Approximately 40% of patients with ER-positive metastatic breast cancer acquire resistance to endocrine therapy due to the acquisition of LBD-ER mutations34. Importantly, this group of patients often have a more aggressive disease and worse prognosis35. Chemoresistance continues to be a major obstacle that impairs the efficacy of cancer therapy and nearly 90% of patients fail to respond to chemotherapy due to resistance of cancer cells, either de novo or acquired resistance36. As most patients with endocrine resistance are treated with chemotherapy, we hypothesized that reduced response to chemotherapy may contribute to the worse prognosis seen in patients harboring these mutations. Yet, no studies have assessed chemotherapy resistance in ER-positive patients harboring LBD-ER mutations in in vitro studies or in in vivo models. In our current study, we examined whether LBD-ER mutations confer resistance to chemotherapeutic drugs, specifically to doxorubicin and paclitaxel, which are among the most commonly used chemotherapy drugs for metastatic breast cancer. We discovered that LBD-ER cells are more resistant to chemotherapy compared to the WT-ER cells as using both in vitro and in vivo models.
In order to study chemoresistance, we employed the well-established MCF-7 cell model that expresses WT, D538G or Y537S –ER, at physiological levels12. Using these cells, we explored the effect of the mutations on viability, colony formation, and apoptosis following treatments with the chemotherapy drugs described above. Our results show that LBD-ER cells are more resistant to chemotherapy compared to the WT-ER cells. Of note, to further investigate chemoresistance, we established an in vivo model, and observed that paclitaxel treatment had no effect on tumor growth in mut-ER group, while this treatment led to tumor shrinkage in the WT-ER group (Fig. 3b). It is important to point that while the in vivo experiment showed a clear resistance of D538G-ER tumors compared to WT-ER tumors, some factors may have contributed to this behavior. First, we employed subcutaneous implantation approach; thus, drug response may be modulated by different microenvironmental stimuli. Second, we used paclitaxel doses suitable for mice tolerance, and perhaps WT-ER cells would respond to higher doses. Yet, the accumulating data we present strongly suggest a relative resistance to chemotherapy in mut-ER cells. Patients with endocrine-resistant MBC often receive CDK and PI3K inhibitors as part of their treatment regimen37. Interestingly, the impact of ESR1 mutations on the response to these inhibitors is still debatable, with some studies indicating that these mutations do not predict treatment response, while others have demonstrated a reduced response associated with ESR1 mutations38. These findings emphasize the necessity for developing novel drugs that can enhance treatment response and potentially prolong survival in patients with ESR1 mutations.
There are several mechanisms that enable BC cells to acquire drug resistance. Among these are the well characterized P-gp transporters, which are encoded by the MDR1 gene39. Doxorubicin is one of the major substrates of P-gp, therefore it plays a role in the internalization of doxorubicin40. Similarly, low doxorubicin accumulation has been reported in MDR cell lines including MCF-7ADR41. Indeed, we have shown that doxorubicin accumulation was lower in the D538G mutated cells compared to WT-ER (Fig. 8b). We also evaluated this gene expression on the mRNA and protein levels, and found that MDR1 protein expression is elevated in LBD-ER mutant cells, though more prominently in D538G, compared to the WT-ER (Fig. 4a-c). In agreement with these results, we found an increase in the P-gp expression level in patients' samples harboring the ESR1 mutations compared to patients without the mutation (Fig. 4f,g). Our results suggest that the mut-ER cells are able to utilize the P-gp transporters as part of their resistance mechanism to overcome chemotherapy treatments. A relationship between endocrine treatment and P-gp was demonstrated previously, showing that tamoxifen increases sensitivity to doxorubicin, and increases accumulation of vinblastine in MCF7 cells expressing MDR142. On the other hand, another study showed that WW domain-binding protein 2 (WBP2) can bind ER, leading increased MDR1 expression43.
We aimed to reveal the mechanism through which mut-ER induces MDR1 expression. Although MDR1 expression was observed primarily in mut-ER cells, we first confirmed that there are no ER binding sites on the MDR1 promoter. In search for a possible mechanism, we focused on the JNK/c-Jun pathway as it has been shown to play a role in chemoresistance of some types of cancer44–46. Thus, activation of JNKs that resulted in transcriptional activity of c-Jun, also led to the transcription of other genes related to drug resistance, like XIAP, c-fos and JunB47,48. Therefore, our goal was to reveal whether the JNK pathway was differentially regulated in LBD-ER cells. We observed activation of the JNK/c-Jun pathway as evidenced by high expressions of p-JNK, p-c-Jun and elevated AP-1 transcriptional activity in the LBD-ER mutant cells compared to WT-ER (Fig. 5). These data suggest a role of the JNK pathway in the LBD-ER mutations. Our results are supported by several studies. Increased JNK signaling was observed following long term acquisition of resistance to endocrine therapies49. Another study showed that activation of the JNK pathway is associated with everolimus resistance in endocrine resistant cells50. Noteworthy, the important role AP-1 plays in endocrine resistance was demonstrated in a study where inhibition of AP-1, in both in vitro and in vivo settings, enhanced anti proliferative effect of endocrine treatments and delayed the onset of tamoxifen resistance in mice model51.
The JNK/c-Jun pathway is known to be associated with MDR1 expression. Several studies have shown that JNK/c-Jun pathway is activated in cancer cells that acquired resistance to chemotherapy in vitro, demonstrating the importance in regulating MDR1 expression21,23,28. To test this, we first showed that SP600125, an inhibitor of the JNK signaling pathway, downregulated the expression of P-gp and MDR1, especially in the D538G cells. To provide additional evidence, we performed a ChIP assay and indeed our results showed an increase in c-Jun occupancy in the MDR1 promoter, leading to an increase in P-gp transporters in the mutated cells. In addition, a bioinformatics analysis of TCGA breast cancer BRCA database showed a strong positive correlation between JUN and ABCB1, confirming our results that c-Jun potentiates the MDR1 expression in BC cells.
The JNK/c-Jun pathway emerged as a differentially activated pathway in mut-ER cells that directly regulates MDR1 expression. Hence, our next aim was to assess whether this pathway may serve as a novel target for mut-ER breast cancer tumors. We studied viability and apoptosis of WT and mut –ER cells treated with JNK inhibitors and found that JNK inhibition decreased viability in both WT and D538G cells but significantly increased apoptosis in the D538G mutated cells. Previously, it was shown that JNK inhibition decreased cell growth in nasopharyngeal carcinoma52 and triple- negative breast cancer53. As we found that c-Jun directly induces MDR1 expression, we aimed to reveal whether inhibition of c-Jun activity would restore cell sensitivity to chemotherapy. Indeed, using a JNK inhibitor (SP600125) we were able to sensitize cells to doxorubicin. This effect was observed in all cell lines, with highest effect in D538G-ER cells. Indeed, it was shown that SP600125 reduced viability of breast cancer cells which acquired resistance to doxorubicin 54. In addition, a different study, in nasopharyngeal carcinoma cancer cells, showed that doxorubicin treatment combined with SP600125 more effectively inhibited cell growth than the single treatments52. Indeed, we showed that JNK inhibition led to increase the accumulation of doxorubicin in the D538G-ER cells. These results imply that doxorubicin resistance can be accounted for by decreased amounts of drug at nuclear targets, which may be mediated by the expression and function of P-gp.
Interestingly, while Y537S and D538G exhibit similar behavior, a detailed analysis suggests some subtle, yet potentially important differences. Thus, we observed less JNK phosphorylation in Y537S cells compared to D538G cells (as verified by analyzing different stable clones, supplementary Fig. 3). Additionally, we noted higher MDR1 expression in D538G cells and lower binding of c-Jun to MDR1 promoter in Y537S cells (Fig. 6c). These results suggest unique characteristics for each mutation despite their similar role in conferring endocrine resistance. Consistent with our observations, previous studies have also revealed notable differences between Y537S and D538G. For example, Y537S confers higher endocrine resistance than D538G cells and is associated with a shorter overall survival55.
While we focused on the JNK-MDR1 axis in mut-ER cells, studies showed that resistance to chemotherapy in breast cancer could evolve from a wide array of mechanisms56, some of them are relevant to mut-ER cells. For example, active PI3K/mTOR pathway may lead to resistance to chemotherapy57, and several studies including ours12,58 showed increased activation of this pathway in mut-ER cells.
Taken together, our current study demonstrates that activating mutations in the LBD-ER not only confer resistance to endocrine therapy but also relative resistance to commonly used chemotherapy. This resistance is, at least in part, mediated by the activation of the JNK/c-Jun pathway, leading to the upregulation of MDR1 expression (Fig. 8e). These findings underscore the urgency of discovering new drugs to effectively treat patients with ESR1-mutated tumors, as these tumors display decreased responsiveness even to subsequent lines of treatment such as chemotherapy. Moreover, the study highlights the significance of targeting the JNK/c-Jun as a strategy to sensitize mut-ER cells to chemotherapy. Furthermore, this study reveals distinct differences between the Y537S and D538G mutations, emphasizing the importance of carefully examining these discrepancies, as they may hold implications for the treatment of patients harboring these specific mutations.