PGF2α is a prostaglandin that binds to the G-protein coupled FP receptor, and is important in parturition (Khan et al., 2008), and is used as an ocular hypotensive in the management of glaucoma. Prostaglandins (PG) have long been implicated in tumor progression, as studies showing that non-steroidal anti-inflammatory agents, which block cyclooxygenase (COX) activity, decrease colorectal carcinoma in men (Fournier and Gordon, 2000). The causative role of PGE2, the major PG released by keratinocytes in response to UVR in non-melanoma skin cancer, is well documented (Muller-Decker and Furstenberger, 2007). The role of PGF2α in skin cancer is less defined; however, the FP receptor is down-regulated in skin papillomas in mouse carcinogenesis models, and its level of expression is inversely correlated with PGF2α production, suggesting that PGF2α regulates levels of the receptor in squamous epithelium (Muller et al., 2000). Melanoma cells express the synthetic enzymes necessary for PG production, and blockade of COX-2 with blocking antibodies has been reported to reduce melanoma proliferation (Scuderi et al., 2008). Analysis of effects of PGF2α on melanocytes show that PGF2α stimulates the activation of tyrosinase, the rate limiting step in melanin synthesis, and the formation of dendrites, which facilitate the transfer of melanosomes from melanocytes to keratinocytes (Scott et al., 2004, 2005). It is not known if the FP receptor is expressed by melanoma.
To examine FP receptor expression in melanoma, total cellular lysates of normal human melanocytes and six human melanoma cell lines were blotted for FP receptor, and real time polymerase chain reaction (PCR) was performed on RNA extracted from parallel dishes (Figure 1A, B). Melanocytes and the primary melanoma cell line WM115 expressed both FP receptor protein and FP receptor message; none of the other melanoma cell lines expressed detectable FP protein or message. Melanocytes cultured from lightly pigmented foreskins expressed slightly more FP receptor compared with melanocytes cultured from darkly pigmented foreskins, as shown by Western blotting and real time PCR (Figure S1A, B). To determine if the FP receptor is lost in melanoma in vivo, tissue microarrays (TMA) of benign nevi (n = 26), primary melanoma (n = 27) and metastatic melanoma (n = 24) were stained with antibodies against the FP receptor (Figure 1C). Examination of stained sections of normal skin revealed expression of FP receptor in melanocytes in the epidermis, in endothelial cells of blood vessels, and in smooth muscle of vessels and pilierector muscles. Examination of nevi showed focal positive staining of nevic cells for FP receptor in only 1/26 nevi, whereas none of the primary or metastatic melanomas expressed FP receptor. Type IV cytosolic phospholipase A2 (cPLA2) is the rate limiting step in PG synthesis. Levels of active (phosphorylated) cPLA2 were determined in melanocytes and human melanoma cell lines by Western blotting (Figure 2A). Steady state levels of active cPLA2 were higher in all melanoma cell lines, compared with normal melanocyte. To determine if melanoma cells synthesize PGF2α, culture supernatants of melanoma cell lines were analyzed for PGF2α by ELISA assay (Figure 2B). Five of six melanoma cell lines synthesize PGF2α at levels approximately 2–4-fold higher than normal human melanocytes, which produced 18 pg/ml/cell × 10−6 PGF2α (Scott et al., 2005). The WM115 melanoma cell line synthesized the most PGF2α (105 pg/ml/cell × 10−6) with the other cell lines ranging from 38 to 62 pg/ml/cell (×10−6).
In an effort to identify additional functions for PGF2α, we determined the effect of PGF2α on melanocyte and melanoma migration, and on melanocyte apoptosis in response to ultraviolet irradiation (UVR). PGF2α did not affect melanocyte migration, even at high doses (Figure S1, C). Similar results were obtained when fluprostenol, a specific FP receptor agonist, was tested (data not shown). PGF2α also had no effect on migration of WM115 cells, which express the FP receptor (Figure S1, C). To determine if PGF2α protects melanocytes from apoptosis in response to UVR, melanocytes were pretreated with PGF2α for 24 h, and then irradiated with 80 mJ/cm2 UVB, and levels of active caspase-3 were measured by fluorescein activated cell sorter (FACS) analysis 24 h later. Because white and black melanocytes express different levels of the FP receptor, pooled cultures of melanocytes isolated from black (n = 4) and white (n = 3) foreskins were tested separately (Table S1). Forty-seven percent (±10% SD) of white melanocytes expressed active caspase-3 following irradiation, compared with 51% (±15% SD) irradiated in the presence of PGF2α (10 nM), a difference that did not reach statistical significance. As expected, melanocytes cultured from darkly pigmented foreskins were more resistant to apoptosis in response to UVR, with only 25% (±12% SD) expressing active caspase-3 following UVR. The inclusion of PGF2α decreased active caspase-3 expression in black melanocytes response to UVR (20%±10% SD); however, this did not achieve statistical significance by Student’s t-test (P = 0.067). Active caspase-3 expression in response to PGF2α did not differ from vehicle treated controls. It is interesting to note that melanocytes cultured from light and darkly pigmented foreskins showed opposing effects of PGF2α following UVR, with an increase in caspase 3 expression seen in white melanocytes (more apoptosis) and a decrease seen in black melanocytes (less apoptosis). While differences between treated and untreated groups did not reach statistical significance, it is interesting to speculate that PGF2α may provide a small photoprotective effect for black melanocytes, and may contribute to UVR-dependent apoptosis in white melanocytes.
Our data indicate that the FP receptor is expressed only by melanocytes, but is lost in melanoma and in nevi. A search of the Oncomine database (https://www.oncomine.org) was performed and virtually all melanoma cell lines reported (n = 501) showed loss of the FP receptor (PTGFR) at the message level, consistent with our results on melanoma in vitro and in vivo. We were unable to identify a role for PGF2α in melanocyte or melanoma migration, or in protection of melanocytes from UVR-induced apoptosis. The significance of loss of FP receptor expression by benign nevi and melanoma remains to be determined. Because nevic cells are generally considered to be modified melanocytes, we were somewhat surprised that nevic cells show essentially total loss of FP receptor expression. While the true nature of nevic cells is still a matter of debate, one generally accepted theory of their ontogenesis is that nevic cells represent a clonal proliferation of melanocytes that acquire the capability to form cohesive nests and traverse the basement membrane. Therefore, loss of FP receptor expression by melanocytes may facilitate, cell–cell adhesion, or invasion through basement membranes. Testing of this notion will require silencing of FP receptor expression in melanocytes followed by analysis for acquisition of a ‘nevic cell’ phenotype. The lack of FP receptor indicates that melanoma will not respond to PGF2α, indeed, the one cell line (WW115) that expressed FP receptor failed to show changes in migration or proliferation (data not shown) in response to PGF2α. Our data do indicate that melanomas synthesize abundant amounts of PGF2α, greater than normal melanocytes (Scott et al., 2005), suggesting that PGF2α may facilitate melanoma progression through effects on accessory cells, such as fibroblasts and endothelial cells, which express the FP receptor.