Endoglin and squamous cell carcinomas

Abstract

Despite the fact that the role of endoglin on endothelial cells has been extensively described, its expression and biological role on (epithelial) cancer cells is still debatable. Especially its function on squamous cell carcinoma (SCC) cells is largely unknown. Therefore, we investigated SCC endoglin expression and function in three types of SCCs; head and neck (HNSCC), esophageal (ESCC) and vulvar (VSCC) cancers. Endoglin expression was evaluated in tumor specimens and 14 patient-derived cell lines. Next to being expressed on angiogenic endothelial cells, endoglin is selectively expressed by individual SCC cells in tumor nests. Patient derived HNSCC, ESCC and VSCC cell lines express varying levels of endoglin with high interpatient variation. To assess the function of endoglin in signaling of TGF-β ligands, endoglin was overexpressed or knocked out or the signaling was blocked using TRC105, an endoglin neutralizing antibody. The endoglin ligand BMP-9 induced strong phosphorylation of SMAD1 independent of expression of the type-I receptor ALK1. Interestingly, we observed that endoglin overexpression leads to strongly increased soluble endoglin levels, which in turn decreases BMP-9 signaling. On the functional level, endoglin, both in a ligand dependent and independent manner, did not influence proliferation or migration of the SCC cells. In conclusion, these data show endoglin expression on individual cells in the tumor nests in SCCs and a role for (soluble) endoglin in paracrine signaling, without directly affecting proliferation or migration in an autocrine manner.

Keywords: BMP-9; TGF-β; TRC105; endoglin; squamous cell carcinoma.

Figure 1

Analysis of squamous cell carcinoma (SCC) primary tumors via immunohistochemistry, all tissues were stained for endoglin expression (brown). The black arrows indicate endothelial endoglin expression. The white arrows indicate epithelial endoglin expression. (A) Representative images of ESCC (n = 9). (B) Representative images of HNSCC (n = 5). (C) Representative images of VSCC (n = 7). Images taken at 100x (left) and 200x (right).

Figure 2

Analysis of HNSCC via imaging mass spectrometry (Hyperion) with a six marker panels. (A) Six images, each depicting the expression of the corresponding marker. (B) A merged image combining pan-cytokeratin, endoglin, and p53 expression. The white arrow indicates cells that co-express pan-cytokeratin and endoglin, which are negative for p53. (C) A merged image combining pan-cytokeratin, endoglin, and CD68 expression. The white arrow indicates cells that co-express pan-cytokeratin and endoglin, which are negative for CD68.

Figure 3

Analysis of ESCC via imaging mass spectrometry (Hyperion) with a six marker panels. (A) Six images, each depicting the expression of the corresponding marker. (B) A merged image combining pan-cytokeratin, endoglin, and p53 expression. The white arrows indicate cells that co-express pan-cytokeratin and endoglin, which are negative for p53. (C) A merged image combining pan-cytokeratin, endoglin, and CD68 expression. The white arrow indicates cells that co-express pan-cytokeratin and endoglin, which are negative for CD68. The blue arrow indicates cells that co-express pan-cytokeratin, endoglin, and CD68.

Figure 4

Analysis of VSCC via imaging mass spectrometry (Hyperion) with a six marker panel. (A) Six images, each depicting the expression of the corresponding marker. (B) A merged image combining pan-cytokeratin, endoglin, and p53 expression. The white arrows indicate cells that co-express pan-cytokeratin and endoglin. (C) A merged image combining pan-cytokeratin, endoglin, and CD68 expression. The white arrow indicates cells that co-express pan-cytokeratin and endoglin, which are negative for CD68. The blue arrow indicates cells that co-express pan-cytokeratin, endoglin, and CD68.

Figure 5

The expression of endoglin by SCC cell lines. Expression by 10 ESCC cell lines, where endoglin expression by TE01 ( p < 0.0001) and TE15 (p < 0.003) significantly differ from all other cell lines (A), OSC-19 and FaDu show significantly different endoglin expression (p = 0.0002) (B), as is also detected in the three VSCC cell lines (*p = 0.0123, **p = 0.0025, ***p = 0.0001) with different morphologies (conventional—VC415-C and VC704; spindle—VC415-S) (C). Endoglin protein levels were determined via western blot (D) and ELISA. Endoglin protein expression by TE01 significantly differs from all other cell lines—(p ≤ 0.0018) (E). Image for western blot analysis is a representative image of n = 2–3 independent experiments.

Figure 6

Analysis of ALK expression by SCC cell lines, determined via qPCR. Gene expression for ALK1—ALK7 by TE10—endoglin low (A), TE11—endoglin low (B), TE01—endoglin high (C), OSC-19—endoglin positive (D), FaDu–endoglin high (E), VC415-C—endoglin low, and VC415-S—endoglin high; ***p = 0.000581 (F). The effect of endoglin overexpression (OE) on ALK expression was analyzed for TE10; ***p=0.000126; **p=0.003291 (G) and TE11 (H), as well as the effect of endoglin knockout (KO) in TE01 (I).

Figure 7

SCC cells were stimulated with either BMP-6 (TE10 and TE11 only), BMP-9 or TGF-β and the level of phosphorylated SMAD1 and SMAD2 (pSMAD1 and pSMAD2) was determined via western blot (A–C). The amount of soluble endoglin in the medium of TE10 and TE11 was determined via ELISA (D,E). Stimulation of OSC-19 (F), FaDu (G), VC415-C (H), and VC415-S (I) with BMP-9/TGF-β/TRC105 was performed, and the levels of pSMAD1 and pSMAD2 were determined via western blot. Western blot images are representative of n = 2–3 per experiment.

Figure 8

SCC cell lines were stimulated with BMP-9 or TGF-β and the proliferation of TE10 (A), TE11 (C), TE01 (E), and VC415-S (G) cells was measured via a MTS assay. To assess the effects of endoglin on cell proliferation, MTS assays were performed on endoglin overexpressing (OE) TE10 (B) and TE11 (D) cells. The effects of endoglin knockout (KO) in TE01 (F) and endoglin knockdown (KD) in VC415-S (H) were also assessed via MTS. n = 2–3 for each experiment.

Figure 9

SCC cell lines were stimulated with BMP-9 or TGF-β and the migration of TE10 (A), TE11 (C), TE01 (E), VC415-S (G), OSC-19 (I), and FaDu (J) cells was measured via a wound healing assay. To assess the effects of endoglin on cell migration, wound healing assays were performed on endoglin overexpressing (OE) TE10 (B) and TE11 (D) cells. The effects of endoglin knockout (KO) in TE01 (F) and endoglin knockdown (KD) in VC415-S (H) were also assessed. Finally, the effects of TRC105 on cell migration was assessed in OSC-19 (I) and FaDu cells (J). n = 2–3 for each experiment.