1) Experiments compared the ability of eye and skin melanoma cells to stimulate T cell proliferation. Primary skin melanoma cells stimulated significant T cell proliferation, while primary eye melanoma cells failed to induce proliferation. 2) Eye melanoma cells expressed normal levels of HLA class I and II molecules but still failed to stimulate T cells. In contrast, eye melanoma cells inhibited T cell proliferation in mixed lymphocyte cultures through direct cell contact. 3) The ability of eye melanoma cells to inhibit T cell proliferation was lost when the cells metastasized from the eye to other sites like the liver, suggesting the unique eye microenvironment alters the immunogenicity of melanoma cells developing there.

1. MELANOMAS THAT DEVELOP WITHIN THE EYE INHIBIT
LYMPHOCYTE PROLIFERATION
David J. VERBIK1*, Timothy G. MURRAY2, Johan M. TRAN1 and Bruce R. KSANDER1
1The Schepens Eye Research Institute and Department of Ophthalmology, Harvard Medical School, Boston, MA, USA
2Bascom Palmer Eye Institute, Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, FL, USA
Experiments were performed to compare the ability of
ocular and skin melanoma cells to stimulate T cells. Primary
melanoma cell lines were obtained from a series of patients
with either eye or skin melanoma. The ability of tumor cells
to stimulate T cells in the absence of exogenous growth
factors was assessed in mixed-lymphocyte tumor cell cultures
in which allogeneic lymphocytes were stimulated with irradi-
ated ocular or skin melanoma cells. Expression of HLA class I
and class II on tumor cells, in the presence or absence of
IFN-␥, was determined by flow cytometry. The ability of
tumor cells to inhibit T-cell proliferation was determined by
adding various concentrations of irradiated tumor cells to
standard mixed-lymphocyte cultures. Our results indicate
that primary skin melanoma cells induce vigorous prolifera-
tion of allo-antigen-specific T cells. By contrast, ocular mela-
noma cells failed to induce significant T-cell proliferation. The
failure of ocular melanoma cells to stimulate lymphocyte
proliferation was not due to low levels of either class I or class
II on tumor cells since tumor cells treated with IFN-␥
expressed high levels of class I and class II but still failed to
induce lymphocyte proliferation. Ocular melanoma cells inhib-
ited lymphocyte proliferation, as shown by experiments in
which a small number of tumor cells prevented proliferation
of T cells in mixed-lymphocyte cultures. Inhibition of lympho-
cyte proliferation required cell-to-cell contact, and superna-
tants from tumor cell cultures did not prevent lymphocyte
proliferation. Moreover, the ability of ocular melanoma cells
to inhibit T-cell proliferation was lost when tumor cells
migrated from the eye and formed hepatic metastases. We
conclude that there is a fundamental difference in the immu-
nogenicity of ocular and skin melanoma cells. Ocular melano-
mas, but not primary skin melanomas, are poorly immuno-
genic tumors that inhibit T-cell proliferation. Our results
imply that the immunogenicity of melanoma cells is altered
when they develop within the unique ocular micro-environ-
ment. Int. J. Cancer 73:470–478, 1997.
௠ 1997 Wiley-Liss, Inc.
Although both eye and skin melanomas are derived from
melanocytes, the clinical presentation and disease progression of
ocular melanomas are distinct from primary melanomas that
develop within the s.c. tissues of the skin. Ocular melanomas are
typically small, slow-growing tumors that do not progress through
radial and vertical growth phases. The incidence of metastatic
spread of tumor cells from patients with large ocular tumors is high,
estimated at 50% (Shields, 1983). Metastases typically are ob-
served first in the liver and can follow a protracted disease-free
interval, which is an unusual characteristic of this disease. How-
ever, once metastatic tumors are detected clinically, they grow at an
accelerated pace and are resistant to conventional forms of therapy
(Rajpal et al., 1983). It is unknown why the malignant transforma-
tion of melanocytes within the choroid of the eye results in a
different clinical disease pattern.
In our previous experiments, we attempted to activate CD8ϩ
cytotoxic T cells specific for ocular melanoma cells by restimulat-
ing peripheral blood lymphocytes with autologous tumor cells in
the presence of exogenous IL-2. Although this technique has been
highly successful in activating cytotoxic T cells specific for skin
melanoma cells, we were unable to stimulate CD8ϩ T cells specific
for ocular melanoma cells using this method (data not shown). The
present experiments were undertaken to determine if this was due
to a fundamental difference in the ability of eye and skin melanoma
cells to stimulate T cells. We hypothesized that melanomas that
develop within the unique ocular micro-environment are unable to
stimulate T cells. The original studies that examined the immunoge-
nicity of skin melanoma cells utilized mixed lymphocyte tumor cell
(MLTC) cultures in which allogeneic lymphocytes were stimulated
with irradiated tumor cells in the absence of exogenous lympho-
kines. These experiments were performed using skin melanoma
cells obtained from either primary or metastatic tumor lesions and
revealed that primary, but not metastatic, tumor cells were able to
induce proliferation of allo-antigen-specific T cells (Alexander et
al., 1989; Taramelli et al., 1988; Fossati et al., 1984). The failure of
metastatic tumor cells to stimulate T cells was associated with
tumor cell–mediated suppression of T cells (Taramelli et al., 1984).
In the present experiments, we used similar methods to compare
the ability of eye and skin melanomas to stimulate T-cell prolifera-
tion. Our results indicate that, unlike primary skin melanomas,
ocular melanomas are poorly immunogenic tumors that inhibit
T-cell proliferation. The ability to inhibit T-cell proliferation is lost
when ocular melanoma cells migrate from the eye, suggesting that
melanomas acquire lymphocyte-inhibitory properties within the
ocular micro-environment.
MATERIAL AND METHODS
Tumor cell lines
Our studies were performed using ocular melanoma cells
(primary and metastatic) and primary skin melanoma cells. Primary
ocular melanoma cell lines (Mel 202, 203, 270, 285 and 290) were
derived from patients whose eyes were enucleated at the Bascom
Palmer Eye Institute (Miami, FL). After enucleation, tumor-
containing eyes were trans-illuminated in the operating room to
determine the exact position of the tumor. The eye was dissected
into 2 portions. The first portion was used for histological
examination, and the second portion was used to generate tumor
cell cultures. Tumor tissue was dissected form the surrounding
normal uveal tissue and enzymatically digested using collagenase
to yield a single-cell suspension, as previously described (Ksander
et al., 1991). Briefly, tumor tissue was minced in a Petri dish
containing 10 ml of collagenase type IV (Sigma, St. Louis, MO) at
150 U/ml and hyaluronidase type V 0.01% (Sigma) in RPMI 1640
medium (BioWhittaker, Walkersville, MD) supplemented with
10% FCS (Hyclone, Logan, UT), 2 mM L-glutamine (BioWhit-
taker), 100 U/ml penicillin and 100 µg/ml streptomycin (BioWhit-
taker), 0.1% fungizone (BioWhittaker), HEPES (0.01 M) and
2-␤-mercaptoethanol (0.5%). Tumor tissue was incubated for 90
min at 37°C. Following incubation, tumor cells were recovered
from the culture supernatant, washed 3 times and examined
microscopically for viable tumor cells. These tumor cells were used
to generate cell lines, which were maintained in culture medium at
37°C and 5% CO2.
Contract grant sponsor: United States Public Health Service; Contract
grant number: EY09294; Contract grant sponsor: Fight for Sight, Inc.
*Correspondence to: The Schepens Eye Research Institute, 20 Staniford
St., Boston, MA 02114, USA. Fax: (617) 720-1069. E-mail:
verbik@vision.eri.harvard.edu
Received 7 March 1997; Revised 26 June 1997
Int. J. Cancer: 73, 470–478 (1997)
௠ 1997 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
Publication de l’Union Internationale Contre le Cancer

2. Metastatic ocular melanoma cells were obtained from patient
270, 3 years after removal of the primary tumor. Four cell lines
were established from 4 different liver biopsies of different tumor
nodules. Biopsies were obtained using standard laparoscopic
techniques. Tumor tissue was treated with collagenase as described
above to obtain tumor cells. Cells were maintained in culture
medium at 37°C and 5% CO2.
The 2 primary skin melanoma cell lines (WM 115 and WM 793),
provided by Dr. M. Herlyn’s laboratory, were derived from vertical
growth phase primary human melanomas (Herlyn et al., 1985).
Collection of peripheral blood lymphocytes from normal donors
Peripheral blood lymphocytes were collected by venepuncture
using heparinized vacutainers from HLA-mis-matched normal
healthy donors. Lymphocytes were isolated using Histopaque 1077
separation medium (Sigma). Briefly, whole blood was layered over
lymphocyte separation medium and centrifuged at 800 g for 15
min. After centrifugation, lymphocytes were removed from the
interface, washed twice in HBSS and resuspended in culture
medium.
HLA tissue typing
The HLA phenotype of ocular melanoma patients was deter-
mined at the Massachusetts General Hospital Tissue Typing
Laboratories, using peripheral blood lymphocytes and the Amos
modified microcytotoxicity test. HLA phenotypes of patients and
donors were as follows: patient 202 (A1, A3, B38, B52), patient
203 (A1, A25, B51, B62, C1, C3), patient 270 (A11, A29, B7,
B52), patient 285 (A24, B7, B53), normal donor 1 (A1, A23, B8,
B44, C4) and normal donor 2 (A11, A33, B35, B46, C4).
MLTC cultures
Allogeneic responder lymphocytes (1 ϫ 105) were cultured with
various concentrations of irradiated (10,000 R) tumor cells in a
round-bottomed 96-well microtiter plate (Beckton Dickinson,
Lincoln Park, NJ) in 200 µl of culture medium. For positive and
negative controls, responder lymphocytes (1 ϫ 105) were cultured
with irradiated (4,000 R) allogeneic or autologous lymphocytes
(1 ϫ 105), respectively. Cultures were incubated at 37°C for 6 days
and pulsed with 3H-thymidine (2 µCi/well) during the final 18 hr.
After incubation, cells were harvested onto glass fiber filters and
radioactivity was measured using a liquid scintillation counter.
Incorporation of 3H-thymidine was calculated as follows: ⌬ cpm ϭ
mean cpm in test wells Ϫ mean cpm of stimulator cells only.
Results are presented as the average ⌬ cpm of 4 test wells Ϯ
standard error of the mean. All experiments were repeated at least
twice, and data from a single representative experiment are
displayed.
Expression of HLA class I and class II
HLA class I and II expression on ocular melanoma cells and
primary skin melanoma cells was determined by flow cytometry
using the Coulter Epics LX analyzer (Coulter, Hialeah, FL). Tumor
cells (1 ϫ 106) were stained with FITC-conjugated mouse anti-
human IgG1, ␬ monoclonal antibodies (MAbs), clone G46-2.6
(PharMingen, San Diego, CA) specific for a non-polymorphic
region of HLA-A, -B and -C class I or FITC-conjugated mouse
anti-human IgG2a, ␬ MAbs, clone TU39 (PharMingen) specific for
a nonpolymorphic region of HLA-DR, -DP and -DQ class II. As a
negative control, tumor cells were stained with a non-specific
isotype-matched FITC-conjugated MAb (PharMingen). To deter-
mine the level of HLA class I and II expression on IFN-␥-treated
tumor cells, tumor cells were treated with IFN-␥ (1,000 U/ml for 48
hr) and evaluated for class I and class II expression as described
above. Data are presented as the percentage of positively stained
tumor cells and/or the intensity of staining shown as mean channel
fluorescence.
Experiments were performed to determine if up-regulation of
class I and class II on tumor cells would promote lymphocyte
proliferation. Tumor cells were treated with IFN-␥ (1,000 U/ml) for
48 hr, washed, irradiated and added as stimulator cells to MLTC
cultures as described above. Preliminary experiments demonstrated
that pre-treatment of tumor cells with IFN-␥ increased and
sustained class I and class II expression for 6 days.
Tumor cell–mediated inhibition of T-cell proliferation
To determine whether ocular melanoma cells inhibit lymphocyte
proliferation, irradiated (10,000 R) primary or metastatic ocular
melanoma cells were added to a standard mixed-lymphocyte
reaction (MLR) and incubated for 6 days at 37°C. Increasing
numbers of tumor cells, either 5,000, 10,000 or 20,000, were added
when MLR cultures were established. Lymphocyte proliferation
was measured by 3H-thymidine incorporation as described above.
As a negative control, irradiated normal human foreskin fibroblasts
were added at the same concentration to MLR cultures. To
determine if fixed tumor cells inhibit T-cell proliferation, primary
ocular melanoma cells were fixed in 2% formalin in PBS and added
to MLR cultures in a similar fashion. In another series of
experiments, tumor cell culture supernatants were collected at
various time points and added to MLR cultures at a 1:1 ratio (v/v).
Statistical analysis
The statistical method used to determine significance was the
Tukey-Kramer Multiple Comparison Test (Huck et al., 1974). This
test was used to determine whether mean levels of proliferation for
each experimental group were significantly different from the
negative control. Differences were considered significant at p Յ
0.001.
RESULTS
T-cell proliferation to primary eye and skin melanoma cells
The first series of experiments examined the ability of primary
eye and skin melanoma cells to stimulate proliferation of allo-
antigen-specific T cells. Ocular melanoma cell lines were estab-
lished from 5 different primary eye tumors (Mel 202, 203, 270, 285
and 290). The HLA phenotype of the ocular melanoma patients was
determined using peripheral blood lymphocytes and standard
tissue-typing methods. This information was used to identify an
HLA-mis-matched donor as a source of responding lymphocytes.
Lymphocytes from this donor were stimulated with increasing
numbers of irradiated melanoma cells (ranging from 2,500 to
100,000 cells) in the absence of exogenous lymphokines. Cultures
were incubated for 6 days, and lymphocyte proliferation was
measured by 3H-thymidine incorporation. As a positive control,
peripheral blood lymphocytes were stimulated with a similar
number of irradiated allogeneic lymphocyte stimulator cells in a
standard MLR. As a negative control, lymphocytes were stimulated
with irradiated autologous stimulator cells. Results from the first
series of experiments confirmed the original observations of
Fossati et al. (1984) and indicate that primary skin melanoma cells
stimulate significant proliferation of allogeneic lymphocytes (Fig.
1a). By contrast, when similar experiments were performed using
ocular melanoma cells from 5 different patients, no significant
lymphocyte proliferation was detected (Fig. 1b). We conclude that
primary ocular melanoma cells are unable to stimulate proliferation
of allo-antigen-specific T cells.
Expression of class I and class II on eye and skin melanoma cells
The failure of ocular melanomas to stimulate allo-antigen-
specific T cells may be due to low expression of HLA class I and
class II. To compare the expression of class I and class II on eye and
skin melanoma cells, the tumor cell lines used in the previous
experiments were stained with MAbs specific for HLA class I or
class II as described in ‘‘Material and Methods’’. Flow cytometry
was used to determine the percentage of positively stained tumor
cells, and the intensity of staining was measured by the mean
channel fluorescence (mcf). Ocular and skin melanoma cells from
471OCULAR MELANOMAS AND T CELLS

3. all patients expressed class I on essentially 100% of tumor cells
(Table I). The staining intensity on ocular melanoma cells ranged
4–15 mcf and was within the range observed for skin melanoma
cells (8–17 mcf). Treatment of tumor cells with IFN-␥ (1,000 U/ml
for 48 hr) increased the intensity of class I on ocular melanoma
cells (9–98 mcf) and skin melanoma cells (59–115 mcf). Class II
was expressed at low levels on skin melanoma cells (WM 115, 65%
2 mcf; WM 793, 89% 10 mcf) but not expressed on ocular
FIGURE 1 – Proliferation of lymphocytes stimulated with allogeneic melanoma cells. Lymphocytes (1 ϫ 105) were cultured for 6 days with
increasing concentrations of irradiated (10,000 R) tumor cells from either primary skin melanomas (a) or primary ocular melanomas (b). Positive
and negative controls included lymphocytes cultured with irradiated allogeneic or autologous lymphocytes, respectively. *Significant proliferation
at p Ͻ 0.001 when compared to the negative control.
472 VERBIK ET AL.

4. melanomas. Treatment of ocular melanoma cells with IFN-␥
increased class II expression on tumor cells from patient 290 to
within the range observed on skin melanoma cells (92% 5 mcf).
T-cell proliferation to IFN-␥-treated tumor cells
The following experiments were performed to determine whether
up-regulating class I and class II expression would restore the
ability of ocular melanoma cells to stimulate proliferation of
allogeneic T cells. Eye and skin melanoma cells were treated with
IFN-␥ (1,000 U/ml for 48 hr) prior to being used as stimulator cells
in MLTC cultures. Preliminary experiments demonstrated that
pre-treatment of tumor cells with IFN-␥ resulted in sustained HLA
expression for 6 days. Tumor cells from all patients listed in Table I
were analyzed. Representative results from 2 skin and 2 eye
melanoma patients are shown in Figure 2a and b, respectively. As
in the previous experiments, skin melanoma cells (not treated with
IFN-␥) stimulated lymphocyte proliferation. Lymphocyte prolifera-
tion against IFN-␥-treated primary skin melanoma cells was not
significantly greater than the levels seen using untreated tumor
cells (Fig. 2a). Ocular melanoma cells (not treated with IFN-␥)
failed to stimulate lymphocyte proliferation, and treatment with
IFN-␥ failed to restore lymphocyte proliferation (Fig. 2b). This
occurred even though ocular melanoma cells from patient 290
expressed levels of class I and class II greater than those expressed
by skin melanoma cells that induced lymphocyte proliferation
(WM 115 not treated with IFN-␥, Table I). These results indicate
that the failure of ocular melanoma cells to stimulate proliferation
of allo-antigen-specific T cells cannot be attributed to low expres-
sion of class I or class II on these tumor cells.
Ocular melanoma cells inhibit T-cell proliferation
The following experiments were performed to determine if
ocular melanoma cells inhibit T-cell proliferation. Increasing
numbers of irradiated tumor cells were added to standard MLR
cultures in which lymphocytes were stimulated with irradiated
allogeneic peripheral blood lymphocytes. Ocular melanoma cells
were added, as regulator cells, when cultures were established. As a
negative control, irradiated human foreskin fibroblasts were added
to MLR cultures at the same concentration as ocular melanoma
cells. The results are displayed in Figure 3. In the absence of
regulator cells, lymphocytes proliferated vigorously to allogeneic
stimulator cells. The addition of increasing numbers of irradiated
normal human fibroblasts had no effect on lymphocyte prolifera-
tion. By contrast, the addition of as few as 10,000 ocular melanoma
cells significantly decreased lymphocyte proliferation. The inhibi-
tory properties of ocular melanoma cells were observed consis-
tently in tumor cells obtained from all 5 patients.
The following experiments were performed to determine whether
the inhibition of T-cell proliferation required cell-to-cell contact or
if tumor cells secrete an immunosuppressive factor. Supernatants
were recovered from cultured tumor cells at different time points,
as described in ‘‘Material and Methods’’, and added to standard
MLR cultures, as in the previous experiment. Tumor cell superna-
tants did not significantly reduce T-cell proliferation (data not
shown). We conclude that ocular melanoma cells do not secrete a
factor that inhibits T-cell proliferation. To determine if fixed tumor
cells expressed a cell-surface protein that inhibits T-cell prolifera-
tion, ocular melanoma cells were fixed as described in ‘‘Material
and Methods’’. The addition of increasing numbers of fixed tumor
cells did not inhibit T-cell proliferation (Fig. 4). These results
indicate that viable ocular melanoma cells inhibit T-cell prolifera-
tion by cell-to-cell contact.
Metastatic ocular melanoma cells fail to inhibit T-cell
proliferation
The next series of experiments were performed to determine if
ocular melanoma cells that migrate from the eye and establish
metastatic hepatic tumors inhibit T-cell proliferation. Metastatic
tumor cells were recovered from several different liver tumor
nodules from patient 270, and tumor cell lines were established.
These metastatic tumor cells were irradiated and added to MLR
cultures as in the previous experiments. Surprisingly, metastatic
tumor cells failed to inhibit T-cell proliferation, even though the
original primary tumor cells were potent inhibitors of T-cell
proliferation (Fig. 5). Tumor cells recovered from all 4 metastatic
nodules were consistently unable to inhibit T-cell proliferation to
the extent observed for the primary tumor cells. The single
exception was OMM2.5 at the highest tumor cell concentration
(20,000 tumor cells/well). We conclude that the ability of ocular
melanoma cells to inhibit T-cell proliferation is lost when tumor
cells migrate from the eye and form hepatic metastases.
T-cell proliferation to metastatic ocular melanoma cells
Since metastatic tumor cells failed to inhibit T-cell proliferation
in allo-MLR cultures, we were interested in whether metastatic
tumor cells could directly stimulate proliferation of allogeneic T
cells. Preliminary experiments were conducted to compare the
expression of HLA class I on primary and metastatic tumor cells
from patient Mel 270. Unexpectedly, expression of class I on
metastatic tumor cells was increased compared with the primary
tumor cells (Table II). The mcf of metastatic tumor cells ranged
35–41 and was considerably greater than that of the primary tumor
(4 mcf). The level of class I expression on metastatic tumor cells
increased following treatment with IFN-␥ (1,000 U/ml for 48 hr).
None of the metastatic tumor cells expressed class II on the surface.
These data indicate that metastatic ocular melanoma cells express
sufficient class I to stimulate allogeneic T cells. T-cell proliferation
was measured as in the previous experiments, and the results are
displayed for 2 representative tumor cell lines, OMM2.2 and
OMM2.3 (Fig. 6). The results indicate that metastatic ocular
melanoma OMM2.2 induces significant T-cell proliferation in
MLTC cultures. Following treatment with IFN-␥, both metastatic
cells induced significant T-cell proliferation. We conclude that
metastatic ocular melanoma cells, unlike primary ocular tumors,
have the ability to induce T-cell proliferation.
DISCUSSION
During the past decade, a large number of laboratories have
confirmed that metastatic skin melanoma cells express tumor
antigens that are recognized by autologous CD8ϩ T cells (Van Pel
et al., 1995; Boon et al., 1994; Traversari et al., 1992). Moreover,
these tumor cells can be used to stimulate proliferation of specific T
cells in cultures containing exogenous lymphokines. In a previous
TABLE I – HLA-CLASS I AND CLASS II EXPRESSION ON PRIMARY
MELANOMAS
Primary melanomas
Percent positive cells
(mean channel fluorescence)1
Class I Class II
No IFN-␥ IFN-␥ No IFN-␥ IFN-␥
Ocular
Mel 202 100 (7) 100 (32) 0 0
Mel 203 96 (15) 100 (45) 0 0
Mel 270 100 (4) 100 (9) 0 0
Mel 285 100 (10) 100 (76) 3 (1) 18 (5)
Mel 290 100 (15) 100 (98) 5 (1) 92 (5)
Skin
WM 115 100 (8) 100 (59) 65 (2) 100 (17)
WM 793 100 (17) 100 (115) 89 (10) 100 (66)
1Primary eye and skin melanoma cells were stained with anti-class I
or class II MAbs and analyzed by flow cytometry. Melanoma cells also
were treated with IFN-␥ (1,000 U/ml for 48 hr). As a negative control,
melanoma cells were stained with the appropriate isotype control
antibody. Data are displayed as percent positive cells and mean channel
fluorescence.
473OCULAR MELANOMAS AND T CELLS

5. FIGURE 2 – Proliferation of lymphocytes stimulated with IFN-␥-treated melanoma cells. Tumor cells were treated with IFN-␥ (1,000 U/ml) for
48 hr prior to assay. Lymphocytes were stimulated with IFN-␥-treated tumor cells from either primary skin melanomas (a) or primary ocular
melanomas (b). Positive and negative controls included lymphocytes cultured with irradiated allogeneic or autologous lymphocytes, respectively.
*Significant proliferation at p Ͻ 0.001 when compared to the negative control.
474 VERBIK ET AL.

6. FIGURE 4 – Fixed primary ocular melanoma cells fail to inhibit lymphocyte proliferation. Primary ocular melanoma cells were fixed in 2%
formalin and then added to an allogeneic MLR at increasing concentrations. As a negative control, fixed human foreskin fibroblasts were added to
MLR cultures at the same concentrations as tumor cells. No significant inhibition of proliferation at p Ͻ 0.001 was observed when compared with
proliferation in the absence of regulator cells.
FIGURE 3 – Primary ocular melanoma cells inhibit lymphocyte proliferation. Increasing numbers of irradiated primary ocular melanoma cells
were added to a standard MLR and cultured for 6 days.As a negative control, irradiated (4,000 R) normal human foreskin fibroblasts were added to
an MLR culture at the same concentration as tumor cells. *Significant inhibition of proliferation at p Ͻ 0.001 when compared with proliferation
when no regulator cells were added.
475OCULAR MELANOMAS AND T CELLS

7. series of experiments conducted in our laboratory, we observed that
primary ocular melanoma cells and exogenous IL-2 were unable to
stimulate proliferation of autologous CD8ϩ T cells (data not
shown). To determine whether these results were due to a basic
difference between the immunogenicity of eye and skin melanoma
cells, we performed the present experiments, in which we directly
compared the ability of eye and skin melanoma cells to stimulate
allogeneic T cells.
Within the skin, melanomas progress through a series of discrete
histological stages: nevous, dysplastic nevous, radial growth,
vertical growth and metastases. Early experiments that examined
the ability of skin melanoma cells to stimulate T cells in the
absence of exogenous lymphokines revealed that, in general, the
immunogenicity of melanoma cells decreases during tumor progres-
sion so that primary melanoma cells are more immunogenic than
metastatic tumor cells. Fossati et al. (1984) observed that primary
skin melanoma cells stimulated proliferation of allogeneic T cells.
By contrast, when metastatic skin melanoma cells were used in a
similar series of experiments, tumor cells lost the ability to
stimulate proliferation of allogeneic T cells and actively inhibited
lymphocyte proliferation (Taramelli et al., 1988). Similar results
were observed when primary and metastatic melanoma cells were
used to stimulate autologous T cells (Parmiani et al., 1990;
Taramelli et al., 1988; Guerry et al., 1984, 1987). Thus, although
metastatic tumor cells express a variety of tumor antigens that are
capable of stimulating autologous CD8ϩ T cells in the presence of
exogenous lymphokines, earlier data indicate that, in the absence of
exogenous lymphokines, only primary skin melanoma cells stimu-
late allogeneic T cells. This coincides with histological studies in
situ demonstrating that early primary lesions are infiltrated by
activated T cells but that this infiltrate is reduced in advanced or
metastatic tumors (Kornstein et al., 1983).
Our results indicate that there is a fundamental difference in the
immunogenicity of ocular and skin melanoma cells. We confirm
that primary skin melanoma cells stimulate proliferation of alloge-
neic T cells. By contrast, primary ocular melanoma cells were
completely unable to stimulate similar T-cell proliferation. The
failure to stimulate T cells was not due to a decreased expression of
HLA on ocular melanoma cells since there was no difference in
expression of class I on eye and skin melanoma cells. Moreover,
when the expression of class I and class II was increased on ocular
melanoma cells by treatment with IFN-␥, tumor cells were still
unable to induce T-cell proliferation. In addition, we evaluated
primary ocular melanoma cells for expression of ICAM-1 and
LFA-3 co-stimulatory molecules and found that Mel 202 and 290
expressed significant levels of both (data not shown). Therefore,
the complete absence of these co-stimulatory signals on ocular
melanoma cells cannot account for their failure to stimulate
lymphocyte proliferation. The failure of ocular melanoma cells to
FIGURE 5 – Metastatic ocular melanoma cells fail to inhibit lymphocyte proliferation. Increasing numbers of irradiated metastatic ocular
melanoma cells from patient 270 (OMM2.2, 2.3, 2.5 and 2.6) were added to a standard MLR and cultured for 6 days. The ability of metastatic
tumor cells to inhibit T-cell proliferation is compared with that of primary tumor cells from the same patient (Mel 270). As a negative control,
irradiated (4,000 R) normal human foreskin fibroblasts were added to an MLR culture at the same concentrations as tumor cells. *Significant
inhibition of proliferation at p Ͻ 0.001 when compared with proliferation in the absence of regulator cells.
TABLE II – HLA-CLASS I EXPRESSION ON METASTATIC OCULAR MELANOMAS
Ocular
melanomas
Mean channel fluorescence1
Class I Class II
No IFN-␥ IFN-␥ No IFN-␥ IFN-␥
Primary
Mel 270 4 9 0 0
Metastatic
OMM2.2 37 52 0 0
OMM2.3 35 48 0 0
OMM2.5 41 50 0 0
OMM2.6 37 49 0 0
1Primary and metastatic ocular melanoma cells were obtained from
patient 270, and class I or class II expression was analyzed by flow
cytometry. Melanoma cells also were treated with IFN-␥ (1,000 U/ml
for 48 hr). As a negative control, melanoma cells were stained with the
appropriate isotype control antibody. All tumor cells were positive for
class I; therefore, data are displayed as mean channel fluorescence.
476 VERBIK ET AL.

8. stimulate T cells appears to result from tumor cell–mediated
inhibition. A small number of ocular melanoma cells were capable
of inhibiting proliferation of T cells in a standard MLR. These
results indicate that primary melanomas that develop within the eye
are less immunogenic than melanomas that develop within the s.c.
tissues of the skin. Moreover, primary eye melanomas actively
inhibit T-cell proliferation.
We were surprised that metastatic ocular melanoma cells lost the
ability to inhibit T-cell proliferation and were able to stimulate
lymphocyte proliferation. This was demonstrated using a matched
pair of cell lines isolated from patient 270, in which the primary
ocular melanoma cells were recovered, as well as several meta-
static melanoma cells from different hepatic lesions that developed
18 years after the initial diagnosis of the primary tumor. These data
suggest that ocular and skin melanoma cells are different in another
respect: while progression of skin melanoma is associated with a
decrease in immunogenicity, ocular melanoma cells appear to
become more immunogenic as the disease progresses.
We are currently examining the mechanism by which primary
ocular melanoma cells inhibit T-cell proliferation. One possible
mechanism is that ocular melanoma cells secrete TGF-␤, an
immunosuppressive factor present within normal ocular fluids
(aqueous and vitreous). Metastatic skin melanoma cells also secrete
TGF-␤ (Rodeck et al., 1994; Inge et al., 1992). However, we do not
predict that ocular melanoma cells prevent stimulation of T cells
via secretion of TGF-␤ since supernatants of cultured tumor cells
did not inhibit T cells. We favor a mechanism that requires
cell-to-cell contact. It is possible that primary ocular melanoma
cells express Fas-ligand on the surface and trigger apoptosis in
Fas-positive responding T cells. Griffith et al. (1995) observed that
Fas ligand was expressed constitutively within the uveal tract of
normal mouse eyes, suggesting that Fas ligand may be expressed
on normal choroidal melanocytes. If Fas ligand is expressed
constitutively on ocular melanocytes, it is likely that it also is
expressed during the earliest stages of malignant transformation.
Hahne et al. (1996) observed that Fas ligand was expressed on
metastatic skin melanoma cells but not on normal cutaneous
melanocytes. It is currently unclear when Fas ligand is up-regulated
during transformation of skin melanoma cells. However, it seems
reasonable to suggest that ocular melanomas may express Fas
ligand earlier during tumor progression and that this may account
for the inhibitory property of these primary tumor cells. It is
attractive to speculate that the loss of Fas ligand on metastatic
ocular melanoma cells may account for the increased immunogenic-
ity of these tumor cells. Experiments are currently under way to
determine the expression of Fas ligand during the development of
ocular melanomas.
REFERENCES
ALEXANDER, M.A., BENNICELLI, J. and GUERRY, D., Defective antigen
presentation by human melanoma cell lines cultured from advanced, but not
biologically early, disease. J. Immunol., 142, 4070–4078 (1989).
BOON, T., CEROTTINI, J.C., VAN DEN EYNDE, B., VAN DER BRUGGEN, P. and
VAN PEL, A., Tumor antigens recognized by T lymphocytes. Ann. Rev.
Immunol., 12, 337–365 (1994).
FOSSATI, G., TARAMELLI, D., BALSARI, A., BOGDANOVICH, G., ANDREAOLA,
S. and PARMIANI, G., Primary but not metastatic human melanomas
expressing DR antigen stimulate autologous lymphocytes. Int. J. Cancer,
33, 591–597 (1984).
GRIFFITH, T.S., BRUNNER, T., FLETCHER, S.M., GREEN, D.R. and FERGUSON,
T.A., Fas ligand induced apoptosis as a mechanism of immune privilege.
Science, 270, 1189–1192 (1995).
GUERRY, D., ALEXANDER, M.A., ELDER, D.E. and HERLYN, M.F., Inter-
feron-␥ regulates the T cell response to precursor nevi and biologically
early melanoma. J. Immunol., 139, 305–312 (1987).
GUERRY, D., ALEXANDER, M.A. and HERLYN, M.F., HLA-DR histocompat-
ibility leukocyte antigens permit cultured human melanoma cells from early
but not advanced disease to stimulate autologous lymphocytes. J. clin.
Invest., 73, 267–271 (1984).
FIGURE 6 – Proliferation of lymphocytes stimulated with IFN-␥-treated metastatic ocular melanoma cells. Tumor cells were either untreated or
treated with IFN-␥ (1,000 U/ml) for 48 hr prior to assay. Lymphocytes (1 ϫ 105) were cultured for 6 days with increasing concentrations of
irradiated tumor cells. Positive and negative controls included lymphocytes cultured with irradiated allogeneic or autologous lymphocytes,
respectively. *Significant proliferation p Ͻ 0.001 when compared to the negative control.
477OCULAR MELANOMAS AND T CELLS

9. HAHNE, M., RIMOLDI, D., SCHROTER, M., ROMERO, P., SCHREIER, M.,
FRENCH, L.E., SCHNEIDER, P., BORNAND, T., FONTANA, A., LIE´NARD, D.,
CEROTTINI, J.C. and TSCHOPP, J., Melanoma cell expression of Fas(Apo-
1/CD 95) ligand: implications for tumor immune escape. Science, 274,
1363–1366 (1996).
HERLYN, M., BALADAN, G., BENNICELLI, J., GUERRY, D., HALABAN, R.,
HERLYN, D., ELDER, D.E., MAUL, G.G., STEPLEWSKI, Z. and NOWELL, P.C.,
Primary melanoma cells of the vertical growth phase: similarities to
metastatic cells. J. nat. Cancer Inst., 74, 283–289 (1985).
HUCK, S.W., CORMIER, W.H. and BOUNDS, W.G., Reading statistics and
research, p. 68, Harper and Row, New York (1974).
INGE, T.H., HOOVER, S.K., SUSSKIND, B.M., BARRETT, S.K. and BEAR, H.D.,
Inhibition of tumor-specific cytotoxic T-lymphocyte response by transform-
ing growth factor beta 1. Cancer Res., 52, 1386–1392 (1992).
KORNSTEIN, M.J., BROOKS, J.S.J. and ELDER, D.E., Immunoperoxidase
localization of lymphocyte subsets in the host response to melanoma and
nevi. Cancer Res., 43, 2749–2753 (1983).
KSANDER, B.R., RUBSAMEN, P., OLSEN, K., COUSINS, S. and STREILEIN, J.W.,
Studies of tumor-infiltrating lymphocytes from a choroidal melanoma.
Invest. Ophthalmol. Vis. Sci., 32, 3198–3207 (1991).
PARMIANI, G., ANICHINI, A. and FOSSATI, A., Cellular immune response
against autologous human malignant melanoma: are in vitro studies
providing a framework for a more effective immunotherapy? J. nat. Cancer
Inst., 82, 361–370 (1990).
RAJPAL, S., MOORE, R. and KARAKOUSIS, C.P., Survival in metastatic ocular
melanoma. Cancer, 52, 334–336 (1983).
RODECK, U., BOSSLER, A., GRAEVEN, U., FOX, F.E., NOWELL, P.C., KNABBE,
C. and KARI, C., Transforming growth factor B production and responsive-
ness in normal human melanocytes and melanoma cells. Cancer Res., 54,
575–581 (1994).
SHIELDS, J.A., Introduction to melanocytic tumors of the uvea. In:
Diagnosis and management of intraocular tumors, pp. 65–82, C.V. Mosby,
St. Louis (1983).
TARAMELLI, D., FOSSATI, G., BALSARI, A., MAROLDA, R. and PARMIANI, G.,
The inhibition of lymphocyte stimulation by autologous human metastatic
melanoma cells correlates with the expression of HLA-DR antigens on the
tumor cells. Int. J. Cancer, 34, 797–806 (1984).
TARAMELLI, D., MAZZOCCHI, A., CLEMENTE, C., FOSSATI, G. and PARMIANI,
G., Lack of suppressive activity of human primary melanoma cells on the
activation of autologous lymphocytes. Cancer Immunol. Immunother., 26,
61–66 (1988).
TRAVERSARI, C., VAN DER BRUGGEN, P., LUESCHER, I.F., LURQUIN, C.,
CHOMEZ, P., VAN PEL, A., DE PLAEN, E., AMAR-COSTESEC, A. and BOON, T.,
A nonapeptide encoded by human gene MAGE-1 is recognized on HLA-A1
by cytolytic T lymphocytes directed against tumor antigen MZ2-E. J. exp.
Med., 176, 1453–1457 (1992).
VAN PEL, A., VAN DER BRUGGEN, P., COULIE, P.G., BRICHARD, V.G., LE´THE,
B., VAN DEN EYNDE, B., UYTTENHOVE, C., RENAULD, J. and BOON, T., Genes
coding for tumor antigens recognized by cytolytic T lymphocytes. Immu-
nol. Rev., 145, 229–250 (1995).
478 VERBIK ET AL.