Eosinophil infiltration around tumor nests is a frequent feature of OSC and is accompanied by a mixed lymphocyte response . This infiltration often correlates with deposition of eosinophils cytotoxic proteins with favorable prognosis but the mechanism remains unclear . The de novo secretion of PGD2 by the OSC cell line, SCC-9, the matrigel transmigration experiments, and inhibition with the PGD2 synthase inhibitor, HQL-79, all combine to suggest that PGD2 may be an important mediator in tumor-induced recruitment of eosinophils.
Previous studies have reported PGE2 secretion by OSC [15, 25, 26] but there has been no evidence that this prostaglandin exerts chemoattractive activity on eosinophils; however a closely related prostaglandin, PGD2, is known for its chemotactic activity on eosinophils [16–18]. In this study we report that eosinophils exhibit potent growth-inhibitory activity against the oral cancer cell line, SCC9 which was associated with eosinophil specific EPO release in culture medium. There is no evidence to date that cytotoxic eosinophil granule deposition plays a role in vivo and no data so far exist to support a correlation between granule deposition in OSC and favorable prognosis. In our experiments, we observed that inhibition of OSC growth correlated with detectable cytotoxic granule enzyme EPO activity in culture medium. This association between OSC growth inhibition and eosinophil mediator release was observed regardless of eosinophil viability in the absence of factors that sustain viability (IL-5) thus disrupting eosinophil cell membranes resulting in a non-specific cytolytic release of granular content.
Immunotherapy using IL-2 has been shown to have moderate success against some tumors and is often associated with “unexpected” but significant eosinophilia , which resulted in assumptions suggesting that eosinophils possess anti-tumor activity, at least in vitro. Indeed, IL-2 is recognized as a potent regulator of eosinophil activation, in vitro[29, 30]. The effects of IL-2 include the release of cytotoxic granules, generation of superoxide radicals [31, 32] and production of autocrine IL-2 [29, 30]. IL-2-induced TATE (corollary to treatment of renal cancer), close proximity of activated eosinophils with bladder tumor cells and the subsequent deposition of eosinophil cationic granules were shown to be associated with a favourable outcome . In contrast, the presence of IL-2-induced eosinophilia was considered predictive of the failure of therapy in renal cancer .
Despite the significant in vitro effects we observed with OSC, eosinophils appeared to be mostly recruited around, but not within OSC masses. As well, there was very little evidence of eosinophil granule deposition ex vivo. Regardless, basic proteins from eosinophil granules are extremely cytotoxic, thus, small concentration of free exocytosed granules may be sufficient to exert a potent inflammatory/cytotoxic response against tumor cells . Recent studies from our group suggested that cell-free granules from eosinophils can secrete their content via direct stimulation of functional cytokine and chemokine membrane receptors for present on the granule membrane, in the absence of an intact cell . In addition to these potential cytotoxic effector activities, eosinophils are also capable of exerting an immunoregulatory role in relation to the tumor environment. Eosinophils secrete a wide range of cytokines chemokines and growth factors  and these may further contribute to the biological and immunological role of the eosinophil in OSC.
Finally, the cyclooxygenase-2 (COX-2) inhibitor, NS398, was reported to inhibit OSC proliferation by suppressing PGE2 secretion . However, PGE2 is not a chemoattractant for eosinophils. In contrast, we show that PGD2 is a potent chemoattractant for eosinophils, and may contribute to eosinophilic infiltration via its specific PGD2 receptor on eosinophil, CRTH2, which is also a marker of TH-2 subset of helper T-cells . Whether PGD2 also enhances the potential of eosinophils to kill target cells by inducing exocytosis and subsequent deposition of cytotoxic granule proteins remains unknown and is the subject of a separate study. Thus, our data suggest that eosinophils may contribute to the inflammatory response observed in OSC and may limit tumor progression.