Epac/Rap1 activation suppresses angiogenesis in vivo. (A) Matrigel plugs were implanted into 3 groups, as follows: negative control group received Matrigel with no supplements (left panel), and the remaining 2 received Matrigel supplemented with VEGF (200 pg/mL) and heparin (60 U/mL), with (right) or without (center) 2.5 μmol 8CPT. No further addition of 8CPT was performed during the experiment. Blood vessels were stained with anti-CD31 and rhodamine-conjugated secondary antibodies. Microvessel density was plotted for each group. Bars show ± SE. *P < .05 (n = 4). (B) Biodegradable scaffolds containing HMVECs were implanted in SCID mice, as described in “Methods.” Daily subcutaneous injections of buffer alone (control vehicle, left panel), 0.03 μmol (center), or 0.3 μmol (right) of 8CPT were performed for 5 days. Sections of implants were stained with an anti–human CD31 antibody to detect blood vessels. Secondary antibodies were peroxidase conjugated, and the chromagen was diaminobenzidine. Pictures shown are a representative field from each treatment group. The number of CD31-positive microvessels was counted in 10 random fields per scaffold using optical microscopy and plotted. Bars show ± SE. **P < .01; ***P < .001 (n = 6). (C) SCID angiogenesis assay was performed as in (B), except HMVECs expressed either CA Rap1 or control vector. Bars show ± SE. *P < .05; ***P < .001 (n = 6).