| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Biochemistry and Molecular Biology, Mayo Clinic, Scottsdale, AZ 85259
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The recent description of MUC1 as a target for CTLs has raised interest in using this protein as a target for immunotherapy. It is expressed by most adenocarcinomas of the breast, lung, stomach, pancreas, colon, prostate, ovary, endometrium, and cervix, which makes MUC1 an attractive therapeutic target. In 1999, cancers that expressed MUC1 accounted for about 72% of new cases and for 66% of the deaths (10). However, expression of the underglycosylated MUC1 is not sufficient to stimulate CTL killing, as >90% of existing carcinomas express MUC1 and these tumors progress. Thus, there is a need for studies to devise effective presentation of MUC1 immunogens to stimulate immune cells to kill tumor cells. The mouse up to now has not been a suitable preclinical model for testing vaccines, as human MUC1 differs in sequence from mouse Muc1 and is a foreign Ag in the mouse. We have developed an inbred C57BL/6 mouse strain that expresses human MUC1 in a tissue-specific fashion, driven by its own promoter. These mice transgenic for a foreign protein develop B and T cell compartment tolerance and are refractory to immunization with the protein encoded by the transgene (11). This experimental model enables us to study the effect of endogenous expression of the MUC1 gene on the ability of mice to produce protective immune responses to tumors, and it represents an improved model system for evaluating the efficacy of anti-MUC1 vaccine formulations in vivo within the context of existing tolerance mechanisms.
Adenocarcinoma of the pancreas is currently the fourth leading cause of death in the United States (12). Metastatic pancreatic cancer is uniformly fatal because no effective chemotherapy is available (13). Most of the models for studying human pancreatic tumors have used either injectable tumor cell lines or xenografts of primary human pancreatic tumors placed directly into the pancreas of mice (14). In this study, we have characterized the development of a double transgenic mouse line that expresses human MUC1 as a self-molecule and spontaneously develops MUC1-expressing tumors of the pancreas. These spontaneous tumors arise naturally in an appropriate tissue background and in the context of a viable immune system. They develop more slowly than injected tumors, giving the host immune system time to respond. Animals experience similar physiological events to humans, which may result in the presence of MUC1 in serum and body fluids and may serve to prime CTLs. Loss of polarized tissue architecture in the tumors and high levels of expression of underglycosylated MUC1 result in circulating MUC1, which may be immunosuppressive (15). Likewise, Abs to the tumor MUC1 may be present, as has been found in humans (16, 17). This model allows us to study for the first time the effects of self-tolerance, immunity, and autoimmunity to MUC1 as the tumors develop spontaneously.
We have analyzed the expression of MUC1 in the bitransgenic mouse model (designated MET) as the tumor develops in the pancreas and characterized the biology of the tumors and the MUC1-specific native immune response that develops during tumor progression. In our model, as in humans, there is overexpression and underglycosylation of MUC1, which in turn elicits cytotoxic T cells against the TR protein core. These naturally occurring CTLs are class I restricted and can be stimulated to kill MUC1-expressing cancer cell lines in vitro. Although these CTLs do not eradicate the spontaneously occurring pancreatic tumors, they are effective in eradicating injectable tumors when adoptively transferred. Thus, the MET model appropriately mimics the human condition and can be used to characterize immunotherapy strategies that will be effective against spontaneous tumors.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MUC1 transgenic (MUC1.Tg) mice are bred with oncogene-expressing mice that spontaneously develop tumors of the pancreas (ET mice) and are designated as MET. MUC1.Tg mice were developed in our laboratory (11), and the ET mice were obtained from Dr. Judith Tevethia (18). ET mice express the first 127 aa of SV40 large T Ag under the control of the rat elastase promoter. Fifty percent of the animals develop life-threatening pancreatic tumors by about 21 wk of age (18). All mice are on the C57BL/6 background. Animals were sacrificed and characterized at 3-wk intervals (n = 6 animals/time point) from 3 to 24 wk. Mice were bred and maintained in specific pathogen-free conditions in the S. C. Johnson Medical Research Building animal facility at Mayo Clinic Scottsdale. All experimental procedures were conducted according to the Institutional Animal Care and Use Committee guidelines.
PCR screening
PCR was used to routinely identify MUC1.Tg- and ET-positive mice in the colony. PCR was conducted as previously described (11, 18). The primer pairs for MUC1.Tg are 5'-CTTGCCAGCCATAGCACCAAG-3' (bp 745–765) and 5'-CTCCACGTCGTGGACATTGATG-3' (bp 1086–1065) and for ET are 5'-GCTCCTTTAACCCACCTG-3' (bp 4055–4072) and 5'-CCAACCTATGGAACTGATGAATG-3' (bp 4546–4568). The amplification program consisted of one cycle of 5 min at 95°C and 40 cycles of 30 s each at 95°C, 52°C, and 72°C followed by one cycle of 10 min at 72°C. The PCR product of each reaction was analyzed by size fractionation through a 1% agarose gel. Amplification conditions for MUC1 are the same except for the annealing temperature, which was 61°C. Amplification of MUC1 resulted in a 500-bp fragment, and amplification of ET resulted in a 491-bp fragment.
Tumor weights
The entire pancreas was dissected free of fat and lymph nodes, weighed, and spread on bibulus paper for photography. Nodules were counted under the dissecting scope. Pancreas was fixed in methacarn, processed for microscopy by conventional methods, step sectioned at 5 µm (about 10 sections per mouse pancreas), stained with hematoxylin and eosin, and examined by light microscopy. Mice were carefully observed for signs of ill-health, including lethargy, abdominal distention, failure to eat or drink, marked weight loss, pale feces, and hunched posture.
Cell lines
Cell lines used included: B16 murine melanoma cell line expressing MUC1 (B16.MUC1) and B16 transfected with vector only (Bl6.neo) (11). These cell lines were kindly provided by Dr. Tony Hollingsworth (University of Nebraska Medical Center, Omaha, NE). B16.MUC1 and B16.neo were maintained in DMEM with 10% FBS, penicillin (50 U/ml), and streptomycin (50 µg/ml), supplemented with 300 µg/ml G418. Cells were routinely tested by flow cytometry for the presence of MUC1.
ELISA
Cytokine levels in culture supernatant samples were determined
by specific ELISAs for IFN-
and IL-2. The IFN- assay used a
sandwich technique as described by Samuel (19). Abs used
were R46A2 as catcher and a second biotinylated Ab, XMG1.2, kindly
supplied by Biomira (Edmonton, Canada). The IFN- standard was
obtained from PharMingen (San Diego, CA). Cytokine levels in the test
sample were determined by comparison with reference standards. IL-2
levels were detected using Endogen ELISA kit (Woburn, MA). Serum MUC1
levels were determined using the Truquant BR RIA supplied by Biomira
(20). Detection of Ab to MUC1 was conducted by ELISA using
synthetic peptides (105 mer) of the 5.25 MUC1 TR as previously
described (11).
Immunohistochemistry
Tumors were obtained from MET mice at various time points during tumor progression, fixed in methacarn (60% methanol, 30% chloroform, 10% glacial acetic acid), embedded in paraffin, and sectioned for immunohistochemical analysis. MUC1 Abs used were CT1, a rabbit polyclonal Ab that recognizes mouse and human cytoplasmic tail region of MUC1 (21), HMFG-2, BC2, and SM-3, which have epitopes in the TR domain of MUC1. The TR epitope of HMFG-2 has been mapped to DTR, that of BC2 to APDTR, and that of SM-3 to PDTRP (22). All TR Abs are specific for human MUC1 and do not cross-react with mouse Muc1. All of the TR Abs are glycosylation sensitive in the pancreas. Abs to Fas ligand (FasL) and TGFß2 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Secondary Ab was swine-anti rabbit conjugated to HRP (Dako, Carpinteria, CA). Ab staining was blocked with the appropriate specific peptides supplied by Santa Cruz Biotechnology.
CTL assays
Determination of CTL activity was performed using a standard 51Cr release method after a 6-day in vitro peptide stimulation without additional added cytokines. Splenocytes from individual MET mice were harvested by passing through a nylon mesh followed by lysis of RBC using pharmlyse (0.45% ammonium chloride solution purchased from PharMingen (San Diego, CA) and cultured in DMEM (1 x 106 cells/ml) with MUC1 TR peptide (24 mer, TAPPAHGVTSAPDTRPAPGSTAPP) at 10 µg/ml in a total volume of 5 ml. Target cell lines (B16.MUC1 and B16.neo) were derived from C57BL/6 mice and expressed high levels of MUC1 as determined by FACS analysis using Abs to the TR. Specific 51Cr release at 6 h was calculated according to the following formula: ((experimental release cpm - spontaneous release cpm)/(maximum release cpm - spontaneous release cpm)) x 100. Spontaneous release in all experiments was <15% of maximum release. Ab blocking experiments were performed by preincubating the targets overnight with 1 µg/106 cells of H-2Kb/Db mAb (clone 28-8-6; PharMingen).
Flow cytometry
Single cells from spleens of MET mice were analyzed by two-color
immunofluorescence for alterations in lymphocyte subpopulations: CD3,
CD4, CD8, Fas, FasL, CD11c, and MHC class I and II. Intracellular
cytokine levels were determined after cells were stimulated with MUC1
peptide (10 µg/ml for 6 days) and treated with brefeldin-A (also
called Golgi-Stop; PharMingen) as directed by the manufacturer’s
recommendation (4 µl/1.2 x 107
cells/6 ml for 3 h at 37°C before staining). Cells were
permeabilized using the PharMingen permeabilization kit and stained for
intracellular IFN-, IL-2, IL-4, and IL-5 as described by PharMingen.
All fluorescently labeled Abs were purchased from PharMingen. Flow
cytometric analysis was done on Becton Dickinson FACscan using the
CellQuest program (Becton Dickinson, Mountain View, CA).
Magnetic activated cell sorting (MACS)
Splenic lymphocytes were stained for 30 min on ice with anti-CD8 Ab conjugated to microbeads (Miltenyi Biotechnologies, Auburn, CA). CD8+ cells were positively selected on an RS-type magnetic column using the Vario MACS magnetic device following the protocol provided by the manufacturer (Miltenyi Biotec, Auburn, CA). Purity of the selected cells was checked by flow cytometry and ranged from 92 to 95%.
Adoptive transfer
Two groups of five MUC1.Tg mice were injected (s.c. in the
flank) with B16.MUC1 cells (1 x 106
cells/mouse/200 µl). Simultaneously, one group of mice received (by
i.v. injection) CD8+ CTLs (5 x
106 cells/mouse/100 µl) isolated from an 18-wk
MET spleen. Splenocytes were grown on irradiated B16.MUC1 cells (20,000
rad x 2) with IL-2 (100 U/ml) and IFN- (150 pg/ml) for 2 wk
before sorting for CD8+ cells. The control group
received media alone. Palpations were started 5 days after tumor
challenge. Tumors were measured using a metric dial caliper (Monostat,
Pequannock, NJ), and tumor volume was determined by the formula
(W2 x L)/2.
The experiment was repeated one time.
Peptide affinity assay
Bone marrow cells from C57BL/6 mice were isolated and cultured in DMEM with 10% FBS, penicillin (50 U/ml), streptomycin (50 µg/ml), 1% glutamax, stem cell factor (10 ng/ml) (Stem Cell Technologies, Vancouver, Canada), GM-CSF (60 pg/ml) (PharMingen), and IL-4 (200 pg/ml) (PharMingen) for 7–10 days. Murine dendritic cells (DCs) were purified using the DC purification kit (Stem Cell Technologies) and positively selected using the RS-type column on the Vario MACS device using the manufacturer’s recommended conditions (Miltenyi Biotec, Auburn, CA). The cells were further cultured for 7 days with GM-CSF (60 pg/ml) and IL-4 (200 pg/ml). At this point, cells were tested for DC-specific and nonspecific cell surface markers by flow cytometric analysis. Abs to CD11c, CD14, Mac1, Gr-1, B7.1, B7.2, and MHC class I and II (PharMingen) were used to determine purity by FACS analysis. DCs were pulsed overnight with various MUC1 TR peptides at concentrations ranging from 10-5 to 10-7 M. The peptides tested spanned the entire TR region of MUC1 and ranged in size from 9 to 25 mer. Peptides were kindly provided by Dr. M. Longenecker (Biomira). Irrelevant peptides including MUC4, GP100-1 (human melanoma), and KAS6-13 (human multiple myeloma) were used as negative controls. Peptide-pulsed DCs were labeled with 51Cr (100 µCi/106 cells) for 2 h in a 37°C incubator with slow shaking and served as targets for CTLs from 18 wk MET mice. B16.MUC1 cells were used as positive targets, and B16.neo cells were used as irrelevant targets. A standard 51Cr release assay was performed using the peptide-pulsed DCs as targets and 18-wk MET splenocytes as effectors.
Statistical analysis
All statistical analyses were performed using the two-sample Student’s unpaired t test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We have generated spontaneous pancreatic tumors that express human MUC1 as a self-protein by mating the MUC1.Tg mice with mice that develop tumors of the pancreas (ET mice). The ET mice were developed and kindly donated by Dr. Judith Tevethia (18). They were created by expressing the first 127 aa of SV40 T Ag under control of the elastase promoter. Although the ET mice were made in the B6D2 F1 mouse strain, they have been backcrossed for >20 generations onto B6 mice. These mice have an interesting advantage over a mouse expressing full-length T Ag. The T1-127 protein does not contain any of the previously identified T Ag CTL epitopes that could be used by the C57BL/6 mice for tumor rejection (18). Although mice expressing full-length T Ag should be tolerant to those epitopes, the absence of the epitopes in the T1-127 ET mice avoids any consideration of SV40-specific immune rejection.
These mice exhibit acinar cell dysplasia at birth, which
progresses to microadenomas and single or multiple acinar cell
carcinomas. These pancreatic tumors express large amounts of MUC1
similar to the in vivo situation in humans. Although we have not
reported the survival data for MET mice, it appears similar to that
observed by Tevethia for the ET mice (18), suggesting that
the presence of MUC1 protein does not substantially alter tumor
progression. We have sacrificed animals at 3-wk intervals to monitor
tumor formation and to process the pancreas for microscopy by
conventional methods. Tumors in the pancreas were first grossly
visualized at week 9. By week 12, as many as nine tumors were observed.
Pancreas weights were determined at sacrifice and are shown in Fig. 1
. The weights were a general indication
of tumor burden. However, in many animals 15 wk or older, the weights
did not increase although tumors were present, presumably due to
cachexia.
|
Histologic analysis showed that the pancreas consisted of
dysplastic acinar cells, acinar cell microadenomas, and acinar cell
carcinomas. As the pancreatic tumors developed, there was a definite
change in the epitopes of MUC1 that were exposed. Normal pancreas
expresses MUC1 in acinar and ductal cells as detected by
immunohistochemical analysis using CT1, a rabbit antiserum to the
cytoplasmic tail of MUC1 (Fig. 2, A and B). The CT1 reactivity is not affected by
glycosylation. MUC1 expressed on normal pancreas is heavily
glycosylated (23). This glycosylation masks TR core
protein epitopes for the mAbs BC2 (Fig. 2, C and
D), HMFG-2 and SM-3 (not shown). Although BC2 has been
described as glycosylation insensitive in the mammary gland, which
exhibits a lower level of glycosylation than the pancreas, this clearly
is not the case in the pancreas (22). Normal pancreas in
humans and mice do not exhibit epitopes for BC2. By 3 wk of age,
dysplastic acinar cells had developed, and the glycosylation was
altered. All three glycosylation-sensitive Abs, BC2 (Fig. 2, G and H), as well as HMFG-2 and SM-3 (data not
shown), showed strong reactivity with tumor cells. Each of the Abs is
shown blocked by an appropriate peptide (Fig. 2, B,
D, F, and H). These alterations in
reactive profiles suggested that in the MET mice, as in humans, MUC1
expressed by tumors is underglycosylated, and the protein core, which
is normally covered with carbohydrate, is exposed. This unmasking of
the core protein may reveal the peptide epitopes that are recognized by
cytotoxic T cells that can kill tumor cells expressing this form of
MUC1, similar to what has been observed in humans.
|
, A–D, MET mice at 18 and 21
wk developed large solid tumor masses of less differentiated acinar
cell carcinomas (Fig. 3, E–H). Expression of MUC1 in the
large solid tumors appeared to be greatly decreased when we stained
with our TR Abs (Fig. 3G). However, staining with CT1
antiserum, which is against the cytoplasmic tail of MUC1 and is
glycosylation insensitive, showed high levels of MUC1 expression (Fig. 3E). Specific staining was blocked by appropriate peptides
(Fig. 3, F and H). The staining could represent
either mouse or human MUC1 as CT1 Ab cross-reacts with both mouse and
human. There is no effective methodology to differentiate between the
mouse and human forms of MUC1 using the cytoplasmic tail Ab. The
immunogenic TR epitopes appeared to become masked (presumably with
carbohydrate) in these later stage tumors. At the same time, there
appeared to be decreased amounts of the well-differentiated acinar cell
tumors, as the large tumor mass began to dominate physically. It is
possible that the CTLs detected at week 18 (see below) that were
reactive with underglycosylated MUC1 were eliminating the regions of
tumor expressing aberrantly glycosylated MUC1 and allowing outgrowth of
tumors expressing nonimmunogenic MUC1. Another possibility is that the
CTLs eliminated all tumor cells expressing human MUC1 and tumors that
survived expressed only mouse Muc1 as detected by CT1.
|
A)
and strong expression at 18 wk (Fig. 4E). These results were
confirmed with Western blot analysis of 3- and 18-wk tumor lysates
(data not shown). Within the same pancreas, the well-differentiated
dysplastic acinar cells showed strong expression of FasL especially in
lumenal regions (Fig. 4E), whereas the solid
undifferentiated tumor mass demonstrated much decreased levels in
staining (Fig. 4I). Both 18-wk tumor sections shown in Fig. 4 were from the same slide and thus were stained simultaneously.
Immunohistochemical staining of TGFß in MET pancreatic tumors also
detected increasing levels of TGFß as tumors progressed (Fig. 4, C, G, and K). Expression of TGFß is
a known mechanism by which tumors evade immune recognition
(24).
|
We have analyzed the native immunological responses in the MET
mice as the mice aged and tumors developed. We observed that
nonimmunized MET mice developed MUC1-specific cytotoxic T cells that
lysed 40% of the MUC1-expressing B16 melanoma target cells in the 12
wk MET mice rising to 80% lysis in 18 wk MET mice (Fig. 5A). This lysis compared with
20% lysis of MUC1-negative B16 cells. The level of lysis dropped to
about 40% at week 21. Similar lysis was detected using as targets a
MUC1-expressing mammary gland cell line C57 MG (data not shown). To our
surprise, we detected no CTLs when we used a MUC1-expressing pancreatic
tumor cell line (Panc 02) as targets (data not shown). We have found
that both the B16 melanoma and the C57 MG cells but not the Panc 02
cells expressed low levels of MHC class I molecules
(H-2Kb/H-2Db) on their
surfaces, suggesting that the CTL lysis was MHC restricted. Another
clone of B16.MUC1 cells lacking class I expression was not lysed until
class I was induced by IFN- treatment (data not shown). To confirm
this class I restriction, lysis was blocked completely by an Ab to
class I
(anti-H-2Kb/H-2Db,
clone 28-8-6) (Fig. 5B). The appearance of reactive CTLs by
18 wk in eight of eight MET mice is highly significant
(p < 0.001) and suggests that the high level
of aberrantly expressed MUC1 by tumors may be antigenic and can
mobilize T cells and elicit a cytotoxic response.
|
We have analyzed CTL reactivity to a selection of MUC1 TR
peptides. DCs were loaded with different concentrations of 9-mer and
larger peptides and used as targets in a CTL assay. CTLs recognized the
MUC1 TR sequences with highest affinity to STAPPAHGV epitope (Fig. 6). These results are similar to those
seen in humans, where the CTLs isolated from breast cancer patients
also reacted to the STAPPAHGV epitope with low affinity
(25). The CTLs were of low avidity as they lost their
lytic function by 10-8 M. This could be one of
several reasons why the CTLs had no obvious effect on the spontaneous
growth of pancreatic tumors in vivo. Although our data suggested that
the CTLs were recognizing epitopes on MUC1 TR, we have not yet
determined whether there are CTL epitopes outside of the MUC1 TR
portion, for example, in other regions of extracellular domain or in
the cytoplasmic domain.
|
The CTLs detected in the MET mice should be fully functional at
killing target tumor cells, as they showed increased IFN-
(p < 0.05) and IL-2 (p
< 0.01) expression in culture supernatants, following stimulation for
6 days with MUC1 TR peptide (Fig. 7, A and B). IFN- levels in the supernatant
sample levels ranged from 6,000 to 19,000 pg/ml in 18-wk MET mice (Fig. 7A). IL-4 levels showed no significant increase (data not
shown), suggesting that the immune response is a type 1 response.
|
Increased levels of MUC1 in the sera from mice aged 18 and 24 wk
corresponded directly to the increased levels of CTL activity (Fig. 8). These results suggested that high
levels of tumor-associated MUC1 may activate MUC1-specific CTLs that
are able to lyse MUC1-expressing tumor cells in vitro.
|
CTLs were obtained from splenocytes of 18 wk MET mice and expanded
on irradiated B16.MUC1 cells. CD8+ cells were
selected (98% of the CTLs were CD8+) by MACS and
adoptively transferred into five MUC1.Tg mice (5 x
106 cells in 100 µl PBS, i.v.). At the same
time, these five mice and five control mice received 1 x
106 viable B16.MUC1 cells (s.c. in the flank),
and tumor growth was monitored by palpation. By 34 days, all five mice
that did not receive the CTLs had tumors of 
1.4 g, whereas mice
receiving the adoptively transferred CTLs remained tumor free for 10 wk
(p < 0.05) (Fig. 9A). In vivo adoptive transfer
of the CD8+ CTLs from 18-wk MET mice showed that
the CTLs were effective in eradicating transplanted B16.MUC1 tumor
cells from MUC1.Tg mice. All mice receiving B16.neo tumor cells
developed tumors even in the presence of CTLs (n = 4)
(Fig. 9A), providing further evidence that the immune
response is specific against tumor Ag MUC1. This result suggested that
the naturally occurring immunity in MET mice can be transferred. These
experiments have been repeated with similar results.
|
B).
A third tumor-free mouse was rechallenged with B16.MUC1 cells at week
10 and remained tumor free for an additional 10 wk (data not shown),
suggesting once again that the 18-wk MET CTL effectors were able to
generate memory T cell effectors that provided long-term protection
against MUC1-expressing tumors. These results suggested that the MET
mice developed effective cellular responses that eliminated B16.MUC1
tumor cells in vivo. However, these CTLs were ineffective in
eradicating the spontaneously occurring pancreatic tumors in the
MET mice. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and 3). Modulation of
antigenic epitopes is one way the host can favor emergence of
nonimmunogenic, nonrejectable tumor cell variants (26).
Similarly, as has been described in humans, CTLs were identified in
nearly all of the MET mice between the ages of 16 and 19 wk (Fig. 5),
which effectively lysed MUC1-expressing target cells in vitro.
Importantly, these CD8+ CTLs eradicated
injectable MUC1-expressing tumors when adoptively transferred in vivo
(Fig. 9). The tumor-bearing MET mice, which expressed underglycosylated
MUC1 on tumors and presumably also in serum, developed a CTL response
to MUC1 that differed from the previously described s.c. response
elicited in MUC1.Tg mice bearing B16.MUC1 cells. CTLs from the MET mice
were effective in eradicating injectable tumors in vivo, whereas the
MUC1.Tg mice with the B16.MUC1 s.c. tumor elicited ineffective CTLs
(27). Presumably, the environment in the mice developing
spontaneous tumors greatly influences the immune response. In the
bitransgenic MET mice as in humans, serum MUC1 levels increased
gradually as tumor development progressed, with maximum levels observed
at about 16–19 wk of age (Fig. 8), when mice have a large tumor burden
(Fig. 1) and high CTL activity (Fig. 5).
Thus far our data suggest a strong Th1 cellular immune response as
determined by high CTL activity and high levels of IFN- and IL-2
(Fig. 7, A and B). Intracellular IFN- and IL-2
analyses using FACS confirmed the ELISA results. Interestingly,
intracellular IL-5 also showed a 2- to 6-fold increase as tumors
progressed. It is of interest to note that IL-5 has been correlated to
increased levels of eosinophils, which are known to play a role in
tumor immunity (28, 29, 30, 31).
Characterization of the MET CTL line revealed an MHC class I-restricted
phenomenon (Fig. 5B) with highest affinity to the MUC1 TR
peptide sequence, STAPPAHGV (Fig. 6). Further characterization of TR
peptides is in progress and has revealed that the STAPPAHGV peptide was
presented to the CTLs in the Db but not in the
Kb groove of MHC class I molecule (manuscript in
preparation). Interestingly, this is the same MUC1 peptide sequence
that was recognized by CTL lines isolated from human breast cancer
patients (25). Although our data suggest that the CTLs are
recognizing epitopes on MUC1 TR, we have not yet determined whether
there are CTL epitopes outside of the MUC1 TR portion, for example in
other regions of extracellular domain or in the cytoplasmic domain.
Although we have isolated lytic CTLs from MET mice, their avidity to
the MUC1 TR peptide is low as their lytic activity is lost at
10-8 M peptide concentrations. MUC1 is expressed
in the transgenic mice under its own promoter, and the pattern and
timing of expression of human and mouse MUC1 show complete concordance.
We have detected expression of both mouse and human MUC1 in the mouse
thymus in day 18 embryos and in adult thymus (unpublished data). Thus,
MUC1 is expressed during the time that central tolerance is
established. It is not surprising that high affinity CTLs are not
detectable. Studies from other laboratories have shown the generation
of CTLs with low avidity for tumor cells expressing influenza virus
hemagglutinin (HA) in insulin-dependent HA transgenic mice that express
HA as a self molecule on pancreatic islet ß cells (32).
Significantly, they have shown that vaccination of insulin-dependent HA
mice can activate these low avidity CTLs that are able to reject tumor
cells expressing high levels of HA, without destruction of pancreatic
islet ß cells expressing moderate levels of HA. Studies with other
transgenic mouse models expressing model proteins as self Ags have
shown that T cells with specificity for self proteins can be
demonstrated within the peripheral T cell pool (33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43).
Often, potentially autoreactive T cells manage to escape deletion by
virtue of their lower avidity for self Ag even when the transgene
element is present in, or is available to, the thymus
(44). The appearance of reactive CTLs by 18 wk in most of
the MET mice was highly significant (p <
0.001) and suggested that the high level of aberrantly expressed MUC1
by tumors and in serum was antigenic and able to mobilize these low
avidity T cells that elicited a CTL response in vitro and were able to
eradicate MUC1-expressing injected tumor cells in vivo (Fig. 9).
However, these CTLs were unable to eradicate the spontaneous pancreatic
tumors. The fact that a tumor Ag like MUC1 elicits a tumor-specific
immune response does not necessarily mean that the immune response will
cause the rejection of a spontaneously occurring tumor in vivo. Many
reports have suggested that progressing tumors in cancer patients have
elaborate means of escaping an apparently effective MHC class
I-restricted immune response (45, 46). Other investigators
have found that the CTL response occurs too late to be effective
against the established tumors (47).
Tumors evade host immunity at both the induction and effector phases.
Because MET mice have developed strong anti-MUC1 CTL responses,
these spontaneously occurring pancreatic tumors must have evaded the
existing CTLs. Recent studies have revealed multiple mechanisms by
which tumors have avoided immune destruction. Down-regulating their
surface expression of MHC class I and costimulatory molecules, such as
B7, are a few of the well-studied mechanisms. Preliminary analysis of
several lines of primary pancreatic tumor cells generated ex vivo from
19- to 22-wk MET mice revealed very low level surface expression of
both B7 (3–5%) and MHC class I (7–9%) by FACS analysis (data not
shown). Tumors may evade CTL recognition by changing their antigenic
composition, much as viruses. Clearly, in MET mice, as tumors progress
there is antigenic modulation of MUC1 (Figs. 2 and 3), thereby allowing
escape of MUC1-specific CTL recognition and killing. Other important
escape mechanisms are secretion of immunosuppressive substances or
induction of suppressor cells or cells secreting inhibitory cytokines
in the host immune system and killing of the effector T cells by tumor
cells. Factors implicated in this effect include TGFß
(24). It has been shown previously that TGFß may alter
TCR subcomponent composition and down-regulate CD3
, , and 
but
not 
, thereby reducing T cell signaling and CTL responses against
tumor cells. When TGFß expression by the tumors was reduced using
antisense techniques, CD3 expression was normalized (26).
Pancreatic tumor sections from MET mice showed very high expression of
TGFß2 (Fig. 4C, G, and K) and
TGFß3 (data not shown), suggesting again that the tumors are capable
of down-regulating the function of existing CTLs and evading host
antitumor immunity.
The most radical way for a tumor to induce nonresponsiveness would be
to kill the attacking CTLs. One possible mechanism for T cell killing
involves the Fas/FasL pathway. Several types of tumors (human and
mouse) express FasL, for example primary lung carcinomas, melanomas,
colonic adenocarcinomas, hepatic tumors, multiple myelomas, ovarian
carcinomas, pancreatic adenocarcinomas, astrocytomas, head and neck
squamous carcinomas, and many others (48). Several of
these tumor cells expressing FasL can induce apoptosis of
Fas-expressing CTLs, thereby evading their own killing
(26). In pancreatic tumors from MET mice, we detect high
expression of FasL in the well-differentiated dysplastic acinar cells,
whereas the solid undifferentiated tumor mass demonstrated much
decreased levels in staining (Fig. 4, A, E, and
I). This was surprising as this result would indicate
reduced ability of tumor cells to kill CTLs via the Fas/FasL mechanism.
One could explain this result by presuming that at the undifferentiated
stage, the tumor cells need more protection as they are more accessible
to the existing CTLs, whereas the well-differentiated solid tumor mass
does not need that protection because by this stage they have become
inaccessible to the CTLs. Another possible explanation is that in
certain circumstances, Fas-mediated costimulation rather than
destruction can occur (49). This possibly helps to explain
some instances where expression of FasL results in enhanced destruction
rather than protection of FasL-bearing cells. In any case, Fas/FasL
interactions and apoptosis are complex mechanisms because both T cells
and tumors can express Fas and FasL, making the outcome of the
interaction difficult to interpret.
To determine whether existing CTLs can successfully access the MET tumors, we stained tumor sections for infiltrating lymphocytes. Hematoxylin and eosin staining showed highly vascularized tumors with vessels filled with lymphocytes. However, we did not observe any lymphocytic infiltration into the tumors (data not shown).
Induction of the cellular response to MUC1 was accompanied by a very modest tumor response in the MET mice. We also detected low levels of circulating Abs to MUC1 in the 18-wk MET mice. Two of the four animals had detectable but low levels of Abs to MUC1 (A450 mm readings were 1.5 and 1.7 times the negative control mice, data not shown), suggesting that the aberrant glycosylation and the high level of expression has changed the antigenic profile and elicited a low level humoral immune response to MUC1. Mice from all other time points were defined as negative with A450 mm readings <1.5 times the negative control sera. Abs reactive with MUC1 have been previously reported in a small number of humans with cancer (16, 17). Although a humoral response is often dismissed as being ineffective as eradicating solid tumors it is still interesting that the response in the MET mice once again parallels that in humans.
Several studies in the literature clearly show that there are no simple correlations between the measured CTL responses and clinical responses (50, 51). Clearly, tumor Ags that are capable of eliciting CD8+ CTL responses in vivo will function as important tumor rejection Ags, and their incorporation into effective tumor vaccination protocols is essential. These studies are in progress. Whereas transplantable tumors can be readily cured with immunotherapeutic approaches, similar therapies in cancer patients have been less effective. These apparently contradictory observations between transplantable murine tumor models and cancer patients could be explained in part by the ability of the slowly progressing spontaneous tumor to induce an immune dysfunction as compared with a rapidly growing transplantable murine tumor, which might spare the immune suppression.
The MET model appropriately mimics the human condition and is an excellent spontaneous pancreatic tumor model with which to elucidate the most successful forms of tumor immunotherapy. In humans, both unrestricted and MHC class I-restricted MUC1-specific CTLs have been reported (8, 25); however, we have only detected MHC class I-restricted CTLs in MET mice. It must be realized that the situation in vivo may be quite different from that in vitro. It is very likely that mice do possess unrestricted CTLs in vivo, but these CTLs may be difficult to detect in vitro and therefore have never been described previously in mice (52, 53, 54, 55, 56, 57). Nevertheless, alterations in MUC1 expression and induction of cellular and humoral immune responses in the MET mice recapitulate what has been described in human cancers. These similarities make this an excellent model for testing therapy in a setting relevant to the treatment of human cancer as well as for prevention and delineation of the mechanisms of tolerance, immunity, and autoimmunity.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Sandra J. Gendler, Mayo Clinic Scottsdale, S. C. Johnson Medical Research Center, 13400 East Shea Boulevard, Scottsdale, AZ 85259.
3 Abbreviations used in this paper: MUC1, mucin 1; MUC1.Tg, MUC1 transgenic; TR, tandem repeat; DC, dendritic cell; MET, MUC1-expressing pancreatic tumor mouse model; ET, pancreatic tumor mouse model; FasL, Fas ligand; MACS, magnetic activated cell sorting; HA, hemagglutinin.
Received for publication June 23, 2000. Accepted for publication July 5, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Kottke, J. Pulido, J. Thompson, L. Sanchez-Perez, H. Chong, S. K. Calderwood, P. Selby, K. Harrington, S. E. Strome, A. Melcher, et al. Antitumor Immunity Can Be Uncoupled from Autoimmunity following Heat Shock Protein 70-Mediated Inflammatory Killing of Normal Pancreas Cancer Res., October 1, 2009; 69(19): 7767 - 7774. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Mukherjee, G. D. Basu, T. L. Tinder, D. B. Subramani, J. M. Bradley, M. Arefayene, T. Skaar, and G. De Petris Progression of Pancreatic Adenocarcinoma Is Significantly Impeded with a Combination of Vaccine and COX-2 Inhibition J. Immunol., January 1, 2009; 182(1): 216 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. L. Tinder, D. B. Subramani, G. D. Basu, J. M. Bradley, J. Schettini, A. Million, T. Skaar, and P. Mukherjee MUC1 Enhances Tumor Progression and Contributes Toward Immunosuppression in a Mouse Model of Spontaneous Pancreatic Adenocarcinoma J. Immunol., September 1, 2008; 181(5): 3116 - 3125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Feldmann and A. Maitra Molecular Genetics of Pancreatic Ductal Adenocarcinomas and Recent Implications for Translational Efforts J. Mol. Diagn., March 1, 2008; 10(2): 111 - 122. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. L. Beatty, S. E. Plevy, A. R. Sepulveda, and O. J. Finn Cutting Edge: Transgenic Expression of Human MUC1 in IL-10-/- Mice Accelerates Inflammatory Bowel Disease and Progression to Colon Cancer J. Immunol., July 15, 2007; 179(2): 735 - 739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. D. Basu, T. L. Tinder, J. M. Bradley, T. Tu, C. L. Hattrup, B. A. Pockaj, and P. Mukherjee Cyclooxygenase-2 Inhibitor Enhances the Efficacy of a Breast Cancer Vaccine: Role of IDO J. Immunol., August 15, 2006; 177(4): 2391 - 2402. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. I. Garbe, B. Vermeer, J. Gamrekelashvili, R. v. Wasielewski, F. R. Greten, A. M. Westendorf, J. Buer, R. M. Schmid, M. P. Manns, F. Korangy, et al. Genetically Induced Pancreatic Adenocarcinoma Is Highly Immunogenic and Causes Spontaneous Tumor-Specific Immune Responses Cancer Res., January 1, 2006; 66(1): 508 - 516. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Gerloni, P. Castiglioni, and M. Zanetti The Cooperation between Two CD4 T Cells Induces Tumor Protective Immunity in MUC.1 Transgenic Mice J. Immunol., November 15, 2005; 175(10): 6551 - 6559. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Dwyer, E. R. Bergert, M. K. O'Connor, S. J. Gendler, and J. C. Morris In vivo Radioiodide Imaging and Treatment of Breast Cancer Xenografts after MUC1-Driven Expression of the Sodium Iodide Symporter Clin. Cancer Res., February 15, 2005; 11(4): 1483 - 1489. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Karsten, N. Serttas, H. Paulsen, A. Danielczyk, and S. Goletz Binding patterns of DTR-specific antibodies reveal a glycosylation-conditioned tumor-specific epitope of the epithelial mucin (MUC1) Glycobiology, August 1, 2004; 14(8): 681 - 692. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Xia, Y. Tanaka, S. Koido, C. Liu, P. Mukherjee, S. J. Gendler, and J. Gong Prevention of Spontaneous Breast Carcinoma by Prophylactic Vaccination with Dendritic/Tumor Fusion Cells J. Immunol., February 15, 2003; 170(4): 1980 - 1986. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Cohen, N. Hoffmann, M. Farago, H. R. Hoogenboom, L. Eisenbach, and Y. Reiter Direct Detection and Quantitation of a Distinct T-Cell Epitope Derived from Tumor-specific Epithelial Cell-associated Mucin Using Human Recombinant Antibodies Endowed with the Antigen-specific, Major Histocompatibility Complex-restricted Specificity of T Cells Cancer Res., October 15, 2002; 62(20): 5835 - 5844. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. Wilkinson, E. L. Ross, D. Ellison, W. Zimmermann, D. Snary, and S. J. Mather Evaluation of a Transgenic Mouse Model for Anti-Human CEA Radioimmunotherapeutics J. Nucl. Med., October 1, 2002; 43(10): 1368 - 1376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. VanLith, K. G. Kohlgraf, C. L. Sivinski, R. M. Tempero, and M. A. Hollingsworth MUC1-specific anti-tumor responses: molecular requirements for CD4-mediated responses Int. Immunol., August 1, 2002; 14(8): 873 - 882. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Schnurr, P. Galambos, C. Scholz, F. Then, M. Dauer, S. Endres, and A. Eigler Tumor Cell Lysate-pulsed Human Dendritic Cells Induce a T-Cell Response against Pancreatic Carcinoma Cells: an in Vitro Model for the Assessment of Tumor Vaccines Cancer Res., September 1, 2001; 61(17): 6445 - 6450. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. Wilkinson, E. L. Ross, R. Poulsom, M. Ilyas, J. Straub, D. Snary, W. F. Bodmer, and S. J. Mather Antibody targeting studies in a transgenic murine model of spontaneous colorectal tumors PNAS, August 17, 2001; (2001) 181353498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. W. Wilkinson, E. L. Ross, R. Poulsom, M. Ilyas, J. Straub, D. Snary, W. F. Bodmer, and S. J. Mather Antibody targeting studies in a transgenic murine model of spontaneous colorectal tumors PNAS, August 28, 2001; 98(18): 10256 - 10260. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
