Recent years have seen an explosion in the number of putative biomarkers for most cancers, including melanoma. The identification of novel biomarkers has been aided in large part by the availability of genome-wide approaches to examine the malignant phenotype and the sequencing of the human genome. Numerous prognostic factors for melanoma have been tested and proposed1 that potentially may serve as useful biomarkers. In addition, the development and testing of prognostic markers also may shed light on the mechanisms of tumor progression for a given malignancy. Many (if not most) of these proposed markers have plausible roles in tumor progression by virtue of their potential effects on the major pathways that comprise the “hallmarks of cancer,” as defined a few years ago by Hanahan and Weinberg.2 Thus, most promising prognostic factors for melanoma have either demonstrated or plausible effects on tumor cell proliferation, suppression of apoptosis, cell cycle progression, invasion, or angiogenesis/lymphangiogenesis.

Another pathway that can contribute to tumor progression is escape from immune surveillance. Several molecular mechanisms have been described by which tumors can evade immune responses.3 These include loss of expression of histocompatibility antigens on the tumor cell surface, the production of immunosuppressive cytokines (such as interleukin-10), the production of angiogenic factors that can inhibit the infiltration of immune effector cells into the tumor, the activation of regulatory T cells, and the expression of inhibitors of the immune response, such as cytotoxic T-lymphocyte–associated protein.4

The interaction between the tumor and the immune system has been of particular interest in melanoma for several reasons. First, there were observations of regressing melanomas that were indicative of endogenous antitumor immunity that may be active in a small subset of patients. Second, the finding that the presence of tumor-infiltrating lymphocytes (TILs) conferred an improved prognosis to primary cutaneous melanoma4 further supported the importance of immune surveillance mechanisms in the progression of the disease. Third, the development and approval of immunotherapies for melanoma in its more advanced stages (ie, interferon alpha 2b in the setting of resected, high-risk melanoma and interleukin-2 in the setting of advanced metastatic melanoma) demonstrated that harnessing the immune system can be effective in reversing tumor progression, albeit in a subset of patients and at the cost of significant toxicity.

Consequently, identifying markers of immune responsiveness has been an active area of investigation in melanoma. Genome-wide profiling analyses have attempted to identify differentially expressed genes in tumors from patients undergoing interleukin-2–based immunotherapy.5 Despite these efforts, few (if any) molecular markers that operate in the immune axis reportedly confer a worse prognosis when they are expressed at higher levels in primary melanoma. In this issue of Cancer, Hino et al6 report on the prognostic role of tumor cell expression of programmed cell death-1 ligand 1 (PD-L1) in melanoma. Programmed cell death-1 (PD-1) is a receptor for members of the B7 family of costimulatory molecules and is expressed on lymphocytes. PD-1 acts as a negative regulator of T-cell function that regulates tolerance and autoimmunity. PD-1 has 2 ligands: PD-L1 (also known as B7-H1), which is expressed broadly on hematopoietic cells and aberrantly on certain tumor cells (in which expression is up-regulated by interferon gamma), and PD-L2 (also known as B7-DC), which is expressed on macrophages and dendritic cells. It has been demonstrated that tumor cell-associated expression of PD-L1 promotes the apoptosis of antigen-specific T-cell clones and of tumor-reactive T cells, resulting in enhanced tumor cell growth.7 In addition, PD-L1 expression was identified as a marker of aggressiveness initially in renal cell carcinomas8 and subsequently in other solid tumors. Finally, blockade of the interaction between PD-1 and PD-L1 is being pursued as an immunotherapeutic strategy in several different malignancies, including melanoma.

Hino et al examined the expression of PD-L1 in 59 primary melanomas and in 20 metastatic lesions composed of lymph node and in-transit metastases. Expression of PD-L1 was examined using immunohistochemical staining and was scored using digital imaging analysis, which resulted in the identification of a high-expression group and a low-expression group. High levels of PD-L1 expression were correlated with increased tumor thickness and Clark level but not with ulceration. In univariate analysis, the high-expression group had a significantly worse overall survival compared with the low-expression group in Kaplan-Meier and Cox regression analyses. Multivariate Cox regression analysis indicated that higher levels of PD-L1 were independently predictive of overall survival with the inclusion of several factors, including primary tumor status (low T classification vs high T classification), ulceration, and lymph node metastases.

Although the results certainly are intriguing, significant additional work will be required to advance PD-L1 as a bona fide prognostic factor for melanoma. The relatively small dataset that was amassed and examined in the study was dominated by extremity tumors (78%), acral lentiginous melanomas (66%), and thicker tumors (48% stage IIB or higher) and also included 8 patients with melanoma in situ. Given the distinct biology of acral melanoma, it will be important to examine the role of PD-L1 expression in larger, population-based series of patients with melanoma and in broader subsets of melanoma patients and histologic subtypes. In addition, several aspects of the prognostic analyses warrant attention. Given the prognostic factor analyses performed by the AJCC Melanoma Committee,9 it is important to analyze putative prognostic factors with inclusion of the known factors that are significantly predictive of survival in a large cohort of patients with melanoma. Consequently, it will be important to analyze the expression of PD-L1 in the context of additional factors, such as Clark level (with which it was correlated) and tumor site, and to examine the full range of tumor thickness using either the 4 AJCC scales or as a continuous variable. Furthermore, the digital imaging approach that was used by Hino et al will need to be validated and compared with pathologist scoring. In addition, it would be interesting to examine the potential correlation between PD-L1 expression and TIL levels in primary cutaneous melanoma. Finally, it will be important to analyze T-cell infiltrates in these primary melanomas to determine the level of PD-1 expression in these cells and the extent of regulatory T cells present in the tumor microenvironment.

Although numerous molecular prognostic factors have been described for melanoma, few have been validated in independent datasets with adequate follow-up, and few include complete clinical and histopathologic annotation. Consequently, no biomarkers are used currently to help make patient care decisions. Promising biomarkers include Ki-67, which reportedly has independent prognostic significance in the setting of thin (<1 mm) melanoma.10 Work from our group has identified 3 prognostic markers that were derived from gene expression profiling analyses (nuclear receptor coactivator 3 [NCOA3], secreted phosphoprotein 1 [SPP1], and regulator of G-protein signaling 1 [RGS1]) in a large tissue microarray cohort of primary cutaneous melanoma. More recent studies have analyzed the powerful and independent prognostic efficacy of combined marker analysis and its evaluation in an independent cohort from Germany.11 Thus, it is possible that validated multimarker prognostic assays will emerge for melanoma. In this regard, on further validation, it will be intriguing to assess the potential additional contribution of PD-L1 to these and other prognostic markers in melanoma.

In addition to the examination of PD-L1 tumor expression, Hino et al also present a preliminary analysis of the expression of PD-1 on T cells in the peripheral blood from 10 patients with metastatic melanoma compared with its expression in samples from 5 healthy volunteers. Both CD4-positive and CD8-positive T-cell populations in samples obtained from patients with stage IV melanoma had higher levels of PD-1 than samples from the control population. It will be important to broaden this analysis to patients with all stages of melanoma and to determine whether levels of PD-1 expression on circulating T cells are correlated with disease stage. It also would be interesting to determine the level of PD-L1 tumor cell expression in these patients and ultimately examine whether a correlation exists between PD-L1 expression in tumors and PD-1 expression in circulating T cells. In addition, the authors present a preliminary analysis of 2 patients in whom PD-1 expression on tumor-infiltrating CD8-positive T cells increased along with disease progression. It would be interesting to examine the level of PD-L1 expression in the primary and metastatic tumors from these patients.

Overall, the studies presented are provocative because of their potential clinical implications. These studies, as mentioned above, implicate a marker of immune evasion as a prognostic marker for melanoma, providing an important additional link between immunosuppression and tumor progression. These results have additional practical significance, in that measuring the levels of either PD-L1 in tumor cells, or PD-1 in TILs in situ, or in circulating T cells may emerge as predictive markers of immunotherapy in general or of PD-1 blockade specifically. One of the big obstacles to the expansion of the scope and utility of immunotherapy for melanoma is the dearth of information regarding the subset of melanoma patients or tumors that potentially are amenable and/or responsive to such approaches, especially given the toxicity of many of the currently available therapies. For example, it will be important to examine the level of tumor cell PD-L1 in patients who respond to interleukin-2 treatment. However, the intriguing preliminary findings presented here demonstrate the importance of the PD-1/PD-L1 axis in promoting tumor progression in melanoma and support the further examination of this interaction not just as a therapeutic strategy for melanoma but also in the identification of potential biomarkers of response to immunotherapy.


  1. Top of page

Dr. Kashani-Sabet owns stock in Melanoma Diagnostics, Inc.


  1. Top of page