Effect of TF gene silencing on HCT116 tumor growth and angiogenesis. (A) TF siRNA-expressing clones (SI-2, SI-3, and to a lesser extent SI-9) demonstrate decreased cell surface TF activity consistent with TF gene knockdown. Presence of integrated TF siRNA sequence in the clones was confirmed by PCR (inset). (B) In vitro growth curves for HCT116 (▪) and the TF siRNA-expressing clones (SI-2, ○; SI-3, ▴; SI-9, ×) were found to be similar in monolayer culture. (C) In contrast, TF RNAi results in diminished in vivo tumor growth of HCT116 tumors in SCID mice. (D) After 11 days in vivo, robust formation of VWF⁺ vascular networks was observed in Matrigel plugs containing HCT116 cells. (E) Endothelial cell infiltration was noticeably reduced in plugs containing equivalent numbers of TF-deficient SI-3 cells. (F) The area occupied by VWF⁺ vascular networks was significantly lower in Matrigel plugs containing TF siRNA-expressing cells (after 11 days in vivo), relative to plugs containing parental HCT116 cells (^*P < .0001; area was expressed as a percentage of the total area of a × 20 field). No significant host cell invasion or VWF positivity was detected in pellets consisting of Matrigel alone. Northern blotting of tumors derived from TF down-regulated SI-3 cells revealed significantly increased expression of 2 angiogenesis inhibitors, (G) TSP-1 and (H) TSP-2, relative to their parental HCT116 counterparts, whereas expression of VEGF (I) was only slightly decreased in SI-3 tumors. Numbers indicate intensity of band relative to 28S rRNA loading control. (J) The working model that illustrates how TF may participate in mediating the K-ras-dependent angiogenic phenotype, in part by deregulating expression of angiogenesis inhibitors TSP-1 and TSP-2. Error bars indicate standard deviation. For panels D and E, images were acquired with a Zeiss Axioskop 2 microscope equipped with a Zeiss Achroplan 20×/0.45 objective lens (Zeiss, Oberkochen, Germany). Images were captured with a Sony DXC-350 video system (Sony, San Diego, CA).