The human triple-negative breast cancer cell line MDA-MB-231 was acquired from the American Type Culture Collection (ATCC, HTB-26) and cultured in RPMI-1640 medium supplemented with 1000 IU penicillin-streptomycin and 10% fetal bovine serum (FBS) in a humidified atmosphere at 37°C with 5% CO2. The human mammary gland-derived breast cancer cell line MCF-7 was acquired from Rio de Janeiro’s Cell Culture Bank and cultured in RPMI-1640 medium supplemented with 1000 IU penicillin-streptomycin and 10% FBS in a humidified atmosphere at 37°C with 5% CO₂. The human monocyte cell line THP-1 was acquired from ATCC (TIB-202) and cultured in RPMI-1640 medium supplemented with 1000 IU penicillin-streptomycin and 10% FBS in a humidified atmosphere at 37°C with 5% CO₂. Whole blood samples (10 mL/donor) were isolated from healthy female and male donors, without a defined age range, for experimental purposes only, following the principles proposed by the National Research Ethics Commission and Research Ethics Committee under the protocol CAAE 71085023.1.0000.5440. Human peripheral blood mononuclear cells (PBMCs) were isolated from whole blood samples by density gradient centrifugation using Histopaque^®-1077 (MERCK, USA). The buffy coat containing PBMCs was transferred to a different tube to isolate CD14+/CD16- monocytes by negative selection using magnetic beads from a commercial classical monocyte isolation kit (Myltenyi Biotec, USA). The purification steps and macrophage differentiation are described in Supplementary Table S1 . Purified CD14+/CD16- monocytes were plated in cell culture dishes containing RPMI-1640 medium supplemented with FBS and 10000 IU of penicillin-streptomycin. To induce the differentiation of monocytes into macrophages, 10 ng/mL of granulocyte-macrophage colony-stimulating factor (Peprotech, USA) was added to the monocyte culture for five days in a 37°C and 5% CO₂ incubator. Tamoxifen (TAM) was acquired from AdipoGen Life Sciences (USA) and doxorubicin (DOX) was acquired from Sigma-Aldrich (USA).

The method chosen to evaluate the half-maximal inhibitory concentration (IC50) of sensitive and resistant cells was the reduction of resazurin, following the protocol described by Kumar (13). Briefly, sensitive or resistant cells of MCF-7 and MDA-MB-231 were seeded in 96-well plates, incubated for 24 h, and treated with a wide range of concentrations of TAM (0.62 to 160 μM) and DOX (0.08 to 20 μM) for 72 h. After the treatment, cells were incubated with a resazurin solution (88 μM) for 4 h, and the plates were read with 570 and 600 nm filters. Cell viability was calculated using the absorbance reads on the indicated equation according to the resazurin manufacturer’s protocol, in which cell viability was expressed as a percentage of live cells in each treatment condition compared with the control. Cell viability was observed by standard staining with fluorescein diacetate (FDA) and propidium iodide (PI) to differentiate viable from dead cells. Staining solutions were prepared to a final concentration of 80 μg/mL of FDA and 200 μg/mL of PI. Cells in contact with the staining solutions were incubated for 30 min at 37°C and then washed to remove the staining excess. After washing, the cells were observed under a fluorescence microscope using the filters according to the manufacturer’s protocol. The captured images were overlaid by NIS-Elements BR to create merged images.

TAM- and DOX-resistant breast cancer cell lines were developed following practical guidelines published by McDermott et al. (14). Using the previously calculated IC50 of each drug in MCF-7 and MDA-MB-231 sensitive cells, a high-level laboratory model of stable drug-resistant cell lines was generated by continuous exposure of sensitive cells to stepwise dose progression, following a comparative selection method based on flask cell confluence and cell viability via trypan blue staining. After obtaining the IC50 of TAM- and DOX-resistant cell lines, the fold resistance was determined as the ratio between the IC50 of the resistant cell line and the IC50 of the sensitive cell line. Cell lines with a fold resistance equal to or higher than 2 were considered sufficiently resistant to proceed with the other experiments.

The differential ultracentrifugation method was used to isolate EVs, following the guidelines of Momen-Heravi (15). Briefly, the supernatants from the cell flasks or plate wells were centrifuged for 10 min at 1500 × g to remove dead cells, transferred to new tubes, and centrifuged for 30 min at 15,000 × g to remove cell debris and large particles. Finally, the supernatants were transferred to appropriate ultracentrifuge tubes and centrifuged at a high rotation speed of 100,000 × g for 70 min to isolate EVs. All centrifugations were performed at 4°C. The pellet of EVs was suspended in 100 μL of RNAse and DNAse-free ultra-pure water, following the guidelines suggested by the International Society of Extracellular Vesicles (ISEV) (16). Aliquots were prepared for characterization using transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA). The remaining suspended EVs were stored at -80°C. To observe the internalization of EVs upon contact with cancer cells, a lipophilic membrane stain was used to label the EVs following the manufacturer’s instructions. Previously isolated EVs were suspended in 200 μL of phosphate-buffered saline (PBS) with 1 μM of Vybrant Dil dye (Ex, 530 nm; Em, 580 nm, Thermo Fisher Scientific, USA) for 30 min at 37°C to obtain the labeling. The labeled EVs were centrifuged at 100,000 × g for 70 min and washed using another ultracentrifugation step in PBS. Labeled EVs were incubated with sensitive breast cancer cells for 24 h at 37°C and 5% CO2. Images were then captured in a bright field and under the appropriate fluorescence measurements in a Nikon Eclipse Ts2 with an attached Digital Sight 10 camera, followed by an overlay of the bright field and fluorescence images using ImageJ software to obtain merged images (17).

Messenger RNA (mRNA) was extracted from MCF-7 and MDA-MB-231 cell lysates using a spin column supplied with the RNeasy kit (QIAGEN, Germany). Both cell lines were co-cultured with TAM- and DOX-resistant EVs for 24 h. All extractions were performed in triplicate for each cell line following the manufacturer’s protocol by a single researcher to obtain high-quality mRNA. The mRNA capture and construction of 150 base pair double-stranded mRNA-Seq libraries were performed using Zymo-Seq RiboFree Total RNA (Zymo Research, USA). The protocol included the capture, elution, and fragmentation of mRNA, followed by synthesis of the first and second 150 strands. Illumina adapters were ligated to the ends of the mRNA-derived cDNA fragments, followed by library PCR amplification. The libraries were sequenced in the SP lane of an Illumina NovaSeq6000 sequencer.

A quality control test was performed before alignment using FastQC software (18). All low-quality reads and Illumina adapters were removed by the 0.39 version of the Trimmomatic tool (19). Reads that passed the quality filter were aligned to the reference genome index (Homo sapiens Ensembl GRCh38, release 109) using STAR 2.7.0a alignment software (20) at the University of California, Los Angeles UNIX-based cluster. The alignment rate and total number of mapped reads are shown in Supplementary Table S1 . Differential expression analysis was performed using the RStudio platform with 4.3.1 R (21). The DESeq2 package was used to detect differentially expressed genes (DEGs) by using negative binomial generalized linear models (22). The comparison between sensitive EVs vs. resistant EVs treatment for each cell line was estimated by the logarithmic fold-change log₂ (sensitive vs. resistant). The lists with the DEGs of each condition were filtered using a 5% false discovery rate (FDR) cutoff and an absolute fold change > 1.5 for upregulated and downregulated genes.

The initial gene ontology (GO) slim analysis was performed using the DEGs lists of each condition in the web-based tool WebGestAlt (23) with a 1% FDR and Benjamini-Hochberg multiple test adjustment. To investigate the signaling pathways associated with the DEGs, we performed an over-representation analysis in GSEA software 4.3.2, with a 1% FDR cutoff, using the Molecular Signatures Database (24, 25). The online tool g:Profiler was used to understand the relationship between the major enriched signaling pathways reported in the prior analysis (26). Functional interactions between potential protein expression among DEGs were performed using STRING networks with a protein-protein interaction FDR of 5% (27).

The detection of apoptotic cells was based on annexin V binding and PI uptake by flow cytometry using the Alexa Fluor^® 488 Annexin V/Dead Cell Apoptosis (Thermo Fisher Scientific, USA), following the manufacturer’s instructions. Briefly, MCF-7 and MDA-MB-231-sensitive cells were exposed to sensitive EVs as controls and to TAM- or DOX-resistant cells for 24 h. After exposure to EVs, the supernatant was removed from each well, and fresh media containing different concentrations of TAM or DOX, up to the previously calculated sensitive cell IC50, was incubated for another 72 h. The cells were harvested, stained with Annexin V-FITC and PI, and incubated for 15 min at room temperature. Annexin V-FITC/PI signals were assessed using a FACSCanto flow cytometer and analyzed using the FACSDiva software (BD Biosciences, USA).

Total RNA was isolated from sensitive cells after 24 h of exposure to TAM- or DOX-resistant EVs using a spin column supplied in the RNeasy kit (QIAGEN, Germany), and cDNA was generated from 1 μg of RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA), according to the manufacturer’s protocol. The conditions used in the qPCR reactions were performed using the Power SYBR™ Green PCR Master Mix (Applied Biosystems, USA) in an ABI 7500 FAST system (Applied Biosystems, USA) and following the manufacturer’s instructions. All samples were analyzed in triplicates. Relative expression of target genes was calculated by comparing threshold cycle (Ct) values of GAPDH as the housekeeping gene, according to the 2-ΔΔCT method (28). Acquired resistance (ABCB1, ABCG2, TGFBR1, TGFBR2, TGFBR3) and apoptosis-related genes (BCL2, CD44, PIK3CA, PTEN) were selected as target genes. The primers used are described in Supplementary Table S2 .

The clonogenic survival of sensitive cells previously exposed to TAM- or DOX-resistant EVs was evaluated using a colony-forming assay using an in vitro cell protocol published by Franken (29). Briefly, MCF-7 and MDA-MB-231 cells were cultured with sensitive (control) or resistant EVs for 24 h. Each condition involved treatment with different concentrations of TAM or DOX for 72 h. The cells were resuspended, counted, and reseeded in a low-density manner (250 cells/well) in six-well plates for 7 days. The colonies were fixed with 6% (v/v) glutaraldehyde and stained with 0.5% (w/v) crystal violet for 30 min. Colonies containing 50 cells were counted.

Interleukin (IL)-1β, IL-10, interferon-gamma (IFN-γ), and tumor necrosis factor (TNF)-α levels from macrophages supernatant after exposure of sensitive or resistant derived EVs for 24 h were measured using a commercial enzyme-linked immunosorbent assay (ELISA) kit following the manufacturer’s instructions (BD Biosciences, USA). The results were measured at an optical density of 450 nm using a SpectraMax 190 spectrophotometer (Molecular Devices). The results are reported as pg/mL of experimental and biological triplicates for each condition.

The first step was to calculate the IC50 values for sensitive cells exposed to a wide range of concentrations of TAM and DOX for 72 h. We estimated the IC50 of TAM and DOX in MCF-7 and MDA-MB-231 cells using a cell viability assay based on resazurin reduction. Nonlinear regression analysis showed that MCF-7 cells had lower TAM and DOX IC50s when compared to MDA-MB-231 cells ( Figures 1A–D ). The generation of drug-resistant breast cancer cell lines capable of overcoming and proliferating under exposure to higher TAM or DOX concentrations than the previously calculated IC50 of sensitive cells took nine months ( Figures 1E, F ). After the resistant cells reached this level of survival, we evaluated the IC50 values of these cell populations using the same resazurin protocol. We observed that the IC50 of MCF-7 exposed to TAM increased from 7.5 to 25.9 μM ( Figure 1H ), and the IC50 of MCF-7 exposed to DOX increased from 0.4 to 1.1 μM ( Figure 1I ). We observed an increase in TAM and DOX IC50 in resistant MDA-MB-231 cells, and in this case, the IC50 scaled from 9.5 to 23.4 μM in TAM-exposed cells ( Figure 1J ) and from 0.5 to 1.5 μM in DOX-exposed MDA-MB-231 cells ( Figure 1K ). TAM-resistant MCF-7 (TAM-R MCF-7) cells showed the highest fold resistance (3.45), followed by the sharp 3-fold resistance of DOX-resistant MDA-MB-231 cells (DOX-R MDA). DOX-resistant MCF-7 (DOX-R MCF-7) and TAM-resistant MDA-MB-231 (TAM-R MDA) cells displayed 2.75- and 2.46-fold resistance, respectively ( Figure 1G ). FDA/PI staining revealed that the resistant MCF-7 and MDA-MB-231 cell lines showed fewer PI-positive (dead) and more FDA-positive (viable) cells, even when exposed to high concentrations of TAM and DOX ( Figures 1L, M ). This dynamic was observed in 3D MCF-7 spheroids, in which sensitive cells were profoundly affected by drug treatment; however, resistant-derived spheroids showed more viable cells ( Figure 1N ).

The overall production of EVs was tracked using the supernatant from each cell line and the cells were plated for 24 h. The NTA results showed that sensitive and resistant cell populations produced approximately the same amount of EVs, averaging between 1.08 and 2.17 × 10¹⁰ EVs per mL ( Figure 2A ). Additionally, the average EV size in the parental and resistant cell lines ranged from 167 to 200 nm ( Figure 2B ). The TEM observation revealed the classic round “cup-shaped” morphology of parental and resistant breast cancer-derived EVs ( Figure 2D ). Furthermore, there were no differences in EV production between sensitive and resistant cells, and between cell lines. Each sensitive and resistant breast cancer cell line produced approximately 60 EVs per hour over 24 h ( Figure 2C ). Collectively, these results showed that there were no differences in the concentration, production, size, or shape between sensitive- and resistant-derived EVs.

After incubating sensitive breast cancer cells with fluorescently labeled EVs from parental and resistant cell lines, a fluorescent signal was observed inside the cells, indicating that these sensitive cells could take up TAM-R and DOX-r EVs ( Figures 3A, B ). Sensitive MCF-7 and MDA-MB-231 cells were exposed to TAM- and DOX-resistant EVs for 24 h, and subsequently treated with TAM or DOX for another 72 h. In this experiment, our results showed that the resistant-derived EVs preexposure increased the viability of sensitive cells in three different scenarios: the first and second are MCF-7 treated with 2.5 μM of TAM and 0.2 μM of DOX, and the last one is MDA-MB-231 with an increased number of viable cells after the resistant EVs preexposure and treatment with 2.5 μM of TAM ( Figure 3C ). Under our experimental conditions, high concentrations of TAM and DOX (near the original IC50 or above) showed no differences between sensitive cells exposed to sensitive or resistant EVs ( Figure 3D ). We generated MCF-7 spheroids and exposed them to sensitive or resistant EVs, which were then treated with TAM or DOX. In the images, more dead cells (red labeled) can be observed in the spheroid previously exposed to sensitive EVs and treated with 2.5 μM of TAM than the one previously exposed to TAM-R EVs. Moreover, the spheroid that received 0.2 μM of DOX and was previously exposed to sensitive EVs showed more dead cells than the one exposed to DOX-R EVs ( Figure 3E ).

To better understand the influence of drug-resistant EVs on the sensitive MCF-7 and MDA-MB-231 transcriptomes, we performed RNA sequencing (RNA-Seq) analysis in these two cell lines after exposure to TAM-R and DOX-R EVs for 24 h. After sequencing and mapping, 306 genes were considered DEGs of MCF-7 cells exposed to TAM-R EVs, of which 165 were upregulated and 141 were downregulated, and 574 genes were labeled as DEGs of MCF-7 cells exposed to DOX-EVs, separated into 437 upregulated and 137 downregulated genes, using the previously explained cutoff filtering ( Figure 4A ). In MDA-MB-231 cells exposed to TAM-R, we identified 160 DEGs, of which 91 were upregulated and 69 were downregulated. Additionally, 211 genes were identified as DEGs and were separated into 125 upregulated and 86 downregulated genes ( Figure 4A ). We analyzed the relationships to cellular components, and molecular functions using GO slim subsets to obtain a comprehensive overview of the DEG lists. This broad analysis is the first step in the transcriptome analysis pipeline, giving general information about which locations, mechanisms or cellular functions have more DEGs in breast cancer cells exposed to EVs derived from resistant cells for further investigation. This GO analysis revealed that the most enriched term for all conditions in which MCF-7 cells were exposed to resistant EVs was related to the membrane, followed by the “nucleus” term. Additionally, in the Molecular Function category, the most enriched term in MCF-7 cells exposed to resistant EVs was related to protein binding ( Figures 4B, C ). We observed the same pattern of enriched GO terms in the DEG list of MDA-MB-231 cells exposed to TAM-R- or DOX-R-derived EVs ( Figures 4D, E ). We generated a heatmap containing the top 10 most upregulated and downregulated genes under each resistant EV exposure condition ( Figure 4F ). Inside this narrow list, we observed the upregulation of DEGs related to solute transport (solute carrier family SLC), ABC transporters (ABCC12 and ABCB4), tumor growth transcription factors (FOSL1 and FOSB), and the downregulation of genes induced by interferon (IFI44L, IFI44, and IFIT1). In addition, the tumor suppressors STAT1 and human leukocyte antigen class I (HLA-A) were downregulated.

To further understand the biological function of the DEGs, we performed a more in-depth GO analysis using the GSEA over-representation parameter with a 1% FDR significance level, and functionally related gene sets were enriched. The upregulated DEGs in MCF-7 cells exposed to TAM-R and DOX-R EVs were related to transport activity in the plasma membrane region via ABC transporters, gated channel activity, and G protein-coupled receptors ( Figures 5A, B ). The downregulated DEGs of MCF-7 cells exposed to TAM-R enriched terms associated with innate immune response and MHC I activity, especially the response to type I interferon, while the downregulated DEGs of MCF-7 cells exposed to DOX-R enriched terms associated with immune response. However, this list enriched terms related to cell death processes, including apoptosis ( Figures 5A, B ). The DEGs from MDA-MB-231 cells exposed to TAM-R and DOX-R EVs followed the same pattern as in MCF-7 cells; the upregulated genes were enriched in terms of the transport of molecules via ABC transporters, G-protein receptors, and potassium ion transmembrane activity on the cell surface. In contrast, the downregulated DEGs of MDA-MB-231 cells exposed to both TAM-R and DOX-R EVs showed enrichment of terms related to cell death, apoptotic processes, and innate immune responses ( Figures 5C, D ).

ABC transporters are involved in acquired resistance in cancer cells. Using the DEGs lists in STRING to generate physical and biological interaction networks between the proteins encoded by these genes, we observed many upregulated ABC transporters in both cell lines after exposure to resistant EVs, including ABCB1 (P-glycoprotein); several ABC1 transporters (ABCA1, ABCA5, ABCA6, ABCA8, ABCA9, ABCA13, and ABCB5); and multidrug resistance-associated proteins (MRPs: ABCC5 and ABCC12), especially in cells exposed to TAM-R EVs ( Figures 5E, G ). Our results indicate the upregulation of genes is directly related to the transport of molecules across membranes. We observed SLC genes in all upregulated DEG, especially in cancer cells exposed to DOX-R EVs ( Figures 5F, H ). Additionally, the networks showed the upregulation of genes associated with voltage-gated channels in MCF-7 cells that received DOX-R EVs, and the upregulation of genes related to the G-protein coupled receptor in MDA-MB-231 cells that were exposed to DOX-R EVs. When looking at the downregulated networks, we can observe the reduced expression of interferon-induced proteins, such as IFIT1, IFIT2, and IFIT3, in MCF-7 cells exposed to TAM-R and DOX-R EVs. The MHC protein complex was enriched in all the downregulated lists, in which we observed the downregulation of several HLA genes, including HLA-A, HLA-B, HLA-C, HLA-E, and HLA-F.

According to our transcriptome results, resistant EVs influence several cellular mechanisms, particularly the transport of molecules and cell death-/survival-related functions. To further explore these findings, we analyzed the effects of resistant EVs on the expression of significant driver genes related to drug resistance and cell survival. The qPCR results showed that resistant EVs upregulate the expression of ABC transporters (ABCB1 and ABCG2) and transforming growth factor-β receptor genes (TGFBR1, TGFBR2, and TGFBR3) in both sensitive cell lines, except for MCF-7 cells exposed to resistant EVs, which showed no difference in TGFBR1 expression in comparison with sensitive EVs exposure ( Figures 6A, B ). We observed the upregulation of PIK3CA and downregulation of BCL2 in MCF-7 cells after exposure to resistant EVs, which is an interesting result that is directly related to apoptosis. The downregulation of genes associated with apoptosis suggests that resistant EVs may inhibit this cell death mechanism. To further explore this finding, we analyzed the effects of TAM- and DOX-R EVs on sensitive cells using a cellular assay for annexin V labeling. After 24 h of initial exposure to resistant EVs, both sensitive cell lines were treated with TAM or DOX for 72 h. The results showed that TAM-R and DOX-R EVs reduced the population of apoptotic MCF-7 cells ( Figure 6C ). In contrast, we observed neither a difference in PIK3CA and BCL2 expression nor in the apoptosis rate in MDA-MB-231 cells exposed to resistant EVs compared to the counterpart-sensitive EVs.

Taking the upregulation of transforming growth factor-β- (TGF-β) related genes as a reference, we investigate the cytostatic effects of resistant EVs to ascertain whether this upregulation of proliferation-related genes translates into increased cell growth and survivability. A clonogenic assay was performed, and the results showed that resistant EVs increased the survival rate and colony formation in the long term. We observed that MCF-7 TAM-R EVs increased the survivability and proliferation of sensitive cells after TAM treatment (2.5 μM) when compared to MCF-7 that received sensitive EVs and the same TAM treatment ( Figure 6D ); however, we did not find any difference in MCF-7 exposed to DOX-R EVs ( Figure 6E ). MDA-MB-231 cells previously exposed to TAM-R EVs showed a significant increase in colony formation after being treated with 2.5 μM of TAM compared to the cells that received sensitive EVs ( Figure 6F ). Additionally, MDA-MB-231 exposed to DOX-R EVs showed a higher survival fraction when treated with 0.2 μM of DOX compared to cells that received sensitive EVs ( Figure 6G ).

Given that genes directly related to the immune response were modulated by resistant EVs in both breast cancer cell lines, we performed ELISA to detect specific cytokines released by different human macrophages when exposed to TAM-R and DOX-R EVs from MCF-7 and MDA-MB-231 resistant cell lines. The human PBMC macrophages were exposed for 24 h to resistant EVs, and then we collected the supernatant and evaluated the levels of several cytokines. In this experiment, we observed increased levels of TNF-α and IL-1β after all resistant EVs exposure ( Figures 7A, B ); we observed an increase in IFN-γ levels after these macrophages were exposed to TAM-R EVs from the TAM-R MDA cell line ( Figure 7C ); however, we did not find any difference between the sensitive and resistant EVs in the IL-10 levels in all tested conditions ( Figure 7D ). Additionally, using the human monocyte cell line THP-1 transformed into macrophages by PMA, we further analyzed the impact of resistant EVs on macrophage cytokine release patterns. We observed results similar to those obtained for PBMC. We observed increased levels of TNF-α and IL-1β ( Figures 7E, F ) and no differences in IL-10 levels ( Figure 7G ) in all tested conditions. The difference lies in IFN-γ level measurement. We observed a significant increase in IFN-γ concentration in the supernatant from macrophages exposed to TAM-R and DOX-R EVs from MCF-R and TAM-R EVs from MDA-R cell lines ( Figure 7H ).