(B) PFS by serum Ang-2. (C) OS by serum Ang-2. Ang-2, angiopoietin-2. Figure 4 Outcome to bevacizumab-containing therapy in CRC by serum VEGF (n=34), tumour MVD and PC (n=17). (A) PFS by serum http://www.selleckchem.com/products/Oligomycin-A.html VEGF. (B) OS by serum VEGF. (C) PFS by MVD. (D) OS by MVD. (E) PFS by PC. (F) OS by PC. VEGF, vascular endothelial growth factor, MVD, microvessel … Discussion Bevacizumab is a VEGF-targeting antibody that is widely used in combination with chemotherapy to treat metastatic CRC (Cercek and Saltz, 2008; McCormack and Keam, 2008). Although much has been learned about its mechanisms of action, suitable biomarkers predicting patients who are likely to benefit from bevacizumab treatment remain elusive (Jubb et al, 2006b; Sessa et al, 2008; Murukesh et al, 2010).

Expression levels of VEGF, in particular, are not predictive of outcome in CRC patients treated with bevacizumab (Jubb et al, 2006a). So far, the search for outcome predictors in CRC has failed for bevacizumab and antibody-free chemotherapeutic regimens alike (De Roock et al, 2009). Although VEGF is primarily produced by tumour cells, its target structure is the tumour vasculature embedded in the stromal compartment, where the therapeutic effects of the antibody involve extensive changes such as blood vessel pruning and reorganisation of the chaotic tumour vasculature (Willett et al, 2004). Thus, stromal factors controlling the responsiveness of blood vessels to VEGF withdrawal rather than determinants of VEGF availability are attractive candidates as outcome predictors for bevacizumab treatment.

Potentially, stromal factors are also outcome predictors of chemotherapy, because the delivery of cytostatic drugs to tumour cells is controlled by the vascular tumour microenvironment. Angiopoietin-2 has been proposed as a gatekeeper of VEGF function and vascular remodelling. (Hanahan, 1997; Augustin et al, 2009). We here identified Ang-2 as a stromal factor in CRC. In tumour lesions of CRC patients and in a murine xenograft model of CRC, Ang-2 mRNA was expressed exclusively in the tumour stromal compartment, but not in the tumour cell compartment itself. Although these findings are at odds with previous immunohistochemical studies reporting Ang-2 expression in the tumour cell compartment of CRC (Ochiumi et al, 2004; Ogawa et al, 2004; Chung et al, 2006; Gu et al, 2006), the published immunohistochemical data should be interpreted with caution owing to the limited specificity of the available antibodies.

Indeed, careful analysis Batimastat of the tissue localisation of Ang-2 expression in cancer entities and tumour models other than CRC has called into question the tumour cell origin of Ang-2 (Zhang et al, 2003). To further elucidate the clinical impact of stromal-derived Ang-2, we measured serum Ang-2 concentrations in CRC patients. Serum Ang-2 levels were significantly elevated in patients with metastatic disease. Indeed, Ang-2 has been shown to promote metastatic growth (Imanishi et al, 2007).