Dataset DOI: 10.5061/dryad.05qfttfhx

Description of the data and file structure

To characterize the local breast cancer tumor microenvironment (BC-TME), we performed high-dimensional single-cell proteomic profiling on 161 fresh tumor specimens obtained from 65 treatment-naïve patients with triple-negative breast cancer (TNBC). A 50-parameter antibody panel was designed to assess TME-associated proteins, including major lineage markers for distinguishing immune cell subsets (e.g., CD4, CD8, TCRγδ, CD69, CD19, CD14, CD15, CD16, CD203, CD206, CD56, CD11b, CD66b, CD123, CD11c, HLA-DR), immune regulatory molecules (e.g., PD-1, CTLA-4, CD25, CD127, TIM-3, CD204, CD47), epithelial markers (e.g., HER2, Mucin-1), stromal markers (e.g., FAP), and cancer stem cell (CSC) markers (TSPAN8, CD44, and CD24). Specimens were analyzed by spectral flow cytometry and stratified into CSC-high (CSC^High) and CSC-low (CSC^Low) groups based on median TSPAN8/CD44 co-expression levels within the CD45^-FAP^-CD24^- population. High-dimensional data were processed using nonlinear dimensionality reduction (t-SNE) and unsupervised PhenoGraph clustering to resolve the heterogeneity of tumor-infiltrating leukocytes in both individual and aggregated analyses (Figure 1B in Processed_data).

The raw data for Figure 1B consist of four files, which are described as follows:

1. Figure1B_raw_data-CD3_T_tsne.acs: Flow cytometry data for Figure 1B underwent nonlinear dimensionality reduction (t-SNE) analysis and unsupervised Phenograph clustering to analyze T cell subsets in CSC-High and CSC-Low tumors.

2. Figure1B_raw_data-concat_T8_high_T8_low_CD3_T_download_16000_cells.downloading: This flow cytometry dataset consists of T cells from a CSC-high tumor sample and a CSC-low tumor sample, which were downsampled to equal numbers and concatenated for subsequent t-SNE analysis.

3. Figure1B_raw_data-T8_high_HSY_1156553_TUMOR_raw_data_010.fcs: This is a representative raw flow cytometry data from a CSC-high tumor sample.

4. Figure1B_raw_data-T8_Low_GCH_1071022_TUMOR_raw_data_013.fcs: This is a representative raw flow cytometry data from a CSC-low tumor sample.

We hypothesized that the increased Treg infiltration in situ is attributable to a paracrine or systemic effect exerted by CSCs. To test this, an in vitro co-culture system using peripheral blood mononuclear cells (PBMCs) from healthy donors was established. Conditioned medium (CM) collected from fluorescence-activated cell sorting (FACS)-sorted TSPAN8+ human TNBC MDA-MB-231 or BC MCF-7 cells (CM-TS+) and the same cells with TSPAN8 depletion (CM-TS+(kd)) were supplemented into the culture system. As anticipated, co-culturing with CM-TS+ led to a significant increase in the frequency of CD4⁺FOXP3⁺ Tregs compared to co-culturing with CM-TS+(kd), as determined by FCM, without altering the frequency of naïve CD4⁺ T cells (Figure 2B in Processed_data).

The raw data for Figure 2B consist of seven files, which are described as follows:

1. Figure2B_raw_data-20250804_Figure_2B_tsne.acs: This is the data for Figure 2B obtained from flow cytometry, which underwent t-SNE analysis and unsupervised Phenograph clustering. The data were used to generate the Figure 2B image, which includes analysis of T cell subsets in both CM-TS+ and CM-TS+(kd) groups of PBMCs.

2-4. Figure2B_raw_data-Ts-KD1.fcs, Figure2B_raw_data-Ts-KD2.fcs, and Figure2B_raw_data-Ts-KD3.fcs: The raw flow cytometry data from three CM-TS+(kd) groups of PBMCs.

5-7. Figure2B_raw_data-Ts-NC1.fcs, Figure2B_raw_data-Ts-NC2.fcs, and Figure2B_raw_data-Ts-NC3.fcs: The raw flow cytometry data from three CM-TS+ groups of PBMCs.

Using an indirect co-culture system, we treated PBMCs with indicated EVs. Incubation with EVs-TS+ resulted in a marked increase in CD4⁺ T cells, particularly CD4⁺FOXP3⁺ Tregs, and a decrease in CD8⁺ T cells (especially the CD8⁺FOXP3⁻ subset) in PBMCs compared to all control conditions (Figure 2G in Processed_data). These data, consistent with prior observations in patient tumor specimens, indicate that EVs derived from TSPAN8⁺ CSCs mediate the enhanced abundance of Tregs.

The raw data for Figure 2G consist of four files, which are described as follows:

1. Figure2G_raw_data-17-Aug-2025.acs: This is the data for Figure 2G obtained from flow cytometry, which underwent t-SNE analysis and unsupervised Phenograph clustering. The data were used to generate the Figure 2G image, which includes analysis of T cell subsets in EV-depletion, EVs-TS+ and EVs-TS+(kd) groups of PBMCs.

2. Figure2G_raw_data-concat2_1.fcs: This flow cytometry data contains T cells from EV-depletion, EVs-TS+ and EVs-TS+(kd) groups of PBMCs that were downsampled to equal numbers and concatenated for subsequent t-SNE analysis.

The processed data of Figure 1B, 2B, and 2G are presented in Processed_data.pdf.

Code/software

The raw data for Figure 1B, 2B, and 2G can be viewed and analyzed by using FlowJo (version 10.5.3).

Access information

Other publicly accessible locations of the data:

• n/a

Data was derived from the following sources:

• n/a

Human subjects data

Approval for experiments involving human tissue samples was obtained from the Human Ethics Committee at Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China (Approval Letter No. 2019SQ220). Tumor and matched adjacent normal tissue specimens were identified and confirmed by pathologists at the same hospital.