizing the HuR inhibitor KH-3. The MTT-based cytotoxicity assay and colony formation assay were conducted to evaluate the anti-cancer effects of KH-3. The 231-DR cells showed a similar IC₅₀ of KH-3: the IC₅₀ was 4.41 µM for 231-DR cells compared with the 3.31 µM for the parental MDA-MB-231 cells (Fig. 3). Similarly, the colony number decreased with the increasing concentrations of KH-3 (Fig. 3B). KH-3 inhibited colony formation in both cell lines with similar efficacy, with an IC₅₀ of 3.36 µM for MDA-MB-231 and 1.99 µM for 231-DR cells in inhibiting the plating efficiency (Fig. 3C). Results of the trypan-blue staining assay confirmed KH-3-induced cell death in both MDA-MB-231 and 231-DR cells (Fig. 3D). Additionally, the growth curve revealed that KH-3 dose-dependently extended the doubling time of MDA-MB-231 and 231-DR cells (Fig. 3E). Although KH-3 treatment reduced the levels of cFLIP_L, XIAP, and Survivin, the pan-caspase inhibitor z-VAD blocked the subsequent cleavage of Caspase-8, Caspase-3, and PARP (Fig. 3F). Taken together, the results verify that KH-3 inhibits cell proliferation and induces cell death in doxorubicin-resistant 231-DR cells, showing similar effects to parental MDA-MB-231 cells.

KH-3 overcame cFLIP-mediated doxorubicin resistance

It has been reported that cFLIP knockdown induces ligand-independent, Caspase-8- and Caspase-9-dependent apoptosis in MCF-7 breast cancer cells⁴⁰. To determine whether the loss of cFLIP facilitates apoptotic cell death and thereby overcomes resistance to doxorubicin-induced apoptosis, the expression level of cFLIP was silenced using RNA interference. The protein level of cFLIP was depleted by sicFLIP compared to siNC (Fig. 4A). Downstream proteins of cFLIP-regulated apoptosis, including cleaved Caspase-8, and PARP, were markedly increased in siFLIP-treated cells compared to siNC in both cell lines (Fig. 4A). KH-3 induced more apoptosis markers, including cleavage PARP and Casp-8, in 231-DR compared to parental 231 cells.

Fig. 4. KH-3 overcame cFLIP-mediated doxorubicin resistance.

A Western blot analysis of cells transiently transfected with cFLIP siRNA or siNC (negative control). Cells treated with 10 µM KH-3 was served as the positive control. And cells were harvested post 48 h transfection or post 24 h KH-3 treatment. For protein quantification, target protein was first normalized to its corresponding GAPDH followed by a second normalization to the negative treatment control in 231 cells. Image J was utilized for quantification. B Western blot analysis of the protein levels of cFLIP and other proteins in cell lysates of 231-DR cells transiently transfected with control vector or vector containing cFLIP_L ORF. Cells were treated with the vehicle control or 10 µM KH-3 for 24 h post 24 h transfection. C Western blot analysis of cells transiently transfected with cFLIP siRNA or siNC. Cells were treated with the vehicle control or 1 µM doxorubicin for 48 h post 24 h transfection. GAPDH was used as the loading control.

Furthermore, we compared the response of both cell lines to cFLIP depletion. In 231-DR cells, cFLIP silencing significantly reduced cell viability (Fig. 1G) as well as markedly increased apoptotic cell death protein markers, including cleaved Caspase-8, and PARP, compared to parental cells (Fig. 4A). This indicates that 231-DR cells exhibited greater sensitivity to cFLIP depletion than doxorubicin-sensitive MDA-MB-231 cells. However, cFLIP overexpression failed to attenuate KH-3-triggered apoptotic cell death in 231-DR cells (Fig. 4B), possibly due to the concomitant loss of XIAP, Survivin, and BCL2 caused by KH-3. The data above suggest that cFLIP is a key driver of doxorubicin resistance in 231-DR cells, which can be targeted by the HuR inhibitor KH-3.

In 231-DR cells, 40 h and 48 h doxorubicin treatment at the same dose did not noticeably enhance PARP cleavage compared to the vehicle control, as observed in MDA-MB-231 cells (Supplementary Fig. 4 and Fig. 4C). Conversely, cFLIP silencing alone induced increased cleavage of PARP and Caspase-8 in 231-DR cells (Fig. 4C). Combining doxorubicin and cFLIP silencing slightly enhanced Caspase-8, and PARP cleavage (Fig. 4C), suggesting a potential synergy between the HuR inhibitor KH-3 and doxorubicin. Taken together, inhibition of HuR function by KH-3 effectively overcomes cFLIP-mediated doxorubicin resistance in 231-DR cells.

KH-3 showed synergy with doxorubicin

To test our hypothesis that HuR inhibitor KH-3 would sensitize breast cancer cells to chemotherapy, we performed the combined MTT assay using multiple human TNBC cell lines, including MDA-MB-231, 231-DR, and SUM159 (Fig. 5A). KH-3 induced apoptosis in a dose-dependent way in SUM159 cells, similar to that in MDA-MB-231 cells (Supplementary Fig. 5). The cell viability shifted from right to left with increased doses of KH-3, indicating that KH-3 enhanced the efficacy of doxorubicin. To further evaluate the combined effects, the synergy score was calculated and shown as the heatmap (Fig. 5B), where the combined treatment displayed synergistic effects. The combined clonogenic assay was performed to further assess the combined effects. The 30 nM or 600 nM doxorubicin decreased the plating efficiency of MDA-MB-231 and 231-DR cells, respectively. However, when combined with the sub-lethal dose of KH-3, the plating efficiency was further decreased in both cell lines compared to either single-agent treatment (Fig. 5C, and Supplementary Fig. 6 and 7). In addition, the fixed molar ratio (KH-3 to doxorubicin) combined MTT assay was performed, and the combination index (CI) was calculated. The CI values indicated the synergistic effects of the two drugs in both 231 and 231-DR cells (Fig. 5D, and Supplementary Fig. 8). Together, the results indicate that the combination of doxorubicin and KH-3 synergized in reducing cell proliferation and/or cell death.

Fig. 5. KH-3 showed synergy with doxorubicin.

A Cell viability curves of MDA-MB-231 cells, 231-DR cells, and SUM159 cells treated with doxorubicin and KH-3 together for 4 days, which were determined by the combined MTT-based cytotoxicity assay. B The synergy score plots of the combination of KH-3 and doxorubicin in (A). C Relative plating efficiency of MDA-MB-231 cells, and 231-DR cells treated with KH-3, doxorubicin, or the combination treatment. MDA-MB-231 cells and 231-DR cells were treated for 9 or 16 days, respectively. Representative colony imagines were shown in (C). Ordinary one-way ANOVA test, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. D Combination index (CI) at different effect levels (ED95, ED90, ED75, and ED50) for the combination of KH-3 and docetaxel in MDA-MB-231 cells, 231-DR cells. All results are shown as mean ± SD of three replicates.

KH-3 enhanced the efficacy of doxorubicin by inducing apoptosis

To further investigate the molecular mechanisms behind the chemo-sensitization activity of KH-3, we performed the Western blot assay to analyze the combination effects of KH-3 and doxorubicin. Our preliminary data showed that doxorubicin did not induce apoptosis markers (cleaved PARP and Caspase-8) until 24 h and obvious apoptosis was observed after 40 h treatment. KH-3 treatment reduced the protein level of cFLIP as early as 4 h treatment. The downstream Caspase-8 action was triggered after 16 h treatment and peaked after 24 h treatment (Supplementary Fig. 9). Therefore, for the combination treatment, we chose 24 h/40 h treatment time. Treatment with KH-3 alone or combined with doxorubicin for 24 h resulted in near complete depletion of cFLIP_L and evident loss of Survivin and XIAP in both MDA-MB-231 cells and 231-DR cells compared to that of the vehicle control (Fig. 6A). We also examined the proteins engaging in the downstream events of cFLIP_L/XIAP/Survivin-regulated apoptotic cell death. In MDA-MB-231 cells, the combination treatment for 40 h induced an obvious increase in apoptosis protein markers compared to treatment with KH-3 or doxorubicin alone, including the cleavage of PARP, Caspase-8, Caspase-9, and Caspase-3 (Fig. 6A, left). These data suggest that KH-3 enhances the efficacy of doxorubicin in TNBC cells.

Fig. 6. KH-3 enhanced the efficacy of doxorubicin by inducing apoptosis.

A Western blot analysis of whole-cell lysates of MDA-MB-231 cells or 231-DR cells. Before harvesting, cells were treated by DMSO (vehicle control), doxorubicin (0.5 µM or 1 µM), KH-3 (10 µM), and the combination treatment of doxorubicin plus KH-3 for 24 h or 40 h. GAPDH was used as the loading control. B The proposed working model of how KH-3 sensitizes cells to doxorubicin. The HuR functional inhibitor, such as KH-3, may sensitize TNBC cells to doxorubicin by blocking the function of HuR by binding to it. The interactions between HuR and its’ target mRNAs are interrupted, leading to the decreased expression of the targets, including Survivin, XIAP, and cFLIP_L.

Doxorubicin treatment led to a dose-dependent increase in cleavage of PARP, Caspase-8, and Caspase-3 in MDA-MB-231 cells post 40 h treatment (Fig. 6A, left panel), but no such effect was observed in 231-DR cells (Fig. 6A, right panel). However, treatment with KH-3 alone or combined with doxorubicin remained to induce the cleavage PARP, Caspase-8, and Caspse-3 in 231-DR cells post 24 h treatment. These results suggest that KH-3 can overcome doxorubicin resistance, reemphasizing the critical roles of HuR in medicating chemoresistance. As depicted in the proposed working model (Fig. 6B), KH-3 synergizes with doxorubicin by downregulating the protein levels of cFLIP_L, XIAP, and Survivin, leading to the activation of caspases and subsequent apoptotic cell death.

Discussion

Chemotherapy remains the standard treatment for triple-negative breast cancer (TNBC), yet patients frequently develop resistance to it¹. Doxorubicin, a widely used chemotherapy reagent, has been employed in the treatment of various cancer types, including TNBC. Our study aimed to investigate the molecular mechanism of doxorubicin resistance and identify strategies to reverse the resistance. The findings shed light on the mechanism of doxorubicin resistance in TNBC cells involving cFLIP overexpression. They elucidate the role of HuR in cFLIP-mediated chemoresistance and demonstrate the potential of combining a HuR inhibitor with doxorubicin to sensitize TNBC cells to the drug.

Our study identified cFLIP overexpression as a main driver of doxorubicin resistance in TNBC cells. cFLIP, a proteolytically inactive mimic of Caspase-8⁴¹, inhibits Caspase-8 activation^17,18, thereby impeding extrinsic apoptosis. The siRNA-mediated cFLIP knockdown triggers the ligand-independent apoptosis that is dependent on Caspase-8 and Caspase-9 in MCF-7 breast cancer cells⁴⁰. In prostate cancer, doxorubicin sensitizes cells to TRAIL-mediated apoptosis, and this sensitization is related to the loss of cFLIP^42,43. In this study, we found that cFLIP expression was noticeably increased in doxorubicin-resistant 231-DR cells. Overexpression of cFLIP_L attenuated the doxorubicin-induced Caspse-8 activation and PARP cleavage in doxorubicin-sensitive MDA-MB-231 cells. Conversely, siRNA-mediated cFLIP knockdown triggered apoptosis, with more pronounced apoptosis observed in 231-DR cells. These findings indicate that cFLIP overexpression is a key driver of doxorubicin resistance in breast cancer cells, highlighting the potential of inhibiting cFLIP expression as a strategy to overcome this resistance.

We demonstrate, for the first time, the positive regulation of cFLIP by HuR in breast cancer cells. Additionally, cytoplasmic HuR protein levels were upregulated in 231-DR cells compared to MDA-MB-231 cells. One previous study showed that HuR silencing led to a decrease in the stability of cFLIP mRNA in pancreatic cancer cells³⁴. However, the understanding of the relationship between HuR and cFLIP is currently limited and requires further characterization. Our findings reveal that HuR inhibition, either pharmacologically using KH-3 or via RNA interference, reduces cFLIP expression. Conversely, HuR overexpression elevates the protein level of cFLIP. This highlights HuR as a potential therapeutic target to overcome cFLIP-mediated chemoresistance.

Collectively, our results provide a strategy to overcome cFLIP-mediated doxorubicin resistance by inhibiting HuR. Decreasing the cytoplasmic HuR protein levels sensitizes breast cancer cell line MCF-7 to tamoxifen, while HuR overexpression confers tamoxifen resistance in MCF-7 cells²⁷. Interestingly, HuR downregulation is implicated in doxorubicin resistance in MCF-7, and HuR restoration re-sensitizes MCF-7 cells to doxorubicin⁴⁴. In the current study, unlike doxorubicin, HuR inhibitor KH-3 inhibited cell proliferation, induced cell death, and downregulated cFLIP in 231-DR cells. Additionally, KH-3 showed synergy with doxorubicin in both MDA-MB-231 and 231-DR cell lines. Moreover, KH-3 sensitized MDA-MB-231 cells to doxorubicin-induced apoptosis. This study provides additional insights into the roles of HuR in doxorubicin resistance in triple-negative breast cancer cells. We previously reported that KH-3 also restored the efficacy of docetaxel³³. Further studies may help identify HuR inhibition as a potential strategy to overcome chemotherapy resistance in other types of cancer cells.

In addition to cFLIP, treatment with KH-3 in MDA-MB-231 and 231-DR cells also caused the loss of XIAP and Survivin. XIAP was a reported HuR target in MDA-MB-231 cells in our previous study²⁶, and HuR stabilizes the mRNA and stimulates the protein translation of XIAP^45,46. XIAP suppresses the activity of Caspase-9, Caspase-7, and Caspase-3⁴⁷ and promotes the degradation of caspases via ubiquitinating them⁴⁷. Similarly, HuR binds to the 3’-UTR of Survivin mRNA and enhances the protein level of Survivin⁴⁸. Survivin binds to and suppresses Caspase-9, Caspase-7, and Caspase-3⁴⁹. These findings collectively suggest that KH-3 may also induce apoptosis by inhibiting XIAP and Survivin.

One limitation of this study is that only a single doxorubicin-resistant TNBC cell line, along with its parental cell line, was used to investigate cFLIP-mediated resistance, although some experiments included an additional TNBC cell line. While our findings are meaningful within these specific cellular environments, future studies should incorporate a broader range of breast cancer cell lines to validate the observed mechanisms and enhance the generalizability of our conclusions across different TNBC cell lines. Another significant limitation is the lack of in vivo validation, which could provide more physiologically relevant context and further support our hypothesis that HuR inhibition may overcome doxorubicin resistance. Additionally, while our time-course experiments offer valuable insights, they may not fully capture the temporal dynamics of HuR’s role in preventing apoptosis by regulating cFLIP and other targets. More detailed time-point analyses could further elucidate the sequence of molecular events that contribute to enhanced chemosensitivity in TNBC cells.

In conclusion, our study uncovers a novel mechanism of cFLIP-mediated doxorubicin resistance in TNBC and demonstrates that HuR inhibitor KH-3 sensitizes TNBC cells to doxorubicin-induced apoptosis by reducing the expression of cFLIP, as well as XIAP, and Survivin. This study supports a new therapeutic strategy to enhance the efficacy of doxorubicin in TNBC treatment by inhibiting HuR.

Methods

Cell culture and reagents

The Human TNBC cell line MDA-MB-231 was purchased from American Type Culture Collection (Manassas, VA) and maintained in DMEM medium (Corning, cat# 10-017-CV) supplemented with 10% (v/v) fetal bovine serum (FBS, Sigma Aldrich, cat# F4135), 1% (v/v) L-glutamine (Corning, cat# 25-005-CI), and 1% (v/v) penicillin-streptomycin (Corning, cat# 30-002-CI). The human TNBC cell line SUM159 was obtained from Asterand⁵⁰ and cultured in Ham’s F12 medium (Corning, cat# 10-080-CV) with 5% FBS, 10 μg/mL insulin (Sigma-Aldrich, cat# 10516), and 500 μg/mL hydrocortisone (Sigma-Aldrich). The HeLa Tet-Off pTre2 hyg-HuR Flag clone was generously provided by Dr. Dan A. Dixon from the University of Kansas. These cells were stably transfected with the Flag epitope-tagged HuR cDNA, which was inserted into the pTRE2hyg expression vector⁵¹. The absence of doxycycline induces the overexpression of HuR in the above HeLa Tet-Off cells. To induce doxorubicin resistance in MDA-MB-231 cells, doxorubicin (GensiaSicor Pharmaceuticals, cat# NDC 0703-5040-01) was added to the cell culture medium. The drug concentration was increased intermittently until a stable doxorubicin-resistant MDA-MB-231cell sub-line, named 231-DR, was established. The 231-DR cells were maintained in the presence of 150 nM doxorubicin, which was removed before the experiments. When harvesting cells for the followed analysis, both the attached cells and floating cells/cell debris were collected. The small molecule inhibitor KH-3 was synthesized as described before²⁶. All other reagents used in this study are listed in Supplementary Table 1. Cell lines were regularly monitored to ensure that they were free of mycoplasma contamination.

The RNA extraction, RT-qPCR, protein extraction, MTT-based cytotoxicity assay, colony formation assay, synergism analysis, Western blot analysis, immunoprecipitation (RNP-IP), and cell viability assay were performed as described in the previous publication³³. All the primers used in this study are listed in Supplementary Table 2. All the antibodies used in this study are listed in Supplementary Tables 3 and 4.

DNA and siRNA transfection

The establishment of the MDA-MB-231 cells stably transfected with HuR shRNA was described previously²⁶. HuR shRNA expression was induced by adding doxycycline (1 μg/mL) to the cell culture medium. For siRNA transfection, the siRNA smart pool targeting ELAVL1 (HuR) (Dharmacon Reagents, cat# L-003773-00-0005) or CFLAR (cFLIP) (Dharmacon Reagents, cat# L-003772-00-0005) was transfected into cells using the Lipofectamine 3000 Reagent (Invitrogen, cat# L3000-015) following the provided protocol with slight adjustments. Additionally, the pcDNA3.1-C-(k)DYK vector containing the human CFLAR cDNA (transcript variant 1, mRNA, NM_003879.7) ORF was procured from GenScript and transfected into cells using the Lipofectamine 3000 Reagent following the suggested protocol with minor modifications.

Statistical analysis

One-way or Two-way ANOVA test, as well as t test were performed to examine the significant difference using Prism software 8.0 (GraphPad). The specific statistical analysis for each dataset was indicated in figure legends. The P value < 0.05 was considered statistically significant. Unless otherwise stated, all in vitro experiments were performed at least three times.

Author contributions

X.W. and L.X.: conceptualization, supervision, funding acquisition, review, revision, and approval of the manuscript. L.W.: investigation, methodology, validation, original draft preparation, editing, and revision of the manuscript. S.K., J.A., and A.A.: investigation, methodology, review and editing of the manuscript. All authors have approved the final version of this manuscript.

Data availability

All data supporting the findings of this study are presented in the article and the Supplementary Materials. All other raw data are available upon reasonable request from the corresponding author.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41698-024-00780-x.