The tyrosine-kinase inhibitor sunitinib targets vascular endothelial (VE)-cadherin: a marker of response to antitumoural treatment in metastatic renal cell carcinoma

Abstract

Background: Vascular endothelial (VE)-cadherin is an endothelial cell-specific protein responsible for endothelium integrity. Its adhesive properties are regulated by post-translational processing, such as tyrosine phosphorylation at site Y685 in its cytoplasmic domain, and cleavage of its extracellular domain (sVE). In hormone-refractory metastatic breast cancer, we recently demonstrated that sVE levels correlate to poor survival. In the present study, we determine whether kidney cancer therapies had an effect on VE-cadherin structural modifications and their clinical interest to monitor patient outcome.

Methods: The effects of kidney cancer biotherapies were tested on an endothelial monolayer model mimicking the endothelium lining blood vessels and on a homotypic and heterotypic 3D cell model mimicking tumour growth. sVE was quantified by ELISA in renal cell carcinoma patients initiating sunitinib (48 patients) or bevacizumab (83 patients) in the first-line metastatic setting (SUVEGIL and TORAVA trials).

Results: Human VE-cadherin is a direct target for sunitinib which inhibits its VEGF-induced phosphorylation and cleavage on endothelial monolayer and endothelial cell migration in the 3D model. The tumour cell environment modulates VE-cadherin functions through MMPs and VEGF. We demonstrate the presence of soluble VE-cadherin in the sera of mRCC patients (n = 131) which level at baseline, is higher than in a healthy donor group (n = 96). Analysis of sVE level after 4 weeks of treatment showed that a decrease in sVE level discriminates the responders vs. non-responders to sunitinib, but not bevacizumab.

Conclusions: These data highlight the interest for the sVE bioassay in future follow-up of cancer patients treated with targeted therapies such as tyrosine-kinase inhibitors.

Trial registration: ClinicalTrials.gov NCT00619268 NCT00943839.

Conflict of interest statement

The authors declare no competing interest.

Figures

Fig. 1

Sunitinib impairs VEGF-induced VE-cadherin tyrosine phosphorylation and cleavage: primary ECs were grown 3–4 days to reach confluency (3 × 106 cells). After 3 h of serum starvation, ECs were stimulated by VEGF alone (50 ng/mL) or in the presence of SUT (20 μM) or TEM (50 ng/mL) during indicated times. Lysates were analysed by SDS-PAGE on 10% polyacrylamide gels and western blotting (5 μg of total protein lysate/lane). a Analysis of phosphotyrosine-containing protein pattern using the anti-phosphotyrosine antibody 4G10. b, c Analysis of phospho-Y685VE-cadherin d. Conditioned media from untreated ECs (CTL), or VEGF-stimulated-ECs or VEGF-stimulated-ECs treated with SUT or TEM for 20 min (same amount of cells in each condition) were collected, centrifuged to discard floating cells, and concentrated on Centriprep Centrifugal Filter Units with an Ultracel YM-30 membrane. Volume of 5 μL of concentrated medium of each condition was analysed by SDS-PAGE and western blotting using the monoclonal anti-VE-cadherin antibody (BV9 clone). e Same experiment as in d with VEGF-treated ECs in the presence of either INFα2 A (7.5 ng/mL), Irinotecan (IRN, 10 μM) and SN-38 (10 μM). Results are representative of three independent experiments

Fig. 2

SUT impairs cell invasion and migration in heterotypic spheroids. a Representative images of spheroids formed from 786-O (RCC) cells alone at the time of their incorporation into collagen gel (T0) and 24 h later (T24), or RCC + ECs at time 0 and 24 h later. b Area of the spheroids was quantified using Image J software (National Institutes of Health). c Quantification of ECs migration from spheroids co-cultured with RCC using the measurement of the Euclidean distance of each EC (n = 500) from the centre of the spheroid after 24 h of culture. Results are representative of three independent experiments. Values are the mean ± SD of the migration distance in μm. Statistical significance is indicated by the number of asterisks (*)

Fig. 3

Effect of RCC tumour microenvironment on VE-cadherin cleavage: elevated levels of sVE in mRCC patients: a Phase contrast microscopy of RCC cells in culture. b Analysis of RCC culture media by zymography showed the presence of MMP2 and MMP9 activities. c The relative amounts of MMPs were measured by densitometry of autoradiographs using ImageJ software (National Institutes of Health). d Phase contrast microscopy of ECs in culture. e Release of sVE by ECs upon RCC conditioned media treatment (20 min) (left panel) alone or RCC conditioned media pre-incubated or not with bevacizumab (BEV) (5 μg/mL) (right panel). f, g The relative amounts of sVE were measured by densitometry of autoradiographs using ImageJ software (National Institutes of Health). Results are representative of three independent experiments. Values are the mean ± SD. h sVE levels in healthy donors (n = 96) and mRCC patients from the clinical trial TORAVA (n = 115) at diagnosis (Day 0, D0) was performed by ELISA. Statistical significance is indicated by the number of asterisks (*)

Fig. 4

Analysis of sVE- levels in mRCC patients after the first cycle of treatment. a Serum from mRCC patient treated with SUT was analysed for sVE-cadherin level using the monoclonal anti-VE-cadherin antibody. b sVE levels in mRCC patients after the first cycle of treatment. c Variations of sVE expressed as % of the initial concentration at diagnosis (level of sVE after the first cycle/level of sVE at diagnosis × 100) in SUT group with stable disease and relapse. d Variations of sVE expressed as % of the initial concentration at diagnosis (level of sVE after the first cycle/level of sVE at diagnosis × 100) in BEV group with stable disease and relapse. Statistical significance is indicated by the number of asterisks (*)

Fig. 5

Relationship between sVE-cadherin levels and PFS of mRCC patients: a–c Kaplan–Meier analysis of PFS of patients with RCC. PFS was calculated from patient subgroups with variations of sVE expressed as % of the initial concentration at diagnosis (level of sVE after the first cycle/level of sVE at diagnosis × 100) that were less or greater than a cutoff ratio at 120%, for SUT group trial (a), (b) BEV group. Statistical significance (P-value) and the time of the median disease free are indicated

Fig. 6

Proposed mechanisms of microenvironment-induced VE-cadherin structural modifications in kidney cancer. This working hypothesis was mainly built from our results obtained using in vitro EC cultures and VEGF as the major actor in angiogenesis., The involvement of other cytokines in these processes cannot be excluded since we previously showed that TNFα induced sVE release