Identification of Human Cancers Deficient in
Antigen Processing
By Nicholas P. Restifo,* FernandoEsquivel,$YutakaKawakami,*
Jonathan W. Yewde11,~JamesJ. Mul~*,StevenA. Rosenberg,*
andJack R. Bennink~
From the "SurgeryBranch, Division of Cancer Treatment,National CancerInstitute;and the
ILaboratoryof Viral Diseases,National Institute of Allergy and InfectiousDiseases,National
Institutes of Health, Bethesda,Maryland 20892
Summary
Intracellular antigens must be processed before presentation to CD8 + T cells by major
histocompatibility complex (MHC) class I molecules. Using a recombinant vaccinia virus (Vac)
to transiently express the Kd molecule, we studied the antigen processing efficiency of 26
different human tumor lines. Three cell lines, all human small cell lung carcinoma, consistently
failed to process endogenously synthesized proteins for presentation to Kd-restricted, Vac-specific
T cells. Pulse-chase experiments showed that MHC class I molecules were not transported by
these cell lines from the endoplasmic reticulum to the cell surface. This finding suggested that
peptides were not available for binding to nascent MHC molecules in the endoplasmic reticulum.
Northern blot analysis of these cells revealed low to nondetectable levels of mRNAs for MHC-encoded proteasome components LMP-7 and LMP-2, as well as the putative peptide transporters
TAP-1 and TAP-2. Treatment of cells with interferon 3/enhanced expression of these mRNAs
and reversed the observed functional and biochemical deficits. Our findings suggest that
downregulation of antigen processing may be one of the strategies used by tumors to escape
immune surveillance. Potential therapeutic applications of these findings include enhancing antigen
processing at the level of the transcription of MHC-encoded proteasome and transporter genes.
I
t has long been known that elements of the cellular immune system are capable of specifically recognizing and
destroying tumor cells (1, 2). In part, this reflects the activity of CD8 + T cells (TcI~s+) (3-7), which recognize class
I molecules of the MHC-bearing peptides of 8-10 residues
derived from proteins located in the cytosol (8-11). There
are now numerous examples of both mouse and human
TcDs+ that specifically recognize tumor cells and have therapeutic activity after adoptive transfer, in some cases inducing
a complete remission (12-16).
Despite the potential for T cells to eradicate neoplasms,
it is obvious from the progressive growth of most cancers
that many tumors escape recognition by TcDs+. The reasons
for this are only partly understood. There is evidence that
some tumor cells express low levels of class I molecules in
vivo and in vitro (17). Poor class I expression by tumor cells
in mice and humans have generally been attributed to low
levels of class I c~chain gene transcription (18, 19). Attempts
have been made to enhance c~chain transcription by transfection or transduction of class I ol chains into tumors (20, 21),
or by use of DNA-hypomethylating or -alkylating agents (22).
These studies, directed at increasing the immunogenicity of
the tumors involved, have met with mixed success. Studies
by Weis and Seidman (23) reported that after the transfection of tumor cells with MHC class I genes there was no
increase in cell surface expression of class I molecules despite
a 20-60-fold increase in mRNA for the inserted class I genes.
Based on recent discoveries, it seems possible that this was
due to limiting amounts of antigenic peptides, which are required for proper assembly of class I c~ chains with 32"
microglobulin (B2-m)1in the endoplasmic reticulum, before
transport of the complex to the plasma membrane.
Antigenic peptides are believed to be generated from a cytosolic pool of proteins. Association of these peptides with class
I c~ chains and 32-m is thought to occur in an early secretory compartment (24-26). There is circumstantial evidence
that two MHC gene products, called LMP-7 and LMP-2 (formerly known as RING 10 and RING 12, respectively; see
reference 27), physically associated with a large nuclear and
cytosolic proteolytic structure (termed the proteasome),
somehow alter the function or location of the proteasome
to favor either the production of antigenic peptides or their
1Abbreviations used in this paper: B2-m, B2-microglobulin; CM, culture
media; HA, hemagglutinin; NP, nucleoprotein; SCLC, small cell lung
carcinomas.
265 The Journal of Experimental Medicine 9 Volume 177 February 1993 265-272
delivery to class I molecules (28-30). Two other gene products, now designated TAP-1 and TAP-2 (for transporter associated with antigen processing), are encoded in the region
of the MHC, and are members of the ABC transporter family
(31-36). TAP-1 (previously known as KING 4, Y3, and PSF-1)
and TAP-2 (previously known as KING 11, Y1, and PSF-2)
are clearly needed for cells to efficiently process antigen.
Whether these proteins directly transport peptide from the
cytosol, or act in another manner, remains to be established.
To examine the capacity of human tumor cell lines to process cytosolic antigens for Tcos+ recognition, we devised a
method for screening a large number of tumor cell isolates
that is independent of both the HLA type of the tumor and
the presence, or absence, of specific cellular proteins. This
method exploits the capacity of vaccinia virus (Vac) to infect
a wide variety of human tumor cells (37). Using a recombinant Vac encoding the mouse H-2 Ka class I molecule (Ka-Vac), we could test human tumor cell lines for presentation
of viral antigens to mouse Ka-restricted, Vac-specific TCD8+
populations and thus study antigen processing capabilities
of human tumor cells per se.
Materials and Methods
Tumor Cells. All cell lines used in these studies were generated
by A. F. Gazdar, J. D. Minna, and their coworkers (University of
Texas Southwestern, Dallas, TX), with the exception of SW480,
LS174T, HT-29, WiDr, MDA-231, MCF7, BT-20, Hs578T, SK-BR-3, and MDA-468, which were obtained from American Type
Culture Collection (Rockville, MD), and CY 6T, which was generated in our laboratory.
Viruses. The production of a Vac recombinant containing the
Kd gene (Ka-Vac)has been described (38). Note that this recombinant differs from a previously published Vac-Kd (39) at position
114. The gene we used represents the corrected version of the Kd
gene in plasmid pH33 in which the wild-type GLN at position
114 is replaced by a HIS. This residue is located in the floor of the
Kd Ag binding site, and greatly influences the ability of Kd to
present viral antigens. The pKCKdwt construct was provided by
Jaulin Kourilsky (Institut Pasteur, Paris, France). The A/Puerto
Rico/8/34 (PK8) influenza virus infectious stock was generated
using the allantoic cavity of embryonated hen eggs and virus concentration using chicken red cell agglutination. The Vac recombinants containing the PK8 nucleoprotein (NP) and hemagglutinin (HA) genes have been described (40).
Effector Cells. Polyclonal TcDs+ populations were generated
from 6-8-wk-old female BALB/c mice by intravenous injection of
5 x 106PFU of Vac or Vac-NP virus. After at least 2 wk, spleens
were removed, dispersed to single cell suspensions with a
homogenizer (Dounce), and stimulated in vitro with either Vacor PK8-infected BALB/c splenocytes at a ratio of 2:1. Cells were
then cultured in culture media (CM) consisting of Iscove'smodified
medium with 7.5% FCS (Biofluids, Kockville, MD) to generate
Tcos+. LAK cells were prepared as previously described (41).
Cytotoxicity Assays. Target cells were infectedwith 10 PFU/cell
of vacdnia virus expressing the Kdclass I molecule for 60-90 min,
incubated in CM at 37~ for 3 h, then labeled with NaSlCrO4
for 1 h. Target cells were then mixed with TcDs+. After 4 h of
incubation the amount of released SlCr was determined by 3'
counting. Percent specificlysiswas calculated asfollows: 100x [(experimental cpm - spontaneous cpm)/(maximal cpm - spontaneous cpm)].
266
FACS| Analysis andAntibodies. Cytofluorographic analysiswas
done using a FACScan 440@(Becton Dickinson & Co., Mountain
View, CA). Cultured tumor cell lines were harvested with 0.02%
EDTA, washed, then stained for 30 min with culture supernatant
containing mAb supernate from H28-E23 (anti-HA antigen).
Where designated, cellswere stained with the mAb TW2.3, which
recognizes an intracellular vaccinia-specific antigen (J. Cox et al.,
manuscript in preparation). To enable antibody access to the antigen, cells were fixed with paraformaldehydeand made permeable
by including 0.1% saponin (wt/vol) during all manipulations. In
all cases, cells with the appropriate isotype-matched control antibody were used. mAb binding to cells was followed by binding
with goat anti-mouse FITC-conjugated antibody (Boehringer
Mannheim Biochemicals, Indianapolis, IN).
Metabolic Radiolabeling Experiments. Pulse-chase experiments
were performed as previously described (42). Briefly, 6 x 106cul-
tured tumor cells were incubated in methionine-free DMEM for
30 min at 37~ then incubated with 100/zCi [3sS]Met (Amersham Corp., Arlington Heights, IL) for 10 min. Ice-cold PBS was
then immediately added to a portion of the cells. These cells represent the 0-min time point. Warmed medium containing 2 mg/ml
of unlabeled Met was added to the remaining cells, which were
then incubated at 37~ for the amount of time specified. Detergent extracts from radiolabeled cells were then normalized to contain equal amounts of acid-precipitable counts, and incubated with
protein A-Sepharose previously loaded with the mAb specified. One
half of each sample was digested overnight with 5 mU Endo H
(Boehringer Mannheim Biochemicals). The other halfof the sample
was mock digested. Sampleswere then analyzedby SDS-PAGEusing
a 12% polyacrylamide gel and the buffer system of Laemmli. Gels
were fixed, incubated with Amplify (Amersham Corp.), dried, and
exposed to preflashed Kodak XAR-5 x-ray film (Kodak, Rochester,
NY) for autoradiography.
Northern Blot Analyses. To generate specific probes for transporter and proteasome genes, total RNA was isolated by the guanidine isothiocyanate-cesium chloridecentrifugation method. Specific
probes were generated from RNA isolated from an EBV-transformed
B cell line, or from the 501 melanoma cell line (both cell lines were
established in our laboratory). First-strand eDNA was synthesized
from 10 mg total RNA with an oligo(dT) primer. 30-35 cycles
of PCK amplification was performed using the conditions: 94~
for 30 s, 60~ for 30 s, 72~ for 1 min followed by an extension
cycle of 10 rain at 72~ The specific primers used had sequences
as follows: GACAAGAGCCACAGGTATTTGG and TGATGA-
GAAGCACTGAGCGG for TAP-1 (formerly RING 4, Y3, or
PSF-1), TACCTGCTCATAAGGAGGGTGC and ATTGGGATA-
TGCAAAGGAGACG for TAP-2 (formerly KING 11, Y1, or
PSF-2), TCGCCTTCAAGTTCCAGCATC~ and CCAACCATC-
TTCCTTCATGTGG for LMP-7 (formerly RING 10), and TTG-
TGATGCa3TTCTGATTCCCGand CAGAGCAATAGCGTCTGT-
GG for LMP-2 (formerlyRING 12). PCK products were subjected
to electrophoresis in 1.5% agarose gels, the correct sizes of bands
were isolated, purified by glass powder methods (Genedean; BIO
101, LaJolla, CA), and used as probes for Northern hybridizations.
PCK product identity was confirmed by cutting with restriction
enzymes and comparing predicted fragment sizeswith those found
in our preparations. The/~-actin cDNA probe was purchased from
Clonetech (PaloAlto, CA). The probes were labeled by the random
priming method using random hexamers. For Northern blot, 10
#g of total KNA was subjected to electrophoresis in a 1% agarose
formaldehyde gel and transferred to a nylon membrane (Zeta-Probe;
Bio-Rad Laboratories, Richmond, CA). Hybridization was done
in a 40% formamide hybridization solution (Northern hybridization buffer; 5 Prime ~ 3 Prime, Inc., West Chester, PA) at 42~
Identification of Human Cancers Deficient in Antigen Processing
i
70
(D
60
_r
~50
nO 40
0
~30
q~
~ 20
o9
10
A
i(c)%
21B)
~x~
3(C)
~k~\
~C)~
30103 1
7B 12(A}lk~CJ C.]~
548)
t 13(Q)ek\ ,,
,,,?,M,\ \
"~/ ~ 118(N)I~\ 9
154L) !I'(O)~N~"~
~S)Ik \\~
)llB)~
T2-- - 9 : 1Ol;
3010 3 1 3010 3 1
Effector: Target (Ratio)
C..I
J21(S)L
,22(S)
"%,'~
i
I24|S)1~.,..~.N,b
,,254s)~
26(S~0103 1
Figure 1. Screening of human tumor lines for antigen
processing defects. Cell lines studiedwere cultured human
tumor cells tested for their abilityto present Vacantigens
to Kd-restricted, Vac-specific TCDS+. Numbers correspond to the following tumor lines and histologies: C,
colon; B, breast; A, adenocarcinoma of the lung; Q, squamous cell carcinoma of the lung; L, large cell carcinoma
of the lung; N, neuroendocrine tumor of the lung; M,
mesothelioma; O, carcinoid of the lung; and S, SCLC
of the lung. Tumor designations are asfollows: I, SW480;
2, LS174% 3, HT-29; 4, WiDr, T2 (see reference 9); 5,
MDA-231; 6, MCF7; 7, BT-20; 8, Hs578T; 9, SK-BR-3;
10, CY 6"1~11,MDA-468; 12, H23; I3, H157; 14, H358;
I5, H1334; I6, Hl155; 17, H28; 18, H460; 19, H720;
20, Hmesol; 21, H187; 22, H510A; 23, N417; 24, H146;
25, H1092; 26, H82.
overnight. Membranes were then washed three to four times in
2x SSC at 60~ for 30 min, and autoradiography was then
performed.
PepticlePulsingExt~ments. The peptide used (NP 147-155 from
influenza A/Puerto Rico/8/34) was synthesized on a peptide synthesizer (Milligen/Biosearch, Burlington, MA) and HPLC purified
with confirmation of sequence by fast atom bombardment (M-Scan
Inc., West Chester, PA). 3 x 106 tumor cells in a volume of 1
ml of CM were pulsed with 1 #M of peptide during SlCr labeling
for 90 min at 37~ Cells were then washed three times and used
in a StCr release assay as described above.
Results
Screening the Antigen Processing Capabilities ofHuman Tumor
Lines. Fig. 1 depicts a composite of three experiments (A,
B, and C) in which 26 different tumor cell lines were infected for 4 h with Kd-Vac,and tested for lysis by Vac-specific
TCDS+. In each experiment, the T2 cell line was included
as a negative control. T2 cells lack a one-megabase region
of the MHC that contains the portion coding for the TAP
genes and proteasome component molecules, and are known
to be deficient in their ability to process viral antigens for
Tc,s+ recognition (43, 44). These experiments revealed that
tumor cell lines vary widely in their abilities to process and
present viral antigens to TCDS+. While some cell lines from
a variety of tumor histologies were lysed at high levels by
Vac-specific CTLs, a number of the lines were lysed at low
levels. These included tumor cell lines derived from several
histologies, including colon (WiDr and CY6T), breast (MDA
468), lung mesothelioma (Hmesol), and most consistently,
lung cancers of the small cell histology. Notably, three of
these small cell lung carcinomas (SCLC), H82, H146, and
H1092, were consistently (six/six experiments) recognized
at levels similar to, or even lower than, T2 cells (Fig. 1 Cli O.
The failure of Vac-specific TCDS+ tO lyse the various
tumor cells cannot be attributed to low levels of expression
of Vac gene products. Cytofluorography after indirect im-
munofluorescencestaining was used to assesswhether tumor
lines were infected with Vac and productive of viral protein.
As Kd on the cell surface was potentially dependent on the
function of antigen processing machinery, the surfaceexpression of a marker gene, the PR8 HA glycoprotein (Vac-HA),
267 Restifo et al.
was studied. The HA molecule was transported by the infected cell to the cell surfacewhere its expression was assessed
using an HA-specific mAb. Representativeresults are depicted
in Fig. 2. Tumor lines poorly recognized by TcDs+ expressed
similar or more viral antigens than those well recognized by
TcDs+, and no obvious correlation between viral gene expression and degree of TcDs+ recognition was observed. Indeed, by this analysis, SCLC generally expressed high levels
of viral antigens relative to the other tumor cell lines studied.
Evidence that SCLC were well infected by Vac was corroborated by a second approach (not shown). In several SlCr release assays, a portion of target cells were assayed for binding
50
40
Io~ 30
03
o_
u_ 20
0
u.J
O..
rJ)
~ 10
Figure 2.
Hl155
H28
H460
H187
H-MESO-1 9
H146
H82 H1O92
5~
~ :r
loo 1;o = 2o0 2so
SURFACE EXPRESSION OF HA
(MCN)
Representative experiment showing that expression of viral
antigen does not correlate with lysabilityof lung cancer lines. In this experiment, cells were infected with 10 PFU/cell of the Vac-HA virus or
with Vac-Ka virus. Vac-HA-infected cells were stained for surface expression of the HA molecule and the mean channel numbers (MCN) by FACS9
analysis are plotted on the x-axis. Isotype-matched control Ab staining
was always below a MCN of 10 (dottedline). Plotted on the y-axis is the
percent specificSlCr released (E/T 30:1) by cells infected by Vac-Kd. Note
that the SCLC cell lines H82, H146, and H1092 were productive of the
viral gene product HA, indicating that these cells are infectable by Vac,
but that they were not lysed by anti-Vac TcDs+ cells.
Figure 3. Assemblyand transport ofMHC classI moleculesin the H82 SCLCcellline. (A) Experiment to study
synthesis and transport of Kd molecules. Detergent extracts from pS]Met-labeled, Ka-Vac-infectedcells were
immunoprecipitated with the HB-159 Ka-specificmAb,
and left untreated, or digested with Endo H as indicated
before analysisby SDS-PAGE. Cells were chased for 0,
40, or 80 min after pulse labeling to monitor class I exocytosis. (B) Experiment to examine synthesis and transport of Ka moleculesafter SCLC line was pretreated with
IFN-% Experiment was performed as in A but cellswere
pretreated with 1,000 U/ml of rhlFN-y for 48 h before
assay.(C) Experiment to study the behavior of nativeclass
I molecules. Experiment was as in B, except cells were
uninfected and were immunoprecipitated with W6/32.
(D) Experiment to test assembly of native class I molecules,with andwithout pretreatment with IFN-3,. Uninfected cells radiolabeled as in A and chased for 0 (pulse)
or 120 rain (chase) were immunoprecipitated with the
W6/32 mAb, specificfor c~chains complexedwith B2-m
or the B2-m-specificmAb L368 (ATCC designation HB
149). Where designated, cellswere pretreated with 1,000
U/ml of rhlFN-'y for 48 h before assay.
of mAb TW 2.3, which recognizes an intracellular vaccinia
virus protein (Yewdell, J. W., unpublished observation). All
SCLC lines were found to be well infected with Vac relative
to controls. Additionally, all of the SCLC lines were efficiently
lysed relative to other cell lines by LAK cells, as determined
by SlCr release assay (not shown). This demonstrates that
the failure of virus-specific TcDs+ to lyse SCLC is not due
to a general resistance of the cells to lysis by immune cells,
or their failure to release SlCr after a lethal hit has been received. Thus, the failure of TcDs+ to lyse Vac-infected SCLC
appeared to be due to a genuine defect in antigen presentation, and not other factors.
Assembly and Transportof Class I Moleculesin SCLC Cell
Lines. To further characterize the antigen presentation deficit
in SCLC, pulse-chase methodology was utilized to study the
biosynthesis and intracellular transport of class I molecules.
Class I molecules were immunoprecipitated from detergent
extracts from cells pulse radiolabeled with [3sS]Met for 10
min, and chased for up to 80 min. In Fig. 3 A, detergent
extracts from Kd-Vac-infectedcells were immunoprecipitated
with the anti-Ka-specific mAb SF1-1.1.1 (ATCC designation
HB 159), which is directed against the ol chain of the Ka
molecule. One half of each immunoprecipitate was then
digested with endo-~-N-acetylglucoseaminidaseH (Endo H),
which cleaves N-linked oligosaccharides in the simple, high-mannose forms that exist before the modifications associated
with transport of class I o~chains through the later portions
of the Golgi complex. Samples were then analyzed by SDS-PAGE for the characteristic increase in ot chain mobility observed after removal of N-linked oligosaccharides. As seen
in Fig. 3 A, Ka molecules remained sensitive to digestion
with Endo H throughout the 80-min chase period.
NorthernBlotAnalysisof TAP andLMP. Retention of class
I molecules in an Endo H-sensitive form has been observed
in cells that fail to express TAP or MHC-proteasome subunit
gene products (45-47). It was therefore of interest to measure the expression of these genes in SCLC. Lacking suitable
antibody reagents to directly measure the expression of these
gene products, we performed Northern blots prepared from
total mlkNA that were then probed for the expression of
the genes indicated. As seen in Fig. 4, mRNA for each of
the four genes was detected in a cell line with normal antigen presentation capacity, but not in T2 cells, whose deletion encompasses all four genes, mlkNA was also not detected in each of the SCLC cell lines deficient in antigen
presentation. Low or absent steady-state levels of mlkNA indicates either that transcription is downregulated or that there
was a shortening of the half-life of these messages. Lack of
detection of message by Northern blot analysis, however,
Figure 4. Northern blot analysesof TAP and LMP genes. Lanescor-
respond to various tumor cells as follows: (a) Hl155 (positive control),
(b) Hl155 treated with 1,000U IFN-3~for 48 h, (c)T2 (negative control),
(d) T2 with IFN-3,, (e) H82, ~ H82 with IFN-% (g) H146, (h) H146
with IFN-% 0) H1092, and (1~H1092with IFN-"/. The sameblots stripped
and reprobed with B-actin are shown below. Note that mRNA levelsfor
TAP and proteasome components were not detected in SCLC lines unless
these lines were pretreated with IFN-%
268 Identification of Human Cancers Deficient in Antigen Processing
dearly does not indicate a deletion of the genes in these cells,
since treatment of ceils with IFN-3~for 48 h induced the expression oflarge quantities ofmRNA encodingthe four genes.
Note also that IFN-3, increased mRNA levelsin control cells
(Hl155 is shown), but not in T2 cells. These findings are
consistent with prior reports that expression of these genes,
like class I ol chain genes, are IFN-~/inducible (34, 48).
EffectoflFN-~"onClassI Asserablyand Transport. In addition to enhancing mRNA levels of LMP and TAP genes,
IFN-7 treatment of SCLC cells greatly altered the intracd-
lular trafficking ofdass I molecules. As seenin Fig. 3 B, IFN-~/
induced the transport of a substantial portion of Kd molecules from the early portion of the secretory pathway through
the Golgi complex, since approximately half of the class I
molecules immunoprecipitatedbecame resistant to Endo H
digestion within 80 min of their synthesis (Fig. 3, C and
D). In the absence of IFN-% SCLC cells do not synthesize
dass I molecules reactivewith the W6/32 mAb, which recognizes only tx chains associated with Bz-m (Fig. 3 D). Consistent with this finding, a mAb specific for 152-mprecipi-
tates only B2-m without any complexed cz chains. Thus,
these cells either fail to assemble class I c~chains, fail to produce a chains (18), or both. After treatment of cells with
IFN-% a large amount of assembled moleculesreactive with
W6/32 or anti-Bz-m mAbs are detected (Fig. 3 D), and
these molecules rapidly (tt/2< 20 min) acquire Endo H resistance (Fig. 3 C). The near absence of radiolabded/32-m
at time 0 in class I o~chain immunoprecipitatedby W6/32
(Fig. 3 D, lane 3) was consistently observed. This may be
due to the binding of nascent, radiolabeled a chains to a relatively large pool of unlabeled B2-m already present in the
endoplasmic reticulum. Thus, the vastmajorityof the W6/32-reactive class I appears to be bound to nonradiolabeled B2-m
at time 0, although the pool of radiolabeled Bz-m is ulti-
matelybound by W6/32-reactive a chains after 120 min (Fig.
3, D chase). The radiolabeled ~z-m in Fig. 3, lane 4,
coprecipitates radiolabled class Iot chain, which is apparently
not recognized by the W6/32 mAb, and thus may not yet
be fully folded at time 0.
IFN-y EnhancesPresentationofEndogenouslyProducedAn-
tigen. We next tested the effect of IFN-'), on the capacity
of SCLC to present viral antigens to Ka-restricted, Vac-specific TcDs+. As seen in Fig. 5, IFN-7 greatly enhanced
the antigen-presenting capacity of the SCLC, but not T2,
cells. In this experiment, the Ka restriction of TcDs+ recognition was demonstrated by the failure of TcDs+ to lyse
IFN-3,-treated cellsinfected with a Vacrecombinant that does
not express Kd. Thus, IFN-3' did not act by making the cells
more susceptible to lysis by NK cells that might be present
in the secondary splenic effector populations.
In addition to the interaction between TCR and MHC
class I-peptide complexes, interactions between other mole-
culeson T celland target cellsurfacescan contributeto delivery
of the lytic signal (49). While such interactions might be
expected to be minimized in the interaction of mouse TcDs+
with human target cells, it remained possible that the effect
ofIFN-3, on SCLC antigen presentation reflectedthe enhanced
expressionof accessory adhesion molecules. We therefore ex-269 Restifo et al.
H82 H146 H1092 CONTROLS
=>5O
"~40
n"
330
.o_ 20
N,--
o
r
, L t , i ' li0 3
i i i i i i I II0 1
10030 10 3 10030 10030 10 3 10030 3
Effector: Target (Ratio)
Figure 5. IFN-3"enables SCLC calls to present endogenous antigen.
Vac-Kd-infectedSCLC lines H82, H146, and H1092 ( 9 were not lysed
by Ka-restricted, anti-VacTc,s +, although a control cell line was lysed
(representedby Hl155 in this experiment [controls]).Treatment with 1,000
U IFN-7 for 48 h before infection enabled SCLC calls (A), but not T2
cells (control) to present viral antigens. Cell lines pretreated with IFN-7
but infected with a control Vac not expressing Kd (A) were not killed,
controllingfor nonspecifickillinginduced by IFN-'y pretreatment alone.
(Control) Hl155 without IFN-3, pretreatment after infection with wild-type Vac is shown ((>).
amined the effect of IFN-~/on the presentation of an exogenously added synthetic peptide to Ka-restricted TcDs+
specific for the peptide. As seen in a representative experiment using H82 cells, IFN-3/had no discernible effect on
peptide presentation (Fig. 6, left), despite the fact that it was
clearly able to enhance presentation of Vac antigens to Vac-specific Tcos+ (Fig. 6, right).
While the ability of non-IFN-3,-pretreated, Vac-Ka-infected cells to present exogenouslyprovided peptide seems paradoxical since Ka is inef~ciently transported to the cell surface under these circumstances, a similar phenomenon is
observed in other antigen processing-deficientcells. This observation has been attributed to surfaceexpression of "empty"
IN VlVO:
IN VITRO
6O
so
~ 40
~ 30
N~
"~ 20
O.
03 10
-5
VAC-NP VAC-NP
FLU-PR8 VAC-WT
I i i i
402010 5
9 H82. + PEPT
9 1,H82 + PEPT
o H82
o 3,H82
402010 5
Effector: Target (Ratio)
Figure 6. IFN-y pretreatment
does not enhance presentation of
exogenously provided Peptide. In
this experiment, H82 cells(_+ exogenously pulsed peptide) were
tested for lysisby TcDs+ specific
for influenza virus NP Peptide
(left, effectorcellsgenerated by in
vivo primingwith Vac-NPandin
vitro secondary stimulationwith
FLU-PRS) or Vacantigens (right,
effector cellsgenerated by in vivo
primingwith Vac-NPand invitro
secondary stimulationwith Vac-WT). (+ pept) Cells were pulsed
with a synthetic Peptide corre-
spondingto NP residues147-155
that represent the bonafidedeter-
minant recognized by Ka-restricted, NP-specific TCDS+. (7) Cells were
pretreated with IFN-3~for 48 h before the assay.Similar results were obtained in a repeat assay.The same experiment was performed twice in
experiments that included other SCI.C cell lines (H146and H1092) with
similar results.
class I molecules (i.e., devoid of natural peptides) (50), the
idea being that the transport of such molecules, while too
low to detect by FACS| staining or immunoprecipitation,
is sufficient to present the 200 or so peptides required for
a TcDs+ to lyse a target cell (51).
Discussion
Class I molecules are poorly expressed on many different
types of human tumor cells in vivo (17). Our findings indicate that this poor expression can be accompanied by a
markedly decreased capacity to present endogenously synthesized proteins, and for SCLC at least, greatly diminished
expression of four MHC genes that encode proteins thought
to be necessary for efficient antigen processing. Although it
is generallyconsidered that most cell types are capable of constitutively processing and presenting antigens to TcDs+, this
has been directly established only with cells of immune lineage. The present studies do not address the question of
whether poor antigen processing is associated with transformation or outgrowth of the tumor, or whether this poor
processing is representative of the natural regulation of class
I expression in the tissue of origin. This question may be
particularly difficult to answer with SCLC since the identity
of the precise cell of origin is not known. Whatever the case
may be, the fact remains that SCLC cell lines processed antigen poorly in the studies reported here. The therapeutic
implications of these findings are that poor class I expression
and poor antigen processing capability must be meliorated
if SCLC and other possible tumor histologies sharing these
characteristics are to be made susceptible to TcDs immunotherapy. It is unlikely that the poor antigen processing
capacity exhibited by SCLC cells is an artifact resulting from
cell culture, since immunohistochemical studies show that
SCLC expresses very low or undetectable levels of class I in
vivo (52).
The great enhancement of antigen processing upon IFN-3,
treatment of SCLC suggests that TCDS+ recognition of
tumor cells could be enhanced by the specific upregulation
of the antigen processing machinery in tumor cells. Such
TcDs+ might recognize tumor-specific antigens, or perhaps
tissue-specific antigens in caseswhere tumors arise from tissues
that are deficientin antigen processing. We recentlydescribed
a mouse tumor (MCA 101)whose antigen processing capacity
was similarly enhanced by IFN-y treatment (53). Expression
of IFN-3' in this tumor from a transfected gene resulted in
the autocrine enhancement of antigen processing, and allowed
the cells to induce a tumor-specific Tcns+ response. Most
importantly, TcDs+ cells grown out of IFN-3,-transduced
tumors were therapeutic against nonmodified tumor cells (54).
Thus, it might be possible to obtain a beneficial TcDs+ response by enhancing antigen processing machineryin a subset
of tumor cells in vivo, or by injection of tumor cells that
have been gene modified to enhance antigen processing in
vitro. Such strategies might also be appropriate for infectious
diseases in cases where the fore!gn organism infects cells with
low antigen processing capability.
Genetic or pharmaceutical therapies directedat the enhancement of antigen processing may exploit the beneficial effects
of IFN-7. However, since IFN-7 has been shown to be antiproliferative to T and NK cells, other ways of enhancing
antigen processing and presentation might prove to be more
therapeutically useful. It seems likely that the genes involved
in antigen processing and presentation share common regulatory elements. When these regulatory elements are elucidated,
new therapies could be developed for specific upregulation
of antigen processing in cancer and infectious disease. Conversely, downregulation of these gene products may prove
useful in tissue transplantation or in autoimmune disease.
In addition to SCLC, we found in recent studies that some
tumor cell lines of other histologies were unable to present
viral antigens to TcDs+. In a number of lines studied the
defects have turned out to be nonfunctional 32-m (Restifo,
N. P., manuscript in preparation). The underlying deficits
in other cell lines being studied remain to be established. The
robust nature of T2 cells (and other antigen processing-deficient mutants) suggests that none of the antigen processing
machinery is required for cell viability. In this event, it is
to be expected that some tumor cells will possess mutations
or deletions resulting in a functional deficiency in one or more
of the dedicated components of antigen processing. Characterization of naturally occurring antigen processing mutants
may help to identify novel gene products that function in
the efficient processing of endogenously synthesizedproteins.
We thank S. Topalian,J. Weber, and P. Hwu for helpful discussions; H. Oie, A. Gazdar, and J. Minna
for celllines;J. Cox for the TW 2.3 hybridoma; and C. Delgado andJ. Stephensfor technicalassistance.
Address correspondence to Nicholas Restifo, Surgery Branch, Division of Cancer Treatment, National
Cancer Institute, Building 10, Room 2B42, National Institutes of Health, Bethesda, MD 20892.
Receivedfor publication 19 August 1992 and in revisedform 26 October 1992.
Noteaddedinproof:Tworecent reports indicatethat theLMP-2 andLMP-7 gene products are not necessary
for the presentation of someantigens (55, 56). The function of these gene products in antigen processing,
therefore, remains to be established.
270 Identification
of Human CancersDeficientin AntigenProcessing
1. Klein, G., H.O. Sj6gren,E. Klein, and K.E. Hellstr6m. 1960.
Demonstration of resistance against methylcholanthrene-induced sarcomasin the primary autochthonous host. Cancer
Res. 20:1561.
2. Prehn, R.T., and J.M. Main. 1957. Immunity to methylcholanthrene-induced sarcomas,f Natl. CancerInst. 18:769.
3. Shimizu, K., and F.-W. Shen. 1979. Role of different T cell
set in the rejection of syngeneicchemicallyinduced tumors.
f lmmunol. 122:1162.
4. Greenberg, P.D., M.A. Cheever,and A. Fefer.1981.H-2 restriction of adoptive immunotherapy of advancedtumors,f
lmraunol. 126:2100.
5. Rosenberg, S.A., P. Spiess,and R. Laffeniere.1986. A new
approachto the adoptiveimmunotherapyofcancerwith tumor-infiltrating lymphocytes. Science(Wash. DC). 233:1318.
6. Itoh, K., C.D. Platsoucas,and C.M. Balch. 1988.Autologous
tumor-specific cytotoxic T lymphocytes in the infiltrate of
human metastaticmelanomas.Activationbyinterleukin 2 and
autologoustumor cells,and involvementofthe T cellreceptor.
J. Extx Med. 168:1419.
7. Schreiber,H., P.L.Ward, D.A. Rowley,and H.J. Stauss.1988.
Unique tumor-specificantigens. Annu. Rev. Imraunol.6:465.
8. Bjorkman, p.J.,M.A. Saper,B. Samraoui,W.S.Bennett, J.L.
Strominger,andD.C. W'rley.1987.The foreignantigenbinding
siteand T cellrecognition regionsofclassI histocompatibility
antigens. Nature (Lond.). 329:512.
9. Townsend, A., and H. Bodmer. 1989. Antigen recognition
byclassI-restrictedT lymphocytes.Annu. R~. Immunol.7:601.
10. Van Bleek, G.M., and S.G.Nathenson. 1990. Isolation of an
endogenouslyprocessedimmunodominant viralpeptide from
the class I H-2K~molecule. Nature (Lond.). 348:213.
11. R6tzschke, O., K. Falk, K. Deres, H. Schild, M. Norda, J.
Metzger, G. Jung, and H. Rammensee. 1990. Isolation and
analysisof naturally processedviralpeptides as recognizedby
cytotoxic T ceils. Nature (Lond.). 348:252.
12. Greenberg,P.D.1991.AdoptiveT celltherapyoftumors: mech-
anismsoperativein the recognition and elimination of tumor
cells. Adv. Immunol. 49:281.
13. Rosenberg,S.A., B.S.Packard,P.M. Acbersold,D. Dolomon,
S.L.Topalian,S.T.Toy,P. Simon, M.T.Lotze,J.C. Yang,C.A.
Seipp,et al. 1988. Use of tumor-infiltrating lymphocytesand
interleukin-2in the immunotherapyofpatientswith metastatic
melanoma. N, Engl.J. Med. 319:1676.
14. van der Bruggen, P., C. Traversari,P. Chomez, C. Lurquin,
E. De Plaen, B. Vanden Eynde,A. Kruth, and T. Boon. 1991.
A gene encoding an antigen recognized by cytolytic T lym-
phocyteson ahuman melanoma.Science(Wash.DC). 254:1643.
15. Van den Eynde, B., B. Leth6, A. Van pel, E. De Plaen, and
T. Boon. 1991.The gene coding for a major tumor rejection
antigen of tumor P815is identical to the normal gene of syngeneic DBA/2 mice.J. Ext~ Med. 173:1373.
16. Schreiber,H., P.L.Ward, D.A. Rowley,and H.J. Stauss.1988.
Unique tumor-specificantigens. Annu. Rev. lmmunol. 6:465.
17. ELliot,B.E., D.A. Carlow, A. Kodricks, and A. Wade. 1989.
Perspectiveson the roleofMHC antigensin normaland malignant cell development. Adv. CancerRes. 53:181.
18. Doyle,A., W.J. Martin, K. Funa, A. Gazdar,D. Carney,S.E.
Martin, I. Linnoila, F. Cuttitta, J. Mulshine, P. Bunn, and
J. Minna. 1985. Markedly decreased expression of class I
histocompatibility antigens, protein, and mRNA in human
small-ceLllung cancer.J. Exl~ Ailed. 161:1135.
19. Lassam,N., and G. Jay. 1989. Suppression of MHC class I
RNA in highlyoncogeniccellsoccursat the leveloftranscrip-
tion initiation. J. Iramunol. 143:3792.
20. Weber,J.S., G.Jay,K. Tanaka,and S.A. Rosenberg. 1987.Immunotherapy ofa murine tumor with interleukin 2: increased
sensitivity after MHC class I gene transfection.J. Exp. Med.
166:1716.
21. Tanaka,K., K.J. Isselbacher, G. Khoury, and G. Jay. 1985.
Reversalofoncogenesisby the expressionofa majorhistocom-
patibility complex class I gene. Science(Wash. DC). 228:26.
22. Carlow, D.A., R.S. Kerbel, and B.E. Elliot. 1989. Failureof
expressionofclassI major histocompatibilityantigens to alter
tumor immunogenicity of a spontaneous murine carcinoma.
J. Natl. CancerInst. 81:759.
23. Weis,J.H., andJ.G. Seidman. 1985. The expressionof major
histocompatibility antigens under metaUothioneingene promoter control. J. Immunol. 134:1999.
24. Yewdell,J.W., and J.R. Bennink. 1992. Cell biology of antigen processing and presentation to MHC class I molecule-restricted T lymphocytes. Annu. Rev, Immunol, 52:1.
25. Cox, J.H., J.W. Yewdell,L.C. Eisenlohr, P.R. Johnson, and
J.R. Bennink. 1990. Antigen presentation requires transport
of MHC class I moleculesfrom the endoplasmicreticulum.
Science(Wash. DC). 247:715.
26. Yewddl,J.W., andJ.R. Bennink. 1989.BrefeldinA specificaLly
inhibits presentation of protein antigens to cytotoxic T lymphocytes. Science(Wash. DC). 244:1072.
27. Bodmer,J.G., S.G.E.Marsh, E.D. Albert, W.F. Bodmer, B.
Dupont, H.A. Erlich, B. Mach, W.R. Mayr,P. Parham, T.
Sasazuki, et al. 1992. Nomenclature for factors of the HLA
system, 1991. TissueAntigens. 39:161.
28. Brown, M.C., J. Driscoll, andJ.J. Monaco. 1991.Structural
and serologicalsimilarityofMHC-linked LMPandproteasome
(multicatalyticproteinase)complexes.Nature(Lond.).353:355.
29. Glynne, R., S.H. Powis, S. Bech, A. Kelly,L.-A. Kerr, and
J. Trowsdah.1991.A proteasome-relatedgenebetweenthe two
ABC transporterlociin the classII regionofthe humanMHC.
Nature (Lond.). 353:357.
30. Goldberg,A.L., andK.L. Rock. 1992.Proteolysis,proteasomes
and antigen presentation. Nature (Lond.). 357:375.
31. Deverson,E.V.,I.R. Gow,W.J. Coadwell,J.J. Monaco,G.W.
Butcher,andJ.C. Howard. 1990.MHC classII regionencoding
proteins related to the multidrug resistancefamily of transmembrane transporters. Nature (Lond.). 348:738.
32. Monaco,J.J., S. Cho, and M. Attaya. 1990.Transportprotein
genes in the murine MHC: possibleimplications for antigen
processing. Nature (Lond.). 250:1723.
33. Spies,T., M. Bresnahan, S. Bahrain, D. Arnold, G. Blanck,
E. Mellins, D. Pious, and R. DeMars. 1990. A gene in the
human major histocompatibilitycomplexclassII region con-
trollingthe classI antigenpresentationpathway.Nature(Lond.).
348:744.
34. Trowsdale,J., I. Hanson, I. Mockridge, S. Beck, A. Townsend, and A. KeLly.1990.Sequenceencodedin the classII region of the MHC related to the 'ABC' superfamilyof transporters. Nature (Lond.). 348:741.
35. Bahram, S., D. Arnold, M. Bresnahan,J.L. Strominger, and
T. Spies. 1991.Two putative subunits of a peptide pump encoded in the human major histocompatibility complex class
II region. ProaNatl. Acad. Sci. USA. 88:10094.
36. Powis,S.J., E.V.Deverson, W.J. Coadwell, A. Ciruela, N.S.
271 Restifo
et al.
Huskisson, H. Smith, G.W. Butcher, andJ.C. Howard. 1992.
Effect of polymorphism of an MHC-linked transporter on the
peptides assembledin aclassI molecule. Nature (Lond.).357:211.
37. Moss, B. 1991. Vacciniavirus: a tool for research and vaccine
development. Science (Wash. DC). 252:1662.
38. Chakrabarti, S., K. Brechling, and B. Moss. 1985. Vaccinia
virus expression vector: coexpression of B galactosidase provides visual screening of recombinant virus plaques. Mol. Cell.
Biol. 5:3403.
39. Coupar, B.E.H., M.E. Andrew, D.B. Boyle, and R.V.Blanden.
1986. Immune responsesto H-2 Kd antigenexpressedbyrecom-
binant vaccinia virus. Proc.Natl. Acad. Sci. USA. 83:7879.
40. Smith, G.L., J.S. Levin, P. Palese, and B. Moss. 1987. Synthesis and cellular location of the ten influenza polypeptides
individually expressedby recombinant vacciniaviruses. Virology.
160:336.
41. Barth, R.J., S.N. Bock, J.J. Mul~, and S.A. Rosenberg. 1990.
Unique murine tumor-associated antigens identified by tumor
infiltrating lymphocytes. J. Immunol. 144:1531.
42. Yewdell,J.W., A. Yellen, and T. Bachi. 1988. Monoclonal antibodies localize events in the folding, assembly, and intracellular transport of the influenza virus hemagglutinin glycoprotein. Cell. 52:843.
43. Hosken, N.A., and M.J. Bevan. 1990. Defective presentation
of endogenous antigen by a cell line expressing class I molecules. Science (Wash. DC). 248:367.
44. Cerundolo, B.,J. Alexander, K. Anderson, C. Lamb, P. Cresswell, A. McMichael, F. Gotch, and A. Townsend. 1990. Presentation of viral antigen is controlled by a gene in the major
histocompatibility complex. Nature (Lond.). 345:499.
45. Townsend, A., T. Elliott, V. Cerundolo, L. Foster, B. Barber,
and A. Tse. 1990. Assembly of MHC class I molecules analyzed in vitro. Cell. 62:285.
46. Townsend, A., C. Ohlen, J. Bastin, H.G. Ljunggren, L. Foster,
and K. K~rre. 1989. Association ofclass I major histocompatibility heavy and light chains induced by viral peptides. Nature
(Lond.). 340:443.
47. Elliot, T., V. Cerundolo, J. Elvin, and A. Townsend. 1991.
Peptide-induced conformational change of the class I heavy
chain. Nature (Lond.). 351:402.
48. Ortiz-Navarrete, V., A. Seelig, M. Gernold, S. Frentzel, P.M.
Kloetzel, and G.J. H~mmerling. 1991. Subunit of the '20S'
proteasome (multicatalytic proteinase) encoded by the major
histocompatibility complex. Nature (Lond.). 353:662.
49. Springer, T.A. 1990.Adhesion receptors of the immune system.
Nature (Lond.). 346:425.
50. Ljunggren, H., M.J. Stare, C. 6hlen, J.J. Nee~es, P. H6glund,
M.-T. Heemels, J. Bastin, T.N.M. Schumacher, A. Townsend,
K. K~irre,and H.L. Ploegh. 1990. Empty MHC class I molecules come out in the cold. Nature (Lond.). 346:476.
51. Christinik, E.R., M.A. Luscher, B.H. Barber, and D.B. Williams. 1991. Peptide binding to class I MHC on living cells
and quantitation of complexes required for CTL lysis. Science
(Wash. DC). 352:67.
52. Funa, K., A.F. Gazdar, J.D. Minna, and R.I. Linnoila. 1986.
Paucity of B2-microglobulin expression on small cell lung
cancer, bronchial carcinoids, and certain other neuroendocrine
tumors. Lab Invest. 55:186.
53. Restifo, N.P., F. Esquivel, A.L. Asher, H. StStter, R.J. Barth,
J.R. Bennink, J.J. Mul~, J.W. Yewdell, and S.A. Rosenberg.
1991. Defective presentation of endogenous antigens by a murine sarcoma: implications for the failure of an anti-tumor immune response. J. Immunol. 147:1453.
54. Restifo, N.P., P.J. Spiess,S.E.Karp,J.J. Mul~,and S.A. Rosenberg. 1992. A nonimmunogenic sarcoma treatment with the
cDNA for interferon 3~elicits CD8 + T cells against the wild-type tumor: correlation with antigen presentation capability.
J. Ext~ Med. 175:1423.
55. Arnold, D., J. Driscoll, M. Ardrolewicz, E. Hughes, P. Cresswell, and T. Spies. 1992. Proteasome subunits encoded in the
MHC are not generally required for the processing ofpeptides
bound by MHC class I molecules. Nature (Lond.). 360:171.
56. Momburg, F., V. Ortiz-Navarrete, J. Neefjes, E. Goulmy, Y.
van de Wal, H. Spits, S.J.Powis, G.W. Butcher, J.C. Howard,
P. Walden, and G.J. H~mmerling. 1992. Proteasome subunits
encoded by the major histocompatibility complex are not essential for antigen presentation. Nature (Lond.). 360:174.
272 Identification
of Human CancersDeficientin Antigen Processing

