Uveal and Cutaneous Melanoma: Shared Expression
Characteristics of Melanoma-Associated Antigens
Leonie C. van Dinten,1 Nicolien Pul,1 A. Frans van Nieuwpoort,2 Coby J. Out,2
Martine J. Jager,3 and Peter J. van den Elsen1,4
PURPOSE. Downregulation of melanoma-associated antigens
(MAAs), against which natural cytolytic T lymphocytes
(CTLs) exist in humans, is one of the mechanisms that aids
in evasion of immune surveillance. In view of putative reexpression strategies for MAAs during immunotherapy, this
study was conducted to investigate MAA silencing in malignant melanoma.
METHODS. The expression of the MAA Melan-A/MART-1 was
analyzed in 10 uveal and 10 cutaneous patient-derived melanoma cell lines by Western blot analysis and RT-PCR. Expression characteristics of four other MAAs—Tyr, Tyrp1, Dct, and
gp100/Pmel17—were analyzed by RT-PCR. DNA methylation
patterns at the Melan-A/MART-1 promoter region were investigated by methylation-sensitive restriction enzyme digestion
and subsequent Southern blot analysis. Exogenous promoter
activity was assessed in all 20 melanoma cell lines to correlate
the DNA methylation patterns with Melan-A/MART-1 expression.
RESULTS. MAA expression was observed in 15 of the 20 melanoma cell lines. Furthermore, there is a direct correlation
between DNA methylation patterns at the Melan-A/MART-1
promoter region, exogenous Melan-A/MART-1 promoter activity, and Melan-A/MART-1 protein expression. These data reveal
the division of patient-derived melanoma cell lines into two
distinct subsets, which are identical for both uveal and cutaneous tumor types.
CONCLUSIONS. The authors propose a categorization of melanoma cell lines into two different panels based on shared
MAA-expression characteristics: panel I, MAA-expressing cell
lines, and panel II, MAA-deficient cell lines. This categorization
can be used to obtain knowledge about the regulation of
MAA-expression and for further research concerning MAAbased immunotherapy. (Invest Ophthalmol Vis Sci. 2005;46:
24 –30) DOI:10.1167/iovs.04-0961
From the 1Department of Immunohematology and Blood Transfusion, Division of Molecular Biology, and the Departments of 2Dermatology and 3Ophthalmology, Leiden University Medical Center, Leiden,
The Netherlands; and the 4Department of Pathology, Vrije Universiteit
Medical Center, Amsterdam, The Netherlands.
Supported by Grant NKB/RUL 2001-2522 from the Dutch Cancer
Society. AFvN is supported by The Netherlands Organization of Scientific Research.
Submitted for publication August 9, 2004; revised September 27,
2004; accepted September 30, 2004.
Disclosure: L.C. van Dinten, None; N. Pul, None; A.F. van
Nieuwpoort, None; C.J. Out, None; M.J. Jager, None; P.J. van den
Elsen, None
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Corresponding author: Peter J. van den Elsen, Department of
Immunohematology and Blood Transfusion, Division of Molecular Biology, Building 1, E3-Q, Leiden University Medical Center, Albinusdreef
2, 2333 ZA Leiden, The Netherlands; [email protected].
24
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
U
veal and cutaneous melanoma are highly malignant diseases that are difficult to treat. One of the novel lines of
treatment is immunotherapy, which targets melanoma-associated antigens (MAAs).1,2 However, in many cases, immunoselection of antigen-negative tumor cells and silencing of MAAs
occurs, resulting in evasion of immune surveillance and tumor
escape3,4 (for a review, see Ref. 5). In the light of the improved
efficacy of immunotherapeutic approaches, we evaluated MAA
expression in a panel of uveal and cutaneous melanoma cell
lines, with the future prospect of developing strategies that can
modulate MAA expression, to obtain optimal immunorecognition.
Melanoma-associated antigens (MAAs) are highly immunogenic human antigens that are recognized by cytotoxic T lymphocytes (CTLs). Both ocular and skin melanomas are derived
from normal melanocytes that originate from the neural crest,
although the melanocytes of the eye are functionally and morphologically distinct.6,7 Both types of melanomas carry a wide
range of MAAs, including Melan-A/MART-1,8,9 gp100,10 tyrosinase (Tyr),11,12 tyrosinase-related protein 1 (Tyrp1),13,14 and
dopachrome tautomerase (Dct).15,16 At the same time, these
tumors carry low immunogenic melanoma-specific antigens
(expressed only in tumor tissue), including members of the
MAGE family.17,18 The expression of these melanoma-specific
antigens by melanoma cells in primary tumor tissue and in
cultured tumor cells has been described to be variable for both
ocular and skin melanoma, whereas the MAAs are more ubiquitously expressed (reviewed in Refs. 7,19). One of the important questions that awaits clarification is whether ocular and
skin melanomas express similar or distinct MAAs. We therefore
analyzed MAA expression in multiple uveal and cutaneous
melanoma cell lines and focused on one of the MAAs, MelanA/MART-1.
Currently, little is known about the function of Melan-A/
MART-1 in melanocyte development and differentiation.
Melan-A/MART-1 is expressed in normal immature melanocytes
and in the majority (⬎90%) of fresh melanoma tumors and
melanoma cell lines, but it is not expressed in other cells and
other tumors. Because of the high percentage of melanoma
tumors that show Melan-A/MART-1 expression and due to its
immunogenicity, this antigen is one of the targets for cellular
immunotherapy against malignant melanoma. However, when
immunoselection of antigen-negative tumor cells and silencing
of MAAs occurs, their absence in the tumor cells impairs
immune recognition by antigen-specific host CTLs, thereby
reducing the efficacy of the immunotherapy. Unfortunately,
up to now, attempts to correlate Melan-A/MART-1 expression
and tumor staging have yielded conflicting results. Nevertheless, its expression can be used as a predictive factor for the
selection of patients eligible for Melan-A/MART-1– based immunotherapy.20 –24
In this study, we investigated Melan-A/MART-1 expression
in 10 uveal melanoma cell lines and 10 cutaneous melanoma
cell lines (Table 1). Based on Melan-A/MART-1 protein and
mRNA expression, the melanoma cell lines could be divided
into two distinct groups: Melan-A/MART-1-protein– expressing
and Melan-A/MART-1-protein– deficient cell lines. Interestingly,
Investigative Ophthalmology & Visual Science, January 2005, Vol. 46, No. 1
Copyright © Association for Research in Vision and Ophthalmology
Shared Expression of Antigens in Melanoma
IOVS, January 2005, Vol. 46, No. 1
TABLE 1. Uveal and Cutaneous Melanoma Cell Lines
Uveal Melanoma
Cell Lines
92-1†
92-2†
Mel270§
Mel285
Mel290
OCM1
OCM3
OMM1
OMM1.3§
OMM1.5§
Derivation*
Cutaneous
Melanoma
Cell Lines
Derivation
Primary
92-1
Primary
Primary
Primary
Primary
Primary
Metastasis
Metastasis
Metastasis
136.2
453A0‡
453B‡
513D
EW
IGR39D
MA
MO
Mu89㛳
Mu96㛳
Metastasis
Primary
Primary
Metastasis
Metastasis
Primary
Metastasis
Metastasis
Metastasis
Mu89
* Primary, primary tumor mass.
† 92-2 is a tissue culture– derived cell line from 92-1.
‡ Derived from the same tumor mass.
§ OMM1.3 and OMM1.5 are liver metastases from the Mel270
primary tumour.
㛳 Mu96 is a tissue culture-derived cell line from Mu89.
melanoma cells expressing Melan-A/MART-1 also expressed
the MAAs Tyr, Tyrp1, Dct, and gp100/Pmel17, whereas cells
that lacked expression of Melan-A/MART-1 failed to express
these additional MAAs. Moreover, we showed distinct MelanA/MART-1 DNA methylation patterns that correlated with the
typical Melan-A/MART-1 protein/mRNA expression patterns as
observed in the two groups of melanoma cell lines.
In conclusion, our results demonstrate the division of patient-derived uveal and cutaneous melanoma cell lines into two
panels on the basis of MAA expression characteristics, MelanA/MART-1 DNA methylation patterns, and Melan-A/MART-1
transcriptional activity. More important, comparable MAA expression patterns were observed in both ocular and skin melanomas, rendering both types of malignant melanoma suitable
for identical immunotherapeutic approaches.
MATERIAL
AND
METHODS
Cell Lines, Melan-A/MART-1 Protein Analysis,
and RT-PCR
Patient characterization and origin of the uveal melanoma cell lines
92-1, 92-2, Mel270, Mel285, Mel290, OCM1, OCM3, OMM1, OMM1.3,
and OMM1.5,25 and the cutaneous melanoma cell lines 136.2, 453A0,
453B, 513E, IGR39D,26 MA, MO,27 Mu89, Mu96, and EW28 have been
described previously (see also Table 1). All melanoma cell lines were
grown at 37°C in a 5% CO2 humidified incubator in Iscove’s modified
Dulbecco’s medium (IMDM; BioWhittaker Europe, Verviers, Belgium),
supplemented with 100 U/mL penicillin, 100 ␮g/mL streptomycin, and
10% fetal calf serum (FCS; Integro, Amsterdam, The Netherlands).
Initial passage numbers of the obtained cell lines were unknown.
However, all cell lines were kept in culture for not more than 3 months
(maximum, 30 passages) and all analyses were performed at least twice
at different culture time points. Conditions of establishment and culture of normal melanocytes, M0101, M9228, M0311, M0312, and
M0313, have been published elsewhere.29 Three uveal melanocyte
cultures, Mel1A, Mel1B, and Mel2, were established from two normal
human donor eyes from different individuals.30
For Western blot analysis, Melan-A/MART-1 protein samples were
prepared by lysing approximately 1 to 2 ⫻ 107 cells in 1 mL lysis buffer
(20 mM Tris/HCl [pH 7.6], 150 mM NaCl, 0.1% SDS, 0.5% DOC, 1%
NP40, 1⫻ protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO).
After centrifugation, a 1:100 volume 0.5 M EDTA was added to the
supernatant. Protein concentrations were determined, equal amounts
were used for SDS-PAGE analysis, and the lysates were subsequently
assessed for the presence of Melan-A/MART-1 by Western blot, using
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
25
the Melan-A/MART-1 monoclonal antibody (clone A103, 1:1000; NeoMarkers, Fremont, CA). A 50⫻ dilution was used for lysates of melanocyte cultures to obtain similar exposure times (30 seconds) as for
the melanoma cell lysates. ␤-Actin staining was used as an internal
control (1:10,000; Oncogene, Boston, MA).
Melan-A/MART-1 mRNA expression was determined by RT-PCR.
cDNA was prepared by using random hexamers (Promega, Madison, WI)
and oligo-dT primers (Amersham Biosciences, Piscataway, NJ). The PCR
reaction was performed by using the previously described exon 2 and 5
primers for 30 cycles at 60°C.28 For Tyr, Tyrp1, Dct, and gp100/Pmel17
RT-PCRs, the following primer pairs and conditions were used: Tyr: sense
5⬘-GCTTTTCAGAGGATGAAAGCTTAAG-3⬘, antisense 5⬘-GTACTCCTCCAATCGGCTACAG-3⬘, 30 cycles, 64°C; Tyrp1: sense 5⬘-CTCTTATTTCAAGCAGAATGAGTG-3⬘, antisense 5⬘-GCCACAGCGGTCTGTCCCAG3⬘, 30 cycles, 64°C; Dct: sense 5⬘-GTAACCTCTGTGATTCTTGTGGG-3⬘,
antisense 5⬘-CACTGGTGGTTTCTTCCGCTCG-3⬘, 30 cycles, 60°C; gp100/
Pmel17: sense 5⬘-CCCAGAAACCAGGACTGGCTTG-3⬘, antisense 5⬘-GCTTCTCTTCTGAGACCAAGAGCC-3⬘, 30 cycles, 60°C.
Southern Blot Analysis
Genomic DNA was isolated by lysing 106 to 107 cells in 1 mL DNA lysis
buffer (100 mM Tris/HCl [pH 8.0], 5 mM EDTA, 0.2% SDS, 200 mM
NaCl, 200 ␮g/mL ProtK) and overnight (o/n) incubation at 56°C. DNA
was purified by phenol and subsequent phenol-chloroform-isoamylalcohol extractions and was precipitated with 100% ethanol. Pellets
were resuspended in 100 to 200 ␮L TE (10 mM Tris/HCl [pH 8.0], 1
mM EDTA).
For testing the methylation status of genomic DNA, the DNA was
digested with HindIII alone, HindIII and NruI, HindIII and HpaII, or
HindIII and MspI. Digestions were performed in a total volume of 50
␮L, using 10 ␮g (melanocytes) or 20 ␮g (melanomas) of DNA and 40
to 50 U of each restriction enzyme. Samples were incubated for 5
hours at 37°C before they were loaded onto 0.8% agarose gels. Gels
were run o/n in 1⫻ TBE (89 mM Tris-base, 89 mM boric acid, and 2 mM
EDTA) at 30 to 40 V.
The DNA was blotted onto predampened nylon transfer membranes (Hybond N⫹; Amersham Pharmacia Biotech) in 0.4 M NaOH
o/n at room temperature (RT). Blots were briefly washed in 2⫻ SSC
before hybridization in 0.5 M phosphate buffer [pH 7.2], 7% SDS, 10
mM EDTA at 65°C using the appropriate probe. Probes were prepared
from the Melan-A/MART-1 promoter region and were purified from
agarose gel. DNA (100 –200 ng) was labeled by the method of random
priming. After o/n hybridization, blots were washed several times
before autoradiography.
Melan-A/MART-1 Promoter Reporter Assays
For the generation of the various Melan-A/MART-1 promoter reporter
constructs, PCRs were performed on genomic DNA that had been
isolated from melanoma cell lines Mu96, IGR39D, and MA. The following primer pairs were used for the generation of various promoter
constructs in the sense and antisense (A) orientation (see Fig. 3A):
p286 and p286A: sense 5⬘-GTGACATGGCAATCCTATGGAGGAGGGAC-3⬘ and anti-sense 5⬘-AGTCCTCTGTCTGCTGGCTGGCCGCGTGTATGAAGATGCT-3⬘; p1593 and p1593A: sense 5⬘-AGCATCTTCATACACGCGGCCAGCCAGCAGACAGAGGACT-3⬘ and anti-sense 5⬘ATCTTGTAGGGTCAGGGCACAGGACACC-3⬘; and p1373 and p1373A:
sense 5⬘-TCATGCCTGTAATCCCAGCACTTTGGGAGG-3⬘ and antisense
5⬘-TGAAGATGCTTCTCTGGCTCTTAATCGTTTTGACTTATTT3⬘.
PCR fragments were cloned into the pGEM-T Easy vector, and their
nucleotide sequence was verified, revealing identical nucleotide sequences. The PCR fragments were subsequently transferred from the
pGEM-T Easy vector to pGL3-basic containing the luciferase gene and
an ApaI-SacII-AatII linker in either the KpnI or the HindIII site. Fragments were transferred by ligating the ApaI-SpeI– digested sequenced
inserts to the ApaI-NheI digested pGL3-basic vector. Constructs p2956
and p2956A were cloned by swapping a BlnI-HindIII fragment derived
26
van Dinten et al.
IOVS, January 2005, Vol. 46, No. 1
RESULTS
MAA Expression in Uveal and Cutaneous
Melanoma Cell Lines
FIGURE 1. Melan-A/MART-1 protein analysis and MAA expression. (A)
Immunoblot of total lysates of melanocyte (M0101, M9228) and melanoma cell cultures stained with a monoclonal antibody against MelanA/MART-1. For the cutaneous and uveal melanoma cell lines, 6 and 10
␮g were loaded, respectively. The amount of cell lysate of the melanocyte cultures was diluted 50⫻ (200 ng). The position of ␤-actin
(internal control) and Melan-A/MART-1 is indicated on the left. M,
Melanocyte cell culture. (B) MAA expression. RT-PCR analysis of
Melan-A/MART-1, Tyr, Tyrp1, Dct, and gp100 and the GAPDH control.
from p1839 and p1839A to the BlnI-HindIII digested vectors p1373
and p1373A, respectively.
Transfection assays were performed by using the above-described
constructs and construct pRL(␤-actin), containing the Renilla gene
behind the ␤-actin promoter. Cells were seeded at 2 ⫻ 105 cells/10
cm2 the day before transfection. For each transfection (in quadruplicate) 4 ␮g of promoter construct and 0.4 ␮g of pRL (␤-actin) construct
were transfected by using the CaCl2 coprecipitation method.31 Cells
were incubated at 37°C for 2 days before lysis in 200 ␮L passive lysis
buffer (Dual-Luciferase Reporter Assay System; Promega) per 10 cm2.
Of each sample, 10 ␮L was analyzed for luciferase and Renilla
activity using the reporter assay. Activity was measured on a Victor2,
1420-012 multilabel counter (Wallac, Oy, Finland). Experiments were
performed at least twice to confirm reproducibility.
We first evaluated the expression of Melan-A/MART-1 by Western blot analysis in our panel of melanoma cell lines derived
from tissue samples of various melanoma patients (Table 1, Fig.
1); normal melanocyte cultures served as a control. ␤-Actin
staining was used as an internal control for equal protein
loading. The results of the Western blot analysis (Fig. 1A)
showed that Melan-A/MART-1 expression could easily be detected in several melanoma cell lines (cutaneous: 136.2,
453A0, 453B, 513D, MA, MO, and Mu89; uveal: 92-1, 92-2,
Mel270, OCM1, OCM3, OMM1, OMM1.3, and OMM1.5),
whereas other melanoma cell lines lacked detectable levels of
Melan-A/MART-1 protein expression (cutaneous: EW, IGR39D,
and Mu96; and uveal: Mel285 and Mel290), in normal skinderived melanocytes (M0101 and M9228), Melan-A/MART-1
was abundantly expressed (Fig. 1A).
RT-PCR analysis (Fig. 1B) revealed that the Melan-A/MART-1
protein–negative uveal and cutaneous melanoma cell lines
were also deficient for Melan-A/MART-1 mRNA synthesis. Furthermore, uveal (not shown) and cutaneous melanocyte cell
cultures (Fig. 1B) showed strong Melan-A/MART-1 expression.
We next evaluated the expression characteristics of other
MAAs (Tyr, Tyrp1, Dct, and gp100/Pmel17) by RT-PCR in these
melanoma cell lines and melanocyte cell cultures. The results
of these analyses showed that cell lines that expressed MelanA/MART-1 also expressed the other MAAs, whereas cell lines
that were deficient for Melan-A/MART-1 expression also lacked
expression of Tyr, Dct, and gp100/Pmel17, although minute
levels of Tyrp1 were detected in the otherwise negative cell
lines (Fig. 1B, Table 2). Similar to the Melan-A/MART-1– expressing melanoma cell lines, uveal (not shown) and cutaneous
melanocyte cell cultures (Fig. 1B) showed expression of the
other MAAs.
Based on these analyses, a division was made into two
different panels of uveal and cutaneous melanoma cell lines:
panel I, MAA-expressing cell lines, and panel II, MAA-deficient
cell lines. These data are indicative of a process whereby
melanocytes transform into different types of tumor cells,
which can be distinguished on the basis of MAA expression or
deficiency.
Specific DNA methylation Pattern Associated with
Expression of Melan-A/MART-1 in Uveal and
Cutaneous Melanoma Cell Lines
To characterize further the two subsets of melanoma cell lines,
we next investigated whether we could correlate the observed
differences in Melan-A/MART-1 protein and mRNA expression
with DNA methylation patterns. To analyze the methylation
TABLE 2. Characteristics and Division of Melanoma Cell Lines into Two Panels
Panel I
Melan-A/MART-1
Positive
Protein expression
Yes
mRNA expression
Yes
Methylation pattern
HpaIIm/NruIu
Promoter activity
Yes
MAA mRNA expression (Tyr, Tyrp1*, Dct, gp100/Pmel17) Yes
Cutaneous melanoma cell lines
136.2, 453A0, 453B, 513D, MA, MO, Mu89
Uveal melanoma cell lines
92-1, 92-2, Mel270, OCM1, OCM3, OMM1, OMM1.3, OMM1.5
m, methylated restriction site; u, unmethylated restriction site.
* Minute expression was observed for Tyrp1 in panel II cell lines.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
Panel II
Negative
No
No
HpaIIu/m/NruIm
No
No
EW, IGR39D, Mu96
Mel285, Mel290
IOVS, January 2005, Vol. 46, No. 1
status of the Melan-A/MART-1 gene, we performed Southern
blot analyses on genomic DNA by using methylation-sensitive
restriction enzymes. Figure 2A shows two representative samples of Melan-A/MART-1– expressing and – deficient melanoma
cell lines of uveal and cutaneous origin, and two melanocyte
cultures derived from the skin. Figures 2B and 2C depict a
schematic representation of the Melan-A/MART-1 locus and
the region that was investigated, respectively. The latter encompasses the region of the Melan-A/MART-1 gene, which is
involved in high promoter activity, and includes the position of
the restriction sites and the probe that was used for restriction
fragment detection.
The panel I melanoma cell lines (Fig. 2A, left panel), show
methylation of the HpaII site in the 5⬘ upstream region (lane
2), as they lacked the typical 3640-bp HpaII-HindIII fragment
(compare with control digestions on the melanoma cell lines
with the HpaII methylation unsensitive isoschizomer MspI,
lane 3). The observed fragments in lane 2 were the result of
partial methylation of upstream HpaII sites. Furthermore, they
showed unmethylated DNA at the NruI site in intron 1, where
a 786-bp NruI-HindIII fragment was observed (lane 4).
In contrast, the panel II melanoma cell lines (Fig. 2A, middle
panel), showed methylation of the intronic NruI site (they lack
the 786-bp fragment, lane 4) in all cell lines. Digestion of the
genomic DNA with HpaII (lane 2) resulted in a mixture of
patterns. For the cutaneous melanoma cell lines (EW, IGR39D),
partial methylation was observed, which generated 3640-bp
fragment (lane 2), indicating changes in methylation of the
upstream region when compared with panel I cell lines. In the
panel II uveal melanoma cell lines (Mel285, Mel290), the HpaII
site was predominantly methylated. Notably, the methylation
pattern for the two melanocyte cell lines, M0101 and M9228
(Fig. 2A, right panel, no isoschizomer control digestion), was
similar to that observed in panel I melanoma cell lines: methylation of the HpaII site (lane 2) and an unmethylated NruI site
(lane 4). Based on these DNA methylation analyses, showing a
Shared Expression of Antigens in Melanoma
27
contrasting pattern at the intronic NruI site, the division into
the two panels of melanoma cell lines was highlighted.
Exogenous Melan-A/MART-1 Promoter Activity
and Endogenous Melan-A/MART-1
Expression Levels
To test for the possible involvement of DNA methylation in the
transcriptional regulation of the Melan-A/MART-1 gene we
generated a set of promoter-reporter constructs (Fig. 3A). The
smallest construct, p286, contained virtually all the Melan-A/
MART-1 exon1 sequences, since, as previously established, this
exon comprises the Melan-A/MART-1 regulatory sequences.32
The other constructs contained sequences up- and/or downstream of this core promoter region, comprising the intron 1
region (Fig. 3A). The intron 1 region contains the NruI site that
was analyzed in the DNA methylation assay, and for which a
difference in methylation pattern was observed between all
panel I and II melanoma cell lines. Each of these (unmethylated) constructs was first tested in two panel I cell lines (a
uveal, 92-2, and a cutaneous, Mu89, melanoma cell line) and a
panel II cell line (cutaneous, Mu96) for transcription activation
of the luciferase reporter gene in transient transfection assays
(Fig. 3B). If DNA methylation is the principle mechanism for
Melan-A/MART-1 silencing, promoter activity of the unmethylated exogenous promoter should be observed in both panels of
cell lines.
The panel I cell lines (Mu89, 92-2) showed activity of the
core promoter construct, p286. Extension of this region with
upstream sequences (p1373) resulted in significant downregulation of promoter activity and the intron 1 sequence itself
(p1593, containing the NruI site) showed no activity at all. The
highest activity was observed for construct p2956 containing
the complete 3-kb promoter region, including the intronic
NruI site. In contrast, the panel II cell line (Mu96), showed no
significant activity for any of the constructs. Together, these
FIGURE 2. DNA methylation assay.
(A) Southern blot analyses of DNA
digested with methylation-sensitive
restriction enzymes are shown for 10
cell cultures. Cell lines are indicated
below the blots, and restriction fragments are indicated on the left. Lane
1: HindIII, lane 2: HindIII⫹HpaII,
lane 3: HindIII⫹MspI (HpaII isoschizomer control), lane 4:
HindIII⫹NruI. The absence of the
786-bp fragment in the HindIII/NruI
lanes is indicative of methylated DNA
at the NruI site. The absence of the
3640-bp fragment in the HindIII/
HpaII lanes is indicative of methylated DNA at the HpaII site. For the
melanoma cell lines, MspI was used
as a control for HpaII (identical recognition sites) and digested the DNA
at methylated and unmethylated
sites. Partial digestion for HpaII indicates differential methylation of the
upstream region. (B) Genome organization of the Melan-A/MART-1
gene. The four introns (unshaded
boxes) and five exons (shaded
boxes) are indicated as well as the
mRNA coding for the ⬃13-kDa protein (top). The 5⬘-region of the locus
and part of the upstream region is depicted. The core promoter was localized to exon 1. Arrow: the transcription start site. (C) A schematic
overview of the HindIII region. The approximately 3-kb promoter region is indicated above and in the HindIII region. Restriction sites and fragment
sizes are indicated as well as the probe used for restriction fragment detection (black bar) and its location in relation to the Melan-A/MART-1 locus.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
28
van Dinten et al.
IOVS, January 2005, Vol. 46, No. 1
data suggest an important role for the intronic region of the
Melan-A/MART-1 gene in promoter activation.
Because the p2956 construct showed the highest activity
and contains the region where a difference in DNA methylation
pattern was observed, we tested this (unmethylated) construct
for promoter activity in the complete panel of cell lines. In
both uveal and cutaneous panel I melanoma cell lines (Fig. 3C)
p2956 promoter activity was observed, albeit at various levels
(3000 –183,000 relative luciferase units [RLU]/s). In contrast,
panel II cell lines showed no to background levels of p2956
promoter activity (Fig. 3C, ⬍900 RLU/s). The results of these
promoter activation studies corroborate the division of melanoma cell lines into two different panels on the basis of MAA
expression characteristics.
DISCUSSION
FIGURE 3. Transient transfection assays. (A) Set of promoter-reporter
constructs. The promoter region is indicated over the constructs. The
luciferase gene is depicted in black, the upstream region in light
shading, the exon 1 (core promoter region) in dark shading, and the
intron 1 region, containing the NruI site, in white. Arrow: transcriptional start site. (B) Transfection of a uveal (92-2) and a cutaneous
(Mu89) panel I and a cutaneous (Mu96) panel II melanoma cell line
with the complete set of promoter-reporter constructs. Luciferase
values are depicted in relative luciferase units/s (RLU/s). (C) Uveal and
cutaneous melanoma cell lines transfected with construct p2956.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
The data presented in this study, show that two subsets of
melanoma cell lines can be discerned on the basis of MAA
expression: panel I, melanoma cell lines that display MAA
expression, and panel II, melanoma cell lines that are deficient
in MAA expression (Table 2). The origin of the tumor cell lines
(primary or metastatic) was not found to correlate with their
classification as a panel I or II cell line (Table 1, 2), in the sense
that not all panel II cell lines were derived from metastases nor
were all panel I cell lines derived from primary tumor material.
Moreover, the uveal melanoma cell lines OMM1.3 and -1.5 are
derived from different liver metastases from the same patient as
the primary Mel270 uveal melanoma cell line, and these three
cell lines showed identical MAA expression characteristics.
Because one of the arguments for metastasis is that immune
detection is evaded by silencing of tumor antigens, allowing
settlement in a distant organ, the finding that Melan-A/MART-1
silencing is not correlated with metastatic behavior presents a
paradox. One possible explanation is that these features
change (not necessarily sequentially) during the process of
transformation of normal cells into often-metastatic cancer
cells, as discussed by Bernards and Weinberg.33 This process
relies on an apparent random succession of genetic and epigenetic changes that occur relatively early in tumorigenesis and
that alter the expression profiles of a myriad of genes. Because
of these successive changes in gene expression profiles, individual neoplastic cells acquire selective advantages in the process of tumor progression.33,34 It infers heterogeneity in expression profiles of specific genes in subsets of cells within the
original primary tumor cell population.33 It also infers that the
expression profiles of specific genes in different tumor metastasis could differ from those of cells in the primary tumor mass,
whereas the overall gene-expression patterns may be strikingly
similar.33,35
Of interest is the notion that, at the level of DNA encompassing the Melan-A/MART-1 regulatory region, there is a clear
difference in DNA methylation pattern of the intronic NruI
site, between melanoma cells that express this antigen and
cells that are deficient for the antigen. At the same time, cells
lacking the Melan-A/MART-1 antigen also failed to support the
activity of an exogenous, unmethylated, Melan-A/MART-1 promoter-reporter construct, as demonstrated in transient transfection assays (see Fig. 3). This is in line with previous reports28,32 suggesting that the lack of Melan-A/MART-1
expression is caused by absence of transcription factors involved in Melan-A/MART-1 promoter activation.
Recent work on the Melan-A/MART-1 and Pmel17 promoters36 and work on the Tyrp1 and Dct promoters37 showed the
involvement of the master regulator of melanocyte development, microphthalmia-associated transcription factor (Mitf38),
in their transcriptional regulation. Although Mitf is capable of
Shared Expression of Antigens in Melanoma
IOVS, January 2005, Vol. 46, No. 1
activating the exogenous Tyr and Tyrp1 promoters, Mitf is not
sufficient to re-induce endogenous expression once these
genes have been silenced.39 These data suggest a role for
chromatin remodeling in combination with specific transcription factors in the transcriptional regulation of these genes in
vivo.39 Our data are in line with these speculations, because
differences in DNA methylation patterns are observed in vivo,
together with the absence of required transcription factors in
panel II cells (Fig. 3). As a whole, our results from the promoter
reporter assays and DNA methylation studies indicate the existence of multiple regulatory regions and mechanisms that
control MAA expression and point toward a transcriptionfactor–regulated silencing process during the conversion of
normal melanocytes to melanoma cells, with a putative role for
chromatin remodeling.
For the analyses described in this article, we used a panel of
cell lines in view of future reverse genetic approaches and
re-expression strategies for MAAs. Because cell lines are derived from tumor tissue by single cell dilutions and subsequent
clonal expansion, panel I cell lines must be derived from
MAA-expressing homogeneous or heterogeneous tumor
masses. At the same time, panel II cell lines must be derived
from MAA-deficient or heterogeneous tumors or must have
arisen during expansion in tissue culture. Of the 20 cell lines
analyzed, only one panel II cell line, Mu96, arose in tissue
culture on high-density growth and TIL immunoselection.28
Cell line 92-2 was also derived from another cell line (92-1) in
tissue culture. However, this cell line retained the panel I
phenotype (Table 1, 2) but, in contrast to its ancestor, was
unable to form metastases in mice (Jager et al., unpublished
observation, 2003). Whether cell lines are more prone to undergoing changes in tissue culture when derived from heterogeneous tumor masses than those derived from homogeneous
tumor masses can only be speculated on, because the immunophenotype of the original tumor masses could not be traced.
On the whole, our data, summarized in Table 2, show the
division of patient-derived melanoma cell lines into two panels,
based on MAA expression. At the same time, differences in
methylation patterns were observed that have to be confirmed
in freshly isolated tumor cells, to investigate the possibility of
therapeutical strategies whereby DNA methylation can be limited or blocked in vivo. Finally, a distinction between uveal and
cutaneous melanoma cell lines could not be established, indicating that identical transformation processes take place in
uveal and cutaneous melanocytes and that a general approach
can be used concerning MAA-based immunotherapy strategies.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Acknowledgments
20.
The authors thank Peter I. Schrier, Marieke Griffioen, James T. Kurnick,
Bruce R. Ksander, June Kan-Mitchel, and Gregorius P. M. Luyten for the
donation of melanoma cell lines and Nienke van der Stoep for a critical
reading of the manuscript and valuable comments.
21.
References
22.
1. Zhai Y, Yang JC, Kawakami Y, et al. Antigen-specific tumor
vaccines: development and characterization of recombinant adenoviruses encoding MART1 or gp100 for cancer therapy. J Immunol. 1996;156:700 –710.
2. Gattoni-Celli S, Cole DJ. Melanoma-associated tumor antigens and
their clinical relevance to immunotherapy. Semin Oncol. 1996;23:
754 –758.
3. Cormier JN, Hijazi YM, Abati A, et al. Heterogeneous expression of
melanoma-associated antigens and HLA-A2 in metastatic melanoma
in vivo. Int J Cancer. 1998;75:517–524.
4. de Vries TJ, Fourkour A, Wobbes T, Verkroost G, Ruiter DJ, van
Muijen GN. Heterogeneous expression of immunotherapy candidate proteins gp100, MART-1, and tyrosinase in human melanoma
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
23.
24.
25.
29
cell lines and in human melanocytic lesions. Cancer Res. 1997;57:
3223–3229.
Komenaka I, Hoerig H, Kaufman HL. Immunotherapy for melanoma. Clin Dermatol. 2004;22:251–265.
Boissy RE. The melanocyte: its structure, function, and subpopulations in skin, eyes, and hair. Dermatol Clin. 1988;6:161–173.
Ksander BR. The tumor-specific T cell response to ocular melanomas. In: Zierhut M, Thiel HJ, eds. Immunology of the Skin and
Eye. Buren, The Netherlands: Aeolus Press, 1998:269 –288.
Kawakami Y, Eliyahu S, Delgado CH, et al. Identification of a
human melanoma antigen recognized by tumor-infiltrating lymphocytes associated with in vivo tumor rejection. Proc Natl Acad
Sci USA. 1994;91:6458 – 6462.
Coulie PG, Brichard V, van Pel A, et al. A new gene coding for a
differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas. J Exp Med. 1994;180:35– 42.
Kawakami Y, Eliyahu S, Delgado CH, et al. Cloning of the gene
coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumor. Proc Natl Acad Sci USA.
1994;91:3515–3519.
Brichard V, van Pel A, Wolfel T, et al. The tyrosinase gene codes for
an antigen recognized by autologous cytolytic T lymphocytes on
HLA-A2 melanomas. J Exp Med. 1993;178:489 – 495.
Robbins PF, El Gamil M, Kawakami Y, Stevens E, Yannelli JR,
Rosenberg SA. Recognition of tyrosinase by tumor-infiltrating lymphocytes from a patient responding to immunotherapy. Cancer
Res. 1994;54:3124 –3126.
Cohen T, Muller RM, Tomita Y, Shibahara S. Nucleotide sequence
of the cDNA encoding human tyrosinase-related protein. Nucleic
Acids Res. 1990;18:2807–2808.
Wang RF, Robbins PF, Kawakami Y, Kang XQ, Rosenberg SA.
Identification of a gene encoding a melanoma tumor antigen recognized by HLA-A31-restricted tumor-infiltrating lymphocytes. J
Exp Med. 1995;181:799 – 804.
Yokoyama K, Suzuki H, Yasumoto K, Tomita Y, Shibahara S.
Molecular cloning and functional analysis of a cDNA coding for
human DOPAchrome tautomerase/tyrosinase-related protein-2.
Biochim Biophys Acta. 1994;1217:317–321.
Wang RF, Appella E, Kawakami Y, Kang X, Rosenberg SA. Identification of TRP-2 as a human tumor antigen recognized by cytotoxic T lymphocytes. J Exp Med. 1996;184:2207–2216.
van der Bruggen P, Traversari C, Chomez P, et al. A gene encoding
an antigen recognized by cytolytic T lymphocytes on a human
melanoma. Science. 1991;254:1643–1647.
Gaugler B, van den Eynde B, van der Bruggen P, et al. Human gene
MAGE-3 codes for an antigen recognized on a melanoma by autologous cytolytic T lymphocytes. J Exp Med. 1994;179:921–930.
Castelli C, Rivoltini L, Andreola G, Carrabba M, Renkvist N, Parmiani G. T-cell recognition of melanoma-associated antigens. J Cell
Physiol. 2000;182:323–331.
Berset M, Cerottini JP, Guggisberg D, et al. Expression of MelanA/MART-1 antigen as a prognostic factor in primary cutaneous
melanoma. Int J Cancer. 2001;95:73–77.
Kageshita T, Kawakami Y, Ono T. Clinical significance of MART-1
and HLA-A2 expression and CD8⫹ T cell infiltration in melanocytic
lesions in HLA-A2 phenotype patients. J Dermatol Sci. 2001;25:
36 – 44.
Hofbauer GF, Kamarashev J, Geertsen R, Boni R, Dummer R. Melan
A/MART-1 immunoreactivity in formalin-fixed paraffin-embedded
primary and metastatic melanoma: frequency and distribution.
Melanoma Res. 1998;8:337–343.
Dalerba P, Ricci A, Russo V, et al. High homogeneity of MAGE,
BAGE, GAGE, tyrosinase and Melan-A/MART-1 gene expression in
clusters of multiple simultaneous metastases of human melanoma:
implications for protocol design of therapeutic antigen-specific
vaccination strategies. Int J Cancer. 1998;77:200 –204.
Kageshita T, Kawakami Y, Hirai S, Ono T. Differential expression
of MART-1 in primary and metastatic melanoma lesions. J Immunother. 1997;20:460 – 465.
Hurks HM, Valter MM, Wilson L, Hilgert I, van den Elsen PJ, Jager
MJ. Uveal melanoma: no expression of HLA-G. Invest Ophthalmol
Vis Sci. 2001;42:3081–3084.
30
van Dinten et al.
26. Versteeg R, Noordermeer IA, Kruse-Wolters M, Ruiter DJ, Schrier
PI. c-myc down-regulates class I HLA expression in human melanomas. EMBO J. 1988;7:1023–1029.
27. Ramirez-Montagut T, Andrews DM, Ihara A, et al. Melanoma antigen recognition by tumour-infiltrating T lymphocytes (TIL): effect
of differential expression of Melan-A/MART-1. Clin Exp Immunol.
2000;119:11–18.
28. Kurnick JT, Ramirez-Montagut T, Boyle LA, et al. A novel autocrine
pathway of tumor escape from immune recognition: melanoma
cell lines produce a soluble protein that diminishes expression of
the gene encoding the melanocyte lineage melan-A/MART-1 antigen through down-modulation of its promoter. J Immunol. 2001;
167:1204 –1211.
29. Pavel S, Smit NP, van der Meulen H, et al. Homozygous germline
mutation of CDKN2A/p16 and glucose-6-phosphate dehydrogenase deficiency in a multiple melanoma case. Melanoma Res.
2003;13:171–178.
30. Zuidervaart W, van der Velden PA, Hurks MH, et al. Gene expression profiling identifies tumour markers potentially playing a role
in uveal melanoma development. Br J Cancer. 2003;89:1914 –
1919.
31. Chen C, Okayama H. High-efficiency transformation of mammalian
cells by plasmid DNA. Mol Cell Biol. 1987;7:2745–2752.
Downloaded From: http://iovs.arvojournals.org/ on 06/17/2017
IOVS, January 2005, Vol. 46, No. 1
32. Butterfield LH, Stoll TC, Lau R, Economou JS. Cloning and analysis
of MART-1/Melan-A human melanoma antigen promoter regions.
Gene. 1997;191:129 –134.
33. Bernards R, Weinberg RA. A progression puzzle. Nature. 2002;
418:823.
34. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;
100:57–70.
35. van der Velden PA, Zuidervaart W, Hurks HM, et al. Expression
profiling reveals that methylation of TIMP3 is involved in uveal
melanoma development. Int J Cancer. 2003;106:472– 479.
36. Du J, Miller AJ, Widlund HR, Horstmann MA, Ramaswamy S, Fisher
DE. MLANA/MART1 and SILV/PMEL17/GP100 are transcriptionally
regulated by MITF in melanocytes and melanoma. Am J Pathol.
2003;163:333–343.
37. Bertolotto C, Abbe P, Hemesath TJ, et al. Microphthalmia gene
product as a signal transducer in cAMP-induced differentiation of
melanocytes. J Cell Biol. 1998;142:827– 835.
38. Goding CR. Mitf from neural crest to melanoma: signal transduction and transcription in the melanocyte lineage. Genes Dev.
2000;14:1712–1728.
39. Gaggioli C, Busca R, Abbe P, Ortonne JP, Ballotti R. Microphthalmia-associated transcription factor (MITF) is required but is not
sufficient to induce the expression of melanogenic genes. Pigment
Cell Res. 2003;16:374 –382.