Melanoma-restricted genes

Abstract

Human metastatic cutaneous melanoma has gained a well deserved reputation for its immune responsiveness. The reason(s) remain(s) unknown. We attempted previously to characterize several variables that may affect the relationship between tumor and host immune cells but, taken one at the time, none yielded a convincing explanation. With explorative purposes, high-throughput technology was applied here to portray transcriptional characteristics unique to metastatic cutaneous melanoma that may or may not be relevant to its immunogenic potential. Several functional signatures could be identified descriptive of immune or other biological functions. In addition, the transcriptional profile of metastatic melanoma was compared with that of primary renal cell cancers (RCC) identifying several genes co-coordinately expressed by the two tumor types. Since RCC is another immune responsive tumor, commonalities between RCC and melanoma may help untangle the enigma of their potential immune responsiveness. This purely descriptive study provides, therefore, a map for the investigation of metastatic melanoma in future clinical trials and at the same time may invite consideration of novel therapeutic targets.

Figure 1

Eisen's clustering based on 2,044 genes up-regulated in metastatic melanoma lesions compared with all other tumors. Signatures include growth regulation (maroon vertical bar); a signature of genes similarly expressed by melanoma and RCC (blue vertical bar) including a sub-cluster of genes predominantly expressed by RCC (double vertical blue bar and blue arrow); an immunological signature (orange vertical bar); a signature specific for genes predominantly expressed by cutaneous and subcutaneous melanoma metastases (green vertical bar); gene related to blood contamination in fine needle aspirates (FNA; red arrow) and a signature specific for melanoma differentiation antigens (MDA; blue arrow). Genes were identified by a two-tailed Student's t test comparing all melanoma lesions with other tumors (with the exception of RCC) applying as cut off of significance a p₂-value < 0.001. Up-regulation was defined as a positive value after subtracting the average of other tumor samples Cy5/Cy3 ratios from that of melanoma samples.

Figure 2

Eisen's clustering of genes already reported to be preferentially expressed by melanomas. The analysis was performed on 180 cancer samples as described in the Results section and ordered according to Table 1. In particular, renal cell cancer (RCC, orange), melanoma (blue), Epithelial Ovarian Cancer (EOC, yellow), Esophageal Cancer (green), Primary Colorectal Cancer (CRC, dark brown) and lymph nodal metastases of CRC (light brown) are shown. Melanoma samples are further subdivided in cutaneous metastases (CM, light blue) from frozen sections (continuous line) or fine needle aspirates (FNA, dashed line) and lymph nodal metastases (LN, darker blue line). Below is the distance among the various genes based on Eisen's clustering.

Figure 3

Eisen's clustering of immunologically relevant genes selected from clusters e and f-j. To the right the identity of the genes included in cluster e is shown.

Figure 4

Eisen's clustering of genes similarly expressed by RCC and melanoma lesion. To the right the identity of genes most prominently expressed by RCC lesions and cutaneous or subcutaneous melanoma lesions is shown.