Work

Melanoma from bench to bedside: meeting report from the 6th international melanoma congress

Year: 2009

Type: article

Abstract: The 2009 International Melanoma Congress sponsored by the Society for Melanoma Research brought together a large collection of laboratory researchers, students, and clinicians. Presentations were made in numerous areas spanning biological analyses to clinical management of the disease. The meeting was characterized by numerous presentations of provocative, unpublished data, including multiple aspects which impact greatly on our understanding and management of melanoma. This meeting summary has been compiled by the authors in an effort to convey some of the newly presented data despite the fact that a comprehensive treatment of the meeting contents is not feasible. Therefore, the authors apologize to the many speakers and poster presenters whose interesting and important discoveries could not be included here. The Second Melanoma Pathology Symposium was organized by Raymond Barnhill (Hospital Saint-Louis and University of Paris VII, Paris, France) and convened as a satellite session just prior to the beginning of the Congress. It covered a diverse spectrum of topics pertinent to the biology and histopathology of melanoma during this 1-day program. These subjects in the Symposium included: Regression in melanoma: what significance does it have in 2009 – C. Shea, Melanocytic lesions of the scalp region – A. Spatz, Sentinel lymph node biopsy and minimal tumor burden – M. Cook, Histological evaluation of melanonchia S. Binder, Spitzoid lesions and sentinel lymph node biopsy – K. Busam, Angiotropic melanoma and extravascular migratory metastasis: An update – R. Barnhill, Reproducibility of histologic diagnosis of atypical nevi and melanoma with a 20-yr interval- L. Duncan, Blue nevi and pigmented epitheliod melanocytomas – A. Zembowicz, A study of acral melanoma in Brazil – G. Landman, The pigmented macular lesions in chronically sun-exposed skin – S. McCarthy, Melanoma and the pre-metastatic niche – M. Mihm, Melanocytic lesions: discordance in diagnosis – M. Piepkorn, Immunohistochemistry and melanocytic lesions – V. Prieto, A critique of ancillary diagnostic techniques in melanoma – D. Elder, The reporting of melanoma and melanoma prognostic factors in 2009 – R. Scolyer, and A multi-observer study in to validate genotype–phenotype correlations to improve melanoma classification – B. Bastian. As a biological phenomenon in melanoma, regression is of considerable importance but nonetheless continues to be very poorly understood. A major problem in the pathology of regression is the ongoing failure to apply standardized and reproducible criteria for its histopathological assessment. Thus, to a significant degree the conflicting data associated with regression and outcome in melanoma can be ascribed to the difficulty and lack of consensus in how regression is evaluated histologically. From this review, it was also concluded that at present there are no objective data that the presence of regression in a primary melanoma is a definite indication for sentinel lymph node biopsy. The scalp constitutes an important site for particular melanocytic lesions such as small cell melanoma in children, both arising de novo and in congenital nevi, blue nevus-like melanomas (malignant blue nevi), and atypical melanocytic nevi that are frequently misclassified as melanoma. There is ongoing debate about whether a minimal tumor burden may exist in metastatic melanoma, as for example, micrometastases measuring <0.1 or 0.2 mm, and in particular, in sentinel lymph nodes (SLN), which has predictive value for patient management and outcome. On the other hand, there is also evidence that documented metastases of any size may potentially correlate with melanoma progression. A review of the application of SLN biopsy for spitzoid melanocytic lesions has demonstrated SLN deposits in about 50% of patients undergoing this procedure. Interestingly, although tentative, virtually all of the latter patients have failed to show any evidence of disease progression with follow-up. There has been controversy surrounding the nomenclature of pigmented epithelioid melanocytic neoplasms described as epithelioid blue nevi in the context of Carney complex and the subsequent entity pigmented epithelioid melanocytoma (PEM). However, recent documentation of the mutations in the protein kinase A regulatory subunit 1alpha gene and loss of protein R1alpha expression in the latter lesions provides support for PEM as a distinct entity and a close relationship with epithelioid blue nevi. Angiotropic melanoma and extravascular migratory metastasis (EVMM) has been the subject of serious research for more than a decade. Accumulating evidence indicates that angiotropism in melanoma is a prognostic factor, and that extravascular migratory metastasis (EVMM) is an important mechanism of melanoma spread. Recent molecular analysis and gene-expression profiling are beginning to yield new information about the molecular basis of EVMM. Although not studied to any significant degree in melanoma thus far, the potential creation of a 'premetastatic niche' by the primary tumor has relevance to melanoma. Pigmented macular lesions involving chronically sun-exposed skin constitute a challenging spectrum of entities ranging from pigmented actinic keratosis to in situ or thin invasive melanoma. Because of the striking heterogeneity of many lesions, great attention to adequate sampling and careful assessment of margins are needed. The rates of discordance in the pathological diagnosis of atypical melanocytic lesions are well-established in the literature. However, the basis for this lack of inter-observer agreement has not yet been subjected to rigorous study. One explanation offered for the continued striking increase in incidence of melanoma has been that of 'diagnostic drift', or in more precise terms, a diminished threshold for the diagnosis of melanoma by pathologists. A recent study examining intra-observer rates of agreement for interpreting a series of severely atypical melanocytic lesions and thin melanomas with a 20-yr interval has suggested a trend to reduced thresholds for melanoma diagnosis. This outcome was also supported by comparable rates of early melanoma diagnosis among young pathologists evaluating the same study set. Immunohistochemistry continues to be an important adjunctive technique for confirming melanocytic differentiation and to aid in the distinction of benign lesions from melanoma. In particular, Ki-67, which measures proliferation rate, and markers such as HMB45(gp100) are among the most useful, particularly when assessing topographical expression. For example, benign nevi show reduced expression of the two markers usually in absolute terms but especially with increasing depth, whereas melanomas usually express the two markers throughout. A critical review of the use of biomarkers in melanoma has revealed that Ki-67 is among the very few biomarkers that have been subjected to fairly rigorous investigation. There is accumulating evidence that the routine use of Ki-67 may potentially provide important prognostic information for clinical management of patients. Among the routine prognostic indices for patients with primary localized melanoma, mitotic rate per mm2 has recently been introduced as a major prognostic factor in the new melanoma clinical staging guidelines of the American Joint Committee on Cancer. Finally, with the ultimate goal of obtaining fairly pure or well-characterized melanoma patient populations, a recent study has investigated the reproducibility (inter-observer agreement rates) for particular histomorphological attributes known to correlate with particular genetic alterations, such as BRAF mutations. While the latter studies are encouraging with respect to good inter-observer agreement for many histomorphological parameters, the critical oncogenic alterations equivalent to BRAF are currently not known for a significant proportion of melanomas impairing the utility of genotype–phenotype correlation to improve classification. The Stem Cell Satellite symposium commenced with short talk by a leader in the field of reprogramming, Rudolf Jaenisch (Whitehead Institute, MIT), who discussed the barriers and technical challenges of iPS reprogramming. It appears that one of the major barriers to reprogramming is the cellular p53 response pathway. Jaenisch showed that blockade of this pathway, via downregulation of p53, increased the proliferation rate of cells, which in turn dramatically increased the efficiency of reprogramming. Alternatively, overexpression of NANOG could also boost reprogramming efficiency. Jaenisch proposed two different models for acquisition of pluripotency. The deterministic model, which posits that only elite cells (i.e., somatic stem cells) have the capacity to regain pluripotency, versus the stochastic model, which posits that all cells have the potential to generate iPS. Studies from his laboratory favor the stochastic model, indicating that any cell can be reprogrammed to iPS. The session shifted gears from reprogramming to melanoma and melanoma stem cells with the next speaker, Marcus Bosenberg (Yale Medical School). Mouse melanomas harvested from Tyr::CreER; Braf CA; Ptenlox/lox mice, in which melanomas spontaneously arise, were analyzed with various stem cell associated markers. Of the many markers analyzed, only CD34, a hematopoietic marker, and p75, a neural crest marker, consistently defined three subpopulations of melanoma cells harboring distinct tumorigenic properties; a CD34+ subpopulation, which had high colony forming efficiencies and high rates of tumor formation but only gave rise to CD34+ cells; a p75+ subpopulation which had low colony forming efficiencies and low rates of tumor induction that were highly sensitive to temozolamide and cisplatin, and a CD34−/p75− subpopulation that gave rise to tumors with a heterogeneous population of CD34 and p75 positive and negative cells. These studies concluded that more than one distinct population of melanoma propagating cells could be isolated from the same tumor and that tumor heterogeneity does not have to be re-established for a cellular subset to form tumors. While Bosenberg's mouse model of melanoma development avoids the artificial setting in which human melanomas are typically studied, namely injection of human melanoma cells embedded in Matrigel into immunodeficient mice, the importance of the appropriate mouse models for analyzing the potential of human melanoma were brought forth in the following two talks. Markus Frank (Brigham and Women's Hospital, Harvard Medical School), who has identified ABCB5 as a marker of melanoma initiating cells, found that ABCB5+ cells downregulate MHC class I molecules and preferentially express the T cell immunosuppressive cytokine IL-10, suggesting that ABCB5+ cells exert immunodulatory functions and have immune evasion mechanisms that are relevant to disease progression. Additionally, ABCB5+ cells produce extracellular matrix factors, particularly laminin that influence the tumor microenvironment and correlates with tumor progression. Based on these results, Frank pointed out that use of Matrigel, composed of a milieu of extracellular matrix molecules, in subcutaneous injection of human melanoma cells provides an artificial advantage to these cells which may not normally produce these factors but that rely on them for tumor propagation. Moreover, the argument was brought forth that the use of NOD/SCID Il2rg−/− mice, which are used by many to assess the tumorigenic capacity of cancer stem cells, provides an advantage to melanoma cells that would otherwise be targeted by the immune system. Sean Morrison (University of Michigan) agreed that while a greater number of melanoma cells were required to initiate tumors in the NOD/SCID than NOD/SCID Il2rg−/− mice, the former model failed to determine the actual tumorigenic potential of melanoma cells. Indeed, despite their extensive search of over 50 potential melanoma stem cell markers, no single marker or combination of markers, including ABCB5, could distinguish tumorigenic from non-tumorigenic cells in NOD/SCID Il2rg−/− mice. All cells gave rise to tumors, even at the single-unsorted cell level. Morrison's findings suggest that all melanoma cells have the potential to generate tumors given the appropriate time and environment. Morrison reasoned that while some cancers, including leukemia, fit the hierarchy of the cancer stem cell model, some cancers, melanoma being the prime example, fail to fit this model. The Congress officially opened with a Keynote Address by David Baltimore (California Institute of Technology). Dr. Baltimore, Nobel Laureate for the discovery of reverse transcriptase, provided a detailed review of his laboratory's efforts to deploy a strategy of 'immunologic engineering' against melanoma. He described three mechanistic opportunities: modulation of dendritic cells, engineering of T cell receptor-modified autologous T cells, and breaking of immunologic tolerance. Of these, his own laboratory efforts have focused on the first two. Baltimore described the design and testing of a novel lentiviral dendritic cell vector which modulates antigen stimulation in a B16 mouse melanoma model by two orders of magnitude. He also described use of recombinant T cell receptor vectors that selectively target tumor associated T cells, to be used in combination with the dendritic cell vector. A clinical trial in man is currently underway (in collaboration with Baltimore's clinical colleagues at UCLA), and significant responses were described to have occurred in all five patients treated to date. A particularly striking observation was made with respect to the occurrence of 'mixed responses' seen in several patients. In one example many metastatic lesions were seen to respond in a patient, whereas one specific lesion had simultaneously progressed despite the immunologic adoptive transfer therapy. The T cell receptor employed in the trial is derived from the laboratory of Dr. Steven Rosenberg (NCI) and is directed against the melanocytic antigen MART1. Biopsy of the isolated 'progressing' lesion revealed that this metastasis was essentially devoid of MART1 expression. While loss of antigen expression may thus represent a significant challenge to the success of this approach, the selection for loss of MART1 expression appears to represent a striking 'proof-of-principal' supporting the therapeutic strategy's mechanism of action and in vivo efficacy. The recent improved understanding of the regulatory pathways of the immune system provides great hope for significant clinical impact. Glenn Dranoff (Dana Farber Cancer Institute) elucidated the important role for granulocyte-macrophage colony stimulating factor (GM-CSF). Through study of GM-CSF deficient mice, GM-CSF was seen to be required for expression of the phosphatidylserine binding protein milk fat globule epidermal growth factor-8 (MFG-E8) in antigen presenting cells. Interestingly, MFG-E8 within the tumor microenvironment is secreted and contributes to disease progression via αvβ3 integrin signaling. The uptake of apoptotic cells by phagocyte-derived MFG-E8 stimulates Treg through TGFβ, MHC Class II, and CCl22. MFG-E8 limits the efficacy of GM-CSF in vaccination with GM-CSF secreting tumor cells (GVAX) through upregulation of Treg. The use of a dominant negative mutant (RGE) potentiates therapeutic immunity resulting in decreased numbers of Treg. This has specific implications for melanoma, as MFG-E8 expression is linked with progression of the vertical growth phase in primary tumors. Anti MFG-E8 antibodies plus chemotherapy were found to enhance DC activation in tumors by improvement in antigen cross presentation. This proposes blockade of MFG-E8 as a means to combinatorial strategies for melanoma immune therapy. Continuing the discussion with GM-CSF, Craig Slingluff (University of Virginia) reported on their clinical experience involving vaccination with peptides to tyrosinase, MART-1, gp100, and MAGE family members in the presence of IFA and GM-CSF. The studies included comparisons of 12 peptides to four peptides with local GM-CSF. Monitoring of peripheral blood revealed decreases in peptide specific immune responses in the presence of GM-CSF. This raised the question of the significance of immune monitoring in the peripheral blood versus immune activity occurring at the sites of disease. Biopsies of pre-existing sites of disease following vaccination revealed potent perivascular immune infiltrates with accumulation of both T and B cells. Eosinophils were also present deep within the tumor deposits. At vaccination sites, there was also consistent inflammation with the presence of CD83 dendritic cells. These studies provide further evidence for use of adjuvants in the combination approaches to immune therapy for melanoma and the role of GM-CSF. A particularly exciting translational clinical investigation currently underway involves B7-H1/PD-1 interactions. B7H1 expression in human cancers is associated with poor prognosis particularly in renal cell cancer and ovarian cancer. Suzanne Topalian (Johns Hopkins School of Medicine) reported early clinical experience with an antibody that blocks PD-1, demonstrating clinical activity in a number of cancers including melanoma. Of particular interest is the detection of persistent PD-1 receptor occupancy following treatment with the antibody. Testing of PD-1 blockade continues. Adoptive T cell therapy involving the ex vivo expansion of generated antigen-specific T cells offers another promising avenue of investigation for immune therapy. As persistence of transferred cells remains a primary concern, Cassian Yee (Fred Hutchinson Cancer Research Center) described the use of high dose cyclophosphamide as a conditioning regimen for adoptively transferred CD8+ cells against melanosomal differentiation antigens, followed by low dose IL-2. In vivo persistence of cells for up to 10 months was observed, as well as clinical responses including complete response in previously refractory patients. In a separate study, adoptively transferred CD4+ cells also revealed clinical responses. NY-ESO-1 specific CTL infusions induced responses that interestingly also induced immunity to MART-1 and MAGE antigens. The cytokine and chemokine profile significantly influences the helper/trafficking functions of cells. Meenhard Herlyn (Wistar Institute, University of Pennsylvania) presented data that chromatin regulator JARID1B plays a critical role in long-term maintenance of melanoma. Abrogation of JARID1B leads to exhaustion melanoma cells in vitro and in vivo. Only a small subpopulation (∼5%) of melanoma cell expressed JARID1B, and these cells were characterized as slowly proliferating cells. As opposed to the cancer stem cell theory, JARID1B did not follow a hierarchical model since JARID1B negative melanoma cells turned into positively expressing (JARID1B) cells, suggesting that environmental cues initiate JARID1B expression. Based on the data, it was proposed that melanoma maintenance is a dynamic process and that the slowly proliferating JARID1B-positive subpopulation is essential for continuous tumor growth. Mary J.C. Hendrix (Northwestern University) discussed influence of the human embryonic stem cell (hESC) microenvironment on the reprogramming of metastatic phenotype of aggressive melanoma. hESCs express a morphogen Nodal and its natural inhibitor Lefty. Aggressive melanoma cells highly express Nodal, while is to of Nodal by matrix or antibodies metastatic melanoma cells a more invasive non-tumorigenic Hendrix also showed that of Nodal expression in melanoma. data suggested that the and Nodal contributes melanoma cell S. presented a model in which melanoma metastases following the of primary tumors. This model was human melanoma and the selection of aggressive melanoma cell which can to and harboring metastases were treated with was but metastases Cell derived from these metastases can to the in tumor This model has the of novel therapeutic which can be via small to proliferation of metastatic cells. Sean Morrison (University of Michigan) discussed whether melanoma a stochastic model or a cancer stem cell model. cancers including some cell cancers and some are organized and only small subpopulations of cells can transfer disease into mice. Only in a melanoma cells form tumors in NOD/SCID mice, one in four can form tumors in a more mouse model when with Matrigel, suggesting tumorigenic potential is a of melanoma cells. (Johns Hopkins School of Medicine) discussed cell metastatic melanoma cells with cells showed specific of cell a approach, identified a of the family of as a gene that is in metastatic melanoma cells during with cells. is in cell during development and is in and through its in signaling. cell thus an important role in cell to Dr. University) the session by about the of specific in studies have been focused on induction of melanoma in the growth factor have induction of melanoma by but no significant role for in in the In this reported his efforts to these studies by examining the of in pigmented mice. showed that of on a genetic with melanoma at the same rate in pigmented and However, isolated was only to produce melanoma in pigmented mice. showed that were generated in both pigmented and with while only in pigmented on his mouse model of human melanoma, that both and can melanoma but by two different In the next Cancer Institute) continued on this by his to in vivo mechanisms by which melanoma. described a mouse that the and of from the mouse following He used this model to the in vivo molecular responses of to and both and In with known activity in his model, found that induced a while to was reported that a highly response gene subset was from associated with a of This was in turn associated with a activation of melanocytic cells, characterized by proliferation and the showed that this activation was by He also presented in vivo data indicating that can enhance in vivo melanoma growth by suggested that to could be in by the action of these which could also potentially David followed with a of his on of the mechanisms that responses of the skin to involves both cell and respect to of the regulatory which plays a role in expression of described the discovery of in melanoma, as well as responses in the skin associated with and Dr. his discovery that to in the of in mice, but that can be by application of a induced against suggesting a clinical strategy for small of also expression by of which can be to required for showed that induction is and p53, which as a for can also rise to which may to human Institute of Medical discussed his to low melanoma associated with skin and described the of studies single to provide assessment of the melanoma and the human As a in described the of melanoma from known to be associated with to the of novel The at including the protein low melanoma have been found to their through melanocytic nevi This is striking this is the most melanoma Melanoma studies are being used to correlate to and (University of about discoveries the role of in and reported that is to of and the of through showed that the is required for of essential for maintenance of at the sites of the same on as the for Interestingly, was reduced by and by data suggested that can by and provided a explanation as to the role of in melanoma Medical discussed that in the in a it is not an as an critical for the development and of melanoma cells, of activity would be highly Cell have that of by melanoma cells to the of a lentiviral for of activity in cells, the a The (in collaboration with David found that of and to upregulation of suggesting in these cells, by decreases the of the Interestingly, a in cells identified but not or indicating that the of a may cell focused their studies on the that to downregulation of to their potential therapeutic of and showed that the identified exert their by the of of these diminished the several melanoma suggesting of the for and the cellular its for and The University of discussed to melanoma development and

Source: Pigment Cell & Melanoma Research

DOI: 10.1111/j.1755-148x.2009.00655.x