13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Scand. J. Immunol. 46, 379–387, 1997

Interferon-g-Induced MHC Class I Expression and Defects in

Jak/Stat Signalling in Methylcholanthrene-Induced Sarcomas
I. M. SVANE*, A.-M. ENGEL*, M. NIELSEN & O. WERDELIN
Institute for Medical Microbiology and Immunology, University of Copenhagen, Panum Institute,Copenhagen, Denmark

(Received 10 April 1997; Accepted in revised form 26 June 1997)

Svane IM, Engel A-M, Nielsen M, Werdelin O. Interferon-g-Induced MHC Class I Expression and Defects in
Jak/Stat Signalling in Methylcholanthrene-Induced Sarcomas. Scand J Immunol 1997;46:379–387
Seventy-eight uncloned tumour cell lines, each established from a primary sarcoma induced with methyl-
cholanthrene in immunocompetent nu/þ BALB/c and C.B.-17 mice or in immunodeficient nu/nu BALB/c and
severe combined immunodeficient (SCID) mice, were examined for sensitivity to interferon-g (IFN-g) as
measured by tumour cell augmentation of major histocompatibility complex (MHC) class I expression. The
tumour cells were cultured with IFN-g and their expression of Kd, Dd and Ld was measured by fluorescence-
activated cell sorter analysis. All but three of the 78 tumour lines up-regulated Kd, Dd and Ld to a variable
degree in response to IFN-g, indicating that IFN-g resistance is not a common property of these sarcomas. The
tumour cell lines varied greatly in their MHC class I expression before as well as after IFN-g stimulation.
There was a tendency towards a higher MHC expression after IFN-g stimulation in tumour lines from
immunocompetent mice compared to immunodeficient mice, but no common maximum MHC class I
expression level was found for the 78 tumour cell lines. Three of the tumour lines, all from immunodeficient
mice, completely failed to respond to IFN-g by up-regulating MHC class I expression. The same three also
displayed absence of IFN-g-induced Stat1b tyrosine phosphorylation and low Stat1a tyrosine phosphoryla-
tion, indicating a defect in the signal transduction pathway affecting phosphorylation of Stat1. These findings
strongly suggest a link between defects in Stat1 phosphorylation and the failure to up-regulate MHC class I. In
all tumour lines tested, the Stat1 Western blotting revealed a 78 kDa protein (p78) not previously described.
Ole Werdelin, Institute for Medical Microbiology and Immunology, University of Copenhagen, Panum
Institute Room 22.5.24, Blegdamsvej 3C, DK-2200 N, Copenhagen, Denmark

caused by low levels of MHC class I expression may enable

INTRODUCTION
Major histocompatibility complex (MHC) class I molecules are
known to play an important role in tumour growth and metastasis
[1–8]. Several reports indicate that the level of MHC class I
expression on the surface of tumour cells has a decisive influence
on whether a tumour will be rejected or accepted when trans-
planted to histocompatible immunocompetent recipients [9–11].
Tumour cells expressing MHC class I above a certain threshold
are capable of effectively presenting tumour-specific antigens in
the form of peptides bound to MHC class I molecules on the
tumour cell surface [12], and will be susceptible to recognition
and lysis by tumour antigen-specific cytotoxic T cells [12–14].
On the other hand, insufficient presentation of tumour antigen
*Inge Marie Svane and Anne-Marie Engel have contributed equally to this
work.
tumour cells to escape detection by the T-cell system of the host.
It has been proposed that many tumours developing in hosts with
a normal immune system undergo immunoselection at a very
early stage of their development, as a result of which, clones
of immunogenic tumour cells become eliminated by tumour-
specific cytotoxic T cells [8, 15–17]. This idea is supported by
the fact that the cells of many human tumours express relatively
little MHC class I on their surface [18–20].
In preceding reports we have described and partially charac-
terized a panel of tumours induced with the chemical carcinogen
3-methylcholanthrene (MCA) in nude BALB/c mice and normal
littermates [21]. Nude mice developed tumours with a higher
incidence and with shorter induction times than normal mice. To
explain this we suggested that nude mice, due to their lack of T
cells, are more susceptible to chemical carcinogenesis than are
normal mice. The tumours were found to be very heterogeneous

q 1997 Blackwell Science Ltd 379


380 I. M. Svane et al.
regarding their expression of MHC class I, and tumours from
nu/þ mice expressed more MHC class I than tumours from
nude mice in general. No connection between MHC expression
and immunogenicity was observed, but tumours originating
from nude mice were more immunogenic as they were rejected
after transplantation to immunocompetent recipients at a higher
rate than were tumours originating from immunocompetent
mice [22]. In parallel, we have induced a panel of tumours in
severe combined immunodeficient (SCID) and normal C.B.-17
mice, SCID mice lacking T and B cells due to non-functional
rearrangement of antigen-receptor genes [23]. Like the tumours
from nude mice, tumours from SCID mice were found to be
more immunogenic than tumours from normal C.B.-17 mice, as
witnessed by a higher rate of rejection of SCID tumours than
non-SCID tumours after transplantation to immunocompetent
recipients [24].
It is of interest to study if the regulation of MHC class I
expression is abnormal in these MCA tumours. Interferon-g
(IFN-g) is an important cytokine in the regulation of MHC
class I expression. It is important in cellular immune defence
against viruses and malignancies [25–29] and it is secreted by
activated T cells (cytotoxic and Th1) and natural killer (NK)
cells. IFN-g also stimulates the expression of several cellular
proteins in addition to MHC class I. Among these are MHC class
II molecules, proteasome subunits LMP-2 and -7, heat-shock
proteins, TAP-1 and -2 and intracellular adhesion molecule-1
(ICAM-1), all believed to be involved in antigen processing and
presentation [30–33]. In this way IFN-g increases the amount of
viral or tumour antigen that can be presented to T cells. IFN-g
displays its biological activity through binding to the IFN-g
receptor which is constitutively expressed on almost all cell types
[34]. This interaction is thought to induce a dimerization of the
receptor leading to association of Jak1 and Jak2, two members of
the Janus tyrosine protein kinase (Jak) family, resulting in their
activation by tyrosine phosphorylation. Subsequently, the IFN-g
receptor and the signal transducer and activator of transcription
molecule, Stat1, are tyrosine phosphorylated by Jak1 and Jak2.
Following phosphorylation, Stat1 forms homodimers that
migrate to the nucleus, where they bind to promotors of IFN-g
responsive genes. Through this binding Stat1 activates genes
encoding proteins responsible for mediating the various actions
of IFN-g [35–37]. Mutations affecting the function of any
component of this signalling pathway would be expected to
reduce the ability of a tumour cell to process and present
tumour-specific antigens by reducing its responsiveness to
IFN-g. Jak and Stat proteins are not only involved in IFN-g
signalling, but seem to be involved in the signalling induced by
many different cytokines [35, 38].
In the present experiments, we have studied the reaction to
IFN-g of the MCA tumours described above from nude and
congenic nu/þ mice and SCID and congenic normal mice. The
experiments were performed on 78 tumour cell lines, each
derived from a separate tumour induced with 3-methylcholan-
threne in one of the four types of mice. The tumour cell lines
were exposed to IFN-g in culture and the expression of MHC
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
class I Kd, Dd and Ld was determined by fluorescence-activated
cell sorter (FACS) analysis. Furthermore, the IFN-g receptor
expression and IFN-g-induced activation of Jak2 and Stat1 were
examined on those tumour cell lines that were found to be
incapable of up-regulating MHC class I expression when treated
with IFN-g.
MATERIALS AND METHODS
Mice. Female mice, 6–8 weeks of age, were used. They were immuno-
competent, normal BALB/c (nu/þ) or congenic, T cell-deficient, athy-
mic, BALB/c (nu/nu) mice, and immunocompetent, normal C.B.-17
(non-SCID) mice, or congenic SCID/SCID (SCID). All mice were
purchased from Gl. Bomholtgaard, Ry, Denmark, and kept under
specific pathogen free housing conditions.
Tumour cell lines. The present experiments were performed on a panel
of sarcomas induced with either 0.5% (40 mg) or 0.1% (8 mg) 3-methyl-
cholanthrene (MCA). The method used for induction of the tumours was
that of Bartlett [39]. Briefly, tumours were induced by subcutaneous
implantation of filter discs, containing the MCA dissolved in paraffin, on
the back of the mice. Each mouse produced only one tumour and a
tumour cell line was established from this tumour by taking into culture
freshly excised tumour tissue from a mouse carrying a tumour approxi-
mately 10 × 10 mm as described [21]. No cloning or deliberate selection
pressure has been applied to the established tumour cell lines, and the
expression ‘tumour cell line’ is used below only to remind the reader of
the fact that the tumour cells in these experiments were cells that had
been propagated in culture. Cells were propagated in complete tissue
medium [RPMI-1640 supplemented with 10% fetal calf serum (FCS)] at
378C in an atmosphere of 5% CO2 and relative humidity of at least 95%.
All plasticware used for tissue culture was from A/S Nunc (Roskilde,
Denmark).
Analysis by flow cytometry. FACS analyses were carried out as
previously described [21]. In brief, 5 × 105 cells were washed twice in
phosphate-buffered saline (PBS) supplemented with 2% FCS and 0.1%
sodium azide, a FACS buffer used throughout the experiment. Incuba-
tion with primary antibody in saturating concentration lasted 30 min. The
primary antibodies used were anti-H-2 Kd, 1 : 10 (31-3-4S, cat.no. CL
9011-A), anti-H-2 Dd, 1 : 320 (34-5-8S, cat.no. CL 9008-A), anti-H-2 Ld,
1 : 40 (34-5-7S, cat.no. CL 9009-A), all obtained from Cedarlane
Laboratories Ltd (Hornby, Ontario, Canada). Purified anti-mouse IFN-
receptor antibody, GR20, 1:20 was a kind gift from Dr S. W. Russel,
(Wilkinson Laboratory for Cancer Research, Kansas City, KS, USA)
[40]. Secondary fluorescein isothiocycanate (FITC)-conjugated antibody
(rabbit anti-mouse immunoglobulin) was used in 1 : 20 dilution for a
30 min incubation period at 408C in the dark. Isotype-matched, non-
binding antibodies were used as controls for specific binding. FACS
analyses were standardized using Fluorescein Microbeads Standards
(cat.no. 891, Flow Cytometry Standards Corp., Leiden, the Netherlands)
and analysed on a FACScan flow cytometer (Becton Dickinson, San
Jose, CA, USA). The FACScan programme was used to calculate the
median fluorescence intensity of 5000–10 000 tumour cells.
IFN-g stimulation of cells. To determine the optimal conditions for
IFN-g stimulation of MHC class I expression, tumour cell lines were
treated with different concentrations of IFN-g ranging from 0 to 40 U/ml
for 48 h, 10 U/ml was found to be optimal (data not shown). Accord-
ingly, IFN-g stimulation prior to FACS analysis was carried out by
incubating cells with 10 U/ml recombinant mouse IFN-g (code: MG-
IFN, Genzyme, Cambridge, UK) in complete tissue medium for 48 h at

378C. Cells for Western blotting analysis were stimulated with either
10 U/ml medium IFN-g for 48 h or 1000 U/ml medium IFN-g for 10 min
at 378C.
Sample preparation and immunoprecipitation. Samples were pre-
pared and immunoprecipitated as decribed [41]. Twenty million cells/
experiment, stimulated or unstimulated, were lysed in ice-cold lysis
buffer [10% nonidet-P40 (NP-40), 20 mM Tris pH 8.0, 137 mM NaCl,
10% glycerol with the following inhibitors added: 5 mM ethylenediami-
netetraacetic acid (EDTA), 1 mM Na3VO4, 10 mg/ml aprotinin, 4 mM
iodoacetamide and 1 mM phenylmethylsulphonyl fluoride (PMSF)].
The lysates were precleared using protein A–Sepharose. Anti-phospho-
tyrosine antibody (4G10, Upstate Biotechnologies, New York, NY,
USA), 8 ml/sample, was added to the precleared lysates and the lysates
were mixed for 3½ h at 48C. After 2 h of mixing, 55 ml of protein A–
Sepharose was added to the mixture. The immunoprecipitated proteins
were pelleted and washed five times in lysis buffer.
Western blotting. Immunoprecipitated proteins were subjected to
analysis by SDS–PAGE. The separated proteins were electrophoreti-
cally transferred to nitrocellulose [1 mA/cm2] using a semi-dry blotting
apparatus. The nitrocellulose was blocked for 1 h in blocking solution
[3% skimmed milk powder, 1% bovine serum albumin (BSA) in PBS]
and subsequently probed with murine monoclonal anti-Stat1 antibody
[1 : 250] (cat.no. S21120, Transduction Laboratories, Lexington, KY,
USA) or rabbit polyclonal anti-Jak2 antibody [1 : 1000](cat.no. 06-255,
Upstate Biotechnology) in blocking solution for 2 h. After washing, the
nitrocellulose was incubated with rabbit anti-mouse or goat anti-rabbit
antibody conjugated to horseradish peroxidase in blocking solution for
1 h. After a final wash the blots were evaluated using enhanced chemi-
luminescence (ECL), according to the manufacturer’s instructions
(Amersham, Bucks, UK).
Statistics. Student’s t-test was used for calculation of statistical
significances.
RESULTS
IFN-g-induced MHC class I expression
Of the 78 tumour cell lines examined, 20 had developed in nu/þ
mice, 15 in nude mice, 23 in non-SCID mice and 20 in SCID
mice. They were grown in culture with or without 10 U/ml IFN-g
for 48 h, before being subjected to flow cytometric analysis of the
surface expression of the MHC class I molecules Kd, Dd and Ld.
Figure 1 depicts the MHC class I expression of the tumour cell
lines before and after treatment wth IFN-g. The median expres-
sion of Kd, Dd and Ld on unstimulated, as well as IFN-g-
stimulated, tumour cell lines is very heterogeneous and there
was no common maximal MHC expression level after IFN-g
treatment of the individual tumour cell lines. The sensitivity to
IFN-g was very variable. Overall, the tumour lines expressing
low levels of MHC class I when unstimulated displayed the
lowest levels of MHC after IFN-g stimulation, while the tumour
lines expressing a high level of MHC class I when unstimulated
attained the highest MHC class I levels upon stimulation.
Although not statistically significant, the tumour cell lines from
tumours induced in nu/þ mice generally expressed more Kd, Dd
and Ld than did the tumour cell lines from tumours induced in
nude mice before, as well as after, IFN-g stimulation (Table 1).
In contrast, unstimulated tumour cell lines from non-SCID and
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Defective IFN-g Signalling in Sarcoma Cells 381
SCID mice expressed almost equal amounts of Kd, Dd and Ld,
while IFN-g-stimulated tumour lines from non-SCID mice had a
higher mean expression of the three MHC class I molecules than
tumour cells from SCID mice (statistically significant for Dd,
P < 0.05) (Table 1).
Three tumour cell lines were incapable of increasing
their MHC class I expression in response to IFN-g and
two displayed very low responses (these results were repro-
ducible). All had a low basic MHC class I expression and all
originated from immunodeficient mice (Table 2). Six tumour
cell lines were found to have an extremely high MHC class I
expression after IFN-g stimulation. All of these tumour lines
had a high basic MHC class I expression and all except one
originated from immunocompetent, normal mice (Table 2). A
total of seven tumour cell lines (Table 2) exhibiting high, low, or
no response to IFN-g stimulation were subjected to further
investigations, in order to reveal possible defects in the IFN-g
signalling pathway.
IFN-g-induced ICAM-1 expression
First, we tested the expression of another IFN-g-inducible
molecule, ICAM-1, to find out whether the tumour cell lines
were able to respond to IFN-g by induction of this molecule.
ICAM-1 is the ligand for lymphocyte function-associated anti-
gen-1 (LFA-1), expressed on T lymphocytes and it can be
expressed by many types of cells. ICAM-1 has been shown to
be inducible on fibroblasts by IFN-g [42]. The ability of IFN-g to
induce ICAM-1 expression was tested by FACS analysis on the
seven selected tumour cell lines (data not shown). It was found
that IFN-g was able to induce surface expression of ICAM-1 in
low but significant amounts in some, but not all, of the tumour
cell lines capable of up-regulating MHC class I. None of the
three tumour lines insensitive to IFN-g with respect to up-
regulation of MHC class I expression were able to express
ICAM-1 in response to IFN-g.
IFN-g receptor expression
The IFN-g receptor surface expression of the seven tumour cell
lines exhibiting very high, very low, or no increase in MHC class
I expression after IFN-g stimulation (Table 2) was then measured
by FACS analysis. As seen from the FACS profiles in Fig. 2,
which are representative of all tumour lines tested, the tumour
cell lines displayed the same low surface expression of IFN-g
receptor. There was no difference in the IFN-g receptor level
among tumour cell lines displaying different MHC class I
responsiveness to IFN-g. We therefore examined the IFN-g-
induced activation of the Jak/Stat signalling pathway.
IFN-g-induced activation of Jak2
Activated Jak2 is believed to be responsible for the phosphoryla-
tion of Stat1 [43]. The ability of IFN-g to induce phosphorylation
of Jak2 was investigated in the same seven tumour cell lines
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387

unstimulated tumour cells (– IFN) and on tumour cells stimulated with 10 U/ml medium of IFN-g for 48 h (þ IFN).
Fig. 1. Expression of Kd, Dd and Ld for tumour cell lines originating from nu/þ, nude, non-SCID and SCID mice. The MHC class I expression was measured by flow cytometry on
382 I. M. Svane et al.
 13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Defective IFN-g Signalling in Sarcoma Cells 383

Table 1. Induction of MHC class I by IFN-g stimulation of tumour did not correlate with the ability to up-regulate MHC class I
cells originating from MCA-induced tumours in various mouse strains expression in response to IFN-g.

MHC Mean MHC Mean MHC

Number class I expression expression IFN-g-induced activation of Stat1
of tumours molecule – IFN-g þ IFN-g
As no defects in tyrosine phosphorylation of Jak2 were found, we
20 nu/þ K d
102 343 continued by examining the tyrosine phosphorylation of Stat1 in
Dd 113 654 the same seven tumour cell lines with high, low and no response
Ld 35 174 to IFN-g stimulation. Cells were either not stimulated (Fig. 4,
15 nude Kd 67 249 lanes 1, 3 and 5) or stimulated with 10 U/ml IFN-g for 48 h (Fig.
Dd 84 503 4, lanes 2, 4 and 6), as the cells tested for up-regulation of MHC
Ld 24 142 class I expression. All cell lines had a relatively high background
level of tyrosine phosphorylated Stat1b proteins and weakly
23 non- Kd 67 277
phosphorylated Stat1a proteins. Unexpectedly, the Stat1 anti-
SCID Dd 77 660
Ld 21 170
body detected three distinct bands: A weak band at 91 kDa
corresponding to Stat1a, one at 84 kDa corresponding to
20 SCID Kd 73 247 Stat1b and an additional band at approximately 78 kDa (p78).
Dd 86 469 The p78 protein was detected in both unstimulated and IFN-g-
Ld 22 111
stimulated tumour cells as well as in the fibroblastic cell lines
3T3 and L-cells (data not shown).
In tumour cell lines responding to IFN-g by pronounced or low
exhibiting either very high, very low, or no MHC class I up- up-regulation of MHC class I, stimulation with IFN-g induced a
regulation in response to IFN-g stimulation. The tyrosine phos- strong increase in the tyrosine phosphorylation of Stat1a and
phorylation of Jak2 in unstimulated tumour cells and tumour Stat1b (Fig. 4, lanes 2 and 4). However, in the three tumour cell
cells stimulated with 1000 U/ml IFN-g for 10 min was examined lines that completely failed to up-regulate MHC in response to
by Jak2 immunoblotting of anti-phosphotyrosine immunopreci- IFN-g, tyrosine phosphorylation of Stat1b did not increase at all
pitates. From Fig. 3 it can be seen that Jak2 was not tyrosine and tyrosine phosphorylation of Stat1a only increased slightly
phosphorylated in unstimulated tumour cells, but that IFN-g (Fig. 4, lane 6). This points to a correlation between Stat1
induced a strong tyrosine phosphorylation of Jak2 in all tumour phosphorylation and MHC class I up-regulation. The phosphor-
cell lines. The IFN-g-induced tyrosine phosphorylation of Jak2 ylation level of p78 did not change in response to IFN-g in any of
was equally strong in all tumour cell lines tested and therefore the cell lines.

Table 2. Tumour cell lines displaying no, low, or high response to IFN-g stimulation

Kd Dd Ld
IFN-g Mouse Cell expression expression expression
response strain line –/þ IFN –/þ IFN –/þ IFN

No nude* n93.2 14/28 14/24 7/13

nude* n94.1 26/50 4/16 1/3
SCID* s87.2 23/30 1/3 1/1

Low SCID* s71-1 35/105 128/250 36/111

SCID* s73.4 13/100 9/226 2/43

High nu/þ þ 80.4 103/673 187/1408 62/398

nu/þ þ 116.2 107/708 42/1198 7/310
nu/þ þ 120.1 296/873 129/1189 11/165
non-SCID* ns76.4 243/404 370/1191 104/362
non-SCID* ns76.5 84/429 169/1248 52/342
SCID s86.2 135/497 141/1096 31/294

*These seven tumour lines were used for further experiments.

q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387


384 I. M. Svane et al.
DISCUSSION
The major source of IFN-g in a normal host is the T-lymphocyte
population and recognition of antigen in the context of either
MHC class I or II by a T cell is the primary external stimulus
which initiates IFN-g production. One of the most important
effects of IFN-g is its ability to induce or increase MHC class I
and II expression on many cell types by up-regulating transcrip-
tion of MHC genes. The up-regulation of MHC expression,
together with the IFN-g-stimulated synthesis of several other
molecules involved in antigen processing and presentation, is
probably crucial during the inductive as well as the effector
Fig. 2. Expression of IFN-g receptor of tumour cell lines responding
with high (a), low (b) or no (c) up-regulation of MHC class I to IFN-g
stimulation. Tumour cells stained with an irrelevant, isotype-matched
primary antibody were used as control (in grey).
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
phase of T-cell responses to tumour antigens. This has been
demonstrated most convincingly in experiments where IFN-g
gene-transfected tumour cells exhibited decreased tumorigeni-
city and metastatic growth, and in experiments where vaccina-
tion with such cells protected mice against subsequent challenge
with untransfected wild-type tumour cells [44–46].
To find out whether insensitivity to IFN-g is a frequent
mechanism by which tumour cells escape the host immune
system, we tested the sensitivity to IFN-g of a large panel of
sarcomas induced in normal and T cell-deficient mice. We found
that the tumour lines were very heterogeneous with respect to the
level of MHC class I expression before as well as after IFN-g
stimulation. No common maximum expression value was
reached, as tumour cell lines expressing low basic levels of
MHC class I in general also displayed the lowest levels of MHC
after IFN-g stimulation, while tumour cell lines with high basic
levels of MHC attained even higher levels. We find it remarkable
that tumours identically induced in syngeneic individuals exhibit
this pronounced heterogeneity in IFN-g sensitivity and MHC
Fig. 3. Induction of tyrosine phosphorylation of Jak2 by IFN-g
stimulation. Cells were incubated with medium (control) or in medium
supplemented with IFN-g (1000 U/ml) for 10 min. Cells were lysed
and the cytosolic fractions were immunoprecipitated with anti-
phosphotyrosine antibody. The precipitated proteins were subjected to
SDS–PAGE and immunoblotted with anti-Jak2. (A) Lysates from a
tumour cell line with high IFN-g response; (B) lysates from a tumour
cell line with low IFN-g response; and (C) lysates from a tumour cell
line not responsive to IFN-g stimulation. Lanes 1, 3 and 5,
unstimulated tumour cells; lanes 2, 4 and 6, IFN-g-stimulated tumour
cells. Only parts of the immunoblots are shown, being representative.
Fig. 4. Induction of tyrosine phosphorylation of Stat1 by IFN-g
stimulation. Cells were incubated with medium (control) or in medium
supplemented with IFN-g (10 U/ml) for 48 h. Cells were lysed and the
cytosolic fractions were immunoprecipitated with anti-phosphotyrosine
antibody. The precipitated proteins were subjected to SDS–PAGE and
immunoblotted with anti-Stat1 antibody. (A) Lysates from a tumour
cell line with high IFN-g response; (B) lysates from a tumour cell line
with very low IFN-g response; and (C) lysates from a tumour cell line
not responsive to IFN-g stimulation. Lanes 1, 3 and 5, unstimulated
tumour cells; lanes 2, 4 and 6, IFN-g-stimulated tumour cells. Only
parts of the immunoblots are shown, being representative.

expression. It may reflect that MCA induces mutations randomly
in genes involved in IFN-g signalling and MHC expression
during the development of the individual tumours. Almost all
tumour lines were able to increase MHC class I expression
significantly in response to IFN-g, with a tendency towards
higher MHC expression after IFN-g stimulation of tumour
lines from immunocompetent mice compared to immunodefi-
cient mice. This observation indicates that resistance to IFN-g is
not a necessary trait for a primary sarcoma to develop in normal
or immunodeficient mice, and that selection of tumour cells
based on IFN-g sensitivity is not predominant.
Only three of the 78 tumour cell lines were completely
insensitive to IFN-g stimulation and another two tumour lines
exhibited a very low response. It is striking that these five tumour
lines all originated from tumours induced in immunodeficient
mice, while the six tumour cell lines expressing the highest levels
of MHC class I after IFN-g treatment all except for one
originated from immunocompetent mice, but an explanation to
this has not been found. It may be important to notice that nearly
all of the tumours from nude mice studied previously had severe
discrepancies between levels of mRNA for MHC class I and cell
surface proteins, indicating an abnormal gene regulation, while
the large majority of tumours from nu/þ mice had concordant
levels of mRNA and surface protein [47]. Examination of these
high- and low-responding tumour lines by flow cytometry
showed that they all expressed low but equal amounts of IFN-
receptor on the surface. This excludes diminished expression of
IFN-g receptor as a cause of the defective IFN-g response, but a
defective signal transmission through the receptor was still
possible. Jak2, a protein kinase activated by ligand binding to
the IFN-g receptor was found to be unphosphorylated in unsti-
mulated tumour cells and equally tyrosine-phosphorylated in
IFN-g-stimulated tumour cells, indicating a normal first-step
activation. An important subsequent step in the IFN-g signalling
pathway is tyrosine phosphorylation of Stat1, which is normally
present in an inactive unphosphorylated form. It has recently
been found that primary leukaemia cells resistant to IFN-a were
unable to activate Stat1, and that cells from Stat1 –/– mice did
not respond to IFN-g and IFN-a [48,49]. In the present study, all
tumour cell lines responding to IFN-g by MHC class I up-
regulation showed a strong increase in Stat1a and Stat1b
phosphorylation after IFN-g treatment. However, the three
tumour cell lines which completely failed to respond to IFN-g
with up-regulation of MHC were remarkable in that they exhib-
ited no increase in Stat1b phosphorylation and a lower increase
in Stat1a phosphorylation after IFN-g. This strongly indicates
that these three tumour lines from immunodeficient mice have a
defect in the signal transduction pathway affecting Stat1.
In addition to the different phosphorylation patterns, two
unexpected observations were made. Firstly, all tumour
lines expressed three different tyrosine-phosphorylated Stat1
molecules, Stat1a, Stat1b and an additional Stat1-related mole-
cule not previously described with a molecular weight of
approximately 78 kDa (p78). This p78 was also found to be
present in immortalized connective tissue cell lines such as 3T3
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Defective IFN-g Signalling in Sarcoma Cells 385
and L-cells. The tyrosine phosphorylation of p78 was unaffected
by IFN-g stimulation, while as already discussed both Stat1a and
Stat1b displayed strongly increased tyrosine phosphorylation
after IFN-g stimulation in IFN-g-responsive tumour lines.
Stat1a and Stat1b have been shown to be alternatively spliced
products of the same gene [38]. It is possible that p78 found in
our tumour lines is an additional splice variant of Stat1, perhaps
connected with transformation like the Jak3 molecule described
by Lai et al. which exhibits tissue preference [50]. The p78-band
could also appear due to a cross-reaction of the antibody with a
Stat protein not yet characterized [51]. No additional phosphor-
ylation of p78 was seen after IFN-g stimulation. It therefore
seems less likely that p78 should be active in the IFN-g signal-
ling pathway.
Secondly, it became clear that all tumour lines had a relatively
high background level of phosphorylated Stat1. This was the case
also for the fibroblast-related cell lines 3T3 and L cells. Con-
stitutively activated Jak-Stat pathways have mainly been
observed in a cutaneous T-cell lymphoma and in virus-trans-
formed T cells [52, 53]. Our tumours are sarcomas derived from
subcutanous connective tissue by in vivo treatment with MCA,
and the dysregulated Stat1 phosphorylation observed might be
connected with the transformation.
ACKNOWLEDGMENTS
These studies were supported by grants from the Danish Cancer
Society, the Danish Biotechnological Research Program,
Mogens og Jenny Vissings Fond, Wedell-Wedellsborgs Fond
and the Danish Medical Research Council. We are grateful for
the financial support of Handelsgartner Ove Villiam Buhl Olesen
& ægtefælle Edith Buhl Olesens Mindelegat, Direktør Jacob
Madsens & hustru Olga Madsens Fond, Dagmar Marshalls Fond,
Beckett Fonden, Kong Christian den Tiendes Fond, Direktør J.P.
Lund og hustru Wilhelmine Bugges Legat, Anna og Jakob
Jakobsens Legat, direktør Einar Willumsens Mindelegat, Wei-
manns Legat, Fru Astrid Thaysens Legat and Lykfeldts Legat.
We thank Mette Nordahl and Flemming Høg for their expert
technical assistance.
REFERENCES
1 Tanaka K, Yoshioka T, Biberich C, Jay G. Role of the major
histocompatibility complex class I antigens in tumor growth and
metastasis. Annu Rev Immunol 1988;6:359–80.
2 Wallich R, Bulbuc N, Hämmerling G, Katzav S, Segal S, Feldman
M. Abrogation of metastatic properties of tumour cells by de novo
expression of H-2K antigens following H-2 gene transfection. Nature
1985;315:301–5.
3 Plaksin D, Gelber C, Feldman M, Eisenbach L. Reversal of the
metastatic phenotype in Lewis lung carcinoma cells after transfec-
tion with syngeneic H-2 Kb gene. Proc Natl Acad Sci USA
1988;85:4463–7.
4 Austin FC, Finerty JF, Sogn JA. Cancer immunology: highlights of
the NCI extramural immunology program. Cancer Invest
1990;8:207–28.

386 I. M. Svane et al.
5 Elliott BE, Carlow DA, Rodricks A, Wade A. Perspectives on the
role of MHC antigens in normal and malignant cell development.
Adv Cancer Res 1989;53:181–245.
6 Garrido F, Ruiz-Cabello F. MHC expression on human tumors – its
relevance for local tumor growth and metastasis. Sem Cancer Biol
1991;2:3–10.
7 Schmidt W, Festenstein H. Resistance to cell mediated cytotoxicity
is correlated with reduction of H-2K gene products in AKR
leukemia. Immunogenetics 1982;16:257–67.
8 Garrido F, Cabrera T, Lopez-Nevot MA, Ruiz-Cabello F. HLA class
I antigens in human tumors. Adv Cancer Res 1996;67:155–95.
9 Hui KM, Kim BS. Tumor specific immunity induced by somatic
hybrids. Cell Immunol 1984;87:591–600.
10 Höglund P, Ljunggren HG, Kärre K, Jay G. Role of major histo-
compatibility complex class I molecules in tumor rejection. New
insights from studies with synthetic peptides and transgenic mice.
Immunol Res 1990;9:298–313.
11 Weber JS, Jay G, Tanaka K, Rosenberg SA. Immunotherapy of a
murine tumor with interleukin 2. Increased sensitivity after MHC
class I gene transfection. J Exp Med 1987;166:1716–33.
12 Christinck ER, Luscher MA, Barber BH, Williams DB. Peptide
binding to class I MHC on living cells and quantitation of complexes
required for CTL lysis. Nature 1991;352:67–70.
13 Wallny H, Deres K, Faath S et al. Identification and quantification of
a naturally presented peptide as recognized by cytotoxic T lympho-
cytes specific for an immunogenic tumor variant. Int Immunol
1992;4:1085–90.
14 Kageyama S, Tsomides TJ, Sykulev Y, Eisen HN. Variations in the
number of peptide MHC class I complexes required to activate
cytotoxic T cell responses. J Immunol 1995;154:567–76.
15 Browning M, Bodmer WF. MHC antigens and cancer: implications
for T-cell surveillance. Curr Opin Immunol 1992;4:613–18.
16 Cohen EP, Kim TS. Neoplastic cells that express low levels of MHC
class I determinants escape host immunity. Sem Cancer Biol
1994;5:419–28.
17 Klein G. Immune and non-immune control of neoplastic develop-
ment: Contrasting effects of host and tumour evolution. Cancer
1980;45:2486–524.
18 Lopez-Nevot MA, Esteban F, Ferron A et al. HLA class I gene
expression on human primary tumors and autologous metastases:
demonstration of selective losses of HLA antigens on colorectal,
gastric and laryngeal carcinomas. Br J Cancer 1989;59:221–6.
19 McDougall CJ, Ngoi SS, Goldman, I et al. Reduced expression of
HLA class I and class II antigens in colon cancer. Cancer Res
1990;50:8023–7.
20 Garrido F, Carara T, Concha A, Glew S, Ruiz-Cabello F, Stern PL.
Natural history of HLA expression during tumor development.
Immunol Today 1993;14:491–9.
21 Engel A, Svane IM, Mouritsen S, Rygaard J, Clausen J, Werdelin O.
Methylcholanthrene-induced sarcomas in nude mice have short
induction times and relatively low levels of surface MHC class I
expression. APMIS 1996;104:629–39.
22 Svane IM, Engel A, Nielsen M, Ljunggren H, Rygaard J, Werdelin
O. Chemically induced sarcomas from nude mice are more immu-
nogenic than similar sarcomas from congenic normal mice. Eur J
Immunol 1996;26:1844–50.
23 Bosma GC, Custer RP, Bosma MJ. A severe combined immuno-
deficiency mutation in the mouse. Nature 1983;301:527–30.
24 Engel A, Svane IM, Rygaard J, Werdelin O. MCA-sarcomas induced
in scid mice are more immunogenic than MCA sarcomas induced in
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
congenic, immunocompetent mice. Scand J Immunol 1997;45:463–
70.
25 Thomson A. The Cytokine Handbook. 2nd Edn. San Diego: Aca-
demic Press, 1994.
26 Knop J. Immunologic effects of interferon. J Invest Dermatol
1990;95:72S-4S.
27 Barth RJ, Mule JJ, Spiess P, Rosenberg SA. Interferon gamma and
tumor necrosis factor have a role in tumor regressions mediated by
murine CD8þ tumor infiltrating lymphocytes. J Exp Med
1991;173:647–58.
28 Kohonen-Corish MRJ, King NJC, Woodhams CE, Ramshaw IA.
Immunodeficient mice recover from infection with vaccinia virus
expressing interferon gamma. Eur J Immunol 1990;20:157–61.
29 Dalton DK, Pitts-Meek S, Keshav S, Figari IS, Bradley A, Stewart
TA. Multiple defects of immune cell function in mice with disrupted
interferon gamma genes. Science 1993;259:1739–42.
30 Restifo NP, Esquivel F, Kawakami Y et al. Identification of human
cancers deficient in antigen processing. J Exp Med 1993;177:265–
72.
31 Aki M, Shimbara N, Takashina M et al. Interferon gamma induces
different subunit organizations and functional diversity of
proteasomes. J Biochem 1994;115:257–69.
32 Klar D, Hämmerling G. Induction of assembly of MHC class I heavy
chains with beta-2 microglobulin by interferon gamma. EMBO J
1989;8:475–81.
33 Anderson SL, Shen T, Lou J et al. The endoplasmic reticular heat
shock protein gp96 is transcriptionally up-regulated in interferon-
treated cells. J Exp Med 1994;180:1565–9.
34 Farrar MA, Schreiber RD. The molecular cell biology of interferon
gamma and its receptor. Annu Rev Immunol 1993;11:571–611.
35 Silvennoinen O, Inle JN, Schlessinger J, Levy D. Interferon induced
nuclear signalling by Jak protein tyrosine kinases. Nature
1993;366:583–7.
36 Resnitzaky D, Tiefenbrun N, Berissi H, Kimchi A. Interferons and
interleukin 6 suppress phosphorylation of the retinoblastoma protein
in growth-sensitive hematopoietic cells. Proc Natl Acad Sci USA
1992;89:402–6.
37 Shuai K, Horvath CM, Huang LHT, Qureshi SA, Cowburn D,
Darnell JE. Interferon activation of the transcription factor stat 91
involves dimerization through SH2-phosphotyrosyl peptide
interactions. Cell 1994;76:821–8.
38 Pellegrini S, Schindler C. Early events in signalling by interferons.
TIBS 1993;18:338–42.
39 Bartlett GL. Effect of host immunity on the antigenic strength of
primary tumors. J Natl Cancer Inst 1972;49:493–504.
40 LeClaire RD, Basu M, Pinson DM et al. Characterization and use of
monoclonal and polyclonal antibodies against the mouse IFN gamma
receptor. J Leuc Biol 1992;51:507–16.
41 Nielsen M, Svejgaard A, Skov S, Ødum N. Interleukin 2 induced
tyrosine phosphorylation and nuclear translocation of stat 3 in human
T lymphocytes. Eur J Immunol 1994;24:3082–6.
42 Dustin ML, Rothlein R, Bhan AK, Dinarello CA, Springer TA.
Induction by IL 1 and interferon gamma: tissue distribution, bio-
chemistry and function of a natural adherence molecule (ICAM-1). J
Immunol 1986;137:245–54.
43 Briscoe J, Rogers NC, Witthuhn BA et al. Kinase-negative mutants
of JAK1 can sustain interferon gamma inducible gene expression but
not an antivrial state. EMBO J 1996;15:799–809.
44 Restifo NP, Spiess PJ, Karp SE, Mule JJ, Rosenberg SA. A
nonimmunogenic sarcoma transduced with the cDNA for interferon

gamma elicits CD8þ T cells against the wild-type tumor: Correlation
with antigen presentation capability. J Exp Med 1992;175:1423–31.
45 Roth C, Rochlitz C, Kourilsky P. Immune response against tumors.
Adv Immunol 1994;57:281–351.
46 Porgador A, Bannerji R, Watanabe Y, Feldman M, Gilboa E,
Eisenbach L. Antimetastatic vaccination of tumor-bearing mice
with two types of IFN gamma gene-inserted tumor cells. J
Immunol 1993;150:1458–70.
47 Engel A-M, Svane IM, Madsen MW, Pedersen M, Werdelin O.
Molecular aberrations in the MHC class I-restricted pathway for
antigen presentation in methylcholanthrene sarcomas from nude
mice: discrepancies between MHC mRNA and surface protein.
Clin Exp Immunol 1997; 109: (in press).
48 Xu B, Grander D, Sangfelt O, Einhorn S. Primary leukemia cells
resistant to alpha interferon in vitro are defective in the activation of
the DNA-binding factor interferon stimulated gene factor 3. Blood
1994;84:1942–9.
49 Durbin JE, Hackenmiller R, Simon MC, Levy DE. Targeted
q 1997 Blackwell Science Ltd, Scandinavian Journal of Immunology, 46, 379–387
13653083, 1997, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1046/j.1365-3083.1997.d01-141.x by Readcube-Labtiva, Wiley Online Library on [13/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Defective IFN-g Signalling in Sarcoma Cells 387
disruption of the mouse STAT1 gene results in compromised
innate immunity to viral disease. Cell 1996;84:443–50.
50 Lai KS, Jin Y, Graham DK, Witthuhn BA, Ihle JN, Liu ET. A kinase-
deficient splice variant of the human JAK3 is expressed in hemato-
poietic and epithelial cancer cells. J Biol Chem 1995;270:25028–36.
51 Boulton TG, Zhong Z, Wen Z, Darnell JE, Stahl N, Yancopoulos
GD. STAT3 activation by cytokines utilizing gp130 and related
tranducers involves a secondary modification requiring an H7-
sensitive kinase. Proc Natl Acad Sci USA 1995;92:6915–19.
52 Migone T, Lin J, Cereseto A et al. Constitutively activated Jak-
STAT pathway in T cells transformed with HTLV–1. Science
1995;269:79–81.
53 Nielsen M, Kaltoft K, Nordahl M, Röpke C, Geisler C, Mustelin T,
Dobson P, Svejgaard A, Ødum N. Constitutive activation of a slowly
migrating isoform of Stat3 in mucosis fungoides: Tyrphostin AG490
inhibits Stat3 activation and growth of mucosis fungoides tumor cell
lines. Proc Natl Acad Sci USA 1997;94:6767–9.