Você está na página 1de 8

0022-1767/87/1391-0305$02.

00/0

Vol. 139.305-312,No. 1. July 1. 1987


Prtnted Ln U.S.A.

THEJOURNAL
OF I"UNOUX;Y
Copyright 0 1987 by The American Assoclatlon of Immunologists

INTERFERON-7 REGULATES THE T CELL RESPONSE TO PRECURSORNEVI


AND BIOLOGICALLY EARLY MELANOMA'
DUPONT GUERRY, IV,2* MICHAELA. ALEXANDER,* DAVID E. ELDER,'
MEENHARD F. HERLYN'

AND

From the *Pigmented Lesion Study Group, the Hematology-Oncology Section, Department
of Medicine. and the 'Department
of Pathology and Laboratory Medicine, University
of Pennsylvania School of Medicine; and the 'Wistar Institute
of Anatomy
and Biology, Philadelphia,P A 191 04

To examine the potential regulatory role of inter- mas contained the highest numbers of activated
feron-y in the cellular immune response to mela- lymphocytes and stained positively for interferonnoma and its precursor lesions, we have tested the ?.
These results suggest that interferon-7 plays a
capacity of this lymphokine to enhance HLA class
I1 antigen-dependent T lymphocyte blastogenesis, central role inthe regulation of the cellular immune
its in vitro production by autologous T cells stimu- response to melanoma. It is produced by T cells,
lated by melanoma, and its presence in melanocyticlikely activated by tumor antigens seen in the conlesions in situ. Cell lines derived from a dysplastic text ofHLA class 11 antigens. In turn, interferon-?
nevus, a radial growth phase primary tumor,a ver- production enhances expression of HLA class 11antical growth phase primary, and metastatic lesions tigens by melanoma and precursor cells, and such
were induced by recombinant interferon-7 to ex- enhancement is associated with additional T cell
press increased amounts of HLA class I1 antigens. activation in a positive feed-back loop.
Such cells were then examined in radioimmunoassay for expression of HLA-DR antigens and in coAnalyses of frozen sections of melanocytic lesions by
culture for their ability to stimulate proliferation of
the
immunoperoxidasetechniquehavedemonstrated
autologous T cells. Interferon-?treatment of melanocytic cells increased their expression of HLA-DR that dysplastic nevi (potential precursors of melanoma)
antigens threefold to sixfold. In parallel with these and biologically early melanoma(radial growth phase
findings, co-culture of T cells with interferon- melanoma, a stage in tumorprogression associated with
treated cells of a dysplastic nevus and a radial phase invasion but without the capacity for local tumorigenic
melanoma ledto augmented T cell incorporation of growth or metastasis) (1) are invariably infiltrated by T
tritiated thymidine, and this stimulation was in- cells,approximately equally distributed between the
hibited with a monoclonal antibodyto HLA-DR an- helper and cytotoxic/suppressor phenotypes (2). In contigens. Despite augmented expressionof HLA class trast, lesions of biologically latemelanoma(vertical
I1 antigens (HLA-DR. -DQ, and -DP),vertical growth growth phase and metastases) are characterized by a
phase and metastatic melanoma cells failed to stim- paucity of infiltrating lymphocytes (2).Additionally, melulate autologous T cells. When T cells wereco- anocytes of melanoma and its precursors commonly excultured with stimulating melanoma cells, culture press class I1 major histocompatibility complex (MHC)
supernatants contained significantly
increased
antigens on their surface (3, 4). These histological findamounts of interferon-? (12 U/ml) in comparison ings are paralleled in functional assays. Melanoma cell
with supernatants of T cells alone (4 U/ml). No inter- lines derived from biologically earlylesionsstimulate
feron wasdetectable in cultures of melanoma cells
coalone. To link these in vitro phenomena to in situ autologous T cells to undergo blastogenesis when
Such stimulation is dependent, in
events, we used murine monoclonal antibodies to cultured in vitro (5,6).
interferon-7, the interleukin 2 receptor, and HLA- part, on the expression ofHLA-DR class I1 antigens by
DR antigens in an immunoperoxidase system to de- the transformed melanocyte, because it is inhibited by
tect interferon production and lymphocyte activa- allospecific antiserum to these HLA antigens (5).Further,
tion in frozen sections of lesions representative of there is a quantitative relationship between the expresmelanocytic tumor progression. In these studies, sion of HLA-DR antigens by tumor cells, as measured in
precursor dysplastic nevi and radial phase melano- radioimmunoassay, and thedegree of lymphocyte prolifReceived for publication July 25,
1986.
Accepted for publlcation April 15, 1987.
The costs of publication of this article were defrayed in part by the
payment of page charges. Thls article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' Thls work was presented in part at the American Association for
Cancer Research meeting. Houston, TX. May 1985 (Proc. Am. Assoc.
Cancer Res. 26:302].This work was supported by Grants CA 29200,CA
25874 and CA25298 from the National Institutes of Health.
'Address correspondence to DuPont Guerry. M.D.. Cancer Center, 7
Sllverstein Bldg., Hospital of the University ofPennsylvania. 3400Spruce
St.. Philadelphia, PA 19104.

eration, as measured by tritiated thymidine incorporation


(5).Cell lines from advanced lesions are not stimulatory,
whether or not they express HLA-DR antigens [5,6).
Interferon-y (IFN-y) is a lymphokine produced by T cells
in response to activation by antigen or mitogen (7). In
vitro, recombinant IFN-y (rIFN-y) has been shownto
increase the synthesis and
cell surface expressionof both
MHC class I and I1 antigens by melanocytes culturedfrom
normal skin, as well as by melanocytic cells from all
classes of melanocytic lesions (8-1 1). Generally, it does
not increase theexpression of non-MHC cell surface an-

305

306

INTERFERON-y AND MELANOMA

tigens associated with melanocytes or melanoma


cells
and recognized bya large panel of monoclonal antibodies
(Mab)(9, 10).
Here, we show that augmented expression of HLA-DR
class I1 antigens by neoplastic melanocytes from a dysplasticnevus and a radialgrowth phase melanoma
treated with rIFN-y is associated with a n increased capacity to stimulate autologous T cells, and that such
stimulation is inhibited by a Mab specific for HLA-DR
antigens. These findings
provide additional evidence that
such tumor-associated MHC antigens areinvolved in provoking T cell blastogenesis. In addition, we find that T
cells activated by co-culture with autologous melanoma
cells generate measurableIFN-y. Finally, to link these in
vitro studies to in situ events, we have used an irnmunoperoxidase technique with Mab to demonstrate in frozen sectionsof melanoma and its precursors that T cells
in situ areactivated (as measured by their expression of
HLA-DR antigens and the interleukin 2 (IL 21 receptor)
and produce IFN-y.
MATERIALS AND METHODS

Murfne Mab. Mab 691-13-17 recognizes a determinant on the


light chain of HLA-DR (12. 13). As tested in binding assays onHLA
class I1 loss mutants described by Kavathas et al. (14, 15).
it is
specific for HLA-DR antigens[personalcommunication,Susan
Radka. Ph.D.. Genetic Systems Corp.. Seattle, WA). L243 (Becton
Dickinson Monoclonal Center, Mountain View, CA) also recognizes
a monomorphic determinant on HLA-DR antigens (16). N297 [clone
HKl9) (Mallinckrodt, St. Louis, MO) recognizes a monomorphic determinant ofHLA-DQ antigens (17) and B7/21 (Ekcton Dickinson)
binds a n epitope present on HLA-DP molecules expressing DP 1-5
specificities (18). W6/32 (Pel Freeze. Rogers. AR) binds a monomorphic determinant ofMHC class 1 antigens. Mab OKT4 (antiinducer/helper T cells), OKT8 [anti-suppressor/cytotoxicT cells),
and OKTll (anti-T cell receptor for sheep erythrocytes) were obtained from Ortho Diagnostics Systems [Raritan, NJ). Both Mab
113.4 and 128.4 (Inter-Yeda Ltd.. Ness-Ziona, Israel) are IgGl antibodies that specifically bind IFN-y; unlike Mab 113.4, 128.4 is a
neutralizing antibody (1 ml of ascitic fluid neutralizes 200,000U of
IFN-y) (19).Anti-Tac recognizes the IL 2 receptor (20)
and wasa gift
of P. Nowell, University of Pennsylvania. Philadelphia. PA. P 3 is the
culture supernatantof the murinemyeloma cell line P 3 ~ 6 3 A g 8 . ~
Cell lines. Nevus and melanoma cell lines were established by
described methods (21) from a dysplastic nevus (line 814). from a
radial growth phase primary (line WM35). from the vertical growth
phase of an advanced primary lesion [line WM793). and from three
metastatic lesions [lines
WM222. WM310, and WM806). The fidelity
of cell lines was monitored by MHC typing and karyotype. Raji is a
B lymphoblast(Burkitt's lymphoma) line.Cell lines were maintained
in T-150 culture flasks (Corning Glass Works., Coming, NY) in a
medium consisting of 70%Ham's F10 and 30%Dulbecco's modified
Eagle's medium (GIBCO Laboratories. Grand Island. NY) supplemented with 10% heat-inactivated fetal calf serum [Armour Pharmaceutical Co.. Kankakee, IL).
Quantitatfon of cell surface antfgens.
Expression of MHC class 1
and I1 antigens was measured by a radioimmunoassay previously
described in detail (5).Here. '251-labeled F(ab)', fragments of sheep
anti-mouse immunoglobulin antibodies (New England Nuclear, Boston. MA) were used to detect the appropriate murineMab.
rIFN-y treatment of cell lfnes.Pre-confluent
cell lines were
treated with 500 U/ml of rIFN-y (Genentech. San Francisco,CA) in
30 mlof tissue culture medium for 48 hr, because this produces
near maximal augmentation ofHLA-DR expression (8-1 1). Cells
were then harvestedby brief trypsinization.
Lymphocyte proliferation. The assay for the
proliferativeresponse of T cells co-cultured with melanocytic (or other) cell lines
has been described (5).Briefly, rIFN-y-treated or untreated tumor
cells (1 X 106/ml) were incubated with 100 &/ml of mitomycin C
[Sigma Chemical Co., St. Louis, MO) a t 37C in RPMI 1640 (GIBCO)
with 10% heated fetal calf serum. After four washes, cells were
resuspended to 2 x 106/ml in RPMI 1640 with 10%heat-inactivated

human AB serum. Peripheralblood lymphocytes enriched in T cells


were obtained from the mononuclearcell population of heparinized
whole blood after density-gradient centrifugation, plastic adherence,
rosetting withneuraminidase-(CalbiochemBiochemicals, San Diego,
CA) treated sheep erythrocytes, and lysis of adherent erythrocytes.
At least 98%of these cells bound sheep erythrocytesupon re-rosetting. Such "T cells" were twice washed and resuspended to 2 X 106/
ml. One-tenth milliliter of "stimulating" tumor cells and of "responding" T cells werethen added to quadruplicatewells of a 96-well, flatbottomed plate (Coming). Tumor cells
and T cellswere also incubated
separately in medium as controls. After 6 days at 37C. 0.6 pCiof
13H]thymidine[New England Nuclear) was added to each well 8 hr
prior to harvesting the
cells onto glass fiber paper with
a n automatic
cell harvester (Brandel, Gaithersburg, MD). 13H]Thymidineincorporation was recorded as counts per minute (cpm) in a scintillation
counter (Searle Radiographics Inc., Des Plaines. IL). Data were expressed as a stimulation index (S.1.). in which S.I. = mean cpm in
test wells + mean cpmof T cells alone.The mean cpmof tumor cells
alone was always<loo. and the meancpm of T cells was generally
<200 and never>300. We take an S.I. of >2.5 to be significant.
Inhfbftion of lymphocyte stfrnulatfonby anti-HLA-DR Mab. The
mixed melanoma cell-lymphocyte assay wasdone as described above
with rIFN-y-treated stimulator cells except that 0.05 ml of medium,
anti-HLA-DR antibody (L243). or anti-HLA class I antibody (W6/32)
was added to suchcells priorto co-culture withT cells (final concentration of antibody of 1 pg/ml].
Quantftatfon of supernatant IFN-y. Supernatants of tumor cells
alone, T cells alone. and these cells in co-culture were assayed for
IFN-.I in the laboratory of G. Trinchieri (Wistar Institute, Philadelphia, PA), as described (22). In this assay, the antiviral activity of
supernatants was tested by inhibition of the cytopathic effect of
vesicular stomatitis virus on the human fibroblast strain, Detroit
532. TheIFN concentration inducing50%protection from the cytopathic effect corresponds to 1 U of the National Institutes of Health
IFN-y standard Gg-23-901-530. Samples showing IFN activity were
also retitered after treatment with a Mab specific for IFN-y, and in
all casesIFN activity was abolished.
Immunoperoxidase analysisof frozen sectfons. Fresh tissue was
obtained from melanocytic lesionsbiopsied during the management
of patients of the Pigmented Lesion Group, University of Pennsylvania. Portions not required for histologic diagnosis were dissected
out. snap frozen. and stored a t -70C until required. Cryostat sections (5 pm) were cut and dried overnight. Sections were then fixed
in acetone (10 min) anddried. After a 3-min wash in PJNaCl at pH
7.4, theprimary antibody was applied a t a nappropriate dilution for
15 min. Each antibody was titrated to allow for crisp staining of
reactive cells againstminimal interstitial background. Afteranother
P,/NaCI wash, the color reaction was developed as described (2) by
using biotinylated anti-mouse antibody, avidin:biotinylated peroxidase complex (Vector Laboratories, Burlingame. CA). and aminoethylcarbazole (AEC)(Aldrich Chemical Co.. Milwaukee. WI). The water
soluble, red AEC pigment was chosenbecause it contrasts well with
brown melanin pigment. Sections weremountedin
Elvanol (E.I.
Dupont Co.. Wilmlngton, DE). Negative control slides had the myeloma supernatant P3 substituted
for the primary antibody.Additionally, isotype-matched (IgGl and IgG2a) control antibodies to irrelevant antigens (theepidermal growth factor receptor and a 120,000
dalton melanocyte antigen) had noreactivity with lymphocytes in a
concurrent series of five dysplastic nevi, five primary melanomas,
and five metastases. As a positive control. primary Mab OKTl 1 and/
or OKT8 and OKT4 were used. In addition. the antibody 691-13-17
to HLA-DR antigens was positive on at least one cell type in every
lesion studied. Positive results consisted of red staining of cells. The
sections were examined by one author(D. E. E.) and were scored for
intensity of staining (1 to 4-t) and the proportion (0 to 100%) of
positive cells of various relevant types [lymphocytes, histiocytes,
Langerhans cells. stromal cells, structural, and lesional epithelial
cells). Among responding lymphocytes. two patterns were distinguished, namely "infiltrating" (lymphocytesamong lesional melanocytic cells) and "noninfiltrating" (lymphocytes at the base of the
lesion). Peripheral, presumably membrane-associated, andcytoplasmic (diffuse) staining patternswere also distinguished.
RESULTS

Augmentation by rlFN-y of HLA-DR antigen expressionbyprecursorandbiologicallyearlymelanoma


cells is associated with increased T cell response in
uitro. Two cell lines, one established from a precursor
SAbbreviations used in this paper: P3. the supernatant of mouse dysplastic nevus (line 814. Table I) and one established
myeloma P3X63Agi3; Tac. the IL 2 receptor: AEC. aminoethylcarbazole:
from a primary melanoma in radial growth phase (line
S.I.,stimulatlon index.

307

INTERFERON-y ANDMELANOMA
TABLE I
rlFN-y treatment of line 81 4 (dysplastic nevus): effecton MHC antigen expression and autologous lymphocyte
stimulation
Line 814
Treated with

Antigen Expression
(cpm RIA + 1 SD)

in

Class F

Class I I ~

Lymphocyte
Stimulation
(S.I./cpmf 1 SD)

P3

Medium alone
374f 20
6.542f 109
349
f3.41630
13
f 110
IFN-r
1,154f
7.933
126
f 156
416f 10
8.0/1501f 405
a Cell line 814 was incubated with medium alone or with rIFN-7 for 48 hr and then wasassayed by radioimmunoassay
and the anti-HLA class I antibody, W6/32.The
(RIA) for the binding of two Mab: the anti-DR antibody. 691-13-17,
background binding of the P3 control was not subtracted here or in Table 11. Aliquots of medium and IFN-y-treated cells
were also mitomycin C treated andcultured with autologous T cells. After 6 days, lymphocyte proliferation was assessed
by [3H]thymidine incorporationand then reported as the S.I. and raw cpm f 1 SD (cpm of T cells alone= 187 f 39).
b691-13-17.
W6132.
TABLE I1
rlFN-y treatment of line WM35 (primary melanoma): effecton MHC antigen expression and autologous lymphocyte
stimulation
Antigen Expression
(cpm in RIA f 1 SD)

Line WM35
Treated with:
Class I I ~

Experiment 1
Medium alone
IFN-7
489
IFN-y f 128.4d
Experiment 2
Medium alone
IFN-y
IFN-7 + 128.4

Class F

426 f 268
1,152
1,897f 3503,859
f 210
1.281
195f 1 1
1,238f 82
219f 21
1,440

P3

107
194
175303
f 140
353

f 1.31219
14
f 1722213.643
f 14
2.01324

k
k

1.006 f 160
5.825f 442
f 149

Lymphocyte
Stimulation
(S.I./cpm2 1 SD)

26
f 754
f
80
f

6
2.71531 f
41.517.922f 1.204
9.7f 1,857f 207

184f 24
219f 9
216 f 13

Cell line WM35 was incubated with medium alone, with rIFN-y preincubated with P3.or with rIFN-y preincubated
with the neutralizing anti-IFN-y antibody. 128.4 PI(50
(5000U) of rIFN-y were incubated for 18 hr at 4C with 10 PI 128.4
ascites (20.000neutralizing U)). This mixture was brought up to 10 ml and added to the preconfluent cells for 48 hr and
then was assayedby radioimmunoassay (RIA)for the binding of the same HLA class I1 and class I antibodies as in Table
I. The capacity of such cells to stimulateautologous T cellswas measured as described in Table I (cpm of T cells alone in
Expt. 1 = 164 f 7,in Expt. 2 = 191 f 15).
b691-13-17.
W6f32.
128.4= Anti-IFN-y.

WM35, Table 11), were incubated in medium alone or


medium containing rIFN-y. Incubation of nevus or melanoma cells with rIFN-y alone resulted in a threefold to
sixfold increase in expressionof HLA-DR antigens. This
increase was abrogated in the case of WM35 by preincubatingthe
neutralizing anti-IFN-y antibody (Mab
128.4) withrIFN-y prior to treating the
cell line (Table11).
When aliquots of these dysplastic nevus cells or primary
melanoma cells were tested a s stimulators, the IFN-ytreated cells stimulated autologous T cell incorporation
of [3H]thymidine by 2.5- to 17 times that provokedby
medium-treated cells (Tables I and 11). A s expected, rIFNy also enhanced the expression of HLA class I antigens.
Anti-HLA-DR antibody blocks stimulation by
rlFN-ytreated melanoma cells of T cells. To demonstrate that
HLA-DR antigens participate in the enhanced capacity
of rIFN-y-treated melanoma cells to stimulate autologous
T cells, we sought to block stimulation with the antiHLA-DR Mab L243. As shown in Figure 1 (right panel),
rIFN-y treatment of line WM35 increased HLA-DR antigen expression 1.4-fold. This wasassociated with a twofold increase in lymphocyte stimulation (1 2,614f 1757
cpm to 26,143 f 10,488 cpm). Addition of the anti-HLADR Mab to melanoma cells treated with rIFN-y inhibited
stimulation 78% with respect to the medium control and
79% with respect to the anti-HLA class I control (Fig. 1,
left panel). In a replicate experiment (not shown), inhibition with respect to the anti-HLA class I control was
96% (medium control not done).
Biologically late melanoma does
not stimulate T cells

LYMPHOCYTE
STIMULATION

HLA CLASS I and -DR


EXPRESSION(RIA)

cpm

40000-

30000.

20000

10 000.

NO Anti- AniiHLA HLA- HLA


AB HLADR Class I

p3

Anti-

Anti-

DR Class I

Figure 1 . Inhibition of melanoma cell stimulation of T cells by antiHLA-DR Mab. Line WM35 melanoma cells were incubated with medium
or rIFN-y. and their relative expression ofHLA-DR antigens (using the
anti-HLA-DR antibody L243) and class I antigens (using the anti-HLA
class I antibody W6/32)was measured in radioimmunoassays (right
panel, mean cpm f 1 SD).IFN-y-treated melanoma cells were then
incubated for 1 hr with each Mab at a final concentration of 1 d m 1 or
in medium alone, and were then co-cultured with autologous T cells.
Stimulation was measured as described in Table I (mean cpm f 1 SD)
(left panel).

308

INTERFERON-y AND MELANOMA

in vitro despite expressing HLA-DR, -09, and -DP.


Three of the four melanoma lines established from advanced disease (WM793, WM222, and WM806) constitutively expressed HLA-DR antigens, and their expression was increasedby 1.6-to 3.2-fold with rIFN-y treatment. As shown in Figure 2, rIFN-y treatment of line
WM806 augmented not only the expression ofHLA-DR
antigens but also that
of the additional HLA class I1
antigens DQ and DP (1.5-and 2.3-fold respectively). Nevertheless, suchcells failed to stimulateautologous T cells
(S.I. in two experiments c2.0). Line WM310 did not express HLA-DR antigens, and rIFN-y treatment failed to
induce their expression. In accord with our previous observation in cell lines from vertical growth phase and
metastases (5),none of the additional linesin thepresent
study, whether or not rIFN-y treated, stimulated proliferation of autologous T cells(data not shown).
T cells stimulated by autologous melanoma in vitro
produce IFN-7. To establish whether or not
autologous T
10000 1

Figure 2. Constltutive and augmented expression of HLA class 1 and


class 11 antigens byrIFN-7-treated metastatic melanoma cells. Line
WM806 melanoma cells were incubatedwith medium or rIFN-y and their
relative expression of antigens was measured by radioimmunoassay
(mean log cpm f 1 SD) by using the anti-HLA-DR antibody L243. the
anti-HLA-DQantibody N247. the antl-HLA-DP antibodyB7121.and the
anti-HLA class I antibody W6/32.

cells stimulated by primaryradialphasemelanoma


(WM35) produce IFN-y, 2 X lo5 melanoma cells (preincubated in medium alone or with rIFN-y) were washed,
mitomycin C treated, and co-incubated with T cells at a
1:1 ratio inquadruplicate wells of a microtest plate for6
days. Fifty percent of the supernatant (0.1 ml) was collected from each of four wells and pooled for assay of
IFN-y. [3H]Thymidinewas thenadded to the wells and its
incorporation by T cells assayed in the usual manner.
As
shown inTable 111,T cells cultured withautologous, radial
phase melanoma had increased thymidine incorporation
and produced IFN-y. No detectable IFN-y was produced
by the melanoma cells cultured alone, and only a small
amount was generated by T cells alone (unstimulated T
cells had a mean cpmof 156 k 44 ( 1 SD) compared with
background cpm = 23 f 2). As expected, the allogeneic
cell line (Raji) stimulated a brisk lymphocyte response,
and this was associated with the highest level of IFN-y
production. No IFN-y activity could be detected in the
cultures of cell lines from advanced disease or in their
co-cultures
with
autologous lymphocytes (data
not
shown).
Tcells infiltrating dysplastic nevt and melanomaare
activated and produce IFN-y in situ. To probe for evidence of T cell activation and IFN-y production in the
lesions of precursor nevi and melanoma, we used a Mab
to HLA-DR antigens (691- 13-17). the IL 2 receptor (antiTac), and IFN-y (113.4 and 128.4) in a n avidin:biotin
immunoperoxidase system. Serial frozen sections from
five common acquired nevi, 24 dysplastic nevi, 11 primary melanomas, and seven metastatic deposits were
examined for antibody binding.
Common acquired nevi contained only rare lymphocytes, and none of these few cells showed reactivity to
the panel of antibodies. A s shown in Table IV, the lymphocytes of all other lesions stained for HLA-DR antigens. The percentage of positive-staining cells ranged
from 5 to 90%, and all cases showed "strong" (2+ or
greater) reactivity. Anti-Tac binding to lymphocytes was
seen in the majority of lesions. In positive lesions, the
proportion of reactive lymphocytes ranged from 1 to 90%,
and 80%of the lesions showed strong binding.
Reactivity with both or either of the Mab to IFN-y was
seen in lymphocytes from the majority of lesions, with
from 1 to 100% of such cells positive. In the majority of
lesions studied with both antibodies
(17 of 31).both were
positive on infiltrating lymphocytes. In 31 of 39 lesions
studied with both antibodies or, for technical reasons,
one antibody,lymphocytes bound at least oneof the two
antibodies toIFN-7.

TABLE I11
Generatlon of ZFN-y by lymphocytes respondlng to autologous melanoma"
S.I./cpm i 1 SD

Stlmulating
Preincubatlon
CellResponding
Cell

Melanoma WM 35
None

um

None
Medium
IFN-7
Medium
IFN-y

T cells

None
Rail

3415,258
6219.694

324

IFN-I

Wmlt
<4

f 1.162
f 1.858

<4
12
12
4

997/155.464 f 15.983

~~~

"Cell line WM35 or the allogeneic lymphoblast Raji was incubated ln the presence or absence of rlFN-y and then
washed. After mitomycin C treatment. these cells were cultured with T cells autologous to line WM35. Lymphocyte
proliferation was assessed by [3H]thymidine incorporatlon
and then reported as the S.1. and raw cpm f 150 (cpmof T cells
alone = 156 2 44). In addltion. Supernatants wereassayed for IFN-y as described In Materials and Methods, and then
reported as units/mllliliter.

MELANOMA

AND

309

INTERFERON-7

TABLE IV
Immunoperoxldase analysls of lymphocytes ln melanocytlc leslons
Type of Leslon
[no. leslonswlth posltive lymphocytes/no. tested]

Antlbody
Dysplastlc nevl

Antl-IFN-y
113.46
128.4*

Both'
EitheP
Anti-TAC'
Anti-DRg

12/22, (35. 1 100)'


619.
14/22. (25. 1-90)
10/18
18/24
14/20. (50.718.
1-80)
24/24. (75. 5-100)
10/10.
5-100)
(75.

Prlmary melanoma
(50. 20-80)
7/10. (70. 10-100)
519
911 1
(60.2-90)

Metastatlc melanoma'
415. ( 10. 3- 10)
315. (50. 10-90)
214
414
315. (50. 20-80)
515.5-100)
(75.

Seven lesions were examlned but only


flve had any lymphocytlc Infiltrate.
b"Strong" (r2+)stainlng intensity and25% posltlve lymphocytes In 80% of leslons.
Mean percent lymphocytes positlve. and range.
Strong intensity and25% posltlve lymphocytes in 80%.
e l n some cases. results avallable for only one antlbody due to technlcal
problems.
'Strong intensity and25% positlve lymphocytes In >80%.
g Strong intensltyin 100%of cases.
a

Flgure3. Frozen sections of a prlmary


radialgrowthphasemelanoma,stalned
with AEC and a Mab to IFN-y (Mab 1 13.4)
in panels A (x150) and C ( ~ 1 0 0 0 )the
. IL 2
receptor (anti-Tac) ln panel D (x1000). or
with P3 control (x200) In panel E. The reaction product ts seen as a black rlm (red in
the orlginal sections) about the
lymphocyte
nuclei inC and D. easlly distinguishedfrom
brown melanln pigment or blue hematoxylin nuclear counterstainIn the orlglnal sections. In control sectlons(panel B).negative
lymphocytes (bottom rlght of panel E)are
in close relation to the larger tumor cells
(upper portfon of panel B).

The lesional nevus or tumor cells also reacted with


anti-IFN-y antibodies in six of 24 dysplastic nevi, eight
of 11 primary melanomas,and four of seven metastases.
In 16 of 16 lesions in which such cells were positive and
lymphocytes were present, the lymphocytes were also
reactive with anti-IFN-y antibodies. Further, the inten-

sity of staining was generally less on lesional cells than


on lymphocytes. These findings suggest that IFN-y is
exported to dysplastic nevus or tumor
cells from activated
lymphocytes. Anti-IFN-y antibodies did not bind to other
structures contained within the sections, including normal melanocytes, squamous epithelia of the epidermis

310

AND

INTERFERON-7

MELANOMA

that melanoma provokes a cellular immune response.


and that the nature
of this response is dependent on the
lesional step in tumor progression. Two of the proximal
steps in tumor progression, melanocytic dysplasia and
radial growth phase melanoma, are characterized by a
lymphocytic infiltrate largely composed of T cells (1, 2)
and often associated withevidence of tumor cell destruction (regression)(24). The later lesions in tumor progression have few infiltrating lymphocytes, suggesting
that a host immune response is lost or suppressed with
the acquisition of the fully malignant phenotype, a step
marked by the evolution of a nodule of vertical phase in
the plaque of radial growth phase (1).
An additional observationof potential immunologic relevance is that normal melanocytes and thecells of common acquired nevi do not constitutively express HLA
class I1 antigens in situ (although in
vitro the former can
with culture
be induced to do so with IFN-y and the latter
alone) (10). whereas melanocytes from lesions throughout the tumor progression pathway, from dysplastic nevi
to metastatic disease, frequently are
HLA class I1 antigen
positive (4,9, 10, 25). Thus,expression of HLA-DR antigens and other class
11, la-like molecules might be viewed
as reflecting a qualitativestep in the evolution of neoplasia of melanocytes. Further, thedirect role ofHLA class
I1 antigens in afferent (26) and efferent (27, 28) target
cell-T cell interactions suggeststhat they may play a role
in melanoma a s well. We and othershave established a n
in vitro model to explore the functional interaction of
melanoma and autologous T cells which delineates arole
for MHC class I1 antigens in theproliferative response of
T cells to melanoma. Thus, cell lines established from
biologically early melanomas stimulate proliferation of
autologous T cells, whereas tumor cells of late disease
are not stimulatory. In vitro, the expression of HLA class
I1 molecules by melanoma cells is necessary but insufficient to provoke a T cell response, suggestingthat T cells
recognize as yet unidentified, non-MHC-related, melanoma-associated antigens (MAAS)-perhaps lost in advanced disease-in association with self MHC class I1
antigens (5,6).
The data presented here expand
these observations
and additionally characterize the in vitro and in vivo
immune response to neoplastic melanocytes proximal in
the pathway of tumor progression. That cells from a
dysplastic nevus stimulate proliferation of T cells (Table
I) parallels the characteristichistologic finding of infiltrative lymphocytes at this step in tumor progression, and
suggests that suchcells, a s well a s those of radial growth
phase melanoma, express a n immunogenic MAA. That
there is a proportionate increase in HLA class I1 antigen
expression and autologous T cell stimulation by neoplastic melanocytes treated with IFN-y (Tables I and 11) is
additional evidence for the direct role of such tumorDISCUSSION
associated MHC antigens in immune recognition and reIn common with other malignancies, melanoma
evolves sponse. This conclusion is bolstered by the demonstrain stepwise fashion. The evident lesions in tumor
pro- tion (9. 10) that IFN-y appears to increase class I and I1
gression in melanoma include common acquired melan- expression without modulating the cell surface expresocytic nevi, dysplastic nevi, a cutaneous plaque of inva- sion of a t least themajority of non-MHC associated MAA.
In the studyof Houghton et al.,IFN-y didnot increase the
sive melanocytes with no apparent capacity for metasprevious
tasis (theradial growth phase), acomplex tumor addition- cell surface expressionof 17 MAA (9),and in our
ally containing cells capable of local tumorigenic growth study none of the seven MAA measured was increased
and of metastasis (the vertical growth phase),
and meta- (10).Although it is likely that T cell proliferation is trigstatic disease itself (1). A variety of observations suggests gered by seeinga melanoma-associated (non-MHC)an-

and skinappendages, nerves, blood vessels, and arrectae


pilae muscles.
The morphology of the lymphocytes stained with antibodies to IFN-7, to HLA-DR antigens, or to IL 2 receptors
was similar to that observed with antibodies to T cell
markers (2). The reactive cells were similar in size and
nuclear andcytoplasmic characteristics to cells staining
only with T cell markers. With all of the antibodies, the
staining appeared to be either membrane associated or
confined to a thin rim of cytoplasm that completely surrounded the unstainednucleus. The intensityof staining
varied among the antibodies, being consistently strongest
with the antibody to HLA-DR antigens. The two antibodies to IFN-y showed some variation of staining intensity from case to case, but were definitely positive (Fig.
3C) or clearly negative. With anti-Tac (Fig. 3D), staining
intensity was similar to that of the anti-IFN-y reaction.
Equivocal staining was occasionally seen with each of
the antibodies, and was considered negative.
The patternof distribution of the reactive lymphocytes
was similar with all the antibodies. When infiltrative
(amongst lesional cells) and noninfiltrative (at the base
of the lesion) compartments of the host response were
distinguished, reactive cells appeared to be distributed
equally between the two.
The antibodyto HLA-DR antigens stainedlymphocytes
as described above, but also stained cells larger than
lymphocytes with more abundant cytoplasm and having
the morphology of macrophages (though melanin-laden
macrophages were generally negative). It alsostained
endothelial cells. In the epidermis, this antibody stained
both high-level dendritic cells consistent with Langerhans cells and some atypical lesional cells of dysplastic
nevi and melanomas, a s previously reported (4). Such
staining appeared to be diffuse throughoutthe cytoplasm
of randomly scattered lesional cells.
When the classof lesion was considered,no clear qualitative differences in the antigenic profile of responding
lymphocytes were observed, except a s noted in the few
lymphocytes associatedwith
common acquired nevi,
which were negative with all the antibodies except those
marking T cells. In contrast and as previously reported
(1, 2, 4.23).there were clear differences in the number
of lymphocytes infiltrating the lesional types. The density
of the response was less inthe dysplastic nevi compared
with the brisk response characterizing primary melanomas. Among melanomas, the response was more dense
in radial than in vertical growthphase, and was least in
the metastases (two of the seven had no detectablelymphocytes). In each of these lesional classes, however, cells
positive for IFN-y, IL 2 receptor, and HLA-DR antigens
were observed in about equal proportions and staining
intensities.

31 1

INTERFERON-7 AND MELANOMA

tigen in the context of HLA class I1 antigens (5). in the


experiments reported here it is likely that only MHC
antigens are increased, and it
is this increase that is
associatedwith
a n augmented lymphocyte response.
That stimulation is significantly blocked by a Mab to
HLA-DR antigens is also evidence of the participation of
these molecules in T cell recognition and response. Although the studiesreported here demonstratethe participation of melanoma-associated HLA-DR antigens inlymphocyte stimulation, they do not address the potential
role of HLA-DQ and -DP molecules. Given the functionof
these other class
I1 antigens inallogeneic T cell responses
(29).they are likely functional in the response of autologous T cells to tumorcells.
The information contained in our immunoperoxidase
study is insufficiently quantitative to reliably compare
the density of lymphocytic infiltration with the number
of dysplastic nevus cells and melanoma cells expressing
HLA class I1 antigens. However, lesions that do not express such antigens (common acquired nevi) invariably
had no activated lymphocytes. By contrast, occasional
nevus cells expressing HLA class I1 antigens have been
reported in those regions of halo nevi that areinfiltrated
by lymphocytes (30).
In all studies to date, cell lines from advanced melanoma have failed to provoke a proliferative lymphocyte
response (5, 6, 31).and this finding wasobtained in the
fouradditionallinesstudiedhere.
Although our data
provide no information on why such cells are incompetent stimulators, they demonstrate that isitnot the failure of such cells to express HLA-DR antigens, because
three of the four lines
both constitutively expressedsuch
molecules and their expression was augmented by IFN--y
treatment. Our studies with line
WM806 also demonstrate constitutive and augmentable expression of other
HLA class I1 antigens (HLA-DQ and -DP) by melanoma
cells of advanced disease, and others have shown that
a
significant proportion of metastatic lines also expresses
these additional specificities(32).
These datasuggest that
it is unlikely that failure to express these antigens is a
general cause of the failure to stimulate lymphocytes in
vitro. That the majority of metastatic lesions contain at
least a few activated T cellsand IFN-7 activity (although
cultured metastases did not stimulate IFN-y production
by T cells) is a discordantfinding,but
likely reflects
mechanisms unexplored here, including the possibility
of antigen-independent recruitment of lymphocytes by
cytokines generated by melanoma cells (33).
Both the in vitro and insitustudies
reported here
suggest a relevant role for the lymphocytic infiltrate associated with melanomaand itsmost common precursor,
the dysplastic nevus. That the T cells of dysplastic nevi
and primary melanoma are activated (HLA-DR antigen
and IL 2 receptor positive) indicates that they are not
merely passengers, a s they appear to be when found in
small numbers in common acquired nevi. Further, the
generation of IFN--y in vitro by T cells incubated with
cells of a biologically early melanoma, and the finding
of
IFN-y reactivity in lesional lymphocytes, are evidence
that T cell activation is likely consequential.
Taken together,the datareported here suggestthat the
lymphocyte infiltration that characterizes lesions early
in tumor progression in melanoma is relevant to this
evolutionary process and is provoked by non-MHC anti-

gens that first appear at the lesional step of dysplastic


(precursor)nevi. Late lesions in tumorprogression have
a n attenuated immune response. IFN--y may play a central role in theregulation of the immune response to early
lesional steps. It is produced by T cells, likely activated
by melanoma-associated antigens seen in the context
of HLA class I1 antigens. In turn, IFN-y enhances the
expression of HLA class I1 antigens, and such enhancement is associated with additional Tcell activation.

Acknowledgments. We thank Ms. Rosemary Stewart


for excellent technical work and specimen collection and
Dr. Georgio Trinchieri for performing the assays of supernatant interferon.
REFERENCES
1. Clark, W. H., D. E. Elder, D. Guerry, M. N. Epstein. M.H. Greene,
and M. VanHorn. 1984. A study of tumor progression:the precursor

2.

3.

4.

5.

6.

7.
8.
9.

10.

11.

12.

13.
14.

15.
16.

17.
18.
19.

20.

lesions of superficial spreading and nodular melanoma. Hum. Pathol. 15:1147.


Kornstein, M. J.,J. S. J. Brooks. and D. E. Elder. 1983. Immunperoxidase localization of lymphocyte subsets in the host response to
melanoma and nevi. Cancer Res.43:2749.
Natali, P. G.. P. Cordiali-Fei. R. Cavaliere, F. Filippo, V. Q u a r a n t a ,
M. A. Pellegrino, and S . Ferrone. 1981. la-like antigens on freshly
explanted human melanoma. Clin. Irnmunol. Imrnunopathul.
19:250.
Thompson. J. J.. M. F. Herlyn, D.E. Elder, W. H. Clark, 2. Steplewski, and H. Koprowski. 1982. Expression ofDR antigens in
freshly frozen human tumors. Hybridoma 1; 1 6 1 .
Guerry, D., M. A. Alexander, M. F. Herlyn, L. M. Zehngebot. K. F.
Mitchell, C. M. Zmijewski, and E. J. Lusk. 1984. HLA-DR histocompatibility leukocyte antigens permit cultured human melanoma cells
from early but notadvanced disease to stimulate autologous lymphocytes. J . Clfn. Invest.73:267.
Fossati, G.. D. Taramelli, A. Dalsari, G . Bogdanovich. S . Andreola.
and G . Parmiani. 1984. Primarybutnot metastatic human melanoma expressing DR antigens stimulate autologous lymphocytes. Int.
J . Cancer 33:591.
McKimm-Breschkin.J. L.. P. L. Mottrem. W. R. Thomas, andJ. F.
A. P. Miller. 1982. Antigen-specific production of immune interferon
by T cell lines. J . Exp. Med. 1 5 5 1 204.
Basham, T. Y.,and T. C. Merigan. 1983. Recombinant interferon-y
increases HLA-DR synthesis and expression. J . Irnmunol. 130:1492.
Houghton, A. N.. T. M. Thompson, D. Gross, H. F. Oettgen. and L.
J. Old. 1984. Surface antigens of melanomas and melanocytes.
Specificity of induction of la antigens by human y-interferon. J . Exp.
Med 160:255.
Herlyn, M. F., D. Guerry. and H. Koprowski. 1985. Recombinant yinterferon induces changes in expression and shedding of antigens
associated with normal human melanocytes, nevus cells, and primary and metastatic melanoma cells. J . Irnmunol. 134:4226.
Taramelli, D., G. Fossati. A. Mazzocchi, D. Delia, S. Ferrone. and
G . Parmiani. 1986. Classes I and I1 HLA and melanoma-associated
antigen expression and modulation on melanoma cells isolated from
primary and metastatic lesions. Cancer Res.46:433.
Mitchell, K. F., J. P. Furhrer, 2 . Steplewski, and H. Koprowski.
1980. Biochemical characterization of human melanoma cells: disUSA
section with monoclonal antibodies. Proc.Natl.Acad.Scl.
77:7287.
Lloyd, K. 0..J. Ng, and W. H. Dippolo. 1981. Analysis of the
biosynthesis of HLA-DR glycoproteins in human malignant melanoma cell lines. J . Irnrnunol. 126~2408.
Kavathas. P.,F. H. Bach. andR. DeMars. 1980. Gamma-ray induced
loss of expression of HLA and glyoxylase I alleles in lymphoblastoid
cells. Proc. Natl. Acad. Sci. USA 77:4251.
Kavathas. P.. R. DeMars,and F. H. Bach. 1980. Homozygous HLA
mutants of lymphoblastoidcells: a new cell type for histocompatibility testing. Hum. Irnmunol. 1:317.
Lampson. L. A.. and R. Levy. 1980. T w o populations of la-like
molecules on a human B cell line. J . Irnmunol. 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 DRrelated 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 class I1 HLA antigen. Nature 304:358.
Novick. D.. 2. Eshhar. D. G. Fischer, J. Friedlander, and M. Rubinstein. 1983. Monoclonal antibodies to human interferon-y: production, affinity purification and radioimmunoassay. EMBO J.
2:1527.
Uchiyama, T..S . Broder, and T. A. Waldmann. 1981. A monoclonal

312
21.

22.

23.

24.
25.
26.

MELANOMA
INTERFERON-? AND

antibody (anti-Tac]reactive with activated and functionally mature


human T cells. J . Irnmunol. 125: 1393.
Herlyn, M. F..D. Herlyn. D.E. Elder, E. Bondi, D. LaRossa, R.
Hamilton, H. F. Sears, G. Etalaban, D. Guerry. W.H. Clark, and H.
Koprowski. 1983. Phenotypic characteristics of cells derived from
precursors of human melanoma. Cancer Res. 43~5502.
Dayton, E. T.. M. Matsumoto-Kobayashi. B. Perussia. and G. Trinchieri. 1985. Role of immune interferon in themonocyte differentiation of human promyelocytic cell lines induced by leukocyte conditioned media. Blood 66:583.
Ruiter. D. J.,A. K.Bahn, T. J. Harrist, A. J. Sober, and M. C. Mihm.
1982. Major histocompatibility antigens and mononuclear inflammatory infiltrate in benign nevomelanocytic proliferations and malignant melanoma. J . Immunol. 129:2808.
Cooper, P. H., H. J. Wanebo. and R. w. Hagar. 1985. Regression in
thin melanoma: microscopic diagnosis and prognostic importance.
Arch. Dermatol. 121:1127.
Natali, P. G., C. DeMartino. V. Quaranta, A. Bigotti, M. A. Pellegrino, and S. Ferrone. 1981. Changes in la-like antigen expression
on malignant human cells. Irnrnunogenetfcs 12:409.
Issekuntz, T., E. Chu, and R. Geha. 1982. Antigen presentation by
human B-cells: T cell proliferation induced by Epstein-Barr virus B

lymphoblasticd cells. J. Irnrnunol. 129: 1446.


27. Reinherz. E. L., S. Meuer, K. A. Fitzgerald, R. E. Hussey. H. Levine,
and S. F. Schlossman. 1982. Antigen recognition by human T lymphocytes is linked to surface expression of the T3molecular complex.
Cell 30:735.
28. Fleischer, B. 1984. Acquisition of specific cytotoxic activity by human T4+T lymphocytes in culture. Nature308:365.
29. Myers, L. K., E. J. Ball, G. Nunez, and P. Stastny. 1986.Recognition
of class I1 molecules by human T cells. I. Analysis of epitopes ofDR
and DQ molecules in a DRwl 1. DRw52, DQw3 haplotype. Immunogenetics 23: 142.
30. Bergman. W., R. Willemze. C. deGraaf-Reitsma, and D. J. Ruiter.
1985. Analysis of major histocompatibility antigens and the mononuclear cell infiltrate in halo nevi. J . Invest. Derrnatol. 8525.
31. Pollack, M. s. 1981. Thedetection of weak allogeneic stimulation by
DR-positive tumor cell lines. Transplant. Proc. 13~1947.
32. Natali. P., A. Bigotti, R. Cavalieri. M. R. Nicotra. R. Tesse, D.
Manfredi. Y.Chen. L. M. Nadler, and S.Ferrone. 1986. Gene products of the HLA-D region in normal and malignant tissues of nonlymphoid origin. Hum. Irnmunol. 15220.
33. Luger, T. A., A. Kock, M. Danner, and M. Micksche. 1986. Human
melanocytes and melanoma cells produce distinct cytokines. Clin.
Res. 343764A.

Você também pode gostar