Regional differences in prostaglandin E2 metabolism in human colorectal cancer liver metastases.

BACKGROUND: Prostaglandin (PG) E2 plays a critical role in colorectal cancer (CRC) progression, including epithelial-mesenchymal transition (EMT). Activity of the rate-limiting enzyme for PGE2 catabolism (15-hydroxyprostaglandin dehydrogenase [15-PGDH]) is dependent on availability of NAD+. We tested the hypothesis that there is intra-tumoral variability in PGE2 content, as well as in levels and activity of 15-PGDH, in human CRC liver metastases (CRCLM). To understand possible underlying mechanisms, we investigated the relationship between hypoxia, 15-PGDH and PGE2 in human CRC cells in vitro. METHODS: Tissue from the periphery and centre of 20 human CRCLM was analysed for PGE2 levels, 15-PGDH and cyclooxygenase (COX)-2 expression, 15-PGDH activity, and NAD+/NADH levels. EMT of LIM1863 human CRC cells was induced by transforming growth factor (TGF) ß. RESULTS: PGE2 levels were significantly higher in the centre of CRCLM compared with peripheral tissue (P = 0.04). There were increased levels of 15-PGDH protein in the centre of CRCLM associated with reduced 15-PGDH activity and low NAD+/NADH levels. There was no significant heterogeneity in COX-2 protein expression. NAD+ availability controlled 15-PGDH activity in human CRC cells in vitro. Hypoxia induced 15-PGDH expression in human CRC cells and promoted EMT, in a similar manner to PGE2. Combined 15-PGDH expression and loss of membranous E-cadherin (EMT biomarker) were present in the centre of human CRCLM in vivo. CONCLUSIONS: There is significant intra-tumoral heterogeneity in PGE2 content, 15-PGDH activity and NAD+ availability in human CRCLM. Tumour micro-environment (including hypoxia)-driven differences in PGE2 metabolism should be targeted for novel treatment of advanced CRC.

Immunohistochemical measurement of endothelial cell apoptosis and proliferation in formalin-fixed, paraffin-embedded human cancer tissue.

There is a need for simple, reproducible methodology for measurement of endothelial cell (EC) apoptosis and proliferation in formalin-fixed, paraffin-embedded (FFPE) human tissue obtained in clinical trials of potential anti-angiogenic agents. Therefore, we developed colorimetric, dual-label immunohistochemical techniques for use on FFPE tissue, based on the use of single-stranded (ss) DNA and Ki-67 as markers of EC apoptosis and proliferation, respectively. Digital image analysis was used to obtain the total tumour microvessel density (MVD), from which the EC apoptosis index (AI) and proliferation index (PI) were derived manually as the number of positive ECs per vessel. Immunohistochemical measurement of EC apoptosis and proliferation was validated on human colorectal cancer liver metastases from a randomized, placebo-controlled trial of the selective cyclooxygenase-2 inhibitor rofecoxib. Proliferating and apoptotic ECs clustered in discrete areas of the tumour vasculature. ECAI (median [inter-quartile range, IQR] 0.0018 [0.0003-0.0064]) and ECPI (median [IQR] 0.0043 [0.002-0.014]) values were low and highly variable in tumours from both placebo- and rofecoxib-treated patients. Our novel dual-label immunohistochemical methods will be generally applicable in FFPE human cancer tissue and should prove invaluable for measuring the anti-angiogenic activity of experimental therapies in clinical trials. The low absolute level of and variability in EC apoptosis and proliferation detectable in human colorectal cancer liver metastases indicates that similar intervention studies using these end-points will need to ensure adequate size in order to be able to detect anti-angiogenic activity by immunohistochemical methods.