Galectin-3 secreted by triple-negative breast cancer cells regulates T cell function

Abstract

Triple-negative breast cancer (TNBC) is an aggressive subtype that accounts for 10-15 % of breast cancer. Current treatment of high-risk early-stage TNBC includes neoadjuvant chemo-immune therapy. However, the substantial variation in immune response prompts an urgent need for new immune-targeting agents. This requires a comprehensive understanding of TNBC's tumor microenvironment. We recently demonstrated that Galectin-3 (Gal-3) binding protein/Gal-3 complex secreted by TNBC cells induces immunosuppression, through inhibiting CD45 signaling in T cells. Here, we further investigated the interaction between secreted Gal-3 and T cells in TNBC. Using CRISPR/Cas9 gene editing of the TNBC MDA-MB-231 cell-line, we obtained Gal-3 negative^(neg) clones. We studied these in an in-vitro model, co-cultured with peripheral blood mononuclear cells (PBMC) to imitate immune-tumor interaction, and in an in-vivo model, when implanted in mice. Gal-3^neg tumors in mice had decelerated tumor growth after PBMC inoculation. In contrast, the Gal-3 positive^(pos) tumors continued growing despite PBMC inoculation, and tumor T regulatory cell (CD4/FoxP3+) infiltration increased. RNA sequencing of T cells from women with TNBC with elevated plasma levels of Gal-3 revealed significantly lower expression of oxidative phosphorylation genes than in T cells from healthy women. Similarly, in our in-vitro model, the decreased expression of oxidative phosphorylation genes and mitochondrial dysfunction resulted in a significant increase in CD8 intracellular reactive oxygen species. Consequently, T exhausted cells (CD8/PD1/Tim3/Lag3+) significantly increased in PBMC co-cultured with Gal-3^pos TNBCs. To conclude, we revealed a novel TNBC-related Gal-3 suppressor mechanism that involved upregulation of CD4 T regulatory and of CD8 T exhausted cells.

Keywords: Galectin-3; Oxidative phosphorylation; T exhausted cells; Triple negative breast cancer.

Graphical abstract

Figure 1

Gal-3 expression. A. Sequencing analysis of three edited clones (ICE sofware). B. Gal-3 mRNA expression studied by RT-PCR, showing less Gal-3 mRNA in the Gal-3^neg than the Gal-3^pos clones. C. SDS-Page gel showing total expression of the Gal-3 protein in three mutated clones (Gal-3^neg) compared to unedited MDA-MB-231 and two Gal-3^pos clones evaluated by Western blot. The bar graph compares the mean±SD of the Gal-3^neg and the Gal-3^pos clones. D. A representative histogram showing Gal-3 cell surface expression measured by flow cytometry analysis. In red, Gal-3^pos expression. The bar graph shows the mean±SD Gal-3 total expression of the Gal-3^neg clones and the Gal-3^pos clones. E. The mean±SD of the secreted Gal-3 protein by the mutated and the Gal-3^pos clones, evaluated by ELISA. F. Optical density (OD) of secreted Gal-3 in the plasma of seven healthy women and six women with TNBC, measured by ELISA.

Figure 2

The effect of Gal-3 on cell proliferation. A.Tumor cell proliferation evaluated by XTT assay. Mean±SD values are presented for Gal-3^neg vs Gal-3^pos (n = 10). B. A representative picture (light microscope, scale bar=100μ m, showing spheroid formation by Gal-3^neg and Gal-3^pos clones. On the right, a quantitative analysis of spheroid size, compares Gal-3^neg to Gal-3^pos clones. The bar graph shows mean±SD in μm² of 3D spheroids (n = 3). C. The bar graph shows peripheral blood mononuclear cell (PBMC) proliferation following culturing with supernatants of Gal-3^neg, Gal-3^neg + rGal-3, and WT clones. Bar graph, mean±SD, n = 5.

Figure 3

The effect of Gal-3 on tumor growth and infiltrated T reg cells in vivo. A. The scheme of the mice experiment protocol. B. Tumor volume is compared following the injection of Gal-3^neg vs. Gal-3^pos (mean±SE of 8-10 mice in each group, significance was calculated by the T-test). PBMCs of healthy women were intravenously inoculated one week after tumor injection. Below, four samples per group of gross tumor images. C. Confocal microscope immunofluorescent staining images of the expression of Gal-3 in Gal-3^pos and Gal-3^neg. In green, HLA-A was used to identify human tumor cells. In red, Gal-3. Dapi (blue) was used for nucleai staining. Scale bar=20μm. D. Confocal microscope immunofluorescent staining images of T-reg. Human HLA-A+ tumor cells are in red. CD3, used to identify T cells, are in purple. In green is FOXP3 staining. In yellow are T-reg and tumor cells that express FOXP3 and CD3. The scale bar=20μm. Arrows show T-reg cells in both groups. E. The bar graph shows mean±SE of CD3/FOXP3 positive cells in three slides of four mice injected with Gal-3^pos clones compared to Gal-3^neg clones (mean±SE, significance calculated by the T-test). F. IL-35 concentration (pg/ml) in the serum of six mice in each group (mean±SE, the significance was calculated by the T-test).

Figure 4

The effect of Gal-3 on T reg cells in vitro. A. The percentage of T-reg cells in co-cultures of PBMCs with Gal-3^neg ± r-Gal-3 or with Gal-3^pos clones after 3 days incubation. Mean±SE, n = 10, the significance was calculated by ANOVA. Below, a representative flow cytometry study shows CD4 gated cells, positive for CD25 and FOXP3. B and C. Assessments of IL-35 and IL-10 in supernatants of the same co-cultures. d. The percentage of activated CD4+/CD69+ cells obtained from co-cultures of PBMCs with Gal-3^neg ± r-Gal-3, or with Gal-3^pos clones. Mean±SE, n = 5. Below, a representative flow cytometry analysis showing a lower percentage of activated CD4+ cells after Gal-3 compared to the other cultures.

Figure 5

The effect of Gal-3 on oxidative phosphorylation genes. A. Gene Set Enrichment Analysis of transcriptome data of PBMCs of women with TNBC and healthy controls, in bulk RNA-Seq before deconvolution. B. T-cell transcriptomic gene expression after deconvolution. The oxidative phosphorylation hallmark MSigDB gene set is significantly enriched in the downregulated genes of the T-cell transcriptome, but not in the bulk transcriptome. C. Heatmap of the expression of five selected genes from PBMCs of four women with TNBC and four healthy female volunteers. D. Validation of gene expression in isolated T cells from PBMCs of six patients compared to six healthy donors. Mean±SD and significance were calculated by the T-test. E. Oxidative phosphorylation gene expression in T cells isolated from PBMCs co-cultured with Gal-3^neg compared to Gal-3^pos clones. Mean±SE, n = 6. The signifcance in each of the five genes is calculated by the T-test.

Figure 6

The effect of Gal-3 on intracellular reactive oxygen species (ROS) and T exhausted cells. A. Flow cytometry analysis of the percentage of intracellular ROS in gated CD8+ cells from co-cultures of PBMCs with Gal-3^neg ± r-Gal-3 or with Gal-3^pos clones. The mean±SE was calculated from seven experiments using PBMCs from different healthy women. The statistical significance was calculated by ANOVA. Below, a representative flow cytometry analysis of gated CD8+/ROS+ cells. B. Flow cytometry analysis of the percentage of CD8/PD1 gated cells positive for Tim-3 and Lag-3 (T exhausted cells) in the three groups after 5 days incubation. Mean±SE of 14 experiments. The statistical significance was calculated by ANOVA.