Sulfatide decreases the resistance to stress-induced apoptosis and increases P-selectin-mediated adhesion: a two-edged sword in breast cancer progression

Abstract

Background: We have previously shown that galactosylceramide (GalCer) affects the tumourigenic and metastatic properties of breast cancer cells by acting as an anti-apoptotic molecule. Since GalCer is a precursor molecule in the synthesis of sulfatides, the present study was aimed to define the role of sulfatides in apoptosis and breast cancer progression.

Methods: Expression of GAL3ST1 in breast cancer cell lines and breast cancer tissue specimens was analysed using real-time PCR, western blotting and immunohistochemistry analysis. The amount of sulfatide, GalCer and ceramide was analysed by thin-layer chromatography binding assay and by the modified hydrophilic interaction liquid chromatography coupled with electrospray mass spectrometry methodology. The tumourigenicity of cancer cells was analysed by an in-vivo tumour growth assay. Apoptotic cells were detected based on caspase-3 activation and the TUNEL assay. The interaction of breast cancer cells with P-selectin or E-selectin was analysed using the flow adhesion assay. The ability of sulfatide-expressing cells to activate and aggregate platelets was studied using the flow-cytometry-based aggregation assay.

Results: Using two models of breast cancer, T47D cells with blocked synthesis of sulfatide and MDA-MB-231 cells with neosynthesis of this glycosphingolipid, we showed that high sulfatide levels resulted in increased sensitivity of cancer cells to apoptosis induced by hypoxia and doxorubicin in vitro, and decreased their tumourigenicity after transplantation into athymic nu/nu mice. Accordingly, a clinical study on GAL3ST1 expression in invasive ductal carcinoma revealed that its elevated level is associated with better prognosis. Using MDA-MB-231 cells with neosynthesis of sulfatide we also showed that sulfatide is responsible for adhesion of breast cancer cells to P-selectin-expressing cells, including platelets. Sulfatide also acted as an activating molecule, increasing the expression of P-selectin.

Conclusions: This study demonstrates that increased synthesis of sulfatide sensitises cancer cells to microenvironmental stress factors such as hypoxia and anticancer drugs such as doxorubicin. However, sulfatide is probably not directly involved in apoptotic cascades, because its increased synthesis by GAL3ST1 decreased the amounts of its precursor, GalCer, a known anti-apoptotic molecule. On the other hand, our data support the view that sulfatides are malignancy-related adhesive molecules involved in activating and binding P-selectin-expressing platelets to breast cancer cells.

Keywords: Adhesion; Apoptosis; Breast cancer; Galactosyloceramide; P-selectin; Platelets; Sulfatide.

Fig. 1

Expression of GAL3ST1 and sulfatide in breast cancer cell lines and characteristics of human breast cancer cell lines with overexpression or blocked expression of GAL3ST1 and sulfatide. a Expression of GAL3ST1 mRNA in breast cancer cell lines. Real-time PCR used to analyse GAL3ST1 mRNA. GAL3ST1 expression levels normalised against β-actin and MDA-MB-231 cells served as calibrator sample. Results expressed as mean. b Western blotting analysis of GAL3ST1 expression in breast cancer cell lines. Anti-GAL3ST1 rabbit polyclonal antibodies used to detect GAL3ST1 in cell lysates. c Immunostaining of gangliosides from human breast cancer cell lines, separated by HP-TLC, with mouse monoclonal antibody against sulfatide. d Western blotting analysis of GAL3ST1 expression in parental MDA-MB-231 cells, control MDA-MB-231 cells (MDA.CTR) and MDA-MB-231 cells overexpressing sulfatide (MDA.SUL). e Immunostaining of gangliosides from MDA-MB-231, MDA.CTR and MDA.SUL cells. f Western blotting analysis of GAL3ST1 expression in parental T47D cells, control T47D cells (T47D.CRISPR.C) and T47D cells with knockout of GAL3ST1 (T47D.Δ.GAL3ST1.1). g Immunostaining of gangliosides from T47D, T47D.CRISPR.C and T47D.Δ.GAL3ST1.1 cells. For western blotting, 40 μg of proteins separated by SDS-PAGE under reducing conditions on a 12% gel and electrophoretically transferred onto a nitrocellulose membrane. β-actin served as internal control. For immunostaining, aliquots of total gangliosides corresponding to 1 × 10⁷ cells applied to HP-TLC plate. Gal3ST1 galactosylceramide sulfotransferase

Fig. 2

Sensitivity of breast cancer MDA-MB-231 and T47D cells with varying expression of GAL3ST1 to apoptosis induced by doxorubicin and hypoxia. Cells grown in presence of doxorubicin at concentration of 0.5 μM (MDA-MB-231 cells) or 1 μM (T47D) and in hypoxic conditions (1% O₂) for 48 h. Cellular response measured by presence of active form of caspase-3 (a, b) or staining with Annexin V and propidium iodide (c, d). Statistically significant differences (**p < 0.01, *p <0.05). Flow cytometry dot plots show percentage of early apoptotic cells (Annexin V⁺/PI⁻, lower right) and late apoptotic cells (Annexin V⁺/PI⁺, upper right). Levels of galactosyloceramide (GalCer) and UGT8 in MDA-MB-231 and T47D cells with varying expression of GAL3ST1. Immunostaining of neutral glycolipids from MDA-MB-231, MDA.C and MDA.SUL cells (e) and from T47D, T47D.CRISPR.C and T47D.Δ.GAL3ST1.1 cells (f) separated by HP-TLC. For immunostaining, aliquots of neutral glycosphingolipids corresponding to 1 × 10⁷ cells applied to HP-TLC plate and stained with anti-GalCer rabbit polyclonal antibodies. g Real-time PCR used to analyse expression of UGT8 mRNA in MDA-MB-231 and T47D cell lines with varying expression levels of sulfatide. UGT8 levels normalised against β-actin and T47D.CRISPR.C cells assigned as calibrator sample. Results expressed as mean. h Western blotting analysis of UGT8 expression in breast cancer cell lines with different expression levels of sulfatide. Anti-UGT8 rabbit polyclonal antibodies used to detect UGT8 in cell lysates. FITC fluorescein isothiocyanate, GalCer galactosyloceramide, PI propidium iodide

Fig. 3

Xenograft tumour growth of MDA.SUL cells with overexpression of GAL3ST1 and control MDA.C cells. a Tumour growth recorded once a week by measuring diameter with a calliper. Data shown as mean tumour volume for group of mice (n   =  8 for MDA.SUL cells and n = 9 for MDA.C cells) ± SE at each indicated time point. Data analysed using Bonferroni multiple comparison test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. b TUNEL assay after subcutaneous implantation of MDA.SUL (I) and MDA.C (II) cells. Arrows indicate the apoptotic cells. Numbers of apoptotic cells in MDA.SUL and MDA.C tumours compared (III) by Mann–Whitney U test (p = 0.0494). TUNEL terminal transferase dUTP nick end labelling

Fig. 4

Relative quantification of ceramides in selected cell lines. a Extracted ion chromatograms for 264.269 ± 001 m/z in cell line samples with peak at RT = 2.1 min corresponding to HexCer. RI stands for Relative Intensity. b Structure of reporter ion selected for relative quantification. c Relative quantities of Cer (dark grey) and HexCer (light grey) in breast cancer cell lines. 100% value corresponds to highest concentration of particular analyte. Error bars represent standard deviations of three biological replicates. Statistically significant differences (*p < 0.01) in analyte levels

Fig. 5

Binding of sulfatide-expressing (a) MDA.SUL cells and (b) MDA.C cells to P-selectin-expressing CHO-Pro-5 (CHO-Pro5/SELP) cells or E-selectin-expressing CHO-Pro-5 (CHO-Pro5/ELAM) cells and control CHO-Pro-5 cells under defined laminar flow conditions. Points on graphs represent mean numbers of breast cancer cells interacting with CHO-Pro-5 cells obtained from three independent experiments. Formation of heterotypic tumour cell–platelet aggregates by breast cancer MDA-MB-231 cells and activated platelets. Turbidimetric platelet aggregation assay with activated platelets and MDA-MB-231, MDA.C and MDA.SUL cells. Platelets (10⁶) and tumour cells (10⁶) in 1 ml of Tyrode’s buffer incubated in absence (c) or presence of 1 μM (d) or 5 μM ADP (e). ADP adensoine diphosphate, PPP platelet-poor plasma, PRP platelet-rich plasma

Fig. 6

Flow cytometry analysis of tumour cell–platelet aggregates. a MDA.C and (b) MDA.SUL cells labelled with DiD (red) and platelets labelled with DiO (green) incubated for 1 h at room temperature. Box marks mixed aggregates formed by cancer cells and platelets. Sulfatides present on breast cancer cells stimulate activation of platelets. Flow cytometric analysis of P-selectin expression using monoclonal antibody against P-selectin human platelets (10⁶) (c) non-activated and activated with ADP at concentration of (d) 1 μM or (e) 5 μM incubated with (10⁶) parental breast cancer MDA-MB-231 cells (solid line), control MDA.C cells (dotted line) and MDA.SUL with overexpression of sulfatide (grey). Platelets incubated only with secondary antibodies used as control (dashed line). ADP adensoine diphosphate

Fig. 7

Expression of GAL3ST1 in IDC tissue specimens. Immunohistochemical staining of breast cancer tumours representing malignancy grades: (a) G1 and (b) G3 (magnification 400×). c Reaction intensities with anti-GAL3ST1 antibody calculated on basis of semi-quantitative IRS scale of Remmele and Stegner [21] and represented as mean. **p < 0.01 for G1 grade breast tumours compared to G3 grade breast tumours (Mann–Whitney U test). d Mantel–Cox analysis distinguished breast cancer patients who were GAL3ST1-positive and GAL3ST1-negative. Patients expressing GAL3ST1 had longer overall survival (*p < 0.05). e Overall survival of GAL3ST1-positive and GAL3ST1-negative breast cancer patients who received chemotherapy (Mantel–Cox analysis, p = 0.1206). Gal3ST1 galactosylceramide sulfotransferase, IRS semi-quantitative immune reactive score, SD standard deviation

Fig. 8

Sulfatides act as a two-edged sword in breast cancer progression