The Response of microRNAs to Solar UVR in Skin-Resident Melanocytes Differs between Melanoma Patients and Healthy Persons

Abstract

The conversion of melanocytes into cutaneous melanoma is largely dictated by the effects of solar ultraviolet radiation (UVR). Yet to be described, however, is exactly how these cells are affected by intense solar UVR while residing in their natural microenvironment, and whether their response differs in persons with a history of melanoma when compared to that of healthy individuals. By using laser capture microdissection (LCM) to isolate a pure population of melanocytes from a small area of skin that had been intermittingly exposed or un-exposed to physiological doses of solar UVR, we can now report for the first time that the majority of UV-responsive microRNAs (miRNAs) in the melanocytes of a group of women with a history of melanoma are down-regulated when compared to those in the melanocytes of healthy controls. Among the miRNAs that were commonly and significantly down-regulated in each of these women were miR-193b (P<0.003), miR-342-3p (P<0.003), miR186 (P<0.007), miR-130a (P<0.007), and miR-146a (P<0.007). To identify genes potentially released from inhibition by these repressed UV-miRNAs, we analyzed databases (e.g., DIANA-TarBase) containing experimentally validated microRNA-gene interactions. In the end, this enabled us to construct UV-miRNA-gene regulatory networks consisting of individual genes with a probable gain-of-function being intersected not by one, but by several down-regulated UV-miRNAs. Most striking, however, was that these networks typified well-known regulatory modules involved in controlling the epithelial-to-mesenchymal transition and processes associated with the regulation of immune-evasion. We speculate that these pathways become activated by UVR resulting in miRNA down regulation only in melanocytes susceptible to melanoma, and that these changes could be partially responsible for empowering these cells toward tumor progression.

Fig 1. Capture of melanocytes from the epidermis of human skin after exposure to ssUVR.

(a) UV-irradiation scheme. On day one and two, the posterior shoulder of each volunteer received an ssUVR dose (6 mm circle) of 4X their baseline MED (see Methods). On day three, a punch biopsy removed the irradiated site, as well as an adjacent un-irradiated site. (b) Immunohistochemical staining of human skin. Antibody to MART-1 (Melan-A) was used for identifying melanocytes present in the epidermis of skin. Brownish red cells indicate the melanocytes. (c) Stained slides before and after LCM. A pure population of melanocytes was captured by using the Arcturus XTTM LCM system.

Fig 2. MicroRNAs differentially expressed in melanocytes of skin after exposure to ssUVR.

(a and b) A color-coded heat map representations of UV-miRNAs in melanocytes commonly expressed in melanoma patients or healthy individuals. The map in (a) ranks each miRNA by their fold-change (RQ), beginning with those having the most negative RQ value (top) to the highest negative RQ value (bottom), whereas the map in (b) ranks each from the most negative to the most positive RQ value. The color bars at the top of (a) and (b) represent the normalized intensity of down-regulated or up-regulated UV-miRNAs with the shaded blue corresponding to a decrease in expression and that of shaded red to an increase in expression. The left margin shows the representative miRNAs and the bottom of the columns show each individual sample. MicroRNAs in the heat map of (a) and which match those previously identified in melanoma tissue are marked by a star. (c) An identical set of UV-miRNAs in the melanocytes of patients or healthy persons is effected differently by ssUVR. The y-axis indicates relative gene expression, and the whisker bars indicate the standard error of the mean. MicroRNAs that match those previously identified in melanoma tissue are marked by a star.

Fig 3. Verification of the expression levels of UV-miRNAs by qRT-PCR.

(a) The human microRNA array cards were validated by performing qRT-PCR analysis on the same pre-amplified RNA (see Methods) that was used on these cards. Pre-amplified RNAs corresponding to the irradiated melanocytes of healthy persons C and G (HC and HG) or melanoma patients C, D and H (PC, PD, and PH) were chosen at random for monitoring the expression of relevant UV-miRNAs. In all cases, the direction of the fold change is the same for both qRT-PCR and Microarray. Fishers exact test on values dichotomized as up/down regulated yields a p-value = 1.29E-05. (b) The expression levels of twelve UV-miRNAs were further validated by conducting qRT-PCR analysis on RNA without pre-amplification and extracted from irradiated melanocytes newly derived by LCM from the samples of healthy persons A and C (HA and HC) or melanoma patients C, D, and H (PC, PD, and PH). Fishers exact test in this panel yields a p-value = 1.323E-05.

Fig 4. Four interlocking regulatory networks showing statistical relationships between a subset of repressed UV-miRNAs and EMT signature genes.

Highlighted in networks a and b is a hypothetical scheme illustrating the modulation of EMT signaling by TGF-β and/or WNT3a/5a, and the shift from E-cadherin to N-cadherin expression (“cadherin switching”). These events are fundamental to the EMT. A potential gain-of-function in genes within networks c and d are respectfully associated with cells acquiring stem-cell like properties and resistance to apoptosis, additional hallmarks of the EMT. FZD* represents three of the WNT receptor genes (Frizzled), FZD2, FZD5, and FZD7. The blue colored notches and circles represent the down-regulated UV-miRNAs listed in Fig 2a and 2c, respectively, while the yellow colored circles signify the potentially up-regulated genes.

Fig 5. A regulatory network showing statistical relationships between 14 repressed UV-miRNAs and 13 immuno-evasive signature genes.

Superimposed on the network is a hypothetical mechanism resulting from a potential gain-of-function in CCL2 and CCL8, both of which are known to attract M2-type macrophages (M2-M), and in turn the release of INF-γ, a cytokine reportedly involved in up-regulating critical immuno-evasive genes such as PD-L1, PD-L2, and B7-H2. The blue colored notches and circles represent the down-regulated UV-miRNAs listed in Fig 2a and 2c, respectively, while the yellow colored circles signify the potentially up-regulated genes.