The study protocol was evaluated and approved by the Siriraj Institutional Review Board (approval nos. Si520/2010 and Si321/2016; Bangkok, Thailand). Written informed consent was obtained from all individual participants prior to enrollment in the study between January 2010 and December 2018. A fresh luminal BC tissue sample (0.5×0.5×1.0 cm³) isolated from BC tissues surgically resected from a 48-year old Thai female patient with T2N1aM0 (stage IIB) BC was incubated in 10X antibiotic mixture (1 U/ml penicillin G sodium and 1 mg/ml streptomycin; Thermo Fisher Scientific, Inc.) diluted in DMEM/F12. Following incubation, the cells were washed three times with PBS to remove red blood cells (RBC) and debris. The obtained tissue was chopped into 0.1×0.1×0.1 cm³ sections and plated in a culture dish. Cell trypsinization was performed using 0.25% trypsin and 0.9 mM EDTA (Thermo Fisher Scientific, Inc.) to activate cell proliferation. The cancer cell line obtained using the aforementioned protocol was termed BCA55-121, and karyotype analysis was performed by Giemsa staining (Division of Medical Genetics, Office for Research and Development, Faculty of Medicine Siriraj Hospital). Chromosomes were observed and counted under microscope. Growth curves of BCA55-121 were generated using a Live-Cell imager (IncuCyte^® Zoom; Sartorius AG) according to the manufacturer's instructions. HLA typing was performed at the Department of Transfusion Medicine, Faculty of Medicine Siriraj Hospital. BCA55-121 cells were cultured in DMEM/F12 (Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Inc.) and placed in a 5% CO₂ incubator at 37°C. The presence of protein markers including pan-cytokeratin (pan-CK), programmed death ligand 1 (PD-L1), fibroblast-activation protein (FAP) was confirmed in BCA55-121 cells by immunofluorescence following incubation with anti-panCK (1:100; cat. no. sc-8018; Santa Cruz Biotechnology, Inc.), anti-PD-L1 (1:100; cat. no. ab205921; Abcam), and anti-FAP (1:100; cat. no. ab53066; Abcam) for 2–3 h at room temperature. Cells were then incubated with appropriate secondary antibodies, including anti-mouse horseradish peroxidase (HRP) conjugate Cy3 (1:2,000; cat. no. 115-166-071) and anti-rabbit HRP conjugated FITC (1:400; cat. no. ab6717; Abcam) at room temperature for 1 h in the dark. Nucleus was stained using Hoechst 33342 (1:1,000; Invitrogen; Thermo Fisher Scientific, Inc.) for 1 h at room temperature. The proliferation of BCA55-121 cells was analyzed by Incucyte^® and analyzed in Incucyte^® Zoom System, and was compared with the proliferation of the commercial cell lines MCF-7 and MDA-MB-231.

BCA55-121 cells were incubated with Allophycocyanin (APC)-labeled anti-CD44 (1:5; cat. no. 21270446; ImmunoTools GmbH) and FITC-labeled anti-CD24 (1:5; cat. no. 21270443; ImmunoTools GmbH) antibodies for 30 min at 4°C. A CytoFLEX flow cytometer (Beckman Coulter, Inc.) was used for flow cytometric analysis using CytExpert software version 2.1 (Beckman Coulter, Inc.). For CSC isolation, 1×107 BCA55-121 cells labeled with the fluorescent antibodies were suspended in 2 ml 2% FBS/PBS, and CD44⁺CD24- cells were collected using a FACSAria II cell sorter (BD Biosciences). The purity of the obtained cells was assessed using the CytoFLEX flow cytometer.

A 0.5% base agar layer was prepared by mixing 1% agarose with 2X DMEM (or DMEM/F12) supplemented with 20% FBS in a 1:1 ratio. An over-layer of 0.35% agar cell suspension was prepared by mixing 0.7% agarose with 2X DMEM (or DMEM/F12) in a ratio of 1:1. BCA55-121 cells and their CSCs were suspended (4×104 cells/ml), in this mixture, which was incubated at 37°C in 5% CO₂ for 14–21 days. Subsequently, cells were fixed with 5% glutaraldehyde and stained with 0.5% crystal violet at room temperature for 15 and 30 min, respectively. Colonies with a diameter >500 µm were counted and surface areas of the formed colonies were measured using an inverted microscope IX71 (original magnification, ×100) at 1 h (baseline) and at the end of the experiment. The data obtained from CSCs are presented compared with the whole population of cultured BC cells.

Proteins (20 µg) extracted from BCA55-121 cells using extraction buffer [0.25 M Tris-HCl pH 6.8, 5% glycerol, 4% SDS, 10% β-mercaptoethanol, 0.001% (w/v) bromophenol blue] were separated using 12% polyacrylamide gels (Sigma-Aldrich; Merck KGaA). Proteins were then transferred to 0.45-µm nitrocellulose membranes. The membranes were blocked for 1 h with 5% BSA in TBST (25 mM Tris-HCl, pH 7.5; 125 mM NaCl; 0.05% Tween 20) at room temperature and incubated overnight at 4°C with anti-human SOX2 rabbit monoclonal antibody (1:1,000; cat. no. 3579; Cell Signaling Technology, Inc.), anti-human OCT4A rabbit monoclonal antibody (1:1,000; cat. no. 2840; Cell Signaling Technology, Inc.), anti-human homeobox protein Nanog (NANOG) rabbit monoclonal antibody (1:1,000; cat. no. 4903, Cell Signaling Technology, Inc.) or anti-human PD-L1 rabbit monoclonal antibody (1:500; cat. no. ab205921; Abcam). The membranes were then washed three times with TBST and incubated with a HRP-conjugated goat anti-rabbit IgG antibody (1:1,000; cat. no. ab6721; Abcam) for 2 h at room temperature. The proteins were visualized using SuperSignal West Pico Chemi-luminescent Substrate (Thermo Fisher Scientific, Inc.) under a Gel Documentation System (G:Box; version EF; Syngene). Anti-β-actin antibody (1:5,000; Santa Cruz Biotechnology, Inc.; cat. no. sc-47778) was incubated for 1 h at room temperature to detect β-actin as a loading control.

Venous blood (50 ml) from three healthy donors with aged between 25 and 30 years with no underlying diseases was collected in 0.1% heparin solution and peripheral blood mononuclear cells (PBMCS) were isolated using Lymphosep^® lymphocyte separation media (Biowest) according to the manufacturer's instructions. Briefly, heparinized blood was overlaid on Lymphosep^® media and centrifuged at 400 × g at 20°C for 30 min. PBMCs were collected and washed three times at room temperature using 200 × g centrifugations for 10 min with PBS or RBC lysis buffer (150 mM NH₄Cl, 10 mM NaHCO₃ and 1.26 mM EDTA), and then resuspended in AIM-V^® medium (Thermo Fisher Scientific, Inc.) and plated onto 6-well plates at 5×106 cells/well. After 2 h of incubation at 37°C and pH 6.8–7.3, non-adherenT cells were gently removed and suspended in 900 µl human type AB serum (Merck KGaA) with 100 µl DMSO and cryopreserved at −80°C as a source of lymphocytes. The adherent monocyte-enriched cells were cultured at 37°C in 5% CO₂ incubator in AIM-V^® medium containing granulocyte-macrophage colony stimulating factor (50 ng/ml) and interleukin (IL)-4 (25 ng/ml) (ImmunoTools GmbH) for 6 days to produce immature dendritic cells (iDCs). The iDCs were then matured to mature dendritic cells (mDCs) after incubation at 37°C in 5% CO₂ with tumor necrosis factor-α (50 ng/ml) and interferon (IFN)-γ (50 ng/ml) (ImmunoTools GmbH) in AIM-V^® medium for 2 days. Subsequently, wells containing ~5×105 mDCs were pulsed with 10 µg cancer cell-derived RNA. For phenotypic analysis of DCs, 5×105 DCs were detached with 5 mM EDTA, washed in 2% FBS/PBS and incubated with fluorescence-conjugated monoclonal antibodies (1:50) at 4°C for 30 min. Following washing in 2% FBS/PBS twice, cells were analyzed using a CytoFLEX flow cytometer. To determine the phenotype of monocytes and DCs, anti-CD11c APC (1:50; cat. no. 21487116; ImmunoTools GmbH), anti-CD14 APC (1:50; 21620146; ImmunoTools GmbH), anti-CD14 FITC (1:50; cat. no. 21620143; ImmunoTools GmbH), anti-CD40 FITC (1:50, cat. no. 21270403; ImmunoTools GmbH), anti-CD83 phycoerythrin (PE; 1:20; cat. no. 12-0839-42; Thermo Fisher Scientific, Inc.), anti-CD86 PE (1:20; cat. no. 12-0869-42; Thermo Fisher Scientific, Inc.) and anti-HLA-DR FITC (1:50; cat. no. 21278993; ImmunoTools GmbH) monoclonal antibodies were used. To determine human leukocyte antigen-A2 (HLA-A2) expression, PBMCs from healthy donors were collected using the method described above and analyzed using anti-HLA-A2 APC antibody (1:50; cat. no. 17-9876-42; ImmunoTools GmbH) and the CytoFLEX Flow cytometer.

Cryopreserved lymphocytes were thawed and resuspended in RPMI medium Thermo Fisher Scientific, Inc.) supplemented with 5% human AB serum (Merck KGaA). The lymphocytes were activated by co-culture with RNA-pulsed mDCs at a ratio of 10:1 for 1 day. Subsequently, lymphocytes were cultured at 37°C in RPMI medium supplemented with 5% human AB serum, IL-2 (20 ng/ml), IL-7 (10 ng/ml) and IL-15 (20 ng/ml) (ImmunoTools GmbH) for 9 days. The culture medium was replaced every other day.

Cytotoxic T cells were isolated using a CD8⁺ T cell Isolation kit (Miltenyi Biotec, Inc.) according to the manufacturer's protocol. In brief, 1×10⁷ lymphocytes were centrifuged at 300 × g for 10 min at room temperature and resuspended in 40 µl buffer containing PBS, 0.5% bovine serum albumin (Capricorn Scientific) and 2 mM EDTA. The cell suspension was mixed and incubated with 10 µl of a CD8⁺ T cell biotin-antibody cocktail (Miltenyi Biotec, Inc.) for 5 min, followed by 20 µl of a CD8⁺ T cell micro-bead cocktail (Miltenyi Biotec, Inc.) for 10 min at room temperature. The cell mixture was then magnetically separated using a magnetic-activated cell sorting (MACS) column and MACS separator (Miltenyi Biotec, Inc.), and the flow-through enriched in CD8⁺ T cells were collected. To determine the phenotype of effector immune cells, anti-CD3 FITC (cat. no. 21850033; ImmunoTools GmbH), anti-CD4 APC (cat. no. 21850046; ImmunoTools GmbH), anti-CD8 APC (cat. no. 21620086; ImmunoTools GmbH), anti-CD16 APC (cat. no. 21278166; ImmunoTools GmbH) and anti-CD56 PE (BioLegend, Inc.) monoclonal antibodies were used. Phenotype markers of lymphocytes, including CD3, CD4, CD8, CD16 and CD56 were analyzed using a FACSCalibur flow cytometer (BD Biosciences) using Flow Jo (Treestar) software version X.

Amounts of IFN-γ secreted by CSC RNA-pulsed DC-activated T cells in comparison with naïve T cells was measured using an IFN-γ ELISA kit (cat. no. DIF50, R&D Systems, Ltd.) according to the manufacturer's protocol. Cytokine concentration was determined by measuring optical density using a microplate reader at 450/570 nm.

BC target cells (T), either the whole culture population (BCA55-121-WP) or the enriched BCA55-121-CSCs, were plated into separate wells of 96-well plates (~3,000 cells/well) for 24 h for the assessment of immune cell killing activity. Subsequently, 50 µl Caspase-3/7 green reagent (Sartorius AG) at a 20 µM concentration was added to detect apoptosis. Total lymphocytes (E) from healthy donors with HLA matched to that of BCA55-121 cells (HLA-A2) were added as effectors and co-cultured with the tumor target cells (T) at effector to target (E:T) ratios of 5:1, 10:1 and 20:1 for 24 h. Cancer cells with apoptotic activity were detected using an IncuCyte^® live-cell analysis system.

All data are presented as the mean ± standard deviation (SD). The data were tested for normal distribution by the Shapiro-Wilk test. The data from two groups were analyzed by paired Student's t-test and from multiple groups by one-way repeated-measure analysis of variance (ANOVA) followed by Tukey's post-hoc test using GraphPad Prism software version 7.04 (GraphPad Software, Inc.) or SigmaPlot 16.0 (Systat Software, Inc.). P<0.05 was considered to indicate a statistically significant difference.

The tumor size measured 4.5×4×2 cm³, and was diagnosed as an invasive ductal carcinoma (moderately-differentiated) positive for estrogen receptor (>75%) and progesterone receptor (<10%) and negative for HER2/neu. Angiolymphatic invasion was absent, however, one lymph node was positive for macro-metastatic cancer cells. The cancer cells were visualized as cobblestone shaped under phase-contrast microscopy (data not shown). The primary cell line was positive for pan-cytokeratin, confirming that the cells were of epithelial origin (18) and without stromal fibroblast contamination in the culture (Fig. 1A). The cell line also tested negative for fibroblast activation protein (FAP), and PD-L1 was expressed at a very low level (Fig. 1A). Chromosomal analysis revealed an aneuploid aberration of chromosome copies. BCA55-121 cells exhibited a higher proliferation rate compared with the MCF-7 cell line, but lower than that of MDA-MB-231 cells (Fig. 1A). Breast CSCs, defined as CD44⁺/CD24⁻ cells, were separated from the BCA55-121-WP cells by fluorescence-activated cell sorting. These CSC markers were expressed in ~75.2±0.45% of the BCA55-121-WP population (Fig. 1B). CSCs were subsequently sorted to enrich the cells with this phenotype to 92.9±5.83% purity (Fig. 1B). The tumorigenic ability of the BCA55-121-CSCs were tested, and the results were compared with BCA55-121-WP cells. Soft agar colony formation assay, a 3D spheroid cell culture assay that measures cell proliferation in semi-solid matrices (11) was conducted. The assay selectively enriches cells that exhibit a malignant transformation, anchorage-independent proliferation and progenitor cell-like tumorigenic properties. Significantly greater proliferative and tumorigenic abilities (2–3 folds) of the BCA55-121-CSC population were indicated by colony count and colony size compared with unsorted BCA55-121-WP cells (Fig. 1C). The level of OCT4 was increased in BCA55-121-CSCs compared with the unsorted cells (Fig. 1D). NANOG expression did not change in BCA55-121-CSC compared with BCA55-121 WP, and the level of SOX2 was very low. These results confirmed the greater tumorigenic and stemness properties of the CSC-enriched population. In addition, the expression of PD-L1 was significantly increased in the CSC population of the CSC culture compared with the WP culture (Fig. 1E).

DC phenotype was analyzed on days 1 (monocytes), 6 (iDCs) and 9 (mDCs). Monocytes are the only cell type that expresses CD14 (19), and this was diminished during subsequent maturation and replaced by a DC marker (CD11c), a DC maturation marker (CD83), an antigen-presenting molecule (HLA-DR) and co-stimulatory molecules (CD40 and CD86; Fig. 2A) demonstrating DC maturation in vitro. The subsets of cells in the activated lymphocyte population following co-culture with DCs were analyzed by flow cytometry. The highest proportion of cytotoxic T lymphocytes in the total lymphocytes was observed after co-culture with RNA-pulsed DCs followed by CD4+ T cells (Fig. 2B). NK cell and NKT cell populations were also identified among the activated lymphocytes (Fig. 2B). In addition, a significantly increased number of IFN-γ-positive CD8+ T cells were observed on day 18 of the lymphocyte activation protocol (day 9 after co-culture with RNA-pulsed DC; Fig. 2C). For each of the three healthy donors, higher levels of secreted IFN-γ were present in the supernatants of activated T cells compared with the supernatants of naïve T cells (Fig. 2D).

Total RNA from BCA55-121-CSC was loaded onto DCs, and the efficacy in activating effector lymphocytes to kill cancer cells was measured using Caspase-3/7 reagents for apoptosis with the Live-cell imager. To prevent the killing of cancer cells by other immune cells, not effector CD8⁺ T cells, HLA-A2 matched lymphocytes and BCA55-121 cells were used. The efficacy of tumor cell killing between lymphocytes activated by BCA55-121-WP and CSC RNA-pulsed DCs was compared (Fig. 3). The killing activity in the CSC RNA-activated lymphocytes was greater compared with that of BCA55-121-WP or CSCs, and this was at either 10:1 or 20:1 E:T tumor cells with significance reached at 20:1 E:T ratio (Fig. 3). Green fluorescence, representing cells with activated Caspase-3/7 at different E:T ratios, is displayed in Fig. 4A. Although increasing apoptotic activity over time was observed in both experiments, a 10:1 ratio of total activated lymphocytes did not result in enhanced killing of CSC targets, although BCA55-121-WP cells were killed (Fig. 4B and C). Additionally, at a 20:1 ratio, more BCA55-121-WP cells were killed compared with CSCs at all time points (Fig. 4B and C). These results demonstrated that apoptotic resistance may be a property of CSCs. In addition, CD8⁺ effector T cells isolated from the CSC RNA-activated lymphocyte population were less effective compared with the whole lymphocyte population against either BCA55-121-WP or BCA55-121-CSC targets (Fig. 4D and E). Limited killing activity was observed by CD8⁺ T cells that were activated by DCs without antigen exposure, which suggested that cancer cell killing was antigen specific.