BAP1 deficiency causes loss of melanocytic cell identity in uveal melanoma

Abstract

Background: Uveal melanoma is a highly aggressive cancer with a strong propensity for metastasis, yet little is known about the biological mechanisms underlying this metastatic potential. We recently showed that most metastasizing uveal melanomas, which exhibit a class 2 gene expression profile, contain inactivating mutations in the tumor suppressor BAP1. The aim of this study was to investigate the role of BAP1 in uveal melanoma progression.

Methods: Uveal melanoma cells were studied following RNAi-mediated depletion of BAP1 using proliferation, BrdU incorporation, flow cytometry, migration, invasion, differentiation and clonogenic assays, as well as in vivo tumorigenicity experiments in NOD-SCID-Gamma mice.

Results: Depletion of BAP1 in uveal melanoma cells resulted in a loss of differentiation and gain of stem-like properties, including expression of stem cell markers, increased capacity for self-replication, and enhanced ability to grow in stem cell conditions. BAP1 depletion did not result in increased proliferation, migration, invasion or tumorigenicity.

Conclusions: BAP1 appears to function in the uveal melanocyte lineage primarily as a regulator of differentiation, with cells deficient for BAP1 exhibiting stem-like qualities. It will be important to elucidate how this effect of BAP1 loss promotes metastasis and how to reverse this effect therapeutically.

Figures

Figure 1

Transient loss of BAP1 causes a decrease in proliferation in uveal melanoma cells. (a) Western blot of two uveal melanoma cell lines, 92.1 and Mel290, showing the levels of BAP1 after transient knockdown of BAP1 with two independent siRNAs compared to control siRNA. α-tubulin was used as a loading control (b-c) Cell proliferation of the indicated melanoma cell lines transfected with BAP1 (grey line) or control (black line) siRNA. Proliferation was measured by BrdU incorporation and shown as fold change compared to Day 1 (b) 92.1 cells (c) Mel290 cells.

Figure 2

Stable loss of BAP1 does not promote cell proliferation in uveal melanoma cells. (a) Western blot showing the levels of BAP1 after stable knockdown of BAP1 with two independent lentiviral shRNA constructs in three uveal melanoma cell lines, OCM1A, 92.1, and Mel290. α-tubulin was used as a loading control (b) Cell viability and proliferation of the indicated BAP1-deficient or control stable cells (Left panel) cell viability measured by MTS assay and shown as fold change over day 0 for each stable cell line (Right panel) cell proliferation measured by BrdU incorporation after 24 hrs (c-d) Cell cycle analysis of the indicated cell lines using flow cytometry on propidium iodide stained cells; x-axis represents DNA content and y-axis represents cell number (c) 92.1 stable cells (d) OCM1A stable cells.

Figure 3

Loss of BAP1 does not promote in vitro tumorigenicity in uveal melanoma cells. (a) Wound healing assays were performed on the indicated BAP1-deficient or control stable cells. (Left panel) representative pictures of Mel290 cells after stable knockdown at Day 0 (time of scratch) and Day 1. (Right panel) quantification of wound healing assays shown as percent of total area of the initial wound (b) BAP1-deficient or control Mel290 stable cells were monitored every 15mins for 16 hrs using live cell imaging. Individual cells were manually tracked and the average distance per frame (left panel) and the total distance traveled (right panel) were calculated (c) Soft agar assays performed on BAP1-deficient or control OCM1A stable cells after one week of growth. (Left panel) representative pictures of soft agar plates stained with crystal violet. (Right panel) quantification of the number of small and large colonies per plate determined using ImageJ software. * denotes P < 0.05 and ** denotes P < 0.01 based on Student’s t-test.

Figure 4

Loss of BAP1 does not promote growth of uveal melanoma in mouse flank and tail vein xenograft models. (a) Mouse xenograft flank injections in which 1000 or 500 BAP1-deficient or control 92.1 or OCM1A stable cells, respectively were injected into the flanks of NSG mice. (Left panel) 92.1 stable cells (Right panel) OCM1A stable cells. Tumor volume was measured at time of necropsy 64 days or 35 days after injection for 92.1 and OCM1A cells, respectively. (b) RNA expression levels of BAP1 in flank tumors formed from BAP1-deficient of control stable cell lines. RNA was isolated at time of necropsy. Expression is shown as fold change compared to control shRNA cells (c-d) Mouse xenograft tail vein injections in which 10,000 BAP1-deficient or control 92.1 stable cells were injected into the tail veins of NSG mice. Necropsy was performed 29 days after injection and liver metastasis was assessed (c) Representative cartoons of liver metastasis made from merged images of liver sections stained with H&E. Images were merged using Adobe Photoshop. Normal liver tissue is shown in grey and metastases are in black. Original H&E merged images are shown in Additional file 2. (d) Quantification of liver metastasis. Shown as the percent of tumor area compared to the total liver area. ImageJ software was used to calculate area (e) Mouse xenograft tail vein injections in which 500,000 BAP1-deficient or control OCM1A stable cells were injected into the tail veins of NSG mice. Liver and lung metastases were measured 30 days after injection. Metastases are shown as a percent of the total liver or lung area.

Figure 5

Gene expression profiling reveals altered RNA metabolism and developmental transcriptomic signatures in BAP1-deficient uveal melanoma cells. (a) Gene expression heatmap generated in Partek of significantly altered genes (P < 0.01) in BAP1-deficient stable cells compared to control stable cells after 4 weeks of shRNA expression. (b) The top six gene ontology categories represented by the significantly changing genes in BAP1-deficient cells. Significant Analysis of Microarrays (SAM) was performed and a cutoff of q < 10% was used to determine the most highly significant genes. Three of the genes in the ubiquitin system category are deubiquitinases (DUBs). (c) The top nine categories represented after gene set enrichment analysis (GSEA) of BAP1-deficient cells using a threshold of P < 0.005. GSEA was performed using a pre-ranked file generated after SAM analysis.

Figure 6

Loss of BAP1 induces a dedifferentiated, stem-like phenotype in uveal melanoma cells. (a) RNA expression of melanocyte differentiation markers in BAP1-deficient or control 92.1 stable cells, primary melanocytes, and primary class 1 tumor cells. Fold change is compared to control knockdown cells. (b) Cell morphology of BAP1-deficient or control primary uveal melanocytes. (Left panel) representative pictures. (Right panel) dendritic arborizations measured as the number of branch points per cell, with a bipolar cell having zero aborizations; shown as percent of total cells. (c) Cell spindle morphology of BAP1-deficient or control primary uveal melanocytes after four weeks of shRNA expression, measured as the ratio of cell length to cell width. (d) RNA expression levels of stem cell genes NANOG and OCT4 in BAP1-deficient or control 92.1 and OCM1A stable cells. Fold change is compared to control knockdown cells. (e) Clonogenic assay using BAP1-deficient or control OCM1A stable cells. One cell/well was grown in serum-free stem cell media in non-adherent 96-well plates. The number of wells producing colonies was measured after 5 days. Data is shown as fold change over control. (f) Assay measuring the ability of BAP1-deficient or control stable cells to grow in stem cell culture conditions on non-adherent plates in serum-free media. Colony size was measured after 5 (OCM1A) or 7 (92.1) days using ImageJ software. (g) RNA expression of melanocyte differentiation markers in BAP1-deficient 92.1 stable cells after 72 hr treatment with 0.5 mM, 1.0 mM, or 2.0 mM HDAC inhibitor (HDACi) as indicated. Fold change is compared to control knockdown cells.

Figure 7

Physical interaction between BAP1 and HCF-1 in uveal melanoma cells. Immunoprecipitations (IP) for endogenous BAP1 and HCF-1 from the combined lysates of UM cell lines. Westerns were performed on IP material and lysate supernatant that was collected after the IP (cleared lysates). α-tubulin was used as a loading control for the cleared lysate samples only. Densitometry was performed using ImageJ and showed ~75% of the total BAP1 was complexed with HCF-1. * denotes P < 0.05 and ** denotes P < 0.01 based on Student’s t-test.