Transcriptional regulation of CYR61 and CTGF by LM98: a synthetic YAP-TEAD inhibitor that targets in-vitro vasculogenic mimicry in glioblastoma cells

Abstract

Glioblastoma (GBM) is a highly angiogenic malignancy of the central nervous system that resists standard antiangiogenic therapy, in part because of an alternative process to angiogenesis termed vasculogenic mimicry. Intricately linked to GBM, dysregulation of the Hippo signaling pathway leads to overexpression of YAP/TEAD and several downstream effectors involved in therapy resistance. Little is known about whether vasculogenic mimicry and the Hippo pathway intersect in the GBM chemoresistance phenotype. This study seeks to investigate the expression patterns of Hippo pathway regulators within clinically annotated GBM samples, examining their involvement in vitro regarding vasculogenic mimicry. In addition, it aims to assess the potential for pharmacological targeting of this pathway. In-silico analysis of the Hippo signaling members YAP1 , TEAD1 , AXL , NF2 , CTGF , and CYR61 transcript levels in low-grade GBM and GBM tumor tissues was done by Gene Expression Profiling Interactive Analysis. Gene expression was analyzed by real-time quantitative PCR from human U87, U118, U138, and U251 brain cancer cell lines and in clinically annotated brain tumor cDNA arrays. Transient gene silencing was performed with specific small interfering RNA. Vasculogenic mimicry was assessed using a Cultrex matrix, and three-dimensional capillary-like structures were analyzed with Wimasis. CYR61 and CTGF transcript levels were elevated in GBM tissues and were further induced when in-vitro vasculogenic mimicry was assessed. Silencing of CYR61 and CTGF , or treatment with a small-molecule TEAD inhibitor LM98 derived from flufenamic acid, inhibited vasculogenic mimicry. Silencing of SNAI1 and FOXC2 also altered vasculogenic mimicry and reduced CYR61 / CTGF levels. Pharmacological targeting of the Hippo pathway inhibits in-vitro vasculogenic mimicry. Unraveling the connections between the Hippo pathway and vasculogenic mimicry may pave the way for innovative therapeutic strategies.

Fig. 1

Increased gene expression of YAP1, TEAD1, and downstream effectors AXL, CTGF, and CYR61 in clinically annotated glioblastoma tumor tissues. In-silico analysis of transcript levels was performed using RNA extracted from clinical samples from glioblastoma (GBM) and low-grade glioma (LGG) (red boxes) and compared with healthy tissue (gray boxes) (*P < 0.05).

Fig. 2

Expression profiles of YAP1, TEAD1, AXL, CTGF, CYR61, and NF2 in four human glioblastoma cell lines, and in cDNA tissue arrays. Total RNA was extracted from four different human glioblastoma cell lines (U87, U118, U251, and U138). The gene expression level for YAP1, TEAD1, AXL, CTGF, CYR61, and NF2 was analyzed and compared by real-time quantitative PCR. (a) Primer validation and single amplicon products were confirmed using agarose gel electrophoresis for the U87 cell line. (b) The gene expression levels of the six selected genes were analyzed and quantified by quantitative PCR as described in the Methods section. Triplicates from a representative quantification, out of three independent experiments, are shown. (c) Cancer tissues (red boxes) and normal tissues (gray boxes) cDNA arrays covering 43 clinical samples of the pooled four stages of pathologist-verified brain cancer tissues as well as normal tissues were screened to assess TEAD1, YAP1, CYR61, and CTGF gene expression levels by quantitative PCR. *P < 0.05.

Fig. 3

In-vitro human glioblastoma cell lines, capillary-like structure formation on Cultrex. Cells were trypsinized and seeded on top of Cultrex to generate three-dimensional capillary-like structures as described in the Methods section. (a) Representative phase-contrast pictures were taken to monitor structure formation at 24 h in U87, U118, U138, and U251 (×4 magnification). (b) CYR61 and CTGF levels are induced in in-vitro vasculogenic mimicry (VM) of the human U87 glioblastoma cell line. Total RNA was extracted either from cells cultured on Cultrex (black bars) or cells cultured as two-dimensional adherent monolayers (white bars). Real-time quantitative PCR analysis was next used to study the expression of the Hippo pathway downstream effectors. (c) Representative phase-contrast pictures were taken (upper panels), and Wimasis analysis of the structures represented in blue (lower panels). (d) VM parameters were extracted from the Wimasis analysis of (c) and representative quantification provided for total tube length, total branching, or total loops. *P < 0.05.

Fig. 4

Silencing of CTGF or CYR61 alters the U87 capillary-like structure formation. (a) U87 glioblastoma cell monolayers were transiently transfected with a scrambled sequence (siScrambled) or a small interfering RNA (siRNA) directed against CTGF (siCTGF) or CYR61 (siCYR61). Total RNA was then extracted, and gene silencing efficiency was validated using real-time quantitative PCR. Gene levels were normalized on glyceraldehyde 3-phosphate dehydrogenase expression. (b) Transfected U87 cells were seeded on top of Cultrex, and three-dimensional capillary-like structures were generated for 24 h. Representative phase-contrast pictures were taken (upper panels) and Wimasis analysis (lower panels) was performed. (c) Vasculogenic mimicry parameters were extracted from the Wimasis analysis of (b) and representative quantification was provided for total tube length, total branching points, or total loops. PPiA, peptidylprolyl isomerase A.

Fig. 5

Transcriptomic analysis reveals that SNAI1 and FOXC2 expression levels are induced in U87 capillary-like structure formation. (a) Indirect target proteins of YAP1 were retrieved from STRING as described in the Methods section. (b) Human U87 glioblastoma cells were seeded as either two-dimensional (2D) monolayers (2D, white bars) or as three-dimensional (3D) capillary-like structures on top of Cultrex (3D, black bars) for 24 h. Total RNA was extracted from triplicate samples for each condition, and gene expression modulation for two of the predicted targets (FOXC2 and Snail) was assessed by real-time quantitative PCR as described in the Methods section. *P < 0.05.

Fig. 6

Silencing of SNAI1 or FOXC2 alters the U87 capillary-like structure formation. (a) U87 glioblastoma cell monolayers were transiently transfected with a scrambled sequence (siScrambled, white bars) or small interfering RNA (siRNA) directed against SNAI1 (siSNAI1, gray bars) or FOXC2 (siFOXC2, black bars). Total RNA was then extracted, and gene silencing efficiency was validated using real-time quantitative PCR. (b) Transfected U87 cells were seeded on top of Cultrex and three-dimensional capillary-like structures were generated for 24 h. Representative phase-contrast pictures were taken.

Fig. 7

Hippo pathway pharmacological targeting alters U87 capillary-like structure formation and nuclear expression of SNAI1. (a) Chemical structure of the tested TEAD inhibitors. (b) Human U87 glioblastoma cells were seeded on top of Cultrex and three-dimensional capillary-like structures were generated for 24 h in the absence (vehicle) or presence of 10 μmol/l Hippo inhibitors HC258, LM41, and LM98, respectively. Representative phase-contrast pictures were taken (upper panels) and Wimasis analysis (lower panels) was performed. (c) Vasculogenic mimicry (VM) parameters were extracted from the Wimasis analysis of (b) and representative quantification was provided for total tube length, branching points, or total loops. (d) The effect of compound LM98 on CTGF, CYR61, FOXC2, and SNAI1 gene expression was assessed by real-time quantitative PCR as described in the Methods section. LM98-treated (black bars) or nontreated (white bars) U87 glioblastoma cells were seeded as three-dimensional capillary-like structures on top of Cultrex for 24 h. Total RNA was extracted from triplicate samples for each condition. (e) Real-time cell migration was assessed using the xCELLigence as described in the Methods section. (f) Nuclear fractionation protocol was performed from two-dimensional or three-dimensional U87 treated or not with LM98. Total proteins were extracted and representative blots for Snail, Fibrillarin, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) are presented. Western blotting of nuclear fibrillarin enrichment and undetectable contamination from cytosolic GAPDH. *P < 0.05.