Figure 2. Characterization of in vitro healthy and tumourigenic vasculature.

(a) HDFs and HUVECs (stromal only cultures) were cultured for 21 days under normal atmospheric oxygen (normoxia ~21% O₂) in 10% collagen gels with and without laminin and in physiological hypoxia (~5% O₂) with laminin only. EC morphology was assessed using immunofluorescence of CD31 (green) and DAPI (blue). (b) ImageJ was used to quantify the length, width, and number of branches, loops and junctions of vascular networks in the stromal only cultures. Normoxia cultures without laminin formed longer and wider networks than in the presence of laminin, however this may have been due to the presence of the EC cobblestones, which led to a significantly lower level of connectivity among the branches. Stromal cultures containing laminin had a significantly higher number of branches, loops and junctions in comparison to normoxia without laminin and hypoxia with laminin. (c) In the presence of cancer cells (biomimetic tumouroids), ECs formed two distinct morphologies, cobblestones and end-to-end vascular networks in the stroma as confirmed by immunofluorescence of CD31 (green). HT29 cells were stained for CK20 (red) while the unstained nuclei in the stroma are HDFs. (d) ImageJ was used to analyze the differences in tubule length and branch connectivity in the presence and absence of HT29 cancer cells and showed vascular network integrity was maintained in the absence of HT29 cancer cells. (e) Quantitative RT-PCR analysis of angiogenic genes downregulated in biomimetic tumouroids in comparison to the stromal only cultures and (f) the expression profiles of angiogenic genes expressed solely in either the stroma or in biomimetic tumouroids. Data is presented mean ± SD (n = 6). *p < 0.05, **p < 0.01, ***p < 0.0001. Scale bar – 100 μm.