Transcriptional profiling of human uveal melanoma from cell lines to intraocular tumors to metastasis

Liquid Biopsy for Solid Ophthalmic Malignancies: An Updated Review and Perspectives

Simple Summary

To date, there is no treatment for metastatic uveal melanoma. Identifying its metastatic spread is essential. Liquid biopsy can identify patients at risk of metastatic spread early. Here, we aim to summarize the current knowledge of liquid biopsy in ophthalmic malignant tumors, including uveal melanoma. Our objective is to establish the current state of liquid biopsy in the ophthalmic field, as well as its perspectives and limitations.

Abstract

Tissue biopsy is considered the gold standard when establishing a diagnosis of cancer. However, tissue biopsies of intraocular ophthalmic malignancies are hard to collect and are thought to be associated with a non-negligible risk of extraocular dissemination. Recently, the liquid biopsy (LB) has emerged as a viable, non-invasive, repeatable, and promising way of obtaining a diagnosis, prognosis, and theragnosis of patients with solid tumors. LB refers to blood, as well as any human liquid. The natural history of uveal melanoma (UM) and retinoblastoma (RB) are radically opposed. On the one hand, UM is known to disseminate through the bloodstream, and is, therefore, more accessible to systemic venous liquid biopsy. On the other hand, RB rarely disseminates hematogenous, and is, therefore, more accessible to local liquid biopsy by performing an anterior chamber puncture. In this review, we summarize the current knowledge concerning LB in UM, RB, conjunctival tumors, and choroidal metastases. We also develop the current limitations encountered, as well as the perspectives.

Novel therapeutic strategies and targets in advanced uveal melanoma

PURPOSE OF REVIEW

Currently, there are no U.S. Food and Drug Administration-approved or effective treatment options for advanced-stage uveal melanoma. In this article, we focus on therapeutic targets in pathways/mechanisms associated with common mutations in uveal melanoma. We review the challenges associated with targeting of these pathways and novel treatment strategies.

RECENT FINDINGS

Common mutations that promote uveal melanoma initiation and progression include alterations in G protein subunit alpha q/11 (GNAQ/GNA11) and breast cancer gene 1-associated protein 1 (BAP1). Mutant GNAQ/GNA11 induces constitutive activation of tumorigenic pathways such as extracellular signal-regulated kinase (ERK)1/2 and yes-associated protein. Inhibition of mitogen-activated protein kinase kinase (MEK) downstream of ERK1/2, however, was shown in trials to have limited clinical benefit. Recent reports suggested that combination therapies of MEK inhibition and modulators of mechanisms of drug resistance may improve tumor responses to MEK inhibitors. BAP1 has been shown to be involved in modulating chromatin dynamics and deubiquitination of proteins. Hence, epigenetic inhibitors are being investigated in BAP1 mutant uveal melanoma. However, other functions of BAP1, such as in DNA damage repair and cell cycle regulation, indicate additional targets for treatment of BAP1 mutant uveal melanoma. In addition, the frequent delayed development of uveal melanoma macrometastases is likely due to cellular dormancy mechanisms. Nuclear receptor subfamily 2, group F, member 1 and transforming growth factor beta 2 were among factors that have been shown in other cancers to induce dormant phenotypes.

SUMMARY

Findings from studies in uveal melanoma and in other cancers provide evidence for potential strategies that may be tested preclinically and clinically in advanced-stage uveal melanoma to improve treatment outcome and overall survival of patients.

The Results of Stricter Inclusion Criteria in an Immunomagnetic Detection Study of Micrometastatic Cells in Bone Marrow of Uveal Melanoma Patients - Relevance for Dormancy

Approximately 50% of uveal melanoma patients develop metastases. We want to evaluate the effect of stricter criteria on our data from our previous study correlating survival and bone marrow (BM) micrometastasis results using our immunomagnetic separation (IMS) method. Mononuclear cell fractions (MNC) isolated from BM were examined for tumour cells and the patients were classified as BM positive (BM+) or BM negative (BM-). The study originally included 328 consecutive patients with uveal melanoma from 1997 to 2006. The cohort was limited to 217 patients when we introduced cyto- or histopathological verification of melanoma cells in the patient as a main new criterion for inclusion. Tumour cells were found in BM-samples in 38.7% (95% CI, 32-45) at enrolment. Until the latest work-up 43.8% (95% CI, 38-50) of patients had developed melanoma metastases. After a minimum follow-up time of 8.5 years, 60.4% (95% CI, 54-66) of patients had died. The causes were: melanoma metastases 69.5%, another type of cancer 5.4% and non-cancerous causes 19.5%. Overall median survival was shorter for the BM- patients (11.3 years) (95% CI, 10-12) compared to the BM+ (16.5 years) (95% CI, 12-14), p = 0.04, log rank test. All-cause mortality and specific melanoma mortality estimated after 12 year follow-up showed a highly significant difference comparing BM- and BM+, p = 0.010 and p = 0,017, respectively. IMS yields a high fraction of BM+ samples due to micrometastasis at diagnosis and these cells appear to have a positive prognostic impact strengthening our previous report. The late recurrences support the concept of tumour dormancy.

Genomic Gain of 16p13.3 in Prostate Cancer Predicts Poor Clinical Outcome after Surgical Intervention

Identifying tumors with high metastatic potential is key to improving the clinical management of prostate cancer. Recently, we characterized a chromosome 16p13.3 gain frequently observed in prostate cancer metastases and now demonstrate the prognostic value of this genomic alteration in surgically treated prostate cancer. Dual-color FISH was used to detect 16p13.3 gain on a human tissue microarray representing 304 primary radical prostatectomy (RP) cases with clinical follow-up data. The results were validated in an external dataset. The 16p13.3 gain was detected in 42% (113/267) of the specimens scorable by FISH and was significantly associated with clinicopathologic features of aggressive prostate cancer, including high preoperative PSA (P = 0.03) levels, high Gleason score (GS, P < 0.0001), advanced pathologic tumor stage (P < 0.0001), and positive surgical margins (P = 0.009). The 16p13.3 gain predicted biochemical recurrence (BCR) in the overall cohort (log-rank P = 0.0005), and in subsets of patients with PSA ≤10 or GS ≤7 (log-rank P = 0.02 and P = 0.006, respectively). Moreover, combining the 16p13.3 gain status with standard prognostic markers improved BCR risk stratification and identified a subgroup of patients with high probability of recurrence. The 16p13.3 gain status was also associated with an increased risk of developing distant metastases (log-rank P = 0.03) further substantiating its role in prostate cancer progression.Implications: This study demonstrates the prognostic significance of the 16p13.3 genomic gain in primary prostate tumors, suggesting potential utility in the clinical management of the disease by identifying patients at high risk of recurrence who may benefit from adjuvant therapies. Mol Cancer Res; 16(1); 115-23. ©2017 AACR.

Animal Models of Uveal Melanoma: Methods, Applicability, and Limitations

Animal models serve as powerful tools for investigating the pathobiology of cancer, identifying relevant pathways, and developing novel therapeutic agents. They have facilitated rapid scientific progress in many tumor entities. However, for establishing a powerful animal model of uveal melanoma fundamental challenges remain. To date, no animal model offers specific genetic attributes as well as histologic, immunologic, and metastatic features of uveal melanoma. Syngeneic models with intraocular injection of cutaneous melanoma cells may suit best for investigating immunologic/tumor biology aspects. However, differences between cutaneous and uveal melanoma regarding genetics and metastasis remain problematic. Human xenograft models are widely used for evaluating novel therapeutics but require immunosuppression to allow tumor growth. New approaches aim to establish transgenic mouse models of spontaneous uveal melanoma which recently provided preliminary promising results. Each model provides certain benefits and may render them suitable for answering a respective scientific question. However, all existing models also exhibit relevant limitations which may have led to delayed research progress. Despite refined therapeutic options for the primary ocular tumor, patients' prognosis has not improved since the 1970s. Basic research needs to further focus on a refinement of a potent animal model which mimics uveal melanoma specific mechanisms of progression and metastasis. This review will summarise and interpret existing animal models of uveal melanoma including recent advances in the field.

Animal Eye Models for Uveal Melanoma

Animal models play an important role in understanding tumor growth and may be used to develop novel therapies against human malignancies. The significance of the results from animal experiments depends on the selection of the proper model. Many attempts have been made to create appropriate animal models for uveal melanoma and its characteristic metastatic behavior. One approach is to use transgenic animal models or to implant tumor cells. A variety of tumor types have been used for this purpose: tumor cells, such as Greene melanoma, murine B16 melanoma, and human uveal melanoma cells, may be implanted in the eyes of hamsters, rats, rabbits, and mice, among others. Various inoculation routes, including into the anterior chamber and posterior compartment, and retro-orbitally, have been applied to obtain tumor growth mimicking ocular uveal melanoma. However, when we choose animal models, we must be conscious of many disadvantages, such as variable tumor growth, or the need for immunosuppression in xenogeneic grafts. In this paper, we will discuss the various eye models.

Uveal melanoma: Ocular and systemic disease

Although rare, uveal melanoma is the most common intraocular tumor in adults. Most cases arise from the choroidal layer of the uvea, displaying a discoid, collar-button, or mushroom shaped growth. Histopathologically, neoplasms are classified by the dominant cell type: spindle, epithelioid or mixed spindle cell type. The most important prognostic factors are cell type, nucleolar size, largest tumor dimension, and mitotic figures. Patient prognosis is poor when metastases occur in the liver, one of the main reasons that despite advances in the diagnosis and treatment of uveal melanoma, the mortality rate has not change significantly since 1973.

Host pigment epithelium-derived factor (PEDF) prevents progression of liver metastasis in a mouse model of uveal melanoma

Uveal melanoma (UM) has a 30 % 5-year mortality rate, primarily due to liver metastasis. Both angiogenesis and stromagenesis are important mechanisms for the progression of liver metastasis. Pigment epithelium-derived factor (PEDF), an anti-angiogenic and anti-stromagenic protein, is produced by hepatocytes. Exogenous PEDF suppresses metastasis progression; however, the effects of host-produced PEDF on metastasis progression are unknown. We hypothesize that host PEDF inhibits liver metastasis progression through a mechanism involving angiogenesis and stromagenesis. Mouse melanoma cells were injected into the posterior ocular compartment of PEDF-null mice and control mice. After 1 month, the number, size, and mean vascular density (MVD) of liver metastases were determined. The stromal component of hepatic stellate cells (HSCs) and the type III collagen they produce was evaluated by immunohistochemistry. Host PEDF inhibited the total area of liver metastasis and the frequency of macrometastases (diameter >200 μm) but did not affect the total number of metastases. Mice expressing PEDF exhibited significantly lower MVD and less type III collagen production in metastases. An increase in activated HSCs was seen in the absence of PEDF, but this result was not statistically significant. In conclusion, host PEDF inhibits the progression of hepatic metastases in a mouse model of UM, and loss of PEDF is accompanied by an increase in tumor blood vessel density and type III collagen.

Uveal melanoma dormancy: an acceptable clinical endpoint?

Uveal melanoma is a rare but life-threatening malignancy. Over the past decades, the morbidity of uveal melanoma has been markedly reduced as a result of advances in the diagnostic ability to detect smaller tumors at an earlier stage. This has allowed for the use of more conservative treatments, avoiding enucleation. Mortality, however, has remained unchanged. This indicates that life expectancy is independent of local tumor control. Metastatic disease, the leading cause of death, is usually diagnosed many years later, despite successful treatment of the primary tumor, and at a late stage, when no effective therapy is available. These observations suggest that the disease was already disseminated at the time of tumor diagnosis. The detection of circulating malignant cells in the bloodstream of patients at different time points in the course of the disease supports this observation. Tumor dormancy has been considered as the leading theory for this intriguing delayed appearance of metastasis. Recent knowledge gained about the biological behavior of uveal melanoma as well as novel potential therapeutic targets are presented in this review.

miRNA-7-5p inhibits melanoma cell migration and invasion

Aberrant expression of microRNAs (miRNAs), a class of small non-coding regulatory RNAs, has been implicated in the development and progression of melanoma. However, the precise mechanistic role of many of these miRNAs remains unclear. We have investigated the functional role of miR-7-5p in melanoma, and demonstrate that miR-7-5p expression is reduced in metastatic melanoma-derived cell lines compared with primary melanoma cells, and that when ectopically expressed miR-7-5p significantly inhibits melanoma cell migration and invasion. Additionally, we report that insulin receptor substrate-2 (IRS-2) is a target of miR-7-5p in melanoma cells, and using RNA interference (RNAi) we provide evidence that IRS-2 activates protein kinase B (Akt), and promotes melanoma cell migration. Thus, miR-7-5p may represent a novel tumor suppressor miRNA in melanoma, acting at least in part via its inhibition of IRS-2 expression and oncogenic Akt signaling.

Effects of interleukin-1 receptor antagonist on tumor stroma in experimental uveal melanoma

PURPOSE

In contrast to many malignancies showing evidence that interleukin-1 (IL-1) promotes progression through effects on tumor vascularity and myeloid suppressor cell populations, in uveal melanoma there is evidence that IL-1 can inhibit progression.

METHODS

The effects of the IL-1 receptor antagonist IL-1ra against the aggressive/invasive MUM2B and the nonaggressive/noninvasive OCM1 uveal melanoma models were examined in vitro and in vivo in mouse xenografts. Vascularity and myeloid suppressor cell populations and their regulators were assessed.

RESULTS

In vitro, IL-1, and IL-1ra did not affect the proliferation of the uveal melanoma cells or their production of IL-1, IL-6, transforming growth factor (TGF) β, or VEGF. In vivo, IL-1ra treatment resulted in substantial growth inhibition of MUM2B tumors; less inhibition was observed against OCM1 tumors. Periodic acid-Schiff loops and CD11b⁺ macrophages within the tumor stroma decreased in vivo; CD31⁺ blood vessels were not altered. IL-1ra treatment in vivo did not affect tumor-derived IL-1, IL-6, TGF-β, or VEGF. In contrast, host IL-1β, IL-6, and tumor necrosis factor decreased. Host VEGF was not altered. Intratumoral IL-12(p40) and CXCL10, markers of host M1 polarization, increased, and intratumoral arginase and CD206, markers of myeloid-derived suppressor cells (MDSC) and M2 macrophage polarization, decreased. IL-1ra treatment in vivo also reduced splenic CD11b⁺Gr1⁺ MDSC.

CONCLUSIONS

IL-1 may play a role in promoting uveal melanoma progression. Inhibiting IL-1 with IL-1ra inhibits tumor growth in vivo but not in vitro. Tumor stroma is modified, myeloid suppressor cells are reduced, and M1 macrophage polarization is increased in vivo.

Prognostic relevance of circulating tumor cells in metastatic uveal melanoma

OBJECTIVE

Uveal melanoma primarily metastasizes hematogenously with metastases often confined to the liver. The aim of this study was to investigate the presence of circulating tumor cells (CTC) in patients with metastatic disease as a marker for systemic disease and to determine their prognostic relevance.

METHODS

Blood samples from 68 patients were collected at the time of initial treatment of metastases. mRNA expression of tyrosinase and MelanA/MART1 as a surrogate marker for the presence of CTC was analyzed by real-time RT-PCR and compared with patient characteristics.

RESULTS

CTC were detected in 63% of all patients and in 67% of the 48 patients with only liver metastases. Univariate and multivariate analyses revealed PCR results and serum lactate dehydrogenase as independent prognostic factors for progression-free (hazard ratios 2.2/3.5) and overall survival (hazard ratios 4.0/6.5). Combination of PCR and lactate dehydrogenase divided the patient cohort into 3 groups with distinct prognosis.

CONCLUSION

CTC as evidence for systemic disease can be found in the majority of patients with metastatic uveal melanoma, including patients with visible disease confined to the liver. Detection of CTC-specific mRNA transcripts for tyrosinase and MelanA/MART1 by PCR is a poor prognostic factor for progression-free and overall survival. Characterization of CTC could improve the understanding of their biology.

Expression profiling of formalin-fixed paraffin embedded primary human uveal melanomas using DASL matrices

PURPOSE

Fresh biopsied ocular tumor tissues are difficult to obtain for the purpose of performing microarray experiments on extracted nucleic acids. Present technology allows for extraction of total RNA from formalin-fixed paraffin embedded (FFPE) tissue analyzed by the cDNA mediated Annealing Sectioning and Ligation (DASL) method. We aimed to correlate gene transcript differences between two uveal melanoma (UM) clinical-histopathological parameters (metastasis, cell type).

METHODS

A total of 43 FFPE UM were used. The expression of RPL13a, a ribosomal protein gene, for each sample was used to evaluate the quality of RNA extracted from FFPE tissue. Gene expression values generated from the array were analyzed using the GeneSpring GX software (Agilent). Immunohistochemistry was used in order to validate transcriptional findings at the protein level.

RESULTS

A total of 106 genes were identified with (P < 0.05, Welch ANOVA test) a difference in transcript abundance for the metastasis clinical parameter. Furthermore, we identified 64 genes with a statistically significant (P < 0.05) difference in transcript abundance between the spindle and epithelioid cell types. Each individual sample for both groups (metastasis, cell type) exhibited distinct transcriptional profiles that were separated on a PCA. Positive nuclear immunostaining for LIG4-metastasis, ErbB3-cell type was found to be associated with better patient prognosis and outcome.

CONCLUSIONS

To the best of our knowledge, this is the first time that a successful retrospective analysis has been done with UM FFPE RNA. This data may lead to future customized therapeutic targets, which may improve the now unchanged mortality rate of this particular malignancy.

Molecular pathways mediating liver metastasis in patients with uveal melanoma

Uveal melanoma arises from melanocytes located in the uveal tract of the eye and is the most common primary intraocular tumor in adults. Metastatic liver disease is the overwhelming cause of death in uveal melanoma patients, with almost 50% of patients developing liver metastases up to 15 years after diagnosis. Most of these patients do not present with any evidence of overt metastasis at the time of initial diagnosis although it is assumed that they have undetectable micrometastases. Currently, there are no therapeutic modalities to prevent or efficiently treat the metastatic disease in uveal melanoma patients. Recent discoveries have shed light on the molecular pathways that may contribute to the progression of liver metastasis. The aim of this review is to describe new insights into the genetic and molecular pathways that may play a role in the development of liver metastases in uveal melanoma patients.

In-vivo xenograft murine human uveal melanoma model develops hepatic micrometastases

The purpose of the study is to develop a mouse ocular melanoma model with human uveal melanoma cells that forms hepatic micrometastases. Human uveal melanoma Mel290 cells were transfected with a lentiviral-enhanced green fluorescent protein (EGFP) expression vector. Proliferation assays were performed by comparing Mel290-EGFP and Mel290 cells. After stable expression of EGFP and proliferation was ascertained, 1 x 10 Mel290-EGFP cells were introduced into NU/NU mice by posterior compartment (PC) inoculation or tail vein injection. Control groups were inoculated or injected with Mel290 cells. Ocular and hepatic frozen sections were examined by fluorescence microscopy, and the number of hepatic micrometastases was determined. EGFP expression was observed at 24 h after transfection. At 72 h after transfection, more than 70% of Mel290 cells expressed EGFP. At 45 days (six passages), 90% of Mel290 cells stably expressed EGFP. Histologic examination showed that Mel290-EGFP cells formed hepatic micrometastases after either PC inoculation or tail vein injection. A significant difference in the number of hepatic micrometastases between PC inoculation and tail vein injection (P<0.01) was observed. Mel290-EGFP cells stably expressed green fluorescent protein in vitro at 45 days (six passages). These cells formed hepatic micrometastases in NU/NU mice after PC inoculation or tail vein injection, with significantly more micrometastases developing in the PC inoculation model than after tail vein injection.