October 2016
Volume 57, Issue 13
Open Access
Biochemistry and Molecular Biology  |   October 2016
Effects of Long-term Serial Passaging on the Characteristics and Properties of Cell Lines Derived From Uveal Melanoma Primary Tumors
Author Affiliations & Notes
  • Frédéric Mouriaux
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Département d'ophtalmologie, Faculté de Médecine, Université Laval, Québec, Canada
    CNRS, UMR 6301 ISTCT, CERVOxy. GIP CYCERON, Caen, France
    CHU de Rennes, Service d'Ophtalmologie, France
  • Karine Zaniolo
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
  • Marjorie-Allison Bergeron
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
  • Cindy Weidmann
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
  • Arnaud De La Fouchardière
    Département de biopathologie, Centre Léon Bérard, Lyon, France
  • Frédéric Fournier
    Département de médecine moléculaire, CHUL, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
  • Arnaud Droit
    Département de médecine moléculaire, CHUL, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
  • Mohib W. Morcos
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
    Service de pathologie, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
  • Solange Landreville
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Département d'ophtalmologie, Faculté de Médecine, Université Laval, Québec, Canada
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
  • Sylvain L. Guérin
    Centre universitaire d'ophtalmologie-Recherche (CUO-Recherche), Axe médecine régénératrice, Hôpital du Saint-Sacrement, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada
    Département d'ophtalmologie, Faculté de Médecine, Université Laval, Québec, Canada
    Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX, Québec, Canada
  • Correspondence: Sylvain L. Guérin, CUO-Recherche, Hôpital du Saint-Sacrement, Centre de recherche FRQS du CHU de Québec, Québec, QC, Canada;Sylvain.Guerin@fmed.ulaval.ca
  • Footnotes
     SLG and SL are joint senior authors.
Investigative Ophthalmology & Visual Science October 2016, Vol.57, 5288-5301. doi:10.1167/iovs.16-19317
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Frédéric Mouriaux, Karine Zaniolo, Marjorie-Allison Bergeron, Cindy Weidmann, Arnaud De La Fouchardière, Frédéric Fournier, Arnaud Droit, Mohib W. Morcos, Solange Landreville, Sylvain L. Guérin; Effects of Long-term Serial Passaging on the Characteristics and Properties of Cell Lines Derived From Uveal Melanoma Primary Tumors. Invest. Ophthalmol. Vis. Sci. 2016;57(13):5288-5301. doi: 10.1167/iovs.16-19317.

      Download citation file:


      © 2017 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

Purpose: Development of liver metastasis remains the most common cause of mortality in uveal melanoma (UM). A few cell lines cultured from primary UM tumors have been used widely to investigate the pathobiology of UM. However, the translation of basic knowledge to the clinic for the treatment of the metastatic disease has remained incremental at best. In this study, we examined whether the properties of UM cell lines at various passages were similar to their corresponding primary tumors.

Methods: Gene expression profiling by microarray was performed on UM primary tumors and derived cell lines cultured at varying passages. Expression of UM protein markers was monitored by immunohistochemical analyses and Western blotting. The in vivo tumorigenic properties of UM cultures were evaluated using athymic nude mice.

Results: Cell passaging severely reduced the expression of genes encoding markers typical of UM, including those of the prognostic gene signature. Marked differences between gene expression profiles of primary tumors and cell lines could be linked to the infiltrating immune and stromal cells in situ. In addition, the tumorigenic properties of UM cell lines also increased with cell passaging in culture as evaluated by their subcutaneous injection into athymic mice.

Conclusions: Together, these findings demonstrate that the short-term UM primary cultures exhibit molecular features that resemble the respective surgical material and, thus, represent the best model for in vitro–assessed cancer treatments.

For decades, tumor cells cultured in Petri dishes have been used to advance our understanding of molecular cancer biology and are now well recognized as a cornerstone of anticancer drug development. Cell lines acquire infinite replication when they are grown in the appropriate culture conditions. They then become an invaluable tool for the cellular/molecular biologists because of the limited availability of resected human tumors. Cell lines offer numerous advantages, including cost-effectiveness, limitless replicative ability, and ease of manipulation for the fundamental study of altered cellular pathways and epigenetic modifications, as well as for anticancer drug testing.1 Results achieved using continuous cell lines as biologic models often are extrapolated to the tumors from which they originate. However, there is a widespread debate among the scientific community on their validity owing to the possibilities of clonal selection and tissue culture artifacts.13 Indeed, cell lines exhibit a lack of differentiation and genomic instability after accumulating new chromosomal rearrangements and mutations during their adaptation to an artificial environment without stromal, vascular, and inflammatory cells.18 
Uveal melanoma (UM) is the most common primary ocular malignancy in adults, accounting for 70% of all eye cancers. The overall mean age-adjusted incidence in the United States is 4 to 6 per million individuals.9 The actuarial 15-year metastatic mortality is approximately 50% despite efficient treatment of the primary ocular tumor.10 Factors associated with metastatic disease include pathologic features11; cytogenetic abnormalities, such as monosomy 3 and 8q gain12,13; BAP1 mutations14; and the class 2 gene signature.15 Because UMs seldom are surgically resected and biopsies dedicated to research very small, laboratory studies are limited by the lack of material. Despite that the establishment of permanent cell cultures from UM tumors is difficult, a number of cell lines have been characterized to date that can be used for research purposes.1639 The availability of these UM cell lines allows ophthalmic researchers to advance from correlative studies to mechanistic investigations using advanced molecular methodologies. However, there often are discrepancies in treatment response between UM cell lines and UM primary tumors and metastases. For instance, based on promising in vitro studies,40,41 our group and others conducted phase II clinical trials with imatinib mesylate as a pharmacologic treatment of metastatic UM.42,43 However, the encouraging in vitro results were not reproduced in patients, as imatinib mesylate did not improve overall survival.42,43 A subsequent study to understand the molecular mechanisms behind this resistance, using UM metastatic tissues and cell lines, demonstrated that expression of the stem cell factor (SCF) in the tumor environment might promote resistance of metastatic UM to imatinib mesylate.44 Interpreting the potential clinical significance of the data obtained from cell lines also requires to determine the extent to which a given cell line and its molecular features resemble the respective surgical material. Such features have been investigated in several studies for other types of cancers using genomic and transcriptomic analyses.1,36,8,4549 
So far, only a few detailed comparisons of the cell lines properties with those of the UM primary tumors from which they were derived have been published.19,21,31,33,34,37,38 We wished to compare the properties of UM primary tumors to those of their corresponding cell lines at various passages. Morphologic characteristics, gene expression profiling, and protein expression were examined by immunohistochemical analyses, DNA microarray, and Western blotting. Moreover, we compared the in vivo tumorigenic properties of short- and long-term UM primary cultures using athymic nude mice. 
Materials and Methods
This study followed the tenets of the Declaration of Helsinki and was approved by our institutional human experimentation committee. Written informed consent was obtained from the enucleated subjects. All animal experiments were conducted in voluntary compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and our institutional animal care and use committee approved all procedures. 
Primary Tumor Samples and Tissue Culture
Six archival paraffin-embedded and frozen human UM tumor tissues (105, 122, 127, 130, 142, and 143) and their corresponding cell lines (T105, T122, T127, T130, T142, and T143) were studied. All samples were collected from enucleated eyes and confirmed by the Service de pathologie du CHU de Québec to be UMs by immunohistochemistry.35 Clinicopathologic characteristics, survival data, chromosomal anomalies and mutation status are compiled in Supplementary Table S1. Pathologic data were assessed as described previously.50 All UM cell lines established in our laboratory were derived from postenucleation biopsies performed by an ocular pathologist and were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich Corp., St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich Corp.) and 0.002% (vol/vol) gentamicin, under 5% CO2 at 37°C as described previously.35 Normal melanocytes or vascular endothelial cells cannot grow in this medium, and tumor-infiltrating noncancerous fibroblasts are scarce in UM. Purity of the cultures was confirmed by microscopic inspection and MART-1 staining (prediluted; Dako Canada, Mississauga, ON, Canada). All cell lines were tested routinely for mycoplasma infection using the Mycoplasma Detection Kit (Biotool, Houston, TX, USA). To test their anchorage-independent growth, the cell lines were plated in 1.3% methylcellulose (diluted in Iscove's modified Dulbecco's medium [IMDM]; R&D Systems, Minneapolis, MN, USA) on untreated 35-mm tissue culture dishes (10,000 cells per triplicate). After 14 days of growth, colonies were observed under a dissecting microscope after staining with 0.1 mg/ml MTT (Sigma-Aldrich Corp.). 
Cell Count and Doubling Time
Uveal melanoma cell number was compared between low and high passages using a cell and particle counter (Z2; Beckman Coulter, Mississauga, ON, Canada). Cell counts were done (N = 3) at three consecutive passages. The doubling time was determined using the initial number of cells, final number of cells, and duration of culture in hours (Roth V. 2006: available in the public domain at http://www.doubling-time.com/compute.php). 
Immunohistochemical Analyses
Cells were grown until they reached 80% confluency and harvested during log phase, then fixed in 10 volumes of phosphate-buffered formalin for 10 minutes before being embedded in paraffin. Cell pellet sections were analyzed for structural features, including cell morphology, nuclear atypia, and mitosis, using hematoxylin and eosin stain. Immunohistochemical analyses were performed on cell line pellets and primary tumors sections following a technique of diaminobenzidine (DAB) biotin-streptavidin complex formation.50 Immunohistologic analyses were performed using antibodies directed against the proteins MART-1 (gene symbol, MLANA; prediluted; Dako Canada), S100 (prediluted; Dako Canada), HMB45 (prediluted; Dako Canada), VEGF (dilution 1:100; Santa Cruz Biotechnology, Dallas, TX, USA), cyclin D1 (gene symbol, CCND1; dilution 1:50; Cell Signaling Technology, Danvers, MA, USA), and p16 (gene symbol, CDKN2A; dilution 1:400; Santa Cruz Biotechnology). A negative control consisting of the omission of the primary antibody was performed for each case. The immunohistochemical staining results were evaluated using the blinded fashion technique by two experienced investigators without knowledge of clinicopathologic data on each individual case. Staining intensity was graded as negative (−), weak (+), moderate (++), or strong (+++). 
Western Blotting
Western blots were conducted as described previously35,51 using the following primary antibodies directed against the proteins HTR2B (dilution 1:1000; Sigma-Aldrich Corp.), PARP-1 (dilution 1:2000), and PARG (dilution 1:2000; both kind gifts of Serge Desnoyers, Centre de Recherche FRQS du CHU de Québec-Université Laval, Québec, Canada), integrin α5 subunit (10 μg/ml; Chemicon, Temecula, CA, USA), Sp1 (dilution 1:2000; Santa Cruz Biotechnology), VEGF (dilution 1:200, Santa Cruz Biotechnology), cyclin D1 (dilution 1:1000; Cell Signaling Technology), and p27 (gene symbol, CDKN1B; dilution 1:1000; Cell Signaling Technology). A β-actin antibody was used as loading control (dilution 1:40,000; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). The membrane then was incubated with peroxidase-conjugated goat secondary antibodies against either mouse or rabbit IgG (1:10,000; Jackson ImmunoResearch Laboratories), and immunoreactive complexes were revealed using a Western blot detection kit (Amersham, Baie d'Urfé, QC, Canada). 
Tumorigenicity Assays in Athymic Nude Mice
Seven-week-old female Crl:CD1-nuBR athymic nude mice (Charles River, Saint-Constant, QC) were inoculated with the UM cell lines T122, T130, T142, and T143 (1 × 106 cells injected in each flank) cultured at low (P4/P6), intermediate (P20/P21), or high (P60/P62) passages. Six female mice were used (both flanks) for each cell line and at each passage (total of 12 injections). All mice were killed at 28 days after injections. Tumors were excised and their volume precisely determined using the following formula: V = (L + W2) × 0.5. They then were embedded in paraffin for histologic analyses using hematoxylin and eosin (H&E) staining and MART-1 antibody as described previously.35 
Gene Expression Profiling and Unsupervised Bioinformatics Analyses
Total RNA was isolated directly from two UM primary tumors (patients 142 and 143) and the corresponding cell lines (T142 and T143) cultured to near 100% confluency at passages P4, P20, and P60 using the RNeasy Mini Kit (Qiagen, Hilden, Germany). Gene expression profiling was performed by the Plateforme de génétique moléculaire du CUO-Recherche (Québec, QC, Canada). Quantity and quality of total RNA were assessed using the 2100 Bioanalyzer and the RNA 6000 Nano LabChip kit (Agilent Technologies, Mississauga, ON, Canada). Cyanine 3-CTP–labeled cRNA targets were prepared and hybridized on a SurePrint G3 Human GE 8 × 60K array slide (60,000 probes, Agilent Technologies) as reported previously.51 Data were analyzed using the ArrayStar software (version 4.1; DNASTAR, Madison, WI, USA). All data generated from the array also were analyzed by RMA (“Robust Multiarray Analysis”) for background correction of the raw values. They then were transformed in log2 base and quantile normalized, before a linear model was fitted to the normalized data to obtain an expression measure for each probe set. The gene list of stromal and immune signatures was determined previously by Yoshihara et al.52 Linear signals and functions of genes plotted in heatmaps are listed in Supplementary Table S4. Unsupervised bioinformatic analyses for the enrichment of metabolic pathways were performed with the R package clusterProfiler (version 2.4.3), the library ReactomePA (version 1.14),53 and the Agilent annotation file 028004_D_GeneList_20150612.txt as the comparison universe. For these analyses, we used lists of differentially expressed genes (>4-fold change) between the tumors and passages P4/P20/P60 of their derived cell lines. All data with an adjusted P value <0.05 are listed in Supplementary Table S5. All microarray data presented in this study comply with the Minimum Information About a Microarray Experiment (MIAME) requirements. The gene expression data have been deposited in NCBIs Gene Expression Omnibus (GEO, available in the public domain at http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE61228. 
Quantitative PCR (qPCR)
Reverse transcription was performed using random hexamer primers following the manufacturer's protocol (Superscript II; Invitrogen, Carlsbad, CA, USA). Equal amounts of cDNA were run in quadruplicate and amplified in a 20 μl reaction containing 10 μL of 2X Brillant III Ultra-Fast SYBR Green QPCR Master Mix (Agilent Technologies), 250 nM of upstream and downstream primers, and 10 ng of cDNA target. The mixture was incubated at 95°C for 3 minutes, and then cycled 40 times at 95°C for 5 seconds and at 60°C for 10 seconds using the Rotor-Gene Q real-time cycler (Qiagen). Amplification efficiency of the HTR2B transcript was validated and normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcript and quantity of HTR2B was calculated using the standard curve method. Specific primers were designed using Primer3 software (available in the public domain at http://biotools.umassmed.edu/bioapps/primer3_www.cgi): HTR2B_Forward-TCTTTTCAACCGCATCCATCA and HTR2B_Reverse-TGCTGTAGCCCGTGAGTTATA (107 base pairs [bp]); GAPDH_Forward-AAGGTCGGAGTCAACGGAT and GAPDH_Reverse-GGAAGATGGTGATGGGATTTC (220 bp). 
Statistical Analyses
Student's t-test was used to determine statistical significance for comparison of the groups in qPCR analyses and the measurement of the cell doubling times. The values are represented as mean ± SEM and P < 0.05 was considered significant. The sample size included data from three independent experiments performed in triplicate. Xenograft tumor growth was compared using 1-way ANOVA with Tukey's multiple comparisons test and P < 0.05 was considered significant. 
Results
Main Characteristics of the UM Cell Lines and the Primary Tumors From Which They Originate
Examination of UM cell lines under phase contrast microscopy indicated that they all exhibit an epithelioid morphology (with the exception of T105 and T143 that have a mixed spindle/epithelioid and a spindle phenotype, respectively; Fig. 1A). All cell lines formed colonies (Fig. 1B) compared to melanocytes (data not shown) demonstrating their anchorage-independent proliferation. Moreover, they all expressed the melanocyte differentiation antigen MART-1 (Fig. 1C; Table), a pattern very similar to that of the primary tumors from which they originated (Fig. 2; Table). Doubling times were measured for the UM cell lines T105, T122, T127, T130, T142, and T143 at three consecutive low- (ranging from P6–P11 depending on the cell line used) and high (ranging from P60–P62) passages. The average doubling time ranged from 23.65 ± 0.35 (with T130) to 34.32 ± 3.46 (with T142) at low passages, but then significantly decreased as UM cells progressed to higher passages (from 14.46 ± 1.06 with T142 to 20.71 ± 2.48 with T127; Supplementary Table S2), which is quite faster than UVM doubling time (>78h).28 The most dramatic differences were observed for the T105 and T142 cell lines, whose respective doubling times decreased from 29.64 ± 2.04 and 34.32 ± 3.46 at low passage to 16.55 ± 0.15 (P < 0.005) and 14.46 ± 1.06 (P = 0.005) at high passage, respectively. We were able to analyze the cytogenetic abnormalities and mutation status in four of our primary tumors (Supplementary Table S1). The primary tumor 105, isolated from a patient who died of other cause, was positive for trisomy 8, and was mutant for BAP1 and GNA11 (Q209L). The primary tumors 130 and 143, derived from patients alive without metastasis, were negative for trisomy 8 and were wild type (WT) for BAP1. The primary tumor 142, established from a patient who died of metastasis, was positive for monosomy 3 and trisomy 8, and was mutant for BAP1 and GNA11 (Q209L). Short Tandem Repeat (STR) profiling could be completed for four of our UM cell lines (Supplementary Table S3) and revealed no relationship to any other human cell line. However, many loci differed from the patient's blood or tumor samples. To determine the effect of long-term culture on the stability of DNA fingerprinting was beyond the scope of this study. However, experiments presently are underway to investigate the potential limitations of STR profiling for UM cell line authentication (such as genetic drifting, and presence of endogenous PCR inhibitory contaminants, like melanin).54 
Figure 1
 
Characteristics of cell lines derived from UM primary tumors. (A) Phase contrast micrographs of monolayer cultures of normal uveal melanocytes (UVM) or UM cell lines T105, T122, T127, T130, T142, and T143. Scale bar: 20 μm. (B) Phase contrast micrographs of the MTT-stained colonies formed by UM cell lines from (A). Scale bar: 20 μm. (C) Hematoxylin and eosin (H&E) staining for morphologic features and expression of MART-1 by immunohistochemistry on UM cell lines from (A). Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 1
 
Characteristics of cell lines derived from UM primary tumors. (A) Phase contrast micrographs of monolayer cultures of normal uveal melanocytes (UVM) or UM cell lines T105, T122, T127, T130, T142, and T143. Scale bar: 20 μm. (B) Phase contrast micrographs of the MTT-stained colonies formed by UM cell lines from (A). Scale bar: 20 μm. (C) Hematoxylin and eosin (H&E) staining for morphologic features and expression of MART-1 by immunohistochemistry on UM cell lines from (A). Dark brown indicates positive staining. Scale bar: 20 μm.
Table
 
Mitotic Rate and Immunostaining Data (Graded According to Intensity*) in UM Primary Tumors (PT) and Derived Cell Lines at Low (P3/P4) and High (P60) Passages
Table
 
Mitotic Rate and Immunostaining Data (Graded According to Intensity*) in UM Primary Tumors (PT) and Derived Cell Lines at Low (P3/P4) and High (P60) Passages
Figure 2
 
Immunohistochemical analyses of UM markers in the primary tumor 122 and its derived cell line T122 upon cell passaging. H&E staining, and expression of MART-1, HMB45, S100, VEGF, cyclin D1, and p16 markers monitored by immunohistochemistry in the UM primary tumor 122 as well as its derived cell line T122 at low (P4) and high (P60) passages. Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 2
 
Immunohistochemical analyses of UM markers in the primary tumor 122 and its derived cell line T122 upon cell passaging. H&E staining, and expression of MART-1, HMB45, S100, VEGF, cyclin D1, and p16 markers monitored by immunohistochemistry in the UM primary tumor 122 as well as its derived cell line T122 at low (P4) and high (P60) passages. Dark brown indicates positive staining. Scale bar: 20 μm.
We next determined by immunohistochemical analyses to which point expression of the UM (MART-1, S100 and HMB45), and vascular (VEGF) markers was affected by cell passaging in UM primary tumors 105, 122, 127, and 130, and their corresponding cell lines at passages P3/P4 and P60 (Fig. 2; Table). Expression of the p16 tumor suppressor and the G1/S cell cycle-specific protein cyclin D1 also was monitored as these proteins were reported to be frequently upregulated in UM.5557 As shown in Figure 2 (and summarized in the Table), the primary tumor 122 expressed all markers at varying levels, the most intense labeling being observed for MART-1, S100, and p16. Interestingly, expression of MART-1, S100, and p16, although detectable in more than 50% of T122 cells at a low passage (P3), became strongly reduced (for MART-1) or entirely lost (for S100 and p16) when T122 cells were grown to P60. On the other hand, VEGF and cyclin D1, whose expression was low at P3, sought their expression considerably increased at P60. Very similar results also were obtained with the T105, T127, and T130 cell lines (Table). Therefore, culturing UM cell lines at high passages also altered expression of many of these proteins compared to primary tumors. 
Influence of Cell Passaging on the Gene Expression Signatures of UM Cell Lines
Culturing cells directly from the primary tumor tissues also may force them to change their properties to survive in such a different environment. To evaluate such a possibility, we conducted gene expression profiling on microarray using total RNA prepared from UM cell lines T142 and T143 cultured at different passages (P4, P20, and P60), and compared their pattern of expressed genes to that of the original primary tumors. Scatter plot analysis revealed that the pattern of genes expressed by the 142 and 143 primary tumors changed dramatically as the cell lines are grown until passage P4 (Fig. 3A). Culturing T143 cells further from P4 to P60 did not lead to modifications as dramatic as those observed between the primary tumor and P4. On the other hand, the profile of expressed genes for T142 cells kept changing relative to the pattern observed in the primary tumor, as indicated by the variation in the slope of the regression curve (R2 = 0.476 at P4 to R2 = 0.193 at P60). Comparison of T142 and T143 transcriptional profiles between passages P4 and P20 (R2 = 0.768 for T142 and R2 = 0.661 for T143), as well as P20 versus P60 (R2 = 0.661 for T142 and R2 = 0.636 for T143) revealed a more modest global change in the patterns of genes expressed by these cells (Fig. 3B) indicating that most of the variations occurred between the uncultured primary tumor and passage P4. 
Figure 3
 
Gene expression profiling by DNA microarray. (A) Scatter plots of log2 signal intensity from 60,000 different targets covering the entire human transcriptome of UM primary tumors 142 and 143 (y-axis) plotted against their derived cell lines T142 and T143 (x-axis) grown at passages P4, P20, and P60. The identity and 2-fold change in intensity lines are shown in green. (B) Scatter plots of UM cell lines T142 and T143 (y-axis) at low passage (P4) plotted against T142 and T143 cells grown at higher passages (P20 and P60; x-axis). (C) Heatmap representation of the 4-fold or more variations for UM cell lines T142 (left) and T143 (right) cultured at passages P4, P20, and P60 against the primary tumor from which they were derived (UM tumors 142 and 143).
Figure 3
 
Gene expression profiling by DNA microarray. (A) Scatter plots of log2 signal intensity from 60,000 different targets covering the entire human transcriptome of UM primary tumors 142 and 143 (y-axis) plotted against their derived cell lines T142 and T143 (x-axis) grown at passages P4, P20, and P60. The identity and 2-fold change in intensity lines are shown in green. (B) Scatter plots of UM cell lines T142 and T143 (y-axis) at low passage (P4) plotted against T142 and T143 cells grown at higher passages (P20 and P60; x-axis). (C) Heatmap representation of the 4-fold or more variations for UM cell lines T142 (left) and T143 (right) cultured at passages P4, P20, and P60 against the primary tumor from which they were derived (UM tumors 142 and 143).
We next generated a heatmap for all the genes showing a 4-fold or more expression variation unique to each cell line at passages P4, P20, and P60 compared to its corresponding primary tumor (Fig. 3C). A total of 4986 (for T142) and 6771 (for T143) genes fitted into that category of differentially regulated genes. Consistent with the results from the scatter plots, the patterns of deregulated genes in the T143 cell line at passages P4, P20, and P60 were very similar to each other, unlike those obtained for the T142 cells that kept changing with increasing passages. Yet, the most striking change in these patterns was no doubt observed when T142 and T143 cells grown to P4 were compared to their corresponding primary tumor. Interestingly, a major proportion of genes expressed at low levels in the primary tumors (blue/light yellow) became highly expressed in their derived cell lines (orange/red). 
Influence of Cell Passaging on the Expression of UM Class 1 and 2 Markers
A gene expression signature comprising 12 genes that can discriminate between UM primary tumors at low (class 1 genes; FXR1, LTA4H, ID2, ROBO1, MTUS1, LMCD1, STAB1, and EIF1B) or high (class 2 genes; HTR2B, CDH1, RAB31, and ECM1) risk of evolving towards the formation of liver metastases has been described.58 Therefore, we examined whether the expression of these markers was affected by cell passaging in T142 and T143 cell lines when compared to their primary tumor. As shown in Figure 4A (also refer to Supplementary Table S4), expression of the FXR1, LTA4H, and ID2 genes was the least affected by cell passaging among the class 1 genes although they were most actively transcribed in the cell lines than in the primary tumors. On the other hand, expression of MTUS1, LMCD1, SATB1, and EIF1B genes, which were moderately to highly expressed in the primary tumors 142 and 143, was rapidly lost with cell passaging in both cell lines. Although class 2 genes were highly to moderately expressed in tumor 142, their expression dramatically decreased (with the exception of ECM1) with the passages in culture (Fig. 4B; Supplementary Table S4). The tumor 143 had high levels of CDH1 and RAB31 genes but only very low levels of HTR2B and ECM1 genes. Again, CDH1 and RAB31 gene expression was entirely lost at a very early passage (P4). The dramatic extinction of HTR2B expression at the transcriptional level in T142 cells (Figs. 4B–C; Supplementary Table S4) was further validated by qPCR (Fig. 4D; P < 0.001). However, these alterations did not necessarily translate into corresponding changes at the protein level as Western blot analysis detected less HTR2B protein, the most discriminant gene of the molecular signature, in the primary tumor than in the T142 cell line (Fig. 4E). Its expression, however, decreased in T142 cells as they are passaged up to P60. Therefore, although this 12-gene signature may prove useful as a tool to discriminate between classes 1 and 2 primary tumors, it is useless when cell lines are selected as the source of biologic material for these analyses. 
Figure 4
 
Influence of cell passaging on the expression of classes 1 and 2 genes. Heatmap representation of (A) class 1 and (B) class 2 genes expressed by the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes β-2-microglobulin (B2M), C-terminal binding protein 1 (CTBP1), and golgin A1 (GOLGA1; Ctrl). (C) Data obtained for the HTR2B gene in T142 cells presented as linear signals normalized to B2M. (D) Quantitative PCR analysis of HTR2B gene expression normalized to GAPDH in T142 cells grown at passages P4, P20, and P60. *P < 0.001. (E) Western blots conducted on protein extracts from UM primary tumors 142 and 143, as well as derived cell lines at passages P4, P20, and P60 using antibodies against HTR2B and actin (loading control).
Figure 4
 
Influence of cell passaging on the expression of classes 1 and 2 genes. Heatmap representation of (A) class 1 and (B) class 2 genes expressed by the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes β-2-microglobulin (B2M), C-terminal binding protein 1 (CTBP1), and golgin A1 (GOLGA1; Ctrl). (C) Data obtained for the HTR2B gene in T142 cells presented as linear signals normalized to B2M. (D) Quantitative PCR analysis of HTR2B gene expression normalized to GAPDH in T142 cells grown at passages P4, P20, and P60. *P < 0.001. (E) Western blots conducted on protein extracts from UM primary tumors 142 and 143, as well as derived cell lines at passages P4, P20, and P60 using antibodies against HTR2B and actin (loading control).
Influence of Cell Passaging on the Expression of Immune, Stromal and Melanocyte Signature Genes, as Well as on Other Metabolic Pathways
Specific signatures related to the infiltration of normal cells in tumor tissues have been described recently.52 Therefore, we examined whether the expression of these immune and stromal genes signatures (141 genes in each), as well as genes associated to melanocyte differentiation (13 genes), was affected by cell passaging in T142 and T143 cell lines when compared to their primary tumor (Fig. 5; Supplementary Table S4). Although immune and stromal genes were highly expressed in primary tumors 142 and 143, their expression dramatically decreased with cells passaging in culture (Figs. 5A–B; Supplementary Table S4). Our unsupervised bioinformatic analyses identified “Interferon signaling” (Reactome ID 913531), “Cytokine signaling in immune system” (Reactome ID 1280215), “Immune system” (Reactome ID 168256), “Class I MHC mediated antigen processing & presentation” (Reactome ID 983169), and “MHC class II antigen presentation” (Reactome ID 2132295) among the enriched pathways (Supplementary Table S5). Thus, pronounced differences between gene expression profiles of primary tumors and cell lines could be attributed partly to the absence of immune and stromal components in vitro. In addition, as shown in Figure 5C (also refer to Supplementary Table S4), melanocyte differentiation genes, such as melanin biosynthetic enzymes, are mostly repressed after cell passaging. Other enriched pathways found by unsupervised bioinformatic analyses (Supplementary Table S5) were associated with “Respiratory electron transport” (Reactome ID 163200), “Citric acid (TCA) cycle” (Reactome ID 1428517), “Mitochondrial translation” (Reactome ID 5368287), “Gene expression” (Reactome ID 74160), “Cell Cycle-Mitotic” (Reactome ID 69278), and “Beta-catenin independent WNT signaling” (Reactome ID 3858494). 
Figure 5
 
Influence of cell passaging on the expression of immune, stromal, and melanocyte gene signatures. Heatmap representation of (A) immune, (B) stromal, and (C) melanocyte gene signatures of the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes B2M, CTBP1, and GOLGA1 (Ctrl).
Figure 5
 
Influence of cell passaging on the expression of immune, stromal, and melanocyte gene signatures. Heatmap representation of (A) immune, (B) stromal, and (C) melanocyte gene signatures of the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes B2M, CTBP1, and GOLGA1 (Ctrl).
Cell Passaging Alters the Expression of Genes Previously Studied in UM and the Tumorigenic Properties of UM Cell Lines
We next addressed whether passaging UM cell lines also alters the expression of genes whose products are important to the properties of UM. Among them, we demonstrated that PARP-1 and PARG enzymes, as well as the integrin α5β1, may have critical functions in the tumorigenic properties of UM.35,51 As these three genes were strongly regulated by the transcription factor Sp1, we also included it into this analysis. In addition, we also monitored the expression of the melanosomal marker MART-1, as well as VEGF, cyclin D1, p16, and p27. Microarray analyses clearly indicated that expression of the gene encoding MART-1, elevated in the primary tumors 142 and 143, was lost very rapidly with cell passaging in T142 and T143 cell lines (Fig. 6A; Supplementary Table S4). Similar results also were obtained for PARP-1, PARG, cyclin D1, and p16 (only in T143 cells) genes. On the other hand, expression of the α5 integrin subunit encoding gene ITGA5, Sp1 and VEGF genes increased with cell passaging compared to their corresponding primary tumor. However, these alterations at the transcriptional level did not necessarily translate into corresponding changes at the protein level as revealed by Western blot analyses. Indeed, although the protein expression of the α5 integrin subunit, Sp1, and p27 matched quite well with their transcriptional profile, that of PARP-1 and PARG, as well as cyclin D1 was inversely correlated with their mRNA expression pattern (Fig. 6B). Consistent with the data from the microarrays, VEGF expression increased dramatically when cells reached P4 in culture (Fig. 6B). However, it then decreased rapidly from P4 to P20 and remained low at P60 in T142 and T143 cells. 
Figure 6
 
Influence of cell passaging on the expression of UM markers and the tumorigenic properties of UM cell lines. (A) The pattern of expression of genes typically reported as important UM markers is shown in a heatmap for the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data for the housekeeping genes B2M, CTBP1, and GOLGA1 also are shown (Ctrl). (B) Western blots conducted on protein extracts from UM primary tumors 105, 142, and 143 and their derived cell lines using antibodies against PARP-1, PARG, integrin α5 subunit, Sp1, VEGF, cyclin D1, p27, and actin (loading control). (C) Scatter plots depicting the tumor volume (mm3) produced in athymic nude mice injected subcutaneously with the UM cell lines T122, T130, T142, and T143 cultured at low (P4/P6), intermediate (P20/P21), or high (P60/P62) passages (line, mean). The number of tumors relative to the number of injected sites is indicated below the graph. *P < 0.05.
Figure 6
 
Influence of cell passaging on the expression of UM markers and the tumorigenic properties of UM cell lines. (A) The pattern of expression of genes typically reported as important UM markers is shown in a heatmap for the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data for the housekeeping genes B2M, CTBP1, and GOLGA1 also are shown (Ctrl). (B) Western blots conducted on protein extracts from UM primary tumors 105, 142, and 143 and their derived cell lines using antibodies against PARP-1, PARG, integrin α5 subunit, Sp1, VEGF, cyclin D1, p27, and actin (loading control). (C) Scatter plots depicting the tumor volume (mm3) produced in athymic nude mice injected subcutaneously with the UM cell lines T122, T130, T142, and T143 cultured at low (P4/P6), intermediate (P20/P21), or high (P60/P62) passages (line, mean). The number of tumors relative to the number of injected sites is indicated below the graph. *P < 0.05.
To evaluate to which point cell passaging may alter the tumorigenic properties of cell lines, T122, T130, T142, and T143 cells grown at passages P4/P6, P20, and P60 were injected into athymic mice and tumor formation was monitored over a 28-day period (Fig. 6C). At passage P4/P6, the T122 and T130 cell lines were by far the most tumorigenic as all injected mice developed detectable tumors (12/12 and 11/12 tumors, with a mean tumor volume of 127 and 110 mm3, respectively). Mice injected with T142 and T143 cells also yielded detectable tumors (both 4/12 tumors, with a mean tumor volume of 5 and 21 mm3, respectively). Most interestingly, the mean tumor volume for all cell lines but T130 increased considerably when P20 or P60 cells were injected. The results were statistically significant for T122 (P4 vs. P20, P < 0.05; P4 vs. P60, P < 0.005), T142 (P4 vs. P20, P < 0.05), and T143 (P6 vs P60, P < 0.01) cells. 
Discussion
Although cell lines maintain expression of some hallmarks of cancer, most exhibit reduced or altered key functions and often no longer represent reliable models of their original source material after several passages in culture.1,59 To the best of our knowledge, no studies have been performed to examine at the molecular level the similarities between UM primary tumor surgical samples and their derived cell lines at short- and long-term passage in culture. In the present study, we demonstrated that cell passaging causes, at the very first passages, dramatic changes in genes expressed in UM cells, including most of those that belong to the prognostic signature. Changes also occurred in their growth properties (evaluated through the measurement of the cell doubling time) as well as in their ability to produce tumors when injected in athymic nude mice, with increased tumorigenic properties at higher cell passages. 
Through immunohistologic analyses, we found that expression of the melanosomal protein MART-1 decreased after short-term passages. This is consistent with the observed decrease in the melanin content during cell passaging in the majority of UM cell lines,35,60,61 as well as our previous demonstration of a loss of melanin biosynthetic enzymes in cultured UM.62,63 It might provide a growth advantage by reducing the energy and oxidative stress associated with pigment production.64 Vascular endothelial growth factor, cyclin D1, and p16 have been studied widely in UM by our group.61,65,66 In all our UM cell lines, expression of p16 decreased in cells grown at high passage compared to primary tumors, and in contrast, expression of cyclin D1 remained stable upon cell passaging compared to primary tumors. Expression of VEGF was variable according to passage or cell line. These results suggested that selection of cell lines that best mirror UMs will optimize the clinical relevance of future discoveries. 
One major problem of tissue culture is that cell lines evolve as they are grown continuously over a large number of passages. In fact, genotypic and phenotypic variants constantly appear in cancer cell lines, and selection of subpopulations with superior growth properties occurs frequently.67 This may change drastically the characteristics of a cell line, as shown in our study with the UM cell lines examined. For instance, at high passage, the Caco2 colon cancer cell line presents phenotypic changes affecting morphology, expression of the glucose transporters, metabolic activity, and transepithelial electrical resistance.68 Passage-related changes that lead to altered androgen response and prostate-specific antigen secretion also were observed in the prostatic LNCaP cell line. Most importantly, a previous comparison of gene expression patterns between several normal tissues and corresponding tumors and cell lines demonstrated a greater discrepancy in the number of differentially expressed genes between the cell line and tumor than between the tumor and normal tissue.3 In our study, such discrepancy depends on passage, cell line, or markers studied. Our research indicates that loss of the immune and stromal components during cell passaging also accounts for the transcriptional difference noted between cell lines and tumors. 
Genomic instability is a characteristic feature of many tumors and it may develop early during pathogenesis.69 It persists after tumor development and results in appearance of multiple subclonal populations.70 As cell lines have shorter population doubling time (hours) than tumors (months), it is expected that cell lines acquire more genetic and epigenetic changes over the same period. In 2004, Onken et al.58 confirmed the presence of two molecular classes of UMs using gene expression profiling, in which the 8-year survival probability was 95% in class 1 patients versus 31% in class 2 patients. In our study, we investigated whether monitoring classes 1 and 2 genes would be as informative for UM cell lines as they are for the primary tumors. However, our results clearly showed that as cells are grown in culture, the class 2 phenotype of tumor 142 is lost in favor of more differentiated traits, a type of in vitro transition that would result in greater transcriptional distance between the class 2 cell line and its respective tumor. 
Consistent with similar studies conducted with other types of cancers,1,67,68 our results provided evidence that short-term UM primary cultures show characteristics much closer to the primary tumors than to long-term cultures. This is most likely due to the fact that low-passage primary cultures still support a culture population made up of different UM cell subtypes that are genetically distinct from one another. Consequently, some investigators have preferred to work with short-term primary cultures to circumvent the limitations of continuous cell lines, but the former cannot be reliably shared between investigative groups, making it difficult to obtain independent validation of results using common biologic samples.32 However, we found that genetic instability also occurred in short-term passages in UM cell lines. Moreover, the genetic alterations seemed to be more significant at early passages than at late passages, which is reflected by the high variability of the STR profile observed for most of our UM cell lines at low cell passages. However, such a high microsatellite instability is not uncommon and has been reported for other types of cancer cell lines.7174 Interestingly, unstable STR has been shown to be the consequence of alterations in the expression/activity of several DNA mismatch repair enzymes, including MLH1, MSH2, MSH6, and PMS2.71 Therefore, examination of the expression status for these proteins should prove very interesting in deciphering whether some of them contribute to the genomic instability noted in our UM cell lines. 
The major limitations of cancer cell lines are their lack of stromal, vascular, and immune components.4 Tumors are, indeed, composed of an ecosystem of evolving clones, competing or cooperating with each other and with other normal cells infiltrating their microenvironment.75 This raises the intriguing question of whether these clones have been selected through their growth into culture medium or over time through cell passaging. Thus, cell lines derived from one clone are not necessarily representative of the diversity found in the original tumor. These deficiencies have been partially overcome by xenograft transplantation of tumor fragments or tumor cell line suspensions into immunocompromised mice.7678 However, xenografts also have certain limitations, including the presence of stromal components of xenogeneic origin that support the growth of human cancer cells. 
Interestingly, despite often dramatic variations in gene expression profiles between UM primary tumors and their derived cell lines, the protein products showed moderate modifications by Western blotting. Recently, we demonstrated the critical functions played by the PARP-1 and PARG enzymes on the tumorigenic properties of UM.51 In the present study, the mRNA transcripts encoding PARP-1 and PARG decreased to very low levels when tumors 142 and 143 were passaged in culture. In contrast, expression of PARP-1 and PARG proteins increased, suggesting that they might negatively regulate transcription of their own respective genes through mechanisms involving negative feedback loops as suggested recently by Vidakovic et al.79 for the mouse PARP-1 gene. 
In conclusion, our analyses provided valuable information relevant to the suitability of cell lines in modeling UM primary tumor heterogeneity. Because it is difficult to obtain sufficient biologic material from metastases, we could not derive cell lines from liver metastatic lesions of UM patients. We concede, therefore, that their characteristics may considerably differ from those of primary tumors. The goal of our study was by no mean to question the data previously obtained using UM long-term cultured cell lines. It is obvious that cell lines may represent valuable tools in functional assays, as well as pharmacologic and genetic studies to better understand the numerous cellular and molecular mechanisms involved in UM development and progression. However, we must be aware of the limitations (such as their elevated genomic instability at low cell passages) of using such cell lines as biologic materials. It remains important to define to which point the in vitro model closely resembles particular features of the original tumor and whether these cellular and molecular properties are maintained in basic culture conditions for primary and metastatic cell lines. 
Acknowledgments
The authors thank Dan Bergeron, Alain Rousseau, and Marcelle Giasson for clinical follow-up data; Daniel O. Black and Yvonne Molgat for the UM tumor surgical samples; the nurses Marthe Mercier, Sylvie Marcoux, Suzanne Boulianne, Louise Monroe, and Jocelyne Boivin for the recruitment of patients; Renée Paradis, Florence Pagé-Larivière, Julie Bérubé, Michèle Orain, Véronique Haddad, Steven Richebourg, and Patrick Laplante for their technical assistance. 
Supported by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC) (SLG, #138624-2012) and from the Fondation du CHU de Québec (SLG/SL); by a research scholarship from the Fonds de Recherche du Québec - Santé (FRQS; SL); a scholarship from the Vision Health Research Network (VHRN; MAB); and by scholarships from the Faculté de médecine de l'Université Laval, the Fondation du CHU de Québec and the Centre de recherche en organogénèse expérimentale de l'Université Laval/LOEX (CW). The Québec Uveal Melanoma Infrastructure is supported by the VHRN from the FRQS. 
Disclosure: F. Mouriaux, None; K. Zaniolo, None; M.-A. Bergeron, None; C. Weidmann, None; A. De La Fouchardière, None; F. Fournier, None; A. Droit, None; M.W. Morcos, None; S. Landreville, None; S.L. Guérin None 
References
van Staveren WC, Solis DY, Hebrant A, Detours V, Dumont JE, Maenhaut C. Human cancer cell lines: experimental models for cancer cells in situ? For cancer stem cells? Biochim Biophys Acta. 2009; 1795: 92–103.
Borrell B. How accurate are cancer cell lines? Nature. 2010; 463: 858.
Ertel A, Verghese A, Byers SW, Ochs M, Tozeren A. Pathway-specific differences between tumor cell lines and normal and tumor tissue cells. Molec Cancer. 2006; 5: 55.
Gazdar AF, Gao B, Minna JD. Lung cancer cell lines: useless artifacts or invaluable tools for medical science? Lung Cancer. 2010; 68: 309–318.
Millau JF, Mai S, Bastien N, Drouin R. p53 functions and cell lines: have we learned the lessons from the past? BioEssays. 2010; 32: 392–400.
Esquenet M, Swinnen JV, Heyns W, Verhoeven G. LNCaP prostatic adenocarcinoma cells derived from low and high passage numbers display divergent responses not only to androgens but also to retinoids. J Steroid Biochem Molec Biol. 1997; 62: 391–399.
Hill RP. Identifying cancer stem cells in solid tumors: case not proven. Cancer Res. 2006; 66: 1891–1895; discussion 1890.
Kao J, Salari K, Bocanegra M, et al. Molecular profiling of breast cancer cell lines defines relevant tumor models and provides a resource for cancer gene discovery. PLoS One. 2009; 4: e6146.
Singh AD, Topham A. Incidence of uveal melanoma in the United States: 1973-1997. Ophthalmology. 2003; 110: 956–961.
Kujala E, Makitie T, Kivela T. Very long-term prognosis of patients with malignant uveal melanoma. Invest Ophthalmol Vis Sci. 2003; 44: 4651–4659.
Singh AD, Shields CL, Shields JA. Prognostic factors in uveal melanoma. Melanoma Res. 2001; 11: 255–263.
Prescher G, Bornfeld N, Hirche H, Horsthemke B, Jockel KH, Becher R. Prognostic implications of monosomy 3 in uveal melanoma. Lancet. 1996; 347: 1222–1225.
White VA, Chambers JD, Courtright PD, Chang WY, Horsman DE. Correlation of cytogenetic abnormalities with the outcome of patients with uveal melanoma. Cancer. 1998; 83: 354–359.
Harbour JW, Onken MD, Roberson ED, et al. Frequent mutation of BAP1 in metastasizing uveal melanomas. Science. 2010; 330: 1410–1413.
Harbour JW. A prognostic test to predict the risk of metastasis in uveal melanoma based on a 15-gene expression profile. Methods Mol Biol. 2014; 1102: 427–440.
Albert DM, Ruzzo MA, McLaughlin MA, Robinson NL, Craft JL, Epstein J. Establishment of cell lines of uveal melanoma. Methodology and characteristics. Invest Ophthalmol Vis Sci. 1984; 25: 1284–1299.
Kan-Mitchell J, Mitchell MS, Rao N, Liggett PE. Characterization of uveal melanoma cell lines that grow as xenografts in rabbit eyes. Invest Ophthalmol Vis Sci. 1989; 30: 829–834.
Ksander BR, Rubsamen PE, Olsen KR, Cousins SW, Streilein JW. Studies of tumor-infiltrating lymphocytes from a human choroidal melanoma. Invest Ophthalmol Vis Sci. 1991; 32: 3198–3208.
Aubert C, Rouge F, Reillaudou M, Metge P. Establishment and characterization of human ocular melanoma cell lines. Int J Cancer. 1993; 54: 784–792.
Massarelli R, Belkhou R, Dunel-Erb S, et al. [Morphological characterization of cell lines of human uveal melanoma]. C R Acad Sci III. 1994; 317: 25–33.
De Waard-Siebinga I, Blom DJ, Griffioen M, et al. Establishment and characterization of an uveal-melanoma cell line. Int J Cancer. 1995; 62: 155–161.
Soulieres D, Rousseau A, Tardif M, et al. The radiosensitivity of uveal melanoma cells and the cell survival curve. Graefes Arch Clin Exp Ophthalmol. 1995; 233: 85–89.
Ma D, Luyten GP, Luider TM, Niederkorn JY. Relationship between natural killer cell susceptibility and metastasis of human uveal melanoma cells in a murine model. Invest Ophthalmol Vis Sci. 1995; 36: 435–441.
Luyten GP, Naus NC, Mooy CM, et al. Establishment and characterization of primary and metastatic uveal melanoma cell lines. Int J Cancer. 1996; 66: 380–387.
Chen PW, Murray TG, Uno T, Salgaller ML, Reddy R, Ksander BR. Expression of MAGE genes in ocular melanoma during progression from primary to metastatic disease. Clin Exp Metastasis. 1997; 15: 509–518.
Verbik DJ, Murray TG, Tran JM, Ksander BR. Melanomas that develop within the eye inhibit lymphocyte proliferation. Int J Cancer. 1997; 73: 470–478.
Maniotis AJ, Folberg R, Hess A, et al. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry. Am J Pathol. 1999; 155: 739–752.
Beliveau A, Berube M, Rousseau A, Pelletier G, Guerin SL. Expression of integrin alpha5beta1 and MMPs associated with epithelioid morphology and malignancy of uveal melanoma. Invest Ophthalmol Vis Sci. 2000; 41: 2363–2372.
Seftor EA, Meltzer PS, Kirschmann DA, et al. Molecular determinants of human uveal melanoma invasion and metastasis. Clin Exp Metastasis. 2002; 19: 233–246.
Labialle S, Dayan G, Gambrelle J, et al. Characterization of the typical multidrug resistance profile in human uveal melanoma cell lines and in mouse liver metastasis derivatives. Melanoma Res. 2005; 15: 257–266.
Nareyeck G, Zeschnigk M, Prescher G, Lohmann DR, Anastassiou G. Establishment and characterization of two uveal melanoma cell lines derived from tumors with loss of one chromosome 3. Exp Eye Res. 2006; 83: 858–864.
Folberg R, Kadkol SS, Frenkel S, et al. Authenticating cell lines in ophthalmic research laboratories. Invest Ophthalmol Vis Sci. 2008; 49: 4697–4701.
Nareyeck G, Zeschnigk M, Bornfeld N, Anastassiou G. Novel cell lines derived by long-term culture of primary uveal melanomas. Ophthalmologica. 2009; 223: 196–201.
Burgess BL, Rao NP, Eskin A, Nelson SF, McCannel TA. Characterization of three cell lines derived from fine needle biopsy of choroidal melanoma with metastatic outcome. Mol Vis. 2011; 17: 607–615.
Landreville S, Vigneault F, Bergeron MA, et al. Suppression of alpha5 gene expression is closely related to the tumorigenic properties of uveal melanoma cell lines. Pigment Cell Melanoma Res. 2011; 24: 643–655.
Griewank KG, Yu X, Khalili J, et al. Genetic and molecular characterization of uveal melanoma cell lines. Pigment Cell Melanoma Res. 2012; 25: 182–187.
Suesskind D, Gauss S, Faust UE, et al. Characterisation of novel uveal melanoma cell lines under serum-free conditions. Graefes Arch Clin Exp Ophthalmol. 2013; 251: 2063–2070.
Amirouchene-Angelozzi N, Nemati F, Gentien D, et al. Establishment of novel cell lines recapitulating the genetic landscape of uveal melanoma and preclinical validation of mTOR as a therapeutic target. Mol Oncol. 2014; 8: 1508–1520.
Angi M, Versluis M, Kalirai H. Culturing uveal melanoma cells. Ocul Oncol Pathol. 2015; 1: 126–132.
Lefevre G, Glotin AL, Calipel A, et al. Roles of stem cell factor/c-Kit and effects of Glivec/STI571 in human uveal melanoma cell tumorigenesis. J Biol Chem. 2004; 279: 31769–31779.
Pereira PR, Odashiro AN, Marshall JC, Correa ZM, Belfort RJr, Burnier MNJr. The role of c-kit and imatinib mesylate in uveal melanoma. J Carcinog. 2005; 4: 19.
Penel N, Delcambre C, Durando X, et al. O-Mel-Inib: a Cancero-pole Nord-Ouest multicenter phase II trial of high-dose imatinib mesylate in metastatic uveal melanoma. Invest New Drugs. 2008; 26: 561–565.
Hofmann UB, Kauczok-Vetter CS, Houben R, Becker JC. Overexpression of the KIT/SCF in uveal melanoma does not translate into clinical efficacy of imatinib mesylate. Clin Cancer Res. 2009; 15: 324–329.
Calipel A, Landreville S, De La Fouchardiere A, et al. Mechanisms of resistance to imatinib mesylate in KIT-positive metastatic uveal melanoma. Clin Exp Metastasis. 2014; 31: 553–564.
Wistuba II, Bryant D, Behrens C, et al. Comparison of features of human lung cancer cell lines and their corresponding tumors. Clin Cancer Res. 1999; 5: 991–1000.
Lee J, Kotliarova S, Kotliarov Y, et al. Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell. 2006; 9: 391–403.
Wang H, Huang S, Shou J, et al. Comparative analysis and integrative classification of NCI60 cell lines and primary tumors using gene expression profiling data. BMC Genomics. 2006; 7: 166.
Domcke S, Sinha R, Levine DA, Sander C, Schultz N. Evaluating cell lines as tumour models by comparison of genomic profiles. Nat Commun. 2013; 4: 2126.
Vincent KM, Findlay SD, Postovit LM. Assessing breast cancer cell lines as tumour models by comparison of mRNA expression profiles. Breast Cancer Res. 2015; 17: 114.
Mouriaux F, Sanschagrin F, Diorio C, et al. Increased HIF-1alpha expression correlates with cell proliferation and vascular markers CD31 and VEGF-A in uveal melanoma. Invest Ophthalmol Vis Sci. 2014; 55: 1277–1283.
Molloy-Simard V, St-Laurent JF, Vigneault F, et al. Altered expression of the poly(ADP-ribosyl)ation enzymes in uveal melanoma and regulation of PARG gene expression by the transcription factor ERM. Invest Ophthalmol Vis Sci. 2012; 53: 6219–6231.
Yoshihara K, Shahmoradgoli M, Martinez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013; 4: 2612.
Fabregat A, Sidiropoulos K, Garapati P, et al. The Reactome pathway knowledgebase. Nucleic Acids Res. 2016; 44: D481–487.
Parson W, Kirchebner R, Muhlmann R, et al. Cancer cell line identification by short tandem repeat profiling: power and limitations. FASEB J. 2005; 19: 434–436.
Babchia N, Calipel A, Mouriaux F, Faussat AM, Mascarelli F. The PI3K/Akt and mTOR/P70S6K signaling pathways in human uveal melanoma cells: interaction with B-Raf/ERK. Invest Ophthalmol Vis Sci. 2010; 51: 421–429.
Coupland SE, Bechrakis N, Schuler A, et al. Expression patterns of cyclin D1 and related proteins regulating G1-S phase transition in uveal melanoma and retinoblastoma. Br J Ophthalmol. 1998; 82: 961–970.
Pardo M, Pineiro A, de la Fuente M, et al. Abnormal cell cycle regulation in primary human uveal melanoma cultures. J Cell Biochem. 2004; 93: 708–720.
Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res. 2004; 64: 7205–7209.
Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011; 144: 646–674.
Hu DN, Wakamatsu K, Ito S, McCormick SA. Comparison of eumelanin and pheomelanin content between cultured uveal melanoma cells and normal uveal melanocytes. Melanoma Res. 2009; 19: 75–79.
Mouriaux F, Casagrande F, Pillaire MJ, Manenti S, Malecaze F, Darbon JM. Differential expression of G1 cyclins and cyclin-dependent kinase inhibitors in normal and transformed melanocytes. Invest Ophthalmol Vis Sci. 1998; 39: 876–884.
Bergeron MA, Champagne S, Gaudreault M, Deschambeault A, Landreville S. Repression of genes involved in melanocyte differentiation in uveal melanoma. Mol Vis. 2012; 18: 1813–1822.
Landreville S, Lupien CB, Vigneault F, et al. Identification of differentially expressed genes in uveal melanoma using suppressive subtractive hybridization. Mol Vis. 2011; 17: 1324–1333.
Levy C, Khaled M, Fisher DE. MITF: master regulator of melanocyte development and melanoma oncogene. Trends Mol Med. 2006; 12: 406–414.
Mouriaux F, Kherrouche Z, Maurage CA, Demailly FX, Labalette P, Saule S. Expression of the c-kit receptor in choroidal melanomas. Melanoma Res. 2003; 13: 161–166.
Mouriaux F, Maurage CA, Labalette P, Sablonniere B, Malecaze F, Darbon JM. Cyclin-dependent kinase inhibitory protein expression in human choroidal melanoma tumors. Invest Ophthalmol Vis Sci. 2000; 41: 2837–2843.
Masters JR. Human cancer cell lines: fact and fantasy. Nat Rev Molec Cell Biol. 2000; 1: 233–236.
Hughes P, Marshall D, Reid Y, Parkes H, Gelber C. The costs of using unauthenticated over-passaged cell lines: how much more data do we need? BioTechniques. 2007; 43: 575, 577–578, 581–572 passim.
Loeb LA. Mutator phenotype may be required for multistage carcinogenesis. Cancer Res. 1991; 51: 3075–3079.
Nowell PC. The clonal evolution of tumor cell populations. Science. 1976; 194: 23–28.
Much M, Buza N, Hui P. Tissue identity testing of cancer by short tandem repeat polymorphism: pitfalls of interpretation in the presence of microsatellite instability. Hum Pathol. 2014; 45: 549–555.
Pereira S, Vayntrub T, Hiraki DD, et al. Short tandem repeat and human leukocyte antigen mutations or losses confound engraftment and typing analysis in hematopoietic stem cell transplants. Hum Immunol. 2011; 72: 503–509.
Rubocki RJ, Duffy KJ, Shepard KL, McCue BJ, Shepherd SJ, Wisecarver JL. Loss of heterozygosity detected in a short tandem repeat (STR) locus commonly used for human DNA identification. J Forensic Sci. 2000; 45: 1087–1089.
Somaschini A, Amboldi N, Nuzzo A, et al. Cell line identity finding by fingerprinting an optimized resource for short tandem repeat profile authentication. Genet Test Mol Biomarkers. 2013; 17: 254–259.
Merlo LM, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and ecological process. Nat Rev Cancer. 2006; 6: 924–935.
Cespedes MV, Casanova I, Parreno M, Mangues R. Mouse models in oncogenesis and cancer therapy. Clin Translat Oncol. 2006; 8: 318–329.
Laurent C, Gentien D, Piperno-Neumann S, et al. Patient-derived xenografts recapitulate molecular features of human uveal melanomas. Mol Oncol. 2013; 7: 625–636.
Nemati F, Sastre-Garau X, Laurent C, et al. Establishment and characterization of a panel of human uveal melanoma xenografts derived from primary and/or metastatic tumors. Clin Cancer Res. 2010; 16: 2352–2362.
Vidakovic M, Gluch A, Qiao J, et al. PARP-1 expression in the mouse is controlled by an autoregulatory loop: PARP-1 binding to an upstream S/MAR element and to a novel recognition motif in its promoter suppresses transcription. J Molec Biol. 2009; 388: 730–750.
Figure 1
 
Characteristics of cell lines derived from UM primary tumors. (A) Phase contrast micrographs of monolayer cultures of normal uveal melanocytes (UVM) or UM cell lines T105, T122, T127, T130, T142, and T143. Scale bar: 20 μm. (B) Phase contrast micrographs of the MTT-stained colonies formed by UM cell lines from (A). Scale bar: 20 μm. (C) Hematoxylin and eosin (H&E) staining for morphologic features and expression of MART-1 by immunohistochemistry on UM cell lines from (A). Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 1
 
Characteristics of cell lines derived from UM primary tumors. (A) Phase contrast micrographs of monolayer cultures of normal uveal melanocytes (UVM) or UM cell lines T105, T122, T127, T130, T142, and T143. Scale bar: 20 μm. (B) Phase contrast micrographs of the MTT-stained colonies formed by UM cell lines from (A). Scale bar: 20 μm. (C) Hematoxylin and eosin (H&E) staining for morphologic features and expression of MART-1 by immunohistochemistry on UM cell lines from (A). Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 2
 
Immunohistochemical analyses of UM markers in the primary tumor 122 and its derived cell line T122 upon cell passaging. H&E staining, and expression of MART-1, HMB45, S100, VEGF, cyclin D1, and p16 markers monitored by immunohistochemistry in the UM primary tumor 122 as well as its derived cell line T122 at low (P4) and high (P60) passages. Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 2
 
Immunohistochemical analyses of UM markers in the primary tumor 122 and its derived cell line T122 upon cell passaging. H&E staining, and expression of MART-1, HMB45, S100, VEGF, cyclin D1, and p16 markers monitored by immunohistochemistry in the UM primary tumor 122 as well as its derived cell line T122 at low (P4) and high (P60) passages. Dark brown indicates positive staining. Scale bar: 20 μm.
Figure 3
 
Gene expression profiling by DNA microarray. (A) Scatter plots of log2 signal intensity from 60,000 different targets covering the entire human transcriptome of UM primary tumors 142 and 143 (y-axis) plotted against their derived cell lines T142 and T143 (x-axis) grown at passages P4, P20, and P60. The identity and 2-fold change in intensity lines are shown in green. (B) Scatter plots of UM cell lines T142 and T143 (y-axis) at low passage (P4) plotted against T142 and T143 cells grown at higher passages (P20 and P60; x-axis). (C) Heatmap representation of the 4-fold or more variations for UM cell lines T142 (left) and T143 (right) cultured at passages P4, P20, and P60 against the primary tumor from which they were derived (UM tumors 142 and 143).
Figure 3
 
Gene expression profiling by DNA microarray. (A) Scatter plots of log2 signal intensity from 60,000 different targets covering the entire human transcriptome of UM primary tumors 142 and 143 (y-axis) plotted against their derived cell lines T142 and T143 (x-axis) grown at passages P4, P20, and P60. The identity and 2-fold change in intensity lines are shown in green. (B) Scatter plots of UM cell lines T142 and T143 (y-axis) at low passage (P4) plotted against T142 and T143 cells grown at higher passages (P20 and P60; x-axis). (C) Heatmap representation of the 4-fold or more variations for UM cell lines T142 (left) and T143 (right) cultured at passages P4, P20, and P60 against the primary tumor from which they were derived (UM tumors 142 and 143).
Figure 4
 
Influence of cell passaging on the expression of classes 1 and 2 genes. Heatmap representation of (A) class 1 and (B) class 2 genes expressed by the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes β-2-microglobulin (B2M), C-terminal binding protein 1 (CTBP1), and golgin A1 (GOLGA1; Ctrl). (C) Data obtained for the HTR2B gene in T142 cells presented as linear signals normalized to B2M. (D) Quantitative PCR analysis of HTR2B gene expression normalized to GAPDH in T142 cells grown at passages P4, P20, and P60. *P < 0.001. (E) Western blots conducted on protein extracts from UM primary tumors 142 and 143, as well as derived cell lines at passages P4, P20, and P60 using antibodies against HTR2B and actin (loading control).
Figure 4
 
Influence of cell passaging on the expression of classes 1 and 2 genes. Heatmap representation of (A) class 1 and (B) class 2 genes expressed by the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes β-2-microglobulin (B2M), C-terminal binding protein 1 (CTBP1), and golgin A1 (GOLGA1; Ctrl). (C) Data obtained for the HTR2B gene in T142 cells presented as linear signals normalized to B2M. (D) Quantitative PCR analysis of HTR2B gene expression normalized to GAPDH in T142 cells grown at passages P4, P20, and P60. *P < 0.001. (E) Western blots conducted on protein extracts from UM primary tumors 142 and 143, as well as derived cell lines at passages P4, P20, and P60 using antibodies against HTR2B and actin (loading control).
Figure 5
 
Influence of cell passaging on the expression of immune, stromal, and melanocyte gene signatures. Heatmap representation of (A) immune, (B) stromal, and (C) melanocyte gene signatures of the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes B2M, CTBP1, and GOLGA1 (Ctrl).
Figure 5
 
Influence of cell passaging on the expression of immune, stromal, and melanocyte gene signatures. Heatmap representation of (A) immune, (B) stromal, and (C) melanocyte gene signatures of the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data also are presented for the housekeeping genes B2M, CTBP1, and GOLGA1 (Ctrl).
Figure 6
 
Influence of cell passaging on the expression of UM markers and the tumorigenic properties of UM cell lines. (A) The pattern of expression of genes typically reported as important UM markers is shown in a heatmap for the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data for the housekeeping genes B2M, CTBP1, and GOLGA1 also are shown (Ctrl). (B) Western blots conducted on protein extracts from UM primary tumors 105, 142, and 143 and their derived cell lines using antibodies against PARP-1, PARG, integrin α5 subunit, Sp1, VEGF, cyclin D1, p27, and actin (loading control). (C) Scatter plots depicting the tumor volume (mm3) produced in athymic nude mice injected subcutaneously with the UM cell lines T122, T130, T142, and T143 cultured at low (P4/P6), intermediate (P20/P21), or high (P60/P62) passages (line, mean). The number of tumors relative to the number of injected sites is indicated below the graph. *P < 0.05.
Figure 6
 
Influence of cell passaging on the expression of UM markers and the tumorigenic properties of UM cell lines. (A) The pattern of expression of genes typically reported as important UM markers is shown in a heatmap for the UM primary tumors 142 and 143, and their derived cell lines T142 and T143 cultured at passages P4, P20, and P60 as determined by DNA microarrays. Data for the housekeeping genes B2M, CTBP1, and GOLGA1 also are shown (Ctrl). (B) Western blots conducted on protein extracts from UM primary tumors 105, 142, and 143 and their derived cell lines using antibodies against PARP-1, PARG, integrin α5 subunit, Sp1, VEGF, cyclin D1, p27, and actin (loading control). (C) Scatter plots depicting the tumor volume (mm3) produced in athymic nude mice injected subcutaneously with the UM cell lines T122, T130, T142, and T143 cultured at low (P4/P6), intermediate (P20/P21), or high (P60/P62) passages (line, mean). The number of tumors relative to the number of injected sites is indicated below the graph. *P < 0.05.
Table
 
Mitotic Rate and Immunostaining Data (Graded According to Intensity*) in UM Primary Tumors (PT) and Derived Cell Lines at Low (P3/P4) and High (P60) Passages
Table
 
Mitotic Rate and Immunostaining Data (Graded According to Intensity*) in UM Primary Tumors (PT) and Derived Cell Lines at Low (P3/P4) and High (P60) Passages
Supplement 1
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×