Citation
Zuidervaart, W. (2005, May 25). Genomic and proteomic analysis in uveal melanoma. Retrieved from https://hdl.handle.net/1887/2696
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in theInstitutional Repository of the University of Leiden Downloaded from: https://hdl.handle.net/1887/2696
4
E
PROFILING
REVEALS THAT
METHYLATION OF
TIMP3
IS
INVOLVED IN
UVEAL
MELANOMA
DEVELOPMENT
Pieter A. van der Velden1, Wieke Zuidervaart2, Monique H.M.H. Hurks2, Sandra Pavey5, Bruce R. Ksander4, Elise Krijgsman2, Rune R. Frants3, Cornelis P.Tensen1, Rein Willemze1, Martine J. Jager2, Nelleke A. Gruis1
1Department of Dermatology, 2Department of Ophthalmology, 3Center for Human and Clinical Genetics section Human Genetics, Leiden University Medical Center, Leiden, The Netherlands, 4Schepens Eye Research Institute and the Department of
Ophthalmology, Harvard Medical School, Boston, Massachusetts, 5Human Genetics Laboratory, Queensland Institute of Medical Research, Brisbane, Australia
A
BSTRACTUveal melanoma is associated with a high tumor-related mortality due to the propensity to develop metastases. The mechanisms that are responsible for malignant dissemination are largely unknown and need to be explored in order to facilitate diagnosis and treatment of metastases. To identify molecules involved in dissemination, we used cell lines derived from a primary uveal melanoma and two liver metastases from the same patient as a model for tumor progression. Using a microarray representing 1176 genes, we identified 63
differentially expressed genes. Forty genes showed a higher expression and 23 showed a lower expression in the primary cell line compared to the metastasis cell lines. These genes are involved in processes like angiogenesis, apoptosis, macrophage stimulation, and
I
NTRODUCTIONMelanoma of the uvea is the most common primary intra-ocular malignancy in adults, with a late, but frequent occurrence of metastases. The 5, 10, and 15 year survival rates, based on tumor-related mortality, are reported to be 72%, 59%, and 53%, respectively (Diener-West et al., 1992; Gamel et al., 1993; McLean, 1993). Metastases may be present at the time that the primary tumor is identified, and early recognition and treatment of metastases may improve survival. Identification of the mechanisms that play a role in metastasis formation and identification of prognostic markers may help to select those patients that have the highest chance to develop metastases.
Known prognostic markers for uveal melanoma are cell type, largest tumor diameter, mitotic rate, monosomy 3, and the presence of specific fibrovascular patterns (Mooy and De Jong, 1996). Vascular density was also identified by some authors as a prognostic marker (Foss et al., 1996). Recently, infiltration of uveal melanoma by macrophages was recognized as a prognostic factor, which might be related to microvessel formation (Makitie et al., 2001). The presence or absence of certain surface molecules may play a role in different phases of
metastasis formation: e.g. HLA expression is probably important during hematogenous spreading of metastases, where tumor cells that express HLA Class I are less susceptible to natural killer cell-mediated killing and are subsequently able to settle in the liver (Ma et al., 1995). On the other hand, molecules such as epidermal growth factor receptor may play a role in the growth of uveal melanoma metastases in the liver by determining the sensitivity of metastatic cells to hepatocyte growth factor (Hurks et al., 2000).
A new approach to identify molecules that play a role in metastasic disease is by microarray analysis, a fast and powerful method that simultaneously and quantitatively measures genome-wide gene expression (DeRisi et al., 1996). We set out to compare the expression profiles of a cell line obtained from a primary uveal melanoma with the profiles of two cell lines obtained from two liver metastases from the same patient. Because the cell lines are derived from one person, there is no genetic heterogeneity except for the genetic variations that occurred during tumor progression. In this model, we observed differential expression of genes that are involved in processes like angiogenesis, apoptosis, macrophage stimulation, and extracellular matrix regulation.
M
ATERIALS AND METHODS Cell linesCell lines Mel-270, OMM1.3, and OMM1.5 originated from one person and were derived from a primary uveal melanoma and two liver metastases, respectively. The two liver
metastases were obtained at the same time from two different sites in the liver. The cell lines were grown in RPMI 1640 (Gibco, Paisley, Scotland) supplemented with 3mM L-glutamine (Gibco, Paisley, Scotland), 2% penicillin/streptomycin and 10% FBS (Hyclone, UT). Cell cultures were incubated at 37°C in a humidified 5% CO2 atmosphere. In order to inhibit DNA methylation of daughter cells, the cell lines were grown in the presence of 1µM 5-aza-2′-deoxycytidine for 72 hours. Fresh medium was added every 24 hours. A de-methylated clone of cell line Mel-270 was derived when after 72 hours of 5-aza-2′deoxycytidine treatment and growth arrest for 75 days this cell line continued to proliferate again (Van der Velden et al., 2001).
The tubular network formation assay has been performed as described by Maniotis et al. (1999). From each cell line cells have been seeded on a glass coverslip with polymerized Matrigel or Type 1 collagen (Collaborative Biochemical, Bedford, MA). After an incubation period in conditioned medium of 1 week, the 3-D matrix gels were fixed and stained in order to report the tube formation ability.
cDNA arrays
We used the Atlas Human Cancer 1.2 Array (Clontech, Palo Alto, CA). This array consists of 1176 clones that are associated with cancer and some housekeeping genes that are spotted on a filter. Primer mixtures that accompany the filters ensure that cDNAs are only synthesized for the genes present on the filter. These primers reduce the complexity of the probe and hence result in specific hybridization probes.
Micro array hybridization and data processing
The cells were grown to sub-confluency and than harvested. Total RNA was isolated with the RNeasy kit (Qiagen, Hilden, Germany), followed by mRNA isolation with biotin-labeled oligo-dT and streptavidin-labeled magnetic beads. The mRNA was converted into radio-labeled (α-[32P]dCTP) cDNA using the oligonucleotide primers that were provided with the filters. These cDNA probes were subsequently hybridized with the Atlas Human Cancer 1.2 Array, all according to the manufacturers specifications (Clontech, Palo Alto, CA). Each individual spot (n=1176) on the filter reflects the expression level of a specific gene which was quantified by phosphorimager using the Image Quant software (Molecular Dynamics, Sunnyvale, CA).
By comparing different filter-hybridizations we can identify differentially expressed genes. In comparisons, only those genes were analyzed for which the combined signal of the two filters was at least twice the combined background of the two filters. Consequently, the expression of a gene had to be substantial in at least one cell line in order to enter the analysis. Also, all cell lines were hybridized in duplicate. The level of expression in the primary uveal
metastasis derived cell lines. Assuming that the filters were equally well hybridized, genes that are not differentially regulated will give a ratio of one. Genes that have a low expression in the metastasis cell lines will yield a ratio higher then one whereas genes that are higher expressed in the metastasis cell lines will yield a ratio lower than one (Figure 4.1). When the filters were not equally well hybridized, the median expression ratio was normalized to one. A twofold change in gene expression was considered differential whereas only the genes with a threefold change in expression are presented in the tables.
Figure 4.1 Unfiltered expression ratio of 1176 transcripts on a logarithmic scale in two
representative comparisons. Approximately half of the genes on the array were expressed in our cell
lines and were included for analysis. Normalization was based on the median expression ratio of two filter-arrays in a comparison as these graphs illustrate. (a) Nine genes were reproducibly found to be differentially expressed between the two metastasis cell lines OMM1.3 and OMM1.5. (b) Comparison of primary uveal melanoma cell line Mel-270 and metastasis cell line OMM1.3 revealed expression differences in genes that are possibly involved in metastasis.
Real-Time Quantitative RT-PCR
For validation of our array data we applied Quantitative RT-PCR (QRT-PCR) to TIMP3 and an endogenous reference gene (U1A). One µg of total RNA was first reverse-transcribed with M-MLV reverse transcriptase (Gibco-BRL, Eggenstein, Germany) in a 20 µl reaction volume. The cDNA synthesis was performed in the presence of random hexamer primers (Promega, San Louis Obispo, CA) at 42°C for 50 minutes. First strand cDNA was used as a template in a Taqman assay on an ABI PRISM 7700 Sequence Detection System using PCR master mix as supplied by the manufacturer (PE Applied Biosystems, Foster City, CA). The principle of the Taqman assay is based on emission of light during the elongation step of the PCR due to the nucleolytic activity of Taq polymerase. During elongation, Taq polymerase separates the reporter dye and the quencher dye of the sequence detection probe and hence light will be emitted by the reporter fluorochrome (Heid et al., 1996). The fluorescence signal is monitored during the PCR reaction and the number of cycles at which the signal reaches a certain
threshold level is used for quantitation. The difference between the threshold value of the target transcript and the endogenous reference is used to calculate the relative expression. In order to allow such quantitation, both target and reference amplicons must amplify with an
efficiency close to 100%. In dilution experiments with variable amounts of input cDNA both TIMP3 and U1A were shown to have those nearly optimal efficiencies.
Primers and probes for TIMP3 and U1A were designed using the Primer Express software (PE Applied Biosystems).
TIMP3-F: 5′-TTC GGT TAC CCT GGC TAC CA-3′, TIMP3-R: 5′-CTG CAG TAG CCG CCC TTC T-3′, TIMP3-probe: 6-FAM 5′-TCC AAA CAC TAC GCC TGC ATC CGG-3′ TAMRA.
U1A-F: 5′-GCA GCT TAT GCC AGG ACA GAT-3′, U1A-R: 5′-TTG GTG AGG AAC AAG ATG TGA TTC-3′, U1A-probe: Vic 5′-CCC CTG CCC AGC CTC TTT CTG AGA-3′ TAMRA.
Methylation-specific PCR
With methylation-specific PCR (MSP) we investigated TIMP3 promoter methylation (Bachman et al., 1999). This assay is based on chemical modification of DNA with
sodiumbisulfite. Due to this treatment cytosine residues in the DNA are converted into uracil unless a methyl group at the cytosine residue prevents deamination by bisulfite. The sequence differences that are introduced after modification between methylated and unmethylated DNA are used to design primers specific for methylated DNA with the Primer Express software (PE Applied Biosystems). The sequence of the forward primer, TIMP3-MF, is: 5′-GAA ACG TAC ACA TAC TCG CCC A-3′ and the reverse primer, TIMP3-MR: 5′-GGT GTA GAT TAG CGT GTC GAA GG-3′. After amplification, we determined the melting curve of the PCR fragments and observed melting temperatures that were concordant with methylation of internal CpG’s. Samples were also run on a 1.5% agarose gel and visualized with UV and ethidiumbromide.
Histopathology
We analyzed sections of paraffin-embedded primary uveal melanoma from 20 patients. Histopathological analysis was performed by an ocular pathologist according to standard citeria (Mooy and De Jong, 1996). Fourteen tumors were confined to the choroid and six tumors were located in the choroid as well as in the ciliary body.
Table 4.3 Clinical and histopathologic data.
Patient Cell type Pigmentation Diameter Mitosis1 Vasc. Loops2 TIMP3 Tumor pos.3
1 Spindle <50% 14 3 7 75-100% 2 Spindle <50% 10 2 3 25-50% 3 Mixed <50% 14 8 8 25-50% 4 Spindle <50% 12 3 7 75-100% 5 Epitheloid >50% 12 1 9 5-25% 6 Spindle <50% 12 3 7 <5% 7 Mixed <50% 20 5 8 <5% 8 Spindle >50% 16 5 3 <5% 9 Mixed severe 15 3 3 <5% 10 Mixed >50% 13 3 8 0% 11 Epitheloid >50% 14 3 9 75-100% 12 Spindle <50% 9 4 9 75-100% 13 Epitheloid >50% 12 2 9 <5% 14 Mixed <50% 8 2 9 25-50% 15 Mixed >50% 13 4 9 75-100% 16 Mixed >50% 15 3 7 <5% 17 Spindle <50% 11 5 9 5-25% 18 Mixed >50% 14 5 7 0% 19 Mixed <50% 17 8 8 5-25% 20 Mixed <50% 7 1 1 75-100%
1 Number of mitoses per 15 fields of microscopic examination at a magnification of X320 2 Amount and degree of loops in vascular networks (Folberg et al., 1992)
3 Percentage of tumor cells stained with anti-TIMP3 monoclonal antibody
Immuno-histochemistry
choroid of normal eye. We therefore used sections of normal eye tissue as positive controls (Fariss et al., 1997). As a control for the antibody, we performed immuno-cytochemistry on three normal choroidal melanocyte cell cultures and the tumor cell lines.
Statistical analysis
The relationship between TIMP3 expression and clinical and histopathological parameters was determined with the χ2 test or the nonpaired, two-sided Student’ s t-test, as appropriate.
R
ESULTS
Differential expression in uveal melanoma metastasis model
Cell lines Mel-270, OMM1.3, and OMM1.5 are derived from a primary uveal melanoma and two liver metastases from one patient and provide a progression model for uveal melanoma. In order to detect differentially expressed genes, we hybridized labeled cDNA derived from mRNA from these three cell lines with a filterarray. The array consists of 1176 cDNA clones that have been associated with cancer. Almost 50% of the clones (563 transcripts) were found to be expressed in the primary uveal melanoma cell line and/or the metastasis cell lines. Visual inspection of duplicate experiments showed nearly identical filterarray hybridizations that clearly displayed the fingerprints of the respective cell lines. Signal analysis of individual genes nevertheless revealed that 1-4% of the genes was differentially expressed between duplicate experiments. Five of those genes were discarded from the analysis because they alternately appeared upregulated or downregulated in our cell line model depending on the duplicate experiment.
There was great concordance in the results of the two metastasis cell lines (Figure 4.1). In total nine genes were found to be differentially expressed between these cell lines of which two genes (histone H4 and TIMP1) also showed differential expression between the primary and the metastases cell lines. Expression of H4 was much higher in cell line OMM1.3 than in OMM1.5, and TIMP1 expression was higher in OMM1.5 than in OMM1.3; nevertheless, H4 and TIMP1 expression were higher in both of the metastasis cell lines when compared with the primary cell line.
Table 4.1 Transcripts with at least a three fold change in gene expression between a primary uveal melanoma cell line and two metastasis cell lines; (a) genes with higher expression in the primary cell line or (b) genes with a lower expression in the primary cell line (Mel-270) compared to the metastasis cell lines (OMM1.3 and OMM1.5). A
Gene Name GenBank Accession Primary vs. Metastases
vimentin (VIM) X56134; M14144 ++
macrophage inhibitory cytokine 1 (MIC1) AF019770 ++
alpha-2-macroglobulin (alpha-2-M) M11313 ++
tissue inhibitor of metalloproteinase 3 (TIMP3)* Z30183 ++
c-jun N-terminal kinase 2 (JNK2) L31951 ++
cathepsin D (CTSD) M11233 ++
nucleoside diphosphate kinase B (NDP kinase B; NDKB) L16785; M36981 ++
type II cytoskeletal 8 keratin (KRT8) M34225 +
ATP synthase coupling factor 6 mitochondrial (F6)* M37104 +
integrin beta 8 (ITGB8)* M73780 +
plexin A3 (PLXNA3) X87852 +
IgG receptor Fc large subunit p51 (FCRN) U12255 +
melanocyte protein PMEL 17; antigen GP100 M77348 +
TNFR1-associated protein (TRAP1) U12595 +
cell surface glycoprotein MUC18 M28882 +
fibroblast adenine nucleotide translocator 2 (ANT2) J02683 +
polyhomeotic 2 homolog (HPH2) U89278 +
PM5 protein X57398 +
glutathione-S-transferase-like protein (GSTTLp28) U90313 +
myc proto-oncogene V00568 +
membrane attack complex inhibition factor (MACIF) M34671 +
eukaryotic translation elongation factor 2 (EEF2; EF2) X51466 +
cyclin-dependent kinase inhibitor 1A (CDKN1A) U09579; L25610 +
c-myc-binding protein MM-1 D89667 +
L-lactate dehydrogenase H subunit (LDHB)* Y00711 +
high mobility group protein isoforms I & Y (HMGIY) M23619 +
interleukin 6 (IL6) X04602; M14584 +
bifunctional purine biosynthesis protein U37436 +
integrin alpha E (ITGAE)* L25851 +
B
Gene Name GenBank Accession Primary vs. Metastases
histone 4 (H4) X67081 --
tissue inhibitor of metalloproteinase 1 (TIMP1)∗ X03124 --
wingless-related MMTV integration site 5a protein (WNT5A) L20861 --
40S ribosomal protein S5 (RPS5) U14970 -
smoothened homolog (SMOH) U84401 -
leukocyte interferon-inducible peptide X02492 -
macrophage migration inhibitory factor (MIF) M25639 -
34/67-kDa laminin receptor; laminin receptor 1 (LAMR1) U43901 -
integrin beta 4 (ITGB4) X53587; X52186 -
60S ribosomal protein L10 (RPL10); tumor suppressor QM M73791 -
CD40 receptor-associated factor 1 (CRAF1) U21092 -
tumor necrosis factor receptor 2 (TNFR2) M32315; M55994 -
guanylate kinase (GMP kinase) L76200 -
-, lower expression in the primary compared to the metastasis cell lines (>3fold), -- >5fold ∗ gene transcript that was up-regulated after demethylation (Table 4.2)
TIMP3 protein expression in the cell lines and in primary uveal melanoma
TIMP3 was one of the seven genes that showed a more than fivefold down-regulated gene
expression in the metastases cell lines compared to the primary tumor cell line (Table 4.1a). Real-time QRT-PCR of the TIMP3 mRNA in the cell lines confirmed down-regulation of TIMP3 in the metastasis cell lines. The primary uveal melanoma cell line expressed approximately 20% of the mRNA levels observed in a normal choroidal melanocyte cell culture and the metastasis cell lines 4% and 2% respectively. Since TIMP3 is gradually down-regulated in the progression model and because of its involvement in extracullular matrix maintenance, we performed the tubular network formation assay. The metastasis cell lines showed partial tubular network formation whereas the primary cell line did not reveal any tube formation.
In order to examine the expression level of TIMP3 protein in primary uveal melanoma, a series of 20 tumors was stained with an anti-TIMP3 monoclonal antibody (Table 4.3). Six of the melanomas showed staining of more than 50% of the neoplastic cells (Figure 4.2), whereas in eight melanomas less than 5% of the cells was positive for TIMP3 and six melanomas had an intermediate expression (5-50%). Association with clinical and
Figure 4.2: Differential TIMP3 staining in an uveal melanoma. (a) section of cells that are positive for
TIMP3 staining; (b) section where less than 5% of the cells stains for TIMP3.
Demethylation of primary uveal melanoma cell line
To assess which genes were repressed by promoter methylation we determined the expression profile of the demethylated clone of Mel-270 using the microarray. This clone was derived from Mel-270 that gave rise to proliferating clones after 5-aza- 2’ deoxycytidine treatment and hence provided a rich and homogeneous source of demethylated cells (Van der Velden et al., 2001). Although a substantial time period had passed since the cells were treated with 5-aza-2’ deoxycytidine the effect could still be measured, supposedly because demethylated daughter chromosomes are transmitted with each cell division. It appeared that many genes were up-regulated after demethylation while only a few are down-up-regulated. Table 4.2 shows the 19 genes that were at least threefold up-regulated following demethylation. Among these, five genes had a high expression and one a low expression in the primary cell line compared to the metastasis cell lines (see Table 4.1). The highest increase in gene expression after
Table 4.2: Differentially expressed genes after demethylation of primary uveal melanoma cell line Mel-270.
Gene Name GenBank Accession 5-aza-dC
tyrosinase-related protein 1 (TRP-1) X51420 ++
endothelin 2 (ET2) M65199 ++
calmodulin (CALM; CAM) D45887 ++
elongation factor 1 alpha (EF1-alpha; EF1A) M27364 +
tissue inhibitor of metalloproteinase 3 (TIMP3) Z30183 +
L-lactate dehydrogenase H subunit (LDHB) Y00711 +
ATP synthase coupling factor 6 mitochondrial (F6) M37104 +
glycyl tRNA synthetase (GLYRS; GARS) D30658 +
integrin beta 8 (ITGB8) M73780 +
integrin alpha E (ITGAE) L25851 +
tissue inhibitor of metalloproteinase 1 (TIMP1) X03124 +
cartilage-specific proteoglycan core protein (CSPCP) M55172 +
KIAA0324 AB002322 +
casein kinase I gamma 2 (CSNK1G2) U89896 +
sialyltransferase 4A (SIAT4A) L13972 +
c-src kinase (CSK) X59932 +
cell cycle protein P38-2G4 homolog (EBP1) U59435 +
PIRIN Y07867 +
proliferating cyclic nuclear antigen (PCNA) M15796; J04718 +
+, higher expression in the demethylated cell line (>3fold), ++ >5fold
Figure 4.3 Microarray analysis of Mel-270 and the demethylated cell line clone of Mel-270. LDHB
expression was moderately upregulated whereas the TRP1 transcript was approximately tenfold increased.
LDHB
Mel-270 Mel-270 demethylated
LDHB
Figure 4.4 Effect of treatment with the hypomethylating agent 5-aza-2’deoxycytidine on TIMP3 expression in the cell lines. (a)TIMP3 gene expression measured with microarray in the tumor cell lines
including the clone of Mel-270 that remained proliferating after demethylation. (b) TIMP3 gene expression in the cell lines relative to the expression in a primary choroidal melanocyte culture as measured with Quantitative RT-PCR (Taqman assay) before and directly after 5-aza-2’ deoxycytidine treatment. (c) MSP analysis in a tumor negative for TIMP3 staining, in two of the cell lines and in the primary melanocyte culture. The primary melanocyte is negative for MSP whereas the tumor and the tumor cell lines appear methylated.
D
ISCUSSIONUveal melanoma has a great propensity to metastasize hematogenously and to develop distant metastases, in particular to the liver. Micrometastases may already be present at the time of primary tumor detection. We tested whether we could identify biologically important molecules that can be used as progression markers to facilitate identification of high-risk patients. In order to identify genes associated with the process of metastasis, we compared gene expression profiles from a primary uveal melanoma cell line and two metastasis cell lines from the same patient. From a total of 1176 genes tested with the array, we detected expression of almost half of the genes (n=563) in the primary and/or the metastases cell lines. Sixty-three genes were expressed differentially between primary and metastases cell lines, when a two-fold change of expression was used as threshold. In contrast, comparison of the two metastases cell lines did not reveal much differential expression, showing that a great similarity existed between the liver metastases. In addition, the results indicate that
maintaining the cell lines in culture for a long time had not influenced the expression pattern of the two metastases cell lines. One of the most interesting findings was the fivefold lower expression of TIMP3 in the metastasis cell lines compared to the primary cell line, expression differences that we could confirm with QRT-PCR (Figure 4.4b). Normally, TIMP3 is
expressed in the eye by the retinal pigment epithelium and deposited in Bruch’ s membrane (Fariss et al., 1997). Mutations in the TIMP3 gene cause hereditary blindness (Sorsby’ s fundus dystrophy), which is characterized by thickening of Bruch’ s membrane, choroidal neovascularization and photoreceptor degeneration (Weber et al., 1994). Also, TIMP3
down-T IM P-3 ce ll lin e/ m el an oc yt e
melanocyte Mel-270 OMM-1.3
untreated 5'aza OMM1.5 1.2 1 0.8 0.6 0.4 0.2 0 OMM1.3
metastasis Mel-270 primary Mel-270 demethylated
regulation has been found in cancer from different origins including kidney and lung, but had not been tested in uveal melanoma (McElligott et al., 1997; Michael et al., 1999).
Homeostasis of the extracellular matrix is achieved by keeping degradation and remodeling in balance. A disturbed interaction between metalloproteinases and it’ s inhibitors (TIMPs) can result in degradation of the basement membrane and favor tumor cell invasion and metastasis (Hofmann et al., 2000; Spurbeck et al., 2002). TIMP3 distinguishes itself from TIMP1, another tissue inhibitor of metalloproteinases, that was found to be differentially expressed (Table 4.1b), by the fact that it is water insoluble and that it is bound to the extracellular matrix (Brew et al., 2000). TIMP3 also exerts a unique function by not only inhibiting invasion but also inducing apoptosis (Ahonen et al., 1998). In this way, loss of TIMP3
interferes with normal regulation of apoptosis and might actually promote tumor progression. Future research has to identify the functional role of TIMP3 in uveal melanoma progression. A hint that loss of TIMP3 associates with tumor invasion was provided in vitro with the tubular network formation we observe in the metastasis cell lines. In order to test the biological significance of reduced TIMP3, we stained 20 uveal melanoma sections with a TIMP3 monoclonal antibody. In roughly half of the tumors, more than 50% of the cells stained positive for TIMP3. Since the tumors were recently collected, the effect of TIMP3 expression on survival could not yet be analyzed. However, a negative correlation was found between TIMP3 staining and tumor size, one of the most consistent prognostic factors in uveal melanoma (Mooy and De Jong, 1996). Down-regulation due to hypermethylation of the
TIMP3 gene is frequently observed in all kinds of cancer (Bachman et al., 1999). Recently we
showed that promoter methylation may be a mechanism of CDKN2A inactivation in uveal melanoma including the cell lines that we used in the present study (Van der Velden et al., 2001). Because we have generated a demethylated clone of cell line Mel-270, we were able to evaluate the effect of demethylation on gene expression, including TIMP3, with the array hybridization of this cell line. As expected, TIMP3 expression was markedly increased after demethylation reaching a mRNA expression level in the range we observed in normal choroidal melanocytes. With QRT-PCR on 5-aza-2’ deoxycytidine treated cell lines we revealed up-regulation of TIMP3 in the metastasis cell lines and confirmed upregulation of TIMP3 in the primary uveal melanoma cell line. Subsequent analysis of promoter methylation using MSP supported the observation from the demethylation experiment that TIMP3
without offering the tumor a clear advantage. It is also evident that some of the genes that responded to 5-aza-2’ deoxycytidine were up-regulated through another mechanism than demethylation or were up-regulated as an effect of another demethylated gene. PCNA, for example, has been analyzed for promoter methylation but expression differences were not caused by methylation (Liu et al., 1993). The notion that TIMP3 up-regulation is caused by promoter demethylation however, is supported by the elevated TIMP3 expression in our cell line model directly after 5-aza-2’ deoxycytidine treatment. Indirect mechanisms may be suspected in the demethylated clone but such mechanisms are unlikely to operate directly after treatment. Hypermethylation is easily reversed with hypomethylating agents and this makes gene methylation attractive for intervention. Treatment with 5-aza-2’ deoxycytidine has proven to be effective but is also toxic. Safer treatment modalities may be provided by S-adenosyl-L-homocysteine (SAH), a nontoxic hypomethylating agent. In addition, dietary folate depletion has proven to be effective in causing gene hypomethylation (De Cabo et al., 1994; Kauwell et al., 2000). It remains, however, to be proven that these alternative
treatments result in re-expression of methylated tumor suppressor genes. Another interesting difference was the threefold increase in macrophage migration inhibitory factor (MIF) mRNA in the metastases cell lines OMM1.3 and OMM1.5 compared to the primary cell line, Mel-270. Consistently, Repp et al. (2000) reported a similar increase in MIF protein expression for the metastasis cell line OMM2.3 compared to Mel-270. Uveal melanoma are reported to achieve protection against natural killer cells through secretion of MIF once they leave the immune-privileged ocular environment, and in this way MIF may contribute to dissemination (Apte et al., 1998; Repp et al., 2000).
In conclusion, expression profiling of our uveal melanoma cell line model was effective in recognizing expression differences that could be validated in primary tumors and in revealing a possible mechanism of gene repression. Opposed to “hypothesis based” candidate
progression markers, analysis of this cell line model provides “discovery based” candidate progression markers. Follow up studies of TIMP3 and other differentially expressed genes are warranted and may result in the identification of progression markers that will facilitate recognition of early metastatic disease.
A
CKNOWLEDGEMENTSR
EFERENCESAhonen M, Baker AH, Kahari VM. Adenovirus-mediated gene delivery of tissue inhibitor of metalloproteinases-3 inhibits invasion and induces apoptosis in melanoma cells. Cancer Res. 1998;58:2metalloproteinases-310-15.
Apte RS, Sinha D, Mayhew E, Wistow GJ, Niederkorn JY. Cutting edge: role of macrophage migration inhibitory factor in inhibiting NK cell activity and preserving immune privilege. J Immunol 1998;160:5693-6.
Bachman KE, Herman JG, Corn PG, Merlo A, Costello JF, Cavenee WK, Baylin SB, Graff JR. Methylation-associated silencing of the tissue inhibitor of metalloproteinase-3 gene suggest a suppressor role in kidney, brain, and other human cancers. Cancer Res 1999;59:798-802.
Brew K, Dinakarpandian D, Nagase H. Tissue inhibitors of metalloproteinases: evolution, structure and function. Biochim Biophys Acta 2000;1477:267-83.
De Cabo SF, Hazen MJ, Molero ML, Fernandez-Piqueras J. S-adenosyl-L-homocysteine: a non-cytotoxic hypomethylating agent. Experientia 1994;50:658-9.
DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen Y, Su Ya, Trent JM. Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat Genet 1996;14:457-60.
Diener-West M, Hawkins BS, Markowitz JA, Schachat AP. A review of mortality from choroidal melanoma. II. A meta-analysis of 5-year mortality rates following enucleation, 1966 through 1988. Arch Ophthalmol 1992;110:245-50.
Fariss RN, Apte SS, Olsen BR, Iwata K, Milam AH. Tissue inhibitor of metalloproteinases-3 is a component of Bruch's membrane of the eye. Am J Pathol 1997;150:323-8.
Folberg R, Pe'er J, Gruman LM, Woolson RF, Jeng G, Montague PR, Moninger TO, Yi H, Moore KC. The morphologic characteristics of tumor blood vessels as a marker of tumor progression in primary human uveal melanoma: a matched case- control study. Hum.Pathol. 1992;23:1298-305.
Foss AJ, Alexander RA, Jefferies LW, Hungerford JL, Harris AL, Lightman S. Microvessel count predicts survival in uveal melanoma. Cancer Res 1996;56:2900-3.
Gamel JW, McLean IW, McCurdy JB. Biologic distinctions between cure and time to death in 2892 patients with intraocular melanoma. Cancer 1993;71:2299-305.
Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res 1996;6:986-94.
Hofmann UB, Westphal JR, van Muijen GN, Ruiter DJ. Matrix metalloproteinases in human melanoma. J Invest Dermatol 2000;115:337-44.
Hurks HM, Metzelaar-Blok JA, Barthen ER, Zwinderman AH, De Wolff-Rouendaal D, Keunen JE, Jager MJ. Expression of epidermal growth factor receptor: risk factor in uveal melanoma. Invest Ophthalmol Vis Sci 2000;41:2023-7.
Kauwell GP, Lippert BL, Wilsky CE, Herrlinger-Garcia K, Hutson AD, Theriaque DW, Rampersaud GC, Cerda JJ, Bailey LB. Folate status of elderly women following moderate folate depletion responds only to a higher folate intake. J Nutr 2000;130:1584-90.
Kerstens HM, Poddighe PJ, Hanselaar AG. A novel in situ hybridization signal amplification method based on the deposition of biotinylated tyramine. J Histochem Cytochem 1995;43:347-52.
Kwon BS, Halaban R, Chintamaneni C. Molecular basis of mouse Himalayan mutation. Biochem Biophys Res Commun 1989;161:252-60.
Liu YW, Chang KJ, Liu YC. DNA methylation is not involved in growth regulation of gene expression of proliferating cell nuclear antigen. Exp Cell Res 1993;208:479-84.
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;36:435-41.
Makitie T, Summanen P, Tarkkanen A, Kivela T. Tumor-infiltrating macrophages (CD68(+) cells) and prognosis in malignant uveal melanoma. Invest Ophthalmol Vis Sci 2001;42:1414-21.
Maniotis AJ, Folberg R, Hess A, Seftor EA, Gardner LM, Pe'er J, Trent JM, Meltzer PS. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry. Am.J.Pathol. 1999;739-52.
McLean IW. The biology of haematogenous metastasis in human uveal malignant melanoma. Virchows Arch A Pathol Anat Histopathol 1993;422:433-7.
Michael M, Babic B, Khokha R, Tsao M, Ho J, Pintilie M, Leco K, Chamberlain D, Shepherd FA. Expression and prognostic significance of metalloproteinases and their tissue inhibitors in patients with small-cell lung cancer. J Clin Oncol 1999;17:1802-8.
Mooy CM, De Jong PT. Prognostic parameters in uveal melanoma: a review. Surv Ophthalmol 1996;41:215-28. Repp AC, Mayhew ES, Apte S, Niederkorn JY. Human uveal melanoma cells produce macrophage
migration-inhibitory factor to prevent lysis by NK cells. J Immunol 2000;165:710-5.
Spurbeck WW, Ng CY, Strom TS, Vanin EF, Davidoff AM. Enforced expression of tissue inhibitor of matrix metalloproteinase-3 affects functional capillary morphogenesis and inhibits tumor growth in a murine tumor model. Blood 2002;100:3361-8.
Van der Velden PA, Metzelaar-Blok JA, Bergman W, Monique H, Hurks H, Frants RR, Gruis NA, Jager MJ. Promoter hypermethylation: a common cause of reduced p16(INK4a) expression in uveal melanoma. Cancer Res 2001;61:5303-6.
