Within the past few decades, a standard way to evaluate the effectiveness of NAC has been to identify any tumor residues in the surgical specimens, however it remains undefined before surgery. About 10% of breast tumor is not responding to NAC and resistant breast tumor leads to disease progression and poor prognosis, especially in TNBC. Accordingly, it is necessary to predict early which patients will not respond to standard treatment for reducing disease progression (17). Circulating EVs in the blood, intracellular communication vesicles with a lipid bilayer that ranges in size from 50 to 300 nm, are recognized to be essential for tumorigenesis, development, progression, and metastasis (18). Multiple technologies have been developed for tumor-derived EV detection and analysis (e.g. immunoaffinity-based capture) and have greatly advanced our understanding of tumor characteristics through liquid biopsy, despite the absence of tumor tissues. EVs can shuttle bioactive molecules, such as proteins and a wide variety of genetic materials, from one cell to another, resulting in a molecular transformation in the recipient cells (19–22). There may be an interplay or synergy between tumor cells in the acquisition of drug resistance via EV exchange. Therefore, we suggest a way to focus on the role of EVs in multidrug resistance for early therapeutic response prediction and therapy monitoring.
Our research revealed the dynamics of epigenetic changes that lead to drug resistance after chemotherapy in cell line models. There are various molecular mechanisms involved in chemoresistance; among these, we found significant changes in cancer stemness in drug-resistant TNBC models. For example, dysregulated TGF-β and Wnt signaling pathways may affect the overall progression to malignancy (Figure S5). They are also known to play a positive role in promoting drug-resistant properties in the cancer stem cell (CSC) population (23, 24). Interestingly, the cytotoxicity-induced morphological changes can be related to drug resistance. It has been reported that cytoskeletal reconstruction induces biological changes in cancer cells with drug resistance (25). After long-term treatment with chemotherapeutic drugs, actin stress fibers changed with a distinct feature showing migratory dynamics of cell spreading (26). All these phenotypic features can be predominantly attributed to their tumorigenic potential and multidrug resistance. Moreover, we focused mainly on the overexpression of ABC transporters, which results in drug resistance, a representative feature of CSC. Accumulated evidence from numerous studies, as well as our own, indicates that high expression of transmembrane proteins of this superfamily, such as MDR1, MRP1, and BCRP, is found in breast cancer, particularly in breast cancer stem cells (27).
Despite the clear relevance of ABC transporters playing a critical role in the development of multidrug resistance, clinical approaches to assess these proteins in the development of drug resistance have not been dramatically successful and are yet to be elucidated (28, 29). However, with the availability of the latest technology for EV analysis, we recommend re-evaluating the role of ABC transporters within EVs, not focusing on those in tumor tissues. Two notable singularities of our study are worth discussing. First, we discovered that a certain percentage of cells were stably resistant to NAC as chemotherapy drug had been administered continuously (Fig. 4b). A superficial explanation for this is that drug resistance may be induced by more than simply genetic alterations; the dysregulation of major epigenetic factors in breast cancer can play a more important role. This hypothesis is consistent with several reports on the epigenetic control of CSC and their influence on tumorigenesis, development, and responsiveness to therapy (30, 31). Second, using the strategy of enrichment of tumor-derived EVs, a higher expression of drug efflux transporters was observed in drug-resistant EVs than in wild-type EVs (Fig. 5d) (32–34). This discrepancy has been hinted at previously and may have clinical significance, as it potentially suggests that resistant clones may induce bursts of ABC transport in EVs to enhance drug resistance.
Although there have been a few studies that have utilized miRNAs, EV concentrations, and cancer antigens to monitor and predict response to treatments (35–37), our study is a novel attempt to use tumor-derived EVs to assess ABC transporters isolated from the plasma of patients with breast cancer treated with NAC to our knowledge. In clinical validation, considerable differences in expression of MDR1, MRP1, and BCRP between patients with a pCR and those with residual tumor were detected. Combination of all three parameters offered acceptable sensitivity, specificity, and accuracy in predicting the effectiveness of NAC through cumulative ROC analysis. Apart from the lack of understanding of the role of EVs in cancer stem cells and drug resistance, we suggest this method as a better strategy to repeatedly screen wandering molecular information to predict therapeutic responses. However, two minor limitations of this study merit discussion. There is a shortage of sufficient sample sizes to conduct significant cohort studies with high statistical validity. Another issue is the lack of a longitudinal study to analyze changes of the EV markers over time in the patient. Future studies will advance to larger patient groups and may address how EVs crosstalk with other tumor cells with different phenotypes to obtain drug-resistant characteristics besides MDR1, MRP1, and BCRP.