Interleukin-3-Receptor-α in Triple-Negative Breast Cancer (TNBC): An Additional Novel Biomarker of TNBC Aggressiveness and a Therapeutic Target

Abstract

Tumour molecular annotation is mandatory for biomarker discovery and personalised approaches, particularly in triple-negative breast cancer (TNBC) lacking effective treatment options. In this study, the interleukin-3 receptor α (IL-3Rα) was investigated as a prognostic biomarker and therapeutic target in TNBC. IL-3Rα expression and patients' clinical and pathological features were retrospectively analysed in 421 TNBC patients. IL-3Rα was expressed in 69% human TNBC samples, and its expression was associated with nodal metastases (p = 0.026) and poor overall survival (hazard ratio = 1.50; 95% CI = 1.01-2.2; p = 0.04). The bioinformatics analysis on the Breast Invasive Carcinoma dataset of The Cancer Genome Atlas (TCGA) proved that IL-3Rα was highly expressed in TNBC compared with luminal breast cancers (p = 0.017, padj = 0.026). Functional studies demonstrated that IL-3Rα activation induced epithelial-to-endothelial and epithelial-to-mesenchymal transition, promoted large blood lacunae and lung metastasis formation, and increased programmed-cell death ligand-1 (PD-L1) in primary tumours and metastases. Based on the TCGA data, IL-3Rα, PD-L1, and EMT coding genes were proposed to discriminate against TNBC aggressiveness (AUC = 0.86 95% CI = 0.82-0.89). Overall, this study identified IL-3Rα as an additional novel biomarker of TNBC aggressiveness and provided the rationale to further investigate its relevance as a therapeutic target.

Keywords: interleukin-3/interleukin-3 receptor α; programmed cell death-ligand 1; triple-negative breast cancer; vascular mimicry.

Figure 1

Human TNBC cells express IL-3Rα. Representative tissue microarrays from human TNBC samples were stained with anti-IL-3Rα antibodies. IL-3Rα-positive and -negative TNBC samples are shown. Original magnification 10× and 20×, scale bar: 100 µm and 50 µm, respectively. Panel (A) (red square) refers to Turin samples and panel (B) (black square) to Sassari samples.

Figure 2

Kaplan–Meier of overall survival. Using the Cox regression test, we obtained a p = 0.04, while using the Wilcoxon–Breslow–Gehan test, the p corresponded to p = 0.0245.

Figure 3

IL-3Rα expression in different TNBC cell lines: (A) representative Western blot of IL-3Rα expression in the indicated TNBC cell lines. Actin served as housekeeping gene. The uncropped blots are shown in Figure S2; (B) FACS dot plots of IL-3Rα expression (purple dots) and appropriate IgG controls (left panel in each cell line) of the indicated TNBC cell lines. M07 and MCF10A cells served as positive and negative controls, respectively.

Figure 4

IL-3Rα activation impacts EMT: (A–D) representative Western blot and quantification of Vimentin and N-Cadherin expression levels in MDA-MB-231, Hs-578T, MDA-MB-436, and HCC-1395 cells untreated (Ctrl, circles on diagrams) or treated with IL-3 (1 ng/mL, squares on diagrams or 5 ng/mL, triangles on diagrams). Data are expressed as the mean ± SEM normalised to housekeeping genes (Vinculin, Actin, and GAPDH). (E–H) SLUG, TWIST1, and ZEB 1 mRNA expression levels in MDA-MB-231, Hs-578T, MDA-MB-436, and HCC-1395 cells untreated (Ctrl) or treated with IL-3 (1 ng/mL or 5 ng/mL). Data are expressed as the mean ± SEM normalised to housekeeping genes. The uncropped blots are shown in Figure S3.

Figure 5

IL-3Rα activation reprograms TNBC cells towards an endothelial-like phenotype: (A) phase images of cells plated on growth factor reduced Matrigel in serum-free medium or supplemented with 1 ng/mL or 5 ng/mL of IL-3; (B) quantification of the vessel number was performed over at least three independent replicates for each cell line. MDA-MB-231, Hs-578T, MDA-MB-436, and HCC-1395 cells untreated (Ctrl, circles on diagrams) or treated with IL-3 (1 ng/mL, squares on diagrams or 5 ng/mL, triangles on diagrams) are shown. The results are expressed as mean ± SEM. Original magnification 20×, scale bar: 80 µm. Comparisons were performed using one-way ANOVA followed by Tukey’s multiple-comparison test.

Figure 6

IL-3Rα activation promotes VM and increases the vessel size in vivo: (A) NSG mice were injected with tumour cells (as indicated) and treated from day 0 with saline or IL-3 (20 ng/mL) twice a week for 4 weeks. The figure was partly generated using Servier Medical Art templates, which are licensed under a Creative Commons Attribution 3.0 Unported License; https://smart.servier.com, (accessed on 5 August 2022). Tumours were recovered at day 30; (B) representative images of MDA-MB-23- derived tumours untreated (Ctrl) or treated with IL-3 (20 ng/mL) stained with anti-CD31 antibody (brown colour) and PAS (violet colour); (C,D) relative quantification of VM/endothelium expressed as the number of CD31+/PAS+ per field ± SEM ((C) corresponds to control, while (D) refers to IL-3 treatment); (E) quantification of the vessel size in MDA-MB-231-derived tumours. The results are expressed as mean ± SEM; (F) representative images of MDA-MB-436-derived tumours untreated (Ctrl) or treated with IL-3 (20 ng/mL) stained with anti-CD31 antibody and PAS; (G,H) relative quantification of VM/endothelium expressed as the number of CD31+/PAS+ per field ± SEM ((G) corresponds to control, while (H) refers to IL-3 treatment); (I) Quantification of the vessel size in MDA-MB-436 derived tumours. Circles on diagrams represent the quantity of CD31+ structures per field, squares represent the quantity of PAS+ structures per field. The results are the mean ± SEM. Original magnification 20×, scale bar: 80 µm. Comparisons were performed using Student’s t-test.

Figure 7

IL-3Rα activation boosts PD-L1 expression in primary tumours and lung metastases: (A,B) representative images and quantification of PD-L1 expression in primary tumours derived from mice injected with MDA-MB-231 (A) and MDA-MB-436 (B) cells untreated or treated with IL-3; (C,D) representative images and quantification of PD-L1 expression in the lung tissues of mice injected with MDA-MB-231 (C) and MDA-MB-436 (D) cells left untreated or treated with IL-3. Diagram data are presented as PD-L1+ area/lung ± SEM; (E,G) representative images of scattered PD-L1+ cells in the lung tissues of mice injected with MDA-MB-231 (E) and MDA-MB-436 (G) cells untreated or treated with IL-3; (F,H) diagrams reporting the number of lung metastases counted according to PD-L1 expression in mice injected with MDA-MB-231 (F) or MDA-MB-436 (H) cells untreated or treated with IL-3. Diagram data are expressed as PD-L1+ metastasis/field ± SEM. Circles on diagrams represent PD-L1 quantification in control mice tissues, squares represent the quantity of PD-L1+ structures in the mice treated with IL-3. Original magnification 20× scale bar: 100 µm. Comparisons were performed using Student’s t-test.

Figure 8

Bioinformatics analysis of TCGA dataset: (A) volcano plots of DEGs in TNBC and luminal breast cancers. Log2FC is displayed on the y-axis and the −log10 of p-value on the x-axis. Dots represent upregulated and downregulated genes; the blue points are those whose p-values were statistically significant (p < 0.05); (B) correlation of gene expression between IL-3Rα and genes involved in EMT and CD274 (PD-L1) (correlation test). The closer the RHO corresponds to 1, the more the correlation. Results of the univariate model show the discriminating power of single genes (AUC). The p-value of the multivariate model showed that all selected genes were relevant for the efficiency of the model. * p-value ≤ 0.05; ** p-value ≤ 0.01; *** p-value ≤ 0.001; (C) ROC curve representing the multivariate logistic model. This model, composed of IL-3Rα, SNAI1, ZEB1, VIM, CTNNB1, MMP2, SPRY2, and CD274 (PD-L1), showed the effectiveness in discriminating breast cancer aggressiveness. (AUC = 0.86 95% CI = 0.82–0.89).