| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Trudeau Institute, Saranac Lake, NY 12983
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, however,
control tumor growth less well, whereas Tc2 effectors lacking IFN-
are unaffected. Tc1 from IFN--deficient mice attract fewer host
cells to the draining lymph node, whereas Tc1 cells from
perforin-deficient donors are unimpaired. We conclude that host cell
recruitment is a crucial element in adoptive immunotherapy. The
differences between the EG7 and the previous B16 melanoma model are
discussed. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A number of investigators have used the adoptive transfer model to investigate the role of the immune system in the response to tumor Ags (5, 6, 7, 8, 9, 10, 11, 12, 13, 14), and CD8 T cells have been shown to have a major therapeutic role in antitumor responses in a number of experimental tumor models (5). In these studies, investigators have used in vitro activated tumor-specific CD8 T cells from wild-type mice (6) or CD8 T cells from mice possessing the transgenic TCR specific for an Ag introduced into the tumor (7, 8, 9, 10) or a naturally occurring tumor Ag (12). In some studies naive or memory T cells were transferred (12); in others, CD8 T cells from recently immunized mice were used (13, 14); whereas in yet others, cytotoxic CD8 T cell clones were transferred (5). Variable results were obtained, in part due to the different tumor models used but in part due to the activation state of the transferred cells.
In our own studies (8, 9, 10), we examined the role of
adoptively transferred Tc1 and Tc2 cytokine polarized subsets of CD8
effector T cells in the control of the growth of an already established
tumor, the B16 melanoma, which grows in the lung. We used in
vitro-generated cytokine Tc1 and Tc2 polarized effector cells from
naive CD8 populations from OT-I mice that bear the 
- and 
-chains
of the TCR specific for the OVA-derived peptide presented on
Kb and OVA-transfected B16 melanoma cells. The
tumor cells were injected i.v. 7 days before adoptive transfer. Graded
numbers of effector Tc1 or Tc2 cells were injected i.v., and the
survival times and numbers of lung metastases were determined in each
case. We found that Tc1 cells were more effective than Tc2 in
controlling tumor growth and in prolonging the survival times of
tumor-bearing mice. Approximately 25 times more Tc2 cells were required
to bring about the same degree of control as for Tc1 cells
(8). Under these conditions, the tumors grew out at later
times; but when fewer tumor cells were used, the adoptively transferred
cells were able to prolong survival indefinitely. Tumor cells were
still present, however, in the lungs of healthy mice, apparently
controlled by the presence of CD8 T cells and other cells
(9). In further studies (Ref. 10 and M. J.
Dobrzanski and R. W. Dutton, unpublished observations), we
determined the role of various donor effector molecules by generating
effectors from OT-1 mice crossed to various deficient mice. In these
studies, we found that donor production of IFN- was essential for
full Tc1 activity but not for Tc2 activity (10). Effector
cell-derived IL-4 or IL-5 was equally important for Tc2, but not Tc1,
activity, whereas perforin deficiencies had no effect on either Tc1 or
Tc2 activity (M. J. Dobrzanski and R. W. Dutton, unpublished
observations). IFN- was shown to inhibit tumor growth in vitro and
to up-regulate class I and class II MHC, CD95, and TNFR gene expression
and to induce the secretion of IFN--inducible protein 10 and other
chemokines (10). We also showed that although the
adoptively transferred cells destroyed most of the tumor cells shortly
after transfer, host cells were required to control further growth of
surviving tumor cells (10). We found that Tc2 cells, that
were equally effective whether they came from IFN--deficient or
wild-type mice, were not effective if the host were IFN- deficient
(M. J. Dobrzanski and R. W. Dutton, unpublished observations).
In the current studies, we sought to determine whether the same parameters were important in the control of a different tumor, growing in a different anatomical location. We used the OVA-transfected EL4 thymoma, EG7, and injected the tumor intradermally into the flank of the recipient mice. In the absence of adoptive transfer, the tumor grows progressively, and the mice die if left untreated. The adoptively transferred cells either were injected mixed with the tumor, as in the classic Winn assay (15), or were injected i.v. at the same time as the tumor or 7 days after tumor injection. We found substantial differences in the relative effectiveness of the Tc1 and Tc2 effectors under the different conditions. Tc2 were one-fifth to one-tenth as effective as Tc1 when injected together with the tumor but were only one-twentieth to one-one hundredth as effective when injected i.v. We used Thy-1 congenic donor and recipient combinations to monitor the entry of donor cells into the draining and contralateral lymph nodes and spleen and also determined the accumulation of host cells in these locations. We found that Tc1 cells entered the draining lymph nodes more rapidly than Tc2 cells and caused a more rapid accumulation of host cells at the same site. Both populations brought about an abrupt termination of host cells migration into and exit from the nodes and spleen after elimination of the tumor, but this took place earlier after transfer of Tc1 than Tc2. It appeared that the host cell response was ineffective in the absence of adoptive transfer but was an essential component of the response after transfer.
We examined next which functions of the Tc1 and Tc2 effectors are
involved in the mechanism of tumor rejection by generating effectors
from perforin- or cytokine-deficient mice and comparing their activity
with comparable effectors from wild-type mice. Perforin expression in
the donor cells appears to play no role in either the Tc1 or Tc2
control of the tumor, and neither IL-4 nor IL5 seem to be needed for
Tc1 or Tc2 function. Tc1 cells lacking IFN- gene expression,
however, are severely handicapped in the function, whereas Tc2 lacking
IFN- gene expression are unaffected. In vitro exposure of the EG7 to
IFN- elevates the already high expression of class I MHC but does
not appear to induce any other changes that might render the tumor less
able to grow in vivo. We show that lack of perforin has no effect on
the accumulation of donor or host cells in the ipsilateral lymph nodes.
Cells from IFN--deficient donors, however, are markedly less able to
bring about host cell accumulation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
TCR-transgenic OT-I mice (16) specific for the OVA
peptide, SIINFEKL, were used as the source of donor cells. In some
experiments, donor cells were from OT-1 mice, crossed to C57BL/6
perforin-/-,
IFN--/-, IL-4-/-,
and IL-5 -/- mice to produce OT-1 mice that
were deficient for each of the single genes. Syngeneic C57BL/6,
PL.Thy-1.1, or C57BL/6 SCID mice were used as recipients.
Cell lines
The H-2b cell lines EG7 and EL4 were obtained from American Type Culture Collection (Manassas, VA).
Monoclonal Abs
The following mAbs were used for immunofluorescent staining:
Cy-Chrome anti-CD8 (BD PharMingen, San Diego, CA); FITC
anti-CD4 (BD PharMingen; GK1.5); FITC anti-CD62L (BD
PharMingen; clone MEL-14); FITC anti-CD44 (BD PharMingen; clone
IM7); FITC anti-CD25 (BD PharMingen; IL-2R, -chain, clone
3C7); FITC anti-Ly6C (BD PharMingen; clone AL-21). For details of
additional specific Abs used in these studies, see figure legends.
Isotype-matched Abs were used as negative controls. Cell populations
were analyzed by multiparameter flow cytometry using a Becton Dickinson
FACSCalibur or FACScan instrument (Becton Dickinson Immunocytometry
Systems, Mountain View, CA). Surface marker analysis was performed
using CellQuest software (Becton Dickinson Immunocytometry
Systems).
Cell preparations
CD8 T cells were isolated from the spleen and lymph nodes of the
OT-1 TCR-transgenic mice and enriched by passage through nylon wool and
treatment with anti-CD4 (RL172.4), anti-heat-stable Ag (J11D),
anti-class II MHC (D3.137, M5114, CA4) mAbs and complement. Small
resting CD8 T cells were harvested from the bottom interphase of a
four-layer Percoll gradient (Sigma, St. Louis, MO). The freshly
isolated T cell populations were 90–95%
CD8+V8+ T cells. Splenic
APCs were enriched from C57BL/6 mice by T cell depletion using
anti-Thy-1.2 (HO13.14 and F7D5), anti-CD4 (RL172.4), and
anti-CD8 (3.155) mAbs.
Tc1 and Tc2 effector generation
CD8 cells were cultured in RPMI 1640 (Irvine Scientific,
Santa Ana, CA) supplemented with penicillin, streptomycin, glutamine,
2-ME, and 10% FCS (HyClone Laboratories, Logan, UT). C57BL/6 3-day
LPS-stimulated B cell blasts were used as APCs and were loaded with the
OVA peptide (11 µM) at 37°C for 30 min, treated with mitomycin C
(100 µg/ml; Sigma) at 37°C for 40 min, and washed three times
before use. CD8 effector T cells were prepared from the OT-1-transgenic
mice by 4-day culture under polarizing conditions as previously
described (8). On day 4 of culture, effectors were
95–99% CD8+V2+
cells.
Analysis of cytokine production
Tc1 and Tc2 effector cells were restimulated in vitro and
supernatants were collected at 24 and 36 h. IFN-, IL-4, and
IL-5 cytokine levels were determined by ELISA as previously described
(8).
CTL assays
Tc1 and Tc2 effector cells cytolytic activity was determined in a 51Cr release assay (8).
RNase protection assays
Pelleted cells were resuspended in Trizol reagent (Life Technologies, Gaithersburg, MD). Total RNA was isolated by chloroform extraction and ethanol precipitation and analyzed using the RiboQuant Multiprobe RNase Protection Assay system (BD PharMingen). mCK-1, mCK-3, mCK-4 cytokine; mCR-1 cytokine receptor; mCK-5 chemokine; mCr-5 and mCR-6 chemokine receptor; and mAPO-3 "death gene" mRNA detection multiprobe template sets were then used. Bands were detected and quantified using the Molecular Imager FX (Bio-Rad Laboratories, Hercules, CA), and the Quantity One software analysis package (Bio-Rad). All values reported were normalized as a percent of the band intensity measured for the L32 housekeeping gene.
Intradermal tumor establishment
Syngeneic C57BL/6, B6.PL-Thy-1, or C57BL/6 SCID mice were injected intradermally with EG7 tumor cells (typically 3 x 106 cells) in 100-µl volumes of sterile PBS. Tumor cells for injection were recovered from log phase in vitro growth and were injected into the right flank skin of recipient mice. Tumors were clearly visible after 1 wk and grew progressively, in an encapsulated fashion.
Tumor measurement and volume approximation
Tumors were measured on two perpendicular axes using a vernier caliper. Tumor volumes were approximated by multiplying the measured length by the measured width by the calculated mean of these measured values and were presented as the mean of (typically) five identically treated mice ± SEM. Tumor size was determined weekly.
Adoptive immunotherapy of tumors
Varying numbers of Tc1 and Tc2 effector cells were injection in 500-µl volumes of PBS. Injections were delivered i.v. unless otherwise stated.
In vivo trafficking of transferred effector cells
B6.PL-Thy-1 mice receiving 107 polarized Thy-1.2 effector cells on day 0 were sacrificed (in duplicate) on days 1, 3, 5, 7, 10, and 13 (or 14) after adoptive transfer. Inguinal and cervical lymph nodes were collected and pooled from the tumor-ipsilateral or tumor-contralateral sides of sacrificed mice. Spleens were also collected at this time. Recovered populations were stained with mAbs specific to mouse CD8 (2.43), CD4 (GK1.5), Thy-1.2 (53-2.1), CD19 (1D3), NK1.1 (4D11), Mac-1 (M1/70), and Ly-6G (RB6) and analyzed by three-color multiparameter flow cytometry. Numbers of individually identified cell populations in each organ preparation were calculated by multiplying the total organ cell count by the frequency of cells that stained with each appropriate cell surface phenotype.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
EG7 is a derivative of the H-2b thymoma EL4,
which was transfected with cDNA encoding full-length chicken OVA
(17) and displays the immunodominant epitope of chicken
OVA on its surface, identified as the eight-residue peptide
OVA257–264, with the amino acid sequence
SIINFEKL (18, 19). On intradermal inoculation of C57BL/6
mice with EG7 cells, tumors grew progressively and in a dose-dependent
manner in the skin of the injected animals (Fig. 1
). Although detectable tumors could be
produced with as few as 105 EG7 cells per mouse,
reliable tumor establishment required injection of at least 3 x
106 tumor cells. Accordingly, 3 x
106 cells was adopted as the minimum dose for
reliable intradermal establishment of tumors and became the standard
dose for producing tumors in vivo as part of the therapeutic model
system. Tumors established at this dose grew progressively in untreated
animals, usually producing death of recipient mice within 40 days of
tumor cell injection (data not shown). Mice were routinely sacrificed
no later than day 28.
|
The CD8 T cells of OT-I TCR-transgenic mice specifically recognize
the SIINFEKL peptide in the context of Kb class I
MHC (16, 17) and were used to generate Tc1 and Tc2
effector populations as described in Materials and Methods.
The cell surface phenotype of the effector cells was monitored to
confirm that the cells were CD8+,
CD4-, CD62 ligandlow,
CD44high, CD25+, and
Ly6C-. The functional properties of these
effector cells were as previously described. Tc1 secreted IFN- on
restimulation in vitro but no IL-4 or IL-5, whereas Tc2 made much
smaller amounts of IFN- and made large amounts of IL-4 and IL-5. The
two populations showed equal Ag-specific cytolytic activity against EG7
targets.
Effect of adoptively transferred CD8 effector cells on tumor growth
In the first experiments, the effects of the Tc1 and Tc2 effectors
on EG7 were compared in a Winn assay (15). Polarized
effector cell populations were prepared as described, and graded
numbers of effector cells were mixed with a constant number (3 x
106) of EG7 cells before coinjection into the
skin of recipient C57BL/6 mice. After injection, recipient mice were
monitored for the presence of tumors and the tumor sizes were measured
as described in Materials and Methods. It can be seen (Fig. 2A) that both Tc1 and Tc2 were
effective at controlling tumor growth but that Tc1 were more effective
than Tc2. Approximately 5–10 times as many Tc2 were required to
achieve the same measure of control as with Tc1. Approximately 3
x 105 Tc1 cells were required to give complete
suppression of tumor growth.
|
B, but under these conditions Tc2
cells were considerably less effective than Tc1, requiring 20–100
times more Tc2 cells than Tc1 cells to achieve the same measure of
control. However, Tc1 cells were almost as effective when injected i.v.
as when coinjected with the tumor.
In a third set of experiments (Fig. 2C), we determined the
effect of transferred cells on established tumors, injected 7 days
before adoptive transfer. Again, Tc2 were less effective than Tc1, with
20–25 times as many cells being required to achieve the same
effect. More than 10 times as many Tc1 cells were now required when the
adoptive transfer was delayed to day 7 rather than day 0. The transfer
of even 107 naive CD8 T cells had no effect on
tumor growth (data not shown).
Trafficking of donor and host T cells
In a further series of experiments, we investigated the entry and the accumulation of donor and host cells into the draining and contralateral lymph nodes and the spleen. Donor cells came from the OT-1 Thy-1.2-positive donors which were transferred into C57BL/6 Thy-1.1 recipients (PLThy-1.1). The lymph nodes and spleen were harvested from recipient mice at various times after adoptive transfer of donor CD8 effectors and analyzed by flow cytometry for donor (Thy-1.2+) and host (Thy-1.2-) cell content, and the total number of cells was determined. The host (Thy-1.2-) cells were further characterized as CD4, CD8, B cells, NK cells, or neutrophils, using the Abs indicated in the figure legends.
It can be seen that, in the absence of adoptive transfer,
there is a substantial accumulation of host cells into the draining
lymph node with time but not the contralateral node (Fig. 3, 
). There is also an increase of
host cells in the spleen. The analyses of cell phenotype showed that
CD4, CD8, B cells, and macrophages were recruited in substantial
numbers to the ipsilateral node. Fewer NK cells and neutrophils were
also seen to accumulate. All of the categories of cells measured were
seen to increase in number in the spleen.
|
, •), Tc1 donor cells
entered the ipsilateral or draining lymph nodes and spleen very rapidly
but did not enter the contralateral nodes. In mice injected with Tc2
effectors (
), Tc2 cells also entered the draining nodes and spleen
but in smaller numbers, and the entry into the draining LN (but not the
spleen) was delayed compared with Tc1 cells. Donor Tc1 T cells appeared in the draining lymph nodes within 3 days of transfer and were tetramer positive when stained with Kb SIINFEKL tetramers (data not shown). The arrival of Tc2 cells was delayed for several days and peaked at day 10. The rate of accumulation host B cells, NK cells, and neutrophils into the draining LN was accelerated after adoptive transfer of Tc1 cells but not Tc2 cells. The accumulation of cells in the draining lymph node and the spleen is severely truncated after elimination of the tumor.
Role of host response.
There is clearly a host response in the absence of adoptive transfer as
seen in Fig. 3, but the growth of the tumor does not appear to be
limited. This was further investigated by comparing wild-type and SCID
mice as recipients (Fig. 4). The tumors
grew at the same initial rate in the SCID mice as in wild-type controls
(data not shown). It can be seen that the initial control of the tumor
by adoptively transferred cells is the same as in the wild-type
recipients but that the tumor grows out in SCID recipients at later
times (day 22), in contrast to the situation in wild-type recipients,
even after the transfer of 107 Tc1 effectors.
|
, IL-4, or IL-5). The
effector populations were examined for cytolytic activity and for
the ability to secrete cytokines on restimulation with Ag
in vitro. Each of the populations was defective only in the function
controlled by the deleted gene (data not shown). Effectors from all but
the perforin-deficient mice had comparable cytolytic activity, whereas
effectors from only IFN-, IL-4, or IL-5-deficient mice lacked the
ability to produce the cytokine controlled by the deleted gene. We next
determined how each deletion affected the ability of the adoptively
transferred cells to control tumor growth. It can be seen (Fig. 5A) that the performance of
both Tc1 and Tc2 cells was independent of perforin expression, there
being no difference in the number of defective and wild-type cells
required to achieve a given level of control of tumor
growth. The same finding was true for effectors prepared from IL-4 or
IL-5-deficient mice as can be seen in Fig. 5, B and
C. Tc1 and Tc2 from IL-4- or IL-5-deficient mice were as
effective as the corresponding populations from wild-type mice. When we
examined the performance of CD8 effector T cells prepared
from IFN--deficient mice, we found that the effect of Tc1 cells from
deficient mice was markedly impaired (Fig. 5D).
Approximately 20–50 times as many Tc1 effectors from IFN--deficient
mice were required to achieve the same level of control as with cells
from wild-type mice. The performance of the Tc2 cells was unaffected,
and the Tc1 populations from IFN--deficient mice were
indistinguishable from that of Tc2 populations from either wild-type or
IFN--deficient mice in their ability to control tumor growth.
|
on EG7 tumor cells
We had shown in previous studies in a related tumor model using
the OVA-transfected B16 melanoma that IFN- had a number of effects
on the tumor cultured in vitro that could be expected to hamper tumor
growth or survival. Thus tumor growth was inhibited and the expression
of class I and class II MHC, CD95, and TNFR55 were up-regulated when
B16 melanoma cells were cultured with 1000 U/ml IFN- for 24–36 h.
The tumor was also induced to express message for IFN--inducible
protein 10 and other chemokines. Experiments were therefore undertaken
to identify what direct effects Tc1-derived IFN- might have on EG7
cell expansion and protein expression. In these experiments, EG7 cells
were harvested from log phase in vitro growth and resuspended at
105 cells/ml in complete RPMI medium ± 1000
U/ml IFN-. After additional culture for 24, 48, or 72 h, the
tumor cells were recovered from culture, counted, and stained for cell
surface H-2Kb class I MHC,
I-Ab class II MHC, and Fas protein expression
(Fig. 6B). Whereas exposure to
1000 U/ml IFN- has been shown to profoundly suppress cell recovery
of an OVA-transfected melanoma tumor line used in similar studies in
the laboratory (10), no deleterious effect was seen in the
growth kinetics of EG7 thymoma cells (Fig. 6A). However, the
presence of exogenous IFN- was seen to induce a roughly 10-fold
enhancement of class I MHC expression in EG7, and a modest
up-regulation of Fas. Neither of these effects was believed to
significantly impact the overall immunogenicity of these tumor cells,
however, because EG7 cells already demonstrated ample expression of
both class I MHC and Fas in the absence of added IFN- (Fig. 6B). In this same experiment, 107
cells from each treatment group were collected at 24, 48, and 72 h
of culture for use in the preparation of total RNA. Using these
samples, RNase protection assays were then conducted to determine the
expression of the genes of these cells encoding a fairly comprehensive
list of cytokines and cytokine receptors and chemokines, chemokine
receptors, and so-called death genes. Generally, very little expression
of these genes was observed among EG7 cells, regardless of their
exposure to IFN- in vitro. Still, bands were identified that
corresponded to hybridized IL-2, TNF-, TGF-1, IL-11,
monocyte-CSF, and leukocyte-inhibitory factor cytokine mRNA. Also, some
expression was seen for the IL-7, IL-9, IL-13, IL-15, IL-4, and IL-2R
-chains, common -chain, and IL-2R -chain. Although EG7 cells
did not appear to express any of the chemokine gene products assayed,
these cells were observed to express significant levels of the
chemokine receptor CXCR4, as well as more limited amounts of CCR4 mRNA.
The expression of death genes did not appear to be affected by culture
in IFN-.
|
-deficient effectors
We showed above that the adoptive transfer of Tc1 effector cells
led to an increased rate of host cell entry into the draining lymph
nodes and that this correlated with their superior efficiency in the
control of tumor growth. We next determined the effect of gene
deficiencies in the Tc1 effector cells on the recruitment of cells into
the draining lymph nodes as described in Materials and
Methods. Tc1 4-day effectors were prepared from perforin-deficient
OT-1 mice and were injected into mice that had been injected with tumor
on day -7. The accumulation of host cells in the ipsi- and
contralateral nodes and the spleen was determined as before. It can be
seen in Fig. 7 that the accumulation of
host cells was the same in mice that received wild-type (•) or
perforin-deficient effector cells (), with a possible slight
diminution of the accumulation of B cells, neutrophils, and
macrophages. In contrast, mice receiving Tc1 effectors from
IFN--deficient OT-1 showed a reduced rate of host cell entry ().
The entry of CD4 cells, CD8 cells, and macrophages was markedly
reduced. The entry of B cells was slightly reduced, and the entry of
neutrophils was delayed.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A). When the CD8 effectors are injected i.v., however,
whether simultaneously (Fig. 2B) with the tumor or 7 days
later (Fig. 2C), the Tc2 are 20–100 times less effective
than Tc1. More cells were required, however, if the adoptive transfer
was delayed to day 7, presumably because there were more tumor cells to
control. We presume that the difference between i.v.-injected Tc1 and
Tc2 may be due, in part, to the relative effectiveness of the two
populations in gaining entry to the draining lymph node and in the
ability to recruit host cells to the site. Indeed, we have shown (Fig. 3) that although both populations enter the spleen early, the entry of
Tc2 cells into the draining lymph node is less rapid for Tc2 and that
they are less effective at recruiting host cells to the nodes. We have
shown, however, that the in vitro killing in a 4-h chromium release
assay is perforin mediated (20) whereas the effectiveness
of CD8 T cells injected i.v. appears to be independent of perforin
secretion (Fig. 5A), and the actual method of tumor
eradication in vivo is unclear. It is therefore possible that factors
other than lymph node homing may also be important causing the reduced
effectiveness of Tc2 cells.
The measurement of host cell accumulation in the absence of
adoptive therapy shows that there is a considerable immune response to
the tumor Ag, which is, however, unable to limit tumor growth (Fig. 3).
This has been seen in a number of tumor models and it has been
suggested that tumor Ags elicit low affinity responses because high
affinity responses, to what is essentially a self Ag, have been
deleted. The Ag here, however, is OVA, and there is no a priori reason
why this should elicit T cells with low affinity receptors. It is
possible that the T cells are in some other way qualitatively different
in the response of untreated mice, and this is currently the focus of
further study. The accumulation of host cells in the untreated mice
occurs both in the draining lymph nodes and in the spleen but is not
seen in the contralateral node. This accumulation includes CD4 and CD8
T cells, B cells, NK cells, neutrophils, and macrophages. Adoptive
transfer of Tc1 effectors accelerates the accumulation of host cells,
especially host B cells, NK cells, and neutrophils in the draining
lymph nodes, and it is tempting to believe that these cells may have a
role in controlling tumor growth. There are more CD4 and CD8 host cells
at the height of the response in the untreated mice than there are when
the Tc1-treated mice undergo tumor rejection (Fig. 3). NK cell and
eosinophils, however, reach higher levels in the treated mice than the
untreated and are perhaps the key to successful control of the tumor
growth. The adoptive transfer of Tc2 cells (Fig. 3) has little effect
on the accumulation of host cells (probably because they themselves
were delayed in the their entry into the draining nodes), and this may
correlate with their lesser effectiveness in controlling tumor growth.
Histological analyses of frozen sections showed that Tc1 and Tc2 cells
also infiltrated the tumor (data not shown), but no definitive
conclusion could be drawn about the kinetics or relative amounts of
infiltration of the two populations. The positive correlation between
the degree of host cell migration into the draining lymph node and the
control of tumor growth in both wild-type Tc1 vs Tc2 and Tc1 wild-type
vs Tc1 IFN--/- provides circumstantial
evidence that factors that control both donor and host cell migration
play a very important role in tumor rejection.
Although the host response in the absence of adoptive transfer was not
able to control the growth of the tumor, it clearly played a role
after adoptive transfer in that cells accumulated more extensively in
the draining lymph nodes and the host cells departed abruptly after
tumor rejection. The role of the host response was also seen when
C57BL/6 SCID mice were used as recipients (Fig. 4). The initial control
of tumor growth was indistinguishable from that in wild-type host, but
at later times the tumor growth was not controlled in SCID recipients
and tumors grew out after a delay of only 7–10 days.
The departure of both host and donor cells after the initial control of the tumor is reminiscent of the results reported by Shrikant and Mescher (7) in which adoptively transferred cells first migrated into the peritoneal cavity where the tumor was growing but left subsequently, allowing the tumor to grow out at later times. One can also speculate that the failure of adoptively transferred CD8 T cells to control more established tumors while rejecting newly injected tumors, seen by Hanson et al. (11) may be dependent on differences in cell migration into the tumor and may be a consequence of the fact that the older tumor no longer secretes chemoattractant factors. Alternatively, the secretion of excessive concentrations of chemokine might lead to the down-regulation of chemokine receptors and the failure of further migration before the donor or host cells reached the tumor.
Finally, the effector cells used in this study were very efficient in controlling tumor growth, and this may be because they were effector cells requiring no additional activation. We have shown in a different model that effector CD8 T cells, expressing chemokine receptors CCR5 (Tc1) or CCR4 (Tc2), migrate rapidly into sites of inflammation (21) whereas naive cells and memory cells require activation before they can migrate (22). In the current studies, naive OT-1 CD8 T cells were totally unable to control tumor growth on adoptive transfer even at the highest cell number, whereas memory cells were less effective than memory cells on a per cell basis (data not shown).
We have begun here the first phase of an analysis of the factors that
control these events and lead to more effective control of tumor growth
by using effector cells generated from gene knockout mice. OT-1 mice
were crossed to perforin, IFN-, IL-4, or IL-5 knockout mice and bred
to produce OT-1-positive mice that were homozygous for each of the gene
defects. We showed that for each knockout combination, the mice were
defective only for the deleted gene and that the ability to kill target
cells (except for the perforin knockouts) or secrete other cytokines
was not affected.
The first factor that we examined was perforin; it was somewhat
surprising to find that effectors from Tc1 or Tc2 from perforin
knockout mice were as effective as those from wild-type donors, and it
is clear that control of tumor growth is not dependent on direct
perforin-mediated lysis (Fig. 5A). This was surprising
because the effectors clearly killed EG7 cells in an in vitro assay by
perforin-mediated lysis. Kagi et al. (23) have shown a key
role for perforin-mediated lysis in CD8 and NK lysis of a variety of
target cells, whereas others have shown that Fas ligand
(24)- or TNF- (25)-mediated killing may be
important in responses to tumors. We are currently breeding the OT-1
transgene onto the Fas ligand-deficient (GLD) and the
TNF--deficient backgrounds to test for the role of these alternate
cytolytic mechanisms in this model.
Similar experiments were conducted with Tc1 and Tc2 effectors
prepared from OT-1 IL-4- and IL-5-deficient mice. No effect was seen
with either Tc1 or Tc2 for either cytokine, and effectors from
wild-type and knockout mice were equally effective as shown in Fig. 5, B and C. This finding is contrast to our previous
finding with the B16 melanoma tumor model (2) in which we
had shown that the effectiveness of the Tc2 (but not Tc1) effectors was
severely compromised in IL-4 or IL-5 knockout mice. Whether the
difference in findings is dependent on a difference between the two
tumors (melanoma vs thymoma) or is a consequence of the different
locations of the two tumors (lung vs intradermal) is not known.
In contrast, Tc1 effectors from OT-1 mice deficient in IFN-
were significantly less able to control tumor growth (Fig. 5D) than those from wild-type mice, as had been seen in the
B16 melanoma model (10). The Tc2 cells from the IFN-
knockout mice, however, were as effective as Tc2 effectors from
wild-type mice, indicating that IFN- played no role in the function
of the Tc2 cells. The effectiveness of Tc1 effector cells from Tc1
IFN--deficient mice was equal to that of Tc2 effectors from
wild-type mice, suggesting that there might be some basic mechanism
common to both Tc1 and Tc2 effectors and that the Tc1 had an additional
IFN-mediated effect.
It is clear that Tc1 secretion of IFN- is a significant factor in
the ability of the effector cells to control tumor growth. In addition
to its role as a powerful immunoregulatory cytokine, IFN- has also
been shown to exert a variety of direct inhibitory effects on a number
of different tumor cell types. For example, in tumors that express only
meager amounts of cell surface class I MHC, exposure to IFN- has
been shown to markedly up-regulate class I expression, enhancing the
functional antigenicity of the tumor cells (26). Also,
IFN- can induce cell cycle arrest of tumor cells (27)
and promote an overall state of tumor dormancy in vivo
(28). Furthermore, exposure to IFN- has been shown to
also directly enhance the expression of Fas on the surface of renal
carcinoma cells, heightening their susceptibility to Fas/Fas
ligand-mediated mechanisms of killing (29). In the B16
model we showed that high concentrations of IFN- prevented in vitro
growth of the tumor and led to an up-regulation of class I and class II
MHC, Fas (CD95), and TNF receptors, all potentially increasing the
vulnerability of the tumor to mechanisms inducing apoptosis
(10). IFN- also induced the up-regulation of message
for IFN--inducible protein 10 (IP10) and other chemokines, which had
the potential to attract CD8 effectors to the tumor.
The studies presented here showed that the presence of exogenous
IFN- induce a substantial enhancement of EG7 cells class I MHC
expression, and a modest up-regulation of Fas (Fig. 6B).
Neither of these effects is likely to significantly impact the overall
immunogenicity or vulnerability of these tumor cells, however, because
EG7 cells already demonstrated ample expression of both class I MHC and
Fas in the absence of added IFN-. Generally, very little expression
of the additional genes studied (cytokines, cytokine receptors,
chemokines, and chemokine receptors) was observed in EG7 cells,
regardless of their exposure to IFN- in vitro. The expression of
death genes was not affected. It seems likely that the diminished
ability of the Tc1 cells to control tumor growth was dependent on
different mechanisms in the two models.
We found, however, that effectors from IFN--deficient mice were
significantly impaired in their ability to recruit host cells into the
draining lymph nodes, whereas effectors from perforin-deficient mice
were unimpaired. It seems likely that the IFN- plays a significant
role in the recruitment of host cells and that these are essential for
complete control of tumor growth. This provides yet another example of
the correlation between the ability of donor cells to cause a host cell
accumulation in the draining node and the ability to limit tumor growth
observed in the accompanying study, and it suggests that the control of
migration is a key factor in the ability of the immune response to
bring about tumor rejection. The role of chemokines and chemokine
receptor expression in this process is currently under
investigation.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Richard W. Dutton, Trudeau Institute, Box 59, Saranac Lake, NY12983. E-mail address: dutton{at}northnet.org
Received for publication December 7, 2000. Accepted for publication March 19, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
in early and late stages of type 1 CD8 effector cell-mediated tumor rejection. Clin. Immunol. 98:70.[Medline]
in CD8 T cell-mediated elimination of 3LL-A9 Lewis lung carcinoma cells in vivo. J. Immunol. 164:3645.-3651. . Annu. Rev. Immunol. 15:749.[Medline]
in establishing and maintaining the tumor-dormant state. J. Immunol. 162:2842.-dependent delay of in vivo tumor progression by Fas overexpression on murine renal cancer cells. J. Immunol. 164:231.This article has been cited by other articles:
|
|
 |
|
  T. Zhang, A. Barber, and C. L. Sentman Chimeric NKG2D Modified T Cells Inhibit Systemic T-Cell Lymphoma Growth in a Manner Involving Multiple Cytokines and Cytotoxic Pathways Cancer Res., November 15, 2007; 67(22): 11029 - 11036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Mathers, O. A. Tckacheva, B. M. Janelsins, W. J. Shufesky, A. E. Morelli, and A. T. Larregina In Vivo Signaling through the Neurokinin 1 Receptor Favors Transgene Expression by Langerhans Cells and Promotes the Generation of Th1- and Tc1-Biased Immune Responses J. Immunol., June 1, 2007; 178(11): 7006 - 7017. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Boissonnas, L. Fetler, I. S. Zeelenberg, S. Hugues, and S. Amigorena In vivo imaging of cytotoxic T cell infiltration and elimination of a solid tumor J. Exp. Med., February 19, 2007; 204(2): 345 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Hollenbaugh and R. W. Dutton IFN-{gamma} Regulates Donor CD8 T Cell Expansion, Migration, and Leads to Apoptosis of Cells of a Solid Tumor. J. Immunol., September 1, 2006; 177(5): 3004 - 3011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D McKenzie, N. L Dudek, L. Mariana, M. M. Chong, J. A Trapani, T. W. Kay, and H. E Thomas Perforin and Fas induced by IFN{gamma} and TNF{alpha} mediate beta cell death by OT-I CTL Int. Immunol., June 1, 2006; 18(6): 837 - 846. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Fadel, L. G. Cowell, S. Cao, D. A. Ozaki, T. B. Kepler, D. A. Steeber, and M. Sarzotti Neonate-primed CD8+ memory cells rival adult-primed memory cells in antigen-driven expansion and anti-viral protection Int. Immunol., February 1, 2006; 18(2): 249 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Olver, P. Groves, K. Buttigieg, E. S. Morris, M. L. Janas, A. Kelso, and N. Kienzle Tumor-Derived Interleukin-4 Reduces Tumor Clearance and Deviates the Cytokine and Granzyme Profile of Tumor-Induced CD8+ T Cells Cancer Res., January 1, 2006; 66(1): 571 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Gao, J. Tao, M. O. Li, D. Zhang, H. Chi, O. Henegariu, S. M. Kaech, R. J. Davis, R. A. Flavell, and Z. Yin JNK1 Is Essential for CD8+ T Cell-Mediated Tumor Immune Surveillance J. Immunol., November 1, 2005; 175(9): 5783 - 5789. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Song, X. Tang, K. M. Harms, and M. Croft OX40 and Bcl-xL Promote the Persistence of CD8 T Cells to Recall Tumor-Associated Antigen J. Immunol., September 15, 2005; 175(6): 3534 - 3541. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Kienzle, S. Olver, K. Buttigieg, P. Groves, M. L. Janas, A. Baz, and A. Kelso Progressive Differentiation and Commitment of CD8+ T Cells to a Poorly Cytolytic CD8low Phenotype in the Presence of IL-4 J. Immunol., February 15, 2005; 174(4): 2021 - 2029. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
|
