Re7107

For tissue microarray construction, 20 ovarian carcinoma tissue samples were retrieved from the Pathology Department of the Academic Hospital of Reims (Reims, France). Immunohistochemistry (IHC) analysis was performed on paraffin-embedded tissue, using anti-TSP-1 (#ab1823, Abcam, Cambridge, UK) and anti-CD47 (#ab192827, Abcam) primary antibodies, followed by standard streptavidin–biotin–peroxidase complex method (Novolink Polymer and Rabbit Specific HRP/DAB [ABC] detections systems from Leica Biosystems and Abcam, respectively). Sections were counterstained with hematoxylin/eosin (#RE7107, Leica Biosystems, Danaher Corporation, Washington, WA, USA). Control isotypes (#ab172730 and #ab81216, Abcam) were used to confirm specificity of immunostaining, while paired para-carcinoma tissues were used as non-tumor control samples. Microarrays were reviewed by a pathologist (blinded to clinical information) for evaluation of intensity of positive staining in order to examine the expression of TSP-1 and CD47, as well as the protein cellular/subcellular localization. Apart from Reims Hospital tissue microarray, the pattern of TSP-1 staining in normal ovary and ovarian carcinoma tissue was also characterized in the Human Protein Atlas web portal (available from www.proteinatlas.org, accessed on 11 January 2018) [30 (link)].

Histological analyses of paraffin-embedded tissues were performed on hematoxylin, eosin and saffron (HES)-stained 3 μm–thick sections that were prepared by using routine histological methods. Anti-CD31 (#ab182981, Abcam, Cambridge, UK; 1:1000), anti-Lyve-1 (#ab33682, Abcam, Cambridge, UK; 1:200), anti-CD4 (#ab183685, Abcam; 1:2000), anti-CD8 (#ab203035, Abcam; 1:250), anti-CD163 (#ab182422, Abcam; 1:500) and anti-TSP-1 (LS-C402945, LifeSpan Biosciences Inc.; 1:100) antibodies were used to perform immunostaining together with biotin-labeled secondary antibodies and streptavidin-HRP DAB (3,3’-diaminobenzidine) detection system (#64261, Abcam, Cambridge, UK), and then followed by hematoxylin/eosin counterstain (#RE7107, Leica Biosystems, Danaher Corporation, Washington, WA, USA). Negative controls were performed by omitting the primary antibody. The number of positive cells for each marker, as well as vessel density, was determined by an experienced pathologist who was blinded to the group’s treatment during the whole analysis procedure in 10 consecutive high-power fields (×400; 0.283 mm²) in the area with the highest density of staining. Moreover, automated quantitative image analyses across the whole s.c. allograft sections were performed, using ImageJ software binarization and thresholding tools.