Differential Impact of Membrane-Bound and Soluble Forms of the Prognostic Marker Syndecan-1 on the Invasiveness, Migration, Apoptosis, and Proliferation of Cervical Cancer Cells

Abstract

Cervical cancer ranks fourth among the most commonly diagnosed malignant tumors in women worldwide. Previously published evidence suggested a possible connection between the expression of the membrane-bound heparan sulfate proteoglycan syndecan-1 (Sdc-1) and the development of cervical carcinoma. Sdc-1 serves as a matrix receptor and coreceptor for receptor tyrosine kinases and additional signaling pathways. It influences cell proliferation, adhesion, and migration and is seen as a modulator of the tumor microenvironment. Following proteolytic cleavage of its extracellular domain in a process called shedding, Sdc-1 can act as a paracrine effector. The loss of Sdc-1 expression is associated with low differentiation of cervical carcinoma and with an increased rate of lymph node metastases. Here, we analyzed the clinical impact of Sdc-1 expression by analysis of public gene expression datasets and studied the effect of an overexpression of Sdc-1 and its membrane-bound and soluble forms on the malignant properties of the human cervical carcinoma cell line HeLa through functional analysis. For this purpose, the HeLa cells were stably transfected with the control plasmid pcDNA3.1 and three different Sdc-1-DNA constructs,encoding wild-type, permanently membrane-bound, and constitutively soluble Sdc-1. In clinical specimens, Sdc-1 mRNA was more highly expressed in local tumor tissues than in normal and metastatic cervical cancer tissues. Moreover, high Sdc-1 expression correlated with a poor prognosis in Kaplan-Meier survival analysis, suggesting the important role of Sdc-1 in the progression of this type of cancer. In vitro, we found that the soluble, as well as the permanently membrane-bound forms of Sdc-1 modulated the proliferation and the cell cycle, while membrane-bound Sdc1 regulated HeLa cell apoptosis. The overexpression of Sdc-1 and its soluble form increased invasiveness. In vitro scratch/wound healing assay, showed reduced Sdc-1-dependent cell motility which was linked to the Rho-GTPase signaling pathway. In conclusion, in cervical cancer Sdc-1 modulates pathogenetically relevant processes, which depend on the membrane-association of Sdc-1.

Keywords: cervical cancer; metastasis; prognosis; proteoglycan; shedding; syndecan-1.

Figure 1

Sdc-1 is dysregulated in cervical carcinoma, and its high expression correlates with poor overall survival. Boxplots of SDC-1 gene expression in cervical cancer tissue when comparing paired normal and tumor RNA Seq data of three patients. (A) and when comparing normal, tumor and metastasis RNA Seq data (B). The quantile cutoff values (minimum, 1st quartile, median, 3rd quartile, maximum) and the number of analyzed samples are presented. (C) Kaplan-Meier Survival analysis of 304 cervical carcinoma patients stratified by high or low expression of Sdc-1. (A) p=0.181 (not significant), Mann-Whitney-Test. (B) p=0.069 (Kruskal-Wallis test, trend for significance), post hoc-analysis p-values (Dunn’s test): p=1.65x10^-03 normal tissue compared to tumors (significant), p=0.175 tumor tissue compared to metastases (not significant), p=0.115 normal tissue compared to metastatic tissue. (C) p=0.024 (significant), Cox regression analysis log rank p-value.

Figure 2

Characterisation of HeLa cells overexpressing wild-type, constitutively membrane bound (Sdc-388) and constitutively shed (Sdc1-392) Sdc-1. (A) Schematic representation of the plasmid Sdc-1 cDNA inserts, CMV, cytomegalovirus promoter; yellow box, juxta-membrane domain; Black box, non-cleavable CD4 sequence; blue box, transmembrane and cytoplasmic domain, orange box, poly-A-tail; Sdc1-WT murine wild-type form; Sdc1-388, uncleavable construct 388;Sdc1-392, constitutively shed construct 392. (B) Quantitative PCR analysis of murine (mu Sdc-1 and human (Hu Sdc-1) Sdc1 expression. Sdc-1 expression was related to the housekeeping gene 18SrRNA. n ≥ 3, error bars = SEM. (C) Detection of human Sdc-1 protein expression at the surface of the transfected HeLa cell lines using flow cytometry. Cells were stained for isotype control mouse IgG1-PE and mouse anti-human Sdc-1 (CD138)-PE and the cells were subjected to flow cytometry. Human Sdc-1 is expressed at the cell surface of all cell types. (D) Immunocytochemistry for murine Sdc-1, demonstrating expression of murine Sdc-1 in Sdc1-WT, Sdc-1-388 and Sdc-1 392 transfected cells (brown-red staining). Original magnification 10x. (E, F) Detection of shed human (E) and murine (F) Sdc-1 in cell culture supernatants of the transfected cell lines. Conditioned media were collected from the cell lines indicated and 600 µl were subjected to a dotblot assay and quantified by Image J densitometric analysis. Left panels = quantification, right panels= representative dot-blots, n>3, *= p<0.05 Sdc1-392 compared to vector control (t-test). The cell lines shed comparable amounts of human Sdc-1 into the culture media, with a moderately, yet significantly enhanced amount in Sdc1-392 cells. Shed amounts of murine Sdc-1 were variable, with Sdc1-WT cells showing significantly increased levels of shed murine Sdc-1 compared to vector control (n>3, *p<0.05, t-test).

Figure 3

Role for soluble and membrane bound forms of Sdc-1 in HeLa cell proliferation, cell cycle progression, apoptosis and invasion. (A) Differential effect of membrane-bound and soluble Sdc-1 on breast cancer cell proliferation. Control vector-transfected HeLa and HeLa cells stably overexpressing WT (Sdc1-WT), constitutively membrane-bound (Sdc1-388) or the soluble ectodomain (Sdc1-392) of Sdc1 were subjected to an Alamar Blue cell proliferation assay *P < 0.05 (One-way ANOVA with Dunn’s post hoc test) for Sdc1-388 compared to vector control and Sdc1-392 compared to vector control n ≥ 3, error bars = SEM. Changes in apoptosis (B) and cell cycle progression (C) after the stable transfection of HeLa cells as quantified by using Annexin V/propidium iodide (B) and by DNA staining (C) and followed by flow cytometry. *P < 0.05 (One-way ANOVA with Dunn’s post hoc test) for Sdc1-388 compared to vector control, and Sdc1-392 compared to vector control, n ≥ 3, error bars = SEM. (D) quantitative Real-Time PCR of the expression of the apoptosis markers Bad, Bak and Bcl-2 in the HeLa Sdc1-388 transfected cells. Data are expressed as fold change versus control vector-transfected cells. ns, no significant p value for all group comparisons (one-way ANOVA with Dunn’s post hoc test). n≥3, error bars = SEM. (E) Stably transfected HeLa cells were subjected to a matrigel invasion assay. Quantification of invasive cells relative to control vector-transfected cells. *p <0.05 Sdc1-WT compared to vectror controls and SDc1-392 compared to vector controls, (one-way ANOVA with Dunn’s post hoc test), n≥4, error bars = SEM. (F) quantitative RT-PCR analysis of E-cadherin, MMP2 and TIMP1 compared with vector controls. *p <0.05, MMP2 Sdc1-388 compared to vector controls, E-cadherin Sdc1-388 compared to vector controls, ***p <0.001, E-cadherin Sdc1-392 compared to vector controls (one-way ANOVA with Dunn’s post hoc test), n≥3, error bars = SEM.

Figure 4

Migration of HeLa cells is decreased by all forms of Sdc-1 in a Rho-GTPase dependent manner. (A) The Scratch/wound area of the four cell lines is shown as a percentage of the wound area at time 0h, 6h, 8.5h and 23h. The ability of all three cell constructs to migrate is reduced compared to the control cell line. * = p <0.05 (nonparametric Friedman’s test with Dunn’s posttest) for Sdc1-388 compared to vector control (t=6h, 8.5h, 23h), for Sdc1-WT vs vector control (t=23h) and for Sdc1-392 vs vector control (t=6h, 23h), n≥3, error bars = SEM. (B) Stably transfected HeLa cells were treated with the Rho kinase (ROCK) inhibitor Y-27632 and then a Scratch/wound area was created. The area of the three cell constructs and vector was quantified at 0h, 6h, 8.5h and 23h. ns, no significant difference compared to vector control (nonparametric Friedman’s test with Dunn’s posttest). n≥3, error bars = SEM. (C) Confocal immunofluorescence microscopic analysis of RhoB protein in HeLa transfected cells. green, RhoB; red, actin-binding protein phalloidin for cytoskeletal staining; blue, DAPI staining for nucleus. Representative images are presented. 40x magnification.