0022-1767/89/14211-4070$02.

00/0

THEJOURNAl. OF iHMUNOI.OGY
Copyright 8 1989 by The American Associationof Immunologists

Voi. 142.4070-4078. No. 11. June 1. 1989

Printed in U . S . A .

DEFECTIVEANTIGENPRESENTATION
BYHUMANMELANOMA
CELL LINES
CULTURED FROMADVANCED, BUT NOTBIOLOGICALLY EARLY,DISEASE'
MICHAEL A. ALEXANDER,2* JEANNETTE BENNICELLI,'

AND

DUPONTGUERRY

IV*+

From the "Graduate Group in Immunology and the Pigmented Lesion Group,
Study the Universityof Pennsylvania Schoolof
Medicine, and the 'Hematology-Oncology Section, Department
of Medicine, Hospitalof the Universityof Pennsylvania,
Philadelphia. P A 19104

The present studies were undertaken to characterize Ag presentation by cultured human melanoma cell lines. Celllines established from "biologically early" lesions of malignant melanoma were
able to present the soluble Ag tetanus toxoid ( T T ) to
autologous and HLA-DR-matched allogeneic, TT-immune T cell clones.Proliferation of T cell clones in
response to Ag presented by primarymelanoma
peaked on day 2 of culture with Ag. Ag presentation
was blocked by pretreatment of TT-pulsed and fixed
melanoma cells with mAb against HLA-DR, but not
HLA-DQ, HLA-DP, or HLA-ABC. Ag processing and
presentation were inhibited by treating the melanoma cells with ammonium chloride. In parallel
with previous findings from
this laboratory demonstrating theinability of cell lines cultured from "advanced" primary or metastaticmelanoma to induce
autologous T cell proliferation, such cell lines also
failed to present this exogenous Ag despite the presence of cell-surface HLA-class I1 molecules. Thus,
in contrast to thefinding in biologically early melanoma, none of the multipleTT-immune, T cell
clones from autologous patients or HLA-DR
matched donors was able to respond to TT presented by melanoma cells cultured from advanced
disease. Co-incubation studies revealed that metastatic melanoma cells did not secrete inhibitory
substances during the APC assay, however, they
were able to process TT, rendering it "immunogenic"
in the presence of fixed, autologous non-T cells.
When fixed, autologous melanoma cells were assayed fortheir ability to present processed Ag; fixed
cells of early but not advanced disease were able to
present Ag in this setting, indicating that the presenting limb becomes flawed in
the evolution ofthe
metastatic phenotype.Finally, studies of chloroquine inhibition of the capacity of melanoma cells
derivedfrom early primary disease to stimulate
autologous peripheral blood T cells suggest that
such cells process and present tumor-associated Ag
in the same fashion as the "modelnAg TT.
~~

Ag-specific responses by T cells are largely dependent

on the participation ofAPC that express MHC class I or
class I1 molecules identical to those of the responding T
cell ( 1 ) . Helper T cells recognize Ag and MHC class I1
molecules present on the surfaceof such cells (2).Recent
studies indicate that Ag is physically associated with
class I1 molecules on the APC membrane (3-5). Before
presentation, nominalprotein Ag are internalized by the
APC and processed into small peptide fragments which
are then associated with class I1 molecules during their
transit through the endocytic pathway (6).Recognition of
the combined product (Ag-class I1 molecule) by specific
receptors on the Th cell is the basis of many cellular
immuneresponses.The
cells of the monocyte-macrophage series arewidely recognized as themodel APC (7).
Other cells capable of this function are activated (8),
viral
transformed (9), and malignant (10) B cells; endothelial
cells (1 1): astrocytes ( 1 2): and a variety of other normal
and malignant cell types (13). In some instances, the
capacity to present Ag has been correlated with the ability
of the cell to stimulate Tcells in a primary MLR [ 14. 15).
Melanoma evolves in stepwise fashion. Clinicopathologically defined lesions of tumor progression include
early primary disease(the radial growth phase), late primary disease(the vertical growth phase), and metastasis
(16). Previous studies from this laboratory have demonstrated that cell lines cultured from lesions early in the
pathway of melanocytic tumor progression stimulate
autologous T cell proliferation in vitro,whereas cell lines
derived from late melanocytic lesions do not ( 1 7, 18).This
is paralleled by in situ studies demonstratingthat certain
precursors and early lesions of primary melanoma are
characterized by a n activated T cell infiltrate, which is
attenuated or absent in advancedprimary or metastatic
lesions (18, 19).Furthermore, it has been shown that the
in vitro autologous T cell response to cell lines established
from early primarymelanomas is dependent on the
expression of HLA-class I1 Ag by the melanoma cell ( 1 7).
Here we report that cell lines established from early
primary melanoma are able to present the soluble Ag TT3
to both autologous and HLA-class 11-matchedallogeneic,
Ag-specific T cell clones. In contrast, cell lines derived
from advanced diseaseare unable to present Ag to such
T cell clones. I n addition, cell lines from both early and
late melanoma are able to process Ag. These observations
suggest that early primary melanoma cells mediate Agspecific T ceII proliferation via their ability to both process Ag and associate it with HLA-class I1 molecules. Fur-

Received for publication June 15. 1988.

Accepted for publication March 13, 1989.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' This work was presented in part at the American Association for
CancerResearch meeting. Atlanta, GA.May
1987 {Proc. Am. Assoc.
Cancer Res. 28: 1401)and was supported
by Grants CA 29200, CA 25874,
and CA 25298 from the National Institutes of Health. Bethesda, MD.
ZAddress correspondence and reprint requests
to Michael A. Alex3Abbreviations used in this paper: MAA. melanoma-associated Ag:
ander, Cancer Center, 7 Silverstein Bldg.. Hospital of the University of
MC. mitomycin C : S.I.. stimulation index: T T , tetanus toxoid.
Pennsylvania, 3400 Spruce St.. Philadelphia.
PA 19104.

4070

ANTIGEN PRESENTATION BY HUMANMELANOMA

ther tumor progression (metastasis) may produce cells

that fail toproductively associate processed Ag and HLAclass I1 molecules leading to their inabilityto immunologically interact withautologous, Ag-specific T cells. These
results furthersuggest that T cells may respond to a s yet
unidentified MAA specifically associated withHLA-class
I1 molecules on the surface of primary melanoma cells.
Therefore the role of Ag presentation in the HLA-class I1
dependent, autologous T cell response to early primary
melanoma cells was also investigated. This proliferative
response was inhibited
by pretreatment of the tumor cells
with a concentration of chloroquine that did not affect
HLA-DR expression. This finding suggests that autologous T cells recognize melanoma-associated tumor Ag
that are processed and linked to HLA-class I1 molecules
on melanoma cells in a manner similar to the handling
of exogenous Ag by classical APC.
MATERIALSANDMETHODS

Murine mAb. mAb L243 (Becton Dickinson Monoclonal Center.

Mountain View. CA) recognizes a monomorphic determinant on
HLA-DR Ag (20). N297 (clone HK19) (Mallinckodt, St. Louis, MO)
recognizes a monomorphic determinant onHLA-DQ Ag (21) andB7/
21 (Becton Dickinson) binds to a n epitope present on HLA-DP molecules expressing DP 1-5 specificities (22). W6/32(Pel-Freeze, Rogers, AR) binds to a monomorphic determinant on MHC class I Ag.
mAb OKT3 (anti-pan T cells, CD3). OKT4 (anti-inducer/helper T
cells, CD4). OKT8 (anti-suppressor/cytotoxic Tcells,
CDS), and
OKDR (anti-HLA-DR)were obtainedfrom OrthoDiagnostics Systems
(Raritan. NJ). mAb Leu-M3 is specific for humanmonocytes (Becton
Dickinson). B73.1 recognizes NK cells (23). and was a gift from G.
Trinchieri (Wistar Institute, Philadelphia, PA). MEDA3 binds a melanoma-associated oncofetal Ag (24) and wasa gift from M. Herlyn
(Wistar Institute). SFR16-DR7M. which recognizes a polymorphic
determinant on HLA-DR7 molecules (25). was a gift from S. Radka
(Genetic Systems Corp., Seattle. WA).
Celllines. Melanoma cell lines were established by described
methods (26)
from early primarymelanoma (lines
W M 3 5 and WM98).
late primarymelanoma (line WM278), and metastatic melanoma
(line WM9).In addition, two melanoma cell lines from the same
individual. oneestablishedfromearly
primarymelanoma[line
WM75)and the other
from a metachronous metastasis(line WM373)
were also used in this study. Othercell lines used were EBV-transformed B lymphoblastoid cellsderived from patient peripheral blood.
The cell lines were monitored by routine MHC typing, karyotyping,
and mycoplasm testing. Cell lines were maintained in T-150 flasks
(Corning Glass Works, Corning, NY) in a medium consisting of 50%
Hams F10 and 50%RPMl 1640 (GIBCO Laboratories. Grand Island.
NY) supplemented with 10% heat-inactivated FCS (Armour Pharmaceutical Co., Kankakee, IL).
Quantitation of cell surface Ag. Expression ofMCH class I and
class 11 Ag and melanoma-associated Ag was measured by a RIA
previously described in detail(17). Here. 1251-labeledF(ab),fragments of sheep anti-mouse Ig antibodies (New England Nuclear,
Boston, MA) were used to detect the appropriate murine mAb. The
phenotype of T cell clones was determined by flow cytometry by
using the Spectrum I11 cytofluorograph (Ortho Diagnostics). T cells
(3X lo5)were stained with theindicated murine mAb, washed, and
developed withfluorescein-conjugatedF(abIilgoat
anti-mouse Ig
(Tago. lnc., Burlingame, CA).
Treatment ofcell lines with lysosomotrophic agents.
MC (Sigma
Chemical Co.. St. Louis, MO) treated melanoma cells in suspension
were treated with indicated concentrations of ammonium chloride
[J.T. Baker Chemical Co., Phillipsburg, N J ) or chloroquine (Sigma)
a t a concentration of 2 x lo6 cells/ml in RPMI 1640 + 10% FCS for
1h at 37C. Treated cellswere washed four times in the same
medium before use.
In vitro immunization. cloning,and maintenance of TT immune
T cells. Washed PBL from healthy normal donors or melanoma
patients were resuspended in RPMI 1640 with 10% heated AB sera
a t a concentration of 3 X 106/ml. One milliliter was added/well of a
24-wellplate(Corning). To these wells,1 ml of either the same
Culture medium or medium containing 40 pg/ml TT (Connaught
Laboratories, Willowdale, Ontario, Canada) was added. The plate
was incubated at 37C for 6 days. At the end of this incubation T
cells were purified as described below and 20 U/ml of rIL-2 (Amgen
Biologicals, Thousand Oaks,CA) was added for 4 to 6 days. Thecells

407 1

CELL LINES

were either restimulated with fresh MC-treated autologous non-T

cells (T cell-depleted PBL) and 20 pg,hl TT for 48 h before the
addition of 20 U/ml rIL-2 or cloned directly by limiting dilution with
the use of lo4 autologous non-T cells, 20 pglml TT. and 20 U/ml
rIL-2 in 96-well round bottom plates (Corning). Wells of the cloning
plates were supplemented withrlL-2 containing medium after 1 wk
and clones were moved to larger wells after 2 wk. T cell clones were
used for experiments
restimulated as above every 10 to 14 days and
14 to 21 days after their last restimulation.
Before their use in
experiments each T cell clone was phenotyped by flow cytometry
and tested for reactivity to purified protein derivative (Connaught
Laboratories] in the assaydescribed below.
TT AgpresentationassayandmAb-blockingexperiments.
Washed, TT-specific, cloned T cells were resuspended to 0.5 X IO6/
ml in RPMI 1640 with 10% heated AB serum. MC-treated non-T or
melanoma cells useda s APC were resuspended in the samemedium
a t a concentration of 1 X 106/ml. One-tenth milliliter of T cells (5 X
lo4),0.05 ml of APC (5 X lo4).and 0.05 ml of TT (final concentration
= 20 pg/ml) were added to round bottom microtest plate wells. The
plates were incubated for 48 h at 37C in 5%COz. including a 24-h
pulse with 1 uCi 13H]thymidine(New England Nuclear) per well before
harvesting. Filter disks were added to scintillation fluid (New England Nuclear) and counted in a beta counter(Beckman Instruments.
lrvine. CA). Results were expressed as counts per minute (cpm) or
S.I. where S.1. = cpm of T cells + APC TT divided by the cpm of T
cells + TT. The cpm of T cells + TT alone or T cells + APC alone
were generally not more than 1.5 times the cpm of T cells alone.
mAb blocking experiments were performed on pulsed and fixed APC
which were prepared by pulsing MC-treated cells (2 X 106/ml) with
50 pg/ml TT in RPMI 1640 FCS for 1 h at 37C. After three washes
the cells were fixed for 5 min a t room temperature with 1% paraformaldehyde (J. T. Baker Co.) in PBS[GIBCO). followed by two
washes in 0.2 M glycine (Sigma) and two washes in medium. mAb
and APC were incubated 30 min before the addition of T cells and
the assay wasperformed as described above.
Autologous T cell proliferation inducedby melanoma cells.The
assay for the proliferative response of T cells co-cultured with melanocytic cell lines has beendescribedindetailelsewhere
(17).
Briefly, tumor cells (1 X 106/ml) were incubated with 100 pdml of
MC a t 37C in RPMI 1640 with 10% heat-inactivated FCS for 1 h.
After four washes thecells were resuspended to 2 X 106/ml in RPMl
1640 with 10% heat-inactivatedAB serum. Peripheral blood T cells
were obtained from the mononuclear cell population of heparinized
whole blood after density-gradient centrifugation, plastic adherence,
and rosetting with neuraminidase (Calbiochem Biochemicals, San
Diego.CA) treated SRBC. Such T cells were twice washed and
resuspended to 2 X 106/ml. One-tenth milliliter of stimulating
tumor cells and responding T cellswere then added to quadruplicate wells of a 96-well, flat bottomed plate (Corning). Tumor cells
and T cells were also incubated separately in medium as controls.
After 6 days at37C in 5%CO,. 0.6 WCiof [3H]TdRwas added to each
well 8 h before harvesting thecells ontoglass fiber filter paper with
a n automatic cell harvester (Brandel, Gaithersburg. MD). [3H]TdR
incorporation was recorded a s cpm in a scintillation counter (Searle
Radiographics Inc.. Des Plains. IL). Data were expressed a s a S.I.. in
which S.1. = mean cpm of test wells divided by cpm of T cells alone.
The mean cpm of tumor cells alonewas always<loo. and themean
cpm of T cells alonewas generally ~ 2 0 and
0 never >300. We take a
S.I. of >2.5 to be significant.

RESULTS

Phenotype of TT-specific T cell clones and kineticsof

the response toTT presented b y autologous, early primary melanoma cells. T cell clones specific for TT were
generated from HLA-typed normals and melanoma patients. Six clones were phenotyped a s activated helper T
cells by using murine mAb and flow cytometry. Percent
positive cells (mean f 1 SD) for each antibody were a s
follows: OKT3, 98.6%f 0.5; OKT4, 98.4% f 0.5; OKT8,
0.8% 0.3; OKDR, 95.8% f 4.0: Leu-M3, 1.4% t 0.1;
B73.1, 1.3% & 0.4. TT-specific T cell clones were also
unable to respond to purified protein derivative in the
APC assay (datanot shown). The kinetics
of the response
to TT presented by autologous primary melanoma cells
was determined by pulsing with [HITdRand harvesting
the cultures every 24 h over a 5-day period. A s shown in
Figure 1, T cell proliferation peaked at 48 h and only

4072

ANTIGEN PRESENTATION BY HUMANMELANOMA

CELL LINES

contrast, no clones from the metastatic melanoma patient responded to TT presented by the autologous melanoma cell despite significant responses to TT presented
by autologous non-T cells. No responses could be seen in
any of the control cultures containing T cells incubated
with TT alone or APC alone. Similar resultswere obtained
by using Tcell clones and tumor cells from two additional
primary (lines WM35 and WM75) and metastatict (lines
WM373 and WM278) melanoma patients.
We next sought to determine whether
TT-specific T cell
clones derived from a patient with metastaticmelanoma
were defective in theirability to interact withmelanoma
1
2
3
5
cells in general and whether allogeneic, primary and
metastatic melanoma cells (lines WM75 and WM373)
Days incubated
from another individual would present Ag to a compatible
Figure 1. Kinetics of TT Ag presentation by autologous primary mel(HLA-DR matched) T cell clone. Three T cell clones speanoma cells. TT-immune Tcell cloneJ.F.TT1 was incubated wlth medium
alone (0)or 20 pg/ml TT and MC-treated autologous non-T cells m), cific for TT, two from a metastatic melanoma patient
melanoma cells 0).
or allogeneic non-T cells(A). Duplicate cultures were
(CJTTl andCJTT7; HLA-DR6/7) and one from a normal
pulsed and harvested on indlcated days and data are reported a s mean
cpm f 1 SD. Mean cpm for T cells Incubated with autologous non-T and donor (RSTT14; HLA-DR7).were assayed for their ability
melanoma cells wlthout Ag over the 5-day culture period were 132 f 31
to respond to TT presented by non-T cells autologous to
and 163 2 5 1 , respectively.
the T cell clone as well as primary and metastatic melanoma cell lines isolated from the same individual (HLATABLE I
DR4/7)
(Table 11). Both T cell clones derived from the
T cell clonal analysls of 2 7 Ag presentation by autologous melanoma
patient with metastatic melanoma responded briskly to
cells"
TT presented by autologous non-T cells. Additionally all
Tl-Immune T Cell Clones + 20 @/ml TT
and
three clones responded to Ag presented by melanoma
Patlent/Tumor Type
'lone
cells of the allogeneic primary (WM75), but not to its
No'
Autologous
Autologous
Allogenelc
Non-T
MeLb
Non-TC
subsequent metastasis (WM373).
J.F./primary (HLA1
166.0d
251 .O
4.5
MHC-restriction of the response to TT and inhibitton
4
DR-/3)
51.0
154.7
0.8
by pretreatment of primary melanoma cells with rnAb
462.5
7.5
5
27 1.3
against HLA-D region proteins. To further explore the
210.0
1 .o
17
63.0
5.4 A 2306.8 157.6
MHC-restriction of TT-specific T cell clones from melanoma patients, Tcell clones from a patient with primary
1 .o
C.J./metastatic
1.5 1
48.0
melanoma and one with metastatic disease
were assayed
2.0
6
21.5
(HLA-DR 1.5
6/7)
7
1.2
I .o
15.5
for their ability to respond to Ag presented by autologous
1.6
1.3
1la
54.2
melanoma and non-T cells, allogeneic non-T cells, and
17
137.0
1.2
1.5
allogeneic,
HLA-DR matched non-T cells (Table 111). As
a Five T cell clones each from HLA class I1 mismatched patients. J.F.
(source of primary Ilne) and C.J. (source of metastatic line), were Incu- expected, proliferative responses were obtained from
bated wfthTTalone or TT wlth an equal number of MC-treatedautologous
both T cell clones by using autologous non-T cells and
non-T cells,melanoma cells. or allogeneic non-T cells andthe APC assay
the appropriate allogeneic, HLA-DR matched non-T cells
was done a s described.
as APC. However, only primary melanoma cells (WM98)
WM98 Is the primary-derived line of patlent J.F. and WM9 is the
metastasis-derived line of patient C.J.
autologous to the TT-specific clone JFTT7 were able to
Non-T cells from the HLA-class 11 mlsmatched melanoma patlent.
serve as APC.
Stimulation Index = cpm of T cells + APC + TT/cpm of T cells + T T .
To examine the role of HLA-class I1 molecules in Ag
autologous non-T cells and primary melanomacells from presentation by autologous primary melanoma cells,
this patient were able to mediate a TT-specific T cell blocking studies withmAb were done. As shown in Figure
response. As expected, incubation of the T cells withTT
TABLE I1
alone orTT with allogeneic APC was ineffective.
Allogeneic TT-immune T cell recognitton ofprimary and metastatic
Clonal analysis of TT presentation by primary and
melanoma cell linesfrom the sameIndividual"
metastatic melanoma cells. Table I reports a comparison
Addltlon
to
Medium
WM75b
WM373b Autologous
Non-T
Culture
of two sets of five autologous T cell clones specific for
Experlment 1'
T T , each from HLA-DR mismatched patients: J. F. (HLA306f33
1 1 61212391+01312 4
Medlum
DR3 and source of primarymelanoma line) and C.J.
20 g&nl TT 114 2 12 2.807 f 190 216 f 21 23,759 + 1.546
(HLA-DR6/7 and source of metastatic melanoma line). T
Experlment 2
551250
1 8340k545f17052 2 5
Medium
cells were incubated withTT and autologous non-T cells,
20 pg/rnl TT 381 2 43 3,271 f 178 326 f 74 33,984 * 2.675
autologous melanoma cells, or the reciprocal allogeneic
Experiment 3
398 k 2 1
2 7 9 2298 3 f
14 2 6 9 + 17
non-T cell and assayed for T cell proliferation against
Medlum
20 ,g/ml TT 289 k 19 2.987 2 102 302 k 21 25,875 k 1,986
TT. Each of the T cell clones shownfrom the patientwith
T cell clones were incubated with melanoma cells or autologous nonthe cultured primary melanoma line responded to TT in
T cells with or without 20 pg/ml of TT in the APC assay as described I n
the presence of autologous non-T cells and melanoma Materials and Methods. Data are reported as mean cpm f 1 SD.
Autologous primary (WM75)and metastatic (WM373)melanoma cell
cells. Two additional clones (not shown)
responded to TT
HLA-DR 4.7.
presented by non-T cells, but not autologous melanoma Ilnes.
CTT-immuneT cell clones from Experiments 1 and 3 are HLA-DR 6.7
cells. A s expected, allogeneic. HLA class I1 mismatched (metastatic melanoma patient clones C.J. TTl and C.J.TT71 and ExperAPC were incompetent mediators of this response. In iment 2 are HLA-DR 7 (normal Individual clone RSTT14).

ANTIGEN
PRESENTATION

4073

BY HUMAN
MELANOMA
LINES
CELL

TABLE 111
M H C restriction of TT-immune T cell clones responding to autologous
Drirnnru a n d metastatic melanoma Celts"
r

cell lines autologous toa primary melanoma(WM98)

and
a metastatic melanoma (WM278) were used as controls
for polymorphic HLA-class I1 and MAA. mAb against
Clone C.J. TT-6'
Clone J.F. T T - 7 b
APC
HLA-DR
HLA-DR, -DQ, and -DP (L243, N297, and B7/21, respeclC"rn1
lCDml
tively) bound significantly to all cells regardless of the
97 f 12
210 f 24
1. ?T alone
type of cell assayed. Both primary and metastatic mela3111,607 f 1.010
281 C 31
2. J.F.n0n-T'
6/7
107 2 19
14.444 i 889
3. C.J. non-T'
noma cells expressa similar rangeof both monomorphic
3125,031 f 987
164
f 19
4. WM98 (J.F.)b
and polymorphic HLA-class I1 determinants. An antibody
176
f
46
617
107 f 5
5.WM9 (C.J.)'
598 f 81
(SFR16-DR7M) that is specific for a polymorphic deter317,554 f 187
6. Allogeneic non-T
7. Allogeneic non-T
61112 f8.927
30
f 334
minant onHLA-DR7 molecules (25) was detected on both
8. Allogeneic non-T
71129 C 23
769 C 22
primary and metastatic melanoma cells that expressed
"TT-specific T cell clones (J.F. TT-7 and C.J. TT-6) were incubated
the HLA-DR7 haplotype. Also, the antibody MEDA3,
with TT alone, or TT and various autologous or allogeneic non-T and
which detects a 105 Kda oncofetal protein expressed on
melanoma cells in the APC assay as described inMaterials and Methods
and datareported as mean cpm f 1 SD.
most melanoma cells (24). bound to all melanoma cells,
Primary melanoma patientT cell clone, autologous non-Tand melabut not to EBV-transformed B cells. In addition, melanoma cell line.
'MetastaticmelanomapatientT
cell clone.autologousnon-Tand
noma cells failed to bind isotype-matched control mAb
melanoma cell line.
(data not shown).
Lack of inhibition of TT-specific Tcell clone prolgeration by autologous metastatic melanoma cells. To de30,000
termine if autologousmelanomacells
derivedfrom a
metastasis (WM9) secrete inhibitory substances during
L"."""
the APC assay, equal numbersof MC-treated tumor cells
t
\
were co-cultured with TT plus autologous, TT-specific,
cloned T cells and APC (autologous non-T cells) in the
standard APC assay (Table V). TT-specific cloned T cells
were incubated with autologous, MC-treated non-T cells
or metastatic melanoma cells.A s shown, T cells neither
responded to APC without the addition of Ag (Table V,
line 1) nor to TT presented
by metastatic melanoma cells
0.1
1
10
[data column2, line 2). When additional MC-treated nonAntibody conc. [ugirnl]
T cells were added to wells already containing non-T cells
Flgure 2. Inhibition of autologous melanoma cell
Ag presentation with a n increase in the response to TT was apparent [data
anti-HLA-DR mAb. Indicated concentrations of mAb against HLA-DR
column 1, line 3 ) .When non-T cellswere added to wells
(L243)(0).HLA-ABC (W6/32) (5.HLA-DP (B7/21](0).and HLA-DQ (N2971
containing T cells, TT, and melanoma cells (data column
(0)were preincubated in triplicate microtest plate wells with T'I-pulsed.
paraformaldehyde-fixed primary melanoma cells (WM98)for 1 h before
2, line 3).a response to TT equivalent wells
to
containing
the addition of cloned. TT-immune T cells(J.F.TT7) and the APC assay
only
T
cells,
non-T
cells,
and
TT
was
found
(data column
was performed as described. Data represent mean cpm f 1 SD.
1, line 2) demonstrating the lack
of inhibitory substances
2, the indicated concentrationsof mAb directed a t HLA- secreted by melanoma cells. Whenthe reciprocal addition
DR (L243). HLA-ABC (W6/32), HLA-DP (B7/21), and
HLA- was done (MC-treated melanoma cells were added to culas APC) again no inhibition
DQ ( N 2 9 7 ) were preincubated with TT-pulsed. paraform- tures containing non-T cells
was seen (data column 1, line
4).
aldehyde fixed, primary melanoma cells for
1 h before
A
g
processing
by
melanoma
cells derived from methe additionof cloned, autologous TT-specific T cells.
The
standard APC assaywasthenperformed.The
mAb tastasis. To explore the Ag-processing capacity of metfixed,autologous,non-Tcells
against HLA-DR (L243) completelyinhibited the response astaticmelanomacells,
were
added
to
wells
containing
cloned T cells, TT, and
at 10 pg/ml. whereas mAb against HLA class I (W6/32)
and the two other HLA class I1 molecules (HLA-DP and various MC-treated autologous or allogeneic cells as Ag
HLA-DQ)were ineffective. Inhibition
at the APC level was processors. In this setting, allogeneic cells may process
evident since blocking was only achieved if mAb were Ag, but not present it, since the response to TTis MHCpreincubated with the APC 1 h before the addition of restricted. A s shown in TableVI, T cells were incubated
responding T cells. Simultaneous addition of antibodies with either medium,TT, or TT fixed, autologous, nonT cells. To these wells either medium,autologous non-T
and T cells abrogated inhibition.
[HLA-DR6/7),or various
Quantitative expression of class I1 and MAA by pri- cells, metastatic melanoma cells
a s Ag processors. The remary and metastatic melanoma cells used a s APC. We allogeneiccellswereadded
next asked if the lack of HLA-class 11-dependent APC sponse to APC by the TT-specific T cells without added
Ag was minimal [data column1).However, when TT was
function by metastatic melanoma cells was associated
cells were able to support a
with a quantitative difference in expressionof cell-sur- added only autologous non-T
face HLA-class I1 Ag. In Table IV three primary melanoma strong response toAg (data column2, line 2). Consistent
autologous metastatic
cell lines capable of presenting TT and stimulatingpe- withpreviousexperiments,the
melanoma cell was unable to presentAg (data column2,
ripheral blood autologous T cell proliferation (17) and
three cell lines derived from advanced disease and unable
line 4).As expected. bothallogeneic cells(non-T cells and
to either presentAg or stimulate autologous T cells (1 7) primary melanoma cells: data column 2, lines 3 and 5,
were examined by RIA for expression of monomorphic respectively) induced levels of proliferation in response
and polymorphic determinants of HLA-class I1 molecules, to TT far lower than that mediated by the autologous
as well as a MAA. Two EBV-transformed B lymphoblast APC. However, both the autologous metastatic melanoma
"~
Z~

I-r---,

4074

ANTIGENPRESENTATIONBYHUMANMELANOMACELLLINES
TABLE IV
MHC class I 1 expression by primary and metastatic melanoma cell lines measured
by RIA"
Cell Line
HLA-DR type

WM35*
2/7
WM9gb
3iWM75b
4/7

against
P-3

HLA-DR

207 f 8

2,732 f 189
2.474

316 f 32
24,100

f 154
24.001

12.499
347 f 70

f 885

WM373'
4/7
WM278'
6i7
WM9'
6/7

17.959
546 f 13

f 2,673
3.028

98EBV-B
3/278EBV-B
6/7

16
168 f 17,938

f 1.788
2.338

195 _C 2916.636

f 785

I 4 6 f 21
2.21
166f 1

mAb
HLA-Dg

24.117f
188

HLA-DR7

MAA

f 813

1.220 f 140
24.237

f 214
5.992

f 22
552

f 43

f 641
2,307

f 65
4,832

f 128

f 132
11.861

f 387
2.755

f 175
6.576

f 702

f 172
3.259

f 446
2.823

f 121
22.863

f 1.022

f 1664,393

f 11024,484

f 93

f 183
398

f 395
27

f 41

983 f 12
6,177

1 f 440
1,365

HLA-DP

f 4,435
87

2.738 f 70119.972
f 163
8,559

6 2 65f ,2923 4 2 2
4 .3101 3 f 3
12
59
7

f 332

5.382 f 239

f 39

a Melanoma and EBV-B cells were incubated with

mAb against HLA-DR (L243).HLA-DQ (N297).HLA-DP (B7/27). HLADR7 (SFRI6-DR7M). melanoma-associated Ag (MEDA3). and control mouse myeloma culture supernatant P-3 (P3 x
63Ag8). TheRIA was performed a s described in Materials and Methods and data reported as cpm
meanf 1 SD.
*Primary melanoma.
Metastatic melanoma

TABLE V
Lack of inhibition by metastatic melanoma cellson the APC.function
of autologous non-T cells"

of the response was not seen.

Ag presentation by fixed primary and metastatic
melanoma cells. To further assign the defect in
APC
Autologous TT-specific clone (C.J.TT7) and
function
of
metastatic
melanoma
cells
to
presentation
Additions to Culture
MC melanoma
rather thanprocessing, both primary and metastatic
melMC non-T
lWM91
anoma cells were fixed with paraformaldehyde and as1. Medium alone
306
f 33
91 2 3
sayed for Ag presentation to autologous, Ag-specific T
2. TT alone
25.759 f 1.546
271 f 57
3. MC non-T + TT
33.680 f 1,453
cell clonesafterTT-processing
by HLA-class I1 mis28.401 f 1,406
f 360
4. MC melanoma + TT 29.830
289 f 5 9
matched cells. A s shown in Table VII, equal numbers of
fixed melanoma cells and HLA-DR mismatched, alloge+ medium
= 116 f 29
Experimentalcontrols:C.J.TT7
C.J. TT7 + TT
= 1 1 4 2 12
neic, MC-treated cells were incubated with autologous,
a Cloned TT-immune T cells (C.J.TT7)from
TT-specific, T cell clones in the standard APC assay. A s
metastaticmelanoma
patient C.J. were added to wells containing a n equal amount ofMCalso seen in the previous experiment,
MC-treated allogetreated autologous non-T cells or autologous metastatic melanoma cells
neic
cells
or
fixed
melanoma
cells
alone
were unable to
(WM9). Addedto these wells were medium alone. TT alone, or TT and an
equal number of MC-treated autologous non-T or melanoma cells. The
present Ag. However, fixed, autologous. primary melaAPC assay was performed a s described in Malerials a n d Methods and
noma cells, but not allogeneic, (mismatched) melanoma
data is expressed as mean cpm f 1 SD.
cells, were able to present
Ag processed by allogeneic
TABLE VI
cells in a n MHC-restricted fashion (data columns2 and
Antigen processing by autologous metastatic melanoma
cells"
3, lines 3 and 4). In contrast, fixed autologous metastatic
T I - l m m u n e T Cell Clone lnrubated with
melanoma cells were unable to present Ag under these
Processing cells
conditions (data column5, lines 3 and 4).
20 &/ml TT
Fixed non-T + T"
Medium
Inhibition of TT Ag presentation by ammonium chlo410 f 97
I . None
1 1 01
f 21 01 f 5 5
f 55 24.371 f 1.02127,707 f 2,043
2. Autologous non-Tb466
ride. To further explore the requirement forAg process7,643 f 1.034
253 f 28
1.057 f 141
3. Allogeneic non-T'
ing by primary melanoma cells used
as APC, the effect of
6,340 f 335
I66 f 13
365 f 55
4. Melanoma (WM9)
ammonium chloride, which inhibits processing
by in4, I67f 238
5. Allogeneic melanoma
205 f 3 9
530 f 52
IWM981
creasing intracellularpH (27),on APC function was stud"Cloned, TT-immune T cells from metastatic melanoma patient C.J.
ied. In the experiment shown in Figure 3 , MC-treated,
were added to wells containing either medium, 20 pglml of TT alone. or
autologous
primary melanoma cells (A) and non-T cells
20 gg/ml of TT and fixed autologous non-T cells. Added to these wells
(B) were treated with the indicated concentrations
of
were medium ( I ] . or a n equal number of MC-treated autologous non-T
cells (2).allogeneic non-T cells(3).autologous metastatic(4).or allogeneic
ammoniumchloridefor
1 h a t 37C. To determine
primary melanoma cells
(5)a s APC. The assay was
performed a s described
whether these cells
could recover fromthe treatment and
in Malerials and Methods and data represent mean
cpm f 1 SD.
'HLA-DR 6.7.
present Ag added during the assay, the cells were treated
HLA-DR 2.9.
with or withouta simultaneous pulseof TT. After washHLA-DR 3.
ing, the cells were added tothe wells with cloned T cells,
to
(data column 3, line 4) and the allogeneic primary mela- and TTor TT-pulsed cells were added with T cells the
Ag. The response to TT prenoma (data column 3, line 5) allowed a response to TT wellswithoutadditional
presented by fixed APC. A s shown, all of the allogeneic, sented by autologous primary melanoma and non-T cells
was inhibited at
every concentration of ammonium chloMC-treated cells were able to process Ag despite their
inability to presentit. If such cells were fixed before theirride used. However, unlike non-T cells, melanoma cells
appeared not tobe able torecover from ammonium chloaddition, or if fixed, allogeneic. non-T cells were used
ride treatment to utilize excess Ag added tothe wells after
instead of fixed, autologous, non-T cells, augmentation

ANTIGENPRESENTATION

4075

BY HUMANMELANOMACELLLINES

TABLE VI1
Ag presentation by paraformaldehydefixed melanoma cells"
TT T Cell Clone J . F . TTAZ
Ag

Processing Cells

T Alone

None
None
Allogeneic
non-T
Allogeneic melanoma
Autologous
non-T

+
+

985 f 85
907 f 22
953 f 6 8

688 f 57
11.751

TTT

Cell Clone C.J.TT1 l a

Presenting Cells

Presenting Cells
Autologous

Allogeneicb

(WM9RJ

(WM75)

626 f755
104
8.951 f
861
760

+ 584
740

f 66
f 77
f 29
160

Alone

Autnlogous

Aiiogrneirb

PM91

(WM98)

125 t 1 3
133 t 131
2
100 f 4
3 1 13f 0
91
0+40

t 28

172 + 18

t3
288i:6
l 5 6 ? 26

f 5 18

27,993
5,632
f 628

Two TT-immune T cell clones, J.F. TTAZ (primary melanoma patient J.F.) and C . J . TT1 l a (metastatic melanoma
fixed autologous or fixed allogeneic melanoma cells a s presenting
patient C.J.). were incubated with medium (T alone) or
(+) or without (-) TT. MC-treated allogeneic non-T cells (allogeneic noncells. To these cultures were added medium with
T). and allogeneic melanoma cells (allogeneic melanoma] as APC. Autologous non-T rells with Ag serve a s a positive
control. The standardAPC assay was performed and the data represent mean
cpm f 1 SD.
* Fixed allogeneic primary melanoma cells

R
looool*

40000

30000

20000

8ooo

10

20

30

40

50

TABLE VIlI
Autologous T cell response to chloroquine-treated primary melanoma
cells"
Pretreatment n i Primary
Melanoma Cells (WM35)
Response

PBS

40 pM/ml chloroquine

Autologous T Cell
(S.I.]

11.2 f 1.6
3.8 f 0.4
12,325

H1.A Class I1 RIA

(wml

1 1,450 f 1.050
f 1.217

a Washed, MC-treated melanoma cells were incubated with 40 pM/ml

of chloroquine or a n equal volume of PBS in culture medium for 1 h at
37C. After incubation, washed melanomacells were ro-cultured with a n
and assayed a s described. A n identical
equal numberof autologous T cells
aliquot of melanoma cells was assayedby RIA for quantitative expression
of HLA-DR with mAb L243 a s described. Tcell response data represents
mean S . I . k 1 S D and RIA data are reporteda s mean cpm f 1 S D for three
separate experiments. Autologous T cell response mean IT3HITdRinrorporation control (T cells + medium)was239 f 46. In the RIA mean
background P-3 binding forPBS and chloroquine-treated melanoma cells
was 182 ? 24 and 184 15, respectively.

DISCUSSION

Previous reports from this laboratory have shown that

cell lines cultured from melanoma precursors biologand
ically early lesions of primary melanoma stimulate proliferation of autologous T cells in vitro. In contrast, cell
2000
lines culturedfrom advanced disease are unable to stimulate T cells (17,
18). An in situ parallelof these findings
0
is that lesionsof early primary melanoma are
typified by
10
20
30
40
50
0
a lymphocytic infiltrate that is absent or attenuated in
Ammomum chlorlde conc.
late melanoma lesions (16).This lymphocyte infiltrate is
[ mM/ml]
comprised primarily of activated (IL-2R and HLA-class I1
Figure 3. Ammonium chloride inhibition of TT presentation by autolpositive) T cells approximately evenlydistributed among
ogous primary melanoma and non-T cells. Indicated concentrations of
(19).The
ammonium chloride were added to melanoma cells( A ) or non-T cells (8) the helper and cytotoxic/suppressor phenotypes
with or without TT for 1 h a t 37C. TT-pulsed cells were added to wells
autologous T cell response to primary melanoma in situ,
containing Tl-specific.cloned T cells (closed symbols). Those incubated as well a s in vitro, has been shown toinvolve the secrewithout TT were added to wells containing TT and TT-specific T cells
tion of IFN-7 by the responding T cell, causing subse(open symbols),Data are in mean cpm t 1 SD.
quentaugmentation
of HLA-class 11 synthesisand
expression by the melanoma cell (18).We (1 7) and others
the treatment.
(28) have demonstrated that the
vitro
in autologous Tcell
The Tcell response to autologous primary melanoma response to cultured primary melanomais dependent on
is dependent on Ag processing. The role of Ag process- the expression of HLA-class I1 molecules by the stimulating in the proliferative response
of autologous peripheral ing tumorcell. These phenomena are consistent with the
blood T cellstoprimarymelanomawasinvestigated.
hypothesis that melanoma cells
of early disease can presTable VI11 records the results of three experiments in
ent melanoma-associated Ag in the contextof HLA-class
which primary melanoma cells were treated with either I1 molecules to host helper T cells that are subsequently
PBS orthe Ag-processing inhibitor,chloroquine,and
activated.
tested for their capacity to stimulate peripheral blood T
To characterize melanoma cells as APC and address
cells. A s shown, chloroquinesignificantlyinhibited
one of several mechanisms that might underlie the failautologous T cell stimulation by primary melanoma cells ure of advanced melanoma to provoke an immune rewithout affectingHLA-class I1 expression by these cells.
sponse, the present study asks whether
cells cultured

4076

ANTIGEN
PRESENTATION
HUMAN
MELANOMA
BY
LINES
CELL

from advanced lesions differ from cells of early lesions

in their capacity to process and present Ag. A s neither
yet
the MAA involved in immunity to melanoma have
been identified, nor HLA restricted, tumorcell-specific T
cell clones generated, we have used the Ag TT and tetanus-specific Tcell clones tomodelthe function of melanoma cells as APC.
Analysis of the response to TT by individual T cell
clones from three patients with
biologically early primary
melanoma revealed significant proliferation to TT presented by celllines establishedfrom their primary tumor.
A s mentioned, two clones were established from one of
the primary melanoma cases which responded to TT
presented by non-T cells but not by the primary melanoma cells. The mechanismof this failure remains unexplained. In contrast to these findings, cell lines established from three advanced melanomas failed to present
TT. That T cell clones established from patients with
advanced melanomawere nevertheless capableof proliferation was evident in theirresponse to TT presented by
autologous non-T cells.
T cell clones from patients with both early and advanced disease were also shown to MHC-restricted
be
in
their response to TT (Table 111). Here cloned, TT-specific
T cells from the metastatic melanoma patient, although
unable to respond to TT in the presence of autologous
melanoma cells, were able to respondto Ag presented by
autologous, a s well as allogeneic, HLA-DR matched nonT cells. The MHC-restriction of TT-immune Tcell clones
is well documented. In fact, this response has been used
to study the nature
of restriction (29)and Ag presentation
(9).Several groups haveused blocking by mAbto identify
HLA-class I1 restriction determinants seen by helper T
cells specific for various Ag (29, 30)and alloantigen (31).
The HLA-class I1 dependence of the response of TTspecific T cell clones presented by autologous primary
melanoma cells was confirmed by experiments in which
pretreatment of the melanoma cell with mAb against
HLA-DR completely blocked the response (Fig. 2).
The datareported in TableI1 suggest that the transition
from early to advanced diseasein the sameindividual is
associated with the loss of APC function. In addition, T
cell clones isolated from a metastatic melanoma patient,
while unable to respond to Ag presented by autologous
melanoma cells, were as competent as normal T cell
clones in their ability to respondto the sameAg presented
by allogeneic, HLA-DR-matched, primarymelanoma
cells. Heretwo autologous melanoma cell lines established from a primary and a subsequent metastasiswere
asked to presentAg to allogeneic, HLA-DR-matched, TTspecific T cell clones. One clone was derived from a
normal donor and the other two from a patient with
metastaticmelanoma.Theseclonesshared
HLA-DR
specificities with both melanoma cell lines. Both clones
were able to respond to Ag presented by the HLA-DRmatched primary melanoma cell line but not by the cell
line established from a subsequent metastasis. The patient from whom these cell lines were established died
before the generation of autologous TT-specific T cell
clones limiting us to the study of allogeneic (HLA-DR
matched) TT-specific T cell responses.
Despite serologic identity at the HLA-DR locus (HLADR7) between the T cell clones and the melanoma cells
in Table 11. the primary melanoma cells appear to func-

tion poorly a s APC when compared with autologous nonT cells. One possible explanation for these resultsis that
the HLA-DR7 haplotype contains at least four T celldefined HLA-Dw specificities [32). Differences in expression of these specificities by the melanoma cellmay
contribute to the lower responses to Ag presented by
them. In addition, HLA-Dw specificities have been shown
to control T cell responses to Ag presented by allogeneic,
HLA-DR matched APC (30).Consequently, the data in
Table I11 may also reflect a disparity in HLA-Dw specificities where non-T cells, matched
for HLA-DR bystandard
HLA typing, are not a s efficient at presenting Ag when
compared with autologous non-T cells. The fact that
primary melanomacells present Ag a s efficiently a s nonT cells in a n autologous system (see TableI) supports the
thesis offered above. Nevertheless, the experiments depicted in Table I1 do not formally rule out, on the one
hand, defective recognition of Ag presented on melanoma
cells by T cells from patients with metastatic disease or,
on the other, that line WM75 is an inefficient APC for
reasons unrelated toits stepin tumor progression.
A number of studies were done to approach the question of why cells from advanced disease were incompetent a s APC. First, this is not a n artifact of experimental
clonal selection a s bulk cultures of TT-immune T cells
from several patients with advanced disease repeatedly
failed to respond to Ag in the presence of autologous
melanoma cells.
Second, cells from metastatic melanoma did not have
a quantitative abnormality
of HLA-class I1 expression. As
Ag presentation is dependent on the expression of MHCclass I1 molecules by the APC (1). it was necessary to
examine the quantitative expression of HLA-class 11 Ag
by melanoma cells used as APC (Table IV). All melanoma
cells, whether established from primary or metastatic
lesions, bound significantamounts of the threeantibodies to monomorphic determinants on HLA-DR. -DQ, and
-DP. Also,where appropriate,the cells bound an antibody
against a polymorphic determinant on HLA-DR7 molecules (25).
Third, we were unable to find evidence of soluble inhibitors secreted by metastatic melanoma cells during
the APC assay. It has been shown that melanoma cells
are capable of secreting substances that can interfere
with Tcell proliferation in humans(33)and in mice (34).
We addressed this problem directly by incubating live,
MC-treated metastatic melanomacells with cloned, autologous, TT-specific T cells, autologous non-T cells [APC),
and TT in the standard APC assay (Table V). In this
situation, T cells respond normally to TT presented by
autologous non-T cells despite the presence of metastatic
melanoma cells. The same resultwas obtained when the
melanoma cells were fixed with paraformaldehyde. When
autologous non-T cells were fixed and added to cultures
of T cells, metastatic melanoma cells, and TT, a n increased response was detected when compared to the
response of cultures with T cells, melanoma cells, and
TT (data not shown). These
cellsTwere unable to respond
to native TT in the presence of fixed autologous non-T
cells. Fixation with paraformaldehyde under these conditions has been shown to reduce cellular metabolism
(35),a function needed for Ag processing to occur (36,
37). These findings indicate
that melanoma cells derived
from metastasis were able to process Ag, but were unable

ANTIGEN PRESENTATION BY HUMAN

MELANOMA

CELL LINES

4077

noma cells with IFN--, (a potent inducer of HLA-class I1

to present it.
To further test this hypothesis, we took advantage of molecules) and this lymphokine is secreted by T cells in
the MHC-restriction of the response by adding fixed autol- response to autologous primary melanoma (18).To demogous non-T cells to cultures containing cloned TT-spe- onstrate thedependence of the autologous T cell response
cific T cells, T T , and autologous metastatic melanoma to endogenously processed cell-surface-associated Ag on
cells or various allogeneic cells [Table VI). These T cells primary melanoma, we used a n inhibitor of processing,
were unable to respond to TT alone, APC alone, TT + chloroquine. Chloroquine treatment has been shown to
allogeneic APC. TT autologous metastatic melanoma inhibit Tcell responses to Ag-pulsed cells. However, class
cells, and TT fixed autologous non-T cells. However, 11-dependent autoreactive and alloreactive T cell rethese T cells responded briskly to TT presented by met- sponses to the samecells are unaffected (41). Here chloabolically active, autologous non-T cells and to TT in roquine treatment of primary melanoma cells substancultures containing metastatic melanoma
cells or alloge- tially reduced their ability to stimulate autologous T cell
neic cells when fixed autologous non-T cells were added. proliferation. HLA-class I1 expression, a s measured by
These experiments demonstrate that metastatic mela- RIA, was found tobe unaffected by such treatment.
Our studies havenot directly addressed the mechanism
noma can generate
immunogenic Ag fragments.
Immunogenic peptide fragments of OVA, generated in of failure to productively associate processed Ag and
vitro, have been shown to stimulate T cell hybridomas HLA-class I1 molecules by melanoma cells of advanced
when presented by fixed tumor cells (38).Therefore we disease. Immunogenic peptides have been shown to asthen asked whether
fixed autologous melanoma cells sociate or bind to class I1 molecules in cellular (42) and
from both early and advanced disease were able to pre- in vitro (4, 5) systems. Processing either in vitro or intrasent Ag processed by third-party allogeneic cells. The cellularly is necessary to generate appropriate peptide
experiment outlined in Table VI1 shows that only fixed fragments that facilitate both MHC-class I1 and TCR bindautologous primary melanoma cells were able to present ing (38).Therefore, if metastatic melanoma cells were
proper immunogenic fragments in
Ag processed by HLA-DR mismatched cells. Fixed allo- unable to generate the
the processing experiments, then one would expect no
geneic melanomaor autologous metastaticmelanoma
cells were completely ineffective at presenting already stimulation via fixed autologous non-T cells. However, if
processed Ag. These experimentsprovide insight intothe metastatic melanoma cells were unable to associateimnature of the failure of metastatic melanoma cells to munogenic Ag with HLA-class I1 molecules, then procfunction a s APC by separating the functionsof Ag proc- essed Ag would appear, but would not be presented to
essing andAg presentation.
autologous immune T cells. The formation of immunoTo directly assess thedependence on cellular process- genic restriction determinants [part Ag and part MHC
ing for Ag presentation by autologous primary melanoma product) is a n essential function performed by APC beand non-T cells, they were treated with various concen- cause it is a prerequisite forAg-specific T cell recognition
trations of ammonium chloride and added to the APC (43).Restriction determinants for many T cell-recognized
assay (Fig. 3 A,B). Both ammonium chloride and chloro- Ag have been studied in detail (44).Antigenic restriction
quine havebeen shown to inhibit
Ag processing, presum- sites [Ag-binding sites) are thought toreside in the first
ably by increasing intracellular pH and preventing en- domain [mostpolymorphic region) of the HLA-DR /%chain
dosomal-lysosomal fusion (39).In this experiment, TT- in humans (27) and the same
region in the murineI-A pspecific T cells were asked to respond to
TT presented by chain (45). Because melanoma cells express these Ag, it
autologous primary melanomacells that were TT exposed will be of interest to characterize the Ag-binding sites by
either during the treatmentwith ammonium chloride or using purified HLA-class I1 molecules derived from both
afterward during the
assay. Ag was added post-treatment primary and metastaticmelanoma.
to determine if melanoma cells could recover and utilize
One of the critical events associated with metastasis
is
available Ag. In a simultaneous experiment (Fig. 3 B ) by evasion of thehostimmuneresponse.
Although the
using autologous non-T cells as APC, we found that am- mechanisms thatunderlie the suppression or evasion of
monium chloride-treated cells were able to largely recover host anti-tumor immunity are likely to be several, this
the APC function of untreated cells. This is consistent study suggests that failure to present relevant tumorwith previous findings with the use of peripheral blood associated Ag on the surface of tumor cells evolving
monocytes and a primaryAg presentation system(35).In toward the metastatic phenotype is one pathway to subour experiments primary melanoma cells are unable to verting the host.
recover from the treatment. Thismay indicate that melREFERENCES
anoma cells depend heavily on the endosomal-lysosomal
fusion process, or that they are unable to reverse the
1 . Schwartz, R. H. 1985. T-lymphocyterecognition of antigen in assoeffects of higher intracellularpH induced by these agents.
ciation with gene products of the major histocompatibilitycomplex.
Ann. Reu. Irnmunol. 3:237.
Processing of cell-derived macromolecules by malignant
2. Schwartz, R. H. 1985. Associations in T cell activation. Nature
melanoma cells has been previously recognized (40).
31 7:284.
3. Phillips, M. L., C. C. Yip, E. M. Shevach, and T.L. Delovitch. 1986.
However, this is the first time that Ag processing by a
Photoaffinity labeling demonstrates binding between Ia molecules
tumor cell has been related to a n immunologic function.
and nominal antigen on antigen-presenting cells. Proc. N a t l . Acad.
Previous studies from this laboratory have shownthat
Sci. USA 83:5634.
the in vitro autologous T cell response to cultured primary 4. Watts, T.H., and H. M. McConnell. 1986. High-affinity fluorescent
peptide binding to I-Ad in lipid membranes. Proc. N a t l . Acad. Sct.
melanoma cells is HLA-DR dependent (17). Responding
USA 83:9660.
5. Buus, S., A. Sette. S. M. Colon, D. M. Jenis, and H. M. Grey. 1986.
T cells likely seestimulating tumorAg in the context of
Isolation and characterizationof antigen-Ia complexes involved in T
HLA-class I1 molecules expressed by the melanoma cell.
cell recognition. Cell 47:1071,
This response is augmented by treatment of the mela6 . Germain, R. N. 1986. The ins and outs of processing and presenta-

4078

ANTIGEN
PRESENTATION

BY HUMAN
MELANOMA
LINES
CELL

tion. Nature322:687.
7. Shevach. E. M.. and A. S. Rosenthal. 1973. Functionof macrophages
26.
in antigen recognition by guinea pig T lymphocytes. 11. Role of the
macrophage in the regulation of genetic control of the immune response. J . Exp. Med. 138:1213.
8. Kakiuchi. T..R. W. Chestnut, and H. M. Grey. 1983. B cells a s
27.
antigen-presenting cells: the requirement for B cell activation. J .
Immunol. 131: 109.
9. Issekutz. T., E. Chu. and R. S . Geha. 1982. Antigen presentation by
28.
human B cells: T cell proliferation induced by Epstein Barr virus B
lymphoblastoid cells.J . Immunol. 129: 1446.
10. Glimcher, L. H., K. Kim, I. Green, and W. E. Paul. 1982. Ia antigenbearing B cell tumor lines can present protein antigen and alloanti-29.
gen in a major histocompatability complex-restricted fashion toantigen-reactive Tcells. J . Exp. Med. 155:445.
30.
11. Nunez, G., E. J. Ball, and P. Stasny. 1983. Accessory cell function
of human endothelial cells. I. A subpopulation of Ia positive cells is
required for antigen presentation.J . Immunol. 131:666.
31.
12. Fierz, W.. B. Endler, K. Reske, H. Wekerle. and A. Fontana. 1985.
Astrocytes as antigen-presenting cells. I. Induction of la antigen
expression on astrocytes by T cells via immune interferon and its
32.
effect on antigen presentation. J . Immunol. 134:3785.
13. Klein, J. 1986. Recognition of soluble antigens in the context of
class 11 molecules. In Natural Historyof t h e Major Histocompatibll33.
ity Complex. J. Klein, ed. J o h n Wiley & Sons, New York, pp. 474492.
14. Stingl. G., S . I. Katz, L. Clement, I. Green, and E. M. Shevach. 1978.
Immunologic functions of Ia-bearing epidermal Langerhanscells. J .
34.
Immunol. 121:2005.
15. Cowing, C., and J. M. Chapdelaine. 1983. T cells discriminate beon allogeneic accessory cellsand B cells:
tween la antigens expressed
35.
on Ia molecules.
a potential function for carbohydrate side chains
Proc. Natl. Acad. Sci. USA 80:6000.
16. Clark. W. H., D. E. Elder, D. Guerry, M. N. Epstein, M. H. Green,
of tumor progression:the precursor 36.
and M. VanHorn. 1984. A study
lesions of superficial spreading and nodular melanoma. Hum. Pathol. 15:1147.
37.
17. Guerry. D., M. A. Alexander. M. F. Herlyn, L. M. Zehngebot. K. F.

18.
19.

20.
21.
22.
23.

24.

25.

Mitchell, C. M. Zmijewski, and E. J. Lusk. 1984. HLA-DR histocompatibility leukocyte antigens permit cultured human melanoma
cells
from early but not advanced disease to stimulate autologous
lymphocytes. J . Clin. Invest. 73:267.
Guerry, D., M. A. Alexander. D. E. Elder, and M. F. Herlyn. 1987.
Interferon-?regulates the T cell response to precursor nevi and
biologically early melanoma.J . Immunol. 139:305.
Kornstein, M. J.. J. S . J. Brooks, and D. E. Elder. 1983. Immunoperoxidase localization of lymphocyte subsets in the host response
to melanoma and nevi. Cancer Res.43:2749.
Lampson, L.A., and R. Levy. 1980. Two populations of la-like
molecules on a human Bcell line. J . Imrnunol. 125:393.
Shipp. M. A.. B.D. Schwartz. C. C. Kannapell. R. C. Griffith. M. G .
Scott, P.Ahmed, J. M. Davie, and M. H. Nahm. 1983. A unique DR
related B cell differentiation antigen. J . Immunol. 131:2458.
Watson. A. J.. R. DeMars, I. S . Trowbridge, and F. H. Bach. 1983.
Detection of a novel human class11HLA antigen. Nature 304r358.
Perussia. B.. S. Star. S . Abraham, V. Fanning, and G . Trinchieri.
1983. Human natural killer cells analyzed by 873.1. a monoclonal
antibody blocking Fc receptor functions. I. Characterization of the
lymphocyte subset reactive with 873.1.J . Immunol. 130:2133.
Herlyn, M., J. Thurin, G . Balaban, J. Bennicelli, D. Herlyn, D. E.
Elder. E. Bondi, D. Guerry. P.Nowell, W. H. Clark. and H. Koprowski. 1985. Characteristicsof cultured human melanocytes isolatedfromdifferentstages
of tumorprogression.CancerRes.
45:5670.
Radka, S . F., D. B. Amos, L. J. Quackenbush, and P. Cresswell.
1984. HLA-DR7-specific monoclonal antibodies and a chimpanzee
on the same molecule.
anti-DR7 serum detect different epitopes

38.

39.
40.
41.

42.

43.
44.

45.

Immunogenetics 19:63.
Herlyn, M. F., D. Herlyn. D. E. Elder, E. Bondi. D. LaRossa. R.
Hamilton, H. F. Shears, G. Balaban, D. Guerry, W. H. Clark, and H.
Koprowski. 1983. Phenotypic characteristics of cells derived from
precursors of human melanoma. Cancer Res.
43:5502.
Wu. S.. T.L. Saunders. and F. H. Bach. 1986. Polymorphism of
human la antigens generatedby reciprocal intergenic exchange between two DR p loci. Nature 324r676.
Fossati, G., D. Taramelli. A. Dalsari. G . Bogdanovich, S . Andreola.
and G . Parmiani. 1984. Primary but not metastatic human melanoma expressingDR antigens stimulateautologous lymphocytes. Int.
J . Cancer 33:591.
Mellins. E.. M. Woelfel, and D. Pious. 1987. Importance ofHLA-DQ
and -DP restriction elements in T cell responses to soluble antigens:
mutational analysis. Hum. Immunol. 18:211.
Gehrz. R. C., S . Fuad. Y.C. Liu, and F. H. Bach. 1987. HLA class I1
restriction of T helper cell response to cytomegalovirus [CMV). I.
Immunogenic control of restriction. J. Immunol. 138:3145.
Weland. C. M.. J. Goronzy, and C. G . Fathman. 1986. Human T-cell
clones used to define functional epitopeson HLA class I1 molecules.
Proc. Natl. Acad. Sci. USA 83:762.
Karr, W. H. 1986. Immunochemical analysis
of the Ia polymorphisms
among the family
of DR7-associatedHLA-D specificities.J . Immunol.
136:999.
Taramelli, D.. G . Fossati, A. Balsari, R. Marolda, and G . Parmiani.
1984. The inhibition of lymphocyte stimulation by autologous human metastatic melanoma cells correlates with expression of HLADR antigens on the tumorcells. Int. J . Cancer 34:797.
Putnam, J. B., and J. A. Roth. 1985. Identification and characterization of a tumor-derived immunosuppressive glycoprotein from
murine melanoma. CancerImmunot. lmmunother. 19:90.
Scala, G., and J. J. Oppenheim. 1983. Antigen presentation by
human monocytes: evidence for stimulant processing and requirements for interleukin 1. J . Immunol. 131:1160.
McKean, D. J., A. Nilson, A. J. Infante, and L. Kazim. 1983.
Biochemical characterization of B lymphomacell antigen processing
J . Immunol. 131:2726.
and presentation to antigen-reactive cells.
T
Allen, P. M.,and E. R. Unanue. 1984. Differential requirements for
antigen processing by macrophages for lysozyme-specific T cell hybridomas. J . Imrnunol. 132: 1077.
Shimonkevitz, R.. J. Kappler. P. Marrack. and H. Grey. 1983.
Antigen recognition by H-2 restricted T cells. I. Cell-free antigen
processing. J . Exp. Med. 158:303.
Mellman. I., R. Fuchs, and A. Helenius. 1986. Acidification of the
endocytic and exocytic pathways. Ann. Rev. Btochem. 55:663.
Johnson, D., and J. C. Bystryn. 1984. Mechanism of autodegradation of cell-surface macromolecules shed by human melanomacells.
Exp. Cell Res. 152: 179.
Finnegan. A,, B. W. Needleman, and R. J. Hodes. 1985. Antigen
processing requirements for T cell activation: differential requirements for the presentation of soluble conventional antigen vs. cell
surface MHC determinants. J . Immunol. 134:2960.
Falo. L. D., K. Sullivan, B. Benacerraf, M. F. Mescher. and K. L.
Rock. 1985. Analysis of antigen presentation by metabolically inactiveaccessory cells and their isolated membranes. Proc. Natl.
Acad. Sci. U S A 8293647.
Heber-Katz, E.. D. Hansburg, and R. H. Schwartz. 1983. The Ia
molecule of the antigen-presenting
cell plays a critical role in immune
response gene regulation.J . Mol. Cell lmmunol. 1:314.
Ottenhoff. T. H. M.. S. Neuteboom, D. G . Elferink. and R. P. DeVries.
1986. Molecular localization and polymorphism of HLA-class I1 restriction determinants defined by Mycobacterium leprae-reactive
helper T cell clones from leprosy patients. J . Exp. Med. 164:1923.
Germain, R. N.. J. D. Ashwell, R. I. Lechler, D. H. Marguilies, K. M.
Nickerson, G . Suzuki, and J. Y. L. Tou. 1985. Exon-shuffling
maps control of antibody-and T-cell-recognition sites to the NH,terminal domain of the class I1 major histocompatibility polypeptide
Ab. Proc. Natl. Acad. Sci. USA 82:2940.