Letter to the editor

Abstract

Hambek et al1 demonstrated that serum PGE2 levels were inversely correlated with T classification in patients with squamous cell carcinoma of the head and neck. Elevated serum PGE2 levels were primarily seen in patients with T1 tumors. Although they report a significant difference between serum PGE2 levels in patients with T2/3 and T4 tumors, this seems improbable given the large overlap of the box and whisker plots (Figure 3). They hypothesized that this inverse correlation was due to tumor-derived PGE2 associated with the process of neovascularization during the proliferative phase of tumor growth. In our studies of cultured squamous cell carcinoma cell lines, PGE2 production by tumor cells was not associated with tumor site, stage, or 2-year disease-free survival.2 In studies of tumor-infiltrating mononuclear cells (TIMC), advanced N stage (but not T stage) was associated with decreased PGE2 production by TIMC.3 Although we did not observe an association of tumor tissue levels of PGE2 with tumor stage, increased levels of PGE2 were associated with increased 2-year disease-free survival.4 Collectively, these findings also suggest that increased production of PGE2 is a favorable prognostic sign and may be associated with earlier stage disease. In their article, Hambek et al do not establish the source of the PGE2 in the serum. Our own observations suggest that tumor cells are not the major source of PGE2; rather, it is derived from tumor-infiltrating and circulating mononuclear cells.2-4 An alternative explanation for their observations is that an inflammatory response to tumors in the early stage of tumor growth, perhaps induced by tumor-derived factors, results in increased production of PGE2 by inflammatory cells. As tumors progress, there is suppression of the host inflammatory response resulting in less PGE2 production. The results of cell culture analyses (Figure 1) in their study may not reflect the complex environment of tumor–host interactions in vivo. Subsequent studies in vivo using murine models of squamous cell carcinoma also support the hypothesis that PGE2 is derived from a host inflammatory response to the tumor and may be beneficial. Inhibition of tumor growth in mice treated with a cyclo-oxygenase inhibitor was associated with increased tumor tissue levels of PGE2 and lymphocytic infiltration.5 Peritumoral injections of tumors with PGE2 significantly inhibited tumor growth.6 Hambek et al also observed normalization of serum PGE2 levels in follow-up and suggest that serum PGE2 levels may be useful as a tumor marker. Do they have data comparing patients who developed recurrent disease or a second primary in follow-up? If serum PGE2 is derived from mononuclear cells rather than tumor cells, the utility of serum PGE2 as a tumor marker may be limited depending on the status of the immune system in patients who develop recurrent disease. I suspect that persistent immunosuppression will prevent significant PGE2 production by mononuclear cells and serum PGE2 levels will remain low.